"Forestry, Faculty of"@en . "DSpace"@en . "UBCV"@en . "Kopra, Kristin"@en . "2009-10-28T23:25:07Z"@en . "2003"@en . "Master of Science - MSc"@en . "University of British Columbia"@en . "Engelmann Spruce-Subalpine Fir (ESSF) forests of interior British Columbia have increasingly\r\nbecome the target of forest harvesting in the past 40 years. Over the past decade, increasing\r\npublic concern over maintaining biodiversity and ecological health within forests has brought\r\nabout two primary pieces of legislation\u00E2\u0080\u0094The Forest Practices Code of 1995 and The Forest and\r\nRange Practices Act of 2002\u00E2\u0080\u0094that shape the current paradigm directing forest management.\r\nThis paradigm maintains that biological diversity can be preserved by designing forest harvesting\r\npractices that result in regenerated forests that closely mimic naturally disturbed forests. It has\r\nbeen suggested that fire is the primary form of natural disturbance in ESSF forests. To date, there\r\nhas been little knowledge gained about the natural disturbance patterns in ESSF forests, leaving\r\nforest managers with little guidance as to how to emulate these patterns. In this study, I sought to\r\nquantify both landscape-level and stand-level differences between naturally disturbed and\r\nharvested forests in the North Thompson variant of the wet, cold subzone of the ESSF\r\n(ESSFwc2) using the Geographical Information System Arcview and field measurement\r\ntechniques. A decrease in the amount of old-growth forest as well as a decrease in the mean\r\ndisturbance interval from naturally disturbed to harvested forests was found. Both patch size and\r\npatch size variability decreased with harvesting in older age classes and increased with\r\nharvesting in young age classes. A decrease in the amount of coarse woody debris and snags\r\nfrom naturally disturbed to harvested stands, as well as a significant difference in tree species\r\ncomposition between naturally disturbed and harvested stands was found. If emulation\r\nsilviculture is the goal of forest management, this study concludes that the current harvesting\r\nmethod of clearcutting is failing to do so and, therefore, alternative harvesting methods and/or\r\nnon-timber outputs should be considered."@en . "https://circle.library.ubc.ca/rest/handle/2429/14291?expand=metadata"@en . "5436748 bytes"@en . "application/pdf"@en . "EFFECTS OF NATURAL DISTURBANCE AND HARVESTING ON THE LANDSCAPE AND STAND LEVEL STRUCTURE OF WET, COLD ENGELMANN SPRUCE SUB ALPINE FIR FORESTS OF SOUTH-CENTRAL BRITISH COLUMBIA, CANADA by KRISTIN KOPRA B.A., Seattle University, 2001 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES THE FACULTY OF FORESTRY Department of Forest Sciences We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 2003 \u00C2\u00A9Kristin Kopra, 2003 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 writ ten permission. Department of jCTT-^J. The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT Engelmann Spruce-Subalpine Fir (ESSF) forests of interior British Columbia have increasingly become the target of forest harvesting in the past 40 years. Over the past decade, increasing public concern over maintaining biodiversity and ecological health within forests has brought about two primary pieces of legislation\u00E2\u0080\u0094The Forest Practices Code of 1995 and The Forest and Range Practices Act of 2002\u00E2\u0080\u0094that shape the current paradigm directing forest management. This paradigm maintains that biological diversity can be preserved by designing forest harvesting practices that result in regenerated forests that closely mimic naturally disturbed forests. It has been suggested that fire is the primary form of natural disturbance in ESSF forests. To date, there has been little knowledge gained about the natural disturbance patterns in ESSF forests, leaving forest managers with little guidance as to how to emulate these patterns. In this study, I sought to quantify both landscape-level and stand-level differences between naturally disturbed and harvested forests in the North Thompson variant of the wet, cold subzone of the ESSF (ESSFwc2) using the Geographical Information System Arcview and field measurement techniques. A decrease in the amount of old-growth forest as well as a decrease in the mean disturbance interval from naturally disturbed to harvested forests was found. Both patch size and patch size variability decreased with harvesting in older age classes and increased with harvesting in young age classes. A decrease in the amount of coarse woody debris and snags from naturally disturbed to harvested stands, as well as a significant difference in tree species composition between naturally disturbed and harvested stands was found. If emulation silviculture is the goal of forest management, this study concludes that the current harvesting method of clearcutting is failing to do so and, therefore, alternative harvesting methods and/or non-timber outputs should be considered. n TABLE OF CONTENTS Page ABSTRACT ii LIST OF TABLES v LIST OF FIGURES vi ACKNOWLEDGEMENTS vii CHAPTER 1: NATURAL DISTURBANCE AND HARVESTING IN ESSF FORESTS OF BRITISH COLUMBIA 1 1.1 Literature Review 1 1.2 Objectives 17 1.3 Study Area 18 CHAPTER 2: ESSF wc2 FOREST LANDSCAPE CHARACTERISTICS PRE AND POST HARVESTING 21 2.1 Introduction \u00E2\u0080\u00A2. 21 2.2 Objectives 23 2.3 Study Area 24 2.4 Methods 24 2.5 Results 32 2.6 Discussion 37 2.7 Conclusions 42 CHAPTER 3: STRUCTURAL CHARACTERISTICS OF STANDS REGNERATED AFTER NATURAL DISTURBANCE AND THOSE REGENERATED AFTER HARVESTING 43 3.1 Introduction 43 3.2 Objectives 48 3.3 Study Area 48 3.4 Methods 49 3.5 Results 57 3.6 Discussion 74 3.7 Conclusions 79 CHAPTER 4: CONCLUSIONS AND MANAGEMENT RECOMMENDATIONS 80 4.1 Landscape level management 80 4.2 Stand level management 84 4.3 Lessons Learned 89 in LITERATURE CITED 92 APPENDIX I: SITE SERIES, SLOPE, AND ASPECT OF SAMPLE AREAS 100 APPENDIX II: PER-PLOT MEASUREMENTS FOR NATURALLY DISTURBED AGE CLASSES 1-9 AND HARVESTED AGE CLASSES 1 AND 2 101 APPENDIX III: STATISTICAL TEST RESULTS 106 iv LIST OF TABLES Table Page 2-1. Age class categories according to the B.C. Min. For. digital database 26 2-2. Queries for area per age class (pre-harvest) 27 2-3. Pre vs. post harvest area per age class in the ESSFwc2 forests in the study area 35 2-4. Average patch size (km )^ and standard deviation for each age class of ESSFwc2 forests in the study area 35 2-5. Patch size range (km )^ for each age class of ESSFwc2 forests in the study area 35 2-6. Perimeter/Area ratio per age class (km/km )^ for ESSFwc2 forests in the study area 36 2- 7. Fire frequency estimations 37 3- 1. Plot locations within ESSFwc2 forests 49 3-2. Total number of plots established 53 3-3. Statistical variables and the tests used to compare them for natural disturbance vs. harvesting origin ESSFwc2 forest stands 58 3-4. Number of samples (n) required to estimate the true mean value of 4 stand variables to within 20% at the 95% confidence level 74 v LIST OF FIGURES Figure Page 1- 1. Study area within British Columbia 20 2- 1. Pre vs. Post-harvest Old-growth in the study area 34 2-2. Pre-harvest fire frequency 37 2- 3. Post-harvest fire frequency 37 3- 1. Location of sampled plots within the study area 50 3-2. Number of live trees/ha by age class 60 3-3. Mean tree dbh by age class 60 3-4. Species composition by age class 61 3-5. Height class composition by age class 62 3-6. Number of snags/ha by age class 62 3-7. Number of seedlings/ha by age class 63 3-8. Quantity of coarse woody debris by age class 63 3-9. Number of trees/ha in fire and harvest origin stands in each of the two age classes 65 3-10. Mean tree dbh of fire and harvest origin stands in each of the two age classes 65 3-11. Species composition (number of individuals) of live trees in fire and harvest origin stands in each of the two age classes 66 3-12. Treatment by location interaction for the number of fir in age class 1 67 3-13. Percent of the total number of trees occupied by each species in each of the two age classes 68 3-14. Number of snags/ha in fire and harvest origin stands in each of the two age classes 69 3-15. Number of seedlings/ha in fire and harvest origin stands in each of the two age classes. 69 3-16. Number of seedlings by species in fire origin and harvest origin stands in each of the two age classes 70 3-17. Percent of the total number of seedlings occupied by each species in each of the two age classes 71 3-18. Treatment by location interaction for the percent of the total number of seedlings represented by spruce in fire origin and harvest origin stands in age class 1 72 3-19. Quantity of coarse woody debris in fire origin and harvest origin stands in each of the two age classes 72 3-20. Treatment by location interaction for the quantity of coarse woody debris in fire origin and harvest origin stands in age class 1 73 v i ACKNOWLEDGEMENTS This project was made possible through funding from the University of British Columbia and Forest Renewal B.C. My work benefited immensely from discussions with my supervisor, Michael Feller over the past two years. I am grateful to Michael for sharing his knowledge and serving as a source of inspiration and support. I am indebted to Ryan Gandy for his unwavering support, help, sense of humor, and never-ending patience with my prehistoric computer skills. I am grateful to John Lewis for theoretical and philosophical discussions, as well as for help with my GIS work. I thank Jerry Maedel for providing me with access to data and an excellent lab in which to complete my GIS work. My statistical analysis benefited immensely from discussions with Peter Marshall, Ryan Gandy, and Heather Bears. Many thanks to Ryan Strenk and Paul Weary of Weyerhaueser and to Dennis Lloyd of the B.C. Ministry of Forests for providing me with maps that allowed me to locate my study plots. My field work would not have been possible without the help received from my field assistants, Christine Bonish, Erin Pierce, Norah White, Stephanie Pollack, and Derek Olive. Thank you\u00E2\u0080\u0094for help, friendship, and for cheerfully braving unprecedented swarms of mosquitoes for an entire summer. This project has more or less been my life for the past two years. First and foremost, I would like to thank my son, Riley, for being one of the most intelligent, compassionate, and uncommon kids around. I am especially grateful to Zac for allowing the ties that bind us to be loosened just enough to let happiness in. Thank you to Willow and Mariah for support and friendship; Fuzz for the wonderful gift of you; to Susan for laughter, a listening and supportive ear, and ample supplies of ice cream; to Dana for friendship and a home base; and to Rebecca (Beck) for assuring me I could do it and for serving as a reminder of where I come from as well as where I want to be headed. Thank you also to Dr. Gary Chamberlain and Dr. David Brubaker who have been mentors, as well as friends, and who continue to serve as reminders of the importance of living a purposeful life. Lastly, I'd like to thank MLDS for unique friendship, a deeper understanding, pushing me when I needed it, and the very best view. vii CHAPTER 1 NATURAL DISTURBANCE AND HARVESTING IN ESSF FORESTS OF BRITISH COLUMBIA 1.1 LITERATURE REVIEW History of forest harvesting in B.C. Timber harvesting has a long history as an economic base in British Columbia. Logging dates back to the 1850's when the commercial exploitation of forests was introduced to the future province of British Columbia in part as a result of the economic boom created by the California Gold Rush in the United States (Rajala 1998). The Gold Rush provided both a vision of economic prosperity through the exploitation of natural resources and the monetary capital to accomplish such exploitation. By 1867 mills had been established on Burrard Inlet that produced lumber for areas around the world, including South America, Asia, Australia, and the United States (Rajala 1998). Timber harvesting originally was restricted to primarily lowland areas that were located relatively short distances from bodies of water where pulp mills and transportation were located. While short lived consideration of selective logging existed throughout the history of forestry in B.C. (e.g. during the Great Depression when clearcutting was seen as a wasteful, overproductive form of harvesting), clearcutting continued to be heralded as the superior form of harvesting until the mid to late 1990's. As a result of an increase in public awareness and concern over the loss of forest ecosystem functions (including watershed health and global 1 temperature moderation, but primarily focused on biodiversity), the need for research into alternative methods of harvesting became necessary. The Forest Practices Code (FPC) of 1995 largely shaped the current paradigm directing forest management in British Columbia. The FPC was \"streamlined\" to what is now referred to as the Forest and Range Practices Act (FRPA) in December of 2002 (Ministry of Forests 2002). While the FRPA, which shifts forest management from a process based to a results based approach, remains in its infancy (e.g. it remains a set of guidelines; as of yet, it is uncertain how these guidelines will be implemented) there are indications that biodiversity will remain a high priority, although it has not yet been stated whether the new code will attempt to maintain the paradigm created by the initial Forest Practices Code. This paradigm maintains that biological diversity can be preserved by designing forest harvesting practices that result in regenerated forests that more closely mimic natural (unmanaged) forests (Anon 1995). Since forest structure and dynamics are known to be influenced primarily by natural disturbance regimes (e.g. fire and insects), it has become clear that increased scientific knowledge into the natural disturbance regimes of forests is critical. The B.C. Ministry of Forests (MoF) has, in recent years, initiated research into the effects of both natural disturbance regimes and various methods of harvesting on forests (Anon 1995), reflecting the general consensus that sustainable forestry can be recognized only through the understanding of natural processes and the subsequent implementation of appropriate silvicultural concepts and applications. Today, forestry in British Columbia accounts for approximately six percent of total employment in the province (90,000 jobs) and an estimated 15 percent of provincial economic activity (B.C. Min. For. 2002a). These factors, coupled with the fact that most of the forest land in low lying and coastal areas has been harvested, have resulted in an increase in logging in high elevation forests, including the Engelmann Spruce-Subalpine Fir (ESSF) forests of the interior. There are still large gaps in scientific knowledge about these forests, specifically in regards to the 2 effects of fire, and other natural disturbances on stand dynamics and structure. If the goal of management is to harvest in ways that mimic natural disturbance patterns, it is crucial these gaps are filled in order to manage these forests in a fashion that mimics nature as closely as possible. ESSF Zone The ESSF biogeoclimatic zone is one of fourteen zones in British Columbia, each of which represents a distinct combination of climate, physiography, vegetation, and soil (Meidinger and Pojar 1991). The zone is divided into fifteen forested subzones based on regional climatic differences. The ESSF zone incorporates much of the high-elevation forests within the interior of British Columbia and is found adjacent to alpine tundra in the Rocky, eastern Coast, Columbia, and Skeena mountain ranges, and the Interior Plateau (Meidinger and Pojar 1991). It is generally comprised of rugged mountainous terrain and ranges in elevation from 1200-2300 m. Climate is cold, moist, and snowy; growing seasons are cool and short and winters are cold and long. Mean annual temperatures range from -2\u00C2\u00B0C to +2\u00C2\u00B0C with maximum mean monthly temperatures rarely exceeding 10\u00C2\u00B0C (Meidinger and Pojar 1991). Mean annual precipitation ranges from 400 mm (dry subzones) to 2200 mm (wet subzones) (Meidinger and Pojar 1991). Mineral soils are dominantly podzolic and are moderately well drained. Organic soils and wetlands are localized in depressions and poorly drained lower slopes (Meidinger and Pojar 1991). As its name indicates, the dominant tree species in the ESSF zone are Engelmann spruce (Picea engelmannii) and subalpine fir (Abies lasiocarpa). Lodgepole pine (Pinus contorta) and trembling aspen (Populus tremuloides) occur as serai species in warmer parts of the zone, with lodgepole pine occurring the most frequently (Farnden 1994; Meidinger & Pojar 1991). Common understory plants include oval-leafed blueberry (Vaccinium ovalifolium), black 3 huckleberry (Vaccinium membranaceum), false azalea (Menziesia ferruginea), white rhododendron (Rhododendron albiflorum), bunchberry (Cornus canadensis), creeping raspberry (Rubuspedatus), queen's cup (Clintonia uniflora), and red-stemmed feathermoss (Pleurozium schreberi). Wildlife species in the ESSF include ungulates such as moose (Alces alces), mountain goat (Oreamnos americanus), woodland caribou (Rangifer tarandus), mule deer (Odocoileus heminous); amphibians, including the western toad (Bufo boreas), spotted frog (Rana pretiosa), Cascades frog (Rana cascadae), Rocky Mountain tailed frog (Ascaphus montanus), and long-toed salamander (Ambystoma macrodactylum); grizzly bear (Ursus arctos); seed-eating birds such as the red crossbill (Loxia curvirostra), white-winged crossbill (Loxia leucoptera), pine siskin (Carduelispinus), Clark's nutcracker (Nucifraga columbiana), varied thrush (Ixoreus naevius), spruce grouse (Falcipennis canadensis), mountain chickadee (Poecile gambeli), winter wren (Troglodytes troglodytes), Steller's jay (Cyanocitta stelleri), orange-crowned warbler (Vermivora celata), and Cassin's finch (Carpodacus cassinii); and furbearers such as the marten (Martes americana), fisher (Martes pennanti), red squirrel (Tamiasciuras hudsonicus), and wolverine (Gulo gulo) (Meidinger and Pojar 1991). Under the current Species at Risk Act (SARA), the woodland caribou is listed as threatened, while the Rocky Mountain tailed frog is listed as endangered (Environment Canada 2002). The wolverine, grizzly bear, and western toad are all listed as species of \"special concern\", meaning they are in danger of becoming threatened and/or endangered if steps are not taken to conserve their populations and habitats (Environment Canada 2002). While Engelmann spruce can outlive subalpine fir, subalpine fir dominates most stands at higher elevations (Farnden 1994). This has led some ecologists to speculate that subalpine fir will, in the absence of disturbance, out compete spruce, becoming the climax species in a steady state forest. Throughout stand development, fir recruitment appears to be continuous; spruce recruitment, however, is absent during the stem exclusion stage and only reappears after a canopy break up of the colonizing cohort (Knapp & Smith 1982; Aplet et al. 1988; Varga and Klinka 2001). Each of these studies was conducted in drier, warmer ESSF forests, leaving some uncertainty in extrapolating to the ESSFwc subzone. Varga and Klinka (2001) have concluded that subalpine fir trees can survive long periods of suppression, allowing them to outlive vegetative competition and/or closed canopy conditions. On the other hand, Antos and Parish (2001) found that seedling bank is critical to the dynamics of both species and that both species were able to survive long periods of supression; furthermore, their study concluded that spruce was more abundant in the seedling bank than in large trees, indicating that future composition of the canopy would most likely constitute a significant amount of spruce. In addition, Antos and Parish (2001) noted a much lower mortality rate among spruce than fir; this coupled with the longer life span of spruce (Alexander 1987) may ensure the continued co-existence of the two species. While it is still uncertain whether fir will become a climax species in the absence of disturbance, most literature to date seems to indicate that the two species co-exist due primarily to different life histories. Subalpine fir Subalpine fir is the smallest of the true fir species in British Columbia and is distinguished by a long, narrow conical crown ending in a noticeable sharp spike point (Burns and Honkala 1990). While seed production can begin when trees are 1.2-1.5 m tall (=20 years), maximum production usually occurs in canopy trees aged 150-200 years (Alexander et al. 1984). Although lodgepole pine is the dominant pioneering species on disturbed sites in lower elevation ESSF forests, subalpine fir commonly colonizes sites in higher elevations. This is due to the facts that (1) lodgepole pine generally requires drier sites and (2) subalpine fir can establish root systems in harsher conditions than its less hardy associates, Engelmann spruce and lodgepole pine. Subalpine fir is not as good a seed producer as Engelmann spruce, although it can 5 germinate on a much wider range of seed beds. Subalpine fir shows dispersal patterns similar to those of Engelmann spruce. Nearly all seed dispersal is by wind and over half occurs within 30 m of the windward forest edge (Noble and Ronco 1978). Although it was initially thought that optimal seedbeds for subalpine fir consisted only of exposed mineral soil (Burns and Honkala 1990), Alexander et al. (1984) observed subalpine fir germinating and surviving on a wide range of seedbed types including undisturbed forest floor, undecomposed duff and litter, and decaying wood. Feller (1998) found that subalpine fir seedlings preferred mineral soil over burned or undisturbed forest floors. Although in the ESSF subalpine fir is considered to be very shade tolerant (Franklin and Dyrness 1973), Feller (1998) found that shading adversely affected germination and initial seedling survival. Subalpine fir is extremely fire sensitive because of several key factors including: (1) thin bark, (2) a shallow root system, (3) fine needles (which act as fine fuel), and (4) a vertical ladder of fuel created by inordinate amounts of accumulation under trees (due to slow rates of decomposition) and limbs that grow close to the ground (Walstad et al. 1990). Engelmann spruce Engelmann spruce is one of the largest and longest-lived conifers in high-elevation mountainous regions. Average maximum age ranges from 300-450 years and trees 500-600 years are not uncommon (Alexander 1987). It can begin producing seeds between 15 and 40 years; however, in closed canopy stands cone production often starts when trees are much older (Alexander and Shepperd 1990). Seed dispersal is driven by wind, with 40% of seeds falling within 30 m of the windward forest edge (Noble and Ronco 1978). Engelmann spruce is less shade tolerant than subalpine fir although initial establishment and growth are favored by 40-60% shade (Burns and Honkala 1990). This