Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Tree species response to gap formation in partially-cut interior cedar-hemlock forests of northwestern… Coates, Kenneth David 1998

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

Item Metadata

Download

Media
831-ubc_1998-345440.pdf [ 8.36MB ]
Metadata
JSON: 831-1.0088978.json
JSON-LD: 831-1.0088978-ld.json
RDF/XML (Pretty): 831-1.0088978-rdf.xml
RDF/JSON: 831-1.0088978-rdf.json
Turtle: 831-1.0088978-turtle.txt
N-Triples: 831-1.0088978-rdf-ntriples.txt
Original Record: 831-1.0088978-source.json
Full Text
831-1.0088978-fulltext.txt
Citation
831-1.0088978.ris

Full Text

TREE SPECIES RESPONSE TO GAP FORMATION IN P A R T I A L L Y - C U T INTERIOR CEDAR - H E M L O C K FORESTS OF NORTHWESTERN BRITISH COLUMBIA. By K E N N E T H DAVID COATES B.S.F., The University of British Columbia, 1979 M.Sc , Oregon State University, 1988 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES Faculty of Forestry, Department of Forest Sciences We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A June 1998 © K. Dave Coates, 1998 I In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) 11 Abstract Partial cutting is a type of small- to intermediate-scale disturbance that affects forest community dynamics and ecosystem processes by creating canopy gaps of various sizes through the harvesting of trees. Studies of canopy gap dynamics have contributed significantly to our understanding of the role of disturbance in forests, but have been little used by foresters for predicting ecosystem response to partial cutting. The gap dynamics literature was reviewed for insights into ecosystem responses to partial cutting and a series of gap-based studies was undertaken in the interior cedar - hemlock forests of northwestern British Columbia to quantify: (1) environmental conditions after gap formation; (2) species-specific responses to light environments; (3) natural regeneration and growth of planted seedlings in different gap sizes and gap positions; (4) tree species composition and growth in 30-40 year old gaps; and (5) windthrow in partially cut forests. Light levels and soil temperature varied by gap size and gap position, increasing from south to north gap positions, but not dramatically so in larger gaps away from the shady south end. Soil temperature in gaps rose above 10° C early in the growing season and remained there until early autumn. Fifth year size and recent growth of five planted conifer species was quite similar from low to high light. Under field conditions, there was no meaningful variation in whole-plant light compensation points among the species. Low light performance of lodgepole pine was similar to or better than that of the other species. Growth rates at high light were broadly overlapping and there was considerable variability within species. Greatest variability among species was at I l l intermediate light levels (30-70% full sun) where careful matching of tree species to light environment can maximize growth rates. Natural regeneration was abundant in gaps but poor in clearcuts and the forest understory. Density of all species increased from the north to south end of gaps, while growth was best in the middle to northerly positions. There was no evidence of gap partitioning by species during the regeneration phase. Instead, gap partitioning appears to take place gradually through differential mortality in response to gap size and gap position. Partial cutting had little effect on wind damage to merchantable trees (> 17.5 cm diameter). In the partial cuts, 2.2% of the trees were damaged compared to 1.1% in unlogged areas. The greatest wind damage occurred in old-growth stands whether partially cut or not. For 8 of the 9 tree species examined, no individual tree characteristics seemed to predispose them to wind damage. Amabilis fir, trembling aspen and subalpine fir were the most susceptible to windthrow. iv Table of Contents Abstract ii List of Tables vi List of Figures viii Preface xii Acknowledgements ..xiii Introduction 1 Chapter 1 A gap-based approach for development of silvicultural systems to address ecosystem management objectives 5 1.1 Introduction 5 1.1.1 Background - traditional silvicultural systems 7 1.2 Methods 12 1.2.1 Gap dynamics review 12 1.2.2 Partial cutting experiment 12 1.2.3 Retrospective study 13 1.2.4 Light modelling 14 1.3 Results 15 1.3.1 Gaps and tree regeneration 15 1.3.2 Small-scale disturbance, biological diversity and ecosystem processes 22 1.3.3 Date Creek gap size and density in natural and manipulated stands 24 1.3.4 Date Creek retrospective gap study 26 1.3.5 Date Creek gap light environments 29 1.4 Discussion 29 1.4.1 Conclusions 33 Chapter 2 Microclimatic conditions after variable levels of canopy removal in the high latitude forests of northwestern British Columbia 35 2.1 Introduction 35 2.2 Methods 36 2.2.1 Photo synthetically active radiation (PAR) 37 2.2.2 Soil temperature 38 2.2.3 Air temperature and relative humidity 39 2.3 Results 39 2.3.1 Photosynthetically active radiation (PAR) 39 2.3.2 Soil temperature 40 2.3.3 Air temperature and humidity 52 2.4 Discussion 58 Chapter 3 Growth of planted tree seedlings in response to ambient light levels in northwestern interior cedar - hemlock forests of British Columbia 61 3.1 Introduction 61 3.2 Methods 63 3.2.1 Data analysis 66 3.3 Results 67 3.4 Discussion 75 3.4.1 Management implications 79 V Chapter 4 Gap size and position effects on natural and planted tree regeneration in northwestern interior cedar - hemlock forests of British Columbia 81 4.1 Introduction 81 4.2 Methods 84 4.2.1 Gap selection 85 4.2.2 Natural regeneration 86 4.2.3 Planted seedlings 86 4.2.4 Analysis 88 4.3 Results 91 4.3.1 Natural regeneration 91 4.3.2 Planted seedlings, regression analysis 101 4.3.3 Planted seedlings, analysis of variance 114 4.4 Discussion 134 4.4.1 Natural regeneration 134 4.4.2 Planted seedlings 136 4.4.3 Conclusions and management implications 139 Chapter 5 A retrospective analysis of the effects of gap size and gap position on tree regeneration in 34 to 41 year old partially logged northwestern interior cedar - hemlock forests 141 5.1 Introduction 141 5.2 Methods 143 5.2.1 Gap selection 143 5.2.2 Regeneration within gaps 144 5.2.3 Analysis 145 5.3 Results 146 5.3.1 Tree density by gap size and gap position 146 5.3.2 Tree size by gap size and gap position 147 5.3.3 Gap filler trees 150 5.4 Discussion 150 Chapter 6 Windthrow damage 2 years after partial cutting at the Date Creek silvicultural systems study in the Interior Cedar-Hemlock forests of northwestern British Columbia 155 6.1 Introduction 155 6.2 Methods 156 6.3 Results 159 6.4 Discussion 167 Conclusions 172 Bibliography 177 vi List of Tables Table 1.1. Traditional silvicultural systems (modified after Smith 1986; Matthews 1989; Klinka et al. 1990) 8 Table 1.2. Examples of ecosystem characteristics that are affected by gap attributes 18 Table 2.1. Soil heat sums at different depths in open conditions, in different positions in 1785 m and 205 m gaps and in the forest understory. Heat sums are the summation of degree-days based on the difference between the average daily soil temperature and a base temperature of 5°C 53 Table 3.1. Provenance descriptions,characteristics of trees at planting and sample sizes by canopy condition 65 Table 3.2. Parameter estimates for predicted fourth and fifth year height (cm) and diameter (mm) using the equation Y=((ah)/((a/s)+L)) where L is the light level 69 Table 3.3. Parameter estimates for predicted height (cm) and diameter (mm) growth rates (average, last 3-yr) using the equation Y=((a(L-c))/((a/s)+(L-c))) where L is light level 70 Table 4.1. A) Randomized block split-split-plot A N O V A p-values from natural regeneration study and B) Each tree species analyzed separately 98 Table 4.2. Parameter estimates for predicted fifth year height and diameter using the equation Y=((aL)/((a/s)+L)) where L is opening size (m ) or north-south line length (m)....106 Table 4.3. Parameter estimates for predicted height and diameter growth rates using the equation Y=((aL)/((a/s)+L)) where L is opening size (m ) or north-south line length (m) 107 Table 4.4. Randomized block split-split-plot A N O V A for planted seedling study 115 Table 4.5. Observed p-values from a randomized block split-plot A N O V A on seedling size. .116 Table 4.6. Observed p-values from a randomized block split-plot A N O V A on seedling growth rates 117 Table 4.7. Percent of seedlings that have died (mortality) in the forest understory, in small, medium and large gaps, and in the open condition of clearcuts 123 Vll Table 5.1. Randomized block split-plot A N O V A p-values from retrospective study of tree density (stems/m2) in old logging created gaps in northwestern interior cedar-hemlock forests 147 Table 6.1. The extent of wind damage in the Date Creek silvicultural systems study two years after logging (trees > 17.5 cm) 160 Table 6.2. A N O V A table for the wind damage response variables in a randomized block design 161 Table 6.3. Number of stems and basal area per hectare before (1991) and after (1992) partial cutting and the extent of wind damage two years after cutting (1994) for trees > 17.5 cm diameter 163 Table 6.4. Windthrow distribution by tree species, combined across all experimental blocks and treatment units 164 Table 6.5. Tree characteristics from prism cruise plots established immediately after logging and from the windthrown trees, combined across all experimental blocks and treatment units 165 vm List of Figures Figure 1.1. Comparison of gap density (a) and area in gaps (b) in natural forests and light (30%) and heavy (60%) removal partial cutting treatments at the Date Creek silvicultural systems study (ICHmc subzone of northwestern British Columbia...25 Figure 1.2. Density of western hemlock, western redcedar, hybrid spruce, subalpine fir, lodgepole pine, and paper birch trees in gaps created by old partial cutting 27 Figure 1.3. Percent of full open photosynthetic photon flux densities received in gaps between April 15 and September 15 at the latitude of the Date Creek silvicultural system study 28 Figure 2.1. Total daily photosynthetically active radiation (PAR) in full open conditions (17-ha clearcut) during the 1995 and 1996 growing seasons 41 Figure 2.2. Percent of daily full sunlight measured at 3 positions inside a 4700 m 2 gap ^ Percent of daily full sunlight measured at 6 positions inside a 1785 m gap 43 Figure 2.4. Percent of daily full sunlight measured at 4 positions inside a 600 m 2 gap and in the forest understory 20 m to the south of the gap 44 Figure 2.5. Percent of daily full sunlight measured at 6 positions inside a 205 m 2 gap 45 Figure 2.6. Average daily soil temperature by gap position and soil depth (5, 10, 20 and 50 cm) in a 205 m 2 gap 46 Figure 2.7. Average daily soil temperature by gap position and soil depth (5, 10, 20 and 50 cm) in a 1785 m 2 gap 48 Figure 2.8. Daily average, maximum and minimum soil temperature in a 17-ha opening at the Date Creek study area 50 Figure 2.9. Average, maximum, and minimum daily soil temperature in the forest understory 51 Figure 2.10. Average daily soil temperature in the forest understory, in full open (clearcut) conditions and in north and south positions of a 1785 m gap 54 Figure 2.11. Average daily soil temperature in the forest understory, in full open (clearcut) conditions and in northern position of a 205 m (small) and 1785 m" (large) gap..55 Figure 2.12. Air temperature at 1.4 m height in 205 m 2 and 1785 m 2 gaps 56 ix Figure 2.13. Average, maximum and minimum relative humidity at 1.4 m above the ground in 205 m 2 (small) and 1785 m 2 (large) gaps 57 Figure 3.1. Observed values and fitted regression lines of seedling height 5 years after planting vs. percent of full sun for 5 tree species using using Equation [1] and parameter values in Table 3.2 71 Figure 3.2. Observed values and fitted regression lines of seedling diameter 5 years after planting vs. percent of full sun for 5 tree species using Equation [1] and parameter values in Table 3.2 72 Figure 3.3. Observed values and fitted regression lines of height growth rates (cm, average per year for last 3-yrs) vs. percent of full sun for 5 tree species using using Equation [2] and parameter values in Table 3.3 73 Figure 3.4. Observed values and fitted regression lines of diameter growth rates (mm, average per year last 3-yrs) vs. percent of full sun for 5 tree species using using Equation [2] and parameter values in Table 3.3 74 Figure 3.5. Predicted fifth year height (a), height growth rates (b), fifth year diameter (c) and diameter growth rates (d) in response to ambient light levels 76 Figure 4.1. Layout of plots for natural regeneration study 87 Figure 4.2. Average density of natural regeneration across all experimental blocks by opening type and year of assessment 92 Figure 4.3. Average density of natural regeneration in mature experimental blocks by opening type and year of assessment 94 Figure 4.4. Average density of natural regeneration in the old-growth experimental block by opening type and year of assessment 96 Figure 4.5. Average density of natural regeneration by gap size class and gap position 99 Figure 4.6. Mean seedling height five years after planting vs. gap size for the five tree species. The fitted regression line use Equation [1] and parameter values in Table 4.2 102 Figure 4.7. Mean seedling diameter five years after planting vs. gap size for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.2 103 Figure 4.8. Mean seedling height growth rates (average per year for last 3-yrs) vs. gap size for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3 104 Figure 4.9. Mean seedling diameter growth rates (average per year for last 3-yrs) vs. gap size for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3 105 Figure 4.10. Mean seedling height five years after planting vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.2 108 Figure 4.11. Mean seedling diameter five years after planting vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.2 109 Figure 4.12. Mean seedling height growth rates (average per year for last 3-yrs) vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3 110 Figure 4.13. Mean seedling diameter growth rates (average per year for last 3-yrs) vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3 111 Figure 4.14. Predicted fifth year height (cm) as a function of (a) opening size (m2) and (b) north-south line length (m); and predicted fifth year diameter (mm) as a function of (c) opening size (m2), and (d) north-south line length (m), using Equation [1] and parameter values in Table 4.2 112 Figure 4.15. Predicted height growth rate (cm, average per year for the last 3-yrs) as a function of (a) opening size (m2) and (b) north-south line length (m); and predicted diameter growth rate (mm, average per year for the last 3-yrs) as a function of (c) opening size (m2), and (d) north-south line length (m), using Equation [1] and parameter values in Table 4.3 113 Figure 4.16. Mean seedling seedling height (cm) five years after planting by opening type 118 Figure 4.17. Mean seedling diameter (mm) five years after planting by opening type 119 Figure 4.18. Height growth rate (cm, average per year for last 3-yrs) by species and opening type. .120 Figure 4.19. Diameter growth rate (mm, average per year for last 3-yrs) by species and opening type 121 Figure 4.20. Mean seedling height five years after planting by gap size class and gap position 126 Figure 4.21. Mean seedling diameter five years after planting by gap size class and gap position 127 xi Figure 4.22. Height growth rate (cm, average per year for last 3-yrs) by gap size class and gap position 128 Figure 4.23. Diameter growth rate (mm, average per year for last 3-yrs) by gap size class and gap position 129 Figure 4.24. Mean seedling seedling height (cm) five years after planting by species and opening type 130 Figure 4.25. Mean seedling diameter (mm) five years after planting by species and opening type 131 Figure 4.26. Mean seedling height growth (cm, average per year for last 3-yrs) five years after planting by species and opening type 132 Figure 4.27. Mean seedling diameter growth (mm, average per year for last 3-yrs) five years after planting by species and opening type 133 Figure 5.1. Tree density by opening type and species 148 Figure 5.2. Tree density by gap size class and gap 149 Figure 5.3. Height (m) and diameter (cm) of trees most likely to be gap fillers in old partially logged areas by gap size class and gap position 151 Figure 6.1. Frequency distribution of the direction in which windthrown trees fell. Height (m) and diameter (cm) of trees most likely to be gap fillers in old partially logged areas by gap size class and gap position 168 Acknowledgements I thank my employer, the British Columbia Forest Service, for giving me the opportunity to undertake my Ph.D studies at UBC. In particular, I would like to thank Larry Pedersen, Jim Snetsinger and Tony Buckley for their continued support. Many people were involved in the installation, maintenance and measurement of the studies reported here. Thanks to Paula Bartemucci, Saleem Dar, Jody Friesen, Heidi Harlacher, Cameron Heibert, Duncan Moss, Les Priest and Jennifer Penny. P. Bartemucci and Russell Klassen assisted with data synthesis. Wendy Bergurd and Peter Ott provided advice on statistical analysis. These studies were partially supported by the silvicultural systems program of the Ministry of Forests, Silviculture Practices Branch, Victoria, B.C. My committee members (Phil Burton, Gordon Weetman, Ken Lertzman and Andy Black) provided valuable direction and advice throughout, even thought I spent most of my time a long way from U B C . I am especially grateful to Phil Burton who has helped me, a silviculturist, develop a much broader understanding of forest ecology. Phil was so dedicated to this cause that he moved to Smithers to keep a close eye on me. Alan White, the external examiner, provided a thorough review and thoughtful comments that were very helpful. It is difficult to describe how much help someone like Sybille Haeussler, my wife, has been during this period and during my career as a forester - thanks. Thanks as well to Andrea and Sonja, my daughters, during the 'dad never plays with us anymore' period of the last few months. 1 Introduction Silviculture is an applied science that uses knowledge of forest stand development to produce desired forest attributes and products. The discipline of silviculture has strong traditions, rooted in European forest practices (see Smith 1986; Matthews 1989; Weetman 1996). Silvicultural practice is based on theory, research and long-term observations of forest responses to traditional or trial-and-error manipulations (Weetman 1996). Until recently, the science of silviculture has focused primarily on tree- or tree-related issues such as reproduction methods, provenance testing, genetics, growth and yield prediction, protection from pests and wildfire and development of planting, tending and harvesting techniques. Today, however, society expects more from forests than just timber. Forest managers are often asked to achieve what are perceived as conflicting objectives: removal of forest products while maintaining the structure and diversity of an intact forest. Forestry research can help with experiments that lead to improved understanding of forest ecosystem response to disturbance, along with better prediction of the consequences and trade-offs involved in the use of different cutting prescriptions. The greatest challenge for the disciplines of silviculture and forest ecology is to understand how community structure in forests develops over time following disturbance. Many studies search for generalities in community ecology (e.g., Ricklefs 1977; Cherrett 1989), yet few have emerged. Another approach is to ask more focused questions about species-specific responses under certain conditions or across a range of conditions, and not look for broad generalities that may have only weak predictive powers. There is no substitute for sound replicated comparisons of tree community response to different types of disturbance. This is especially the case following human disturbances in forests. 2 Our understanding of the implications of varying frequency, intensity and pattern of tree removal on forest community dynamics and ecosystem processes is limited. I believe it is necessary to undertake an integrated program of empirical studies that characterize species-specific responses to changing canopy conditions in order to develop reliable predictions of the effects of management on long-term forest dynamics. Such empirical studies when linked with forest simulation models (e.g., Pacala et al. 1996; Kimrnins et al. 1997), offer the best chance of providing insight into forest pattern at different spatial scales and time periods. Silviculturists manipulate stand structure and are in a position to predict the short- and long-term consequences of forest disturbance on a host of forest values. The traditional silvicultural systems (Matthews 1989) have served silviculturists well in the past, when the emphasis was on timber production, but lack the needed flexibility, imagination and innovation that is required to manage forests today (Weetman 1996; Kohm and Franklin 1997). The broad objective of this thesis was to undertake empirical research that will lead to a better understanding of how small- to intermediate-scale disturbance to the tree canopy (partial cutting) affects tree community dynamics. Specific attention was paid to the regeneration and growth of tree species after disturbance and to species-specific seedling responses to changing canopy conditions. All experiments were undertaken in the Interior Cedar - Hemlock (ICH) forests of northwestern British Columbia and form the core of the Date Creek silvicultural systems study (Coates et al. 1997). The ICH forests are dominated by a mixture of seven conifer and three hardwood tree species (Banner et al. 1983). Chapter 1 examines the value of a gap-based approach for understanding ecosystem response to tree cutting. It reviews the gap dynamics literature, paying special attention to papers that used gap size or position inside gaps as predictive variables for responses indicative of 3 silvicultural success or maintenance of ecosystem function. It then proposes a gap-based approach for studying stand response to silvicultural manipulation. Chapter 2 examines the effect of canopy removal on resource availability and physical conditions from the open conditions of a clearcut, to gaps of different sizes and different positions inside those gaps, to the forest understory. The objective is to see if there is a strong gradient in resource availability, specifically light, that could be utilized differentially by co-existing species (e.g., Leibold 1995; Bazzaz 1996) or if physical conditions along the tree removal gradient from forest understory to clearcut changes enough to differentially affect species success. In Chapter 3, tree seedling height and radial growth response to variation in ambient light levels are determined for the most commonly planted tree species in the ICH forests. Understanding the response of individual trees to variation in light is fundamental to predicting the effects of partial cutting on future forest composition and growth. Under field conditions, the research addresses how the magnitude of seedling response to light and the shapes of the light response curves vary among the individual tree species and if there are clear trade-offs among tree species in aboveground growth at low vs. high light. Clearly, other resources and physical conditions will vary with ambient light level, but these are more difficult to quantify and to manipulate. My research will answer how good closely correlated ambient light levels are with whole-tree growth in the field and identify which species are best suited to different gap sizes or positions inside or outside gaps based on light response characteristics. Chapter 4 more closely examines the influence of gap size and position inside or outside of a gap on establishment of natural regeneration and performance of planted trees. For the high latitude ICH forests, this study will allow me to address several major hypotheses formulated to explain tree species diversity in tropical forests (e.g., Ricklefs 1977; Denslow 1980; Brokaw 1985, 1987, Whitmore 1989). Do densities of individual naturally regenerating tree species differ significantly by gap size or position within a gap and, if so, are results tied to a species' shade tolerance ranking? Do growth rates of shade tolerant and shade intolerant species vary by gap size and gap position? Is the regeneration niche "concordant" or "discordant" (Schupp 1995), that is, is the best location for germination and early establishment also the best for growth? If gap partitioning is occurring in ICH forests is the regeneration niche or the growth niche most important? Retrospective studies can provide rapid and useful answers to many management questions. In Chapter 5,1 look at gaps created by partial logging 34 to 41 years ago to see how tree species composition, density, and growth vary by gap size and gap position and if the results are consistent with those of the designed experiment at Date Creek (Chapter 4). Many of the questions and hypotheses addressed in Chapter 4 can be examined in these old retrospective gaps. Windthrow along with pest damage is often considered a major deterrent to the use of partial cutting systems, and can frustrate the best of intentions. Chapter 6 examines the amount of windthrow damage that occurred in the undisturbed forest and the two partial cutting treatments at Date Creek and tests the hypothesis that greater wind damage occurs in partially cut units than uncut areas. In a concluding section, I provide some broad recommendations on approaches to forest management in the ICH forests. I further show how the studies reported here tie together with the spatially-explicit model SORTIE (Pacala et al. 1996) that is being calibrated for use in predicting community dynamics in these northern forests (Kobe and Coates 1997; Wright et al. 1998; LePage et al, submitted). 5 Chapter 1 A gap-based approach for development of silvicultural systems to address ecosystem management objectives1. 1.1 Introduction "When a tree falls by itself it is for a good cause; a spirit has struck it, struck it through the elderberry shrubs. Do not be afraid, if it crashes by your side. All dead trees must fall. Falling they yield an opening for the spider-like feet of the sun, that the pale flowers underneath may turn into ripe fruit." -the 1888 exile lament of Kamalmuk (Barbeau 1928; pg 126) Kamalmuk, a Gitxsan native from northwestern British Columbia, understood one role of canopy gaps well before Sernander (1936), Jones (1945), and Watt (1947) first formally recognized the importance of gaps in ecological regeneration and succession. A natural gap is formed by the death or fall of large branches, an individual tree, or a group of trees that results in a canopy opening, usually quantified in terms of projected land area (m2). Many researchers have studied ecosystem processes in canopy gaps or described natural forests by their gap size distribution. Just as a natural forest has a distribution of gap sizes, a managed forest subject to harvesting and silvicultural intervention can have a gap size distribution, though this has rarely been quantified. Foresters have tended to quantify response to silvicultural manipulations at the 1 Coates, K.D. and Burton, P.J. 1997. A gap-based approach for development of silvicultural systems to address ecosystem management objectives. Forest Ecology and Management. 99(3):337-354. 6 scale of the stand (a somewhat arbitrary unit of forest, relatively homogeneous in site, composition, and age structure throughout), often without regard for the fine-scale variation within a stand. Society's desire for a broader range of forest management options, combined with emerging concepts in the ecosystem and landscape ecology of both natural and managed forests, are causing a reassessment of forestry practices (Harris 1984; Aplet et al. 1992; Hagner 1992; Thompson and Welsh 1993). Foresters world-wide are being called upon to implement a wider variety of stand management systems. To meet new silvicultural, ecological and social management objectives, silvicultural systems must evolve beyond their traditional emphasis on timber production to include the broader objectives of protecting sensitive species, sustaining ecosystem functioning (diversity, productivity, nutrient cycling and resilience) and identifying sustainable levels of use for a broad range of renewable resources. These goals collectively define the emerging field of ecosystem management (Slocombe 1993; Galindo-Leal and Bunnell 1995; Christensen et al. 1996). Sustainable ecosystem management should be seen as the collection of protocols and actions that allow us to deliver the essential goods and services of ecosystems in perpetuity (Christensen et al. 1996). The maintenance of structurally and compositionally diverse forests is a major objective of ecosystem management. Organisms and ecosystem processes respond to forest structure manipulation at many spatial and temporal scales. We will explore why the application of traditional silvicultural systems are too narrowly focused to meet ecosystem management objectives. We then review the ecological literature for examples of how a functional understanding of the role of small-scale, low intensity disturbance can aid in developing silviculture systems for ecosystem management. We present preliminary results from a silvicultural systems experiment designed 7 to examine ecosystem processes at multiple scales, and we suggest a framework for implementing a gap-based approach for developing improved silvicultural systems. 1.1.1 Background - traditional silvicultural systems Silviculture is an applied science that uses knowledge of forest stand development to produce desired forest attributes and products. The growing of timber has been the most common objective. The science of silviculture has focused primarily on tree- or tree-related issues such as reproduction methods, provenance testing, genetics, growth and yield prediction, protection from pests and wildfire and development of planting, tending and harvesting techniques. Over the past century or so, foresters have manipulated stands primarily by implementing one or more of the traditional silvicultural systems (Smith 1986; Matthews 1989; Table 1.1). A silvicultural system is a cycle of activities by which a forest stand is harvested, regenerated and tended over time. The names of silvicultural systems reflect the type of forest structure remaining after the initial harvest and the type of reproduction method employed. They have been developed and applied in western and central Europe for several hundred years, and have been implemented in forested regions throughout the world. The greatest strengths of these traditional silvicultural systems are immediate recognition by foresters and longevity of use. They have provided a common communication framework for foresters, and they encompass the complete range of stand manipulations commonly applied. Development and refinement of systems has been based largely on long-term observations of forest responses to traditional or trial-and-error manipulations, as articulated by Troup (1928): "The detailed study of these systems under as many different conditions as possible is the only means of acquiring that special knowledge which will lead to their intelligent application in practice." oo o ON ON a a B eg u CM o rea (ha « cn CW "3 C _u '5 'B. 0> >. a H o Si ^ A A 2 o o V •a V o CD cd C •i—l a ON" 00 ON -O "9 -7! 8. J3 8. o •tf o OO ON c 8. 8. S 8. <U o S a a o a 00 C-H D cd O a. •C — -a cd C O -a cd E o C/J cd S-i 3 3 O o u s o. 8. 8. o s; 8. •c 8. cd H S3 0 O U 9 Silvicultural systems are generally viewed as resulting in the development of either even-aged (clearcut, seed-tree, shelterwood, or coppice methods) or uneven-aged (group- or single-tree selection methods) forest stands (Table 1.1). Alternately, they can be characterized by the degree of canopy influence during the various stages of stand development. Clearcut systems have minimal canopy influence throughout establishment and growth. Seedtree systems are typified by canopy seedfall but minimal protection for establishing regeneration. Shelterwood systems have canopy protection during establishment and early growth of new seedlings. Selection systems are characterized by a dominance of canopy influences throughout the life of the stand. Another way of classifying silvicultural systems is according to the distribution of canopy trees after harvest. Silvicultural systems then fall into two broad groups: uniform applications where canopy trees are evenly distributed after harvest (uniform shelterwood, seedtree, single-tree selection) and patch applications where discrete gaps are created in the tree canopy (group, irregular, strip or wedge shelterwoods and group-selection). Bradshaw (1992) argues that large patch cuts (clearcuts) are at one end of a continuum of the degree of canopy influence with single-tree selection at the other end. It follows that the measurement of gap size and gap dispersion are ways of describing the location of any silvicultural system along the canopy influence continuum. Interestingly, when silvicultural systems are viewed this way the distinction between even-aged and uneven-aged categories becomes blurred. In our opinion, the traditional silvicultural systems have four major limitations for meeting ecosystem management objectives: First, silvicultural systems are, unfortunately, often thought of as a prescribed program of treatments to be applied without thought and adjustment. The systems are usually considered to be locally "prescriptive" (i.e., choosing from a fixed set of possible treatments), rather than 10 globally "descriptive" (i.e., encompassing an unlimited set of possible treatments). Application procedures are often very rigid. Second, foresters typically describe silvicultural systems and the stands they produce by tree-centred attributes such as the method of reproduction, residual basal area, density, diameter distribution or a q-factor and maximum tree diameter retained. These descriptors are useful for timber management, but provide an incomplete framework for understanding how organisms and ecosystem processes are affected by tree removal. Third, traditional silvicultural systems are not process-based. They are empirically-based and provide only very broad predictive powers. The systems are usually just a description of tree reproduction methods and desired stand structure. The fact that a silvicultural system is well known, and the methodology for its implementation well described, does not mean it can be successfully applied on any given site. Silvicultural systems do not, in and of themselves, impart any understanding of ecological processes that would allow their refinement and extension. For example, while it is accepted knowledge that single-tree selection favours regeneration of shade-tolerant over shade-intolerant tree species, the relative success of colonists of each species and their subsequent performance can not be predicted simply by knowing that a single-tree selection method was used. Debate concerning the virtues of different silvicultural systems, especially even-aged versus uneven-aged management, and the conditions under which they can be applied has been ongoing and without any real resolution (Jones 1945; Bradshaw 1992). The long-term nature of silvicultural systems and the difficulty of conducting research on what are fundamentally descriptive management systems is partly to blame. Finally, both "uniform" and "patch" applications of silvicultural systems assume and prescribe stand homogeneity in order to optimize and predict tree growth. In fact, spatial and 11 environmental heterogeneity exists within stands after silvicultural manipulation and may be highly desirable from an ecosystem management perspective. Heterogeneity cannot, however, be predicted based on silvicultural system alone. Methods for both describing and prescribing this within-stand variability are required. Smith (1993) suggests that silvicultural systems themselves are not researchable treatments, but rather that the component parts of the systems should instead be studied. Successful prediction of the effect of a silvicultural manipulation must be based on an understanding of the spatial and temporal dynamics of forest response to different kinds, sizes, frequencies and intensities of disturbance. We think the most important component parts of a silvicultural system that affect stand heterogeneity following tree cutting are: the number (density) of canopy gaps created; the size frequency distribution of the gaps; the longevity of the gaps before filling or expansion by further cutting; and the density, spatial distribution and temporal extent of the retained forest matrix. We propose that the explicit consideration of within-stand variability caused by canopy gaps is an important key to understanding these processes and to designing innovative stand management systems. The functional role of gaps will be greatest in patch applications of silvicultural systems and the forest matrix becomes of primary importance in uniform applications. Gap dynamics have been the focus of considerable ecological study and warrant broader consideration in forest management. 