is supported by Feller (1998) who found that, while the exposed conditions of clearcuts favored germination for 6 Engelmann spruce seeds, they appeared to be less beneficial for initial seedling survival. Engelmann spruce is generally more exacting in its seed bed requirements and prefers exposed mineral soil and humus (Day 1964). However, McCaughey et al. (1991) found that while germination success in clearcut areas was best on mineral soil, partially cut areas had highest germination success on forest floor seedbeds. Engelmann spruce trees are very sensitive to fire due to several factors, including (1) thin bark, (2) shallow root system, (3) heavy lichen growth, (4) resin in bark that promotes easy ignition, and (5) numerous low, dead branches providing vertical fuel ladders (Crane 1982). Fire History of British Columbia In British Columbia and the Pacific Northwest of the United States, fire has played an important role in the development and maintenance of different forest types for many thousands of years. Pyne (1997) documented the fire practices of Native Americans prior to European settlement and the effects these fires had on forested landscapes. Fire was used frequently to clear surrounding woods of underbrush to prevent the attack of hostile forces and to create fields that would attract grazing animals that could then be hunted. So frequent and vast were these burns that were it not for European settlement and the pyrophobia that accompanied it, it is possible that at least the lowland forests of North America would have been converted into and maintained as grasslands or savannas (Pyne 1993, 1997). The role of fire has changed drastically over the last century. Beginning in the early 20th century and continuing until the late 1960's land management institutions saw fire as a threat to forest management and human settlement. Fire was portrayed as an enemy to be conquered, and it was extinguished nearly everywhere. This view of fire, however, resulted in drastic changes to natural fire regimes of forests. Forests that would normally burn once every 20 years were left unburned, affecting the structure and ecology of such forests. Furthermore, fire prevention and 7 suppression actually created a more volatile fire environment because it allowed for the accumulation of large loads of fuel that increased the severity and intensity of burns when they did occur (Agee 1993). While fire began to be reintroduced to forests in the late 1960's, it has only been in the last decade that fire has been valued as an integral part of forest ecology. National Parks such as Banff National Park in Alberta have made efforts to reintroduce fire into the land in hopes of returning natural disturbance regimes to these ecosystems. Unfortunately, these efforts have been unsuccessful due to confounding effects of wildlife (which have been severely restricted in their habitat range due to development in the parks) on newly established seedlings and saplings. Efforts are currently under way to solve these problems, and it is a goal of Parks Canada to return fire into the ecosystems there (personal communication, Rick Searle, University of Victoria, Dept. of Geography, 2003). Fire regimes and ESSF forests Fire regimes refer to the nature of fire occurring over long periods of time and the predominant effects of fire that define an ecosystem (Hubbard et al. 2000). Factors used to describe fire regimes include fire frequency, fire periodicity, fire intensity, size of fire, landscape patterns created by fire, and season of burn (Kilgore 1987). The B.C. Ministry of Forests recognizes five natural disturbance types based primarily on fire periodicity, landscape patterns, area burned, and fire frequency (Anon 1995). ESSF forests fall into several different natural disturbance types depending on which subzone they are in. The rest of this discussion will focus on the ESSF wet, cold (wc) subzone. This subzone is of particular concern for several reasons: (1) it occupies a relatively large area in B.C., (2) this subzone is unique to B.C., and (3) it has become important for harvesting in the interior of British Columbia. 8 The ESSFwc is classified as natural disturbance type (NDT) 1\u00E2\u0080\u0094ecosystems with rare stand-initiating events. Lightning has historically been the major cause of fires in these forests. The mean return interval for fires in these disturbance types is considered to be 250-350 years (Anon. 1995). However, these return intervals can vary greatly and may not be cyclic (Walstad et al. 1990). Hawkes et al. (1997) reported fire cycles ranging from 800 to 2,000 years in the ESSF wet, cool (wk) subzone, supporting the idea that fire cycles in the ESSF zone are irregular and less frequent than believed. These fires have resulted in multi-storied uneven aged stands. Fires have typically resulted in irregular landscape patterns. Fires can be severe, with mineral soils being exposed (organic matter and top layers of soil are burned off). Entire stands are often burned due to: (1) continuous fuel ladders, (2) infrequent fires resulting from large accumulations of fine fuel (e.g. needles, branches), and (3) the extreme fire sensitivity of both dominant tree species. These disturbances result in major shifts in forest structure and function. Large areas of forest can be eliminated with a single fire, and it may take a century or more before herbaceous colonizers give way to tree species (Walstad et al. 1990). Alternatively, fires can burn in a patchy way such that islands of trees are left standing and alive. Such shifts can alter the composition of forests, change production potential, and can result in substantial nutrient release as ash to soils and atmosphere (Walstad et al. 1990). The effects of fire on soils are well known and include a potential increase in erosion and mass wasting as well as fluctuations in soil temperature (Walstad et al. 1990). In the ESSF, erosion is of more concern than in lower elevation zones because of generally higher precipitation and steeper slopes. Mass wasting is erosion by detachment and transportation due to gravity. Mass wasting is a common concern in the wetter biogeoclimatic zones, such as the Interior Coastal Hemlock (ICH), the wetter Sub-Boreal Spruce (SBS) and ESSF, and the Boreal White and Black Spruce (BWBS); but it can be a concern in any zone (B.C. Min. For. 1997). Because of the high amounts of precipitation and the steep slopes often found in wet ESSF 9 forests, the hazard of mass wasting can be increased with fire due to soils being exposed. The effects that fire has on soil temperature are perhaps of greatest importance in determining forest successional patterns. Farnden (1994) suggests that temperature is the most important environmental factor influencing seedling survival. Because of the cold, wet climate of the ESSFwc subzone, the temperature effects of fire on soils are most likely beneficial for plant growth, creating more post-burn vegetative cover and aiding seedling germination and survival through increased soil temperatures. The effects of fire can vary with aspect (Parminter 1983). Soil heating is less on north than on south facing slopes because snow remains longer, resulting in longer fire intervals, potentially less intense fires, and less overall solar radiation. In addition, black charred ground left after a fire absorbs more heat than undisturbed ground, causing an increase in soil temperature on areas burned (more often found on south facing slopes). Research on the impacts of fire in high elevation forests has been limited in British Columbia, most likely because of the low frequency of fire and the difficulties this poses for research. Additional factors such as extreme climate and high cost of research may also contribute to the lack of literature. Literature that is available presents varying scenarios of stand successional stages after fire. Varga and Klinka (2001) found that old growth stands in the ESSF moist cold (mc) subzone were dominated by subalpine fir. They also found similar proportions of spruce in seedlings, saplings, and trees, and they concluded that subalpine fir will eventually become dominant in the absence of disturbance. Antos and Parish (2001), likewise, found subalpine fir dominating old growth stands in ESSFwc forests; however, in contrast to Varga and Klinka's findings, Antos and Parish found a higher proportion of spruce in seedlings than in trees, and they concluded that the two species would continue to co-exist in the absence of disturbance. While Parminter (1983) stated that subalpine fir would become the dominant species in the absence of fire, he contended that fire has usually prevented subalpine fir from becoming dominant in ESSF forests. 10 Severe fires in the ESSF expose mineral soil, providing optimal seedbeds for both Engelmann spruce and subalpine fir. Feller (1998) contended that exposed mineral soil was the optimum seedbed for fir germination and initial growth. This same study showed that exposed mineral soil enhanced germination for spruce, while reducing intial growth. Because fires burn unevenly, leaving some large trees (both living and dead) standing, partial shade (required by both species' seedlings) is provided. It is likely that fire is responsible for the existence of Engelmann spruce in these forests, due to its provision of appropriate conditions for spruce regeneration. Regeneration of both species seems to depend largely on several factors including: (1) The number of trees left for seed recruitment (2) The pattern of the burn. Gaps must be large enough to allow for adequate light for the shade intolerant Engelmann spruce seedlings. (3) The severity of the burn. Exposed mineral soil provides the best seed bed for both Engelmann spruce and subalpine fir. (4) The post-fire climate. If the ground is still snow covered, seed germination will be difficult. Non-fire disturbances in ESSF forests While fire is a major disturbance factor in the structure of ESSF forests, several studies suggest that inter-fire small disturbances (e.g. insect infestation and windthrow) have equally important effects on stand dynamics (Alexander 1987: Parish et al. 1999). While it is beyond the scope of this study to discuss these disturbances in detail, it is necessary to understand their importance in relation to fire in determining stand structure. Some insects such as wood borers and secondary beetles attack only dead trees; thus, as the amount of dead trees declines (e.g. through fire and/or harvesting), so does the incidence of insect outbreaks (Lewis and Lindgren, 2000). However, certain insects such as the western balsam and spruce bark beetles are capable 11 of switching from dead trees as hosts to live trees (Anon 1995; Lewis and Lindgren, 2000). Insect infestation can result in either gaps being created or entire stands being infested and destroyed, depending on the type of insect and the susceptibility of the trees to the insect. Furthermore, host-fungus equilibria have evolved in most forests and, although sometimes unfavorable to management objectives, serve as necessary protection to residual and regeneration trees (Lewis and Lindgren, 2000). Parish et al. (1999) suggested that the composition of ESSFwc2 forests is a result of different types of disturbances with varying intensities. Antos and Parish (2001) concluded that small openings created by single tree mortality (due primarily to insect infestation) may be large enough to encourage and support spruce reinitiation, noting that the narrow, conical shape of both species allows more light to the forest floor than other species in other forest types allow. It is important, when considering disturbance factors in ESSFwc forests, as well as when preparing silvicultural prescriptions, to take into consideration the important role that insects and pathogens play in stand structure and succession. Traditional Management of ESSF forests It is widely believed that harvesting is replacing fire as the dominant disturbance process shaping landscape patterns of forests (Delong and Tanner 1996; Brown and Smith, 2000). As forested areas in low elevations have been harvested, ESSF forests, previously considered uneconomic to harvest, have become the target of harvesting activities. Most harvesting in the last 40 years has consisted of clearcutting (Farnden 1994). Clearcutting is a silvicultural system in which an entire stand of trees is cleared from an area at once. Clearcutting results in wider extremes of temperature ranges which can have large impacts on soil warming and frost protection (Farnden 1994). Soil warming is usually increased due to the exposure of soil to sunlight; however, this effect is reversed at night if soil is exposed (e.g. if slash has been burned) because the soil is less capable of retaining heat. Clearcutting provides 12 less frost protection and can result in a higher frequency of frost heaving, whereby seeds are pushed from the ground due to a cyclical freezing and thawing of exposed soil. While clearcutting usually results in less risk of windthrow than other silvicultural systems, it can pose a threat if careful attention is not paid to locating wind-firm cutting limit boundaries (Alexander 1964, 1967). Both Engelmann spruce and subalpine fir are very susceptible to windthrow because of their shallow rooting systems. There are notable differences in landscape patterns between harvesting and fire. Perhaps the greatest difference affecting patterns is the difference in tree selection. Fire often burns unevenly and is not as selective as clearcutting as to the age of trees it will burn. Some large trees may survive fire and dead trees (snags) are left standing. These standing trees offer wildlife habitat and also provide both seed banks and protection for seedlings. Clearcutting, on the other hand, occurs mostly in older forests due to the fact that trees in these forests are larger and more economical to harvest and their biomass increment is often less than that of younger forests so they are less productive from an economic standpoint than younger forests. Clearcutting can result in the removal of all trees in a stand or in small patches of mature trees being left as seed sources for natural regeneration. In addition, clearcutting leaves large amounts of slash which covers up mineral soil needed for optimal seedling germination. While broadcast burning of slash is a useful practice in other zones, it must be applied with extra caution in clearcut blocks in the ESSF because of the extreme sensitivity to fire of both dominant species. Care must be taken to ensure that adjacent stands are not affected by broadcast burning, as was the case with an escaped slashburn near the Sicamous Silvicultural Systems Project Area in 1998. This slashburn crept into adjacent stands, burning approximately four to five hectares (personal observation 2002). Knowledge of fire weather conditions and making intelligent informed decisions about how, where, and when to begin burning is critical if prescribed burning is to be successful and controlled in ESSF forests. Ignition and spread of fire may be difficult due to the 13 frequently high fuel moisture content and the uneven, broken topography. Broadcast burning can expose mineral soil, creating more favorable seedbeds and adding to the heat absorption capabilities of soils. It can also add nutrients (as \"natural\" fire does) to the soil by depositing ash on the surface. Another difference between harvested and burned landscapes is the size of disturbance. DeLong and Tanner (1996) showed that fires in the Sub Boreal Spruce zone have either consumed large areas (> 500 ha) or small areas (< 50 ha) of forest, supporting the idea that fires rarely create uniform landscape patterns. More often than not, large-scale disturbances such as fire result in irregular landscape patterns, alternating between large and small gaps. Clearcut areas are usually limited in size, creating a much different landscape mosaic than fire. Cutblocks tend to be uniform in size and more evenly spaced than gaps created naturally by fire. Therefore, while patches may be closer together overall, the variety in gap size created by fire is absent in clearcut areas. Clearcut patches are often replanted, which can result in the evolution of even-aged, single species stands which can differ greatly from the mosaic of age and species diversity created by fire. Vertical structural diversity is important for animal habitat, and it influences patch dynamics. Stands resulting from clearcutting may have relatively little vertical structural diversity, thus removing this important attribute of natural ESSF forests. The difference in frequency of disturbance between fires and clearcutting affects forest structure. Hawkes et al. (1997) reported that fires occur in the ESSFwk subzone every 800 to 2,000 years. Parish et al. (1999) found that fire and intermittent disturbances, caused mainly by bark beetles, were responsible for current forest stands. Insect infestations can happen slowly, causing one tree at a time to fall down, or they can result in small to large sized gaps created by many trees dying at once. Disturbance, then, can be an ongoing process occurring on small and large scales all the time in a forest (although this is not always the case). Clearcutting, on the 14 other hand, does not occur on either extremely small (e.g. single tree) scales or on extremely large scales (due to maximum cutblock sizes set by the Ministry of Forests). Fire also results in dead trees (snags) being left standing or on the ground. Snags that remain standing, provide necessary habitat for many animal species and serve as a long-term \"bank\" of nutrients for forest soils through decomposition. Partially burned logs remain on the forest floor, providing organic matter for soil (when they decompose) and seedbeds and shade for fir germinants. Clearcutting adds woody debris in the form of branches, but it leaves relatively few snags for habitat or nursery logs on the ground, thus disadvantaging those species which are enhanced by coarse woody debris, including many lichens, mosses, and small mammals (Song 1997). Alternative Methods of Harvesting Steventon (1997) proposed that if mimicking natural disturbance regimes is the goal of management, even-aged management is acceptable for zones such as the Sub-Boreal Spruce and Sub-Boreal Pine-Spruce (SBPS) zones, but the dominant single tree and patch mortality dynamics in ESSF zones are better simulated by partial cutting. Shelterwood and selection cutting are two possible alternative silvicultural methods for ESSF forests. Shelterwood cutting involves harvesting so groups of trees are left for seed banks and to protect regeneration. While shelterwood systems offer seed protection, soil warming is limited when compared with clearcutting. If broadcast burning occurred after shelterwood harvesting, soil temperatures would increase and mineral soil might be exposed. However, broadcast burning after shelterwood cutting is not recommended because of the severe susceptibility of both dominant tree species to fire. Selection cutting involves cutting individual trees or groups of trees in hopes of maintaining an uneven stand structure. Of the three harvesting methods discussed, selection 15 cutting creates the smallest gaps and results in the least disturbance to the forest canopy. It seeks to strike a balance between harvesting and the \"natural\" forest structure. This type of harvesting will not necessarily create essential gaps for Engelmann spruce recruitment, although recent studies (Antos and Parish 2001) suggested that small gaps may be enough for spruce recruitment. Like shelterwood cutting, selection cutting will result in less soil warming than clearcutting. If selection cutting is applied, it is possible that subalpine fir will become the dominant species and that, eventually, Engelmann spruce will make up an extremely small portion of these forests. Summary Fire has played an important role in determining the structure and landscape patterns of ESSF forests. Harvesting is rapidly replacing fire as the major disturbance in these high elevation forests, and it is creating landscape patterns very different from natural ones (Steventon 1997). If biodiversity, watershed health, and long-term productivity are to be achieved, it is crucial to understand which silvicultural prescription(s) is (are) most appropriate for these forests. Clearcutting may have adverse effects on forest regeneration by creating unnatural gaps between patches and by making seeds more susceptible to death by frost heaving and seedlings more susceptible to death from increased exposure to direct light. Moreover, clearcutting causes an increase in the possibility of soil erosion which negatively impacts both seedlings and established stands. While it adds certain kinds of woody debris, it may leave limited numbers of half burned logs and/or larger logs lying on the ground, decreasing the natural process of decay that is vital to forest nutrient cycling and the additional benefit of good seedbeds. Clearcutting has the potential to create forest edges that are more susceptible to windthrow, if careful attention is not paid to the location of wind-firm cutting limit boundaries. 16 Shelterwood and selection cutting systems are both viable harvesting methods. Both seem to favor the regeneration of subalpine fir over Engelmann spruce (due to the lack of gaps created and the layer of forest floor left remaining). It is important to realize, however, that these silvicultural systems are not a true replacement for natural disturbance. It is assumed that once harvesting of any type begins in these forests, fire will most likely be prevented\u00E2\u0080\u0094due both to the nature of settlement (here, harvesting the land) and the susceptibility of the dominant tree species to it. Fire prevents inordinate amounts of fine fuel from accumulating under spruce and fir (which happens due to the physiology of both species), keeping long continuous fuel ladders from forming. Fire adds to the structural diversity and nutrient cycling in soils within stands. None of these silvicultural systems produce all the effects of fire in their applications. 