12 1.2 Methods 1.2.1 Gap dynamics review We reviewed the gap dynamics literature paying special attention to papers that use gap size or position within a gap as predictive variables for responses indicative of silvicultural success or maintenance of ecosystem function. We searched for relevant papers and reports in both the ecological and forestry literature. Particular attention was paid to parallels between ecological and silvicultural disturbances and the potential for extending ecological intrepretations to the improvement of silvicultural practices. 1.2.2 Partial cutting experiment The Date Creek silvicultural systems study (Coates et al. 1997) was established in 1992, in northwestern British Columbia, Canada, approximately 21 km north of Hazelton, west of the Kispiox River (55°22'N, 127°50'W; 370-665 m elevation). Date Creek is within the moist cold subzone of the Interior Cedar-Hemlock biogeoclimatic zone (ICHmc; Banner et al. 1993), a transitional area between the interior and coastal forests of northwestern British Columbia (Pojar et al. 1987, Meidinger and Pojar 1991). Mature forests (140 yr) in the area are mixtures of conifer and deciduous tree species. Western hemlock (Tsuga heterophylla Raf. Sarg.) dominates; other species include western redcedar (Thuja plicata (Donn. ex D. Don), subalpine fir (Abies lasiocarpa (Hook.) Nutt.), lodgepole pine (Pinus contorta var latifolia Engelm.), hybrid spruce [the complex of white spruce (Picea glauca (Moench) Voss), Sitka spruce (P. sitchensis (Bong.) Carr.) and occasionally Engelmann spruce (P. engelmannii Parry ex Engelm.)], paper birch (Betula papyrifera Marsh.), trembling aspen (Populus tremuloides Michx.), and black Cottonwood (Populus balsamifera ssp. trichocarpa Torr. & Gray). Old-13 growth forests (250-300+ yr) are dominated by western hemlock with varying amounts of western redcedar and some amabilis fir ( A b i e s a m a b i l i s Dougl. ex Forbes). Amabilis fir abundance increases with elevation. Four tree removal treatments (in approximately 20 ha treatment units) were replicated four times (16 units in total), in a randomized block design, with combinations of ecological site type and forest age as the blocking factor (mesic 140 yrs; mesic-submesic 140 yrs; mesic-subhygric 140 yrs; mesic 350 yrs). The treatments involved increasing levels of tree removal: (1) no tree removal; (2) light tree removal where approximately 30% of the stand volume was removed as single stems and small groups (similar to a single-tree or group-selection, or a light initial shelterwood removal; Table 1.1); (3) heavy tree removal where approximately 60% of the stand volume was removed using a combination of small patch cuts (1000-5000 m^) and single-tree to small group-selection within the forest matrix between the patch cuts (similar overall to an irregular shelterwood system; Table 1.1); and (4) total removal of all merchantable trees or clearcut. Across the three treatments where trees were retained, a gradient of gap sizes was created from very small (removal of a single tree) to 5000 m^. We determined gap size, gap density and total gap area within all treatment units except the clearcut. Transect lines were established 50 m apart (approximately 3600 m of line) on maps of each treatment unit. From these lines, approximately 2000 m of transect line was randomly selected. The total length of line in canopy gap and forest matrix was recorded. The size of each canopy gap encountered was determined using methods described by Runkle (1992). 1.2.3 R e t r o s p e c t i v e s t u d y A companion retrospective study was established in old partially cut areas of the ICHmc subzone to determine tree species composition and density within gaps of a similar size range to those created at the Date Creek silvicultural systems study. Five stands, partially logged between 1954 and 1961, were identified from local knowledge, forest cover maps and aerial photographs. Within each stand, we identified all logging-created gaps which had good canopy closure around the entire perimeter of the gap. A total of 86 gaps were found and gap size was determined using the methods of Runkle (1992). From this population 46 gaps were randomly selected (20 less than 90 m^, 21 between 90 and 600 m^ and 5 gaps greater than 600 m^). Sample plots within gaps were established systematically so that various gap positions (i.e., north, south, east, west edges; the northern, middle and southern thirds) were represented in equal numbers of plots for detailed sampling of species composition and density of trees greater than 10 cm tall. We randomly selected 15 of the 46 gaps for destructive sampling to determine tree age. A l l trees taller than 1 m were destructively sampled in gaps with less than 100 trees. In larger gaps, the 10 tallest trees per species occurring in the northern, middle and southern thirds of each gap were selected. Sample sizes were unbalanced because not all species occurred in all gaps and gap positions. In total, 381 western hemlock, 89 western redcedar, 53 hybrid spruce, 33 subalpine fir, 10 paper birch and 1 lodgepole pine were destructively sampled. 1.2.4 Light modelling To model light availability in gaps at the latitude of the Date Creek silvicultural systems experiment, we estimated the integrated photosynthetic photon flux densities received in gaps using the model CLiMo, version 2.0 (Chen et al. 1993). This model considers direct beam and diffuse radiation penetrating gaps after partial removals of overstory. Light is expressed as percentage of above canopy light and may be estimated for any period from one day up to a full year. The value is determined by integration over time, at half-hour time steps, so that the 15 effects of diurnal and seasonal changes in the solar azimuth and elevation angles are taken into account. No account is taken of local cloudiness in the current model. We modelled gaps from 5 m diameter (19.6 m2) to 150 m diameter (17,671.5 m.2) in 5 m increments. A l l modelled gaps were assumed to be circular, on flat ground and to have 30 m tall canopy trees surrounding the gap. Light was estimated for a growing season of April 15 to September 15. 1.3 Results 1.3.1 Gaps and tree regeneration Forests with a wide range of gap sizes offer diverse habitat conditions for the regeneration of trees. Habitat conditions vary among gaps, within gaps, at the gap/closed canopy interface (gap edge), and within the forest stand matrix. Most gap studies of forest regeneration have focused on population or community-level responses within gaps by examining (1) differences in microclimate as gap size, shape, or orientation change, (2) differences in the nature and abundance of specific substrates available for colonization or (3) the effects of root and crown influence zones. The gap edge and stand matrix have received less study in terms of assessing their suitability for tree establishment and growth. Lertzman (1992) has summarized the major alternative hypotheses concerning the relationship between gap environments and the tree species that colonize gaps: (1) species differ in the size of gap in which they are most successful, leading to different species prevailing in gaps of different sizes (Denslow 1980, 1987; Brokaw 1985, 1987; Whitmore 1989); (2) the most successful species will be those that are most abundant surrounding the gap (Sousa 1984); (3) certain species will be most successful in particular locations within gaps (i.e., centre vs. periphery; Orians 1982; Putz 1983; Brokaw 1985; Beatty and Stone 1986); or (4) certain species 16 will be most successful on particular substrates (Putz 1983; Cristy and Mack 1984; Lawton and Putz 1988). The germination, survival and growth of shade-tolerant and shade-intolerant tree species has often been correlated with gap size (e.g., Brokaw 1985; Whitmore 1989; Stewart et al. 1991). In a review of tropical forest gap studies, Denslow and Hartshorn (1994) conclude that variation in light availability is the pre-eminent factor affecting tree seedling establishment and early growth. Light conditions at ground level throughout forest stands are directly related to gap size, shape, canopy height, and latitude (Canham et al. 1990) and since light availability can vary so sharply over short distances within gaps, position within a gap is critically important to the physiology and growth of any individual tree seedling (Wayne and Bazzaz 1993; Sipe and Bazzaz 1995). Light availability is clearly important, but other factors also play a role in regeneration success. Putz (1983) observed that shade-intolerant tree species were more common on mineral soil and concluded that since large treefall gaps are often caused by uprooted trees, the abundance of shade-intolerant tree species in large gaps could be partially explained by the higher incidence of exposed mineral soil. Several other studies have shown different tree species to be concentrated on specific substrate types within gaps (Lawton and Putz 1988; Peterson and Pickett 1990; Stewart et al. 1991; Lertzman 1992; Hofgaard 1993). Gap dynamics can be quite different in areas where advance regeneration (seedlings and saplings established prior to gap formation) comprise most of the gap colonizing species. Uhl et al. (1988) observed only minor new colonization and high subsequent mortality in recently formed Amazonian treefall gaps; advance regeneration fared much better. In the mixed Nothofagus forests of New Zealand, Stewart et al. (1991) found no relationship between pattern 17 of regeneration and gap size. They also conclude that most gap occupants were present prior to gap formation and that little subsequent establishment occurred. In both studies, gap size influenced growth rates of established individuals. For Nothofagus (Stewart et al. 1991), gap size influenced the competitive relations between the two dominant species, with the more light-demanding species gaining an advantage in gaps >400 m^. In situations where advance regeneration dominates gaps, the most important factor affecting tree species diversity and initial stem density within the gap is the relative ability of species to regenerate under a closed canopy. Gap size and the provision of suitable microclimate at any particular location within the gap affect subsequent opportunities for each individual to become a canopy tree (Canham 1988a). We infer from the gap regeneration literature that the size-class distribution of gaps within a forested area after the first harvesting entry and the proportion of each size class characterized by a given type of microenvironment directly affects colonization by individuals, their persistence and growth and the number of species in an area. The subsequent frequency of gap creation or gap expansion by silvicultural intervention will probably play a more important role than initial gap size for many species (particularly the more shade tolerant species; e.g., Canham 1985, 1990). Future timing of silvicultural treatments will have an important influence on population dynamics. Another lesson to be drawn from gap ecology studies is the importance of determining the origin of regeneration found within gaps. When gaps are dominated by colonists established from seed after gap formation, studies that examine the factors that influence successful establishment, survival, and growth of new seedlings (e.g. gap size, gap microclimate, substrate diversity; Table 1.2) are needed. When advance regeneration is abundant, regeneration (i.e. germination and emergence) studies can be restricted to closed canopy conditions (e.g., Harmon T3 CD CU O e o 4-1 cu cu T3 0 3 S "o 1 .s re b CD oj CT) oi c 0 T3 i— re CQ LO co CT3 oo 0) CD > 0. CD CO 3 C L c re c: o o CT3 CT) T- ~ N ca oo cn CO oo cn CT> i -re tr" CC g CD _ CO 3 TJ C CO CO a. Q. O G5 CD 03 CD CL CD CT) CT) T— CT) C 1_ ri CT) CT) CM 7 6 yttebc T— CT) CD o CT) 0 yttebc "ro CT) c 1 X CD ro 1 bo TJ c art CD aylc ytte re ZJ Stew; arik h- X L J Stew; CD co CT3 05 CM c > 03 J= — ^ TJ cz cz ro ro E >-N O <JJ * = > < - £ ? 0 ro 0 J h L t ro CO t : 0 cu " 00 (50 c o c/5 00 ± -V 1 3 s-i w cd C & «» cd _r oo >, oo X) CJ\ T3 w CU -*-> O cd cu & -i—< cd 43 -i—» c/3 O 1/1 fi CU Q oo ON N cd N N e£ PQ (U ° C c/) cd JZ N N cd cd X C Q <U <U 3 '> cd U H Pi SI ° D. 0 ro CD CD < c B ca UJ ro ro h t: ro 3 •§! ° E ° t •£ < § g. CL H = CO < W ^ O o .b! CD co 0 c_> CO  9-0 ro CO CD 0 M— C L ° J=! O o oo s r o >> cu CU OH CO DC UJ H o < DC < X o UJ H CO >-CO o o UJ + + > + + + + + + ) co cz ro 'sz o ro at co 2 E Q. CD E £ O =3 o o > , w * _ +-T CO co c 9? 0 o S S 8 O = TJ m i l -a.<2 « w o 8? « > h- 0 O < CO c i 0 to ro CO E ro cz ro CO Q_ 0 p CO * " 0 T3 c ro ro 0 l> ro aj ro -a. CO 0 < CO I 8 0 ro s °. p to CO 0 £ o H - O 9 T3 ^ E E S t u5 O o o o ro ro =i; o ro bz o o SZ i - o F o. o. H ? » - 5 . O r— r— CD O I— < « re —' 0 c 3 0 CD "> f, 0 0 g > CO «" 0) 0 O o c CO "c re o re o E Q < c i 0 o CO CO T3 CZ re CO o Q . CO co re -= 0 re Q. Q-< + + + + + + + O + + < CO 3 -a B OJ CO 1 s C £ O co U 0 <o E5 (J) o £ a. co <o J5 " I ro CO ro < 0 CO 0 0 g b co >-*— £ ° CL H -0 3 ^ « 1 0 CO 0 ^: Sz .ti O CO CO CO T3 ro ro 0 N g 0 ro E _2 o O ro sz -a .!2 JE -c - CO •o p « H — m 0 ro c ro •<= c c ~ 0 - < CO c?5 i2 0 CO 0 . O X I ^ = o c 0 § ° H •£ ro o 2 ^ o o CO S cz _ o ro o 9 0 o. tz 2 Q. V re JD X I CO CO CD CD w o .2 £ -Q O ON c g ro O 0 SI 3 CO w Q . CD CO O ) D5 < tz •2 o. ro ro zz CT CD co OJ & 0 0 59 co CT) CD CO cz co C\J > 0 0 -n ro X. tz ro > T J CZ CO ^ CD CD 0 cz 3 CO V— o o CD u b_ T J CZ ro 03 cz o OJ CJ) . . T — CO 3 25 CD 22 — co 0 5 CD £ 52 CD J tn o o a> CJ) cz CD CD CO CO LU T J SZ CO cz o CO CO CZ o o O) OJ CD OJ CJJ CO co T J ]T ro ra CO "CD _cz CO 0_ CO. m O) CJ) CD . Q D . E ro O CJ) OJ - r j ^ CZ : : CO CO CO c 0 0 g o o jr, N N £ zz .— Q) Q O C L CM CM CJJ CJJ CO O) 5 co co oi j £ - r -2 CO CO CD TJ CL <° I 00 CD T-I CO c cn 2 £ ra - °> 0 CD w .cz > co o ra cj o co co _i Z CO co O) CO T J ir T J CZ CO CD E o O • r j -CJ) CJJ CO S? ra cn co oo CJ) co co OJ CZ o CO 1 CO X T J CZ CO °8 CO > CO Q • X CD = >- CO CD ZJ 0 SZ -SZ >-co co es CZ g co o 0_ CL CO O J S ra .y CT co ) + + + i + CD o c co CL CD CO CO CJ) CD H -LX ° 1 + 1 + + + + ^ + + + + + + + + + (0 <u T J CD ZJ .cz o CM 0 SD CO Q. O n T J C co CO .Q CD SZ . 0 co o CD CD CZ ~ CL tn co o 1 E -e (fi in sz CD cj cz CO T J CO c CD ZJ n SD § 8 ra — CO >r CO „ 0 O c ^ CD & i « T J > £Z D . CZ ° ° m ra - Q SD -SZZ 3 i _ — i _ 0 CD 3 SZ sz sz CO iZ CO CD CD CD > CO T J sD CT CD p £•= 13 CL _ E 2 co 2 CO - Q CO CO CO CD o CD O -co SD i _ CD _ T J CO CD CO CO CD c 3 E O H _ i - CO o 0 CD H SD CZ i — i — 3 0 0 CZ SZ SZ T J cz g -t—' CO o O -E O CD O 0 CJ O 0 CO 1/5 CZ D . 3 CO 0 CO CO J/J 0 T J CO C C T J ' > •— CJ - Q 3 O CQ ' C CO CT CO _co (0 E E CO E CO 0 CJ 0 _ CL to " J E 5^ CO SD 0 O c ro T J c 3 X ) ^ ra CJ T J £ tD s ^ ° ^ ra , « 1 0 £ ^ to «2 ra o co 0 o -o > 0 1 S -§ >J .a c ro O 0 •*—> a. 2 0 3 to CD £ 00 co CT) co H 03 Qj o O CJ) 03 03 03 T -CD _- T- 4-: 3 ro Q - co N -I—> rj CL m CL CO CD i 0 o c XJ sr ro CO g. 0 O XJ c ro cz o XI cz ro CO CL CO CD 03 LO 1- 03 03 CM CM 03 CD (33 03 CO CO _ l 0_ CD CD O N CO • -OQ Q N CO N N CO CQ XJ tr ro 0 cr >^  CO N CO N N CO CQ XJ c CO CD Q. CO 00 co CO CO CD XI c ro CO tz o o O O 00 r-» CD 0 T — CD LO c 03 CD 0 T — CD (0 CD i _ T — LO c LO CM zzaz CM 00 CO 00 ZJ zzaz CD Sw, CD CD zzaz 03 Sw, •J— CO CD zzaz 03 Sw, XJ cz Ba ro .nd ro CO ro XJ ro CD CO CD -»—» CD CZ CO CD CO CD 0 rik ipe ^ nn XI UU JZ ro ipe CO Z! 0 ZJ C L CQ CO CQ CQ 0 CQ 00 00 03 co CD •s ro Q- ^ CD C\J O C3 " — CD • « CD c -75 O CZ X) CZ ro CO CD "o ro O CQ C L CD CO CD CD < g '•+—• O -ro CO tz CD CD "•— 0 O cz cz 0 Q . CO CO CD 0 C L ., O C L CD CO N CD co + + + + + + ^ + + + + + + + + 1 + + + + I 1 + + ! + 1 + + + I I CD o CO c 2-CD CO CO 03 CD H _ CL ° + + < + - - ( + + + + + + I + XJ CD ZJ c '•4—• cz o CM JD CO < CO 3 < CO cr i £ to ro 0 to CD OJ XJ " O m ° •S 5 XJ XJ X CO CO CO 0 if R t5 g 0 . y o = — CO ro E > . H ° cz 0 ro -c o CD Q . 0 o CO co J= co co" O XJ O c . ro CO 0 CO ZJ O J Z .2 0 X XI 0 ro -•!< to ^ CO ro ^  X) 0 > ro 0 XJ CO o o a co E <= co cn co ^ co c co 0 0 0 -»— T z 2 Oco .9 T3 C L 0 O .>< J3 _ CO 0 ro o C L o J= X < I- I-ro *r CL J2 C L ZJ < co 0 > 0 ro o ro o •5. 0 xi 2 8- ro I— LU X 0 3 0 _ U « E '< £ x co E ZJ E ro XI X) cz c ro ro 0 0 E E 1 x XJ c 5 o c CO 0 l_ 3 CO o E o in 0 u co M— 3 CO CN cz o CD Q. 2 SI ZJ co ° CL CD CO cn co < c B o_ CO CO 22 o S O ! LO 0 5 O) CO ^ T- CO £ p =3 CM x: cn CO 0 3 T J tz CO CO tz o O CO CL N CO N N CO CQ T J cz CO CD CL CO CO CO CD CD CO oo CJJ CJJ T — T — g O o CO CO c cz CD CD CD Q Q oo CO T J T J O) CO CZ cz 1— •»— CO CO M  : "o CO CD 0 tz CO CO 0 CD ZJ ZJ T J o o CO Uh > > oo oo CJ) ^ o ti CJ) 0 CJ) T -o __. b_ T J CJ) CJJ CJJ CJJ CO ^ - CM ^ CJJ — CJJ CO T-3 z iz « a 0 - jg CO cz o o a 5 *-6 0 5 0 CO Jo 2 CL 0 _ ^ CO CO co CJ) CJJ 0 o CJJ O Q_ 00 CJJ CO O) T J T - I " C . . CO 1 I - - CO "2 •S E o o C cz c o CO CO ^ cj) cj) it cz g co o 0. CL co O) + — l ) + ! D J CO 1 + 1 + + + ^ + 0 o CO c 9-0 co CO CO 0 M -ct ° ! + + > 1 i + + + + CZ o T J 0 tz o CM 0 x i co y-3 E 2 | a> . E a. cz M * * cz = ra o 0 co E X X CO CO E g > ^ race 5 CO CO T J QJ 0 E " Xi 05 '5 > CO c CZ 0 O c o o CO ZJ i _ o x: CL CO O T J CZ co cz jo •£ 3 C E E 0 I •5 | ^ CO c Z ° S c N o g O V N 9 V a nterr sqn z CO shr M — O cz CO 0 CO O •*—> T J on ure p 0 CO T J CO CO 'tz T J N 0 o CO E 'cz 0 0 CO CO N c ZJ CO o » o '•E E van' cz 0 z van' c o zz w o E o o 0 JJ CO (0 0 c o CO CO CO E o o o CD CD and Franklin 1989) and subsequent growth can be examined over a range of gap sizes and various positions within gaps to determine how the advance regeneration responds to changes in resource availability (e.g., Pacala et al. 1994). The latter also applies to coppice regeneration. Contrasting the success of natural regeneration and planting over the range of gap attributes provides an opportunity to distinguish the "regeneration niche" (Grubb 1977) from the "growth niche". 1.3.2 Small-scale disturbance, biological diversity and ecosystem processes Maintenance of biological diversity has become an important priority for foresters (Society of American Foresters 1991; Burton et al. 1992). A central tenet of ecosystem management is the maintenance and enhancement of indigenous biological diversity across appropriate scales of space and time (Slocombe 1993). In lowland, evergreen tropical forests, results of gap studies suggest that tree species diversity is promoted by frequent small disturbances that provide a diversity of regeneration niches (Brokaw 1987; Lawton and Putz 1988; Primack and Hall 1992). In forests subject to more frequent catastrophic disturbance (such as wildfires, hurricanes or floods), maximum tree species diversity may be maintained at intermediate levels of disturbance (cf. Connell 1978). In many forested regions of the world, including those historically dominated by stand-destroying disturbance events, societal pressure to implement partial cutting practices (such as the selection methods; Table 1) is increasing (Bradshaw 1992; Kimmins 1993). Evidence is growing that most forest types can be successfully managed by "great cycles" of stand-level disturbance and even-aged stand development, or by "lesser cycles" of gap-level disturbance and uneven-aged stand development. Recent studies (Hytteborn et al. 1991; Hofgaard 1993; Kuuluvainen 1994) suggest that old-growth phases of boreal forests, 23 while rare because of short fire-return intervals, are able to regenerate and maintain themselves through gap dynamics. Forest management must increasingly interpret multiple resource values in terms of stand structural objectives. Table 2 summarizes results of tests for gap presence, size, orientation and age, position within gap and gap substrate effects on diverse elements of the forest ecosystem. Establishment, growth and species diversity of bryophytes and vascular plants have been examined in gaps (e.g., Ehrenfeld 1980; Denslow et al. 1990; Jonsson and Esseen 1990; Phillips and Shure 1990; Reader and Bricker 1991; Dirzo et al. 1992; Reader et. al 1995). Forest gaps also influence plantherbivore relationships (Harrison 1987; Shure and Wilson 1993). Invertebrate faunal abundance and species richness, especially that of butterflies, can be enhanced in forest stands with open areas (Greatorex-Davies et al. 1994). Canopy gaps can also have a profound effect on the structure of vertebrate communities, according to the differential abilities of species to use gaps as opposed to closed canopy areas. For example, Crome and Richards (1988) were able to classify species of bats as either closed canopy specialists, gap specialists or gap incorporators. Similarly, gap and closed canopy forest bird assemblages can be quite distinct, thus treefall gaps can influence the composition and richness of bird communities (Schemske and Brokaw 1981; Blake and Hoppes 1986; Levey 1988; Noss 1991). Ecosystem processes such as hydrological fluxes and nutrient cycling have also been related to forest canopy gaps. Nutrient relations and fine root biomass accumulation have been studied in natural and artificial canopy gaps (Vitousek and Denslow 1986; Mladenoff 1987; Sanford 1989, 1990; Marrs et al. 1991; Spies 1991; Parsons et al. 1994). The results of these studies, summarized in Table 2, have not been conclusive, but as Parsons et al. (1994) indicate, few of the studies considered the effect of gap size. Parsons et al. (1994) created different sized 24 gaps in Rocky Mountain coniferous forests and were able to show an effect of gap size on the level of dissolved nitrogen in ground water. The heterogeneity of canopy influences and microsite conditions, both within gaps and within the surrounding forest matrix, has often been ignored by gap researchers. In particular, many gap studies and modelling efforts have neglected the effects of gaps on ecological processes within the forest matrix around them. The influence of canopy openings on the understory microclimate extends beyond the gap edge, especially at high latitudes where sun angles are low. Thus it may be just as important to document the effect canopy gaps have on the surrounding forest matrix as within the gap itself. Gap size, position within a gap or distance from a gap edge may all be just as important as the gap/non-gap distinction in microsite location. It is really the point-specific irradiance input or (conversely) canopy closure (Lieberman et al. 1989) that is most predictive of tree growth, plant phenology, nitrogen mineralization, suitability as insect habitat and myriad other ecological processes. 1.3.3 Date Creek gap size and density in natural and manipulated stands Figure 1 contrasts gap density and area in gaps of different sizes among undisturbed, light removal and heavy removal treatment units at the Date Creek silvicultural systems study. The majority of gaps in natural stands are <150 m^ with a few up to 600 m^. After light removal logging, in which individual trees or small groups of trees were removed, gap density has increased over the natural forest but the size distribution is still similar to that found in the natural forest; few gaps exceeded 600 m^ in size (Fig. 1.1a). Gap density increased further after heavy removal logging, but the distribution of gaps among size classes <600 m.2 remained similar to that in the natural forest and after light removal logging. In addition, heavy removal 25 18 16 0) 14 iS 12 O 0) JZ 10 OJ CL W O. (0 (3 • Natural forest • Light removal (30%) • Heavy removal (60%) <75 2401-5000 Gap size (m ) <75 75-150 151-300 301- 601-600 1200 .2x 1201- 2401-2400 5000 Gap size (m ) Figure 1.1 Comparison of gap density (a) and area in gaps (b) in natural forests and light (30%) and heavy (60%) removal partial cutting treatments at the Date Creek silvicultural systems study (ICHmc subzone of northwestern British Columbia). Values represent averages from four replicates of each treatment. Treatment units varied in size from 17.4 to 25.2 ha. Area in gaps was: natural forest = 1.45 ha (6.6%); light removal = 2.83 ha (14.4%); heavy removal = 7.49 ha (35.5%). 26 logging extended the maximum gap size well beyond that found in the natural forest or after light removal logging. Within each size class, the total area occupied by gaps increased steadily from the natural forest through to heavy removal logging (Fig. 1.1b). This illustrates that partial cutting can create gap size distributions customised to match or extend those found under natural conditions and that considerable overlap in the distribution of gap sizes (and hence understory conditions) among different partial cutting systems occurs. 1.3.4 Date Creek retrospective gap study Tree regeneration in old logging origin gaps varied as gap size increased and reflected their relative shade tolerance (Kobe and Coates 1997). Figure 1.2 shows stem density and species composition across the range of gap size classes presented in Figure 1.1. At the time of sampling, gap ages were 34-41 yr. Four of the eight tree species commonly found in the ICHmc subzone occurred consistently in the gaps: western hemlock, western redcedar, subalpine fir and hybrid spruce. Trembling aspen and black Cottonwood, two of the most shade intolerant species in the ICHmc subzone, were not found in any of the gaps. Lodgepole pine and paper birch (also shade intolerant species) occurred rarely and mainly in the larger gap sizes (Fig. 1.2b). Of the six tree species found in the gaps, only western hemlock and western redcedar (shade tolerant) were well represented in all gap size classes (Fig. 1.2). Hemlock was by far the most abundant tree species and steadily increased in density as gap size increased (Fig. 1.2a). Western redcedar was well represented in all gap size classes in roughly similar densities as size increased (Fig. 1.2b). Hybrid spruce (moderately shade-tolerant) was most abundant in gaps >1200 (Fig. 1.2b). Spruce density was about double that of redcedar in the two larger gap size classes. Subalpine fir (shade tolerant) occurred at low densities in most gap sizes (Fig. 1.2b). Tree 27 70000 - • hemlock a • redcedar 60000 - • spruce OJ Hsubalpine fir co 50000 - Spine heel • birch heel 40000 -terns per 30000 -20000 -CO 10000 • I I ~L n m "•_ * 0 -<75 75- 151- 301- 601- 1201- 2401-150 300 600 1200 2400 5000 Gap size (m2) 7000 6000 a> CO 5000 •*—I o CD SZ 4000 _^ a> ° - 3000 co E a> 2000 CO 1000 H o EL <75 75-150 151-300 301-600 601-1200 1201-2400 2401-5000 Gap size (m ) Figure 1.2. Density of western hemlock, western redcedar, hybrid spruce, subalpine fir, lodgepole pine, and paper birch trees in gaps created by old partial cutting. Logging was between 1954 and 1961. A l l trees are of natural origin. Graph (a) includes all 6 species, graph (b) excludes hemlock so to better display the distribution of the other 5 species; note difference in stems per hectare scale between the two graphs. Shade tolerance ranking as determined by Kobe and Coates (1997) for the ICHmc is as follows: western redcedar > western hemlock = subalpine fir > hybrid spruce > lodgepole pine > trembling aspen > black cottonwood = paper birch. 28 o -I 1 1—'•—i 1 f 1 1 1 1 1 1 1 1 1 1 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Gap diameter (m) 0 0.01 0.03 0.07 0.13 0.19 0.28 0.38 0.50 0.64 0.79 0.95 1.13 1.33 1.54 1.71 Gap area (ha) Figure 1.3. Percent of full open photosynthetic photon flux densities received in gaps between April 15 and September 15 at the latitude of the Date Creek silvicultural system study. Gaps vary in size from 5 m diameter (0.002 ha) to 150 m diameter (1.77 ha), with a canopy height of 30 m. Photosynthetic photon flux densities were estimated using the model CLiMo, version 2.0 (Chen et al. 1993). species diversity (presence and abundance) increased as gap size increased. These retrospective data clearly show that tree species presence and density are strongly affected by gap size. Most of the trees found in the gaps (81%) established after the logging disturbance. Western hemlock, western redcedar and subalpine fir were the only species with advance regeneration in the gaps. A l l hybrid spruce, lodgepole pine and paper birch established after 29 logging. The percent of hemlock, redcedar and subalpine fir that established after logging was 85, 69 and 60, respectively. 1.3.5 Date Creek gap light environments Figure 1.3 shows the estimated proportion of gap area in each of four different light classes in gaps up to 150 m diameter. Light classes represent the percent of full open photosynthetic active radiation (PAR) received between April 15 and September 15 (the typical growing season in the ICHmc subzone). Gaps must be >50 m in diameter before any gap position receives >75% of open PAR. Gaps <25 m diameter only receive 50% full light or less; those <10 m diameter only receive 25% full PAR or less. At approximately 40 m diameter, half of the gap area will receive light at 50% of open PAR or greater. Levels of PAR begin to stabilize in gaps greater than about 110-120 m diameter. In 110 m diameter gaps about 65% of the area receives PAR greater than 75% full sunlight and that amount of area slowly increases as gap size increases; the gap area receiving <50% full PAR stabilizes around 10%. Note that the maximum diversity of light conditions is found in gaps of 30 to 75 m diameter, or one to 2.5 times the height of the tree canopy. 1.4 Discussion Gap dynamics studies have contributed significantly to our understanding of forest successional patterns and of community and population dynamics, and should be of considerable value to foresters designing silvicultural systems to meet the objectives of ecosystem management. However, we believe that foresters are, for the most part, unfamiliar with the ecological gap dynamics literature, its overlap with silvicultural systems and its application to ecosystem management. The knowledge and hypotheses generated, and the methods used in gap studies 30 can be used by foresters as models for predicting and testing the effects of partial cutting systems on ecosystem processes, and for prescribing novel silvicultural systems. Gap size and position in gap, in particular, have considerable effect on a number of microclimate and biological processes. Our surveys of canopy gaps in undisturbed forests of northwestern British Columbia, and in adjacent treatment blocks subject to different partial cutting prescriptions (Fig. 1.1), confirms that quantifying the gap size distribution is a useful and practical means of describing the internal heterogeneity of a forest stand. Each size of gap has a characteristic combination of light regimes (Fig. 1.3), which are correlated with a number of other microclimate attributes which collectively determine vegetation responses. This was confirmed in the retrospective analysis of old logging gaps, where the density of naturally regenerating western hemlock was positively correlated with gap size (Fig. 1.2a). Likewise, paper birch was found only in gaps >2400 m2, and high densities of hybrid spruce were restricted to gaps >1200 m^ (Fig. 1.2b). Ongoing studies are interpreting the mechanisms of radiation, water balance and nutrient dynamics associated with different gap sizes and gap positions, and their effects on tree seed germination, seedling mortality, tree growth, vegetation development and animal activity (Coates et al. 1997). Disturbance is fundamental to the development of structure, composition and functioning of forest ecosystems (Oliver 1981; Attiwill 1994). Disturbance in forested ecosystems varies spatially and temporally from frequent small-scale, low intensity gap forming disturbances operating at the individual tree scale to infrequent stand- or landscape-scale, high intensity events that can significantly alter single or multiple stands. Both small- and large-scale natural disturbances are common in British Columbia forests depending on location and stand age. 31 Small-scale, low intensity disturbance is common in very wet maritime old-growth forests of coastal British Columbia (Lertzman et al. 1996), in high elevation forests with long fire return intervals, and in low elevation interior sub-boreal spruce stands that have escaped catastrophic fire disturbance for prolonged time periods (Kneeshaw 1992, Clark 1994). Thus, small-scale, low intensity disturbance is found under two dominant conditions: in climatic zones where large-scale disturbances are rare and in dispersed areas that have escaped catastrophic disturbance. A l l forests eventually undergo gap dynamics (small-scale, low intensity disturbance) if they escape large-scale disturbance. Foresters often describe their management practices as emulating the historic natural disturbance regime. In British Columbia, foresters have tended to emulate catastrophic large-scale disturbance by using clearcutting as the dominant silvicultural system in all forested landscapes. As more emphasis is placed on partial cutting systems in order to meet broader management objectives, foresters need to understand and predict the prevalence and consequences of small-scale disturbance in forest stands and landscapes. To illustrate, when the primary management objective is to retain mature or old-growth forest structure and function after tree harvest, a logical first step should be to create opening sizes that fall within the range found following gap dynamics in natural older forests. However, to regenerate specific tree species or achieve economically acceptable growth rates, foresters may then ask how far it is possible to extend the size distribution of small-scale, low intensity disturbance while still maintaining mature or old forest structure and function. Conversely, in partially cut areas where timber production is the primary management objective, it will be important to understand how increasing opening sizes to optimize tree growth affects other organisms and ecosystem processes. We think a gap-based approach to developing silvicultural 32 systems addresses these issues. Results from the Date Creek retrospective study clearly show that tree species presence and abundance in the ICHmc subzone are tied to gap size. Given the quickly changing light environment in gaps as size increases (Fig. 1.3), tree growth will also vary considerably in different gap sizes and positions in gaps. The tree species found in the ICHmc subzone will increase their growth rates over a wide range of light levels (Carter and Klinka 1992; Klinka et al. 1992; Kayahara et al. 1996; Chapter 3). Most trees found in logged ICHmc subzone gaps established after gap formation, indicating that tree establishment studies in logged gaps will provide useful insights for forest management. A gap-based approach for study of stand response to silvicultural manipulation has three major applications for ecosystem management: first, to aid development of cutting prescriptions that maintain functional mature or old growth conditions in stands where small-scale, low intensity disturbance historically dominated; second, to refine and extend our understanding of how biological structures, organisms and ecosystem processes are affected by fine-scale variation within stands; and third, to develop novel silvicultural systems that meet timber production objectives without compromising ecosystem management principles. Our first approximation of a gap-based approach is: 1. Determine the extent and character of gap processes in the landscape relative to other disturbance regimes. 