1.2 OBJECTIVES While there have been some studies of natural disturbances in ESSF forests conducted at the stand level (e.g. Hollstedt and Vyse 1997; Jull and Stevenson 2000), to date, there have been few studies conducted at the larger (landscape) level. As Sachs et al. (1997) point out, stands cannot be managed wisely and efficiently in isolation. Both disturbance regimes and harvesting can fundamentally alter landscape patterns (Franklin and Forman 1987; Ripple et al. 1991; Sachs et al. 1997); furthermore, they do so in different ways. In view of this lack of information, this study commenced in 2001 with the broad objective of quantifying the structure at the landscape and stand level of ESSFwc2 forests subjected to a natural disturbance regime and comparing it to the structure created by harvesting. Specific objectives were to determine: (1) Age class distribution of tree stands (2) Patch characteristics throughout the ESSFwc2 (3) Landscape Diversity 17 (4) The fire frequency for pre and post-harvested landscapes of the North Thompson biogeoclimatic variant of the wet, cold subzone of the ESSF zone (ESSFwc2). (5) Stand structure characteristics in both post-fire and post-harvested stands. These included (a) abundance and horizontal distribution of seedlings, live trees, and snags, (b) percent composition of species, (c) vertical distribution of live trees, and (d) abundance of coarse woody debris. I compared stands that had historically burned with those that had been harvested. Most harvesting within this area had taken place in the last 40 years. Comparison of the two different types of disturbance (i.e. natural vs. human) should indicate what steps, if any, need to be taken to move harvested landscapes into an acceptable range of natural landscape variability. 1.3 STUDY AREA The study area for this project was primarily in the Monashee mountain range in south central British Columbia. The study area encompassed most of the North Thompson wet, cold Engelmann-Spruce Subalpine Fir variant (ESSFwc2). The ESSFwc2 variant is probably a uniquely British Columbian variant (Feller 1998) and is characterized by cold temperatures (mean annual temperatures range from -2\u00C2\u00B0C to +2\u00C2\u00B0C), and high precipitation (up to 2200 mm/yr. with snowpacks ranging from 1-4 m). Engelmann spruce and subalpine fir are the dominant tree species; other characteristic species include oval-leaved blueberry (Vaccinium ovalifolium), oak fern (Gymnocarpium dryopteris), Sitka valerian (Valeriana sitchensis), black huckleberry (Vaccinium membranaceum), and false azalea (Menziesiaferruginea). The variant covers 7.89 x 103 km2; my study area covered a total of 6.46 x 103 km2, representing 82% of the total ESSFwc2 area. The remainder of the ESSFwc2 area could not be included in my study area either because the area fell under a different jurisdiction than the MoF (e.g. B.C. Parks) and, therefore, GIS data were unavailable, or because the data in the spatial files were inaccessible 18 therefore, GIS data were unavailable, or because the data in the spatial files were inaccessible due to errors within the MoF's database that I used. Approximate boundaries for my study area were the towns of Valemount to the north, Clearwater and Kamloops to the west, and Vernon to the south, and the Rocky Mountains to the east (Figure 1-1). 19 2 0 CHAPTER 2 ESSFwc2 FOREST LANDSCAPE CHARACTERISTICS PRE AND POST HARVESTING 2.1 INTRODUCTION As stated in the first chapter, the Forest Practices Code (FPC) of 1995, and the anticipation of its streamlined version, the Forest and Range Practices Act (FRPA) of 2002, have largely shaped the current paradigm directing forest management in British Columbia today. This paradigm maintains that biological diversity can be preserved by designing forest harvesting practices that result in regenerated forests that more closely mimic natural (unmanaged) forests (Anon 1995). This mimicking is commonly referred to as \"emulation silviculture\". Furthermore, there is increasing support for maintaining ecological integrity as a management goal (Panel on the Ecological Integrity of Canada's National Parks 2000; McRae et al. 2001; B.C. Min. For. 2002b). An ecosystem is said to have integrity when \"it is deemed characteristic of its natural region, including the composition and abundance of native species and ecological communities, rates of change, and supporting processes\" (Panel on the Ecological Integrity of Canada's National Parks 2000). It has been suggested that the natural disturbance regime in ESSF forests involves widespread fire followed by smaller disturbances, most likely from insects (Parish, 1997; Parish et al. 1999). If emulation silviculture is possible, and if it is to stand a chance in being effective, it is essential that quantitative data is obtained regarding the effects and dynamics of natural disturbances on the landscape. Delong and Tanner (1996) found significant differences in 21 (SBSmkl) biogeoclimatic unit, including (1) a higher proportion of large (>500 ha) and small (<50 ha) forest patches in burned landscapes and (2) more complex shapes in burned landscapes and patches of mature forest within them. Because ESSF forests have become popular harvesting landscapes, studies such as Delong and Tanner's would be advantageous to determine if significant differences are, similarly, present in ESSF forested landscapes. McRae et al. (2001) make the salient point that fire effects must also be linked to fire behavior in order for emulation silviculture to be successful. In other words, there needs to be proof that the effects seen from natural disturbances are, indeed, caused by the disturbance and not other factors (e.g. chance, climate change, time) before decisions can be made in support of emulation silviculture. While it was beyond the scope of this project to link such effects, it is important to remember this point when making any decisions for emulation silviculture based on the present and subsequent studies. Geographical Information Systems Geographic Information Systems (GIS) are computer-based systems used to store and manipulate geographic information; they are most commonly employed in situations where large quantities of data are involved (Aronoff 1989). GIS have become accepted as necessary tools to effectively store and manipulate geographic data (Aronoff, 1989). Information such as forest cover data used to be available to researchers only by going through a lengthy process of contacting BC Ministry of Forests (MoF) officials and requesting that information be sent. In the last few years, however, the MoF has made this information available through a public access database, in theory making the collection of data a much easier and more efficient process. Researchers now have the ability to identify and relay information about landscape features and patterns of interest before embarking on field work that often follows GIS analysis. The accuracy of studies conducted using GIS to analyze data from the MoF is dependent on the accuracy of the 22 initial data (Aronoff 1989). While this seems an easy pitfall to avoid (i.e. meticulous input of accurate data in the original data source), the occurrence and subsequent consequences of inaccuracy of data will be illustrated and discussed in this and subsequent chapters. 2.2 OBJECTIVES This study had a broad objective of assessing the differences between naturally disturbed and harvested high elevation ESSFwc2 forested landscapes by utilizing the GIS Arcview to analyze data from the MoF database. Specifically, it was my goal to develop a starting point for describing landscape patterns in ESSFwc2 forests from which subsequent studies could expand upon. As McRae et al. (2001) point out, patch sizes created by logging are but a small subset of those created by fire. This is supported by Delong and Tanner (1996), as previously noted. If the goal of forest management is to mimic natural disturbance patterns and regimes, it is essential that we gain a better understanding of the variability of naturally disturbed landscapes (e.g. patch size and arrangement). My initial hypotheses were: (1) Harvesting creates a more uniform landscape than fire, with patch sizes being determined by provincial guidelines. Patch sizes have distinct boundaries and are usually rectangular or square shaped (although this is not always the case), whereas patches created by fire tend to have irregular shapes, due to the fact that fire usually burns in elliptical patterns and often leaves clumps of live (and dead) trees standing. Consequently, naturally disturbed landscapes have greater perimeter/area ratios for forest patches. (2) Landscape diversity increases with harvesting, due to the fact that former old growth gets distributed throughout younger age classes. (3) The mean fire return interval for ESSFwc2 forests is longer than the 250-350 years indicated by the B.C. Min. For. The mean fire return interval is defined by Seymour and 23 Hunter (1999) as the average time between fire occurrences in a given stand. Fire frequency is one of the factors used in determining mean fire return intervals, and it is the average number of fires that occur per unit time at a given point. This hypothesis is based on suggestions by Parminter (1992) and Hawkes (1997) that there may be differences between the MoF published fire return intervals and actual fire return intervals. 2.3 STUDY AREA The study area for this project is found in south central British Columbia and incorporated 82% of the entire ESSFwc2 biogeoclimatic variant. It lay in the Monashee mountain range; approximate boundaries are the towns of Valemount to the north, Clearwater and Kamloops to the east, and Vernon to the south, and the Rocky Mountains to the east. Elevation for the study area ranged from 1200 m to 1800 m. For further details as well as a detailed map of the study area, refer to the description in Chapter 1 and Figure 1-1. 2.4 METHODS There were three major steps involved in this study: data collection, data integration and analysis using GIS, and output evaluation. Data Collection A map of British Columbia, along with a grid representing different mapsheets for the entire province, were provided to me by Jerry Maedel of the FIRMS lab at the University of British Columbia (UBC). Mapsheets that included portions of the ESSFwc2 variant were identified; in the end, forest cover data attribute tables for 168 mapsheets were downloaded using a VPN connection provided by the FIRMS lab at UBC. Each mapsheet represented an 11.2 km x 14.6 km area (Clover Point Cartographies Ltd 2003). Jerry Maedel in the FIRMS lab at UBC 24 provided me with files for spatial data. Attribute tables were then linked to spatial data using the vector-based GIS system, ArcView. Data integration and analysis Once tables had been linked, all mapsheets were brought into ArcView as separate polygon themes. These 168 themes were then merged together to create one coherent theme that represented the study area. This theme was exported to another vector based GIS, Arc/Info, in order to perform an overlay with the ESSFwc2 biogeoclimatic variant theme. Arc/Info was used for this operation because of its superior capacity to perform functions on large quantities of data. Once both themes were imported to Arc/Info, the shapefiles were converted to coverage files and a clip was performed to extract the areas of the mapsheets that fell within the ESSFwc2 variant. This new theme was then converted back into a shapefile and exported back into ArcView for further analysis. Queries were performed to extract ten age classes from the study area theme. These age classes follow those designated by the B.C. Min. For. (Table 2-1). Age classes are determined by stand age which, in turn, is determined either by examining aerial photographs or by aging trees using bored cores and making corrections to obtain total tree age. Each age class extraction was converted to a shapefile, which was then brought into the project as a separate theme. Harvesting has been occurring in the ESSFwc2 for approximately 40 years. I was interested in illustrating the differences between the pre-harvested and post-harvested landscapes. In order to obtain this information, two separate queries were performed\u00E2\u0080\u0094one for pre-harvest old growth and one for post-harvest old growth. As Goward (1994) points out, a clear definition of \"old growth\" has yet to be formulated and agreed upon by members of the ecological community. In general, old growth is usually defined with respect to ecosystem attributes such as structural diversity, canopy heterogeneity, numbers of standing dead trees, presence of large woody debris 25 (Franklin et al. 1987; Goward 1994). MacKinnon and Void (1998) have recommended minimum age criteria for both coastal and interior forest types in British Columbia. According to this designation, and for the purpose of this study, old growth was defined as forests older than 140 years. Two assumptions were made in order to calculate the area occupied by each age class prior to and after harvesting: (1) All harvested stands were old growth prior to being harvested. This assumption was checked for validity by ageing trees that had been cut in harvested plots sampled in the summer of 2002. Stumps in harvested areas were aged by counting rings. While this may not have been the most accurate approach to aging, due to time and funding constraints, it was determined to be adequate for gaining rough estimations (i.e. if the trees were over 140 years old). (2) The percentage of the total amount of old growth occupied by each of the two age classes greater than 140 years old (i.e. age classes 8 and 9) had the same distribution pre and post harvesting. Table 2-1. Age class categories according to B.C. Min. For. forest cover data sets. Age Class Age range (in years) of trees 0 0 1 1-20 2 21-40 3 41-60 4 61-80 5 81-100 6 101-120 7 121-140 8 141-250 9 251 + To obtain pre-harvest statistics, it was necessary to assume that all age classes could be rolled back by 40 years (i.e. age class 3 became age class 1, age class 2 became age class 0, age classes 1 and 0 became old growth, etc.). Because age classes 8 and 9 both have age ranges 26 larger than 40 years, a number of separate queries had to be performed in order to estimate the amount of area occupied by age classes 6, 7, 8, and 9 prior to harvesting (Table 2-2). Table 2-2. Queries for area per age class (pre-harvest). Age Class Query performed 6 7 8 9 140< stand age < 160 160< stand age < 180 180< stand age <290 stand age > 290 The percentage of total old growth occupied by each of the age classes 8 and 9 post-harvest was calculated. I then took these percentages, applied them to age classes 0 and 1 post-harvest to estimate the contribution of age classes 0 and 1 to the pre-harvest area of age classes 8 and 9. These areas were then added to the amount of area calculated by the above queries for age classes 8 and 9 to obtain the total area occupied by these two age classes prior to harvesting. To obtain the post-harvest old growth area, I queried for age classes greater than 7 (i.e. greater than 140 years). Distribution of pre and post harvested old growth was illustrated in the mapview by querying for stand ages less than 40 years or greater than 180 years (for pre-harvest\u00E2\u0080\u0094most of these stands would have been old growth 40 years ago) and by querying for stand ages greater than 140 years (for post-harvest). Both queries were converted to shapefiles and brought into the project as separate themes. Parameters calculated from the GIS pre/post harvesting database (1) Patch Size: The size and shape of patches determine the amount of interior habitat available to support wildlife dependent on it. If patch size falls below a certain minimum or has an especially high perimeter/area ratio, edge effects can permeate the patch, creating undesirable habitat for interior dwelling species (von Sacken 1998). Both average patch size and patch size range figure prominently in preserving habitat for large and small mammals, 27 amphibians, and many species of birds. Patches of different sizes and structure, as well as patches of varying sizes are required by wildlife for sustenance, protection from predators, and reproduction (Sierra Nevada Ecosystem Project Report 1996). In the present study average patch size and patch size range values for pre and post harvested landscapes were determined. (2) Patch perimeter/area ratios: Patch perimeter/area ratios indicate the amount of interior vs. edge habitat that is available for species requiring/utilizing one type of habitat or the other. Four perimeter/area ratio statistics were determined: average, minimum, maximum, and standard deviation. To obtain these statistics, it was necessary to export each attribute table as a dbf file to Microsoft Access in order to perform update queries for perimeter/area ratios. The tables were imported back into ArcView, and the ArcView calculator was used to calculate the average, minimum, maximum, and standard deviations. A challenging case arose in trying to calculate the perimeter/area ratios for age classes 8 and 9 for pre-harvest because age classes 0 and 1 had to be incorporated into the calculations. Because there is no way of knowing how patches were distributed prior to harvesting, I took all of the polygons in age classes 0 and 1 and assigned them randomly to age classes 8 and 9, using a series of IF statements in Microsoft Excel, until the appropriate percentage of total old growth makeup was achieved (i.e. 77% for age class 8 and 23% for age class 9). Calculations were then done on the resultant age classes 8 and 9. (3) Landscape Diversity: The Shannon-Wiener index was used to calculate landscape diversity; it was chosen over other indices because of its wide acceptance among scientific communities and because of its relative ease in use (Magurran 1988). The formula for the index is: H ' = - Spjlnpj where: 28 Pi = proportion of individuals found in the ith species (in this application, pj represented the proportion of landscape found in each age class); and i = the total number of age classes. (4) Mean fire return interval: Two frequently used models in fire frequency studies are the negative exponential and the Weibull model (Johnson and Gutsell 1994). Both of these models can be used to estimate one of two distributions: time-since-fire (survivorship) and fire interval (mortality). The survivorship distribution represents the probability of going without fire for longer than time t while the mortality distribution is the probability of a fire occurring in the time interval t to t + 1. Here, the survivorship distribution was estimated to determine fire frequency in the ESSFwc2. The survivorship distribution for the Weibull model is: A(t) = exp [-(t/b)c] where: A(r)= the cumulative proportion of landscape left unburned t - time; b = fire reoccurrence parameter (years); and c = shape parameter. Note that when f=0, A(0)=1, meaning that at the current time (time 0) 100% of the landscape is unburned. From there, A(f) decreases according to the percentage of landscape remaining at determined times. For this study, A(r) was determined at 20-year intervals. For age classes 8 and 9 that contained more than one 20 year interval, proportions of landscape lying in each 20 year interval were determined for post harvesting by querying for area by stand ages (e.g. stand age = 141-160) in the post harvesting database. In order to determine 29 proportions for pre-harvesting, post-harvesting proportion distributions were assumed and applied to the pre-harvested area that was greater than 140 years old. The Weibull model assumes that c>l, suggesting that the likelihood of a fire occurring is a function of forest age (Johnson and Gutsell 1994). The negative exponential model is a special case of the Weibull model whereby c=l, indicating no correlation between the likelihood of fire and forest age. Both models require that populations have stationarity, in this case meaning that all time periods have been subjected to the same disturbance regime (Johnson and Van Wagner 1985). In recent years, there has been some discussion over the validity of using the reverse cumulative distribution as illustrated above by the Weibull model. Huggard and Arsenault (1999, 2001) claim that the reverse cumulative distribution is incorrectly used to determine survivorship because it is not based on the actual (current) standing age distribution but on a log transformation of data. Huggard and Arsenault (2001) also argue that fire frequency analyses using standing age distributions can only be conducted if age class distributions are stationary and if the probability of fire (tree mortality causing disturbance) is independent of forest age. They consider that neither assumption is likely to be true for most forests. Although there is some evidence that pre-harvest era ESSFwc2 forests were stationary (e.g Antos and Parish 2001), current ESSFwc2 forests are not stationary and current disturbance is primarily in the oldest age classes (old-growth forest) so is very age dependent. Consequently, the significance of the fire, or disturbance, frequency analyses conducted using the age distributions is unclear. Little should be read into the numerical values of the numbers generated. These numbers are presented for anyone wishing to compare them to similar numbers published for other studies. Reed and Johnson (1999), however, argue that using the current standing age distribution is incorrect because current stand distribution potentially represents various past average rates of 30 burning. While there has been no final agreement on which distribution (cumulative/actual standing age or reverse cumulative) is best to use, the reverse cumulative was chosen in my study because of its long-standing history in fire frequency studies. Because I was curious to see if a large difference existed between plotting the actual standing age (as Huggard and Arsenault recommend) and using the reverse cumulative distribution (as this study used), I also analyzed fire frequency by plotting the actual standing age and came within 10 years of the estimates obtained by using the reverse cumulative distribution. It is possible that future discussions will bring about changes to the model used in my study; however, at this point in time there is no clear indication that this will or will not happen. Average fire frequency for pre and post-harvested landscapes was estimated by first plotting the data to determine which model provided the best fit. All data were plotted in the statistical program SYSTAT. For pre-harvested landscapes, it was determined that the Weibull model was appropriate to use. For the post-harvested landscapes, the negative exponential model provided a better fit for the data. Nonlinear regression analysis was used to determine the two parameters (b and c). It is important to note that fire frequency is not the same as natural disturbance frequency in ESSFwc2 forests. While fire frequency represents the amount of landscape left unburned by fire, natural disturbance frequency represents the amount of landscape left unaltered due to any natural disturbance. As previously discussed, both insects and fire play integral roles in the natural disturbance regimes of ESSFwc2 forests. Due to the nature of the current study, specifically that all data used were from digitized maps and not field sampling, the disturbance frequency, rather than fire frequency, was determined. Subsequent parts of the study were intended to discern insect from fire disturbances; however, time and funding constraints, in the end, made this task impossible. 