2. For stands displaying small-scale, low intensity disturbance, determine whether these are in the latter stages of a large-scale, high intensity cycle or whether small-scale, low intensity disturbance is the norm. Determine if these stands develop by patchy gap creation or by uniform species turnover. 33 3. Characterize gap size frequencies and density within small-scale, low intensity disturbance stands. 4. Devise management systems that emulate the historic frequency of gap sizes and density in small-scale disturbance stands. This approach is most appropriate for areas where the management objective is to maintain mature or old-growth forest structure and function while allowing some timber extraction. 5. Establish experiments that evaluate how important ecosystem processes, species colonization, growth rates and diversity responses are affected as gap size and distribution are manipulated. Results will help determine how far outside the historic size distribution of small-scale, low intensity disturbance it is advisable to go in areas where the objective is to maintain mature or old-growth forest function. 6. Extend studies to gap sizes and distributions well beyond those found under historic small-scale disturbance. This will be of great assistance to foresters trying to predict tree composition and growth in patch application cutting systems and aid in the development of ecosystem based and predictive silvicultural systems. 7. In addition to designed experiments, use retrospective study of old logging or natural disturbances to quantify the effect of gap size, position in gap or position within the forest matrix on organisms, biological structures or ecosystem processes of interest. 1.4.1 Conclusions Treefall gaps, whether natural or man-made, are an essential component of stand and landscape heterogeneity. They are a logical level of organization within which to study many of the important ecological processes that underlie stand level response after partial cutting, especially patch applications where distinct canopy openings are created. The distribution of gap sizes, 34 their dispersion and dynamics can explain much of the silvicultural and mensurational behaviour of a stand. Studies that examine the response of forest flora, fauna and processes within gaps (of varying size and configuration) and within the adjacent forest matrix can form the basis for developing predictive silvicultural systems for ecosystem management. Examination of ecosystem responses over an array of gap sizes, gap positions, and levels of retained forest matrix set in forest stands of different ages can address all possible degrees of canopy influence over the course of stand development. Using canopy gaps as a framework thus provides a mechanism for the experimental study of forest stand response to a full range of silvicultural systems, as defined primarily by the degree of canopy retention associated with the first harvesting entry. This framework is presented in the hope of stimulating the widespread establishment of such experiments, and the generation of testable hypotheses that will improve our ability to generalize how forest stands respond to silvicultural manipulations. More refined studies of the response of different organisms and ecological processes to gradients of canopy influence, and better predictions of the spatio-temporal extent of canopy influences in forest stands, hold the most promise for better predicting and managing forest ecosystem attributes in the future. 35 Chapter 2 Microclimatic conditions after variable levels of canopy removal in the high latitude forests of northwestern British Columbia. 2.1 Introduction Small-scale gap-forming disturbance alters resource availability and physical conditions in forest understories (Minckler et al. 1973; Collins and Pickett 1988; Canham et al. 1990; Barik et al. 1992; Parsons et al. 1992). In low latitude forests, especially tropical forests, the environmental heterogeneity in microclimatic conditions found in different sized gaps is thought to play a role in maintaining tree species diversity (Ricklefs 1977; Lieberman et al. 1995). In high latitude, northern forests where sun angles are much lower in the sky, optimal microclimates for different species should occur at different positions inside gaps as well as across a gradient of gap sizes (e.g., Canham et al. 1990). This, in turn, should affect natural regeneration success and growth of established seedlings (Bazzaz 1996). The majority of research on tree response to gap formation, however, has focused on structural attributes of forests with differing small-scale disturbance regimes and response to gap size (e.g., Runkle 1985; see review in Chapter 1). Plants, of course, do not respond directly to structural variables. They respond to changes in resource availability after gap formation by natural disturbance or logging (Bazzaz 1983, 1996). These resources may be light, soil moisture or soil nutrients. The flux and availability of light affects air and soil temperature and humidity levels, which in turn affect moisture and nutrient flux and availability. It is important to quantify how resource patterns and environmental conditions are altered by gap formation in order to 36 understand and predict tree responses to gap size or to specific positions inside or outside of a gap-Trie intent of this study was to see how photosynthetically active radiation (PAR), air temperature, soil temperature and humidity are affected by different levels of canopy removal in the northwestern interior cedar-hemlock forests of British Columbia. Measurements were made in the open conditions of a clearcut, in gaps of varying size, at different positions inside gaps (gap position) and in the forest understory. The research was designed to answer three questions. (1) Is there a strong microclimatic gradient along the north-south axis of high latitude forest gaps? (2) How is microclimate affected as gap size increases? (3) How do PAR and soil temperature vary from the forest understory, to gaps, to the open conditions of a clearcut? These measurements additionally serve to document the abiotic environment and treatment conditions affecting the growth of seedlings (Chapter 4). 2.2 Methods This study is a component of the Date Creek silvicultural systems study (Coates et al. 1997), established in 1992, in northwestern British Columbia, Canada, approximately 21 km north of Hazelton, west of the Kispiox River (55°22'N, 127°50'W; 370-665 m elevation). The area is within the moist cold subzone of the Interior Cedar-Hemlock zone (ICHmc), a transitional zone between the interior and coastal areas of northwestern British Columbia (Banner et al. 1993). Microclimatic data were collected in the full open conditions of a clearcut (at least 3 canopy tree heights from the forest edge), at different positions in gaps of varying size (205, 600, 1785 and 4700 m2) and in the forest understory during 1995 and 1996. Not all microclimatic variables were monitored in all stand conditions. A l l sampling occurred in mature (140 yr old) forests of wildfire origin. Tree species composition, structure and development pattern of these mixed-species forests are thoroughly described in LePage (1995) and Coates et al. (1997). Sample sites were generally flat or gently sloping and of average moisture and nutrient conditions for the ICHmc. 2.2.1 Photosynthetically active radiation (PAR) LI-COR spot quantum sensors (LI-COR Inc., Lincoln, Nebraska, USA) were used to monitor PAR levels. One sensor was monitored in the clearcut during both years. In 1995, 6 sensors were monitored at different positions inside each of two gaps (205 and 1785 m ). Sensors were placed along the north-south axis of both gaps and on the east and west edges. In the small gap (with a north-south axis of 18 m), sensors were placed at the gap centre, halfway between the centre and south end of the gap and 1 m from the canopy edge at each of the northern, southern, eastern, and western sides of the gap. Placement was the same in the larger gap (48 m north-south axis), except the north, south, east, and west side sensors were placed 2 m from the canopy edge. In 1996, sensors were placed in 600 and 4700 m" gaps. In the 600 m 2 gap (33 m north-south axis), sensors were at the gap centre, 4.5 m from the north and south ends, 3 m from the west canopy edge and 6 m into the forest understory off the south end of the gap. The forest understory sensor was placed under closed canopy conditions. In the 4700 m 2 gap (92 m north-south axis), sensors were 5, 20, and 40 m from the south end of the gap. Sensors were attached to data loggers (Campbell Scientific CR-10 or 21-X). PAR levels were recorded every 10 seconds and from these values, hourly and daily totals were calculated and saved by the data logger. A l l sensors used in this study were sent to LI-COR 38 (Lincoln, Nebraska, USA) for calibration before the 1995 field season and again at the end of the 1995 field season before use in 1996. Data from the full open conditions of the clearcut are presented in moles m"2 day"1 (PAR units). A l l PAR data from gaps and the forest understory are presented as percent of PAR recorded in the clearcut (and called percent of full sunlight). This provides a better comparison between years of measurement (1995 and 1996) and with the hemispherical photograph method of estimating percent of full sun in Chapter 3. 2.2.2 Soil temperature Soil temperature was monitored at 5, 10, 20 and 50 cm depths. Temperature profiles were established in open conditions, at three positions in the 205 m gap (at gap centre and 1 m from the north and south edges), and at 4 positions in the 1785 m gap (gap centre and 2 m from the north and south edges and half way between the centre and southern end of the gap). A fifth temperature profile was established 20 m into the forest off the north end of the large gap where no tree removal had occurred, to represent forest understory conditions. Omega EXPP-T-20S copper/constantan thermocouple wire was used at all locations. At each depth, the wire was coiled several times before insertion into the soil to minimize heat transfer from the surface. Sensors were placed 30 cm horizontally into undisturbed soil, at each depth, from the side of the pit excavated for the profile. The pit was then carefully filled and the forest floor layer replaced. Sensors were attached to data loggers (Campbell Scientific CR-10 or 21-X), with a scanning interval of 10 min during the snow-free period and a scanning interval of 1 hr in the winter. Daily minimum, average and maximum values were calculated and recorded. In the large gap, the data logger malfunctioned during the winter of 1995/96 and all data were lost. 39 Growing season heat sums were calculated for each soil temperature profile in 1995 and 1996. Heat sums were the summation of degree-days based on the difference between the average daily soil temperature and a base temperature of 5° C. In both years, there were short periods of missing data. Regression equations were developed to predict values for the periods of missing data, using relationships with another sensor from the same profile. 2.2.3 Air temperature and relative humidity Air temperature sensors were placed at two heights (0.5 and 1.4 m) in the centre of the 205 and 1785 m gaps. Omega TT-T-24SLE copper/constantan thermocouple wire was used to monitor air temperature. Temperature probes were shielded by two layers of sheet metal (20 cm x 20 cm) above the sensor (top surface painted white) and one layer of sheet metal below. Approximately 3 cm separated the two lower shields with the probe placed in-between. Relative humidity was also monitored at the gap centre (1.4 m height) using a shielded Campbell Scientific 207 temperature and relative humidity probe. Scanning interval was every 2 minutes, and daily minimum, average and maximum values were calculated and recorded. 2.3 Results 2.3.1 Photo synthetically active radiation (PAR) PAR levels were higher in the open conditions of the clearcut (Fig. 2.1) than at any position in the gaps. In 1996, PAR levels in the clearcut were consistently lower than those recorded in 1995, reflecting a cloudier summer and, quite possibly, a newly calibrated sensor that recorded slightly lower PAR levels even on clear sunny days. Even in the 4700 m gap, light levels in the 40 middle gap position were only 60-80% of those recorded in the clearcut (Fig. 2.2), and were similar to those recorded in the middle position of the 1785 m 2 gap a year earlier (Fig. 2.3). Light levels inside the 1785 m 2 gap varied little from the northern to the middle and middle-south position, and then decreased sharply at the south position (Fig. 2.3). The east and west positions were very similar, but lower than the north, middle and middle-south positions. Light levels were lowest inside the two smallest gaps (600 and 205 m2). There was very little differences in percent of full sunlight among the gap positions, except at the north end of the gaps where levels were higher (Figs. 2.4 and 2.5). There was a slight trend of increasing sunlight from the south to the middle-south to the middle positions. The north position received direct sunlight consistently throughout the growing season, whereas the other positions were dominated by diffuse light. There was little difference between the east and west positions. Light levels in the forest understory 6 m south of the 600 m 2 gap were about half of those recorded in the south end of the gap. 2.3.2 Soil temperature In both years, average soil temperature at each gap position in the 205 and 1785 m gaps followed the same general trend with soil depth (Figs. 2.6 and 2.7). There was little difference between the 5 and 10 cm depths, followed by consistently lower temperatures at 20 and then 50 cm depths. In the small gap, average soil temperature was a little lower at each gap position than that observed in the larger gap. Average soil temperature in the clearcut was slightly higher than in gaps at all soil depths (Fig. 2.8). As expected, soil temperature was coolest in the forest understory (Fig 2.9). 70 n , o -| 1 1 1 1 1 1 — 16-Apr 11-May 05-Jun 30-Jun 25-Jul 19-Aug 14-Sep 1996 Figure 2.1. Total daily photosynthetically active radiation in full open conditions (17-ha clearcut) during the 1995 and 1996 growing seasons. 42 100 80 60 40 20 South 26-Apr 16-May 05-Jun 1996 25-Jun 100 80 60 40 20 Mid-south 26-Apr 16-May 05-Jun 1996 25-Jun 16-May 05-Jun 1996 25-Jun Figure 2.2. Percent of daily full sunlight measured at 3 positions inside a 4700 m 2 gap. North 100 80 H 60 40 20 0 02-May 01-Jul 30-Aug 29-Oct 1995 tz ZJ CO ~B South 100 80 60 40 20 0 02-May 01-Jul 30-Aug 1995 29-Oct 100 80 60 40 20 A Middle o 02-May 01-Jul 30-Aug 1995 29-Oct sz CO 02-May 01-Jul 30-Aug 1995 29-Oct 100 80 -60 -40 20 East 02-May 01-Jul 30-Aug 1995 29-Oct c zs CO 100 80 -60 -40 20 -I West 02-May 01-Jul 30-Aug 1995 29-Oct Figure 2.3. Percent of total daily sunlight measured at 6 positions inside a 1785 m gap 100 80 60 40 20 North 26-Apr 05-Jun 15-Jul 24-Aug 1996 100 80 60 40 4. 20 0 Middle 26-Apr 05-Jun 15-Jul 24-Aug 1996 c Z i in ~3 sz Z i CO 100 80 60 40 20 0 South 26-Apr 05-Jun 15-Jul 24-Aug 1996 100 80 60 — 40 20 South 0 26-Apr 05-Jun 15-Jul 24-Aug 1996 100 80 60 40 20 0 Forest 26-Apr 05-Jun 15-Jul 24-Aug 1996 Figure 2.4. Percent of daily full sunlight measured at 4 positions inside a 600 m 2 gap and in forest understory 20 m to the south of the gap. 100 80 c w 60 1 40 20 North o 02-May 01-Jul 30-Aug 29-Oct 1995 CO "5 \ 0 100 80 60 40 20 South 0 02-May 01-Jul 30-Aug 29-Oct 1995 100 80 c w 60 -5 40-20 0 Middle 02-May 01-Jul 30-Aug 29-Oct 1995 100 -1 80 -cz w 60 Mid-south 40 20 0 02-May 01-Jul 30-Aug 29-Oct 1995 100 80 cz m 60 5 40 20 0 East 02-May 01-Jul 30-Aug 29-Oct 1995 100 -i 80 cz co 60 20 0 West 02-May 01-Jul 30-Aug 29-Oct 1995 Figure 2.5. Percent of total daily sunlight measured at 6 positions inside a 205 m 2 gap. 18 -i 10cm O o CD CD Q . E CD -i—• O co ~] 1 1 1 1 — r 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22 -Sep 1995 1996 Figure 2.6. Average daily soil temperature by gap position and soil depth (5, 10, 20 and 50 in a 205 m 2 gap. 18 20 cm -2 -I 1 1 1 1 1 1 1 — 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 gure 2.6 continued. o o CD zs co i _ 0 C L E 3 o co O o 4i 12 H -I—' pd CD C L E 0 "o CO 20 n 5 cm 4? 12 H 8 H 4 H 0 17-Apr 20 n 16 A 17-Apr North Centre Centre-south South 16-Jun 15-Aug 14-Oct 1996 10 cm 16-Jun 15-Aug 14-Oct 1996 Figure 2.7. Average daily soil temperature by gap position and soil depth (5, 10, 20 and 50 in a 1785 m 2 gap. o o 3 12 H CD i _ ZJ 03 L a CD C L E Q) O CD O o CD L_ Z5 -i—• pd CD Q . E CD -i—i O CD 20 i 16 H 20 cm 17-Apr 20 i 16 H 4i 12 H 8 H 4^ 0 17-Apr 16-Jun 15-Aug 1996 50 cm 14-Oct 16-Jun 15-Aug 1996 14-Oct Figure 2.7 continued. 50 I I I I I 07-Dec 07-Mar 05-Jun 03-Sep 02-Dec 02-Mar 31-May 29-Aug 1994 1995 1996 -2 -I 1 1 - 1 1 . 1 1 07-Dec 07-Mar 05-Jun 03-Sep 02-Dec 02-Mar 31-May 29-Aug 1994 1995 1996 Figure 2.8. Daily average, maximum and minimum soil temperature in a 17-ha opening at the Date Creek study area. Sampling period: December 1994 to October 1996. 0 01-Jan 31-Mar 29-Jun 12-Sep 11-Dec 11-Mar 09-Jun 07-Sep 1995 1996 Maximum 15 n 0 01-Jan 31-Mar 29-Jun 12-Sep 11-Dec 11-Mar 09-Jun 07-Sep 1995 1996 Minimum _ 15 n O 01-Jan 31-Mar 29-Jun 12-Sep 11-Dec 11-Mar 09-Jun 07-Sep 1995 1996 gure 2.9. Average, maximum, and minimum daily soil temperature in the forest 52 The effects of gap size and position inside a gap on soil temperature are more easily seen by comparing heat sums (>5° C) over the growing season (Table 2.1). At all depths, heat sums in the forest understory were 67 to 70% of those in the clearcut (in the °C degree day >5° C units used). In the north end of the 1785 m 2 gap, heat sums were 92 to 96% of those in the clearcut, whereas in the north end of the 205 m 2 gap, heat sums were 80 to 88% of those in the clearcut. To better contrast the effects of forest cover and gap position, average soil temperature at all depths was compared in the open clearcut, the forest understory, and the north and south positions of the 1785 m 2 gap (Fig. 2.10). The same clearcut and understory soil temperature data were also compared to values monitored at the north positions of the 205 and 1785 m gaps (Fig. 2.11). 2.3.3 Air temperature and humidity There was striking similarity in air temperature at both heights and in the two gaps monitored (205 and 1785 m2). Average air temperature at 1.4 m was slightly higher in the large gap than in the small gap throughout the growing season (Fig. 2.12). Minimum temperature was a little higher in the small gap and maximum air temperature was highest in the large gap (Fig. 2.12). The same trends were observed at 0.5 m height (data not shown). Minimum air temperature dropped below freezing twice in May at both heights. There was also little difference in relative humidity at the centre position of the 205 and 1785 m 2 gaps (Fig. 2.13). 53 Table 2.1. Soil heat sums at different depths in open conditions, in different positions in 1785 m 2 and 205 m 2 gaps and in the forest understory. Heat sums are the summation of degree-days based on the difference between the average daily soil temperature and a base temperature of 5°C. 5 cm depth Clearcut Large gap Small gap Forest North Centre Centre-south South North Centre South Total heat sums 1995 1790 1721 1583 1573 1489 1540 1474 1452 1259 1996 1783 1677 1536 1444 1397 1444 1369 1354 1207 Percent of clearcut heat sums 1995 100 96 88 88 83 86 82 81 70 1996 100 94 86 81 78 81 77 76 68 10 cm depth Clearcut Large gap Small gap Forest North Centre Centre/south South North Centre South Total heat sums 1995 1770 1674 1511 1503 1466 1517 1433 1399 1218 1996 1765 1626 1464 1376 1370 1410 1328 1307 1173 Percent of clearcut heat sums 1995 100 95 85 85 83 86 81 79 69 1996 100 92 83 78 78 80 75 74 66 20 cm depth Clearcut Large gap Small gap Forest North Centre Centre/south South North Centre South Total heat sums 1995 1712 1599 1488 1443 1427 1459 1383 1338 1151 1996 1702 1555 1435 1323 1338 1352 1282 1247 1118 Percent of clearcut heat sums 1995 100 93 87 84 83 85 81 78 67 1996 100 91 84 78 79 79 75 73 66 50 c m depth Clearcut Large gap Small gap Forest North Centre Centre/south South North Centre South Total heat sums 1995 1555 1471 1350 1361 1342 1368 1278 1220 1029 1996 1536 1427 1294 1253 1260 1267 1180 1142 1015 Percent of clearcut heat sums 1995 100 95 87 88 86 88 82 78 66 1996 100 93 84 82 82 83 77 74 66 Notes: 1. Growing seasons - May 4 to Sept. 15, 1995 and April 17 to Sept. 15, 1996. 2. Regression analysis was used to predict values for all missing data prior to calculation of heat sums: 1995, April 15-May 3 in the large gap and May 2-May 22, north position, 5 cm depth in the small gap; 1996, Sept. 4-5, south position, 20 cm depth in the small gap. 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 -i r 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 -2 01 •Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 Figure 2.10. Average daily soil temperature in the forest understory, in full open (clearcut) conditions and in north and south positions of a 1785 m 2 gap. 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 01-Jan 31-Mar 29-Jun 27-Sep 26-Dec 26-Mar 24-Jun 22-Sep 1995 1996 Figure 2.11. Average daily soil temperature in the forest understory, in full open (clearcut) conditions and in northern positions of 205 m 2 (small) and 1785 m 2 (large) gaps. 56 Maximum o o CD Z5 CS i _ CD C L Small gap Large gap 02-May 27-May 21-Jun 16-Jul 10-Aug 04-Sep 1995 40 30 20 B 10 0 Average o o CD 1 _ 13 -•—' cd CD C L E CD 02-May 27-May 21-Jun 16-Jul 10-Aug 04-Sep 1995 4 0 Minimum 30 20 10 0 -10 02-May 27-May 21-Jun 16-Jul 10-Aug 04-Sep 1995 Figure 2.12. Air temperature at 1.4 m height in 205 m 2 and 1785 m 2 gaps. Large gap Small gap 100 t5 80 E zs 60 > 40 CO a5 20 rr W W Average 02-May 21-Jun 10-Aug 1995 29-Sep 100 02-May 21-Jun 10-Aug 1995 29-Sep 100 S 80 5 60 sz > 40 co ai 20 rr Maximum 02-May 21-Jun 10-Aug 29-Sep 1995 > 60 40 co "55 20 rr o Maximum 02-May 21-Jun 10-Aug 29-Sep 1995 Figure 2.13. Average, maximum and minimum relative humidity at 1.4 m above the ground in 205 m 2 (small) and 1785 m 2 (large) gaps. 2.4 Discussion Microclimatic variability was affected by the intensity of forest canopy removal in these high latitude forests, but the magnitude of response differed for light, soil temperature, air temperature and humidity. Photosynthetically active radiation (PAR) levels remained low at all positions in the smallest gap (205 m2) and changed little as gap size increased to 600 m 2 . However, PAR levels in small gaps were much higher than in the understory of the undisturbed forest. This is consistent with findings in other gap studies (e.g., Collins and Pickett 1988; Denslow et al. 1990; Barik et al. 1992). Forest understory light levels under completely closed canopy conditions were 2-5% of those in full open conditions on a sunny day. This single-sensor reading was slightly lower than the 4 to 11% full sun range estimated by analysis of hemispherical photographs (see Chapter 3) taken above seedlings planted in the understory. In contrast to these dense 140 yr old forests, light levels in understories of old-growth (300 yrs +) ICH forests varied from 11 to 30 % full sun (Bartemucci and Coates, unpublished data). Light levels increased along a gradient from the south to north ends of the large gaps (1785 and 4600 m2), with the most dramatic difference being between the south position and the other positions. Similar results were observed by Minckler et al. (1973) in small and large gaps in mixed-hardwood stands in Illinois. In their study, light levels varied little by position in gaps with diameters less than 3/4 the canopy tree height, but in larger gaps the centre and north positions received the most light, followed by the west, east and south. Even the most exposed positions in the large gaps at Date Creek did not receive as much PAR as clearcuts. On a sunny day, the most open areas of large gaps received about 70% of that recorded in the open clearcut. 59 Soil temperature increased steadily, especially near surface depths, from the forest understory to ever more exposed positions in gaps, to the open conditions of the clearcut. However, under all canopy conditions except the closed forest, soil temperature remained above 10° C from early in the growing season until early autumn. Soil temperatures below 10° C are thought to delay bud burst, reduce root or shoot growth, increase water-flow resistance of roots, increase water stress, restrict nutrient uptake and decrease photosynthesis rates in conifers (Kaufmann 1975; Tranquillini 1979; Running and Reid 1980; Lopushinsky and Kaufmann 1979; Goldstein et al. 1985; DeLucia 1986). Except under the undisturbed forest canopy, there was little evidence to suggest soil temperature would differentially affect performance of established seedlings by gap size or position inside a gap. Even in the shadiest position in the smallest gap (205 m2), growing season heat sums were 74 to 81% of those recorded under fully exposed conditions. Growing season soil temperature in the forest understory was low enough to inhibit seedling performance. Air temperature and humidity varied little in the middle of the two gaps examined (205 and 1785 m2). Maximum air temperature during the growing season (30 - 35° C) were below levels thought to be harmful to established conifer tree seedlings (Seidel 1986) or that might affect photosynthesis rates (Mitchell and Arnott 1995). Maximum air temperature in these ICH forests were well below those recorded at 0.1 m above the ground in 20 and 40 m diameter gaps in a northern hardwood forest in Wisconsin (up to 46° C; Strong et al. 1997). In conclusion, light levels and soil temperature varied with gap size and gap position in the high latitude interior cedar - hemlock forests of northwestern British Columbia. As expected, these energy inputs increased from the southern side to the northern side of gaps, although once away from the south end, there were not dramatic differences by gap position in larger gaps. 60 Full light conditions were not found in gaps up to 0.5 ha in size. The variation in light levels associated with partial cutting in ICH forests should be sufficient to create different competitive conditions for the broad range of tree species occurring in these forests. 61 Chapter 3 Growth of planted tree seedlings in response to ambient light levels in northwestern interior cedar - hemlock forests of British Columbia. 3.1 Introduction Predicting the effects of management on long-term forest dynamics involves a complex suite of factors, especially when partial cutting systems are employed that retain variable levels of canopy trees in time and space. Light is arguably the most important factor influencing tree growth in temperate and tropical forest biomes (Denslow and Hartshorn 1994; Pacala et al. 1996), and certainly the easiest to manipulate. Understanding the response of individual seedlings and saplings to variation in light is therefore fundamental to predicting the effects of partial cutting on future forest composition and growth (Marks 1975; Shugart 1984; Carter and Klinka 1992). Ecophysiological studies of leaf-level responses have shown that individual tree species have characteristic photosynthetic responses to variation in incident light (e.g., Grossnickle and Arnott 1992; Sipe and Bazzaz 1994; Burton and Bazzaz 1995; Barker et al. 1997). These differences are not necessarily reflected in whole-plant field growth rates. Whole-tree carbon balance in the field depends on many variables including tree age, crown architecture, leaf number and age, leaf structure and orientation, and availability of soil resources (e.g., Field 1988; Givnish 1988; Klinka et al. 1992; Walters et al. 1993, 1997; Canham et al. 1996; Chen et al. 1996). 62 For British Columbia tree species, available data on whole-plant growth rates in response to ambient light levels are derived mostly from retrospective studies of natural-origin trees. The studies, which utilized either clearcut edges or prior natural disturbances to provide a range of light levels, had little control over location, tree age or growth history (Carter and Klinka 1992; Klinka et al. 1992; Chen et al. 1996; Kayahara et al. 1996; Wright et al. 1998; but see Wang et al. 1994; Chen 1997). Partial cutting can create wide variation in ambient light conditions and, it follows, variation in growth rates among individual tree species. The distribution of light environments within canopy gaps of the northwestern interior cedar - hemlock (ICH) forests change dramatically from small single-tree openings to openings of 110 to 120 m diameter (0.8 to 1.1 ha), after which the distribution of light environments within gaps begins to stabilize (Coates and Burton 1997). The maximum diversity of ambient light conditions is found in openings of 30 to 75 m diameter (0.07 - 0.4 ha) (Coates and Burton 1997). An understanding of how tree species differ in their growth responses to such variable canopy conditions and light levels is essential for prescribing species composition and distribution in partially cut stands, whether the objective is to maximize tree growth or to incorporate structural retention to maintain other forest values, or both. This study was designed to characterize variation in height and radial growth in response to ambient light levels for the five most commonly planted tree species in the northwestern ICH. This research addresses four questions: (1) How do both the magnitude of seedling response to light, and the shapes of the light response curves (i.e., the functional forms of light response), vary among the individual tree species under field conditions? (2) Are there clear trade-offs among tree species in aboveground growth at low vs. high light? (3) Is ambient light level a 63 good predictor of whole-tree growth in the field? (4) Which species are best suited to different gap sizes or gap positions based on their light response characteristics? 3.2 Methods The research was a component of the Date Creek silvicultural systems study (Coates et al. 1997), located near Hazelton, British Columbia, Canada (55° 22' N , 127° 50' W; 370-665 m elevation). The study area lies within the moist cold subzone of the Interior Cedar - Hemlock biogeoclimatic zone (ICHmc) (Pojar et al. 1987; Banner et al. 1993). Mature forests at Date Creek (130-140 years since fire) are dominated by western hemlock {Tsuga heterophylla (Raf.) Sarg.), in admixture with western redcedar (Thuja plicata Donn ex D. Don), subalpine fir (Abies lasiocarpa (Hook.) Nutt.), lodgepole pine (Plnus contorta var latifolia Engelm.), hybrid spruce [the complex of white spruce (Picea glauca (Moench) Voss), Sitka spruce (P. sitchensis (Bong.) Carr.) and occasionally Engelmann spruce (P. engelmannii Parry ex Engelm.)], paper birch (Betula papyrifera Marsh.), trembling aspen (Populus tremuloides Michx.) and black cottonwood (Populus balsamifera ssp. trichocarpa Torr. & Gray). Subalpine fir is commonly replaced by amabilis fir (Abies amabilis Dougl. ex Forbes) at higher elevations. In old-growth stands (350 yrs-t- since fire), western hemlock dominates with minor components of western redcedar, subalpine fir and amabilis fir. Descriptions of these mature and old-growth stages can be found in LePage (1995) and Coates et al. (1997). The treatments applied in the Date Creek silvicultural systems study created a range of overstory canopy structure which included undisturbed forest, two levels of partial cutting and clearcut forests. Within this experimental framework, seedlings of five species (western redcedar, western hemlock, subalpine fir, hybrid spruce and lodgepole pine) were planted in 64 early May, 1993. Seedlings were planted in the understory of the undisturbed forests, into 2 2 canopy gaps ranging in size from 10 m' to 5000 m\ and in full open conditions in clearcut treatment units (at least 3 tree heights from the forest edge). A l l planting stock was PSB 1+0 415 except for subalpine fir seedlings which were PSB 1+0 313. Stock was grown in peat- and perlite-filled styroblocks in greenhouses and open compounds at Kalamalka Nursery in the summer of 1992 and stored at -5° C over winter. Initial sizes were fairly consistent among species except that subalpine fir were shorter and redcedar taller than the other three species (Table 3.1). Sample trees were selected so that similar numbers of seedlings occurred across the full range of light conditions (Table 3.1). Hemispherical photographs were taken above individual trees of each species to quantify light available for growth over the growing season. An index of growing season light availability, GLI, was computed for each photograph using GLIC 2.0 software, following Canham (1988b). This index combines the seasonal distribution of sky brightness with the distribution of canopy openness to calculate a single index of available light in units of percent full sun for a specified growing season (mid-April through mid-September) (Canham et al. 1990). Sample trees were free of overtopping shrubs or herbs and had no evidence of nearby windthrow that might have recently affected light levels. Since growth of tree seedlings can be affected simultaneously by both light and soil resource availability (Latham 1992; Canham et al. 1996; Walters and Reich 1996), seedlings were sampled on mesic sites only to minimize variability in soil resources. When there was a choice of seedling to sample, the best growing individual was always selected. Thus, the light response curves represent the best performance that can be expected at any given light level. 