31 Output Evaluation Following the output of my final pre and post harvesting maps (Figure 2-1), I reviewed the results to check for inconsistencies, accuracy, and overall cohesion that would not have surfaced until the final composite maps were created. In addition, age class data collected and analyzed were checked for accuracy in the field in the summer of 2002. Specifically, 3 trees per plot were aged by counting rings in cores to validate locations of each of the nine age classes given on the B.C. Min. For. forest cover maps. Parish (B.C. Min. For., Res. Br., 2002 pers. comm.) claimed that at least 100 trees would have to be sampled in order to accurately determine the age of a given patch. However, since funding and time were constraining factors in this study, and because the goal was not to determine age classes so much as validate a pre-determination, my method seemed most feasible. Finally, a check on the accuracy of forest cover maps was conducted by querying within each age class for stand age. The purpose of this was to determine if there were any incorrectly designated patches (e.g. stand age = 300 years but age class =0, which should only contain stands at age 0 years); this check was done because of a discrepancy I noticed, by chance, when scrolling through one of the age class attribute tables. 2.5 RESULTS The assumption that all harvested stands were old growth prior to harvesting was validated by aging three stumps in each of the harvested plots in the summer of 2002. Ages of stumps ranged from 150 to 300 years old, although this range is an approximation due to the limitations of aging in the field stumps that have been subjected to weathering. Discrepancies between the MoF forest cover database and the forest cover maps being used for field work were found. These discrepancies existed between age classes and stand ages; of 8,328 records in age class 0 (which should contain only stand ages at age 0 years), 225 incorrect entries of stand age being older than age 0 (specifically, age 300 years) were found. These incorrect entries totaled 32 20.3 km2, or 0.3% of the entire study area. This was not found to be a significant difference in terms of the percent of total landscape occupied by either age class 0 or age class 9. All subsequent analyses were done assuming that the age class was correct and not the stand age; this decision was purely arbitrary as either classification could have been correct. The total amount of old growth dropped from 4.80 x 10-^ km 2 prior to harvesting to 2.90 x 10^ km2 post harvesting (Figure 2-1). Most of this was lost in age class 8, where total area occupied dropped 14.6 x 102 km 2 (Table 2-3). Patch Size The average patch size in older forests (e.g. age classes 5-9) consistently declined after harvesting, while the average patch size in younger forests (e.g. age classes 0-4) consistently increased (Table 2-4). Variability, as described by range (maximum - minimum) in average patch sizes between pre and post-harvested landscapes is shown in Table 2-5. Pre-harvested landscapes showed a noticeable trend of variability increasing with age with the exception of age classes 0-1 and age classes 6-7. Post-harvested landscapes showed no consistent trends. Between the two landscapes, there was a trend of increasing patch size range in young age classes (i.e. 0-2) and a decreasing patch size range in older age classes (i.e. 3-9, with the exception of age class 7). Perimeter/Area ratios There was no discernible pattern in differences between pre and post-harvested perimeter/area ratios of patches either with age or between pre and post-harvested patches (Table 2-6). 33 Post-harvest old-growth Pre-harvest old-growth Scale: 1:1,500,000 N 5 \u00C2\u00A3 0 50 1QQ Ki lometers Figure 2-1. Pre vs. post-harvest old growth in the study area. Pre-harvest includes all post-harvest old growth and is represented, therefore, by both the yellow and green. 34 Table 2-3. Pre vs. post harvest area per age class in the ESSFwc2 forests in the study area. PPvE-HARVEST POST-HARVEST Age Class Area (km?) Pi(%) Area (km?) Pi(%) 0 2.27 X 102 3.5 1.71 x 103 26.4 1 1.70 X 102 2.6 8.00 x 102 12.4 2 2.18 X 102 3.4 2.27 x 102 3.5 3 1.53 X 102 2.4 1.70 x 102 2.6 4 1.41 X 102 2.2 2.18 x 102 3.4 5 1.44 X 102 2.2 1.53 x 102 2.4 6 3.57 X 102 5.5 1.41 x 102 2.2 7 2.49 X 102 3.9 1.44 x 102 2.2 8 37.0 X 102 57.2 22.4 x 102 34.6 9 11.0 X 102 17.0 6.60 x 102 10.2 Table 2-4. Average patch size (km2) and standard deviation (sd) for each age class of ESSFwc2 forests in the study area. PPvE-HARVEST POST-HARVEST Age Class Area (km?) sd Area (km?) sd 0 0.18 0.32 0.21 0.21 1 0.13 0.19 0.18 0.25 2 0.14 0.19 0.18 0.32 3 0.12 0.22 0.13 0.19 4 0.13 0.24 0.14 0.19 5 0.18 0.55 0.12 0.22 6 0.19 0.67 0.13 0.24 7 0.21 0.47 0.18 0.55 8 0.20 0.35 0.20 0.53 9 0.18 0.41 0.15 0.28 Table 2-5. Patch size range (km2) for each age class of ESSFwc2 forests in the study area. Age Class Pre-harvest Post-Harvest 0 5.1 126.8 1 1.6 4.3 2 2.0 5.1 3 3.8 1.6 4 4.3 2.0 5 12.5 3.8 6 16.5 4.3 7 6.5 12.5 8 45.9 16.5 9 124.3 3.8 35 Table 2-6. Perimeter/Area ratio per age class (km/km )^ for ESSFwc2 forests in the study area. PRE-HARVEST POST-HARVEST Age Class Avg. Min. Max. sd Avg. Min. Max. sd 0 0.05 0.004 4.56 0.22 0.07 0.001 4.26 0.22 1 0.06 0.004 3.13 0.18 0.03 0.004 3.82 0.13 2 0.05 0.004 3.98 0.18 0.05 0.004 4.56 0.22 3 0.08 0.004 3.98 0.28 0.06 0.004 3.13 0.18 4 0.05 0.005 3.33 0.17 0.05 0.004 3.98 0.18 5 0.05 0.004 2.66 0.17 0.08 0.004 3.98 0.28 6 0.05 0.003 4.40 0.17 0.05 0.005 3.33 0.17 7 0.04 0.004 4.28 0.19 0.05 0.004 2.66 0.17 8 0.05 0.002 4.67 0.18 0.05 0.003 4.67 0.18 9 0.06 0.001 4.22 0.20 0.06 0.004 3.94 0.19 Landscape Diversity Applying the Shannon-Wiener index formula (see methods) to the percentages in Table 2-3,1 calculated a pre-harvesting value of 1.49 while post-harvesting was 1.80, indicating an increase in landscape diversity after harvesting. Disturbance Frequency Table 2-7 illustrates the disturbance frequency obtained by applying the Weibull model (in the case of pre-harvested landscape) and the negative exponential model (in the case of post-harvested landscape) to the data. The \"b\" parameter represents the mean fire (disturbance) return interval and the \"c\" parameter, as previously described, indicates the likelihood that fire (disturbance) occurrence is a function of forest age. The analyses indicate disturbance frequencies of 207 years for pre-harvested landscapes and 138 years for post-harvested landscapes as can be seen in Figures 2-2 and 2-3. 36 Table 2-7. Fire frequency estimations. Landscape Parameter Estimate Lower CI 95% Upper CI 95% Pre-harvested b 207.3 200.2 214.5 c 3.0 2.6 3.5 Post-harvested b 138.4 117.5 159.2 c 1.0 0.8 1.3 0 100 200 300 400 500 TIME Figure 2-2. Pre-harvest fire frequency. 0 100 200 300 400 500 TIME Figure 2-3. Post-harvest fire frequency. 2.6 DISCUSSION Discrepancies between the MoF forest cover database and the forest cover maps being used for field work were found. For the purposes of this study, it was assumed that the age class classification (not the stand age) was correct. However, depending on which entry was incorrect (e.g. age class or stand age), and depending on whether or not other entries were incorrect, the outcome of all analyses could have been different, leaving the possibility of significant differences between pre and post-harvested landscapes. This source of error illustrates a major constraint of all GIS systems: the information extracted from them is only as good as the data on which the extraction is based. Data entry cannot be done carelessly and should be checked for errors. All of the errors encountered seemed 37 easily identifiable (e.g. age class 0 but stand age 300), yet easy to fix before the data were made available to the public by simply checking the data entry for accuracy. While the area that was mismarked accounted for less than 1 % of the total area, inaccurate data (regardless of the size of its overall impact) raises suspicions as to the validity of any/all data in the database. Old growth forest provides unique and necessary habitat for many wildlife species, including the threatened woodland caribou which feeds on lichens found in old growth forest. Environment Canada (2002) cites habitat destruction as a major cause of the threatened status of the woodland caribou. The loss of old growth determined in this study signifies danger for woodland caribou and other species dependent on old growth habitat. It should be noted that not all of age class 2 would necessarily have been age class 0 prior to harvesting because age class 0 represents a single age (i.e. age 0) and age class 2 represents a 20-year range of ages. While this may partially have been compensated for by the fact that it can take several years for trees to occupy a site after disturbance, it is possible that the area assigned to age class 0 pre-harvest was overestimated. Similarly, the amount of pre-harvest old growth may have been underestimated due to the possibility of some of age class 2 post-harvest being old growth instead of solely age class 0. Table 2-3 illustrates, however, that the total area potentially distributed erroneously was fairly negligible (maximum 3-4% of the total area, or the total area in age class 2 post-harvest). Hypothesis 1: Harvesting creates a more uniform landscape with less variability between patch sizes and perimeter/area ratios. Patch size is significant because it is representative of the landscape created by disturbance patterns (either natural or human induced). It, along with perimeter/area ratios, also suggests the amount of interior vs. edge habitat available to animal and plant species requiring and/or using each type of habitat. 38 Patch size and variability can affect the availability of wildlife habitat, travel corridors and connectivity. The largely homogenous, fragmented landscape created by clearcutting results in large areas being exposed, potentially decreasing the amount of travel corridors and, consequently, the possibility for wildlife to travel from one patch to the next. Clearcutting potentially diminishes specific patch sizes that may be required by various species. The threat of windthrow also seems to be affected by patch size; clearcuts can create problems by increasing wind speed and turbulence (Stathers et al. 1994). Jull and Sagar (2001) found that wind damage was more severe within the first 20 m of the harvest boundary of clearcut edges than in any other harvesting treatment (group retention, irregular shelterwood, or single-tree selection), indicating that wildlife populations, habitat, and potential timber could be adversely affected by clearcutting through an increase in the amount of trees lost due to windthrow. Lastly, patch size and variability can have implications for regeneration at the stand level, which need to be considered in conjunction with habitat and landscape fragmentation. Eastham (1999) found that the availability of seed sources, essential for natural regeneration, was removed when patch sizes were too large (i.e. sources were only within 100 m). Alternatively, patch sizes that are not large enough, while providing adequate seed source, may be detrimental to seedling germination success and subsequent growth, due to shading that directly (via a decrease in sunlight) and indirectly (via a decrease in soil temperature) affects seedling germination and survival (Feller 1998; Eastham 1999). Younger forests occurred in larger sized patches and older forests occurred in smaller patches after harvesting. Generally, with the exception of age class 7 and age classes 0-2, patch size variability decreased with harvesting. While the B.C. Min. For. (2002) acknowledges that patches provide structural heterogeneity, species diversity in vegetation, and better habitat for wildlife than scattered individual trees, the analysis of patch size and variability in my project suggests that current harvesting practices are not succeeding in emulating natural disturbance 39 patch size or variability. The loss of patch size and patch size variability in older age classes could have impacts on wildlife dependent on older forests, although at this time the literature is lacking in studies that have looked at species dependent solely on ESSF forests, thus limiting the conclusions that can be drawn regarding patch size, variability and species dependence. The data did not support the hypothesis that there was a significant difference between perimeter/area ratios on pre and post-harvested landscapes. This suggested that species dependent on interior forest habitat are not being negatively affected by harvesting; similarly, opportunistic species that thrive on edge habitat are not being favored by harvesting practices. While the maximum cutblock size range of 40-60 ha (Anon 1995) implemented by the FPC (and subsequently the FRPA) falls within the estimated 50-500 ha range of mean fire size in the ESSF (Parminter 1992; Hawkes et al. 1997), the range of patch sizes created by harvesting falls severely short of the estimated 0.1 to 10000 ha range (Parminter 1992) found in naturally disturbed landscapes. Small, equal-sized patch distribution is creating an unnatural landscape which may potentially have negative impacts on some wildlife species. This may be remedied by varying cutblock size, although this may have further implications for forests at the stand level, as will be discussed in Chapter 3. Hypothesis 2: Landscape diversity increases with harvesting because old growth is distributed throughout younger age classes. Landscape diversity was assessed using the Shannon-Wiener index and did, in fact, increase with harvesting. While at first glance this may appear to support logging as a way to increase diversity, it is important to look closely at what this rise in diversity is showing. At this time, it is unclear what this increase is indicating. Magurran (1988) notes differences between species richness and evenness and concedes that a more even species (here, landscape) distribution results in greater diversity. A more even landscape distribution, however, may not be \"natural\" and can increase the chance of catastrophic disturbance (Perry 1994) which, in turn, 40 could result in necessary habitat for plant and animal species being lost. Traits of young and old forests (e.g. sources of nutrients and nutrient cycling, habitat for wildlife, etc..) differ greatly. Diversity can be measured on three levels: genetic, species, and ecological, or landscape (Miller 1998). It would be incorrect to conclude that an increase in one type of diversity (here, landscape) indicates an increase in the other two. Because old growth offers unique habitats and is home to unique species of lichen, mosses, plants, and animals, its loss to younger forest results in less diversity among these species and can potentially result in a decrease of genetic diversity, due to the fact that only those species least resistant to change in habitat will survive. Lastly, it should be noted that continuation of old growth harvesting ultimately may lead to a drop in landscape diversity. Hypothesis 3: The mean fire return interval is longer than the 250-350 years currently believed to be characteristic of ESSF wc2 forests. The data, when plotted, showed a mean disturbance return interval as being less than 250 years. This number may not be very meaningful due to uncertainties about the method used to generate it. Also, the estimated intervals signified simply when these forests regenerated. It was not the mean fire return interval, but, rather, the mean natural disturbance interval. While fire is the major disturbance factor in the structure of ESSF forests, some studies suggest that inter-fire small disturbances, specifically insect infestation, have equally important effects on stand dynamics (Alexander 1987; Parish et al. 1999; Lewis and Lindgren 2000). This information suggests that stands could be regenerated following tree mortality from fire, insects, disease, wind, or snow, making the 207 year (pre-harvest) and 138 year (post-harvest) frequencies minimum fire return intervals. Separating disturbance factors (e.g. fire and insects) is a challenging problem that was found to be beyond the scope of this project. While charcoal was collected from old growth 41 stands in the summer of 2002 with the intent of determining, more accurately, fire vs. insect caused disturbances, the samples were not analyzed due to time and funding constraints. In looking at the validity of the assumptions used in using age distributions for disturbance frequency analysis as well as the poor fit of the negative exponential model to the data in Figure 2-3 and the uneven distribution of area among age classes (Table 2-3), caution should be used when interpreting the disturbance interval for pre and post-harvested landscapes. It appears that disturbance depended on forest age for pre- and post-harvested landscapes and that the post-harvested landscape was not stationary. While the exact disturbance interval may be called into question, it is clear that there has been a notable change in the standing age distributions and a change in the disturbance interval between pre and post-harvested landscapes, which suggests that harvesting is not currently mimicking natural disturbance regimes. 2.7 CONCLUSIONS GIS provided a relatively easy way to illustrate differences in pre and post harvested landscapes. Together with Microsoft Access, Microsoft Excel, and Arc/Info, ArcView allowed numerous operations and calculations to be performed in order to analyze various statistical parameters within the Ministry of Forests forest cover database. Significant differences between pre and post-harvested landscapes, including mean disturbance interval, average patch size, patch size variability, and total amount of old growth occurring in ESSFwc2 forests, suggest that changes need to be made to current harvesting practices (i.e. clearcutting) if emulation of natural disturbance patterns and preservation of biodiversity is the management objective. 42 CHAPTER 3 STRUCTURAL CHARACTERISTICS OF STANDS REGENERATED AFTER NATURAL DISTURBANCE AND THOSE REGENERATED AFTER HARVESTING 3.1 INTRODUCTION An understanding of stand structure and stand dynamics is essential to effective forest management (Oliver and Larson 1996). Knowledge of stand structural characteristics complements an understanding of landscape level forest dynamics; this dual-scale approach to understanding forest development and processes provides a comprehensive knowledge base which forest managers as well as policy makers can draw from when making decisions about appropriate activities in forests. In the past two decades, the importance of disturbance in stand dynamics has been recognized. While there have been few studies, to date, conducted at the landscape level in ESSF forests, there have been more studies conducted at the stand level, both in British Columbia and in the United States Pacific Northwest (Aplet et al. 1988; Parish et al. 1999; Antos and Parish 2001; Varga and Klinka 2001), although because most of these studies (Antos and Parish 2001 being an acception) were conducted in ESSF forests that were generally warmer and drier, extrapolation to the wet, cold ESSF forests of B.C. may be problematic. In addition, three silvicultural system projects were initiated in ESSF forests in central and south-central British Columbia in the 1990's: (1) the Lucille Mountain Project in the Prince George Forest Region (Jull and Stevenson 2001), (2) the Grain Creek-Blackbear Creek Project in the Cariboo Forest Region (B.C. Min. For. 1992), and (3) the Sicamous Silvicultural Systems Project in the Kamloops Forest Region (Hollstedt and Vyse 1997). These projects considered various components of stand development including: abundance of trees and seedlings, horizontal 43 (spatial) distribution, vertical distribution, regeneration and growth patterns of trees, spruce-fir interactions in various successional stages, and the retention/importance of standing dead trees (snags) and coarse woody debris (cwd) in stand dynamics over time. For this project, I examined all of these components except for horizontal distribution; data were collected but not analyzed due to funding and time constraints. Kimmins (1997) as well as Perry (1994) offer explanations for the importance of each of the above mentioned structural characteristics. The abundance of trees and seedlings within a stand offers insight into site quality (e.g. soil productivity) as well as vegetation competition for sunlight, nutrients, and water. A lesser number of trees, other factors remaining the same, presumably indicates fewer negative impacts on tree growth due to competition. The abundance of trees, additionally, has implications for timber harvesting. Striking a balance between overstocked stands that result in increased competition and, therefore, slower growth, and stands that have too few trees (from a timber-value point of view), due to any number of factors (e.g. lack of seedbank, inadequate soil temperature, lack of light and/or nutrients) is one of many challenges facing managers of ESSF forests today. The abundance of trees can affect regeneration by providing adequate seedbanks (Eastham 1999) and, alternatively, can have negative effects on regeneration by shading seedlings that are relatively shade intolerant (especially Engelmann spruce). The abundance of seedlings, in turn, reflects seedbed quality, adequate seedbank, and availability of light, nutrients, and water (Feller 1998; Eastham 1999). Tree diameter in ESSF forests sometimes follows a reverse-J distribution (although this is not always the case), as described and supported by Varga and Klinka (2001). There are more stems/ha in smaller diameter classes; the number of stems/ha decreases as trees get larger. Tree diameter can affect the amount of light available for regeneration, can signify site quality, and has obvious implications for harvesting (i.e. larger trees are economically desirable). Antos and Parish (2001) as well as Varga and Klinka (2001) concluded that dbh had no more than a weak 44 relationship with age, suggesting that both Engelmann spruce and subalpine fir can survive long periods of suppression before release is made possible through a break up of the canopy. Horizontal and vertical distribution are defining characteristics of stand structure (Kimmins 1999). Antos and Parish (2001) contended that spatial distribution offered critical insight into small scale processes affecting stand structure and succession. Varga and Klinka (2001) concluded that old-growth stands in the ESSF zone were two storied and uneven/multi-aged. Higher diversity in vertical structure affects forest succession by offering a constant supply of trees during canopy break ups. In addition, vertical diversity supplies habitat for wildlife species at different levels, theoretically reducing competition for resources by providing spatial separation (Kimmins 1999) between competing species. The interaction between spruce and fir trees and seedlings in ESSF forest stands has been one of the major focal points of studies conducted to date; there has been, however, a lack of studies conducted in wet, cold ESSF forests, making extrapolation from most of the studies cited here potentially problematic. Aplet et al. (1988) characterized ESSF stand succession as consisting of: (1) stand development, (2) colonization phase, (3) spruce exclusion phase, (4) spruce reinitiation phase, and (5) second generation spruce-fir forest. Of particular importance in this study, as in subsequent studies (Antos and Parish, 2001; Varga and Klinka, 2001) is the replacement of the generic \"stem-exclusion\" phase (Perry 1994; Kimmins 1999) with a solely spruce exclusion phase. It has been suggested (Knapp and Smith 1982; Aplet et al. 1988; Varga and Klinka, 2001) that, in the absence of disturbance, fir will out compete spruce, becoming the climax species in a steady state forest (although the existence of a \"steady state\" ESSF forest is still highly debated). Antos and Parish (2001), however, found that spruce existed in a larger proportion of the seedbank than of large trees, indicating that spruce would most likely represent a substantial proportion of future stands. Current management of ESSF forests relies on planting spruce because of its economic superiority (Jull and Stevenson 2001); this contrasts with natural 45 regeneration patterns that favor fir over spruce (Aplet et al. 1988; Parish et al. 1999; Antos and Parish 2001). The presence and ecological importance of snags has been recognized in recent years (Voller and Harrison 1998; Kimmins 1999). Snags contribute to community structure by providing a food source, in the form of insects, for birds, providing nesting and shelter sites, and providing perches for large birds. Snags can aid in seedling establishment by providing protection from direct sunlight that can kill seedlings. In a functional role, snags provide sites with low level nonsymbiotic nitrogen fixation and contribute to long-term supplies of organic matter input to the forest ecosystem (Kimmins 1999). Huggard (1997) pointed out that snags have important economic implications for forest management. He found that fall down rates for certain decay classes of snags (specifically, newly dead snags) were probably similar to the fall down rates for live trees, indicating that snag felling of these types of snags may be both an economical and ecological waste. Current management guidelines for snags in British Columbia allow for some snag retention through wildlife patches (Stone et al. 2002). The importance of coarse woody debris has been increasingly acknowledged over the past decade, although direct management guidelines for cwd are still largely lacking in British Columbia (Stone et al. 2002). The decay of coarse woody debris releases nutrients into the soil (Perry 1994; Feller 1997), although overall nutrient supply decreases as wood decays (Kayahara 2000). In addition, coarse woody debris provides food for insects and microorganisms that, in turn, serve essential purposes in forest ecosystem processes (e.g. nutrient cycling, soil formation), although this is an area that needs more studying (Feller 2003). Coarse woody debris serves as sediment storage (Voller and Harrison 1998), protecting mineral soil from extensive erosion during disturbance; coarse woody debris also offers a long-term, steady supply of organic matter for the forest ecosystem. Prescott (1999) acknowledged the importance of organic matter input to ESSF forest ecosystems. Her study specifically found that decomposition rates 46 did not vary between clearcut and unharvested sites, indicating that, contrary to popular belief, clearcutting did not promote faster decomposition rates (by increasing soil temperatures). Kayahara (2000) concluded that in wet coastal forests of British Columbia, coarse woody debris should be maintained in ecosystems if the goal is to preserve biodiversity but that coarse woody debris did not increase overall nutrient supply and, therefore, did not positively benefit long-term productivity on sites. The traditional view that coarse woody debris was \"wasteful\" and \"messy\" (Voller and Harrison 1998; Kimmins 1999) has, in recent years, seen a shift among researchers to the ecologically-based view that coarse woody debris serves important functional and structural purposes in forest ecosystems. In a review of cwd studies conducted in B.C. and elsewhere, Feller (2003) concluded that while attributes of cwd can be well described for the ESSF zone (as in several other zones), more research is needed to understand the functional role (e.g. influence on slope stabilization, vegetation, animals, nutrients, and soils) of cwd in old growth forests of British Columbia. He also contended that future studies should consider both scale and temporal issues, as studies to date have not. Overall, each of these stand components contributes to the complex picture of the entire forest ecosystem. Each characteristic is necessarily, therefore, related to others, and it is necessary to view attributes of each in relation to others when drawing conclusions. Stand level attributes were seen as complementary to the landscape study presented in Chapter 2, as they serve as indicators of forest ecosystem functioning and health, including the ability of the forest to support wildlife (Hunter 1999) and maintain ecological processes such as infiltration, nutrient cycling, and climate regulation (Perry 1994). If the goal of forest management is to mimic the patterns of natural disturbance, knowledge of both landscape and stand level effects of natural disturbance is essential. 47 3.2 OBJECTIVES This project had two broad objectives: (1) to describe the structural characteristics of ESSFwc2 forests over a range of successional stages/ages (i.e. a chronosequence) and (2) to compare structural characteristics of stands regenerated after harvesting with those regenerated after natural disturbances. Specifically, the structural characteristics analyzed in this project were: (1) Abundance of living trees (2) Abundance of regeneration (seedlings) (3) Abundance of snags (4) Species composition of seedlings and live trees (5) Crown class composition (6) Abundance of coarse woody debris By measuring and statistically analyzing these attributes and adding this knowledge to the landscape level study, it was hoped that a comprehensive description of any differences between harvested and non-harvested ESSFwc2 forests would emerge in order to guide future management of these forests. 3.3 STUDY AREA The study area was the ESSFwc2 biogeoclimatic variant which occurs in the Monashee Mountains of south central British Columbia, between the towns of Valemount (north), Clearwater/Kamloops (west), and Vernon (south). My study area incorporated approximately 6.46 x 103 km2, or 82%, of the total ESSFwc2 forested landscape. For the purpose of collecting stand level data, my study area was divided into a northern and southern geographic region with the goal of illustrating the existence, if any, of geographical differences between stand structural characteristics. All plots lay between 1400 and 1700 m elevation. The location of these plots is 48 geographically detailed in Table 3-1 and graphically illustrated in Figure 3-1. The study area is described in further detail in Chapter 2. Table 3-1. Plot locations within ESSFwc2 forests. Area Name Location Age Classes Sampled Raft River (N) Shannon Creek (N) Fowler Lake (N) Vavenby (N) Mabel Lake (S) Sicamous (S) Ashton Creek (S) 51\u00C2\u00B051'00\"N and 119\u00C2\u00B043'00\"W 51\u00C2\u00B045'24\"Nand 119\u00C2\u00B012'31\"W 51\u00C2\u00B047'36\"N and 119\u00C2\u00B013'28\"W 51\u00C2\u00B035'51\"Nand 119\u00C2\u00B033'07\"W 50\u00C2\u00B024'00\"N and 118\u00C2\u00B036'00\"W 50\u00C2\u00B049'00\"N and 118\u00C2\u00B048'00\"W 50\u00C2\u00B036'00\"N and 118\u00C2\u00B058'00\"W 1 & 2 harvest; 4, 5, 6, 8 ND a 1, 9 ND 2ND 3 ND 4,7, 8 ND 1 & 2 harvest; 1,3,5,8,9 ND 4,6 ND aND= naturally disturbed 3.4 METHODS GIS Methods (pre-field) Originally, it was my intention to identify and sample three large study areas, each containing all age classes (Chapter 2, Table 2-1), that covered the range of the ESSFwc2 from north to south in order to provide insight into possible geographical differences within stand structure. Once forest cover data were analyzed in Arcview GIS, however, it became clear that the age classes were not distributed evenly enough to sample three large areas. I, therefore, divided my study area into two geographic areas-one comprised of the northern half of the ESSFwc2 and one comprised of the southern half of the ESSFwc2. 49 Figure 3-1. Location of sampled plots within the ESSFwc2 study area. 50 Identification of potential sampling sites was done by analyzing forest cover data in Arcview GIS (Chapter 2). Once age classes were identified, forest cover maps were obtained from the British Columbia Ministry of Forests regional office in Kamloops and from the district offices in Clearwater (northern sites) and Vernon (southern sites). From these maps, areas for sampling were chosen randomly after the following criteria were met: (1) The sampling area fell within approximately 1 km (-15 min. walk) of the last point able to be reached by driving. This was done for logistical ease in order to keep within time and budget constraints. (2) If sample areas lay alongside a road, plots were chosen randomly after ensuring that there was at least a 100 m buffer between the road and the plot. This was done to eliminate edge effects from the roads that may create errors in diversity estimations. All harvested sites were planted; this was not a purposeful choice on my part, but, rather, reflected the current practice in ESSF forests. As Farnden (1994) asserts, artificial regeneration is recommended over natural regeneration due to low germination rates on undisturbed soils and poor survival (particularly on exposed microsites) that have plagued the success of natural regeneration efforts (Alexander 1969; Alexander et al. 1984; Butt 1990). Simply put, most, if not all, harvested stands in ESSFwc2 forests appeared to have been artificially regenerated to maximize successful tree establishment. Field Methods All mapped stands chosen for sampling were checked for accuracy in the field by aging trees with an increment borer before sampling took place. If the ages of several dominant trees did not correspond to the age class designated on the forest cover map, I either abandoned the stand (if we had the age class sampled already) or I sampled the stand assuming that my age class was correct. The latter occurred only twice. The first instance occurred in age class 7, 51 whereby I was not sure where we were in relation to the forest cover map, and, therefore, had no age class on the map with which to check my aging. The B.C. Min. For. designates age classes by interpreting aerial photographs and corroborating their estimates with ground sampling of at least 100 trees (Bob Macdonald, B.C. Min. For., Growth and Yield Forester, pers. comm.). Due to the practical limitations in aging that many trees, I aged 10 dominant trees and used this as my indication that I was, in fact, in age class 7. The second instance took place in Sicamous, where the forest cover map claimed there was a stand of age class 3; my sampling (again, using 10 trees) indicated the stand was, in fact, age class 5. Whenever possible all plots within an age class were placed in the same patch to minimize travel between sample sites and to obtain samples from age classes that did not have many accessible patches. In actuality, this ended up being the case in about half of the age classes (1 ND, 2 ND, 1 harvest, 2 harvest, 3, and 7); in the rest of the age classes, two or more patches were sampled. For both the northern and southern halves of my study area, I sampled three plots in each of nine age classes resulting from natural disturbance and three plots in each of two age classes resulting from forest harvesting The nine natural disturbance age classes were sampled in order to provide an illustration of how ESSFwc2 forest structure changes over time with a natural disturbance regime. In addition, the two harvesting age classes were sampled and compared to the two corresponding natural disturbance age classes (i.e. age classes 1 and 2) to determine what structural differences, if any, existed between the two types of disturbance. There were several instances when an insufficient number of stands of certain age classes were available for study. This was due to lack of road access and/or a location beyond the designated walking distance (see plot sampling criteria above); this resulted in several age classes having less than six plots. The total number of plots that were established and sampled is illustrated in Table 3-2. 52 Table 3-2. Total number of plots established. Age Class North Plots South Plots TOTAL 1 harvesting 3 3 6 2 harvesting 3 3 6 1 ND a 3 3 6 2ND 3 0 3 3 ND 2 1 3 4 ND 3 2 5 5 ND 3 3 6 6 ND 3 3 6 7 ND 0 3 3 8 ND 3 3 6 9 ND 3 3 6 TOTAL 29 27 56 aND= natural disturbance All plots were situated mid-slope and slopes ranged from 3-62\u00C2\u00B0; aspects varied. Site series were determined both by vegetation identification and the utilization of diagnostic keys (Lloyd et al. 1990) . Most plots were classified as ESSFwc2/Bl- Azalea - Oak fern or ESSFwc2/01, which represented zonal sites. Dominant tree species were Engelmann spruce and subalpine fir; shrub species included white-flower rhododendron and false azalea; and the herb layer was predominantly oak fern and Sitka valerian. When at all possible, care was taken to avoid steep slopes, harsh aspects, and extremely wet or dry site series, as all can affect stand characteristics (Meidinger and Pojar 1991) and, therefore, diversity estimations (i.e. they represent extreme, rather than zonal, sites). Appendix I gives the site series, slope, and aspect for all 56 plots. Random numbers were alternately chosen by one field assistant to determine the angle at which we entered the stand, the number of steps we took to get to a plot starting point (after the criteria listed above were met), and the direction of the first side of the plot. Plots were laid out using measuring tapes and compasses; all plots were square with 20 m sides (400m2). To obtain accurate readings for tree location, two additional tapes were run 53 through the middle of each plot so that it was divided into four-5 m x 5 m sampling quadrants. Two teams consisting of three people each carried out plot measurements; plots were sampled two at a time, with one team/plot. Measurements made within each plot were: (1) Tree location and species. Initially, I had planned on analyzing spatial distribution of live trees and regeneration. (2) Tree diameter. Dbh was measured using calipers. (3) Crown class. For the purposes of this study, dominance was divided into 5 classes: (1) dominant, (2) co-dominant, (3) intermediate, (4) suppressed, and (5) veteran. Crown class was determined by visual estimations which were made by the same team of people for all plots within an age class to ensure consistency in estimation. Originally it was my intention to statistically analyze differences in crown class composition between naturally disturbed and harvested stands. However, upon reflection of the methods used to classify the 5 crown classes, I decided they were insufficient to draw confident conclusions and, thus, the statistical analysis was dropped from my project. While height class was determined by the same team for each age class, visual estimations could have varied from day to day and from team member to team member. Furthermore, there was a change in the height class designation from the beginning of the sampling period to the end. In the beginning I thought it would be easiest, since age classes 1 and 2 are young age classes, to classify trees as \"small, medium, and large.\" Later on, I realized that the classification should be done as it was in all other age classes, at which point the classification was changed to fit the \"dominant, co-dominant, etc...\" criteria. When entering this data into SPSS, I had to decide which classification (e.g. dominant, co-dominant) the \"small, medium, large\" classes fell into, which became somewhat arbitrary and could very well have affected height class diversity estimations. 54 (4) Seedling location and species. Seedlings were considered to be any tree shorter than breast height\u00E2\u0080\u00941.3 metres\u00E2\u0080\u0094which is the suggested height at which to measure diameter according to the B.C. Ministry of Sustainable Resource Management (2002). (5) Snag location and species, if discernible. Species identification was affected by the extent to which the snag had been weathered. Snags had to be greater than 1.3 m in height. (6) Tree age/fire history. In harvested plots, stumps, if present, were aged by counting tree rings to test the assumption (Chapter 2) that all harvested stands were old-growth when cut. This method, as opposed to more accurate dendrochronological methods such as cross dating, was felt to be accurate enough since the end goal was not to determine an exact age but, rather, to determine if the trees were over 140 or 250 years old (age classes 8 and 9, old-growth, as defined in Chapter 2). Charcoal samples, if present within moderate digging distance (approximately 5 cm) of the soil surface, were taken in each plot with the intention of carbon dating the charcoal to determine the date of the last fire disturbance. (7) The abundance of coarse woody debris was visually estimated using photo guides from the U.S. Forest Service (Fischer 1981). Visual estimations were made instead of conducting more precise measurements using line transects because of the time required to carry out these measurements. This was justified by the fact that I was only interested in showing general trends between two types of stands (i.e. naturally disturbed and harvested) and, thus, precise measurements were not considered necessary. In an attempt to minimize error in estimations, the amount of coarse woody debris was determined by taking an average of estimates made (using the photo guides) by three field assistants. 55 (8) Plot characteristics. In addition to collecting data on each tree and seedling, measurements of slope and aspect were taken. Laboratory Methods Upon completion of field work, all data were entered into the statistical program SPSS 11.01 in order to conduct statistical analysis. This analysis was divided into two distinct categories: (1) descriptive statistics derived for the chronosequence of the natural disturbance forest, and (2) statistical analysis for the comparison of natural disturbance to harvesting stands. Natural disturbance Descriptive statistics (mean and standard deviation) were obtained, using the statistical program SPSS, and illustrated, using Microsoft Excel graphing, for each of the following parameters: (1) Abundance of live trees (2) Abundance of snags (3) Abundance of seedlings (4) Dbh (5) Species composition (6) Crown class composition (7) Coarse woody debris Statistical Analysis: natural disturbance vs. harvest origin stands Inferential statistical analysis was performed on the abundance of live trees, snags, and seedlings; mean dbh of live trees; species composition of live trees and seedlings, and amount of cwd (kg/m2). Species composition tests were conducted both for the percent of all trees each species (e.g. spruce, fir) comprised and for the exact number of individual trees of each species. 56 These two different analyses were performed due to the potential difference in outcomes when testing strict counts as compared to the proportion occupied by a species. Because there were no naturally disturbed plots sampled in the south for age class 2, the number of samples differed between variables depending on whether a treatment by location difference was detected by analysis of variance of age class 1 data. If there was a significant treatment by location interaction in age class 1, only the northern plots for harvesting origin stands in age class 2 were used (i.e. a total of three sample points for naturally disturbed and harvesting origin stands); if there was no significant interaction in age class 1, all six harvesting origin stands in age class 2 were used (i.e. a total of three sample points for naturally disturbed and six sample points for harvesting origin stands). Initially, plots within the same age class and treatment (e.g. age class 1 harvesting) were treated as one sample; in other words, northern plots were not separated from southern plots. The mean (for whichever variable was being determined, e.g. number of trees, dbh, species composition) for each plot was determined and these means (1 for each plot) were used as the data points for subsequent parametric tests, giving six data points for each age class and treatment, except for the case of age class 2 natural disturbance for which only 3 plots were sampled, yielding only three data points. For each variable, data were first tested for normality, using the Shapiro-Wilk test, in order to determine if parametric tests were appropriate. If the population was normal, a two-way analysis of variance (ANOVA) test was done to determine if there were statistically significant differences between naturally disturbed and harvested sites and/or between north and south geographic regions. If data were non-normal, transformation (e.g. log, arcsine) was attempted in order to achieve normality. If a transformation worked, it was used; if it did not work, an appropriate non-parametric test was used to test for differences between naturally disturbed and harvested 57 sites. The statistics and the tests used to assess them are listed in Table 3-3. Data were represented graphically to illustrate all statistical test outcomes. All graphs were generated using Microsoft Excel, due to the fact that Excel was easier to use than SPSS 11.01 for graphing. Table 3-3. Measured variables and the tests used to compare them for natural disturbance vs. harvesting origin ESSFwc2 forest stands. Variable Test Mean number of live trees/ha Two-way ANOVA Mean dbh of live trees Two-way ANOVA Species composition of live trees (count)a Two-way ANOVA Species composition of live trees (%)^ Two-way ANOVA (arcsine transformation) Mean number of snags/ha Kruskal-Wallis Mean number of seedlings/ha Two-way ANOVA Species composition of seedlings (count) Two-way ANOVA Species composition of seedlings (%) Two-way ANOVA (arcsine transformation) Mean coarse woody debris (kg/m2) Two-way ANOVA \"count\" refers to the mean number of each species present within a plot. b \"%\" refers to the percent occupied by each species or crown class within a plot. Lastly, upon completing ANOVA's for each variable, I determined an ideal sample size for each age class and treatment combination for all variables where a significant difference was not found with the intention of providing one possible explanation as to why significant difference was not found (i.e. sample size was too small). To determine an ideal sample size, I first calculated a quantity referred to by Zar (1995) as lowercase Greek phi. The equation used to determine this was: 4> = V(n52)/2ks2 where: n= number of samples within each factor 8= detectable difference (difference between and p,2as observed in my samples) k= number of factors (or groups; here, there were two groups\u00E2\u0080\u0094fire origin and harvest origin) s2= MS error from ANOVA table 58 Once (j) was determined, I iteratively determined appropriate sample sizes by using different values for my sample size (n) until an 80% power with a=0.05 was obtained (Zar 1995). Thus, I was able to determine appropriate sample sizes to estimate the true mean with 80% probability, testing at the 0.05 significance level. 