65 Table 3.1. Provenance descriptions, characteristics of trees at planting and sample sizes by canopy condition. Western Western Subalpine Hybrid Lodgepole redcedar hemlock fir spruce pine Provenance description Seedlot number 35033 03437 35005 14577 27792 Subzone ICHmc ICHmc ICHvc ICHmc ICHmc Location Date Creek Suskwa River Bell II Calvin Creek Kitseguecla River Latitude 55° 27' N 55° 17' N 56° 45' N 55° 30' N 55° 06' N Longitude 127° 50'W 127° 17' W 129° 50'W 128° 45' W 127° 50'W Elevation (m) 370 500 700 610 650 Characteristics at planting3 Height (cm) 41.2+0.81 24.5 +0.82 13.4 +0.36 27.4 +0.56 21.8 ±0.81 Diameter (mm) 3.3 +0.07 3.3 +0.07 3.9 ±0.08 3.7 +0.06 4.2 ±0.08 Sample size Under closed canopies" 28 20 8 22 12 gaps < 300 m 2 24 17 32 30 28 gaps 301-1500 m 2 18 19 11 15 19 gaps 1501-5000 m 2 12 11 11 13 12 full open conditions 12 12 12 12 12 Total 94 79 74 92 83 a45 trees randomly selected per species; planted spring, 1993; mean ± SE. ''trees sampled in unlogged treatment units or from under the canopy adjacent to a sampled canopy gap. Hemispherical photographs were taken in the late summer of 1996, at the end of the 4th growing season. Seedlings selected for this study were measured for height increment and total height at the end of the first growing season, and basal diameter and total height at the end of the 2nd, 3rd, 4th and 5th growing seasons. Canopy crown filling into openings is very slow in these forests (D. Coates, personal observation) and was not thought to have significantly altered light levels over the 5 years. 66 3.2.1 Data analysis Non-linear regression was used to develop growth functions for each of the 5 tree species. Total height and basal diameter 4 and 5 growing seasons after out-planting, and average height and radial increments over the last 3 years were used as response variables. For recent growth rates, I analyzed absolute growth rather than relative growth. Initial height and diameter within species was very uniform. Among species, initial diameter was similar, but height varied by up to 9-17 cm at planting (Table 3.1). In a companion study (Wright et al. 1998) where tree sizes varied widely, growth did not show consistent size-dependent relationships. Transformation of the growth data was not required to stabilize variances. Residuals for all growth variables were well balanced. In the Wright et al. (1998) study, three non-linear models were tested for their ability to predict sapling response to ambient light levels: the Weibull, Chapman-Richard's and Michaelis-Menton functions. The Michaelis-Menton equation provided more consistently reliable fits for the growth variables and its parameters were more easily interpreted, allowing comparisons of growth response at high versus low light. In this study, I used two forms of the Michaelis-Menton equation: [1] Y = ((a L)/((a/s) + L)) + e [2] Y = ((a (L-c))/((a/s) + (L-c))) + e where in equation 1, Y is a growth reponse variable, a is the asymptote of the function at high light, s is the slope of the relationship at zero light, L is light level, or relative irradiance (here measured by GLI in units of percent full light received over the growing season) and the error term is assumed to be normally distributed. Equation 2 allows for tests of non-zero intercepts, where a and s are as previously described and c is the light level at which growth equals zero (ie. 67 a whole-plant light compensation point). Equation 1 was used to predict height and diameter of seedlings after 4 and 5 growing seasons. For recent growth rates (average height and radial increments over the last 3 years) both models were tested. Model parameters were estimated for each species using the Nonlin procedure in Systat 6.0 with the simplex estimation method to minimize the loss function (Systat, Inc. 1996). The 95% confidence limits for parameters a, s and c for each species were used as a conservative method for making pairwise comparisons among species at high and low light and for comparing non-zero intercepts. If the 95% confidence limits for two species did not overlap then the parameter estimates were deemed significantly different. 3.3 Results Ambient light levels proved to be a good predictor of tree size and growth rates under field conditions. Goodness-of-fit of the Equation 1 regression models to predict height and diameter of seedlings after 4 and 5 growing seasons was generally good (R2=0.55 to 0.89; Table 3.2). Because there was evidence among the tree species of non-zero intercepts for recent growth, Equation 2 was used to predict average radial and height growth over the last three years. Ninety-five percent confidence limits for the whole plant light compensation point (c parameter) overlapped with zero in only two cases: western redcedar and western hemlock diameter growth (Table 3.3). In all other cases, light compensation points were significantly greater than zero. The goodness-of-fit for Equation 2 predictions of recent growth rates was good (R =0.53 to 0.87; Table 3.3). The rank order of c parameter estimates (Table 3.3) were consistent with the shade tolerance ranking of the tree species, from most to least shade tolerant: western redcear > 68 western hemlock > subalpine fir > hybrid spruce > lodgepole pine (Krajina 1969; Krajina et al. 1982; Kobe and Coates 1997). However, pairwise comparisons of the non-zero intercept parameter were not significantly different among any of the species for either average 3-yr height or diameter growth, suggesting no meaningful variation in whole-plant compensation points among the five tree species when growing under field conditions. No subalpine fir seedlings were planted in the forest understory, hence the c parameter estimates for subalpine fir were derived from data that did not include trees growing at the lowest light levels sampled for other tree species. This may have artificially increased the c parameter estimates for subalpine fir, indicating zero growth at a higher light level than would have been the case if seedlings had been present to sample at lower light levels. The shapes of the light response curves (Figs. 3.1-3.4) were also generally consistent with expectations based on shade tolerance ranking of the species. There were, however, no clear trade-offs in low vs. high light growth. The rank order of the s parameter values (response at low light) did not correlate well with traditional shade tolerance rankings. The most shade tolerant species (western redcedar and western hemlock) had the highest s parameter estimates for recent height growth at low light, followed by lodgepole pine (Table 3.3). Pine had the highest s parameter estimate for low light diameter growth. This was an unexpected result based on lodgepole pine's ranking as the most shade intolerant of these conifer species (Krajina 1969; Kobe and Coates 1997). There were, however, no significant differences in species pairwise comparisons at low light, based on overlapping 95% confidence limits for the s parameter (Table 3.3). Asymptotic height growth of lodgepole pine at high light (a parameter) was significantly greater than that of redcedar and subalpine fir. Hemlock high light growth was superior to that Table 3.2. Parameter estimates for predicted fourth and fifth year height (cm) and diameter (mm) using the equation Y=((aL)/((a/s)+L)) where L is the light level. a parameter s parameter Fourth year height estimate SE 95% C.L. estimate SE 95% C.L. R2 N Western redcedar 152.7 7.8 15.5 12.48 1.72 3.41 0.56 94 Western hemlock 177.0 11.4 22.8 7.80 1.02 2.04 0.66 79 Subalpine fir 86.8 8.1 16.2 2.69 0.40 0.79 0.55 74 Hybrid spruce 177.8 11.9 23.7 4.61 0.41 0.81 0.75 92 Lodgepole pine 281.7 26.2 52.2 4.28 0.35 0.70 0.78 83 Fifth year height Western redcedar 198.5 12.1 24.0 11.92 1.62 3.21 0.60 93 Western hemlock 260.6 23.3 46.3 7.56 1.04 2.07 0.68 77 Subalpine fir 125.7 16.0 31.9 2.77 0.44 0.87 0.55 72 Hybrid spruce 279.0 28.6 56.9 4.39 0.42 0.83 0.76 91 Lodgepole pine 506.8 71.1 141.5 4.55 0.37 0.73 0.81 80 Fourth year diameter Western redcedar 41.2 4.5 8.9 0.78 0.09 0.18 0.71 94 Western hemlock 43.1 6.3 12.5 0.66 0.09 0.18 0.68 79 Subalpine fir 26.4 2.6 5.2 0.53 0.06 0.12 0.68 74 Hybrid spruce 61.8 8.1 16.1 0.52 0.04 0.08 0.86 92 Lodgepole pine 302.6 168.8 335.8 0.48 0.03 0.07 0.87 83 Fifth year diameter Western redcedar 50.4 6.3 12.5 0.89 0.11 0.22 0.69 93 Western hemlock 80.6 20.2 40.3 0.65 0.09 0.18 0.70 77 Subalpine fir 30.9 3.6 7.1 0.60 0.08 0.16 0.64 72 Hybrid spruce 107.3 20.0 39.7 0.57 0.04 0.08 0.89 90 Lodgepole pine 930.8 1390.9 2769.7 0.57 0.04 0.08 0.88 79 Note: a and s, parameters of model; 95% C.L., 95% confidence limits; SE, 1 standard error of estimated parameter values. o co OJ >. I co *—' CO _CO 1_ OJ > o 1 CO CD I CD ED-IT o 21 o c gp 'CD -CZ CD tzn CO CD I* in UJ co CD •I—< CO E w CD IO in 0) LU CO 0 -4—• CO E • CO 0 o 0 0 E co C L LO CD LU CO 0 CO E +-» CO 0 CO CD CM i - CD O) N N CD N S CO N CM CM LO CD LO 00 00 0 0 0 0 0 S CO CD O) Tf - i - oo o p ci ^ i n i o IO CD Tt" CO CO O CO O LO CM CM c\i c\i O) IO N ^ CO ^ CD O N CO CO ^ N CO r-~ i^ - Tf CM Tf O O O C) o O) N CO CM r -C0 CO CM - i - CM o o o d o CO CD CD O CD ^ m ID s i -d d cq q ID s ^ O O - i - 1 LO CD T - i - i - h~ CD O CO r ; 00 iri d iri co ^ f- co co O) N CO CO CO CD Tf CM co co OJ i - co Jar: o o o p 0 .E 2 -«= 0 0) c cz c lo o r-i C L 0 0 — "O (D 5 § co x J J CO 0 >> CO w CO 0 > o 1_ CO 0 1 CD C L IT 0 E 0 o c 1_ 0 •—» 0 E co TJ 0 CO CO 1 0 > < o LU CO 0 -I—' CO E -t—' CO 0 o c> LO CD LU CO 0 CO E to 0 CO LO CM O CD O) N N O) N CO CM CO N LO !D N IO CO CO O CD CD O O r-~ i - - i -CM 00 LO Tf o CD 00 00 Tf Tf CD N S N CO T- CM O O co Tf Tf c\i CM CM i - S CO CM cd iri co CD co _1 d CO o O CD o CD O LO CD d d d Tf d Ul Tf o Tt-o LO o CM O co o CO d d d d d timate CM CD CD CO oo Tf CM o o timate T T— i — T— CM timate CD o CD CD d esl CM CD Tf , - LO o o •Y— CO LO od CD co Tf N (D r-O co o cd co C  CM O CD CD LO co o o CO TJ 0 CJ ~ T3 0 0 CZ c » V— cu 0 00 to 0 0 5 5 0 .!= O ••- Zi 0 cl .<= co J3- -o CO 'CZ .Q .O ZI CO I > -4—> cj s S3 a , -a u o u a ca CL, L) "a K 71 Subalpine fir 3001 1 1 1 1" o £ 200 -0 20 40 60 80 100 Percent of full sun Hybrid spruce 3001 1 1 1 If o E 200 h 0 20 40 60 80 100 Percent of full sun Lodgepole pine 0 20 40 60 80 100 Percent of full sun Figure 3.1. Observed values and fitted regression lines of seedling height 5 years after planting vs. percent of full sun for 5 tree species using using Equation [1] and parameter values in Table 3.2. 72 Lodgepole pine 801 1 1 1 r 0 20 40 60 80 100 Percent of full sun Figure 3.2. Observed values and fitted regression lines of seedling diameter 5 years after planting vs. percent of full sun for 5 tree species using Equation [1] and parameter values in Table 3.2. 73 Lodgepole pine 801 1 1 1 1 0 20 40 60 80 100 Percent of full sun Figure 3.3. Observed values and fitted regression lines of height growth rates (cm, average per year for last 3 yrs) vs. percent of full sun for 5 tree species using Equation [2] and parameter values in Table 3.3. Western redcedar Western hemlock Subalpine fir 201 1 1 1 20 40 60 80 100 Percent of full sun Hybrid spruce 201 1 1 1 Percent of full sun Lodgepole pine 201 1 1 1 1 0 20 40 60 80 100 Percent of full sun Figure 3.4. Observed values and fitted regression lines of diameter growth rates (mm, average per year for last 3 yrs) vs. percent of full sun for 5 tree species using Equation and parameter values in Table 3.3. 75 of subalpine fir. Diameter growth rates at high light were not significantly different among the species (Table 3.3). Growth rates at high light were highly variable within species (Figs. 3.3 and 3.4). Differences in initial starting sizes among the tree species (Table 3.1) made comparisons of predicted size at 5-yrs difficult. Comparisons among species of 3-yr average growth rates were better since planting effects in the first two years can be significant. Predicted growth rates diverged at high light levels, with the least shade tolerant species, lodgepole pine, having the greatest radial and height growth in full sun (Fig. 3.5). Above 40% full sun, lodgepole pine diameter growth rapidly increased compared to the other species. Predicted diameter growth of redcedar, hemlock and spruce was very similar at all light levels (Fig. 3.5). The tree species differed more in height growth than diameter growth across the light gradient (Fig 3.5). Predicted height growth of lodgepole pine was equal to or better than that of other species at all light levels and far better above 60% full sun. Western redcedar and western hemlock height growth diverged above 30% full sun with hemlock maintaining good growth rates and redcedar dropping off. Hybrid spruce height growth increased steadily with increasing light. Subalpine fir had the poorest predicted growth at both high and low light. This result may reflect poor planting stock quality rather than an inherent characteristic of the species (discussed in greater detail in Chapter 4). 3.4 Discussion Partial cutting of forests presents new problems for foresters selecting tree species to plant after logging. In the open conditions of clearcuts, concern centred around stock quality, site preparation and vegetation management techniques to enhance early survival and promote rapid 76 Figure 3.5. Predicted fifth year height (a), height growth rates (b), fifth year diameter (c) and diameter growth rates (d) in response to ambient light levels. 77 growth rates after outplanting (Walstad and Kuch 1987; Lavender et al. 1990). Because all trees were planted in open conditions, little attention was paid to a species' light response characteristics. Until recently, information on tree species light response came from leaf-level, controlled environment studies where light levels were manipulated alone or in combination with other resources. Basing species selection criteria on physiological data assumes that the relationships between leaf-level and whole-plant performance are strong and consistent. In fact, there are very few data comparing leaf-level and whole-plant performance across light gradients (Sipe 1990; Barker et al. 1997), and the few studies that have examined whole-plant light response in the field often have had different results than leaf-level studies, especially at low light (Kitajima 1994; Pacala et al. 1994; Barker et al. 1997; Wright et al. 1998; but see Burton and Bazzaz 1995). The response of planted seedlings to variation in light in this study were consistent with those reported for natural saplings in ICH forests by Wright et al. (1998). Under field conditions, there was no meaningful variation in whole-plant compensation points among the five tree species. The low-light performance of lodgepole pine was similar to or better than that of the other species, which contrasts with expectations from physiological studies (Bassman 1985). Leaf-level measurement of photosynthesis for the ICH tree species indicate that the saturation point is reached at 30-40% of full sunlight (Dykstra 1974; Smith 1985; Lavender 1990; Major 1990; Grossnickle and Arnott 1992; Koppenaal et al. 1995; Mitchell and Arnott 1995). This was not the case for measures of whole-plant growth in the field. Predicted diameter and height growth continued to increase steadily with increasing light, especially for the two species considered most shade intolerant, lodgepole pine and hybrid spruce (Krajina 78 1969; Kobe and Coates 1997). Contrary to predictions from leaf-level studies, growth responses among the species varied considerably above 40% full sun. The results of this study are in general agreement with those of other field studies that have examined growth responses of the five species to variation in light (Carter and Klinka 1992; Klinka et al. 1992; Wang et al. 1994; Chen et al. 1996; Kayahara et al. 1996; Chen 1997; Wright et al. 1998). Goodness-of-fit for the predicted models were often much higher than those previously reported. This was probably due to the use of planting stock with similar initial sizes and better predictions of growing season ambient light levels with hemispherical photographs, than the single day measurement technique used by Klinka and colleagues (see Carter and Klinka 1992). For example, using planting stock improved goodness-of-fits for height growth by 15 to 35% (except for redcedar 5%) in my study compared to natural origin trees of the same ICH species reported in Wright et al. (1998), who also used hemispherical photographs to quantify light levels. Ambient light level was an excellent predictor of total seedling size and current growth rates, even though other resources and abiotic conditions such as soil moisture and nutrients and soil temperature would have varied across the different gap sizes used to create the light gradient. This study and the companion study (Wright et al. 1998) confirm the expectation that ICH species traditionally classified as shade tolerant have the greatest response to an increase in light at low light levels (i.e. high s values), but have relatively low asymptotic growth at high light (i.e. low a parameter values) (e.g. Klinka et al. 1992; Walters et al. 1993; Pacala et al. 1994; Walters and Reich 1996). Also as expected, the two species traditionally classified as more shade intolerant (hybrid spruce and lodgepole pine) had the highest growth (high a parameter values) at high light levels. The exception to the anticipated low light vs. high light 79 trade-off was the very good performance of planted lodgepole pine at low light. Pine growth rates were among the highest at all light levels. Other recent studies have also reported high growth rates of shade intolerant species at low light (e.g. Walters and Reich 1996). Tree species in tropical and lower latitude temperate forests, which have evolved to regenerate primarily through gap dynamics, have been shown to sharply increase their growth rates at low light levels and can reach 50% of maximum growth by 10% of full sun and near maximum growth at 20% full sun (e.g., Fetcher et al. 1993; Pacala et al. 1994). This study, under more controlled conditions than those of Wright et al. (1998), confirms their observation that conifers of the high latitude ICH forest have a fundamentally different response to light than tropical and lower latitude forest species. Growth of the ICH species increased gradually with increasing light and continued to increase even at very high light levels, albeit slowly for the most shade tolerant species. 3.4.1 Management implications Logging disturbance alters resource availability and physical conditions in forests. One of the most important resources for tree performance is light availability, and it is the one resource that silviculturists can easily manipulate through the spatial and temporal extent of canopy tree removal. This study has shown that the five tree species commonly planted in ICH forests have broadly overlapping absolute growth rates from low to high light. The greatest variability in growth rates was at intermediate light levels (30-70% full sun) where careful matching of tree species to light environment can maximize growth rates. Contrary to popular expectations, lodgepole pine grows as well as or better than the other species in absolute growth at all light levels within partial cutting systems. 80 At low light levels all the species exhibited similar growth rates, but at very low radial growth rates (<0.2 mm/yr) mortality rates of hybrid spruce and lodgepole pine will be unacceptably high (Kobe and Coates 1997). Growth rates among the species at high light were broadly overlapping and there was considerable variability within species. Above 70% full sun, silviculturists should expect little difference in growth rates among species and only slightly improved growth as light levels approach full sun. Only lodgepole pine has clearly superior growth rates in full sun. 81 Chapter 4 Gap size and position effects on natural and planted tree regeneration in northwestern interior cedar - hemlock forests of British Columbia. 4.1 Introduction During the 1970's and 80's the majority of forest regeneration practices in British Columbia followed an agricultural model. Forest researchers examined techniques for establishing and promoting early growth of planted seedlings after clearcutting (Lavender et al. 1990; Walstad and Kuch 1987), followed by growth and yield studies that emphasized single-aged and single-species growth with no overstory canopy influences (e.g., Ek et al. 1988). In recent years this industrial forestry model has been questioned. The emphasis has shifted to a more ecologically and socially-based approach to forest management that requires greater structural diversity be maintained in managed stands in order to meet a broader range of forest values (e.g., Gilmore 1997; Kohm and Franklin 1997). As managed forest stands become more complex, foresters will require more complex predictive models of forest regeneration and growth that are capable of accommodating variable levels of canopy retention. Tree regeneration occurring inside canopy gaps after a disturbance has been studied worldwide (see reviews by Bazzaz and Pickett 1980; Denslow 1987; Piatt and Strong 1989). Population- or community-level responses within gaps of varying size are often used to explain how tree species diversity is maintained in forests subject to small-scale disturbance, a process referred to as 'gap partitioning' (e.g., Ricklefs 1977; Denslow 1980). While gap dynamics 82 studies have contributed significantly to our understanding of the role of small-scale disturbance in forest ecosystems, they have been little used by foresters to predict regeneration response following partial cutting. Gap size strongly affects the success of natural regeneration in forests (Denslow 1980; Brokaw 1985, 1987; Whitmore 1989; Stewart et al. 1991). Position either inside or outside of the gap may also influence regeneration success (Poulson and Piatt 1989) but is less studied, perhaps because most gap studies have been carried out in low latitude forests where high sun angles make gap position less likely to have a profound effect on biotic and abiotic conditions. In high latitude northern forests, position inside a gap can have considerable effect on resource levels (such as light availability; see Canham et al. 1990 and Chapter 2) and microclimate conditions (e.g., soil temperature; see Chapter 2), especially along the north-south axis. Such environmental heterogeneity should result in differential success of seedling establishment and subsequent growth (Bazzaz 1984). However, neither gap size nor gap position effects have been experimentally tested in high latitude forests (but see Wright et al. 1999). Relying solely on natural regeneration to separate the effects of gap size and position is problematic. Many factors influence seedling establishment following canopy disturbance. Parent tree proximity and abundance (Ribbens et al. 1994, LePage et al., submitted), seedbed substrate (Garman and Orr-Ewing 1949; Prochneau 1963; Day 1964; Eis 1965, 1967; Waldron 1966; Putz 1983; Lawton and Putz 1988; Stewart et al. 1991; Hofgaard 1993), presence of seed consumers and dispersers (Schupp 1988; Schupp et al. 1989) and climatic and microclimatic variability (Waldron 1966; Hennessey 1968; Noble and Alexander 1977; Alexander 1984) all cause dramatic spatial and temporal variation in seedling recruitment. Planted trees can be used 83 to avoid many of the stochastic events surrounding natural seedling establishment. Also, planted seedling performance can be contrasted with natural seedling establishment across a range of gap sizes and positions to test for "seed-seedling discordancy" in the regeneration niche (Schupp 1995). Partial cutting creates gradients of canopy influence that can affect seedling establishment and growth in many ways. Studies that examine regeneration and growth responses of interior cedar-hemlock forest tree species within gaps (of varying size and configuration) and within the adjacent forest matrix can form the basis for explaining community responses following partial cutting in these high latitude forests. Specifically, this study was designed to address four questions. (1) How does gap size affect natural regeneration and growth of tree seedlings in high latitude forests? (2) Is gap size a good predictor of tree growth? (3) How are natural regeneration and growth of planted trees affected by position inside and outside of gaps in high latitude forests? (4) Do these responses vary with the shade tolerance rankings of tree species? The influence of gap size and gap position on tree performance is addressed here. A number of relationships, not mutually exclusive, are possible: (1) density of shade-intolerant natural regeneration may increase as gap size increases (after Brokaw 1985, 1987; Whitmore 1989); (2) density of shade-intolerant natural regeneration may vary with position inside or outside of the gap, especially within larger gaps; (3) growth of shade-intolerant species may increase as gap size increases; (4) growth of shade-intolerant species may be best in the north end of gaps, especially within larger gaps; (5) the best gap size or gap position for establishment of natural regeneration may not be the best location for growth of planted seedlings, that is, the regeneration niche is "discordant", after Schupp (1995). 84 4.2 Methods This study is a component of the Date Creek silvicultural systems study (Coates et al. 1997), established in 1992, in northwestern British Columbia, Canada, approximately 21 km north of Hazelton, west of the Kispiox River (55°22'N, 127°50'W; 370-665 m elevation). Date Creek is within the moist cold subzone of the Interior Cedar-Hemlock zone (ICHmc), a transitional zone between the interior and coastal areas of northwestern British Columbia. See Pojar et al. (1987) or Meidinger and Pojar (1991) for a description of this system of vegetation classification and Banner et al. (1993) for detailed description of the ICHmc subzone. Mature forests (140 yr since fire) in the area are dominated by a mixture of conifer and deciduous tree species. Western hemlock {Tsuga heterophylla (Raf.) Sarg.) dominates, in mixture with western redcedar {Thuja plicata Donn. ex D. Don), subalpine fir (Abies lasiocarpa (Hook.) Nutt.), lodgepole pine (Pinus contorta var latifolia Engelm.), hybrid spruce [the complex of white spruce (Picea glauca (Moench) Voss), Sitka spruce (P. sitchensis (Bong.) Carr.) and occasionally Engelmann spruce (P. engelmannii Parry ex Engelm.)], paper birch (Betula papyrifera Marsh.), trembling aspen (Populus tremuloides Michx.), and black Cottonwood (Populus balsamifera ssp. trichocarpa Torr. & Gray). Amabilis fir (Abies amabilis Dougl. ex Forbes) is found at higher elevations in the study area. In the old-growth stands (300 yrs+ since fire), western hemlock dominates with minor components of western redcedar, subalpine fir and amabilis fir. Descriptions of these mature and old-growth stages can be found in LePage (1995) and Coates et al. (1997). Canopy trees averaged about 30 m tall, but there were abundant intermediate and understory trees, resulting in full canopies to near ground level. 85 Four tree removal treatments (in approximately 20 ha treatment units) were replicated four times (16 units in total), applied in a randomized block design, with blocking factor being ecological site type and age of forest (mesic 140 yrs; mesic-submesic 140 yrs; mesic-subhygric 140 yr; mesic 350 yr). The removal treatments included undisturbed forests, light and heavy partial cutting and clearcuts. In the light partial cutting treatment, approximately 30% of the stand volume was removed by cutting either single stems or small groups (3-10 trees). In the heavy partial cutting treatment, approximately 60% of stand volume was removed. The cutting pattern utilized both large gaps (0.1-0.5 ha in size), evenly distributed across the treatment units and either single tree or small gaps in the forest matrix between the large openings. In the clearcut treatment units all stems were removed except for scattered residual deciduous trees, mostly trembling aspen and paper birch. Cutting treatments and the pre- and post-treatment stand attributes are fully described in Coates et al. (1997). 4.2.1 Gap selection In spring 1993, logging-created canopy gaps were enumerated at approximately 63 evenly spaced grid points (50 by 50 m) within each light and heavy partial cutting treatment unit. Gaps were enumerated if they could be seen from the grid point, if they had an evenly distributed canopy around their perimeter and if they were dominated by the mesic site series (Banner et al. 1993). From this population of openings, individual gaps were randomly selected in roughly equal numbers from each treatment unit. A total of 109 gaps were selected: 76 less than 1000 m 2 in size, 33 between 1000 and 5000 m 2 , including all 5 suitable gaps greater than 3000 m 2 . The size of each gap was determined using methods described by Runkle (1992). Briefly, canopy gap size was calculated as the area of an ellipse. The major axis of the ellipse was the longest line that could be run from canopy edge to canopy edge inside the gap. The minor axis 86 line of the ellipse was the longest line that could be run from canopy edge to canopy edge perpendicular to the long line. In addition, the length of the longest north to south line that could be run from canopy edge to canopy edge was determined. This was done by standing under the canopy at the most northerly edge of the gap and measuring due south to the most southerly edge of the gap. Each canopy gap was divided into 4 sections or positions along the north-south axis. There were 3 positions inside the gap (northern, middle and southern thirds) and one gap edge position - the immediate understory area off the edge of the canopy gap. Thus the term "gap position" refers to areas both inside and outside of a distinct canopy gap. 4.2.2 N a t u r a l r e g e n e r a t i o n The density of all natural regeneration was assessed in each experimental gap in circular, 0.25 m 2 area subplots, established at each of the gap positions shown in Fig. 4.1. Regeneration subplots were assigned to gap positions categories as follows: north = 1, 2, 7, 8; middle = 3, 4, 9, 10, 21-24; south = 5, 6, 11, 12; and gap edge understory = 13-20. Twenty natural regeneration subplots were also established in each clearcut and undisturbed forest treatment unit. In each subplot the total number of individuals by species and origin was recorded. Although advance regeneration and paper birch stump sprouts were counted, these results are not reported here. Stem counts were taken in the fall of 1994 and 1995, the end of the second and third growing seasons, respectively. The undisturbed forest treatment units were measured in 1995 only. 4.2.3 P l a n t e d s e e d l i n g s In spring 1993, seedlings of five species (western redcedar, western hemlock, subalpine fir, hybrid spruce and lodgepole pine) were planted in gaps and in the undisturbed and clearcut treatment units. Due to shortage of stock, no subalpine fir seedlings were planted in the Figure 4.1. Layout of plots for natural regeneration study. a) The N-S line (plots 1-6) is the longest line that can be run from north to south in a gap. b) Gaps <1000 m 2 have 20 plots (1-20). c) Gaps > 1000 m 2 have 24 plots (1-24). d) In gaps <300 m 2 plots 1 and 6 are 1 m from the canopy edge into the gap. e) In gaps >300 m 2 plots 1 and 6 are 2 m from the canopy edge into the gap. f) Plots 2-5 are equally spaced along the N-S line (i.e. length of N-S line/5 = distance to plot 2 from canopy edg and the distance between plots 2, 3,4, and 5). g) Plots 7-12 are located half-way to the furthermost canopy edge to either the east or west of the N- S line. h) Plots 13-16 and 17-20 are 2, 4, 8, and 16 m into the canopy from the north and south canopy edges, respectively. i) Plots 21-22 and 23-24 are east and west, respectively from plots 3 and 4. They are located 2 m into the gap 88 understory of the undisturbed forest treatment units. Stock types were as described in Chapter 3, section 3.2. Initial sizes were fairly consistent among species except that subalpine fir were shorter and redcedar taller (Table 3.1). In gaps, seedlings were planted in single-species rows oriented north-south; planting rows were spaced 2 m apart and trees within lines spaced 1.5 m apart. In larger gaps, east-west planting lines for each species were also established. In all cases, tree species were randomly assigned to planting lines. Where conditions allowed, up to 10 seedlings were planted in each cardinal direction from the edge of the gap into the understory of the forest matrix. In clearcuts, seedlings were planted at least 3 tree heights from a forested edge to ensure they were in full open conditions. Measurement trees were selected in the fall of 1993. In gaps larger than 300 m 2 , trees were selected so that the number of each species was balanced in each gap position. A l l seedlings were measured in smaller gaps. In the undisturbed forest and clearcut treatment units, 15 of 30 planted seedlings per species were measured. In all, about 1,000 trees per species were measured. The current years height increment, total height and seedling condition were recorded at the end of the first growing season. At the end of the 2nd, 3rd and 5th growing seasons, basal diameter, total height, and seedling condition were recorded. 4.2.4 Analysis A randomized block split-split-plot design was employed to study the influence of 3 size classes of gaps, 4 different gap positions and tree species on density of naturally established seedlings and planted tree growth performance. Main plots were small gaps (10-300 m 2), medium gaps (301-1000 m2) and large gaps (1001-5000 m2). The 4 gap positions (north, middle and southern thirds inside gaps and understory off the gap edge) formed the first split-plot, with the different tree species providing the second split-plot. An individual gap was 89 considered the experimental unit in this design and individual planted seedlings or natural regeneration subplots were sub-samples. Pairwise comparisons were used to test for differences among size classes and gap positions. An adjustment was used to control for Type I error inflation: namely the pre-determined a-level for differences between pairwise comparisons to be deemed significant (0.05) was divided by the number of comparisons being tested (Milliken and Johnson 1992). Thus, an observed p-value had to be less than this preset level of significance for the comparison to be deemed significantly different. Three pairwise comparisons were selected to examine the influence of gap size class: large vs medium, medium vs small and large vs small. For a response variable to be significant between 2 size classes, a p-value of less than 0.0167 was needed. The observed p-value for a gap position pairwise comparison (gap edge vs south, middle vs north, middle vs south and north vs south) was deemed significant if less than 0.0125. The gap edge position caused problems with analysis of the main effects (the 3 size classes of gaps), and were excluded so that response means for individual gaps, and the mean for all gaps within a size class, properly represented average conditions inside the gap. Including canopy edge data in the analysis of the main effects would have artificially lowered the mean for individual gaps as gap size increased. Hence, all statistical tests for differences among or between gap size classes were done with gap edge position regeneration subplots and planted trees removed. For statistical tests of gap position effects, gap edge trees and natural regeneration subplots were retained. The clearcut treatment (representing full open conditions) and the undisturbed forest treatment (representing understory conditions) were analyzed separately as a randomized 90 block, split-plot design because they contain no anologous treatments to gap position. In this analysis, open and understory conditions were the main plots and tree species were the split-plots. Values from this analysis and from the split-split-plot design above, were used to manually calculate pairwise comparisons between full open conditions and large gaps and between small gaps and the understory of the undisturbed forest for each tree species. P-values less than 0.025 were deemed significant. Analyses of variance used the M I X E D Procedure from SAS because the model being tested included both fixed and random factors (SAS Institute Inc., 1989). Data presented in the figures are lsmeans with standard errors, as calculated by SAS. Lastly, non-linear regression was used to develop growth functions for each of the 5 planted tree species as a function of either gap area or the length of the north-south line. Most gap studies in lower latitude and tropical forests, where sun angles are high in the sky, have used gap size as the independent variable in models of tree growth response. In higher latitude forests, sun angles are much lower in the sky resulting in a more pronounced north-south gradient in ambient light levels over the growing season (see Chapter 2). Thus, similar-sized gaps with different orientations (hence different north-south light gradients) could result in different growing conditions for seedlings. To test this hypothesis, I also developed predictive models where the independent variable was the length of the longest north to south line inside the gap. As in Chapter 3,1 used the Michaelis-Menton equation which takes the form: [1] Y = ((a G)/((o/s) + G)) + e 91 where Y is seedling growth response, a is the asymptote of the function, s is the slope of the relationship at zero gap size or north-south line length, G is the measure of gap size or north-south line length and e is the error term of the equation. 4.3 Results 4.3.1 Natural regeneration Across all experimental blocks, natural regeneration was abundant in large, medium and small gaps. Western hemlock and paper birch (4-10 seedlings per m2) were the most common species, followed by western redcedar and hybrid spruce (1 - 2.5 seedlings per m2). A l l remaining tree species had similar but much lower densities than the 4 dominant species (<0.5 seedlings per m ; Fig. 4.2). These trends were consistent in both the mature (Fig. 4.3) and old-growth (Fig. 4.4) forest types with a few notable exceptions. First, birch, pine, aspen and cottonwood regeneration was absent in the old-growth units reflecting the very low densities of parents found in the old-growth forest (Coates et al. 1997). Second, hemlock density in old-growth medium and large gaps was much higher than in the mature forest. Finally, fir regeneration was much higher in the old-growth than in the mature stands. Because of difficulties distinguishing between young seedling of amabilis fir and subalpine fir, seedlings in the old-growth were called amabilis fir and all seedlings in mature forests were called subalpine fir. Few seedlings in mature forests would have been amabilis fir due to lack of parents (Coates et al. 1997). Ssubalpine fir parents were present in low numbers in old-growth areas. 92 Western hemlock E GO C CD T3 CD T3 CD CD CO U S M L O Opening type Paper birch _ 15 CVJ E 10 CO c CD TJ f 5 T3 CD CD CO 0 ~i r U S M L O Opening type Western redcedar 41 1 i 1 1 1— CM E S 3 -U S M L O Opening type Amabi l is and subalpine fir 41 1 1 1 1 1 CM E S 3 ->^  -t—> CO g 2 -T 3 CD C U S M L O Opening type Figure 4.2. Average density of natural regeneration across all experimental blocks by opening type and year of assessment (U= forest understory; S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ; 0= full open; M = end of 2nd year after logging; CU = end of 3rd year after logging). Error bars represent 1 standard error of the mean. 93 Hybrid spruce 2.0 I 1 1 1 1 r CNJ E S 1.5-U S M L O Opening type Black cottonwood 2.01 1 i 1 1 1— CM E S 1.5-U S M L O Opening type Trembl ing aspen Lodgepole pine U S M L O Opening type U S M L O Opening type Figure 4.2 continued. 94 Western hemlock Paper birch U S M L O Opening type E >- 10 00 c CD TD C D C T5 Qi CD 00 U S M L O Opening type Western redcedar U S M L O Opening type Hybrid spruce U S M L O Opening type Figure 4.3. Average density of natural regeneration in mature experimental blocks by opening type and year of assessment (U= forest understory; S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ; 0= full open; ^ = end of 2nd year after logging; EH = end of 3rd year after logging). Error bars represent 1 standard error of the mean. 95 Subalpine fir Lodgepole pine U S M L O Opening type 0.0 U S M L O Opening type Trembling aspen Black cottonwood U S M L O Opening type U S M L O Opening type Figure 4.3 continued. 96 Western hemlock U S M L O Opening type Western redcedar U S M L O Opening type Amabi l is fir U S M L O Opening type Hybrid spruce U S M L O Opening type Figure 4.4. Average density of natural regeneration in the old-growth experimental block by opening type and year of assessment (U= forest understory; S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ; 0= full open; ^ = end of 2nd year after logging; LZ1 = end of 3rd year after logging). Error bars represent 1 standard error of the mean. 97 After 3 years, there was very little evidence that gap size affected natural regeneration density (Table 4.1). Among individual species, only western hemlock showed a slight effect of gap size on regeneration density (p=0.09); for other species p-values were >0.26 (Table 4.1). Gap position effects were also weak and depended mostly on whether or not the gap edge position was included and, interestingly, whether or not the old-growth experimental block was included in the analysis (results not shown). The trends held true when each species was examined individually. For the 3 positions inside a gap, the only pairwise comparison found to be significant was higher western hemlock density in the south compared to north position (Table 4.1). For most of the tree species (except subalpine fir, lodgepole pine and trembling aspen), density was significantly lower in the gap edge position compared to the south gap position (Table 4.1 and Fig. 4.5). Several characteristics of natural regeneration density were consistent in both measurement years (only 3rd year shown, Fig. 4.5). For all species except trembling aspen, seedling density increased from the north to the south end of gaps (especially in large and medium gaps). Density in the gap edge position was consistently low regardless of gap size, but for some species (subalpine fir and aspen) it was similar to that inside small gaps. In general, gap position was least important in small gaps (up to 300 m2). The effect of gap position was least consistent for trembling aspen, perhaps because of the root suckering of this species. Compared to gaps, the understory of the undisturbed forests and clearcuts had low densities of natural regeneration (Figs. 4.2-4.4). There was little difference in the amount of regeneration in clearcuts and the undisturbed forest after 3 years (all experiment blocks, p=0.21; mature blocks, p=0.17), but differences in the abundance of individual species between clearcuts and the forest understory were highly significant (p<0.007 in the mature experimental blocks). Table 4.1. A) Randomized block split-split-plot A N O V A p-values from natural regeneration study and B) Each tree species analyzed separately. P-values with asterisks are significant at p<0.05. To control alpha inflation, p-values of pairwise comparisons had to be <0.0167 for size class or <0.0125 for gap position to be deemed significant (a level 0.05/riumber of comparisons). A) Observed p-values from split-split-plot ANOVA Gap edge plots removed Gap edge plots present All blocks Mature blocks All blocks Mature b Size class 0.2985 0.1217 - -Position 0.1343 0.0238** 0.0072** 0.0004** Size class*Position 0.1514 0.0229** 0.1415 0.0078** Species 0.0009** 0.0001 ** 0.0008** 0.0001** Size class*Species 0.2796 0.0081** 0.2867 0.0218** Position*Species 0.2277 0.0001** 0.0002** 0.0001** Size class*Position*Species 0.3178 0.0003** 0.2141 0.0001 ** B) Pairwise contrasts for individual tree species (mature blocks only) Gap edge understory position not included Redcedar Hemlock Subalpine fir Spruce Pine Birch Aspen Size class 0.5112 0.0886 0.6250 0.9245 0.4297 0.2574 0.8475 Position 0.4578 0.0196** 0.2805 0.3917 0.2785 0.1663 0.7870 Size class*Position 0.9860 0.0336** 0.5510 0.3854 0.7512 0.3368 0.4568 Large vs. Medium 0.2946 0.0375 0.3647 0.7897 0.2207 0.2704 0.8561 Medium vs. Small 0.5691 0.0791 0.5049 0.9883 0.3825 0.1201 0.6057 Large vs. Small 0.4408 0.3252 0.7383 0.7264 0.5575 0.4929 0.7090 Gap edge understory position included Position 0.0076** 0.0010** 0.3305 0.0481** 0.2904 0.0105** 0.7857 Size class*Position 0.1010** 0.0097** 0.7374 0.4617 0.8495 0.3663 0.4921 Gap edge vs. South 0.0019** 0.0002** 0.7753 0.0111** 0.1510 0.0028** 0.9164 Middle vs. North 0.4875 0.1372 0.9356 0.5084 0.3145 0.1611 0.7918 Middle vs. South 0.4783 0.0145 0.1552 0.4678 0.3967 0.6436 0.6286 North vs. South 0.1855 0.0022** 0.1742 0.1899 0.0911 0.0822 0.4620 99 Western hemlock S M L Gap size class Paper birch S M L Gap size class Western redcedar Hybrid spruce S M L S M L Gap size class Gap size class Figure 4.5. Average density of natural regeneration by gap size class (S= small gaps, 10-300 m M= medium gaps, 301-1000 m 2; L= large gaps, 1001-5000 m 2 ) and gap position (from left to right, Gap edge= ^ , South= ^ , Middle= 1=1, North= d ) . Error bars represent 1 standard error of the mean. 100 Subalpine fir Lodgepole pine S M L Gap size class S M L Gap size class Trembling aspen S M L Gap size class Figure 4.5 continued. 101 Hemlock was the exclusive dominant in the forest understory (0.53 seedlings per m 2) whereas in clearcuts, 4 species were present: paper birch, hemlock, trembling aspen and black cottonwood (1.6, 0.13, 0.13 and 0.07 seedlings per m , respectively). Deciduous species were dominant over conifers in the clearcut regeneration plots. With the exception of western hemlock, lodgepole pine and trembling aspen, seedling density in gaps was lower 3 years after logging than after 2 years. For example, paper birch density dropped by 30-50% in large, medium and small gaps but remained constant in the clearcut units (Fig. 4.3). Trembling aspen density increased with time in clearcuts. 4.3.2 P l a n t e d s e e d l i n g s , r e g r e s s i o n a n a l y s i s There were strong and consistent trends in growth response among the tree species as gap size increased. Fifth-year height and diameter, and height and diameter increments averaged over the last 3 years (recent growth) increased rapidly from small single-tree gaps to about 1000 m gaps, but thereafter showed little change in gaps up to 5000 m 2 (Figs. 4.6-4.9). These trends were also present at the end of the third growing season (data not presented). Gap size and north-south line length explained roughly equal amounts of variation in 5-yr old seedling size (R 2 = 0.40 - 0.72; Table 4.2), but gap size was a slightly better predictor of recent growth rates than north-south line length (R 2 = 0.50 - 0.70; Table 4.3). Predictive models of 5th-yr size and 3-yr average growth rates are presented in Figures 4.10 - 4.13. There was little statistical basis for declaring either gap size or length of a gap's north-south line as the superior independent variable for predicting tree growth. However, the shapes of the curves derived from the two approaches provide somewhat different interpretations. For both total size and growth rates, the gap size curves asymptoted near or just above 1000 m 2 , indicating little growth advantage as gap size increases above 0.1 ha (Figs. 4.14 - 4.15). The north-south line 102 Western redcedar Western hemlock 250 250 0 1000 2000 3000 4000 5000 Gap size (m ) Subalpine fir 250 E" 200 o 150 CD 50 0 0 1000 2000 3000 4000 5000 Gap size (m2) 0 1000 2000 3000 4000 5000 Gap size (m ) Hybrid spruce 0 1000 2000 3000 4000 5000 Gap size (m2) Lodgepole pine 250 £f 200 0 1000 2000 3000 4000 5000 Gap size (m2) Figure 4.6. Mean seedling height five years after planting vs. gap size for the five tree species. The fitted regression line use Equation [1] and parameter values in Table 4.2. 103 Western redcedar 501 1 1 1 1— | 40 -co 0 1000 2000 3000 4000 5000 Gap size (m2) Western hemlock 501 1 1 1 1-| 40 -0 1000 2000 3000 4000 5000 Gap size (m2) Subalpine fir 501 1 1 1 r | 40 -CD % 30 -CO T3 0 1000 2000 3000 4000 5000 Gap size (m2) Hybrid spruce 501 1 1 1 r 0 1000 2000 3000 4000 5000 Gap size (m2) Lodgepole pine Figure 4.7. Mean seedling diameter five years after planting vs. gap size for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.2. 104 Subalpine fir 60 r Hybrid spruce 60 rFigure 4.8. Mean seedling height growth rates (average per year for last 3-yrs) vs. gap size for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3. 105 Subalpine fir 15 £ 1 0 5 o S 5 C i i i 1 — 0 1000 2000 3000 4000 5000 Gap size (m2) Hybrid spruce 0 1000 2000 3000 4000 5000 Gap size (m2) Lodgepole pine 151 1 1 1 r E 0 1000 2000 3000 4000 5000 Gap size (m2) Figure 4.9. Mean seedling diameter growth rates (average per year for last 3-yrs) vs. gap size for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3. O H - l C o 3 cr CD (3 a 3 X J c cd J3 . OJ S j3 W 3 S bO • .—I -4—» -*—» O 3 O 1/3 X J 4J & -5 a § CD O .1 s CD CD N CD '3 £ M s •§ cd (3 H CD 3 CL , Cu O . C/J •9| D) 0 >_ co 0 LL o CD o ca CO a> "5 0 Q. w CM O Tl- CM O O O) 00 O) C) co i - oo I S- TI-T]- LO CO CO CD o d d o d CM Tf _ CO LO CM Tf LO O CD Tf Is-CD CO o o Is-CM CD oo oo T- Tf I s- co I LO CM CM CM cn co co o T- CD Tf CO i -CM T - - i -CM CD O LO LO LO Tf 00 co o> Tf cy> co LO Is— CD O ) Tf O ) Tf T— CJ) C3) CO LO 00 CO CD O O Tf - i - LO Tf LO CD CD CD CD CD CD CD CD CO Tf Tf LO O CO - i - CM CO LO CJ) CO - i - O CM 0 0 0 ^ 0 00 CJ) CO LO 00 l - O) O i - CM C O C O I O O ) r ; c\i T— d d T -CM CO LO - i - CO CJ) O ) LO O Tf q co IO N q C O O ) ^ N O ) O LO CM Tf T - CO N t - CD i -LO Tf CO Tf T— CM CO CD T - T f 3 o OJ -S2 CJ " X J o E -CD ® i - SZ QJ E E -CD CD -5" W « o Q) CD "S 5 5 co CD c; C L _0 o C L CD O ) CD o 3 1_ C L CO X J a> CD E ro xi CO a> tz i i o o o CO o CO CD o (D Q. CO CM o Tf CM o CJ) CJ) 00 O) O) Is- Tf Is- CD LO LO LO CD CD CD d d d d d CO Is- CM LO o Tf LO Tf LO I s-CO co CM CM CM d d d d d CJ) Tf CO CM CM CD Tf co CO Tf CO CM Tf O O) Is- O) Is- Is- Tf Tf LO LO CD O) o Is-CO Tf LO od T _ co o CD CD Is- O) O) o 00 Tf LO r-^  d d CM CM CO Tf O Tf I s - CM CD CO CD CD Is-d d d d d LO CM oo LO CM CO CO CM CM CM o o O o o d d d d d Is- CM CO CO co co T— T— o CJ) T— 1— 1— o CD d d d d O) Tf co o oo LO Is- o Tf LO O) o O) LO T ~ o ••- CM O) CM Tf O) o q Is- oq CM d Tf CM O) CM CM CM CM * 8 CD — ° E 0 0 .E c c .g 0 0 8 S | 5 5 co C L X J 0 £.3 CO 4—1 S3 o i — i t-i i CO 4 - » D >> V-i P. g O ISO <-i fl rt C VH --3 4) ^ > X5 « J -a •*—» rt X ! •4—» O S-H S-i <D 4—» e. rt •5 T3 . C rt O fl an c -1—1 e CD OH O T3 S G G O rt O H § H . <u <u X 5 D i-i CD XJ 4^ C/) 15 <u c E o w CO Q) >. 4-< CO CS CS CD >. i_ 0) Q. CD O) (0 t_ CD > CU 4—' cs 5 o D) .SP "55 CD IO 3 o CO X c a, l CD l o m CT) cc IO m cn CD N CO c 'g! CL I O o Lf) 05 to CD "o CD Q. CO CM O - t f CM O O ) O I CO G ) O ) •tf o co o h~ LO tf- co co o o d d o co i n CM O) co co co CM co o CM LO - i - O CO i - CM N -eo co CM CO T - CM i n tf-•tf cri • t f CO CO CM O CM oo co -tf CM CO LO CO O CM CO 3 8 CD — ° F "8 % O O O O CM i - i - CT) CO • t f LO CM CT) CT) LO tf- CO CM o ^ o o o CD LO S! " O) O) N J ° LO CO CM CO CM 00 LO CT) -tf ( D LO N S N CO CO -•- CM CD CM tf" - i - i - CM O O O O O d d c i d d CM O) T— CD CO D5 CO -tf -tf l -O T - O O i -o o o o o CO O) CO CT) CD t— CO CO CM CO CM CT) CD LO CM O tf- tf" "tf S CV1 CM CO CD h~ CO CM CO czCD CD CO v , CD CD CD CD C S Q-l ^ -s s & <o -R -e ^ -> =J >> o > (O I J E E to" _^ CO a> >. CO 4-> CO cs co CD >. CD Q. CD U) co i_ CD > w CD CO o D) i _ CD CD E co 5 ST _ i E, d CD C LO cn X o CO CO X •c o H — o _ i X O 0> c CD LO _ l CT) CC CC CD N CO CJ) ' CD Q . O o CO CD "o CD Q. cn CM o • t f CM o C35 CT) oo CT) CT) CO ,_ cn CM LO LO LO CD CD d d d d d CM CD O oo 'cr LO CO • t f • t f O o o o o d d d d d CO CM o LO CD o • t f CD T— T— i — T — o o o o o 00 CO • t f CM CD r» 1^ o oo co CO CT) C\i CM CO LO 00 OJ CM CT) 00 o • t f CT) CD LO • t f CD CM 1^ CO • t f CM CT) CM CD CO CD CD d d d d d • t f LO CO CO • t f o o o o o o o o o o d d d d d CO CM oo i — CM T— o i— o o o o O d d d d d CD • t f CD CD CM LO CO o o d d d d co LO 00 LO • t f o r»- tf- CT) • r -CD T— o CO co • t f LO CO LO 1^ CJ o CO T 3 CD CJ "D CD C C 0 0) U) 00 0 CD 5 5 CD .!= O ZD CD .E co CD ~ X X 3 > . IV) X 108 Lodgepole pine 2501 1 1 1 1 If 200 -o 0 22 44 66 88 110 North-south line length (m) Figure 4.10. Mean seedling height five years after planting vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.2. 109 Western redcedar 501 1 1 1 1— | 40 -0 22 44 66 88 110 North-south line length (m) Western hemlock 501 1 1 1 1-| 40 -North-south line length (m) Subalpine fir 501 1 1 1 1 40 -CD 0 22 44 66 88 110 North-south line length (m) Hybrid spruce 501 1 1 1 | 40 -0 22 44 66 88 110 North-south line length (m) Lodgepole pine 501 1 1 1 1 i 4 0 -0 22 44 66 88 110 North-south line length (m) Figure 4.11. Mean seedling diameter five years after planting vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.2. 110 Western redcedar 601 1 1 1 1— o 40-0 22 44 66 88 110 North-south line length (m) Western hemlock 601 1 1 1 1— North-south line length (m) Figure 4.12. Mean seedling height growth rates (average per year for last 3-yrs) vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3. I l l Western redcedar 15 | 1 1 1 r— E 0 22 44 66 88 110 North-south line length (m) Western hemlock 15i 1 1 1 1-E North-south line length (m) Lodgepole pine 151 1 1 1 1 E 0 22 44 66 88 110 North-south line length (m) Figure 4.13. Mean seedling diameter growth rates (average per year for last 3-yrs) vs. north-south line length for the five tree species. The fitted regression lines use Equation [1] and parameter values in Table 4.3. 112 250 200 150 CD 2 100 50 a I I I -^_=9—©—&- 9 9 ft=K £ ^ 1> 1 c t> c* C __a____-B B B B B B f i - . f i o o o © © * L - 1 1 1 250 0 1000 2000 3000 4000 5000 Opening size (m2) 0 22 44 66 88 110 North-south line length (m) Figure 4.14. Predicted fifth year height (cm) as a function of (a) opening size (m2) and (b) north-south line length (m); and predicted fifth year diameter (mm) as a function of (c) opening size (m2), and (d) north-south line length (m), using Equation [1] and parameter values in Table 4.2. 113 1000 2000 Opening size (m 3000 4000 2x 5000 15 E E 10 o B c5 E CO Q o Subalpine fir > Western redcedar O Western hemlock o Lodgepole pine • Hybrid spruce 22 44 66 88 North-south line length (m) 110 Figure 4.15. Predicted height growth rate (cm, average per year for the last 3-yrs) as a function of (a) opening size (m2) and (b) north-south line length (m); and predicted diameter growth rate (mm, average per year for the last 3-yrs) as a function of (c) opening size (m2), and (d) north-south line length (m), using Equation [1] and parameter values in Table 4.3. 114 curves, especially average recent growth, do not have as clear an upper limit as line length increases. 4.3.3 P l a n t e d s e e d l i n g s , a n a l y s i s o f v a r i a n c e The split-split-plot A N O V A demonstrated that gap size, seedling position inside or outside a gap, tree species involved and the interactions among size, position and species all influence seedling size or current growth rates (Table 4.4). The observed p-values presented in Table 4.4 were based on analyses with and without gap edge understory trees. Although the presence of gap edge trees had very little effect on gap size class results, all future discussion of the influence of gap size class are based on trees inside gaps only. For all tree species, mean size after 5 years and average 3-yr growth was strongly affected by gap size (large, medium, small) (p<0.004) (Tables 4.5 and 4.6). There was a consistent decrease in mean size, with diameter responding more dramatically than height, from large to medium to small gap size classes. Differences among size classes were greater as seedling shade-tolerance decreased (from redcedar to pine) (Figs. 4.16-4.19). With the exception of redcedar, 5th-yr size of all species was significantly greater in large gaps than medium gaps (Table 4.5). Similar differences existed between medium and small gaps, except that height and diameter of redcedar and hybrid spruce were not significantly different between the two size classes. A l l tree species were larger and growing more rapidly in large than in small gaps (Tables 4.5 and 4.6). Three-year average height and diameter growth of western redcedar varied little between large and medium and between medium and small gaps (Table 4.6). Hemlock height growth was similar in large and medium gaps. Seedling height was generally greatest in open conditions (but see subalpine fir in Fig. 4.16). However, for all species except lodgepole pine (p=0.0012), 5-yr height was not 115 Table 4.4. Randomized block split-split-plot A N O V A for planted seedling study. Observed p-values for gap size class (main plots), gap position (first-split), tree species (second-split) and their interactions. Analysis is presented twice: first, without gap edge understory trees so only trees inside gaps are being considered; second, with gap edge understory trees included, thus tests include trees both inside and outside the gap. Gap edge understory position not included 5th yr height Height growth 5th yr diameter Diameter growth (cm) (cm/yr, last 3-yr) (mm) (mm/yr, last 3-yr) S o u r c e o f v a r i a t i o n Size class 0.0001 0.0001 0.0001 0.0001 Position 0.0065 0.0146 0.0035 0.0054 Size class*Position 0.0373 0.0183 0.0303 0.0695 Species 0.0001 0.0001 0.0001 0.0001 Size class*Species 0.0001 0.0001 0.0001 0.0001 Position*Species 0.0459 0.0355 0.0258 0.0238 Size class*Position*Species 0.0855 0.0633 0.0139 0.0079 Gap edge understory position included S o u r c e o f v a r i a t i o n Size class 0.0002 0.0002 0.0001 0.0002 Position 0.0001 0.0001 0.0001 0.0001 Size class*Position 0.0003 0.0002 0.0001 0.0003 Species 0.0001 0.0001 0.0004 0.0009 Size class*Species 0.0001 0.0001 0.0001 0.0001 Position*Species 0.0056 0.0033 0.0067 0.0054 Size class*Position*Species 0.0557 0.0849 0.0168 0.0140 116 Table 4.5. Observed p-values from a randomized block split-plot A N O V A on seedling size. Each tree species was analyzed separately. P-values with asterisks are significant at p<0.05. To control alpha inflation, p-values of pairwise comparisons had to be <0.0167 for size class or <0.0125 for gap position to be deemed significant (a level 0.05/number of comparisons). Gap edge understory position not included Western Western Subalpine Hybrid Lodgepole redcedar hemlock1 fir1 spruce pine 5th year height Size class 0.0036** 0.0001** 0.0001** 0.0001** 0.0001** Position 0.0144** 0.0657 0.0415** 0.0271** 0.0126** Size class*position 0.2898 0.2483 0.1018 0.5226 0.0066** Large vs. medium 0.1137 0.0117** 0.0055** 0.0004** 0.0025** Medium vs. small 0.0237 0.0041** 0.0012** 0.0168 0.0010** Large vs. small 0.0013** 0.0001** 0.0001** 0.0001 ** 0.0001** 5th year diameter Size class 0.0011** 0.0002** 0.0001** 0.0001** 0.0001** Position 0.0197** 0.1233 0.0800 0.0101** 0.0039** Size class*position 0.2910 0.3360 0.4093 0.0601 0.0037** Large vs. medium 0.0185 0.0073** 0.0016** 0.0007** 0.0006** Medium vs. small 0.0233 0.0052** 0.0020** 0.0477 0.0068** Large vs. small 0.0004** 0.0001** 0.0001** 0.0001** 0.0001** Gap edge understory position included 5th year height Position 0.0051** 0.0657 0.0415** 0.0016** 0.0010** Size class*position 0.0580 0.2483 0.1018 0.0077** 0.0036** Gap edge vs. south 0.0321 - - 0.0091** 0.0092** Middle vs. north 0.0027** 0.0521 0.0251 0.0258 0.2416 Middle vs. south 0.0240 0.0388 0.0311 0.0056** 0.0022** North vs. south 0.2104 0.8487 0.8983 0.3581 0.0149 5th year diameter Position 0.0047** 0.1233 0.0800 0.0007** 0.0005** Size class*position 0.0179** 0.3360 0.4093 0.0028** 0.0007** Gap edge vs. south 0.0243 - - 0.0059 0.0098 Middle vs. north 0.0060** 0.1131 0.0549 0.0114** 0.0476 Middle vs. south 0.0114** 0.0631 0.0487 0.0021** 0.0006** North vs. south 0.7082 0.7066 0.9225 0.2907 0.0164 'Because of inadequate sample sizes comparisions of hemlock and subalpine fir in south vs. gap edge positions was not possible. 117 Table 4.6. Observed p-values from a randomized block split-plot A N O V A on seedling growth rates. Each tree species was analyzed separately. P-values with asterisks are significant at p<0.05. To control alpha inflation, p-values of pairwise comparisons had to be <0.0167 for size class or <0.0125 for gap position to be deemed significant (a level 0.05/number of comparisons). Gap edge understory position not included Western Western Subalpine Hybrid Lodgepole redcedar hemlock1 fir1 spruce pine Height growth (cm/yr, last 3-yr) Size class 0.0024** 0.0002** 0.0001** 0.0001** 0.0001** Position 0.0367** 0.0513 0.0309** 0.0210** 0.0271** Size class*position 0.1808 0.3512 0.1281 0.3662 0.0079** Large vs. medium 0.0734 0.0250 0.0035** 0.0003** 0.0019** Medium vs. small 0.0186 0.0052** 0.0007** 0.0208 0.0011** Large vs. small 0.0008** 0.0001 ** 0.0001** 0.0001** 0.0001** Diameter growth (mm/yr, last 3-yr) Size class 0.0012** 0.0003** 0.0001** 0.0002** 0.0001** Position 0.0300** 0.1405 0.0724 0.0115** 0.0070** Size class*position 0.3832 0.3173 0.5884 0.1030 0.0049** Large vs. medium 0.0177 0.0096** 0.0051** 0.0008** 0.0013** Medium vs. small 0.0284 0.0085** 0.0046** 0.1023 0.0145 Large vs. small 0.0004** 0.0001** 0.0001 ** 0.0001** 0.0001** Gap edge understory position included Height growth (cm/yr, last 3-yr) Position 0.0078** 0.0004** 0.0002** 0.0008** 0.0018** Size class*position 0.0406** 0.2040 0.0179** 0.0064** 0.0032** Gap edge vs. south 0.0220 - - 0.0053** 0.0110** Middle vs. north 0.0060** 0.1660 0.0065** 0.0182 0.1744 Middle vs. south 0.0572 0.0295 0.0217 0.0035** 0.0045** North vs. south 0.2037 0.7396 0.5062 0.3209 0.0466 Diameter growth (mm/yr, last 3-yr) Position 0.0072** 0.0010** 0.0003** 0.0007** 0.0009** Size class*position 0.0195** 0.0980 0.2721 0.00514** 0.0018** Gap edge vs. south 0.0313 - - 0.0070** 0.0119 Middle vs. north 0.0067** 0.1121 0.0357 0.0084** 0.0841 Middle vs. south 0.0198 0.0533 0.0271 0.0019** 0.0014** North vs. south 0.3323 0.6751 0.8570 0.3375 0.0285 'Because of inadequate sample sizes comparisions of hemlock and subalpine fir in south vs. gap edge positions was not possible. 118 Western redcedar Western hemlock 250 250 U S M L O Opening type U S M L O Opening type 250 E 200 o .gi 150 CD S 1 0 0 >. ? 50 Subalpine fir ~i 1 r U S M L O Opening type 250 E 200 o Hybrid spruce u s M L o Opening type Lodgepole pine i r 5° S M L O Opening type Figure 4.16. Mean seedling height (cm) five years after planting by opening type (U= forest understory; S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ; 0= full open). Subalpine fir were not planted in the forest understory. Error bars, +1 SE of the mean. 119 Western redcedar 601 1 1 1 1 r If S 40-E U S M L O Opening type Western hemlock 60 E 2 40-a> E co Opening type Lodgepole pine 601 1 1 1 1 r E Opening type Figure 4.17. Mean seedling diameter (mm) five years after planting by opening type (U= forest understory; S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ; 0= full open). Subalpine fir were not planted in the forest understory. Error bars, ±1 SE of the mean. 120 Western redcedar 601 1 1 1 1 r &_ I 40 -Opening type Western hemlock 601 1 1 1 1 r o 40 h Opening type Lodgepole pine 60 U S M L O Opening type Figure 4.18. Height growth rate (cm, average per year for last 3-yrs) by species and opening type (U= forest understory; S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ; 0= full open). Subalpine fir were not planted in the forest understory. Error bars, ±1 SE of the mean. 121 Western redcedar 201 1 1 1 1 1-o 10-Opening type Western hemlock 201 1 1 1 1 r Opening type Subalpine fir Hybrid spruce u s M L o Opening type S M L Opening type Lodgepole pine u s M L o Opening type Figure 4.19. Diameter growth rate (mm, average per year for last 3-yrs) by species and opening type (U= forest understory; S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ; 0= full open). Subalpine fir were not planted in the forest understory. Error bars, ±1 SE of the mean. 122 significantly different between large gaps and clearcuts. Fifth-yr diameter, however, was significantly greater in clearcuts than large gaps (p<0.0021), for all species but subalpine fir (p=0.11), with lodgepole pine having the greatest differential between open conditions and large gaps (Fig. 4.17). Diameter growth rates of all species showed a steady decline from open conditions to large, medium and small gaps and to the forest understory (Fig. 4.19). This was especially true for the two most light-demanding conifers, lodgepole pine and hybrid spruce. For pine, 3-yr diameter growth rates in open conditions were more than twice that found in large gaps (Fig. 4.19), and height growth rates behaved similarly (Fig. 4.18). Diameter growth of all species except subalpine fir (p=0.2) was greater in the open condition of the clearcuts compared to large gaps (p<0.0025). Height growth of lodgepole pine and hybrid spruce was significantly greater in clearcuts than large gaps (p<0.02); there was little difference in height growth among the other 3 species between clearcuts and large gaps (p>0.16). In the forest understory all species performed poorly and after 2 years all lodgepole pine had died. Fifth-yr height and diameter of redcedar and hemlock was significantly greater in small gaps than in the forest understory (p<0.015). In the last 3-yr, only average hemlock height increment was significantly greater in small gaps than the forest understory (p=0.0127). Neither total size or recent growth of spruce varied between small gaps and the forest understory (p>0.053). Subalpine fir was consistently the poorest performer in all opening types. Planting stock conditions were probably somewhat responsible. At planting, subalpine fir was the smallest of the 5 species (see Table 3.1). This species has a history of poor outplanting performance. After planting, terminal buds often do not flush, resulting in poor apical dominance for the first few 123 years (Pinkerton 1994), a phenomenon observed in this experiment. Many of the stock problems with subalpine fir have recently been resolved (Pinkerton, personal communication, 1998) For these reasons, I do not believe the results presented here should be considered indicative of current subalpine fir stock performance during the first few years after partial cutting. Western hemlock planting stock was also problematic. Many seedlings died during the first summer for no apparent reason and this trend continued for several years. Mortality of the three species that exibited no planting stock problems (redcedar, spruce and pine) increased as gap size decreased, with the exception of high mortality of redcedar in open conditions (Table 4.7). Table 4.7. Percent of seedlings that have died (mortality) in the forest understory, in small, medium and large gaps, and in the open condition of clearcuts. Subalpine fir was not planted in the forest understory due to sortage of stock. Opening type Forest Small Medium Large Clearcut Second year mortality (%f gaps gaps gaps Western redcedar 3.2 0.0 0.0 0.9 13.3 Western hemlock 25.0 19.3 10.3 11.3 10.9 Subalpine fir 5.3 2.7 1.5 0.0 Hybrid spruce 48.4 2.0 0.0 0.0 0.0 Lodgepole pine 100.0 6.1 2.3 0.7 0.0 Five year mortality (%f Western redcedar 17.5 1.2 1.0 2.8 35.6 Western hemlock 48.4 26.2 16.8 20.6 20.0 Subalpine fir 13.3 9.2 5.7 0.0 Hybrid spruce 65.6 6.9 2.8 2.7 0.0 Lodgepole pine 100.0 13.3 3.5 4.6 1.7 'Reflects mortality from fall 1993 to fall 1994; seedlings planted in spring 1993, measurement trees selected in fall 1993. There are no data available on first growing season mortality. Reflects mortality from fall 1993 to fall 1997. Growth of the three most shade tolerant species (western redcedar, western hemlock and subalpine fir) was consistent by gap position in large, medium and small gaps (Table 4.5). Hybrid spruce height and diameter showed no interaction between gap size and gap position 124 inside gaps, but a weak interaction was evident when gap edge seedlings were included. Lodgepole pine exhibited a strong gap size by gap position interaction both with and without gap edge seedlings (p<0.007; Table 4.5), because of poor performance in the south position of large gaps. The poorest seedling performance was usually in gap edge positions, although in small gaps (and for some species in medium gaps) the edge positions differed little from other positions (Figs. 4.20 and 4.21). Medium gaps had the greatest variability in growth with gap position. Somewhat surprising were the poor performances of all species (except pine height) at the north end of medium gaps and the clear superiority of the middle position in the medium gaps, suggesting strong interference from edge canopy trees, even those to the north of the gap. Gap position had little influence on western hemlock and subalpine fir but was important for the 3 other species, which performed well at the middle position, but poorly at the north and south ends. Fifth year height and diameter of all species varied little between north and south positions. Even though the 5th-yr height of all species was best in the middle position, only western redcedar height was significantly better in the middle than north position (Table 4.5). Height of hybrid spruce and lodgepole pine in the middle position was superior to that in the south position. Seedling diameter was more sensitive than height to gap position. Both redcedar and hybrid spruce had larger 5-yr diameters in the middle vs. north or south positions (Table 4.5). Lodgepole pine diameter was similar in middle and north gap positions, but 31% greater in middle than south position. Height and diameter growth rate responses were similar to those observed for total size (Figs. 4.22 and 4.23). Again, there was a tendency for the best growth to be in the middle gap 125 position compared to north and south positions, but in many cases the differences were not great enough to be significant (Table 4.6). A concern of silviculturists prescribing tree species in partially cut forests is which species is best suited to which conditions. Excluding lodgepole pine, which excelled over all other species in full open conditions, a striking feature of these early results (after 5 years) was the remarkable similarity in mean size of all the species in each of the 5 opening types (full open, large, medium and small gaps and the forest understory) (Figs. 4.24 and 4.25). Subalpine fir, because of low initial height and poor stock quality was consistently 9-14 cm shorter than hemlock, spruce and pine. Average yearly growth rates over the last 3 years were similar, with pine outgrowing all other species in open conditions (Figs. 4.26 and 4.27). Western redcedar grew more slowly than hemlock and spruce in the open and was similar in total height after 5 years in clearcuts primarly because of its 14 - 17 cm initial height advantage (Table 3.1). In gaps, height and diameter growth of all species was fairly similar with slightly more variability in height increment than in diameter increment. The equality of species was most striking in the medium and small gaps, especially for diameter increment (Fig. 4.27). Lodgepole pine, a species considered too shade-intolerant for partial cutting prescriptions, grew as well or better than the other species in all gap sizes and gap positions. In the forest understory, height and diameter growth of western redcedar, western hemlock and hybrid spruce were equally poor. 126 Western redcedar S M L Gap size class Western hemlock 250 Gap size class Subalpine fir 250 S M L Gap size class Hybrid spruce S M L Gap size class Lodgepole pine 250 o, 200 Z> 150 CD SZ a 100 CD >> £ 50 «+— LL 0 -i 1 1 1 | 1| 1 +1 -S M L Gap size class Figure 4.20. Mean seedling height five years after planting by gap size class (S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ) and gap position (from South= , Middle= [WW, North= • ) . Error bars, ±1 SE of the left to right, Gap edge= mean. 127 Western redcedar Western hemlock S M L Gap size class _ 40 E E, oj 3 0 OJ E .