3.5 RESULTS Natural disturbance Appendix II gives tables containing per-plot measurements for each variable. Error bars on each of the following figures represent + one standard error, and sample sizes are given above each error bar. Abundance of live trees The mean number of trees increased from age class 1 to age class 2 and then decreased over time, with the exception of an increase between age classes 7 and 8 (Figure 3-2). Mean dbh of live trees The mean dbh of live trees showed a trend of increasing until age class 7, at which time there was a decrease between age classes 7 and 8 and an increase between age classes 8 and 9 (Figure 3-3). Species composition of live trees The proportion of the total number of trees occupied by each species within an age class is illustrated in Figure 3-4. Fir represented over 50% of the total number of trees in each age class. Because of their extremely low presence, western redcedar (Thuja plicata), trembling aspen (Populus tremuloides), and western white pine (Pinus monticola) were omitted from analysis. 59 6000 5500 5000 4500 S 4000 co \" 4500 S 4000 -I B 3500 3 z 3000 2500 2000 1500 1000 500 0 6 6 3 \ 6 5 / \ 3 / 6 / 3 / 6 / Age Class. Figure 3-7. Number of seedlings/ha by age class. Coarse woody debris The mean amount of coarse woody debris decreased from age class 1 to age class 3, increased between age class 3 and age class 7, and declined between age class 7 and age class 9 (Figure 3-8). I 0.5 10.0 9.5 9.0 8.5 8.0 7.5 , ~ , 7.0 M S.O ^ 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1 .5 1.0 0.5 0.0 Q U 3 6 / \ 6 6 1 6 / \ 6 5 < / V 3 > <\u00E2\u0080\u00A2 J Age Class Figure 3-8. Quantity of coarse woody debris by age class. 63 Stands resulting from natural disturbance vs. those resulting from harvesting Comparison between stands resulting from natural disturbance and those resulting from harvesting was done for age classes 1 and 2, which represented the only two age classes regenerated after harvesting. For the purposes of the following analyses, natural disturbance was defined specifically as fire because sampled stands had all arisen after fire (i.e. the fires had happened within the last 40 years and had been recorded). In addition, \"treatment\" refers to the stand originating type of disturbance while \"location\" refers to the geographic location of the sampled plots (Section 3.2). Per-plot data can be found in Appendix II, while the results of all statistical tests are reported in Appendix III. In each of the following figures, error bars represent + 1 standard error from the mean. Different letters given above bars within an age class indicate significant (p<0.05) differences between fire and harvest origin stands; identical letters indicate no significant difference between fire and harvest origin stands. Abundance of live trees The abundance of live trees differed significantly between treatments in age class 1; there was no significant difference between locations nor was there a significant interaction between treatment and location (Figure 3-9). By age class 2, there was no significant difference between treatments or locations. Mean dbh of live trees The mean tree dbh was lower in fire origin stands than harvest origin stands in age class 1 while the opposite was true in age class 2. These differences, however, were not statistically significant (Figure 3-10). 64 Figure 3-9. Number of trees/ha in fire and harvest origin stands in each of the two age classes. ^ F i r e o r i g in E J H a r v e s t o r i g i n i A ge C lass 2 Figure 3-10. Mean tree dbh of fire and harvest origin stands in each of the two age classes. 65 Species composition of live trees: number of individuals There was a significantly greater number of spruce trees in harvest origin than in fire origin stands in both age classes (Figure 3-11); for spruce, there was no significant difference between locations in either age class. There was a significantly greater number of fir trees in fire origin than in harvest origin stands in age class 1; similarly, there was a significant difference between locations and a significant treatment x location interaction (Figure 3-12). Specifically, there were more fir in harvest origin stands in the south while there were more fir in fire origin stands in the north. There was no significant difference in number of fir between fire and harvest origin stands and difference between locations was not checked for due to the lack of fire origin plots in the south (Figure 3-11). 6000 5500 5000 4500 ca 4000 =5 U 3500 H S 3000 CO g 2500 3 ^ 2000 1500 1000 500 0 2 harvested 1 harvested 1 burned burned \u00E2\u0080\u0094i Eka 2 burned 1 harvested 1 burned 12 harvested Spruce Fit Species Figure 3-11. Species composition (number of individuals) of live trees in fire and harvest origin stands in each of the two age classes. 66 550 -500 -450 -400 -350 \u00E2\u0080\u00A2 300 -X l I Z 250 \u00E2\u0080\u00A2 200 -150 -100 -50 -0 -south north Location Figure 3-12. Treatment by location interaction for the number of fir in age class 1. Species composition of live trees: percent composition of the total number of trees The proportion of the total number of trees occupied by spruce was significantly higher in harvest origin than in fire origin stands in both age classes (Figure 3-13). In addition, there was a significant difference of spruce proportion between locations in age class 1; as noted earlier, a difference between locations in age class 2 was not tested for due to the lack of fire origin plots sampled in the south. The proportion of the total number of trees occupied by fir was significantly higher in fire origin than in harvest origin stands in both age classes (Figure 3-13). Lastly, the proportion of the total number of trees represented by fir was significantly different between locations in age class 1. 67 90% 80% 70% C3 O -*-\u00C2\u00BB O c 600 O 500 1 5 400 200 I \u00E2\u0080\u00A2 Fire origin ES! Harvest origin Age Class Figure 3-14. Number of snags/ha in fire and harvest origin stands in each of the two age classes. rue origin H Harvest origin Age class Figure 3-15. Number of seedlings/ha in fire and harvest origin stands in each of the two age classes. 69 Number of seedlings by species For both age classes, spruce was significantly more abundant in harvest origin stands than in fire origin stands; there was no significant difference between locations for spruce. Neither age class had significant differences between fire origin and harvest origin stands for the number of fir seedlings (Figure 3-16). While there was no significant difference between locations for fir, there was a significant treatment x location difference, with fire origin stands containing more fir seedlings than harvested stands in the south and fire origin stands containing less fir seedlings than harvested stands in the north. 6000 -i \u00E2\u0080\u0094\u00E2\u0080\u0094\u00E2\u0080\u0094\u00E2\u0080\u0094\u00E2\u0080\u0094\u00E2\u0080\u0094 \u00E2\u0080\u0094\u00E2\u0080\u0094 5 5 00 - 2 h \u00C2\u00BB \" e s \" = d 5000 - I 4500 Spruce F i r S p e c i e s Figure 3-16. Number of seedlings by species in fire origin and harvest origin stands in each of the two age classes. Percent composition of the total number of seedlings present Fir seedlings constituted a greater percentage of the total number of seedlings in fire origin stands than in harvest origin stands for both age classes (Figure 3-17); however, this difference was not significant in either age class. The percent of the total number of seedlings comprised of spruce was significantly higher in harvest origin stands than in fire origin stands in 70 both age classes (Figure 3-17). There was a significant treatment x location interaction in age class 1, with spruce occupying a higher proportion of the total number of seedlings in fire origin stands than in harvest origin stands in the north and a smaller proportion of the total number of seedlings in fire origin stands than in harvest origin stands in the south. (Figure 3-18). 100% -o Cu \u00E2\u0080\u00A2 Fir \u00E2\u0080\u00A2 Spruce 1 burned 1 harvested 2 burned 2 harvested Age class/treatment Figure 3-17. Percent of the total number of seedlings occupied by each species in each of the two age classes. Coarse woody debris There were significantly greater quantities of coarse woody debris in fire origin than in harvest origin stands in both age classes (Figure 3-19). In age class 1, there was more coarse woody debris in fire origin than harvest origin stands in both the north and the south, but this difference was only significant in the north (Figure 3-20). 71 north south Location Figure 3-18. Treatment by location interaction for the percent of the total number of seedlings represented by spruce in fire origin and harvest origin stands in age class 1. \u00E2\u0080\u00A2 Fire origin B Harvest origin i Age class 2 Figure 3-19. Quantity of coarse woody debris in fire origin and harvest origin stands in each of the two age classes. 72 6 3 ^ Fire origin H Harvest origin Location Figure 3-20. Treatment by location interaction for the quantity of coarse woody debris in fire origin and harvest origin stands in age class 1. Sample size As described in the methods section (page 57), I determined the sample size necessary to allow a reasonable chance of finding statistically significant differences for each variable that did not produce a significant difference as one way of explaining the lack of difference found between fire origin and harvest origin stands. The variables that this applied to were: dbh (each age class), number of live trees (age class 2), number of seedlings (both age classes), and number of fir seedlings. The resultant numbers of plots are listed in Table 3-4. Note that for age class 1, my data showed a significant difference in the mean number of live trees between fire origin and harvest origin plots, indicating that the sample size I used (six plots for each age class/treatment combination) was sufficient. Thus, an appropriate sample size was not determined for this variable for age class 1, but it was determined for age class 2, where no significant difference was detected in my data. 73 Table 3-4. Number of samples (n) required to estimate the true mean value of 4 stand variables to within 20% at the 95% confidence level. Age Class Variable n 1 dbh 48 # seedlings 17 # fir seedlings 105 2 dbh 15 # seedlings 9 # fir seedlings 35 # live trees 105 3.6 DISCUSSION Natural disturbance chronosequence The chronosequence offers an interesting look at ESSFwc2 naturally disturbed forests over time. While no statistical tests were performed on the observations made, the data illustrate some general trends which provide valuable insight into the ecology of these forests. Perry (1995) describes four successional stages of stand development: (1) stand initiation, (2) stem exclusion, (3) reinitiation, and (4) old growth. In my study, age classes 1 and 2 illustrate the stand initiation phase which is characterized by trees colonizing disturbed land in large numbers due to an abundance of sunlight and relatively little competition for nutrients and water. The time between age class 2 and about age class 4 illustrates the stem exclusion phase which, as Aplet et al. (1988) showed (and my study supports), in ESSFwc2 forests consists mainly of spruce stem exclusion. During this phase, trees die off due to the increased competition for nutrients, sunlight, and water that comes with tree growth. Age classes 4 to 7 illustrate the reinitiation stage whereby the mean number of trees remains the same and/or increases due to the opening of the canopy associated with tree die-off. The age class 6 that I sampled in the south had been affected by insect infestation, which is a possible reason for the decrease (although no significant difference was tested for) in mean number of live trees between age classes 5 and 6. The rise in 74 dead trees in age class 6 (Figure 3-6) illustrates the trade-off that took place between live trees (which decreased) and dead trees (which increased) as a result of insect induced disturbance. In addition, the age class 7 that was sampled had been affected by insect infestation as well, which could have created a higher mean number of small trees than would be found in a stand that was not under the influence of such infestation, due to the opening up of the canopy as a result of tree die-off (Figure 3-2). There were less snags in age class 7 than in age class 6; it is unclear as to why this was the case. The mean dbh generally increased until age class 7, at which time there was a decrease. The slight decrease from age class 4 to 5 might be due to two of the six plots in age class 5 being on north facing slopes whereas all age class 4 plots were on south facing slopes. The mean dbh for age classes 6 and 7 might be higher than would be expected, due to the opening of the canopy created by insect infestation. It is important to note, however, that insect infestation is a natural disturbance (like fire) with which these forests have interacted since their beginnings. Thus, even though the mean dbh (or mean number of trees) may be different than a \"typical\" successional stage, the values obtained in this study are reflective of natural fluctuations in ESSFwc2 forests. With regards to species composition, fir was most abundant in every age class, with spruce occupying varying amounts over time; this is consistent with other findings in other studies (Parish et al. 1999; Varga and Klinka 2001). While not statistically analyzed, a general increase in crown class diversity was observed, which is typical of forests as they grow older (Brokaw and Lent 1999), primarily because older forests have had time for multiple layers to develop. The crown class data collected in my study, however, should be interpreted cautiously as the methods used were questionable. A noticeable decline in the number of snags occurred between age classes 1 and 2; this could have been due either to collapse of snags with time or to more severe stand-originating fires in age class 2 which resulted in less snags left standing. Franklin et al. (1987) as well as Raphael and Morrison (1987) concur that mortality rates of live 75 trees are highest in younger serai stages. In my study, age classes 6 and 7 had the highest number of snags (along with age class 1), illustrating, most likely, the effects of insect infestation on these stands. Coarse woody debris plays an important role in ecosystem functioning. Many insects and small mammals use coarse woody debris as a source of food and for denning. These insects and small mammals, in turn, serve as prey for larger mammals and birds, creating a complex forest food web. In addition, coarse woody debris adds thermal and protective cover from prey for many small mammals (Maser et al. 1979; Craig 2002), although this has been found to be scale dependent in some cases (Craig 1995). Thus, it is possible that cwd, while perhaps preferred by certain small mammals, is not a critical habitat component (Feller 2003). Coarse woody debris acts as a nursery site for many bryophytes and plants which, in turn, aid in the collection and dispersal of water and nutrients. While coarse woody debris has been determined to play a major role in nutrient cycling (Harmon et al. 1986; Voller and Harrison 1998), Feller (2003) illustrates that there is some debate over the extent to which coarse woody debris influences nutrient cycling (Heilman 1990; Laiho and Prescott 1999). Coarse woody debris has been found to be highest in early successional stages (herbs and shrubs), lowest in intermediate age classes (e.g. 60-80 years), and high in old-growth stands (Spies et al. 1988; Voller and Harrison 1998). Feller (2003) reports several trends that have been reported in cwd abundance, one of these trends being the U-shaped curve illustrated by my data. The U-shaped curve results from high inputs of cwd into young age classes immediately following disturbance followed by a decrease in cwd due to decomposition. This decrease is then followed by an increase beginning in the stem exclusion phase and continuing through to the old growth phase where old trees begin to collapse (Feller 2003). 76 Fire origin vs. harvest origin stands Between fire origin and harvest origin stands, there was no evidence that geographic location affected abundance of live trees, growth (as indicated by dbh), or seedling abundance in either of the two age classes. While there was a significant difference in the number of live trees between fire origin and harvest origin stands in age class 1, this difference had disappeared by age class 2. Neither dbh nor seedling abundance were found to be significantly different between treatments in either age class. I expected to see a significant difference in dbh due to an increase in competition for nutrients, sunlight, and water created by the presence of more live trees in harvest origin stands than in fire origin stands. While there was no significant difference, it can be seen (Figure 3-10) that in age class 2 the mean tree dbh in fire origin stands is greater than the mean tree dbh in harvest origin stands, indicating that a larger sample size may have detected a significant difference (page 74). More noticeable differences lay in species composition of both live trees and seedlings, in the abundance of snags, and in the volume of coarse woody debris. There was a clear and significant difference in species composition between burned and harvested stands; specifically, harvested sites were planted with spruce, due to its superior economic value, while burned sites regenerated primarily with fir. From an ecological standpoint, this significant difference could eventually result in ESSFwc2 forests that are spruce, rather than fir, dominated. Under a natural disturbance regime, several studies (Aplet et al. 1988; Antos and Parish 2001; Varga and Klinka 2001) have suggested that subalpine fir is far more abundant in ESSF forests and that, in the absence of disturbance, subalpine fir could become the climax species. As previously noted, however, most of these studies were conducted in ESSF forests that differed from the ESSF forests in my study (i.e. my forests were wetter and/or colder), thus, conclusions drawn from these studies do not necessarily hold true for wet, cold ESSF forests. Furthermore, the literature is lacking in studies that have looked at the potential dependence of wildlife and plant species on 77 either subalpine fir or spruce, leaving open the possibility that a change in dominance from one to the other may not have any useful implications from the perspective of preserving biodiversity. Holsten et al. (1999) asserted that where spruce occupied 65% or more of the stand, spruce beetle (Dendroctonus rufipennis) infestation was more likely to occur, although this study was not specific to wet, cold ESSF forests. The spruce composition in both harvested age classes in my study area was around 80%, indicating that these stands may be more prone to spruce beetle outbreak than the naturally disturbed stands, although additional studies are needed to confirm this possibility. In the past 25 years spruce beetle has been estimated to be responsible for the loss of approximately 3 billion board feet (Holsten et al. 1999); based on this, a change in spruce composition seen in my project (e.g. 42% in age class 1 fire origin stands compared to 81% in age class 1 harvest origin stands) has the potential to change the dynamics of spruce presence as well as create a potentially less stable long-term harvesting situation (i.e. if spruce bark beetle are more likely to attack stands, less spruce will be available for harvesting). The lack of snags in harvested stands was another difference between the fire and harvest origin stands in my study area. The importance of snags for wildlife habitat is well documented (Harmon et al. 1986; Bull et al. 2001). Many species, including the fisher, marten, mountain chickadee, and red squirrel (which are all present in ESSFwc2 forests), require snags for food sources, rest, and nesting/home making. Simply put, if snags are not present, these species face increased danger from predation as well as other factors such as exposure to sunlight (for insects) and starvation). Snags can also provide an input of nutrients to the forest ecosystem through slow decomposition after falling down. Without this nutrient input, there may be a decline in nutrients in the soil, however, at this time, there is not enough evidence (from my study or others) to confidently conclude that soil nutrients will decrease significantly with a reduction in snags. The lesser amount of coarse woody debris in harvested sites may have a significant impacts on ESSFwc2 forests from both a wildlife habitat and a nutrient cycling perspective. 78 Feller (2003) makes clear, however, that additional studies into the habitat requirements (in regards to cwd) of small and large mammals, as well as the effects of cwd on nutrient cycling are necessary to better understand the functional role of cwd in forested ecosystems. 3.7 CONCLUSIONS This project illustrates some key differences between fire origin and harvest origin stands in ESSFwc2 forests. The abundance of fir, coarse woody debris, and snags were all significantly less in harvest origin stands than in fire origin stands; these differences have the potential to significantly impact forest ecosystem functioning and long-term sustainable harvesting. Growth did not seem to be affected by geographical location, nor was location a significant factor in the abundance of live trees or in the species composition of seedlings. This is a significant finding for forest managers as it indicates that trees establish and grow similarly throughout the range of the ESSFwc2, supporting the biogeoclimatic delineation of this variant. 