« 20 T3 K CO 5. 10 L I to S M L Gap size class Subalpine fir S M L Gap size class Lodgepole pine S M L Gap size class ~ 40 E a 3 0 OJ E .« 20 T3 CO Si. 10 Hybrid spruce ft • ft ft S M L Gap size class Figure 4.21. Mean seedling diameter five years after planting by gap size class (S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2; L= large gaps, 1001-5000 m 2 ) and gap position (from left to right, Gap edge= ^ , South= ^ , Middle= § , North= • ) . Error bars, ±1 SE of the mean. 128 Western redcedar CD X S M L Gap size class Western hemlock sz 'CD X S M L Gap size class Subalpine fir S M L Gap size class c o 'CD x Hybrid spruce S M L Gap size class 2 40 CD > CO > * I CO E o o CD 'CD X 30 20 h 10 0 Lodgepole pine fc I m ii S M L Gap size class Figure 4.22. Height growth rate (cm, average per year for last 3-yrs) by gap size class (S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L=large gaps, 1001-5000 rn^and gap position (from left to right, Gap edge= , South= ^ , Middle= fH, North= • ) . Error bars, ±1 SE of the mean. 129 Western redcedar Western hemlock S M L S M L Gap size class Gap size class Subalpine fir S M L Gap size class Hybrid spruce S M L Gap size class Lodgepole pine s M L Gap size class Figure 4.23. Diameter growth rate (mm, average per year for last 3-yrs) by gap size class (S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m 2 ) and gap ^ , South= ^ , Middle= 1=3, North= • ) . Error bars, position (from left to right, Gap edge= ±1 SE of the mean. 130 Undisturbed forest Small gaps 250 CW HW BL SX PL Species 250 200 !§> 150h OJ CW HW BL SX PL Species 250 Medium gaps CW HW BL SX PL Species 250 Large gaps CW HW BL SX PL Species Full open 250 CW HW BL SX PL Species Figure 4.24. Mean seedling height (cm) five years after planting by species and opening type; small gaps, 10-300 m 2 ; medium gaps, 301-1000 m 2; large gaps, 1001-5000 m 2 . Species: CW, redcedar; HW, hemlock; B L , subalpine fir; SX, spruce; PL, pine. Subalpine fir were not planted in the forest understory. Error bars, ±1 SE of the mean. 131 Undisturbed forest 60 E E cu cu E CO T3 CO 0 ) B 40r-20 1 1 wimmk CW HW BL SX Species PL Small gaps CW HW BL SX PL Species Medium gaps CW HW BL SX Species Large gaps 60 B c5 E CO T3 i ro 0 3 L L B 40 20 CW HW BL SX Species • w m PL Full open CW HW BL SX PL Species Figure 4.25. Mean seedling diameter (mm) five years after planting by species and opening type; small gaps, 10-300 m 2 ; medium gaps, 301-1000 m 2; large gaps, 1001-5000 m 2 . Species: CW, redcedar; HW, hemlock; B L , subalpine fir; SX, spruce; PL, pine. Subalpine fir were not planted in the forest understory. Error bars, ±1 SE of the mean. 132 Undisturbed forest CW HW BL SX Species Small gaps CW HW BL SX Species Full open 60 CW HW BL SX PL Species Figure 4.26. Mean seedling height growth (cm, average per year for last 3-yrs) five years after planting by species and opening type; small gaps, 10-300 m 2 ; medium gaps, 301-1000 m 2 ; large gaps, 1001-5000 m 2 . Species: CW, redcedar; HW, hemlock; B L , subalpine fir; SX, spruce; PL, pine. Subalpine fir were not planted in the forest understory. Error bars, ±1 SE of the mean. 133 Undisturbed forest cw H W B L sx Species Small gaps C W H W B L SX P L Species Medium gaps 201 1 1 1 1 C W H W B L SX P L Species Large gaps 201 1 1 1 1— C W H W B L SX P L Species 20 E 15 E, % 10 o i— CT> K_ CD 5 E cc b Full open C W H W B L SX P L Species Figure 4.27. Mean diameter growth (mm, average per year for last 3-yrs) five years after planting by species and opening type; small gaps, 10-300 m 2 ; medium gaps, 301-1000 m 2 ; large gaps, 1001-5000 m 2 . Species: CW, redcedar; HW, hemlock; B L , subalpine fir; SX, spruce; PL, pine. Subalpine fir were not planted in the forest understory. Error bars, ±1 SE of the mean. 134 4.4 Discussion 4,4.1 Natural regeneration Natural regeneration was abundant in single-tree to 0.5 ha gaps in these interior cedar-hemlock (ICH) forests, but far less abundant in the full open conditions of large clearcuts and in undisturbed forest understories. The 9 tree species span a wide range of shade tolerances (Krajina 1969; Krajina et al. 1982; Burns and Honkala 1990; Carter and Klinka 1992; Klinka et al. 1992; Wang et al. 1994; Kayahara et al. 1996; Kobe and Coates 1997), yet there was little evidence of species partitioning (e.g., Ricklefs 1977; Leibold 1995) across the range of gap sizes. Within a species, the density of regeneration was roughly equal from small through large gaps, in both mature (140 yr) and old-growth (300 yr+) forests (Figs. 4.2-4.4). Seedling density of individual species varied widely, reflecting differences in parent tree abundance (Ribbens et al. 1994; Coates et al. 1997; LePage et al., submitted). A lack of evidence for gap partitioning in the regeneration phase of ICH forests is consistent with results obtained by Sipe and Bazzaz (1995) in northeastern and Grey and Spies (1996) in Pacific Northwest U.S. forests. There was also little evidence to support the hypothesis that the regeneration of shade-intolerant species continuously increases as gap size increases (after Brokaw 1985, 1987; Whitmore 1989). Paper birch, one of the least shade tolerant species in ICH forests (Kobe and Coates 1997), had greatest densities in medium sized gaps (300 to 1000 m2) and lower densities in large and small gaps. Hybrid spruce, a moderately intolerant species, showed no change in density from small to large gaps. It is difficult to draw conclusions about the remaining shade intolerant tree species - pine, aspen and black cottonwood - because of low numbers and regeneration characteristics. Aspen regeneration is mainly of root sucker origin while much of the cottonwood regeneration was from stem and branch fragments (P. Bartemucci, personal 135 observation; Haeussler et al. 1990), hence initial distribution of these species is more strongly tied to parent tree location and soil surface disturbance levels than to gap size. Lodgepole pine is a serotinous species whose regeneration is not favoured by partial cutting, but densities may increase over time as cones open on the ground. Nevertheless, there was no evidence that regeneration of any of these species increased with gap size. There was good evidence of intraspecific species partitioning by gap position, especially in the medium and large gaps, where a strong gradient in density was observed from the north (low densities) to the south end of gaps. This result was consistent with those of Wright et al. (1999) where the same species were artificially sown at different gap positions in 600 m gaps. The north-south density gradient was not evident in small gaps (<300 m ) where there was little intraspecific variability in seedling density among gap positions. None of the tree species favoured the gap edge position beneath a tree canopy just outside the north or south end of gaps. In gaps sufficiently large to receive direct sunshine, the shadier southern half had the greatest abundance of young seedlings. Higher rates of germination in shaded microsites has been previously reported for many conifer species (Garman and Orr-Ewing 1949; Day 1964; Alexander 1984; Smith 1986; Burton 1997; Farmer 1997;). Results from this and a companion study (Wright et al. 1999) support the hypothesis that successful regeneration of shade-intolerant species varies with gap position, especially in larger gaps. However, the same is also true of shade tolerant species, a result that conflicts with the hypothesis that shade tolerant tree density will not vary with gap size or gap position. In fact, all ICH species, except trembling aspen, showed a similar response to gap position, occurring in greatest numbers in the south end of gaps. This result suggests that, in gaps at least, lack of sunlight is not the dominant factor limiting natural regeneration. 136 4.4.2 Planted seedlings The five planted conifer species responded in a similar and consistent manner to the range of gap sizes tested in this experiment (single-tree to 0.5 ha). A l l species showed asymptotic growth as gap size increased from single-tree gaps to 0.5 ha. Growth increased rapidly from single-tree gaps up to approximately 0.1 ha gaps, after which there was little further improvement in average growth rates by gap size. Gap size and the length of the longest north-south line in a gap were both good predictors of seedling performance, explaining between 40 and 79% of the variation in 5th-yr size and recent growth rates. Although north-south line length may be a more biologically sensitive measure of gap conditions than gap size in high latitude forests (because of lower sun angles resulting in a north-south light gradient), there was little difference in predictive power of these two variables. From a pratical perspective, north-south line length is much easier to measure than gap size and can provide a quick and inexpensive way to characterize the gap distribution of forests after natural or human disturbance. Contrasting seedling performance in the open conditions of clearcuts to different size classes of gaps and the forest understory revealed results consistent with the regression analysis. Tree size and current growth rates were highest in open conditions, especially for lodgepole pine, but for more shade tolerant species, only marginally more so than in large gaps (0.1-0.5 ha). As gap size decreased below 0.1 ha, growth performance among the tree species became even more similar, to the point where there was little difference in growth by gap position in small gaps (10-300 m2). Seedlings of all species planted in the understory off the gap edge performed poorly regardless of gap size or the cardinal direction into the understory. This suggests that the light advantage experienced off the north end of northern latitude gaps 137 (Canham et al. 1990), compared to other cardinal directions, was not an advantage for growth in these forests. Because of the north-south light gradient in high latitude gaps (Canham et al. 1990; Chapter 2), I expected tree growth rates to be best in the north and middle positions of large gaps and show a strong north to south decline in growth as gap size decreased. Instead, the largest trees of all species were found in the middle gap position of large gaps and there was little difference between the north and south positions, except for lodgepole pine which clearly grew poorly in the shady south position. Lodgepole pine is considered the most shade intolerant of these conifer species (Krajina 1969; Krajina et al. 1982; Kobe and Coates 1997). Overall, these results suggest that below-ground competition from canopy edge trees has an important influence on seedling growth. Few of the hypotheses about shade tolerant and shade intolerant species from tropical and lower latitude forests apply to the growth of planted seedlings in the ICH forests of northwestern British Columbia. Species-specific differences in ranking of height and diameter growth response to gap size or gap position were not sufficient to support a species-partitioning model along the lines proposed by Ricklefs (1977) and Denslow (1980) and still actively debated for tropical forests (e.g., Lieberman et al. 1995; Barker et al. 1997). Brown (1996), in the tropical forests of Borneo, also found no evidence of gap partitioning based on the height ranking among 3 species of dipterocarp across a gap gradient. Sipe and Bazzaz (1995) and Grey and Spies (1996) in northeastern and Pacific Northwest U.S. forests, respectively, were also unable to show evidence of gap partitioning. The deciduous species (trembling aspen, black cottonwood and paper birch) which are the most shade intolerant trees in the ICH, were not included in my planting study. There is no reason to think their inclusion would have changed 138 the conclusion about gap partitioning based on growth rates. In a companion study, the growth rates of the deciduous species broadly overlapped those of the other ICH forest species across a low to high light gradient (Wright et al. 1998). I therefore expect the performance of deciduous trees in gaps to parallel the results reported here for the commonly planted conifers. The regeneration niche and growth niche in ICH forest gaps appear to be "discordant", in the sense of Schupp (1995). Emergence and early survival of natural regeneration was clearly highest in the shaded south end of gaps. Growth of planted trees was highest in the sunnier middle to northern end of gaps. The best location for natural regeneration was not the best for growth of planted trees. This discordance appears to be a widespread phenomena among western North American forest species (e.g., Day 1964; Farmer 1997; Tappeiner et al. 1997). One of the most striking results of this study was the similar growth rates among the five tree species in each of the combinations of gap size and gap position monitored. The only major exception was the superior performance of lodgepole pine in open conditions compared to other species. In contrast, early mortality rates varied considerably among the species, with a trend of increased mortality with decreasing shade tolerance especially in the low light of the forest understory, where average growth rates ranged from 0.3-0.5 mm/yr. Kobe and Coates (1997) also documented wide variation among ICH tree species in sapling mortality rates as a function of recent radial growth. At a radial growth rate of 0.4 mm/yr, Kobe and Coates (1997; equation in Table 1), predict a 3-yr mortality rate for redcedar, hemlock, subalpine fir, spruce and pine of 1, 4, 4, 27 and 51%, respectively, a similar rate to that observed in this study. In ICH forests, gap partitioning may take place gradually in response to the different growing conditions associated with different gap sizes and gap positions. Fifth year results from the planting study indicate differential mortality among the species based on shade tolerance 139 ranking in the shadiest portions of larger gaps, in small gaps and in the forest understory. Gap partitioning is a function of survivorship at low light and hence low growth in these forests. Gap partitioning resulting from seedling mortality rather than a shift in height growth hierachy has also been observed in tropical forests (Brown 1996). 4.4.3 Conclusions and management implications Overall, the ICH tree species are generalists across a wide range of gap sizes, with no one species appearing to regenerate or grow better in a given gap size or gap position. To some extent, this changes at the two extremes for seedling establishment and growing conditions. In large clearcuts natural regeneration is dominated by deciduous species. Growth of shade intolerant species, like lodgepole pine, is clearly superior compared to the more tolerant species. In the forest understory, natural regeneration is restricted to a few species and the low but similar growth rates observed in the understory result in widely differing mortality rates. The partial cutting treatments at Date Creek have produced abundant natural regeneration of a wide range of species in all gap sizes and gap positions. LePage et al. (submitted) have shown that successful natural regeneration following partial cutting results from a combination of favourable seedbed substrates (through minor disturbance to the moss layer) and adequate local density of seed trees. Because natural regeneration in clearcuts was poor and restricted mainly to deciduous species, clearcuts will require planting if the more commercially desired conifers are to dominate. Careful matching of species to position in gaps, keeping the most light demanding species away from the shadier positions, will reduce the risk of early mortality and maximize growth rates in partially cut ICH forests. Opening sizes need not be very large (>0.1 - 0.2 ha) in order for most tree species to achieve growth rates similar to those found in the open conditions 140 of clearcuts. Lodgepole pine performed well in the partially cut forests relative to the other species. Even in medium and small size class gaps, absolute growth of pine was as good as or better than that of the other species. However, pine seedlings may be quite stressed at low light levels by trying to maintain high rates of growth. Coopersmith (personal communication 1998) believes at height:diameter ratios greater than 55, lodgepole pine seedlings are exhibiting signs of light stress indicative of poorly balanced seedlings. In small gaps the height:diameter ratio of lodgepole pine averaged 80. The important criterion upon which to base tree species selection in partially cut forests is the ability to survive until growing conditions improve, especially under low light conditions. At radial growth rates above 2 - 2.5 mm per year the mortality risk for all ICH tree species is quite low (Kobe and Coates 1997). Such radial growth rates are achieved at light levels of 20-40% full sun in these forests (Wright et al. 1998; Chapter 3). 141 Chapter 5 A retrospective analysis of the effects of gap size and gap position on tree regeneration in 34 to 41 year old partially logged northwestern interior cedar - hemlock forests. 5.1 Introduction Development and refinement of silvicultural practice is based on theory, research and largely, long-term observations of forest responses to traditional or trial-and-error manipulations (Weetman 1996). In Canada, only a few long-term silvicultural experiments have been established (e.g., Decie 1957; Glew 1963; Lees 1964; Hughes 1967), and even fewer have been followed for more than a decade or two due to changing priorities and funding problems. Silvicultural systems are a cycle of activities in which a stand is harvested, regenerated and tended over time (Troup 1926; Daniels et al. 1979; Smith 1986; Matthews 1989). Silvicultural systems research is particualarly plagued by a lack of good experimental data because of the long-term nature of such studies and the difficulty of conducting research on what are fundamentally descriptive management systems. Smith (1993) suggests that silvicultural systems themselves are not researchable treatments, but rather that the component parts of the systems should instead be studied. I have argued in Chapter 1 that the most important components of a silvicultural system are the number and size frequency distribution of canopy gaps created after harvest, and how natural regeneration and subsequent tree growth rates are affected by those openings (Coates and Burton 1997). Early results from a designed experiment in the interior cedar-hemlock (ICH) forests of northwestern British Columbia, have shown that establishment of natural regeneration and early growth responses of planted seedlings are affected by gap size and position inside a gap (Chapter 142 4). What happens as gaps age is of considerable interest to forest managers and those trying to understand and predict forest community dynamics. There are three possible approaches to determine whether the early observations reported in Chapter 4 will continue in the future: (1) wait and measure the plots again, (2) use a forest model to predict future conditions (e.g., Pacala et al. 1996, Kimmins et al. 1997), or (3) conduct a retrospective study. Retrospective studies, investigations looking backward from the present to infer causes and their impacts, can provide rapid and useful answers to many management questions, and clarify the issues to be addressed directly through designed experiments (Pickett 1989; Powers 1989). During the 1950's and 1960's extensive partial cutting, primarily for cedar poles used for telephone and electric lines, occurred in the ICH forests. The level of activity varied from individual tree or small group removal, to high-grading where all valuable trees were removed in small patches (0.1-1 ha). This logging activity can be viewed as an unplanned experiment which has created a range of treatments along a gap size continum. The logging has provided a population of gaps suitable for retrospective study. The objective of this study was to retrospectively examine the effect of gap size and position inside a gap on tree species composition, density and growth. More specifically, this study addressed two questions: (1) How does tree species composition, density, and growth vary by gap size and gap position in 34 - 41-year-old gaps? (2) Are the results consistent with those of the designed experiment at Date Creek (Chapter 4)? Based on the establishment and growth of tree seedlings in the forest five years after canopy gap creation (Chapter 4), I expected to find the following patterns in the older gaps: (1) little difference in density of individual tree species by gap size; (2) increasing density of all species from the south to north ends of gaps; (3) gaps dominated by western hemlock and paper 143 birch; and (4) presence of all species in all gap sizes and gap positions, with the possible exception of lodgepole pine. 5.2 Methods The study was established in the moist cold subzone of the Interior Cedar-Hemlock zone (ICHmc), a transitional zone between the interior and coastal areas of northwestern British Columbia. See Pojar et al. (1987) or Meidinger and Pojar (1991) for a description of this system of vegetation classification and Banner et al. (1993) for detailed description of the ICHmc subzone. Low elevation, late successional stands in this subzone are dominated by western hemlock (Tsuga heterophylla (Raf.) Sarg.), in mixture with western redcedar (Thuja plicata Donn. ex D. Don), subalpine fir (Abies lasiocarpa (Hook.) Nutt), lodgepole pine (Pinus contorta var latifolia Engelm.), hybrid spruce [the complex of white spruce (Picea glauca (Moench) Voss), Sitka spruce (P. sitchensis (Bong.) Carr.) and occasionally Engelmann spruce (P. engelmannii Parry ex Engelm.)], paper birch (Betula papyrifera Marsh.), trembling aspen (Populus tremuloides Michx.), and black cottonwood (Populus balsamifera ssp. trichocarpa Torr. & Gray). 5.2.1 Gap selection Areas with old partial logging were identified from local knowledge, forest cover maps and examination of aerial photographs. Two major sampling areas partially logged between 1954 and 1961 were selected: 2 stands in the Kitwanga area (55° 20' N , 128° 10' W, 800 m elevation) and 3 stands in the Muldoe area (55° 25' N , 127° 40' W, 600 m elevation). Transect lines 50 m apart were established in each stand. A l l logging-created gaps (identified by cut stumps) with good canopy closure around the entire perimeter were tallied and canopy gap size determined (as 144 described in Chapter 4). A total of 42 and 44 gaps that met the criteria were found at Kitwanga and Muldoe, respectively and as might be expected, the gap size distribution was skewed to many small and few large gaps. Five gaps greater than 600 m were found and these were sampled. From the remaining population of gaps, 45 gaps were randomly selected, so that there would be roughly 25 gaps per study area. Five gaps were dropped following more detailed inspections. The final sample size consisted of 45 gaps, ranging in size from 16 to 4431 m . 5.2.2 Regeneration within gaps Sample plots inside each gap were established as shown in Fig. 4.1. No sample plots were established outside the gap. If the north-south line through the gap was less than 10 m, plots 1-6 were assessed; i f >10m and <30 m, plots 1-12 were assessed, and if >30m, plots 1-12 and 21-24 were assessed (Fig. 4.1). Two plot sizes were used in the study. If the north-south line was less than 15 m then circular 1 m 2 (0.57 m radius) plots were used. In gaps with longer north-south lines, all trees 9 2 <1.3 m tall were tallied in the 1 nT plots and all taller trees were tallied in 3 m (0.98 m radius) plots. Trees greater than 10 cm tall were assessed, and if a tree was greater than 1/3 the canopy height they were considered part of the canopy and not a gap tree; such trees were rare. Species, total height and diameter (of tree >1.3 m tall) of every tree in each sample plot were recorded. A total of 477 plots and 2102 trees were sampled in this manner. In gaps less than 300 m , the tree considered most likely to dominate (i.e. the tallest and/or best growing tree) in each third of the gap (north, south, middle) was recorded (height, diameter, species). In gaps greater than 600 m , two trees per gap position were measured. These gap-filler trees were thought to be mostly regeneration since logging or small advance 145 regeneration at the time of logging. Obvious residual trees were not selected. About 81% of the trees sampled in the retrospective study established after logging (Chapter 1). 5.2.3 Analysis A randomized block split-plot design was employed to study the influence of 3 size classes of gaps and 3 different gap positions on density and size of each individual trees species. These were the same size classes and positions as those tested in Chapter 4. Experimental blocks were the two main sampling areas - Kitwanga and Muldoe. Main plots were small gaps (10-300 m ), medium gaps (301-1000 m2) and large gaps (1001-5000 m2). The 3 gap positions (north, middle and southern thirds inside gaps) formed the split-plot. An individual gap was considered the experimental unit in this design and individual sampling plots within gaps were sub-samples. Pairwise comparisons were used to test for differences among size classes and gap positions. An adjustment was used to control for Type I error inflation (Milliken and Johnson 1992): the pre-determined oc-level (0.05) deemed significant for differences between pairwise comparisons was divided by the number of comparisons being tested. Thus, an observed p-value had to be less than this pre-set level of significance for the comparison to be deemed significantly different. Three pairwise comparisons were selected to examine the influence of gap size class (large vs. medium, medium vs. small and large vs. small) and gap position (middle vs. north, middle vs. south and north vs. south). For a response variable to be significant between two pairwise comparisons, a p-value of less than 0.0167 was needed. Analyses of variance used the M I X E D Procedure from SAS because the the model being tested included both fixed and random factors (SAS Institute Inc., 1989). Data presented in the figures are lsmeans with standard errors, as calculated by SAS. 146 5.3 Results 5.3.1 Tree density by gap size and gap position A portion of this study was reported on in Chapter 1 (Section 1.4.4). As indicated in Fig. 1.2, only four species were consistently found in all gap sizes sampled: western hemlock, western redcedar, subalpine fir and hybrid spruce. Lodgepole pine and paper birch were rare and trembling aspen and black cottonwood were absent. In the gap size classes examined in this study (small, 10-300 m 2; medium, 301-1000 m 2 ; and large, 1001-5000 m ) there was little evidence of a gap size effect on density of hemlock, redcedar and subalpine fir (Table 5.1 and Fig. 5.1). For hemlock, there was an increase in density as gap size increased, but it was not significant (Table 5.1). Redcedar density was greatest in medium gaps (Fig. 5.1), but again not significantly different from small or large gaps. Hybrid spruce density was significantly affected by gap size (Table 5.1) and was much greater in large gaps than medium or small gaps (Fig. 5.1). Lodgepole pine and paper birch occurred rarely and mainly in larger gaps (Fig. 5.1). None of the tree species varied significantly in density by gap position 34 - 41 yr after logging (Table 5.1), although in large gaps spruce density was much higher in the middle position (Fig. 5.2). In large gaps, hemlock density was highest at the south end of gaps, but this trend was weak in medium and small gaps. There were no consistent trends in redcedar density by gap position. Lodgepole pine and paper birch were found only in the most exposed (northerly) parts of gaps (Fig. 5.2). 147 Table 5.1. Randomized block split-plot A N O V A p-values from retrospective study of tree density (stems/m2) in old logging created gaps in northwestern interior cedar-hemlock forests. P-values with asterisks are significant at p<0.05. To control alpha inflation, p-values of pairwise comparisons of size class and gap position had to be <0.0167 to be deemed significant (a level 0.05/number of comparisons). Tree species Western Western redcedar hemlock Subalpine Hybrid fir spruce Paper birch S o u r c e o f v a r i a t i o n Size class 0.1910 0.1825 0.2958 0.0373** 0.8004 Position 0.7133 0.7147 0.8071 0.0767 0.7458 Size class*Position 0.7737 0.6592 0.7857 0.0110** 0.6549 P a i r w i s e c o n t r a s t s Large vs. Medium 0.2243 0.4320 0.2274 0.0395 0.5658 Medium vs. Small 0.1014 0.2488 0.8852 0.5951 0.7782 Large vs. Small 0.6325 0.1044 0.1646 0.0188 0.6121 Middle vs. North 0.6601 0.8124 0.8679 0.0497 0.6501 Middle vs. South 0.4659 0.6081 0.5717 0.0528 0.8021 North vs. South 0.7391 0.4749 0.6775 0.9018 0.5013 5.3.2 Tree size by gap size and gap position This analysis could be performed on only a few species because of unbalanced numbers of seedlings by experimental block, gap size class and gap position. Also, because of high densities of small trees and relatively low densities of big trees, average height and diameter values may be misleading. Hemlock was by far the most abundant tree species allowing the most comprehensive analysis. Mean height and diameter were determined separately for hemlock trees greater than 0.1, 1, 2, 3 and 4 m tall (no diameter analysis was conducted for trees <2 m tall). In no cases were there significant differences in mean hemlock size by gap size (p>0.23) or position inside a gap (p>0.53). No pairwise comparisons were significant. The 148 Western hemlock Western redcedar S M L Opening type S M L Opening type Subalpine fir Hybrid spruce S M L Opening type S M L Opening type 0.10 Lodgepole pine S M L Opening type 0.10 S 0.02 to 0.0 Paper birch S M L Opening type Figure 5.1. Tree density by opening type (S= small gaps, 10-300 m ; M= medium gaps, 301-1000 m 2 ; L= large gaps, 1001-5000 m2) and species. Error bars represent 1 standard error of the mean. 149 Western hemlock 1 o i 1 1 1 CM Gap size class Western redcedar 1.01 1 1 1 C\J Gap size class Subalpine fir S M L Gap size class Hybrid spruce S M L Gap size class Lodgepole pine Paper birch S M L Gap size class S M L Gap size class Figure 5.2. Tree density by gap size class (S= small gaps, 10-300 m ; M= medium gaps,^ 301 1000 m 2 ; and L= large gaps, 1001-5000 m2) and gap position (from left to right, South = Middle = CH, and North = H). Error bars represent 1 standard error of the mean. 150 same result was found for mean height of all western redcedar taller than 0.1 m. Statistical analysis of other size combinations and species was not possible due to missing data. Hybrid spruce showed a clear pattern of decreasing height with decreasing gap size and better height growth in the middle position than at the north or south ends of large gaps, but showed little difference by position in medium and small gaps. For example, in the middle position of large, medium and small gaps, mean height of all spruce trees taller than 0.1 m was 2.5, 1.3 and 0.25 m, respectively. 5.3.3 Gap filler trees Trees considered most likely to be "gap fillers", that is the best trees in each gap position, did vary significantly in height (p=0.03) and diameter (p=0.04) by gap size. However, there was no indication that gap filler trees varied in size by position inside a gap (p>0.31). Gap filler trees were taller in large gaps (p=0.022) and medium gaps (p=0.047) than in small gaps, and were greater in diameter in large compared to small gaps (p=0.027) (Fig 5.3). Gap filler trees averaged 8.9, 7.2 and 3.1 m in height and 10.6, 8.4, and 3.6 cm in diameter in large, medium and small gaps, respectively (Fig 5.3). Most gap filler trees were western hemlock (84.5%), followed by western redcedar (10%), subalpine fir (3.5%), hybrid spruce (1.5%) and paper birch (0.5%). 5.4 Discussion Trees were abundant and shade tolerant species dominated in gaps created by logging 34 to 41 years ago. There was a trend of increased density in larger gaps, but neither gap size nor position inside a gap had much effect on density of individual tree species. Shade intolerant species showed a preference for larger gaps and the more exposed (northerly) positions inside those 151 Gap s ize c lass 15 Gap s ize c lass Figure 5.3. Height (m) and diameter (cm) of trees most likely to be gap fillers in old partially logged areas by gap size class (S= small gaps, 10-300 m 2 ; M= medium gaps, 301-1000 m 2 ; and L= large gaps, 1001-5000 m2) and gap position (from left to right, South = ^ , Middle = LZ1, and North = § ) . Error bars represent + 1 standard error of the mean. 152 gaps, however, sample sizes to determine density of all intolerant species except hybrid spruce (intermediate in tolerance; Kobe and Coates 1997) were too small to draw reliable conclusions. The hypotheses that have been invoked to explain tree species diversity after small-scale disturbance in lower latitude forests, especially tropical forests (e.g., Denslow 1980; Lieberman et al. 1995), are centred around the existence of distinct niches occupied by individual species or species groups. Thus, over a range of gap sizes, specific species should occur only in certain sized gaps or positions inside a gap. Specialized regeneration niches (Grubb 1977) were not observed across a wide range of gap sizes during early establishment of the ICH forest species (Chapter 4). Unfortunately, in the old retrospective gaps it is not possible to know the breadth of species establishment by gap size and gap position 34 to 41 years ago, although species sorting by gap size is evident now. Hence, the pattern observed may be due to a combination of differences in seed rain, germination, early seedling survival and growth, or inter-tree competition [competitive (intraspecific and interspecific) survival and growth]. Tree species diversity increased steadily with increasing gap size (Fig. 1.2), primarily due to the absence of shade intolerant species in small gaps. There are two possible causes. First, there may have been little regeneration of shade intolerant species because of either low parent presence or minimal seedbed disturbance at the time of logging (LePage et al., submitted). Because of the type of equipment used in the old partial cutting (small crawler tractors) and the limited number of trees removed in small gaps, logging operations appeared to disturb small gaps far less than large gaps (D. Coates, personal observation). This would favour regeneration of hemlock over other species in small gaps. Lepage et al. (submitted) have shown that hemlock regeneration dominates in the ICH forest understory even when parents of other species are abundant. Second, low growth rates in low-light small gaps may have resulted in high mortality 153 of shade intolerant species over time (Kobe and Coates 1997). If differential mortality was the basis for species diversity differences observed in old retrospective gaps, then survivorship must be important for gap partitioning among the species. Tree species sorting as a result of seedling mortality has been observed in tropical forests (Brown 1996). In contrast to early establishment results at Date Creek (Chapter 4), where seedling density did not vary by gap size, overall tree density 34 - 41 years after gap formation increased with gap size (Fig 1.2). This might be be explained by higher rates of mortality (associated with lower growth rates) in medium and small gaps compared with large gaps. Seedling density was consistently highest in the south end of gaps three years after gap formation at Date Creek (Chapter 4). If the same regeneration pattern occurred in the retrospective gaps, then the advantage in numbers at southerly positions had diminished or disappeared 30 - 40 years later (Fig. 5.2). This discrepancy or evidence of a shift was most striking in medium gaps, where the middle or north positions typically had the greatest numbers of trees. This result is consistent with expected higher mortality rates in the shady south end of gaps, where growth rates would be lowest and confirms that the best location for early survival of germinants is not necessarily the best location for growth and survival of seedlings and larger trees (e.g., Burton 1997; Tappiener et al. 1997). In other words, the regeneration niche in gaps is "discordant" (Schupp 1995). Tree density in the gaps was high enough (59,730, 43,690 and 20,960 stems per hectare in large, medium and small gaps, respectively) to have constrained growth rates compared to rates typically observed in well-spaced trees. Comparisons of height and diameter growth by gap size and gap position are therefore confounded because inter-tree competition may have been more important than gap size and gap position effects. The largest gap filler trees were 154 found in the large and medium gaps, but the position of a gap filler tree had little effect on growth. Gap size and position appeared to have little effect on average height and diameter of western hemlock and height of western redcedar even when small trees were excluded. This result is in agreement with the expectation that growth of shade tolerant species should vary little by gap size and position (Whitmore 1989). Moreover, the average size of hybrid spruce, a species considered intermediate in shade tolerance, decreased rapidly with decreasing gap size. In conclusion, retrospective research is typified by poor or non-existent controls, predetermined and limited treatment options and uncertainty about causal factors. A l l of these limitations were at least partially present in the retrospective study. However, when expectations for future growth rates (Wright et al. 1998; Chapter 3) and future mortality rates (Coates and Kobe 1997) by gap size and gap position for the tree species at Date Creek are factored in, the observed distribution of trees in the old logging gaps are reasonable. 