79 CHAPTER 4 CONCLUSIONS AND RESULTANT MANAGEMENT RECOMMENDATIONS 4.1 LANDSCAPE LEVEL MANAGEMENT Components of ESSFwc2 forested landscapes that were considered in my study included: the abundance of old-growth forests; disturbance; landscape diversity; and patch characteristics. If the goal of forest management is to utilize forests in a way that closely mimics natural disturbance patterns, it is necessary to understand how these four components exist in naturally disturbed landscapes in order to identify approaches to management that are appropriate. My study highlighted areas where current harvesting seems to be creating landscapes similar to naturally disturbed landscapes as well as areas where harvesting needs to be altered in order to achieve the desired goal of emulation silviculture and, subsequently, the hypothesized preservation of biodiversity. Old-growth forests There has been a significant decrease in the amount of ESSFwc2 old-growth forests\u00E2\u0080\u0094a loss of nearly 2.0 x 103 km2. Old growth forest provides essential habitat for wildlife species, including perhaps the most studied in this variant\u00E2\u0080\u0094the threatened woodland caribou. Old growth provides unique combinations of structural characteristics such as coarse woody debris, large standing snags, and height class diversity that support ecosystem functioning and wildlife. My study shows significantly less coarse woody debris in harvested than in naturally disturbed stands, indicating that harvesting is not effectively mimicking natural disturbance in these forests. This decrease could have significant, long lasting effects on the ability of these forests to 80 cycle nutrients and provide wildlife habitat, although, as previously noted, more studies are required before definitive conclusions are possible. My study also shows a significant decrease in the amount of snags from naturally disturbed to harvested stands. Snags provide vertical diversity (which decreases species competition), provide wildlife habitat, and provide nutrient release as they decay. It is not known yet whether these early effects of harvesting will remain significant as these forests age; discerning this would be an interesting and valuable objective for future studies. It is my recommendation that the harvesting of old-growth cease and alternative, non-timber forest outputs (e.g. wildlife habitat and non-motorized recreational usage) be considered for these forests. Even if the present old growth (i.e. age classes 8 and 9) is left untouched, and, simultaneously, age class 7 (which will be old-growth in 20 years) is left untouched, it will still be over 20 years before pre-harvested levels of old-growth are attained (i.e. the area of age class 7 is only 1.49 x 103 km2 which is insufficient to make up for the 2.0 x 103 km2 loss of old-growth area). While it is acknowledged that there are still significant gaps in our knowledge of habitat requirements (e.g. amount of old-growth necessary to maintain viable populations of species currently found in these forests, dependence of species on ESSF forests specifically) of species found in ESSFwc2 forests, using the precautionary principle until we have definitive conclusions on this subject seems warranted. If these old-growth forests continue to be harvested it is possible that habitat and, therefore, certain species (i.e. those currently found in ESSFwc2 forests) could become threatened, although at this point this is uncertain. Based on this recommendation, the rest of my discussion seeks to make recommendations for younger forests (age class 6 or under). 81 Disturbance frequency The disturbance interval has decreased significantly as a result of increased harvesting in ESSFwc2 forests. The mean disturbance return interval dropped from 207 years pre harvesting to 138 years post harvesting, although, as previously noted, the post-harvesting interval should be interpreted with caution. Nevertheless, it is clear that disturbance is occurring much more often than it used to. It is important to remember that these disturbance return intervals represent solely when these forests are regenerating (i.e. due to fire or insects pre-harvesting and fire, insects, or harvesting post-harvesting), meaning that the mean fire return interval is most likely much longer than the 207 years determined for pre harvested landscapes. The British Columbia Ministry of Forests estimates that the mean fire return interval for ESSF forests is 250-350 years; there has been speculation that the mean fire return interval is, in fact, much longer than this estimate (Hawkes et al. 1997). This has significance for forest management in that it illustrates the difference between harvesting rotation times and natural disturbance \"rotations\". The time between disturbances affects the ability of a forest to recover from the disturbance; most specifically, this time allows the soil to recover, vegetation and wildlife habitat to return to the land, and nutrient cycling (via coarse woody debris and standing snag decay) to return to its pre-disturbance level. It is reasonable to assume that a reduction in this time impedes forest ecosystem recovery. Based on my findings, longer harvest rotations are necessary in order to bring the post-harvested disturbance regime closer to the natural disturbance regime. In addition, since fire is the major cause of disturbance in ESSFwc2 forests, future studies are needed in order to distinguish insect from fire disturbance and, therefore, to determine a fire return interval (rather than a disturbance return interval). It is quite possible that such studies may reveal an even larger disparity between natural and harvesting disturbance regimes, thus supporting my recommendation that harvesting rotation length be increased. 82 Landscape diversity Landscape diversity increased from pre to post harvesting which, upon first glance, may appear to be supporting harvesting as a way of increasing diversity. However, the determination of diversity is complicated by several factors, making a definitive answer difficult at this point. Firstly, diversity increased due to higher evenness, not due to higher richness (number of age classes remained constant). Magurran (1988) acknowledges that a more even distribution of species (here, forest patches) will result in an increase in diversity even if species richness is not affected or diminished. It is well recognized that an even distribution of forest patches does not reflect the natural mosaic created by fire (Fuller 1991) and can result in a decrease in or loss of species that are associated with or dependent on different age forests (Voller and Harrison 1998). Secondly, it is imperative that managers recognize that there are various forms of diversity and that an increase in one does not indicate an increase in another; in other words, an increase in forested landscape diversity does not equal an increase in species diversity within those forests. It is essential, therefore, that the increase in landscape diversity be seen for what it is and that it not be misused as a tool for promoting harvesting in order to increase species diversity (which is most commonly associated with biodiversity in the eyes of the public). Lastly, it is possible that continued harvesting will result, eventually, in a decrease in landscape diversity. Based on my findings, there are, at this time, no direct management recommendations to address landscape diversity (since it is \"increasing\"). I suggest, however, that extreme caution be used when interpreting these results, based on the considerations I have presented and that, in the very least, this increase in diversity not be misconstrued as an overall increase in biodiversity. As harvesting continues in these forests, it will be necessary to re-evaluate landscape diversity frequently in order to identify changes, if any, that occur over time. 83 Patch characteristics Average patch size and patch size variability decreased from pre to post harvested landscapes in age classes 8 and 9 where it was assumed all harvesting has taken place. This assumption was tested for accuracy by aging three stumps in each of the harvested plots and was supported by all trees being aged at between 150 and 300 years. My findings showed no significant decrease in perimeter/area ratios between pre and post harvested landscapes, indicating that, at this point in time, wildlife requiring interior habitat are not being adversely affected by harvesting. However, there were discrepancies within the data used, leaving open the possibility of potential differences that were not detected. In light of my data, there are several strategies that could be employed to bring harvested landscapes into the range of variability seen in naturally disturbed forests. In order to prevent the loss of wildlife habitat, travel corridors, and connectivity found in naturally disturbed forests (Voller and Harrison 1998), harvesting should create a wider range of variability in patch sizes. These patches should be connected to one another with ample amounts of travel corridors in order to provide wildlife movement. If natural regeneration is desired, patch sizes should be small enough to allow for the availability of sufficient seed source (within 100 m) but should be large enough to allow ample amounts of sunlight to reach the forest floor. Single-tree selection (depending on tree spacing) or group retention are two possible alternatives that could meet these criteria. Both of these alternatives would also provide a decreased risk of windthrow, based on findings in other studies and reviews (Stathers et al. 1994; lull and Sagar 2001). 4.2 STAND LEVEL MANAGEMENT In conjunction with landscape level planning, individual stand characteristics need to be considered when managing forests. My study aimed to quantify differences between stands of fire origin and those of harvesting origin in regards to several stand attributes, including: 84 abundance of living trees, snags, and seedlings; height class distribution of living trees; abundance of coarse woody debris; and species composition of both live trees and seedlings (regeneration). Clear and statistically significant differences were observed in both age classes studied (the only two age classes originating from harvesting; age classes 1 and 2) between the abundance of coarse woody debris and snags, as well as the species composition of both live trees and seedlings. These differences have implications for biodiversity preservation and, therefore, forest management. Fire as a management tool Fire suppression has been occurring in the ESSF zone for as long as harvesting has been occurring (i.e. 40-50 years). Fire suppression, while protecting stands of harvestable timber, can lead to inordinate amounts of fuel build up which has the potential to increase the severity of wildfires when they do burn in these forests. In addition, my study found significant differences between naturally disturbed and harvested landscapes; these differences could perhaps be diminished if fires were allowed to run their natural course at some times in some areas. Decreasing fire suppression could potentially alter patch size and patch variability as well as increase the abundance of coarse woody debris and snags, resulting in post-harvested landscapes that more closely represent pre-harvested ones. While slash burning has decreased in recent years in B.C. and elsewhere, I observed it occurring in some of the sites I sampled. Slash burning and prescribed burning should not be used in ESSFwc2 forests due to the susceptibility of both dominant tree species to fire. Coarse woody debris Coarse woody debris has traditionally been viewed as a hindrance to harvesting and reforestation (Voller and Harrison 1998). Clearcutting usually results in a removal of much coarse woody debris in order to reduce fuel loading and to create more desirable seedbeds 85 (Voller and Harrison 1998). There was significantly less coarse woody debris in stands originating from harvesting than in stands originating from fire. This difference could have both short term and long term effects on both wildlife and ecosystem functioning. There would probably be fewer insects and mammals that survive on and in coarse woody debris; this would necessarily affect those higher level animals dependant on these animals for food, although as previously noted, there have not been enough studies conducted to confidently assert much about the dependence of wildlife on ESSF cwd. The decrease in coarse woody debris results in a decrease in available nursery sites for seedlings and bryophytes, decreasing the ability of the forest to catch and slowly release water and nutrients. Given the fact that both spruce and fir seedlings seem to prefer exposed mineral soil as seedbeds (Day 1964; Feller 1998; Eastham 1999) and that they grow better in less shaded areas (spruce more so than fir)(Feller 1998; Eastham 1999), the soil coverage and shade provided by coarse woody debris may not be of great benefit to actual seedling germination and survival. However, coarse woody debris offers other benefits to ESSFwc2 forests (e.g. soil formation, nutrient cycling, water collection and dispersal, wildlife habitat) that are essential to short and long term forest health. With these considerations in mind, it seems wise to attempt to strike a balance between no coarse woody debris (or minimal amounts as is currently seen in harvested stands) and abundant amounts of evenly distributed coarse woody debris that may inhibit seedling germination and survival. Leaving coarse woody debris in some sections of harvested plots seems a viable solution that would both retain essential services offered by coarse woody debris and would eliminate negative impacts on seedling germination and survival. If the desired outcome of forest management is to mimic nature, then my study illustrates the need to retain more coarse woody debris in harvested ESSFwc2 forests and to arrange this coarse woody debris in a manner that minimizes shade to seedlings. This could be done by creating exposed mineral soil seedbed mounds surrounded by coarse woody debris. 86 Snags Standing dead trees (snags) provide vital habitat for many wildlife species (Voller and Harrison 1998). In addition, lichens, fungi, and invertebrates have all been found to be associated with snags (Huggard 1997). Stevenson (1993) found that harvesting ESSF forests caused lichens necessary for woodland caribou foraging to be reduced to one-third their pre-harvest level; this difference was attributed mainly to snag felling. Snags in ESSFwc2 forests most often result from insect infestation or fire. Currently in British Columbia there is some encouragement to retain snags (via wildlife patches) when harvesting forests (Stone et al. 2002). The felling of snags can have detrimental impacts on wildlife species dependant on such snags for habitat; it is also a dangerous and costly undertaking. Huggard (1997) found that recently dead subalpine fir snags had the slowest rate of fall-down for five designated stages of snag decay. The results of Huggard's study led him to conclude that recently dead snags should be retained in harvested areas in one of three manners: (1) reserve areas, such as wildlife patches, that retain groups of snags, (2) single stem retention throughout rotation, or (3) extended rotation lengths that allow younger trees the opportunity to become canopy and subcanopy snags (Huggard 1997). My study has shown that the abundance of snags in naturally disturbed forests is significantly greater than the abundance of snags in harvested forests, indicating that wildlife habitat is, and will continue to be, lost if efforts are not undertaken to increase the amount of standing dead trees in harvested stands. Given that fire in ESSFwc2 forests tends to burn irregularly, leaving patches of standing dead trees, I recommend that patches of recently dead snags be retained when harvesting takes place. In addition, I support Huggard's recommendation of extending rotation lengths in order to allow for younger trees to reach subcanopy and canopy heights before dying and becoming snags. The extension of rotations could have the added benefit of allowing some of the harvested forest to return to an old-growth stage, fitting with the 87 need to increase the amount of old-growth in ESSFwc2 forests (page 80). Leaving snags that are recently dead, as compared to those that are in later stages of decay, would provide wildlife habitat without unnecessarily increasing the potential for harm to worker's. Snag retention would save money that is currently used to pay snag fellers high wages (due to the danger of their job) and that is used for other harvesting procedures (e.g. timber cruising) that are made more dangerous and difficult by piles of felled snags. The size of snags being retained should be considered, as the diameter at breast height strongly influences which species use snags (Huggard 1997; Voller and Harrison 1998). Lastly, planted regeneration should consider the shade provided by standing dead trees and plant accordingly (i.e. out of the shade to allow for better growth). Species composition ESSFwc2 forests are named as such due to the co-existence of Engelmann spruce and subalpine fir trees. Subalpine fir dominate most stands at higher elevations (Farnden 1994), most likely due to a better adaptation to heavy snowpacks and its ability to reproduce vegetatively. Many have speculated that, in the absence of fire, subalpine fir would outcompete Engelmann spruce, becoming a singular dominating species (Knapp and Smith 1982; Aplet et al. 1988; Varga and Klinka 2001). However, because fire opens the canopy and allows for the less shade tolerant spruce to germinate, and because spruce is longer lived than subalpine fir, the presence of spruce in ESSFwc2 forests remains significant in naturally disturbed forests. Human influences in ESSFwc2 forests can be seen when we look at the species composition differences between naturally disturbed and harvested forests. As can be seen from my study, subalpine fir is more abundant than spruce in naturally disturbed forests, while the opposite is true in harvested forests. This is due to the fact that spruce is economically preferable to fir, and, thus, is planted in order to maximize profit from managed stands. As discussed in 88 chapter 3, the spruce beetle becomes a greater threat to ESSF forests that contain more than 65% spruce. While it is clear from my study that there are differences in species composition between naturally disturbed and harvested stands, it is unclear, at this time, what implications this change in species composition will have on these forests and the species living in them. My study makes clear the differences in species composition between naturally disturbed and harvested ESSFwc2 forests. It is unlikely that this will see change in the near future, due to foresters' preference of spruce over fir. However, it should be recognized that this economic preference is not supported by any ecological rationale; in other words, the forests that are being created with harvesting are not the same as the forests being created by fire. Namely, fir is usually dominant in naturally disturbed forests, whereas spruce is dominant in forests originating from harvesting. While it is still uncertain what these changes will result in, I recommend that caution be exerted when planting solely spruce until further research draws definitive conclusions. 4.3 LESSONS LEARNED My study offers an insight into the differences between naturally disturbed and harvested ESSFwc2 forests, both on the landscape and stand levels. These differences could have long-term implications for genetic, species, and ecosystem diversity and should, therefore, be taken into consideration when managing these forests. As is often the case with science, there are things I would do differently the second time around; the shortcomings of my study are shared here in the hopes of providing guidance to future studies that build upon the work I have done. Snags were not taken into consideration when determining diversity; snags represent a component that necessarily affects the overall picture of vertical diversity between naturally disturbed and harvested stands. It is quite possible that the significant decrease in snags between naturally disturbed and harvested stands would have significantly affected height class diversity 89 estimates. I recommend, for future studies, that height class be determined more consistently for every plot in order to minimize human error in estimations. In addition, I recommend that future studies record the height class of snags and include this in diversity analysis. The way in which coarse woody debris was estimated also left room for potential human error. Although I tried to minimize this by taking an average of three estimates per plot, if I were to do it again I would have the same person estimate coarse woody debris for all plots in order to further diminish the potential for human inconsistency and error. Using methods described by Zar (1999), I determined that my sample size for four different variables in each age class should have been larger. Testing in both age classes 1 and 2 for number of trees, dbh, number of seedlings, and number of fir seedlings indicated that between 9 and 105 plots were needed in order to statistically detect differences between fire origin and harvest origin stands for each age class with 80% confidence (testing for 0.05 significance. This project, therefore, would have benefited from two field seasons, rather than one, in order to obtain more samples. As previously noted, taking more time and obtaining more funding in order to distinguish fire from insect initiated disturbances is essential for future studies if accurate fire return intervals are to be determined. Shortcomings of my project that were beyond my control, but should be considered by those appropriate to handle such problems include the inaccuracy of Ministry of Forests digital data used in my landscape study and the inaccuracy of and associated difficulty in using Ministry of Forests forest cover maps. Once in the field, I found it extremely difficult to use these maps with any degree of certainty, as often there were roads on the maps that did not exist on land, and, more frequently, roads that existed that were not on the maps. For the most part, I found the age classes I was searching for, but considerable time was wasted trying to guess if I was where I thought I was. I ended up getting help from employees at several logging companies in the areas 90 I was sampling; these people provided me with maps that were much more accurate and easy to read (e.g. roads were actually named as they are in the field; Ministry of Forests maps often only name the major roads, making the locating of smaller side roads extremely difficult). While I recognize that logging companies are in these forests on a more regular basis and, therefore, have more up-to-date knowledge of, for instance, road names, I think more effort should be exerted towards making Ministry maps more useful. Research is essential to ensuring long-term ecological health and, from an economic point of view, to ensuring long-term availability of harvestable areas. Good research requires the ability to get into the field to make observations in a timely and efficient manner; this is currently being impaired by inadequate documentation of forested land data. With the current climate of major cutbacks within the Ministry of Forests, I am somewhat sympathetic to the contention that there are less people being asked to do more work, and I realize that there are no black-and-white answers to this problem. However, if research, and the tools necessary to carry it out (i.e. accurate maps), continues to place second behind the harvesting of forests, it is likely that neither long-term ecological health nor long-term sustainable harvesting will be achieved and, in the end, everyone -including logging companies, scientists, the general public\u00E2\u0080\u0094will pay the price. It is my hope that my work offers baseline knowledge into the disturbance regimes and patterns of ESSFwc2 forests and that this knowledge is expanded upon in future studies. Furthermore, it is my sincere hope that harvesting in the future is done in a manner that is consistent with preserving biodiversity, a concern that has been, and will surely continue to be, expressed by the people of British Columbia. 