155 Chapter 6 Windthrow damage 2 years after partial cutting at the Date Creek silvicultural systems study in the Interior Cedar-Hemlock forests of northwestern British Columbia 2 . 6.1 Introduction In British Columbia, where clearcutting is the dominant harvesting practice, wind damage generally occurs in uncut mature and old-growth stands and along edges of clearcut units (Stathers et al. 1994). Partial cutting in mature and old-growth forests is increasing in response to new silvicultural, ecological and social management objectives. In part, the success of these partial cutting systems will depend on how the logging alters wind damage risk and whether the extent of subsequent wind damage compromises management objectives. Wind damage in forest stands is affected by internal stand characteristics (age, species composition, diameter and height distributions, presence of root rot), internal stand treatment history (time since last cutting, percent of stand removed during cutting), adjacent stand history (e.g. clearcutting), site conditions (soil moisture and depth, local topography) and storm characteristics (season, wind direction, average and maximum gust wind speed) (see reviews by Hubert 1918, Curtis 1943, Ruth and Yoder 1953, Savill 1983, Harris 1989, Stathers et al. 1994, Navratil 1995, Coutts and Grace 1995, Ruel 1995). There have been numerous post hoc studies of wind damage in partially cut stands (Smith and Weitknech 1915, Weidman 1920, Behre 1921, Gilmour 1926, Kelly and Place 1950, Ruth and Yoder 1953, Worthington 1953, McLintock 1954, Glew 1963, Elling and Verry 1978, 2 Coates, K.D. 1997. Windthrow damage 2 years after partial cutting at the Date Creek silvicultural systems study in the Interior Cedar-Hemlock forests of northwestern British Columbia. Can. J. For. Res. 27:1695-1701. 156 Fleming and Grossfield 1983). The intent was usually to document the magnitude of damage, look for causal factors and make recommendations on the acceptability of the partial cutting practices. Conclusions varied about the suitability of the various cutting methods. Unfortunately, direct comparison of wind damage in uncut stands and adjacent partially cut stands were made in only a few early studies (Smith and Weitknech 1915, Ruth and Yoder 1953, McLintock 1954, Lees 1964), of which only one was a replicated experiment (Lees 1964). Ruel (1995), in a review focusing on how silvicultural practice affects wind damage, concluded that windthrow losses remain mostly unquantified. In this study, a replicated experimental design was used to assess the hypothesis that increasing amounts of tree removal result in higher wind damage in the coast-interior transitional forests of northwestern British Columbia. 6.2 Methods The windthrow study is a component of the Date Creek silvicultural systems study (Coates et al. 1997), established in 1992, approximately 21 km north of Hazelton, west of the Kispiox River (55°22'N, 127°50'W; 370-665 m elevation). Date Creek is within the moist cold subzone of the Interior Cedar-Hemlock biogeoclimatic zone (ICHmc; Banner et al. 1993), a transitional area between the interior and coastal forests of northwestern British Columbia (Pojar et al. 1987, Meidinger and Pojar 1991). Forests in the Date Creek area are wild fire origin stratified mixtures of coniferous and deciduous tree species. In mature stands (140 yr), western hemlock (Tsuga heterophylla (Raf.) Sarg.) dominates; other species include western redcedar (Thuja plicata Donn. ex D. Don), subalpine fir (Abies lasiocarpa (Hook.) Nutt.), lodgepole pine (Pinus contorta var. latifolia Engelm.), hybrid spruce [a complex of white spruce (Picea glauca 157 (Moench) Voss), Sitka spruce (P. s i t c h e n s i s (Bong.) Carr.) and perhaps Engelmann spruce (P. e n g e l m a n n i i Parry ex Engelm.)], paper birch ( B e t u l a p a p y r i f e r a Marsh.), trembling aspen { P o p u l u s t r e m u l o i d e s Michx.) and black cottonwood ( P o p u l u s b a l s a m i f e r a ssp. t r i c h o c a r p a Torr. & Gray). Old-growth forests (250-300+ yr) are dominated by western hemlock with varying amounts of western redcedar and some amabilis fir ( A b i e s a m a b i l i s Dougl. ex Forbes). Amabilis fir abundance increases with elevation. The mountainous nature of the study area results in climatic variation occurring over relatively short distances. Variations correspond to elevation and surrounding topography. The moderate to steep lower slopes of the Kispiox Range make up one half of the total area. The rest of the area is a rolling morainal landscape, dissected by many glacial meltwater channels. Morainal parent materials dominate the area, ranging in texture from loamy sand to clay loam. Eluviated Dystric Brunisols, Orthic Dystric Brunisols, and Orthic Humo-Ferric Podzols are the most common soils (Agriculture Canada Expert Committee on Soil Survey 1987). For the windthrow study, three tree removal treatments (in approximately 20 ha treatment units) were replicated four times (12 units in total), in a randomized block design, with combinations of ecological site type and forest age as the blocking factor (mesic 140 yr; mesic-submesic 140 yr; mesic-subhygric 140 yr; mesic 350 yr). The treatments were: (1) no tree removal; (2) light tree removal where approximately 30% of the stand volume was removed across all species and diameter classes in single stems and small groups; and (3) heavy tree removal where approximately 60% of the stand volume was removed using a combination of small patch cuts (0.1 - 0.5 ha) and single-tree to small group-selection within the surrounding forest matrix. Again, in the forest matrix, trees were removed evenly across all species and diameter classes. 158 Wind damage was sampled in late-August, 1994, two years after logging was completed. Two major wind events occurred, one in mid-October, 1993, and another in mid-August, 1994, just prior to sampling. Both storms caused considerable windthrow over a much wider geographic area than the study sites, and were considered 1 in 5 or 1 in 10 year events by local foresters. Unfortunately, no wind or gust speed data were recorded for either event. To assess wind damage, transect lines were established 50 m apart (approximately 3600 m of line) on maps of each treatment unit. From these lines, approximately 2000 m of transect line was randomly selected. Windthrow originally rooted within 5 m of either side of the transect line was tallied. The total amount of transect line in an individual treatment unit was considered one plot (i.e., 1950 m x 10 m wide transect line = 1.95 ha). To be tallied, a windthrown tree had to have a diameter of 17.5 cm or greater and have been either snapped or uprooted (uprooted included trees leaning at greater than a 45° angle from the vertical). Data recorded for each tree included: species, stem diameter (at 1.3 m), height, height to live crown, crown class (overstory, intermediate or suppressed), direction of fall and type of windfall (uprooted or snapped). The extent of windthrow was calculated both as basal area (m^ ha~l) and stems per hectare (sph) on the ground. In addition, damage was calculated as a percent of the standing basal area and stem density using stand structure data collected immediately after logging (Coates et al. 1997). Individual tree data (mean diameter, height, height:diameter ratio, crown class) of the windthrown trees were compared to data from prism cruise plots established after logging was completed. In each cruise plot all ' in' trees were recorded by species and diameter and 2 trees were randomly selected for height measurement. To compare the 2 populations (except for 159 crown class), I used a folded form F statistic to test the hypothesis of equal variances and found only western hemlock diameter significant at p < 0.05. I then performed a classic 2 sample t-test on all the data; thus all p-values in Table 5, except for hemlock diameter, assume equal variances. In applied experiments of this type, it is important to examine the power the experiment had to detect differences among the treatment means (effect size). Based on an informal survey of local professional foresters, I selected an effect size of greater than 10% as large enough to be operationally significant. In northwestern British Columbia, 10 to 20% of a stand would have to be wind damaged before either management intervention would be considered, or the treatment would be written off as a failure. Finally, all statistical analysis was done with SAS version 6.11 (SAS Institute Inc. 1989). 6.3 Results On average, 6.7 stems per hectare (sph) of windthrow occurred across all treatment units, representing approximately 1.9% of the standing trees (Table 6.1). Over the two years, 0.63 m^ ha"l of merchantable basal area was damaged or 1.54% of the original standing basal area (Table 6.1). There was little evidence that wind damage to merchantable trees (those >17.5 cm diameter) was greater in the light or heavy removal treatments (Table 6.2). This result is true for measures of both basal area and stems per hectare on the ground and for damage expressed as either percent of original basal area or original number of standing stems (Table 6.2; all p-values >0.37). CD o ID- 1 O f s " o .11 E E x CB I E in ZD E co 5 o £ CO LU £Z co CD CO CO c CD CD x rr CO > o i x E CD CD rr co > o E CD rr cz CD CO CD 1^ CM • t f GO CT> CO 0 X CO CO > CD CO c o Q . CO 0 i CD CD| CO E CO "d TD C CM LO CO CM CO N co $2 LO O c o o i o CM N O •tf O CO oo CM o co o j ° . -tf O CM oo CO • t f oo • t f co LO CO oo co d CM o> co CO o T - LO CM d d d LO CO • t f oo CO LO CO d tf-CO t-- CO • t f CO d CM +1 +1 +1 +1 LO CM CO o C- LO CT) cb CM d •>-LO tf-CM CM LO tf; • t f d +1 + 1 +1 + 1 CT) 1— c o T— d co CM CM CO LO CO LO d d +1 + 1 +1 +1 CM CO OD LO CO LO d d CO 0 0 T3 o E r » S T3 M - CZ o > CO CO CO o x; o n. ca > o £ > CB CO > o 6 "3 > O (T) >< O u a* a. C3 •o ca > o E CO at, 00 W) 3 u J3 161 Table 6.2. A N O V A table for the wind damage response variables in a randomized block design. df SS MS F-value P-value Experimental block No of windthrown trees ha"1 3 109.2 36.4 2.3 0.18 % windthrow 3 35.0 11.7 4.7 0.05 Basal area of windthrown trees ha"1 3 2.99 0.99 15.1 0.003 % basal area 3 18.8 6.3 4.9 0.05 Tree removal level (treatment) No of windthrown trees ha"1 2 12.6 6.3 0.39 0.69 % windthrow 2 5.2 2.5 1.06 0.4 Basal area of windthrown trees ha"1 2 0.15 0.08 1.14 0.38 % basal area 2 2.8 1.4 1.09 0.39 Error (block*treatment) No of windthrown trees ha"1 6 96.1 16 % windthrow 6 14.9 2.5 Basal area of windthrown trees ha"1 6 0.4 0.07 % basal area 6 7.6 1.3 Total No of windthrown trees ha"1 11 217.9 % windthrow 11 55.1 Basal area of windthrown trees ha"1 11 3.5 % basal area 11 29.2 Power analysis found the experiment had a 90% reliability (at oc=0.05) to detect an effect size (difference between means) of 5.2% windthrow or more and may have had sufficient power for differences between 4.3 and 5.2% windthrow, but would not have detected differences smaller than 4.3% windthrow among the treatments (Table 6.1). The same detection limits exist for windthrow sph, basal area of windthrow per hectare and percent basal area damage using numbers presented in Table 6.1. The experiment had nearly 100% reliability for detecting differences among treatment means (effect size) of 8% or greater windthrow (6% for basal area damage), well below the effect size considered operationally significant (10%). Among the individual treatment units, windthrow varied from a low of 1.3 sph (0.3%, in a heavy removal unit) to a high of 12.8 sph (8.2%, also in a heavy removal unit) (Table 6.3). Basal area damage among individual treatment units varied from a low of 0.06 m2 ha"l to a high 162 of 1.68 m2 ha~l or between 0.2% and 5.9% of the original basal area, with both extremes again found in heavy removal units (Table 6.3). One clear trend in the data was a higher incidence of wind damage in the old-growth experimental block (Block 1 in Table 6.3). Statistically, it is not possible to conclude whether this trend was due to forest age or to some other factor, such as the physical location of the experimental block, since blocks were not replicated in the experiment. Among the mature forest experimental blocks there were no consistent trends; for example, there was no indication that the wettest sites (Block 4) were least windfirm. There was no evidence to suggest that species susceptibility to wind damage varied at different levels of tree removal (interaction among species and treatment: windthrown sph, p=0.98; percent windthrow, p=0.61; basal area damaged, p=0.97; percent basal area, p=0.53). There were, however, differences in susceptibility among species when data were pooled across all treatments and the observed distribution of windthrow was compared to the predicted distribution (Chi-Square df=6; value=58.1; p<0.001). Although western hemlock was by far the most common tree species and had the greatest number of wind damaged stems (Table 6.4), hemlock was not the most wind-damage-prone species. The rank order of species susceptibility (percent of the original standing population damaged by wind) from most- to least-damaged was amabilis fir (5.3%) > trembling aspen (4%) > subalpine fir (2%) > western hemlock (1.3%) > hybrid spruce (0.8%) > paper birch (0.6%) > western redcedar (0.3%) > lodgepole pine (0.2%) > black cottonwood (0%) (Table 6.4). The majority of wind damage was uprooting of trees (84.4%) rather than stem snapping (15.6%) and this trend was consistent for all species except trembling aspen, however, sample sizes for two of the species, lodgepole pine and black cottonwood, were too small to draw reliable conclusions (Table 6.4). Of the wind-damaged trees, 72.4% were from the overstory C O N O o T t 53 -H o NO >< o o. Q . <L> 3 6 X J fl rt fl 3 ON ON „ 1-1 w <U 4-J >-i CD & £ x ) C fl ^ s —H m ON C A l <U c n LH CD O CD 64—I V-i CD T t ^ 2 CD Cu aa c rt -.a CD 4-1 rt o " f l I-i C/j j_t CO <4-j X> rt c co rt CD £ o | * C/l ^ CD 0 3 t-i CD 1 rtX ) 3 X J C 14-1 o C O N O fl rt X H CD O ) CO co CM CO CM O _o CD iS c 0 E 1_ <D CL X LU CT) CO T f O LO CD 00 O) CM CT) 00 O) T - Is-o Is-LO LO T t Is-CT) CD CO LO CO Is-CO Is-O) CD LO LO co CO CD co CO CD T—" LO 00 CD 00 CD 1-00 CT) LO CO T t CM T t r-: LO Is- LO co T t T -Is-: T t LO CD CT) Is- LO LO T t LO 00 LO CD T t CO t - CD T t O) CO — sz O ) CO E 3 Si cj co S o £ o CD 2 - Q CO sz CO CD CM CO CD T t CD CT) LO Is- o T - CO T t CD CO T t CO CD o T t LO 1-T t CM T t 1- Is-cb cd LO CM CM Is-: Is-CD CD T t o T t CO CT) 00 co CD T - CD LO Is-: 00 T t CM LO LO T t T-Is-: T t LO CD O) Is- LO T t 00 T t T— T t LO 10 co co I « ^— CO °> E • £ CD tJ) 3 o L . _ CD O vfc Z CO co T t CM CM CO CT) CM CM CO sz CO CD v_ CO To CO CO CD CO CD CD T t * -O) sz CD £ 11 I* CM CD CO 0 CM O 0 0 T t T t CT) T— CO CT) 0 O CO CM CM CM 00 0 O 00 LO CM cp CO CO 1 — LO CT) CD CT) T - T t CT) O O 00 LO T t CO Is-C) 0 O CM T t ^ _ CT) co CM 1 ~ d CM T t CO CM LO CM CO r~ CO CT) LO * l — 1-; LO O ' l _ O 1 _ 00 Is- Is-T— CM d O O 00 CM i— T— CM O O d T t CD 00 O) CM O) c\i •*~ 'CO SZ CO CD CD ^—1 c 0 1 XJ c o sz S X) 1 0 •I To g CO -a CO J 3 > o e > c a O > o co >< o t-o-o. ca -a ca la > o 6 <D £ 00 T3 (U 00 _o "S 3 i" c3 c > U 3 H i Tt -VO rt c CD CD P H X OJ o l-l o a X ) CD C o o c/T _CD 'o CD & C/3 CD -t—» >» X e _o *4—» 3 j O °C -*—» c/3 cn .— -t-» 73 -g 2 c X ! CD X J £ S 3 £ fi . X ) ^ s vd 0 3 C/3 H 3 X I C L C L CO C XJ 0 O O * Q_ Z> XJ C CO 0 CD 1 C O zz T J c CO T J CO c 0 O T3 CO CO co • 0 -iii-Q co c C CD "cp D O c ro co CO X J E .gj CO CO 0 o CD C L co L O O CM L O o £ « J £ o o L O O X L O L O C D O Is- L O C O Is- C O CJ) o o O O O C O - i - O L O L O C M O w o r o n o o i n w w q u i ^ r r d c i c i o d CM o ^ co N LO O C M T j - c O g C M C O C O O O L O CO o o T - 1 c\i <°. CM T- °° O) 00 2 co cd oj d n c o c o ^ N o m c M O T j - T i - c o c o c M T j - T r i - p d d d ' t d d d d d +1 +1 +1 +1 +1 +1 +1 +1 +1 O O O O O h - T j - C M L O C O O T - c M C M c o - r - ; C M C M q q o d d L r i d d d d d ^ ^ L q c D ^ c p o c o ^ d d 2 ^ d o o c D L O +1 +1 +1 +1 +| +| +1 +1 +1 c o o c o c o c o a ) c o ° o ^ C v i l ^ d C J ) | v : c M ' r - L r ) ' ^ s - " ? *" O ) L O C D L O c M T f c o r - - T - o CD o co i _ E ^ CO H= CD ^= c n CD co c g c = 7S C L CO CO XJ 2 •£ © X J o 0 O S ._= £ T J C L C X I ± ^ 0 " co a. a co ; 5 T J 0) £ C O o i - C L o 0 0 E 2 3 < H CO >, co -2 o -2 o I D . S J f f l l -i n MD CD c o -a (U a -BO 7? rt 6 0 £ o t 3 x> "3 4—» rt O CD CD rt I § a c O J-J OH fi ™ 4 - 1 CD rt GC CD • g n fi rt O H a II <H rt oo C CJ CD t/3 CD CD O H 5 <u rt X ! O rt CD c/J CD S-l H i CD c CD 3 3 8 rt o H o T3 CD CO .—. CO <D CT-\ '—' Q. Q_ 3 cn £ "co td _ CD Cp E b 0) c o (O • CD ^ > o CO 4 - ; rr X v_ O ) CD CD 0 I E CO b 4 - LU •- +1 CZ CO CD E E CJ £ LU CD CO E CO Q CD N CO 0 CL E CO CO +1 cz CO 0 E CO CD CJ 0 Q. CO tf O T - O T- CM LO O CO CO cn ^ ™ 0 0 m +1 +l<2 d • t f CO ^ rt. co o CM +1 CT) CO X) CO E CZ B "55 "co c: m •tf +1 Is-co d CO CO cn CM in o o o o co co o o o o 2 o CM O CM CO co o o o o o r-~. oo LO i n +1 +lg L O q 0-CM fc; cn oo 0 0 8 oo ^ £ co Is-: co ^ m . d co co ro oo io Is-Is- co CD CM •r^ CO +1 +| tf" , - , 0 0 o co x d d ° co co Is- ro LO • t f CM CO d 1— • t f +1 •tf + 1 LO CM °°. +l CO +1 O _ o T O + 1 Is- +l CD 00 Is-' ° d 0 X 1 d CD CO CO CO s J • t f • t f 00 I s-oo • t f CM co co CM • t f CO « a co" w LO ^ CM 00 CM CT) CT) 4— Q_ T3 CO CO C •2>1 o 5 CO CO T ^ to r — -Q c jrf § E ! > ? • CD >- •= o . £ o •* CO E 0 td "co CD CO g 55 p; CO SZ ? r ? c § 1 -y C= — 1 q) co ^ E o 0 •§ -o tS E c <-- C CO g T - o LO co 1— • t f cn CM T— CM CD +1 +1 0 0 +1 +1 CO tf; +l 1- O 0 0 0 o> d 1 0 • t f " CO d CO 0 0 d 1 ^ 0 0 ro CO ^ co o d CO o CO • t f q CT) • t f CT) CO LO T- Is-; CM CD +1 CO +1 T— +1 d +1 co od cvi +1 +1 cn d +1 d +1 O O +1 + 1 00 +1 +1 LO CT) O d CO CM d T- • t f d cn CD O LO d LO d CM CO cd CM d CM d CM CM co cb CM •tf" CM d CO CO • t f co CM CM CM O CM +1 +1 tf" tf- 00 co 00 0 CM CM CO CM Is-00 CO CO CO E E £ £ CD "g «S -a ui 2 & I -c S I - - o a"> o y ^ o H C « i : < l ) t o " ^ i : < D ' S , s ' - < D  ^ : § > ? ^ : o , ? « * o ) ? § n <: n < n o < o «3 o ^ o >45 5 . E VO VO T3 CD =3 C • »—I c o o vd (D n E co CO T3 CD CO -—-co CD \ —^ Q_ Q_ co "jo T3 _ CD Cp E ° ; £ o CO • i _ , CD ' > o CO ~ oc -c ,_ O) CD 'CD 0 I E CO b LU E E o B LU CD CO I +1 Q co 0 0 N CO CO 0 ' o CD o_ co co tf" CM CD Is-00 00 o' CM' +1 +18 a) d LO tf1 CM CM 00 CM" tf-co CO co E "CJ CD $ TS W H i £ u> g E co £ 0 »- - s o o CM O CM CM CM 00 CM Is-•tf o o o o m o o • t f tf- LO cn CM 1 0 • t f CO CD LO +1 +1 +1 cd +1 O CM C ) Is- CM LO CO T-^ 0 CM CD CD cn cn co +1 CR • t f CO • t f 1— q 1— CD LO + 1 +1 CD +1 q +1 CM tf- ^ CD f- CO CO CM ° CM W CM tf- tf- CO Is-• t f tf co E "S 0 O co |-g > s « 1 = O- CO _CD CO CD O -= 3 & 2 > § CL O O. U l-< >i a* o c ca o CJ o > o u a. cj o ~° '-3 CJ cL J. 2 G ca 3 £ o 00 o '5 cn x; o ca • 00 CO CO ^ in 1— <D ca - a -u ca Q .2 eo -a u s 3 c >, o "a, S ca cj a ca £2 00 CJ > o\ O CN 167 (top third of canopy), 24.5% were intermediate (middle third of canopy) and 3.1% were suppressed (bottom third of canopy). This was very similar to the distribution of merchantable trees immediately after logging - 74.8% overstory, 22.2% intermediate and 3.0% suppressed. Except for western hemlock, there was no evidence (all p-values >0.05) to suggest that wind damaged individual species were different in terms of mean diameter, height or height: diameter ratio from the population of trees sampled in the cruise plots established right after logging (Table 6.5). Wind damaged hemlock trees appear to be smaller in diameter, but still fairly tall, resulting in a high height:diameter ratio compared to the general hemlock population. Wind damaged paper birch, although not significant (p=0.06), tended to have high height:diameter ratios (Table 6.5). The crown class distribution of wind damaged trees by species was also quite similar to that of the population right after logging was completed (Table 6.5). Finally, there was a very consistent trend in the direction of fall of all wind damaged trees (Fig. 6.1). The vast majority of trees fell to the north suggesting winds from the south caused most of the damage. This is reasonable given that the Kispiox River valley is orientated north-south and the major winds coming from the Pacific Coast (up the Skeena River from Prince Rupert) would enter the Kispiox valley from the south. 6.4 Discussion The results demonstrate that partial cutting rates of up to 60% removal can be undertaken in previously unmanaged mixed-species forests with little resulting short-term increase in wind damage. A certain amount of wind damage is acceptable, or even desirable, from an ecosystem perspective, but at some point damage can compromise management objectives. Choosing a threshold level of damage depends on management values. After two years, 2.2% of the trees 168 N S Figure 6.1. Frequency distribution of the direction in which windthrown trees fell. had been damaged in partially cut units compared to 1.1% in unlogged areas. This experiment does not have the power to detect whether a 1.1% difference was due to the partial cutting. However, the 1.1% increase in damage was well below the 10% effect size (the difference between treatment means) considered large enough to warrant either management intervention or to deem the partial cutting a failure. Amabilis fir, trembling aspen and subalpine fir were the most windthrown species in the study. This corresponds well with other studies that have reported true firs among the most susceptible tree species (e.g. Ruel 1995). However, it is possible that our amabilis fir results may be biased by the fact that amabilis fir occurred only in the old-growth experimental block 169 which experienced the highest levels of wind damage. Hardwoods are generally considered to be quite windfirm and except for aspen this was the case in the mixed-species stands. The high incidence of aspen windthrow might be due to their tendency to have high levels of decay when old (Peterson and Peterson 1995). Unlike other studies (e.g. Curtis 1943, Ruth and Yoder 1953), the trees wind damaged in the Date Creek study were simply a subsample of the population of trees found in the forest. Except for western hemlock (and possibly paper birch), where trees with a high height:diameter ratio were most damage prone, there were no individual tree characteristics that seemed to predispose trees to damage. Nor was the susceptibility of individual trees affected by the levels of tree removal tested in the study. The highest levels of damage occurred in the old-growth stands, generally 2 to 3 times greater for each wind damage variable (Table 6.2) than in the mature stands. The heavy removal old-growth stand was by far the most wind damaged stand in the experiment. The overall high level of damage in old-growth stands is most likely because the trees were old and often decayed than because of the location of the units. Advanced age, large size and presence of pathogens are frequently associated with higher windthrow risk (Hubert 1918, Behre 1921, Curtis 1943, Lohmander and Helles 1987). Wind damage in the heavy removal treatment may have been a topographic effect or due to the level of tree removal. The stand was adjacent to the unlogged unit which suffered little damage. This suggests old-growth stands may be more susceptible to wind damage as logging removal rates increase. The mature stands (140 yr), although old by industrial forest management standards, were quite windfirm. In these stands, losses to wind damage were 0.17 m.2 ha~l yr"l over the two year study. To put this in perspective, growth in the mature stands has averaged 0.39 m2 ha"l y r 1 over 140 years (Coates et al. 1997), thus windthrow losses have been less than half of 170 one year's average growth. These losses are similar in magnitude to those found with other partial cutting studies in mixed-eastern U.S. forests (Behre 1921, McLintock 1954), in New Brunswick upland spruce-fir (Kelly and Place 1950), after single-tree selection in northern interior B.C. forests (Glew 1963), in mixed spruce-aspen forests of northern Alberta (Lees 1964), and in selection cuttings of over-mature spruce-fir stands in Quebec (Weetman and Algar 1976). Like Date Creek, all of these studies were in unmanaged stands of natural origin. Windthrow studies that have reported low levels of damage in partially cut stands appear to have certain elements in common. Some sort of dispersed cutting pattern has been used and maximum canopy opening size is relatively small. In my study trees were removed individually or in groups with a maximum opening size of 0.5 ha. The six studies cited above (Behre 1921, Kelly and Place 1950, McLintock 1954, Glew 1963, Lees 1964, Weetman and Algar 1976), had similar levels and patterns of tree removal. A western U.S. study (Worthington 1953), that contrasted small and large opening sizes, found negligible damage around 0.5 to 1.6 ha openings, but severe damage around large clearcut boundaries. Similarly, Glew (1963) reported little damage with single-tree selection but high damage in strip-cut areas. Other strip-cutting studies have also reported high levels of damage (Fleming and Grossfield 1983). In general, wind damage increases when tree removal rates are high (Lohmander and Helles 1987). Windthrow research has shown that damage is greatest in the first few years following logging (Weidman 1920, Fleming and Grossfield 1983, Laiho 1987). Valid long-term comparisons of partially cut and unlogged stands should measure net growth, balancing losses due to windthrow (and other factors such as competition-induced mortality, insects and pathogens) against growth gains in released trees. 171 Wind damage is a natural liability in forest management. The cutting pattern employed in partial cuts appears to be a critical element in windthrow susceptibility. As with any cutting method, careful attention to topographic features and prevailing wind directions is required to mitigate windthrow risk. Although this study presents results from only two years after logging, the area has been subject to two major storms. These early results suggest that fear of wind damage should not prevent forest managers from implementing partial cutting prescriptions. 172 Conclusions Studies of gap dynamics have contributed significantly to our understanding of the role of small-scale disturbance in forest ecosystems, but have been little used by foresters for predicting response to partial cutting. Tree species diversity in the interior cedar-hemlock (ICH) forests of northwestern British Columbia is among the highest in the province (Meidinger and Pojar 1991), making this a good system in which to test a gap-based approach to understanding tree response to partial cutting. Review of the gap dynamics literature confirmed that many ecosystem responses to small- and intermediate-scale disturbance, such as those created by partial cutting of forests, can be studied in gaps (Table 1.2). Both gap size and gap position had a major influence on resource availability and physical conditions in the high latitude ICH forests. Gap position was especially important at gap sizes large enough for direct sunlight to reach the forest floor for at least part of the day, typically in gaps greater than 200 m . Establishment of natural regeneration was strongly tied to gap position. Seedling density of all species increased from the north to south end in gaps, especially in larger gaps receiving direct radiation at the north end. These results were in close agreement with those of Wright et al. 1999 who directly seeded species into different gap positions in the same forests, and with findings in more southerly temperate forests (Farmer 1997; Tappeiner et al. 1997). Exposure to direct sunlight (Chapter 2) appears to constrain natural regeneration even at these relatively high latitudes and even in the supposedly sheltered microclimate of a partially cut forest. Strong interest in the gap partitioning hypothesis in the ecological literature, especially in tropical forests (Ricklefs 1977; Lieberman et al. 1995), led me to design the partial cutting 173 experiment so that species responses across a wide range of gaps sizes (10 - 5000 m2) could be examined. The tree species of the ICH forest include a full spectrum of shade tolerances (Krajina 1969; Krajina et al. 1982; Burns and Honkala 1990; Carter and Klinka 1992; Klinka et al. 1992; Wang et al. 1994; Kobe and Coates 1997), yet there was little evidence to suggest that gap partitioning occurs during the regeneration phase as it does in tropical forests. My results were similar to those of Sipe and Bazzaz (1995) in northeastern US forests and to those of Grey and Spies (1996) in Pacific Northwest U.S. forests. The ICH tree species were generalists across the wide range of gap sizes studied and showed marked species differences in regeneration success only at the two extremes of full open conditions in a clearcut (well away from the forest edge) and in the understory of the undisturbed forest, suggesting that adaptation to the subtleties of gap size are not well developed in these species. Regeneration in the forest understory is inhibited by the undisturbed moss substrate and is restricted mostly to western hemlock (LePage et al., submitted). On suitable substrates, emergence and early establishment of ICH tree species in the forest understory can be high (Wright et al. 1999). Many studies have shown natural regeneration to be limited in large clearcuts due to a combination of unfavourable microclimate and seed limitation (Farmer 1997), and the Date Creek study was no exception. A l l species regenerated well in partial cut areas but only deciduous species were common at the centre of clearcuts. There was little difference among the planted species (western redcedar, western hemlock, subalpine fir, hybrid spruce and lodgepole pine) in whole-plant growth response to ambient light, consistent with the findings of Wright et al. (1998) for older naturally-established trees in the ICH. A l l species exhibited similar average growth rates in various gap positions and in gaps of various size, especially those less than 0.1 ha. For a given species, there was little 174 difference in growth rate or total size once gap size exceeded 0.1 ha. As expected, predicted growth of lodgepole pine above 70% full sun, and five year performance of lodgepole pine in large gaps and clearcuts exceeded that of other conifer species. Contrary to expectations, all species, including lodgepole pine, performed equally well in openings less than 0.1 ha. However, in very small gaps and in the forest understory, differences in mortality were evident, with the most light demanding species having the greatest early mortality, as predicted by the mortality models of Kobe and Coates (1997). In ICH forests, gap partitioning may take place gradually in response to the different growing conditions associated with different gap sizes and gap positions. Fifth year results from the planting study indicate differential mortality among the species based on shade tolerance ranking in the shadiest portions of larger gaps, in small gaps and in the forest understory. These results were supported by the retrospective study of old logging gaps where tree species richness clearly increased with increasing gap size. Taken as a whole, these three studies demonstrate that regeneration niches in ICH forests are discordant, in the sense of Schupp (1995). Favourable locations for emergence and early establishment of natural regeneration are less favourable for growth and survival of established seedlings. Tree species diversity appears to be controlled more by differentiation among growth and survival niches than by the regeneration niches. The risk of windthrow after partial cutting is of concern to ICH forest managers, yet there was little evidence that gap creation at Date Creek increased the risk of wind damage. There was little difference in damage among the uncut forest and the 30 and 60% removal partial cutting areas despite two major windstorms occurring shortly after the experiment was established, a time when partially cut units should have been most vulnerable (Weidman 1920, 175 Fleming and Grossfield 1983, Laiho 1987). This result indicates that partial cutting can be successfully implemented with minimal risk of increased windthrow damage. Silvicultural research must look beyond the establishment and early growth phase to time periods of decades and centuries to assess longer-term effects of stand management activities on forest community dynamics and succession. In order to practically address these long-term effects, empirical studies must be linked with simulation models of future forest development. My gap-based studies in the ICH, combined with the work of Kobe and Coates (1997), Wright et al. (1998) and LePage et al. (submitted), are being used to calibrate SORTIE (Pacala et al. 1996), a spatially-explicit model designed to examine the effects of small-scale disturbance in mixed-species forests. A distinctive feature of SORTIE is that the model requires direct input of data from rigorous field studies. The experimental treatments applied in the Date Creek study have provided ideal conditions for those studies. SORTIE can be used to model the consequences of a wide range of partial cutting strategies, at different spatial scales and over different time periods, an impossible undertaking for field-based research. One of the unique challenges in developing optimal partial cutting strategies is the management of patchiness in the distribution of both harvested and residual trees. As a spatially-explicit model, SORTIE is ideally suited to address this issue. Forest managers need help to determine the optimal spatial pattern of partial cutting (including the optimal mix of sizes and spatial distribution of discrete canopy gaps) in a given stand. The pattern will be a function of existing stand structure and composition, and desired stand structure and composition in the future. The Date Creek research will lead to better growth and yield predictions in complex mixed-species stands and help our understanding of how logging affects tree species composition, and the structural attributes of the forest that may be required for wildlife habitat. 177 Bibliography Alexander, R. R. 1984. Natural regeneration of Engelmann spruce after clearcutting in the central Rocky Mountains in relation to environmental factors. U.S. Dep. Agric , For. Serv., Rocky Mt. For. Range Exp. Sta., Fort Collins, Colorado. Res. Pap. 254. Aplet, G.H., Johnson, N . , Olson, J.T. and Sample, V . A . (eds.). 1993. Defining Sustainable Forestry. Island Press, Washington, D.C. Attiwill, P .M. 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. For. Ecol. Manage. 63:247-300. Banner, A. , MacKenzie, W., Haeussler, S., Thomson, S., Pojar, J and Trowbridge, R. 1993. A field guide to site identification and interpretation for the Prince Rupert Forest Region, Land Manage. Handb. 26., B.C. Ministry of Forests, Victoria, B.C Barbeau, C M . 1928. The Downfall of Temlaham. Macmillan, Toronto. Barden, L.S. 1979. Tree replacement in small canopy gaps of a Tsuga canadensis forest in the Southern Appalachians, Tennessee. Oecologia 44:141-142. Barik, S.K., Pandey, H.N., Tripathi, R.S. and Rao, P. 1992 Microenvironmental variability and species diversity in treefall gaps in a sub-tropical broadleaved forest. Vegetatio 103:31-40. Barker, M.G. , Press, M.C. and Brown, N.D. 1997. Photosynthetic characteristics of dipterocarp seedlings in three tropical rain forest light environments: a basis for gap partitioning? Oecologia 112:453-463. Bassman, J.H. 1985. Selected physiological characteristics of lodgepole pine. In: Baumgartner, D .M. , Krebill, R.G., Arnott, J.T. and Weetman, G.F. Lodgepole pine the species and its management. Symposium Proc, May 14-16, Vancover, B.C. Cooperative Extension, Washington State Univ., Pullman, WA. pp. 27-44. Bazzaz, F.A. 1983. Characteristics of populations in relation to disturbance in natural and man-modified environments. In: H.A. Mooney and M . Gordon (eds.) Disturbance and Ecosystems: Components of Response. Springer-Verlag, Berlin, pp.259-275. Bazzaz, F.A. 1984. Dynamics of wet tropical forests and their species strategies. In: E. Medina, H.A. Mooney and C. Vazquez-Yanes (Editors). Physiological ecology of plants of the wet-tropics. Dr. W. Junk Publishers, The Hague, Netherlands, pp. 233-244. Bazzaz, F.A. 1996. Plants in changing environments. Cambridge University Press, Cambridge, U.K. 178 Bazzaz, F.A. and Pickett, S.T.A. 1980. Physiological ecology of tropical succession: a comparative review. Ann. Rev. Ecol. Syst. 11:287-310. Beatty, S.W. and Stone, E.L. 1986. The variety of microsites created by tree falls. Can. J. For. Res. 16:539-548. Behre, C E . 1921. A study of windfall in the Adirondacks. J. For. 19:632-637. Blake, J.G. and Hoppes, W.G. 1986. Influence of resource abundance on use of treefall gaps by birds. Auk 103:328-340. Bradshaw, F.J. 1992. Quantifying edge effect and patch size for multiple-use silviculture - a discussion paper. For. Ecol. Manage. 48:249-264. Brokaw, N . V . L . 1985. Gap-phase regeneration in a tropical forest. Ecology 66:682-687. Brokaw, N . V . L . 1987. Gap-phase regeneration of three pioneer tree species in a tropical forest. J. Ecology 75:9-19. Brown, N . 1996. A gradient of seedling growth from the centre of a tropical rain forest canopy gap. For. Eco. Manage. 82:239-244. Bunnell, F.L., McNay, R.S., and Shank, C C 1985. Trees and snow: the deposition of snow on the ground - a review and quantitative synthesis. rWTFR Report 17, B.C. Ministries of Environment and Forests, Victoria, B.C. Burns, R. M . and Ffonkala, B.H. (tech. coords). 1990. Silvics of North Amercia. Vol .1 and 2, Agric. Handb. 654, USDA. For. Serv., Washington, D . C Burton, P.J. 1997. Conifer germination on different seedbeds influenced by partially-cut canopies. Abstract, Bulletin of the Ecological Society of America 78(4)Suppl.:230. Burton, P.J. and Bazzaz, F.A. 1995. Ecophysiological responses of tree seedlings invading different patches of old-field vegetation. J. Ecol. 83:99-112. Burton, P.J., Balisky, A . C , Coward, L.P., Cumming, S.G., and Kneeshaw, D.D. 1992. The value of managing for biodiversity. For. Chron. 68:225-237. Canham, C D . 1985. Suppression and release during canopy recruitment in Acer saccharum. Bull. Torrey Bot. Club 112:134-145. Canham, C D . 1988a. Growth and canopy architecture of shade-tolerant trees: response to canopy gaps. Ecology 69:786-795. 179 Canham, C D . 1988b. An index for understory light levels in and around canopy gaps. Ecology 69:1634-1638. Canham, C D . 1990. Suppression and release during canopy recruitment in Fagus grandifolia. Bull. Torrey Bot. Club 117:1-7. Canham, C D . , Denslow, J.S., Piatt, W.J., Runkle, J.R., Spies, T.A. and White, P.S. 1990. Light regimes beneath closed canopies and tree-fall gaps in temperate and tropical forests. Can. J. For. Res. 20:620-631 Canham, C D . , Finzi, A . C , Pacala, S.W. and Burbank, D.H. 1994. Causes and consequences of resource heterogeneity in forests: interspecific variation in light transmission by canopy trees. Can. J. For. Res. 24:337-349. Canham, C D . , Berkowitz, A.R., Kelly, V.R., Lovett, G . M , Ollinger, S.V. and Schnurr, J. 1996. Biomass allocation and multiple resource limitation in tree seedlings. Can. J. For. Res. 26:1521-1530. Carter, R.E. and Klinka, K. 1992. Variation in shade tolerance of Douglas fir, western hemlock, and western redcedar in coastal British Columbia. For. Ecol. Manage. 55:87-105. Chen, H . Y . H . 1997. Interspecific response of planted seedlings to light availability in interior British Columbia: survival, growth, allometric patterns, and specific leaf area. Can. J. For. Res. 27:1383-1393. Chen, H .Y.H. , Klinka, K. and Kayahara, G.J. 1996. Effects of light on growth, crown architecture, and specific leaf area for naturally established Pinus contorta var. latifolia and Pseudotsuga menziesii var. glauca saplings. Can. J. For. Res. 26:1149-1157. Chen, J .M., Black, T.A., Price, D.T., and Carter, R.E. 1993. Model for calculating photosynthetic photon flux densities in forest openings on slopes. J. App. Met. 32:1656-1665. Cherrett, J .M. (ed.) 1989. Ecological Concepts. Blackwell Scientific Publications, Oxford, England. Christensen, N.L. , Bartuska, A . M . , Brown, J.H., Carpenter, S., D'Antonio, C , Francis, R., Franklin, J.F., MacMahon, J.A., Noss, R.F., Parsons, D.J., Peterson, C.H., Turner, M . G . and Woodmansee, R.G. 1996. The scientific basis for ecosystem management. Ecol. Appl. 6:665-691. Christy, J.E. and Mack, R.N. 1984. Variation in demography of juvenile Tsuga heterophylla across the substratum mosaic. J. Ecol. 72:75-91. 180 Clark, D.F. 1994. Post-fire succession in the sub-boreal spruce forests of the Nechako Plateau, central British Columbia. M.Sc thesis, University of Victoria, Victoria, B.C. Cloutts, M.P. and J. Grace (eds.). 1995. Wind and Trees. Cambridge Univ. Press., Cambridge, U.K. Coates, K .D. and Burton, P. J. 1997. A gap-based approach for development of silvicultural systems to address ecosystem management objectives. For. Ecol. Manage. 99:337-354. Coates, K.D. , Banner, A. , Steventon, J.D., LePage, P. and Bartemucci, P. 1997. The Date Creek silvicultural systems study in the Interior Cedar-Hemlock forests of northwestern British Columbia: Overview and treatment summaries. Land Manage. Handb. 38, B.C. Ministry of Forests, Victoria, B.C. Collins, B.S. and Pickett, S.T.A. 1988. Response of herb layer cover to experimental canopy gaps. Amer. Midland. Nat. 119:282-290. Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302-1310. Crome, F.H.J, and Richards, G.C. 1988. Bats and gaps: Microchiropteran community structure in a Queensland rain forest. Ecology 69:1960-1969. Curtis, J.D. 1943. Some observations on wind damage. J. For. 41:877-882. Daniel, T.W., Helms, J.A. and Baker, F.S. 1979. Principles of Silviculture. Second edition. McGraw-Hill Book Company, New York. Day, R. J. 1964. The microenvironments occupied by spruce and fir regeneration in the Rocky Mountains. Publication 1037, Forest Research Branch, Canadian Department of Forestry, Ottawa. Decie, T. 1957. Working plan for the Forest Experimentation Station Aleza Lake: for the period April 1st, 1957 to March 31, 1967. Forest Service, Prince George, B.C. DeLucia, E.H. 1986. Effects of low soil temperature on net photosynthesis, stomatal conductance and carbohydrate concentraction in Engelmann spruce (Picea engelmanni Parry ex Engelm) seedlings. Tree Physiol. 2:143-154. Denslow, J.S. 1980. Gap partitioning among tropical rainforest trees. Biotropica 12 (Supp.):47-55. 181 Denslow, J.S. 1987. Tropical rainforest gaps and tree species diversity. Ann. Rev. Ecol. Syst. 18:431-451. Denslow, J.S., Schulz, J .C , Vitousek, P .M. and Strain, B.R. 1990. Growth responses of tropical shrubs to treefall gap environments. Ecology 71:165-179 Denslow, J.S. and Hartshorn, G.S.. 1994. Tree-fall gap environments and forest dynamic processes. In: McDade, L .A. , K.S. Bawa, H.A. Hespenheide, and G.S. Hartshorn (Editors). La Selva: ecology and natural history of a neotropical rain forest. The University of Chicago Press, pp. 120-127. Dirzo, R., Horvitz, C C , Quevedo, H. and Lopez, M . A . 1992. The effects of gap size and age on the understory herb community of a tropical Mexican rain forest. J. Ecol. 80:809-822. Dykstra, G.F. 1974. Photosynthesis and carbon dioxide transfer resistance of lodgepole pine seedlings in relation to irradiance, temperature and water potential. Can. J. For. Res. 4:201-206. Ehrenfeld, J.G. 1980. Understory response of canopy gaps of varying size in a mature oak forest. Bull. Torrey Bot. Club 107:29-41. Eis, S. 1965. Development of white spruce and alpine fir seedlings on cut-over areas in the Central Interior of British Columbia. For. Chron. 41: 419-431. Eis, S. 1967. Establishment and early development of white spruce in the interior of British Columbia. For. Chron. 43:174-177. Ek, A.R., Shifley, S.R. and Burk, T.E. (Eds.). 1988. Forest growth modelling and prediction, 2 Volumes. Proceedings of the IUFRO Conference, August 23-27, 1987, Minneopolis, Minnesota, Gen. Tech. Rep. NC-120, USDA For. Serv., St Paul, Minnesota. Elling, A .E . and E.S. Verry. 1978. Predicting wind-caused mortality in strip-cut stands of peatland black spruce. For. Chron. 54:249-252. Farmer, R.E. 1997. Seed Ecophysiology of Temperate and Boreal Zone Forest Trees. St. Lucie Press, Delray Beach, Florida, USA. Field, C B . 1988. On the role of photosynthetic responses in constraining the habitat distribution of rainforest plants. Aust. J. Plant. Phys. 15:343-358. 182 Fetcher, N . , Oberbauer, S. F., and Chazdon, R. L . 1993. Physiological ecology of plants. In: L .C . McDade, K.S. Bawa, H.A. Hespenheide, and G.S. Hartshorn (eds.), La Selva: Ecology and Natural History of a Neotropical Rain Forest. University of Chicago Press, Chicago, pp. 128-141 Fleming, R.L. and R . M . Crossfield. 1983. Strip cutting in shallow-soil upland black spruce near Nipigon, Ontario. III. Windfall and mortality in the leave strips: preliminary results. Canadian For. Serv., Great Lakes For. Res. Centre, Info. Rep. O-X-354. Galindo-Leal, C. and Bunnell, F.L. 1995. Ecosystem management: implications and opportunities of a new paradigm. For. Chron. 71:601-606. Garman, E.H., and Orr-Ewing, A . L . 1949. Direct-Seeding Experiments in the Southern Coastal Region of British Columbia. Dep. Lands and Forests, B.C. Forest Service, Victoria B.C. Givnish, T.J. 1988. Adaptation to sun and shade: a whole-plant perspective. Aust. J. Plant. Phys. 15:63-92. Gilmore, D.W. 1997. Ecosystem management - a needs driven, resource-use philosophy. For. Chron. 73:560-564. Gilmour, J.D. 1926. Clear-cutting of pulpwood lands. For. Chron. 2:1-2. Glew, D.R. 1963. The results of stand treatment in the white spruce-alpine fir type of the northern interior of British Columbia. Manage. Note 1, For. Dep. Lands, Forests, and Water Resources, B.C. For. Serv., Victoria, B.C. Golding, D.L. and Swanson, R.H. 1978. Snow accumulation and melt in small forest openings in Alberta. Can. J. For. Res. 8:380-388. Gray, A . N . and Spies, T.A. 1996. Gap size, within-gap position and canopy structure effects on conifer seedling establishment. J. Ecol. 84: 635-645. Greatorex-Davies, J.N., Sparks, T.H., Hall, M . L . 1994. The response of Heteroptera and Coleoptera species to shade and aspect in rides of coniferised lowland woods in southern England. Biol. Cons. 67:255-273. Goldstein, G.H., L .B. Brubaker, and T.M. Hinckley. 1985. Water relations of white spruce (Picea glauca (Moench) Voss) at treeline in north central Alaska. Can. J. For. Res. 15:1080-1087. 183 Grossnickle, S.C. and Arnott, J.T. 1992. Gas exchange response of western hemlock seedlings from various dormancy induction treatments to reforestation site environmental conditions. For. Ecol. Manage. 49:177-193. Grubb, P.J. 1977. The maintenance of species-richness in plant communities: the importance of the regeneration niche. Biol. Rev. 52:107-145. Haeussler, S., Coates, K.D. , and Mather, J. 1990. Autecology of common plants in British Columbia: A literature review. Canada - B.C. Forest Resources Development Agreement, FRDA Rep. 158, B.C. Min. of Forests and Forestry Canada, Victoria, B.C.. Hagner, M . 1992. Silvicultural alternatives. Swedish Univ. Agric. Sci., Dept. Silv., Rep. No. 35., Umea, Sweden. 214p. Harmon, M.E. and Franklin J.F. 1989. Tree seedlings on logs in Picea-Tsuga forests of Oregon and Washington. Ecology 70:48-59. Harris, A.S. 1989. Wind in the forests of southeast Alaska and guides for reducing damage. Gen. Tech. Rep. PNW-GTR-224, U.S. Dept. Agric , For. Serv., Pac. Northwest Res. Stn. Harris, L .D. 1984. The Fragmented Forest: Island Biogeography Theory and the Preservation of Biotic Diversity. Univ. Chicago Press, Chicago, Illinois. Harrison, S. 1987. Treefall gaps versus forest understory as environments for a defoliating moth on a tropical shrub. Oecologia 72:65-68. Hennessey, G.R. 1968. Early survival and growth of planted and seeded white spruce as affected by seedbed types occurring on scalped strips prepared in aspen stands, Manitoba. Intern. Rep. MS-70, Dep. For. Rural Devel., Can. For. Serv., For. Res. Lab., Winnipeg, Manitoba. Hofgaard, A . 1993. 50 years of change in Swedish boreal old-growth Picea abies forest. J. Veg. Sci. 4:773-782. Hubert, E.E. 1918. Fungi as contributory causes of windfall in the northwest. J. For. 16:696-714. Hughes, E.L. 1967. Studies in the standard seedbed treatment to obtain spruce and fir reproduction in the mixedwood slope type in northwestern Ontario. For. Branch Public. 1189, Can. Dept. For. Rural Develop., Ottawa, Ontario. 184 Hytteborn, H. , Liu, Q. and Verwijst, T. 1991. Natural disturbance and gap dynamics in a Swedish boreal spruce forest. In: N . Nakagoshi and F.B. Golley (editors). Coniferous Forest Ecology from an International Perspective. SPB Academic Publishing bv, The Hague, pp. 93-108. Jones, E.W. 1945. The structure and reproduction of the virgin forest of the north temperate zone. New Phytologist 44:130-148. Jonsson, B.G. and Esseen, P. 1990. Treefall disturbance maintains high bryophyte diversity in a boreal spruce forest. J. Ecol. 78:924-936. Kaufmann, M.R. 1975. Leaf water stress in Engelmann spruce. Plant Physiol. 56:841-844. Kayahara, G.J., Chen, H.Y.H. , Klinka, K., and Coates, K.D. 1996. Relations of terminal growth and specific leaf area to available light in naturally regenerated seedlings of lodgepole pine and interior spruce in central British Columbia. Research Note 09, B.C. Ministry of Forests, Victoria, B.C. Kelly, J.A.B. and I.C.M. Place. 1950. Windfirmness of residual spruce and fir. Pulp and Paper Mag. Canada. 51(6):124. Kimmins, J.P. 1993. Ecology, environmentalism and green religion. For. Chron. 69:285-289. IGmmins, J.P., D. Mailly and K . A . Scoullar. 1997. Models and their role in ecology and resource management. In: Forest Ecology: A Foundation for Sustainable Management. 2nd Edit. Prentice Hall, NJ. Kitajima, K. 1994. Relative importance of photosynthetic traits and allocation patterns as correlates of seedling shade tolerance of 13 tropical trees. Oecologia 98:419-428. Klinka, K., Carter, R.E. and Feller, M.C. 1990. Cutting old-growth forests in British Columbia: ecological considerations for forest regeneration. Northwest Envir. J. 6:221-242. Klinka, K. , Feller, M . C , Green, R.N., Meidinger, D.V., Pojar, J. and Worrall, J. 1990. Ecological principles: applications. In: Lavender, D.P., Parish, P., Johnson, C M . , Montgomery, G, Vyse, A, Willis, R.A., and Winston, D. (editors) 1990. Regenerating British Columbia's forests. University of British Columbia Press, Vancouver, B.C. pp. 55-72. Klinka, K., Wang, Q., Kayahara, G.J., Carter, R.E. and Blackwell, B .A. 1992. Light growth response relationships in Pacific silver fir (Abies amabilis) and subalpine fir (Abies lasiocarpa). Can. J. Bot. 70:1919-1930. 185 Kneeshaw, D.D. 1992. Tree population dynamics of some old sub-boreal spruce stands. M.Sc. thesis, University of British, Faculty of Forestry, Vancouver, B.C. Kobe, R.K. and Coates K.D. 1997. Models of sapling mortality as a function of growth to characterize interspecific variation in shade tolerance of eight tree species of northwestern British Columbia. Can. J. For. Res. 27:227-236. Kohm, K . A . and Franklin, J.F. (Eds.) 1997. Creating a Forestry for the 21st Century. Island Press, Washington, D.C. Koppenaal, R.S., Hawkins, B.J. and Mitchell, A . K . 1995. Acclimation of natural and planted Amabilis fir to clearcut and stand edge microclimates on a coastal montane reforestation site. FRDA Rep. 232, Canada Forst Serv. and B.C. Ministry of Forests, Victoria, B.C., 21p. Krajina, V.J . 1969. Ecology of forest trees in British Columbia. Ecol. West. N . Amer. 2: 1-146. Krajina, V.J . , Klinka, K., and Worrall, J. 1982. Distribution and ecological characteristics of trees and shrubs of British Columbia. Faculty of Forestry, University of British Columbia, Vancouver. Kuuluvainen, T. 1994. Gap disturbance, ground microtopography, and the regeneration dynamics of boreal conifer forests in Finland: a review. Annales Zoologici Fennici. 31:35-52. Laiho, O. 1987. Susceptibility of forests to windthrow in Southern Finland. Folia Forestalia 706:23-24. Latham, R.E. 1992. Co-occurring tree species change rank in seedling performance with resources varied experimentally. Ecology 73:2129-2144. Lavender, D.P. 1990 Physiological principles of regeneration. In: Lavender, D.P., Parish, P., Johnson, C M . , Montgomery, G, Vyse, A, Willis, R.A., and Winston, D. (editors) 1990. Regenerating British Columbia's forests. University of British Columbia Press, Vancouver, B.C. pp. 30-44. Lavender, D.P., Parish, P., Johnson, C M . , Montgomery, G, Vyse, A , Willis, R.A., and Winston, D. (Eds.) 1990. Regenerating British Columbia's forests. University of British Columbia Press, Vancouver, B.C. 372 p. Leibold, M . A . 1995. The niche concept revisited: mechanistic models and community context. Ecology 76:1371-1382. 186 Lawton, R.O. and Putz, F.E. 1988. Natural disturbance and gap-phase regeneration in a wind-exposed tropical cloud forest. Ecology 69:764-777. LePage, P. 1995. The structure and development pattern of mixed-species forest stands in the Interior Cedar-Hemlock zone: Moist Cold subzone of northwestern British Columbia. MSc thesis. Oreg. State Univ., Corvallis, Oreg. LePage, P., Canham, C. D., Coates, K.D. and Bartemucci, P. in review. Seed source versus substrate limitation of seedling recruitment in Interior Cedar-Hemlock forests of British Columbia. Ecol. Appl., submitted. Lertzman, K.P. 1992. Patterns of gap-phase replacement in a subalpine, old-growth forest. Ecology 73:657-669. Lertzman, K.P., Sutherland, G.D., Inselberg, A. and Saunders, S.C. 1996. Canopy gaps and the landscape mosaic in a coastal temperate rainforest. Ecology 77:1254-1270. Levey, D.J. 1988. Tropical wet forest treefall gaps and distribution of understory birds and plants. Ecology 69:1076-1089. Lieberman, M . , Lieberman, D. and Peralta, R. 1989. Forests are not just swiss cheeze: canopy stereogeometry of non-gaps in tropical forests. Ecology 70:550-552. Lieberman, M . , Lieberman, D., Peralta, R. and Hartshorn, G.S. 1995. Canopy closure and the distribution of forest tree species at La Selva, Costa Rica. J. Trop. Ecol. 11:161-178. Liu, Q and Hytteborn, H. 1991. Gap structure, disturbance and regeneration in a primeval Picea abies forest. J. Veg. Sci. 2:391-402. Lohmander, P. and F. Helles. 1987. Windthrow probability as a function of stand characteristics and shelter. Scan. J. For. Res. 2:227-238. Lopushinsky, W. and M.R. Kaufmann. 1984. Effects of cold soils on water relations and spring growth of Douglas-fir seedlings. For. Sci. 30:628-634. Major, J.E. 1990. Ecophysiological assessment of western hemlock and western redcedar green-house stocktypes. M.Sc. Thesis. University of Btitish Columbia, Vancouver, B.C. 163 pp. Marks, P. L . 1975. On the relation between extension growth and successional status of deciduous trees of the northeastern United States. Bull. Torrey Bot. Club 102: 172-177. 187 Marrs, R.H., Thompson, J., Scott, D. and Proctor, J. 1991. Nitrogen mineralization and nitrification in terra firme forest and savanna soils on Ilha de Maraca, Roraima, Brazil. J. Trop. Ecol. 7:123-137. Matthews, J.D. 1989. Silvicultural Systems. Oxford Scientific Publications, Oxford, U K . McLintock, T.F. 1954. Factors affecting wind damage in selectively cut stands of spruce and fir in Maine and northern New Hampshire. Sta. Paper No. 70, U.S. Dept. Agric , For. Serv., Northeastern Exp. Stn. Meidinger, D. and J. Pojar (compilers and editors). 1991. Ecosystems of British Columbia. Special Report Series 6, B.C. Min. For., Victoria, B.C. Milliken, G.A. and Johnson, D.E. 1992. Analysis of Messy Data. Volume 1: Designed Experiments. Van Nostrand Reinhold, New York. Chapter 3 (pp. 29-45). Minckler, L.S., Woerheide, J.D. and Schlesinger, R.C. 1973. Light, soil moisture, and tree reproduction in hardwood forest openings. Res. Pap. NC-89, U.S. Dep. Agric , For. Serv., North Central For. Exp. Stn., St. Paul, Minneosta. Mitchell, A . K . and Arnott, J.T. 1995. Effects of shade on the morphology and physiology of amabilis fir and western hemlock seedlings. New Forests 10:79-98. Mladenoff, D.J. 1987. Dynamics of nitrogen mineralization and nitrification in hemlock and hardwood treefall gaps. Ecology 68:1171-1180. Moore, M.R., and VanKat, J.L. 1986. Responses of the herb layer to the gap dynamics of a mature beech-maple forest. Amer. Midi. Nat. 115:336-347. Navratil, S. 1995. Minimizing wind damage in alternative silviculture systems in boreal mixedwoods. Rep. 8033-124, Canadian For. Serv., Northern For. Centre, Edmonton, Alberta. Nemec, A.F .L . 1991. Power analysis handbook for the design and analysis of forestry trials. Biometrics Information Handb. No. 2, B.C. Ministry of Forests, Victoria, B.C.. Noble, D. L. , and Alexander, R.R. 1977. Environmental factors affecting natural regeneration of Engelmann spruce in the Central Rocky Mountains. For. Sci. 23: 420-429. Noss, R.F. 1991. Effects of edge and internal patchiness on avian habitat use in an old-growth Florida hammock. Nat. Areas J. 11:34-47. 188 Oliver, C D . 1981. Forest development in North America following major disturbances. For. Ecol. Manage. 3:153-168. Orians, G.H. 1982. The influence of tree-falls in tropical forests on tree species richness. Trop. Ecology 23:255-279. Pacala, S. W., Canham, C. D., Saponara, J., Silander, J. A. , Jr., Kobe, R. K. and Ribbens, E. 1996. Forest models defined by field measurements: II Estimation, error analysis, and dynamics. Ecol. Monogr. 66: 1-43. Pacala, S.W., Canham, C D . , Silander, Jr., J.A., Kobe, R.K. 1994. Sapling growth as a function of resources in a north temperate forest. Can. J. For. Res. 24:2172-2183. Parsons, W.F.J., Knight, D.H., and Miller, S.L. 1994. Root gap dynamics in lodgepole pine forests: nitrogen transformations in gaps of different size. Ecol. Appl. 4:354-362. Peterson, C J . and Pickett, S.T.A. 1990. Microsite and elevational influences on early regeneration after catastrophic windthrow. J. Veg. Sci. 1:657-662. Phillips, D.L. and Shure, D. 1990. Patch-size effects on early succession in southern Appalachian forests. Ecology 71:204-212. Pickett, S.T.A. 1989. Space-for-time substitution as an alternative to long-term studies. In: Likens, G.E. (editor) Long-term studies in ecology. Approaches and alternatives. Springer-Verlag, N Y . pp 110-135. Pinkerton, G. 1994. Plantation performance and treatments: Abies in the Prince Rupert Forest Region. In: Proceeding of the Abies spp. workshop: problems and solutions, Feb. 8, 1994, Parksville, B.C., Forest Practices Branch, Victoria, B.C. Piatt, W.J. and Strong, D.R. 1989. Special feature: gaps in forest ecology. Ecology 70:535-576. Pojar, J., Klinka, K. and Meidinger, D.V. 1987. Biogeoclimatic ecosystem classification in British Columbia. For. Ecol. Manage. 22:119-154. Poulson, T.L. and Piatt, W.J. 1989. Gap light regimes influence canopy tree diversity. Ecology 70:553-555. Powers, R.F. 1989. Retrospective studies in perspective: strengths and weaknesses. In: W.J. Dyck and C A . Mees (ed.). Research strategies for long-term site productivity. Procc. IE A / B E A3 Workshop, Seattle, W A , Aug., 1988. IEA/B3 Report No. 8. Forest Res. Inst., New Zealand, Bull. 152. pp. 47-62. 189 Primack, R.B. and Hall, P. 1992. Biodiversity and forest change in Malaysian Borneo. BioSci. 42:829-837. Prochnau, A.E . 1963. Direct seeding experiments with white spruce, alpine fir, douglas fir and lodgepole pine in the Central Interior of British Columbia. Research Note No. 37, B.C. Ministry of Forests, Victoria, B.C. Putz, F.E. 1983. Treefall pits and mounds, buried seeds, and the importance of soil disturbance to pioneer trees on Barro Colorado Island, Panama. Ecology 64:1069-1074. Quine, C P . 1995. Assessing the risk of wind damage to forests: practice and pitfalls. In: Cloutts, M.P. and J. Grace (eds.). Wind and trees. Cambridge Univ. Press, pp. 379-403. Reader, R J . and Bricker, B.D. 1991. Response of five deciduous forest herbs to partial canopy removal and patch size. Amer. Midi. Nat. 127:149-157. Reader, R.J., Bonser, S.P., Duralia, T.E. and Bricker, B.D. 1995. Interspecific variation in tree seedling establishment in canopy gaps in relation to tree density. J. Veg. Sci. 6:609-614. Rebertus, A.J. , and Veblen, T.T. 1993. Structure and tree-fall gap dynamics of old-growth Nothofagus forests in Tierra del Fuego, Argentina. J. Veg. Sci. 4:641-654. Ribbens, E., Pacala, S.W., and Silander, J.A., Jr. 1994. Recruitment in northern hardwood forests: a method for calibrating models that describe patterns of tree seedling dispersal. Ecology 75: 1794-1806. Ricklefs, J.R. 1977. Environmental heterogeneity and plant species diversity: a hypothesis. Amer. Nat. 111:376-381. Ruel, J.C. 1995. Understanding windthrow: silvicultural implications. For. Chron. 71:434-445. Runkle, J.R. 1982. Patterns of disturbance in some old-growth mesic forests of eastern North America. Ecology 63:1533-1546. Runkle, J.R. 1992. Guidelines and sample protocol for sampling forest gaps. Gen. Tech. Rep. PNW-GTR-283, U.S. Dept. Agric , For. Serv., Pac. Northwest Res. Stn., Portland, Oregon. Runkle, J.R. and Yetter, T . C 1987. Treefalls revisited: gap dynamics in the southern Appalachians. Ecology 68:417-424. 190 Running, S.W. and C P . Reid. 1980. Soil temperature influences on root resistance of Pinus contorta seedlings. Plant Physiol. 65:635-640. Ruth, R.H. and R.A. Yoder. 1953. Reducing wind damage in the forests of the Oregon Coast Range. Res. Paper No. 7, U.S. Dept. Agric, For. Serv., Pac. Northwest Res. Stn., Portland, Oregon. SAS Institute Inc. 1989. SAS/STAT User's Guide, Version 6, Fourth edition, Volume 1, Cary, N C S A S Institute Inc. Sanford, R.L., Jr. 1989. Fine root biomass under a tropical forest light gap opening in Costa Rica. J. Trop. Ecol. 5:251-256. Sanford, R.L., Jr. 1990. Fine root biomass under forest light gap openings in an Amazon rainforest. Oecologia 83:541-545. Savill, P.S. 1983. Silviculture in windy climates. For. Abs., (Review article) 44:473-488. Schemske, D.W. and Brokaw, N . 1981. Treefall and distribution of understory birds in a tropical forest. Ecology 62:938-945. Schupp, E.W. 1988. Factors affecting post-dispersal seed survival in a tropical forest. Oecologia 76:525-530. Schupp, E.W. 1995. Seed-seedling conflicts, habitat choice, and patterns of plant recruitment. Amer. J. Bot. 82:399-409. Schupp, E.W., H.F. Howe, C.K. Augspurger, and D.J. Levey. 1989. Arrival and survival in tropical treefall gaps. Ecology 70:562-564. Sernander, R. 1936. The primitive forests of Granskar and Fiby. A study of the part played by storm-gaps and dwarf trees in the regeneration of the Swedish spruce forest. Acta Phytogeogr. Suec. 8:1-232 (in Swedish with English summary) Shelly, T.E. 1988. Relative abundance of day-flying insects in treefall gaps vs shaded understory in a Neotropical forest. Biotropica 20:114-119. Shugart, H . H. 1984. A Theory of Forest Dynamics. Springer-Verlag, New York. Shure, D.J. and Wilson, L . A . 1993. Patch-size effects on plant phenolics in successional openings of the southern Appalachians. Ecology 74:55-67. Sipe, T.W. 1990. Gap partitioning among maples (Acer) in the forests of central New England. Dissertation. Harvard University, Cambridge, Mass., USA. 191 Sipe, T.W., and Bazzaz, F.A. 1994. Gap partitioning among maples (Acer) in central New England: shoot architecture and photosynthesis. Ecology 75:2318-2332. Sipe, T.W., and Bazzaz, F.A. 1995. Gap partitioning among maples (Acer) in central New England: survival and growth. Ecology 76:1587-1602. Slocombe, D.S. 1993. Environmental planning, ecosystem science, and ecosystem approaches for integrating environment and development. Envir. Manage. 17:289-303. Smith, W.K. 1985. Western montane forests. In: Chabot, B.F. and Mooney, H.A. (eds.). Physiological Ecology of North American Plant Communities. Chapman and Hall, New York. pp. 95-126. Smith, D . M . 1986. The Practice of Silviculture. 8th Edition, John Wiley & Sons, New York. Smith, D . M . 1993. Silvicultural Systems Program: British Columbia Ministry of Forests. Canada-British Columbia Forest Resource Development Agreement, B.C. Min. For, Res. Br., Victoria, B.C. Smith, K. and R.H. Weitknecht. 1915. Windfall damage in selection cuttings in Oregon. Proc. Soc. Amer. Foresters. 10:363-365. Society of American Foresters. 1991. Task Force Report on Biological Diversity in Forest Ecosystems. Soc. Amer. For., Bethesda, M D . Sousa, W.P. 1984. Intertidal mosaics: propagule availability, and spatially variable patterns of succession. Ecology 65:1918-1935. Spies, T.A. 1991. Response of coniferous ecosystems to tree-fall gaps. Below-ground Ecol. 2:13-15. Spies, T.A., Franklin, J.F. and Klopsch, M . 1990. Canopy gaps in Douglas-fir forest of the Cascade mountains. Can. J. For. Res. 20:649-658. Stathers, R.J., T.P. Rollerson, and S.J. Mitchell. 1994. Windthrow handbook for British Columbia forests. Working Paper 9401, B.C. Ministry of Forests, Victoria, B.C.. Stewart, G.H. 1986. Forest development in canopy openings in old-growth Pseudotsuga forests in the western Cascade Range, Oregon. Can. J. For. Res. 16:558-568. Stewart, G.H. 1988. The influence of canopy cover on understory development in forests of the western Cascade Range, Oregon, USA. Vegetatio 76:76-88. 192 Stewart, G.H., Rose, A .B . , and Veblen, T.T. 1991. Forest development in canopy gaps in old-growth beech (Nothofagus) forests, New Zealand. J. Veg. Sci. 2:679-690. Systat, Inc. 1996. SYSTAT for Windows: Statistics, Version 6 Edition. Evanston, Illinois. Tappeiner, J . C , Lavender, D, Walstad, J, Curtis, R.O. and DeBell, D.S. 1997. Silvicultural systems and regeneration methods: current practices and new alternatives. In: Kohm, K . A . and Franklin, J.F. (Eds.) 1997. Creating a forestry for the 21st century. Island Press, Washington, D . C , pp. 151-164. Taylor, A . H . 1990. Distribution and persistence of Sitka spruce (Picea sitchensis) (Bong) Carr.) in coastal forests of the Pacific Northwest, North America. J. Biogeogr. 17:47-58. Taylor, A . H . , and Qin, Z. 1992. Tree regeneration after bamboo die-back in Chinese Abies-Betula forests. J. Veg. Sci. 3:253-260. Thompson, I.D. and Welsh, D.A. 1993. Integrated resources management in boreal ecosystems - impediments and solutions. For. Chron. 69:32-39. Tranquillini, W. 1979. Physiological ecology of the alpine timberline. Springer-Verlag, Berlin, Heidelberg, New York. Troup, R.S. 1928. Silvicultural Systems. Oxford University Press. Uhl, C , Clark, K. and Maquirno, P. 1988. Vegetation dynamics in Amazonian treefall gaps. Ecology 69:751-763. Vitousek, P .M. , and Denslow, J.S. 1986. Nitrogen and phosphorus availability in treefall gaps of lowland tropical rainforests. J. Ecol. 74:1167-1178. Waldron, R . M . 1966. Factors affecting natural white spruce regeneration on prepared seedbeds at the Riding Mountain Experimental Area, Manitoba. Pub. No. 1169, Canada Dep. of Forestry and Rural Development, Forestry Branch. Dep., Ottawa, Ontario. Walstad, J.D. and Kuch, P.J. 1987. Forest Vegetation Management for Conifer Production. John Wiley & Sons, New York. Walters, M . B. and Reich, P.B. 1996. Are shade tolerance, survival, and growth linked? Low light and nitrogen effects on hardwood seedlings. Ecology 77:841-853. 193 Walters, M . B., Kruger, E. L. , and Reich, P. B. 1993. Growth, biomass distribution and C O 2 exchange of northern hardwood seedlings in high and low light: relationships with successional status and shade tolerance. Oecologia, 94: 7-16. Wang, G. G., Quian, H. , and Klinka, K. 1994. Growth of Thuja plicata seedlings along a light gradient. Can. J. Bot. 72: 1749-1757. Watt, A.S. 1947. Pattern and process in the plant community. J. Ecol. 35:1-22. Wayne, P .M. and Bazzaz, F.A. 1993. Birch seedling response to daily time courses of light in experimental forest gaps and shadehouses. Ecology 74:1500-1515. Weetman, G.F. 1996. Are European silvicultural systems and precedents useful for British Columbia silviculture prescriptions? FRDA Rep. 239, Canadian Forest Service and B.C. Ministry of Forests, Victoria, B.C. Weetman, G.F. and D. Algar. 1976. Selection cutting in over-mature spruce-fir stands in Quebec. Can. J. For. Res. 6:69-77. Weidman, R.H. 1920. A study of windfall loss of western yellow pine in selection cuttings fifteen to thirty years old. J. For. 18:616-622. Whitmore, T.C. 1989. Canopy gaps and the two major groups of forest trees. Ecology 70:536-538. Wilczynski, C.J. and Pickett, S.T.A. 1993. Fine root biomass within experimental canopy gaps: evidence for a below-ground gap. J. Veg. Sci. 4:571-574. Worthington, N.P. 1953. Reproduction following small group cuttings in virgin Douglas-fir. Res. Note 84, U.S. Dept. Agric, For. Serv., Pac. Northwest Res. Stn., Portland, Oregon. Wright, E.F., Coates, K.D. , Canham C D . and Bartemucci, P. 1998. Species variability in growth response to light across climatic regions in northwestern British Columbia. Can. J. For. Res., in press. Wright, E.F., Coates, K.D. and Bartemucci, P. 1999. Regeneration from seed of six tree species in the interior cedar-hemlock forests of British Columbia as affected by substrate and canopy gap position. Can. J. For. Res., in press. 

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.831.1-0088978/manifest

Comment

Related Items