91 LITERATURE CITED Agee, J. 1993. Fire ecology of Pacific Northwest forests. Island Press, Washington, D.C. Alexander, R.R. 1964, 1967. Minimizing windfall around clearcuttings in spruce-fir forests. For. Sci. 10: 130-142. Alexander, R.R. 1969. Seedfall and establishment of Engelmann spruce in clearcut openings: A case history. USDA For. Serv. Res. Pap. RM-53. Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. Alexander, R.R. 1987. Ecology, silviculture, and management of the Engelmann spruce-subalpine fir type in the central and southern Rocky Mountains. USDA For. Serv. Agric. Handb. 659. Washington, D.C. Alexander, R.R., R.C. Shearer, and Wayne D. Shepperd. 1984. Silvical characteristics of subalpine fir. USDA For. Serv. Gen. Tech. Rep. RM-115. USDA For. Serv. Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. Alexander, R.R. and W.D. Shepperd 1990. Picea engelmannii Parry ex Engelm. Engelmann spruce. In: Silvics of North America. Volume 1: Conifers. Burns, R.M. and B.H. Honkala (compilers). 1990. USDA For. Serv. Agric. Handb. 654. Washington, D.C. Anonymous. 1995. Forest Practices Code of British Columbia. B.C. Min. For. Victoria, B.C. Antos, J.A. and R. Parish. 2001. Structure and dynamics of a nearly steady-state subalpine forest in south-central British Columbia, Canada. Oecologia, 130(1): 126-135. Aplet, G.H., R.D. Laven, and F.W. Smith. 1988. Patterns of community dynamics in Colorado Engelmann spruce-subalpine fir forests. Ecology, 69: 312-319. Aronoff, S. 1989. Geographic Information Systems: A Management Perspective. WDL Publications, Ottawa, ON. British Columbia Ministry of Forests. 1992. Old Growth Strategy Project: An Old growth strategy for British Columbia. Victoria, B.C. British Columbia Ministry of Forests. 1997. Silviculture Prescriptions Field Methods Book, Interim Draft. No. SIL411. Victoria, B.C. British Columbia Ministry of Forests. 2002a. http://www.for.gov.bc.ca/hfp/fordev/biodiversity British Columbia Ministry of Forests. 2002b. http://www.for.gov.bc.ca/hre/comonbio/ 92 British Columbia Ministry of Sustainable Resource Management. 2002. http://www.gov.bc.ca/srm Brokaw, N. and R. Lent. 1999. Vertical Structure. In Maintaining Biodiversity in Forest Ecosystems. M. Hunter Jr. (editor). Cambridge University Press, Cambridge, UK. pp. 373-399. Brown, J.K. and J.K. Smith (editors). 2000. Wildland Fire in Ecosystems: Effects of Fire on Flora. USDA For. Serv. Gen. Tech. Rep. RMRS-GTR-42-vol. 2. Fort Collins, CO. Bull, E.L., K.B. Aubry, B.C. Wales. 2001. Effects of disturbance on forest carnivores of conservation concern in Eastern Oregon and Washington. N.W. Sci. 75:180-184. Burns, R.M. and B.H. Honkala (compilers). 1990. Silvics of North America Volume 1: Conifers. USDA Agric. Handb. 654. Washington, D.C. Butt, G. 1990. Forest regeneration in the ESSF subzone: a problem analysis. B.C. Min. For. FRDA Rep. 118, Victoria, B.C. Clover Point Cartographies Ltd. 2003. http://www.cloverpoint.com/forestcovermaps.html Craig, V.J. 1995. Relationships between shrews (Sorex spp.) and downed wood in the Vancouver watersheds, B.C. M.Sc. thesis, University of British Columbia, Vancouver, B.C. Craig, V.J. 2002. Population and habitat use characteristics of forest-dwelling small mammals in relation to downed wood. Ph.D. thesis, University of British Columbia, Vancouver, B.C. . Crane, M.F. 1982. Fire ecology of Rocky Mountain Region forest habitat types. Final Rep. Contract No. 43-83X9-1-884. Missoula, MT. Day, R.J. 1964. The microenvironments occupied by spruce and fir regeneration in the Rocky Mountains. Cited in: Silvics of North America. Volume 1: Conifers. Burns, R.M. and B.H. Honkala (compilers). 1990. USDA For. Serv. Agric. Handb. 654. Washington, D.C. pp. 187-203. Delong, S. and D. Tanner. 1996. Managing the pattern of forest harvest: lessons from wildfire. Biodivers. Conserv. 5: 1191-1205. Eastham, A.M. 1999. Natural regeneration sub-project 6-year results. Lucille Mountain EP 1119ESSF Silvicultural Systems Study. Unpublished report, B.C. Min. For. Prince George, B.C. Environment Canada. 2002. http:Wwww.speciesatrisk.gc.ca/species/search Farnden, C. 1994. Forest regeneration in the ESSF zone of north-central British Columbia. Can. For. Serv. Pacific and Yukon Reg. Inf. Rep. BC-X-351. Pacific Forestry Centre, Victoria, B.C. Feller, M.C. 1997. Coarse Woody Debris in Forests: An Overview of the Coarse Woody Debris Study and the Sicamous Creek Study Area. In Sicamous Creek Silvicultural Systems Project: Workshop Proceedings. April 24-25, 1996, Working Pap. 24/1997. C. Hollstedt and A. Vyse (editors). B.C. Min. For. Victoria, B.C. pp. 134-143. 93 Feller, M.C. 1998. Influence of ecological conditions on Engelmann spruce and subalpine fir germinant survival and initial seedling growth in south-central British Columbia. For. Ecol. and Manage. 107:55-69. Feller, M.C. 2003. Coarse woody debris in the old-growth forests of British Columbia. Environ. Rev.: in press. Fischer, W.C. 1981. Photo guide for appraising downed woody fuels in Montana forests: Lodgepole pine and Engelmann spruce-subalpine fir cover types. USDA For. Serv. Gen. Tech. Rep. INT-98. Franklin, J.F. and C T . Dyrness. 1973. Natural Vegetation of Oregon and Washington. USDA For. Serv. Rep. PNW-8. PNW Forest and Range Experiment Station, Portland, OR. Franklin, J.F. and R.T. Forman. 1987. Creating landscape patterns by forest cutting: ecological consequences and principles. Landscape Ecol. 1:5-18. Franklin, J.F., H.H. Shugart, and M.E. Harmon. 1987. Tree death as an ecological process. Bioscience, 37(8):550-556. Goward, T. 1994. Notes on oldgrowth-dependent epiphytic macrolichens in inland British Columbia, Canada. Acta Bot. Fennica, 150: 31-38. Fuller, M. 1991. Forest Fires: An Introduction to Wildland Fire Behavior, Management, Firefighting, and Prevention. John Wiley and Sons, Inc. New York, N.Y. Harmon, M.E., J.F. Franklin, F.J. Swanson, P. Sollins, S.V. Gregory, J.D. Lattin, N.H. Anderson, S.P. Cline, N.G. Aumen, J.R. Sidell, G.W. Lienkaemper, K. Cromack Jr., and K.W. Cummins. 1986. Ecology of coarse woody debris in temperate ecosystems. Adv. Ecol. Res. 15:133-202. Hawkes, B., W. Vasbinder, C. Opio, and C. Delong. 1997. A Retrospective Fire Study, Interim Report: Fire in the SBS and ESSF Biogeoclimatic Zones of British Columbia-A Literature Review. McGregor Model Forest Association, Prince George, B.C. Heilman, P.E. 1990. Forest management challenged in the Pacific Northwest. J. For. 88: 16-23. Hollstedt, C. and A. Vyse (Editors). 1997. Sicamous Creek Silvicultural Systems Project: Workshop Proceedings. B.C. Min. For. Working Paper 24/1997, Victoria, B.C. Holsten, E.H., R.W. Their, AS. Munson, K.E. Gibson. 1999. The Spruce Beetle: Forest Insect and Disease Leaflet 127, USDA For. Serv. Hubbard, JA, S Archer, TW Boutton, RJ Ansley. 2000. Effects of fire and simulated grazing on root dynamics and soil respiration in a mixed-grass prairie. Ecological Society of America Annual Meetings, Snow Bird, UT. Huggard, D. 1997. Fall-down Rates of Subalpine Fir Snags at Sicamous Creek: Implications for Worker Safety and Habitat Supply. Extension Note 16. B.C. Min. For. Kamloops, B.C. 94 Huggard, D. and A. Arsenault. 1999. Comment-Reverse cumulative standing age distributions in fire-frequency analysis. Can. J. For. Res. 29: 1449-1456. Huggard, D. and A. Arsenault. 2001. Reply-standing age distributions and fire frequency analysis. Can. J. For. Res. 31: 369-371. Hunter, M, Jr. 1999. Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press, Cambridge, UK. Johnson, E.A. and S.L. Gutsell. 1994. Fire frequency methods, models, and interpretation. Adv. Ecol. Res. 25: 239-283. Johnson, E.A. and C E . Van Wagner. 1985. The theory and use of two fire history models. Can. J. For. Res. 15: 214-220. Jull, M. and B. Sagar. 2001. Wind and Windthrow Incidence. In The Lucille Mountain Study: 8-year Results of a Silvicultural Systems Trial in the Engelmann Spruce-Subalpine Fir Zone. M.J. Jull and S.K. Stevenson (editors). Res. Br. Working Pap. 59/2001, B.C. Min. For. Victoria, B.C. pp. 54-65. Jull, M.J. and S.K. Stevenson (Editors). 2001. The Lucille Mountain Study: 8-year Results of a Silvicultural Systems Trial in the Engelmann Spruce-Subalpine Fir Zone. B.C. Min. For. Res. Br. Working Pap. 59/2001, Victoria, B.C. Kayahara, G.J. 2000. The effect of coarse woody debris on site productivity of some forest sites in southwestern British Columbia. Ph.D thesis, University of British Columbia, Vancouver, B.C. Kilgore, B.M. 1987. The role of fire in wilderness: a state-of-knowledge review. USDA For. Serv. Gen. Tech. Rep. INT-220. Intermountain Res. Station, Ogden, UT. Kimmins, J.P. 1997. Forest ecology: a foundation for sustainable management. Macmillan, New York, N.Y. Knapp, A.K. and W.K. Smith. 1982. Factors influencing understory seedling establishment of Engelmann spruce and subalpine fir in southeast Wyoming. Can. J. Bot. 60: 2753-2761. Laiho, R. and C E . Prescott. 1999. The contribution of coarse woody debris to carbon, nitrogen, and phosphorus cycles in three Rocky Mountain coniferous forests. Can. J. For. Res. 29: 1592-1603. Lewis, K.J. and B.S. Lindgren. 2000. A conceptual model of biotic disturbance ecology in the central interior of B.C.: How forest management can turn Dr. Jekyll into Mr. Hyde. For. Chron. 76(3): 443-442. Lloyd, D., K. Angove, G. Hope, and C. Thompson. 1990. A Guide to Site Identification and Interpretation for the Kamloops Forest Region. B.C. Min. For. Res. Br. Land Management Handbook 23. Victoria, B.C. 95 MacKinnon, A. and T. Void. 1998. Old-Growth Forests Inventory for British Columbia, Canada. Nat. Areas J. 18: 309-316. Magurran, A.E. 1988. Ecological Diversity and Its Measurement. Princeton University Press, Princeton, N.J. Maser, C , R.G. Anderson, K. Cromack Jr., J.T. Williams, and R.E. Martin. 1979. Dead and down material. In Wildlife habitats in managed forests: the Blue Mountains of Oregon and Washington. Edited by J.W. Thomas. USDA Agric. Handb. No. 553. pp.78-95. McCaughey W.W., C E . Fiedler, and W.C. Schmidt 1991. Twenty-year natural regeneration following five silvicultural prescriptions in spruce-fir forests of the Intermountain west. USDA For. Serv. Res. Pap. INT-439. Intermountain For. and Range Exp. Station, Ogden, UT. McRae, D.J., L . C Duchesne, B. Freedman, T.J. Lynham, and S. Woodley. 2001. Comparisons between wildfire and forest harvesting and their implications in forest management. Environ. Rev. 9: 223-260. Meidinger, D. and J. Pojar (Editors). 1991. Ecosystems of British Columbia. B.C. Min. For. Res. Br. Victoria, B.C. Miller, G.T. 1998. Living in the Environment. 10th ed. Wadsworth Publishing Company, Belmont, California. Noble, D. L. and F. Ronco Jr. 1978. Seedfall and establishment of Engelmann spruce and subalpine fir in clearcut openings in Colorado. USDA For. Serv. Res. Pap. RM-200. Rocky Mountain For. and Range Exp. Station, Fort Collins, CO. Oliver, C D . and B.C. Larson. 1996. Forest Stand Dynamics. John Wiley & Sons, New York, N.Y. Panel on the Ecological Integrity of Canada's National Parks. 2000. Conserving ecological integrity with Canada's national parks unimpaired for future generations. Min. of Public Works and Gov. Serv. Ottawa, ON. Parish, R. 1997. Age and size structure of the forest at Sicamous Creek Silvicultural Project. In Sicamous Creek Silvicultural Systems Project: Workshop Proceedings. Edited by C. Hollstedt and A. Vyse. B.C. Min. For. Working Paper 24/1997, Victoria, B.C. pp. 16-31. Parish, R., J.A. Antos, and M.J. Fortin. 1999. Stand development in and old-growth subalpine forest in southern interior British Columbia. Can. J. For. Res. 29: 1347-1356. Parminter, J. 1983. Fire-ecological relationships for the biogeoclimatic zones of the Cassiar Timber Supply Area: Summary Report. In Northern Fire Ecology Project, Cassiar Timber Supply Area. B.C. Min. For. Victoria, B.C. Parminter, J. 1992. Typical historic patterns of wildfire disturbance by biogeoclimatic zone. B.C. Min. For. Res. Br. Victoria, B.C. 96 Perry, D.A. 1994. Forest Ecosystems. John Hopkins University Press, Baltimore, MD. Prescott, C E . 1999. Predicting rates of organic matter decomposition of forests and silvicultural systems in B.C. Final Rep. HQ96045-RE, For. Renewal B.C. Victoria, B.C. Pyne, S.J. 1993. Keeper of the Flame: A Survey of Anthropogenic Fire In Fire in the Environment: The Ecological, Atmospheric, and Climatic Importance of Vegetation Fires. P.J. Crutzen and J.G. Goldammer (Editors). John Wiley and Sons, New York, N.Y. pp. 245-266. Pyne, S.J. 1997. The Fire Practices of Aboriginal North Americans. In Ecoforestry: The Art and Science of Sustainable Forest Use. Drengson, A. and D. Taylor (Editors). New Society Publishers, Gabriola Island, B.C. pp. 182-186. Rajala, R.A. 1998. Clearcutting the Pacific Rain Forest. UBC Press, Vancouver, B.C. Raphael, M.G. and M.L. Morrison. 1987. Decay and dynamics of snags in the Sierra Nevada, California. For. Sci. 33(3):774-783. Reed, W.J. and E.A. Johnson. 1999. Reply-Reverse cumulative standing age distributions in fire-frequency analysis. Can. J. For. Res. 29:1812-1815. Ripple, W.J., G.A. Bradshaw, and T.A. Spies. 1991. Measuring forest landscape patterns in the Cascade Range of Oregon, USA. Biodivers. Conserv. 57: 73-88. Sachs, D., P. Sollins, and W. Cohen. 1998. Detecting Landscape Changes in the Interior of British Columbia from 1975-1992 using Satellite Imagery. Can. J. For. Res. 28: 23-26. Seymour, R. and M. Hunter Jr. 1999. Biological Diversity. In Maintaining Biodiversity in Forest Ecosystems. M. Hunter Jr. (editor). Cambridge University Press, Cambridge, UK. Pp. 3-21. Sierra Nevada Ecosystem Project. 1996. Summary of the Sierra Nevada Ecosystem Project Report. Centers for Water and Wildland Resources, University of California, Davis, pp. 46-58. Song, X. 1997. Relationships between coarse woody debris and understory vegetation in six forest ecosystems in British Columbia. M.Sc. thesis, University of British Columbia, Vancouver, B.C. Spies, T.A., JF. Franklin, and T.B. Thomas. 1988. Coarse woody debris in Douglas-fir forests of western Oregon and Washington. Ecology, 69(6): 1689-1702. Stathers, R.J., T.P. Rollerson, and S.J. Mitchell. 1994. Windthrow Handbook for British Columbia Forests, Res. Program Working Paper 9401. B.C. Min. For. Victoria, B.C. Stevenson, S.K. 1993. Alternative silviculture systems to maintain caribou habitat: Lucille Mountain Project. Establishment Report. Unpublished report prepared for BC Hydro. Cited in The Lucille Mountain Study: 8-year Results of a Silvicultural Systems Trial in the Engelmann Spruce-Subalpine Fir Zone. M.J. Jull and S.K. Stevenson (editors). Res. Br. Working Pap. 59/2001, B.C. Min. For. Victoria, B.C. 97 Steventon, D. 1997. Historic Disturbance Rates for Interior Biogeoclimatic Subzones of the Prince Rupert Forest Region. B.C. For. Serv. For. Sci. Extension Note No. 26. Prince Rupert, B.C. Stone, J., J. Parminter, A. Arsenault, T. Manning, N. Densmore, G. Davis, and A. MacKinnon. 2002. Dead Tree Management in British Columbia. USDA For. Serv. Gen. Tech. Rep. PSW-GTR-181. Varga, P. and K. Klinka. 2001. Structure of high-elevation, old-growth stands in west-central British Columbia. Can. J. For. Res. 31: 2098-2106. Voller, J. and S. Harrison. 1998. Conservation Biology Principles for Forested Landscapes. UBC Press, Vancouver, B.C. von Sacken, A. 1998. Interior Habitat. In Conservation Biology Principles for Forested Landscapes. J. Voller and S. Harrison (editors). UBC Press, Vancouver, B.C. pp. 130-143. Walstad, J.D., S.R. Radosevich, and D.V. Sandberg (editors). 1990. Natural and Prescribed Fire in the Pacific Northwest forests. Oregon State University Press, Corvallis, OR. Zar, J.H. 1999. Biostatistical Analysis. 4th ed. Prentice Hall, Upper Saddle River, N. J. 98 APPENDICES 99 APPENDIX I SITE SERIES, SLOPE, AND ASPECT OF EACH 400 m2 PLOT Plot IEF Slope Aspect Site Series Plot ID Slope Aspect Site Series 1-1-B-N 55% SW 01 4-3-N 10% SE 01 1-1-B-S 42% SE 01 5-1-N 24% SW 01 1-2-B-N 37% SW 01 5-1-S 4% SE 01 1-2-B-S 62% SW 5-2-N 16% SW 01 1-3-B-N 62% SW 01 5-2-S 15% NW 06 1-3-B-S SW 5-3-N 24% NE 01 2-1-B-N 18% NW 01 5-3-S 3% SE 01 2-2-B-N 54% SW 01 6-1-N 01 2-3-B-N 4% NE 01 6-1-S 35% NE 01 1-1-H-N 35% SW 01 6-2-N 14% SW 01 1-1 -H-S 8% 01 6-2-S 50% SE 01 1-2-H-N 10% SW 01 6-3-N 14% SW 01 1-2-H-S 8% SW 07 6-3-S 45% SE 01 1-3-H-N 25% SW 01 7-1-S 25% NW 01 1-3-H-S 13% SE 07 7-2-S 30% SW 01 2-1-H-N 4% SE 01 7-3-S 22% NW 01 2-1-H-S 5% SW 01 8-1-N 4% SE 01 2-2-H-N 11% SE 01 8-1-S 13% SE 01 2-2-H-S 16% SW 05 8-2-N 13% SE 01 2-3-H-N 15% SE 01 8-2-S 7% SE 01 2-3-H-S 11% SW 01 8-3-N 12% NE 01 3-1-N 13% NW 8-3-S 36% N 01 3-2-N 11% NW 01 9-1-N 30% SW 01 3-1-S 15% SE 01 9-1-S 15% NW 01 4-1-N 14% SE 01 9-2-N 20% SW 01 4-1-S 25% S 01 9-2-S 18% SE 01 4-2-N 14% SE 01 9-3-N 35% SW 01 4-2-S 25% 01 9-3-S 30% NW 01 a Plot ID is given as age class-plot number-burned or harvested (for age classes 1 and 2)-geographic location. 100 APPENDIX II PER-PLOT MEASUREMENTS OF 8 TRAITS (A-H) IN NATURALLY DISTURBED AGE CLASSES 1-9 AND HARVESTED AGE CLASSES 1 AND 2 For each of the following tables, Plot ID is given as age class-plot number-treatment (for age classes 1 and 2)-location. A. Number of live trees per 400 m plot Plot ID\u00C2\u00AB Trees Plot ID Trees Plot ID Trees 1-1-B-N 2 2-3-H-S 120 6-2-S 68 1-1-B-S 5 2-3-H-N 91 6-3-N 55 1-2-B-N 0 3-1-N 143 6-3-S 38 1-2-B-S 7 3-1-S 94 7-1-S 69 1-3-B-N 11 3-2-N 120 7-2-S 26 1-3-B-S 0 4-1-N 78 7-3-S 20 1-1-H-S 42 4-1-S 77 8-1-N 98 1-1-H-N 42 4-2-N 49 8-1-S 140 1-2-H-S 42 4-2-S 111 8-2-N 87 1-2-H-N 36 4-3-N 87 8-2-S 55 1-3-H-S 58 5-1-N 49 8-3-N 68 1-3-H-N 44 5-1-S 98 8-3-S 17 2-1-B-N 130 5-2-N 54 9-1-N 55 2-2-B-N 58 5-2-S 62 9-1-S 34 2-3-B-N 281 5-3-N 47 9-2-N 66 2-1-H-S 55 5-3-S 134 9-2-S 50 2-1-H-N 195 6-1-N 67 9-3-N 45 2-2-H-S 101 6-1-S 31 9-3-S 31 2-2-H-N 317 6-2-N 43 B. Mean dbh (cm) of live trees in each 400 m2 plot Plot ID Mean SE Plot ID Mean SE Plot ID Mean SE 1-1-B-N 1.85 0.35 1-1-B-S 4.38 1.86 1-2-B-N 0.00 0.00 2-3-H-S 2.93 0.19 2- 3-H-N 4.02 0.34 3- 1-N 4.16 0.25 6-2-S 12.91 0.92 6-3-N 17.72 1.62 6-3-S 21.81 1.62 101 Plot ID Mean SE 1-2-B-S 3.80 1.66 1-3-B-N 4.44 0.40 1-3-B-S 0.00 0.00 1-1-H-S 3.95 0.48 1-1-H-N 1.38 0.10 1-2-H-S 4.10 0.47 1-2-H-N 1.74 0.11 1-3-H-S 4.64 0.58 1-3-H-N 1.89 0.13 2-1-B-N 4.10 0.20 2-2-B-N 6.17 0.47 2-3-B-N 3.65 0.13 2-1-H-S 4.30 0.40 2-1-H-N 2.13 0.11 2-2-H-S 4.15 0.28 2-2-H-N 2.39 0.12 Plot ID Mean SE 3-1-S 9.58 0.96 3-2-N 3.80 0.19 4-1-N 15.44 1.04 4-1-S 13.50 1.23 4-2-N 18.97 1.09 4-2-S 7.75 0.70 4-3-N 16.64 0.88 5-1-N 18.53 1.63 5-1-S 10.08 0.70 5-2-N 15.94 1.41 5-2-S 10.67 1.32 5-3-N 18.22 1.73 5-3-S 8.28 0.65 6-1-N 14.07 0.99 6-1-S 19.27 1.43 6-2-N 20.92 1.77 Plot ID Mean SE 7- 1-\u00E2\u0080\u00A2S 6.40 1.14 7-2-\u00E2\u0080\u00A2s 23.89 1.94 7- 3-\u00E2\u0080\u00A2s 29.96 1.56 8-1--N 11.71 1.33 8--1--s 6.56 0.57 8-\u00E2\u0080\u00A22--N 10.12 1.45 8-\u00E2\u0080\u00A22--S 11.09 1.88 8--3 -N 16.19 1.77 8--3 -S 22.58 5.27 9--1 -N 12.08 1.95 9--1 -S 17.75 2.98 9-\u00E2\u0080\u00A22 -N 12.90 1.72 9--2 -S 12.32 1.93 9 -3 -N 17.52 2.25 9 -3 -S 17.77 2.78 C. Species composition of live trees in each 400 m2 plot Plot ID Spruce Fir % Spruce % Fir Plot ID Spruce Fir % Spruce % Fir 1-1-B-N 1 1 50 50 4-3-N 33 53 38 62 1-1-B-S 0 5 0 100 5-1-N 12 36 25 75 1-2-B-N 0 0 0 0 5-1-S 21 77 21 79 1-2-B-S 1 6 14 86 5-2-N 13 39 25 75 1-3-B-N 9 2 82 18 5-2-S 9 53 15 85 1-3-B-S 0 0 0 0 5-3-N 17 27 39 61 1-1-H-S 31 11 74 26 5-3-S 18 116 13 87 1-2-H-S 18 24 43 57 6-1-N 21 45 32 68 1-3-H-S 42 16 72 28 6-1-S 5 26 16 84 1-1-H-N 42 0 100 0 6-2-N 16 27 37 63 1-2-H-N 36 0 100 0 6-2-S 11 57 16 84 1-3-H-N 44 0 100 0 6-3-N 17 38 31 69 2-1-B-N 0 130 0 100 6-3-S 13 25 34 66 2-2-B-N 7 49 13 87 7-1-S 20 49 29 71 2-3-B-N 7 273 2 98 7-2-S 15 11 58 42 2-1-H-S 55 0 100 0 7-3-S 12 8 60 40 2-2-H-S 72 29 71 29 8-1-N 36 62 37 63 2-3-H-S 109 11 91 9 8-1-S 23 117 16 84 2-1-H-N 148 47 76 24 8-2-N 25 62 29 71 2-2-H-N 253 63 80 20 8-2-S 6 49 11 89 2-3-H-N 52 39 57 43 8-3-N 21 47 31 69 102 Plot ID Spruce Fir % Spruce % Fir Plot ID Spruce Fir % Spruce % Fir 3-1-N 32 111 22 78 8-3-S 2 15 12 88 3-1-S 13 81 14 86 9-1-N 6 49 11 89 3-2-N 16 102 14 86 9-1-S 9 25 26 74 4-1-N 20 58 26 74 9-2-N 9 57 14 86 4-1-S 25 52 32 68 9-2-S 14 36 28 72 4-2-N 15 34 31 69 9-3-N 5 40 11 89 4-2-S 25 86 23 77 9-3-S 5 26 16 84 D. Number of live trees per age class in each of five designated height classes Age Class Dominant Co-Dominant Intermediate Suppressed Veteran 1 Fire origin 18 6 0 0 0 2 Fire origin 177 292 0 0 0 3 6 93 116 139 3 4 96 79 70 157 0 5 88 75 89 192 0 6 32 124 70 76 0 7 18 26 13 58 0 8 47 40 64 306 8 9 28 42 60 150 1 1 Harvest 107 155 0 0 1 2 Harvest 423 455 3 3 0 E. Number of snags per 400 m plot Plot ID Snags Plot ID Snags Plot ID Snags Plot ID Snags 1-1-B-N 1-1-B-S 1-2-B-N 1-2-B-S 1-3-B-N 1-3-B-S 1-1-H-S 1-1-H-N 1-2-H-S 1-2-H-N 1-3-H-S 1- 3-H-N 2- 1-B-N 2-2-B-N 33 25 37 38 20 35 1 0 0 0 4 0 8 7 2-3-B-N 2-1-H-S 2-1-H-N 2-2-H-S 2-2-H-N 2-3-H-S 2- 3-H-N 3- 1-N 3-1-S 3- 2-N 4- 1-N 4-1-S 4-2-N 4-2-S 5 0 0 0 0 0 3 13 4 12 28 5 28 1 4- 3-N 5- 1-N 5-1-S 5-2-N 5-2-S 5-3-N 5- 3-S 6- 1-N 6-1-S 6-2-N 6-2-S 6-3-N 6- 3-S 7- 1-S 61 4 7 6 18 29 4 6 36 18 60 3 77 11 7-2-S 7- 3-S 8- 1-N 8-1-S 8-2-N 8-2-S 8-3-N 8- 3-S 9- 1-N 9-1-S 9-2-N 9-2-S 9-3-N 9-3-S 18 44 3 0 3 12 3 6 8 5 11 13 5 9 103 F. Number of seedlings per 400 m 2 plot Plot ID Seeds 1-1-B-N 12 1-1-B-S 13 1-2-B-N 21 1-2-B-S 70 1-3-B-N 37 1-3-B-S 18 1-1-H-N 39 1-2-H-N 40 1-3-H-N 40 1-1-H-S 73 1-2-H-S 25 1-3-H-S 74 2-1-B-N 83 2-2-B-N 36 Plot ID Seeds 2-3-B-N 156 2-1-H-S 88 2-1-H-N 366 2-2-H-S 274 2-2-H-N 212 2-3-H-S 173 2-3-H-N 194 3-1-N 101 3-2-N 83 3-1-S 301 4-1-N 119 4-1-S 131 4-2-N 104 4-2-S 328 Plot ID Seeds 4-3-N 74 5-1-N 105 5-1-S 122 5-2-N 134 5-2-S 99 5-3-N 66 5-3-S 217 6-1-N 36 6-1-S 144 6-2-N 239 6-2-S 51 6-3-N 216 6-3-S 667 7-1-S 238 Plot ID Seeds 7-2-S 129 7-3-S 116 8-1-N 89 8-1-S 214 8-2-N 132 8-2-S 710 8-3-N 122 8-3-S 207 9-1-N 279 9-1-S 221 9-2-N 102 9-2-S 92 9-3-N 284 9-3-S 81 G. Species composition of seedlings in each 400 m2 plot PLOT ID Spruce Fir % Spruce % Fir PLOT ID Spruce Fir % Spruce % Fir 1-1-B-N 7 5 58 42 4-3-N 1 69 1 99 1-1-B-S 4 7 36 64 5-1-N 14 84 14 86 1-2-B-N 12 8 60 40 5-1-S 1 111 1 99 1-2-B-S 17 22 44 56 5-2-N 1 119 1 99 1-3-B-N 4 33 11 89 5-2-S 3 93 3 97 1-3-B-S 2 16 11 89 5-3-N 15 47 24 76 1-1-H-N 5 34 13 87 5-3-S 21 192 10 90 1-2-H-N 16 24 40 60 6-1-N 1 35 3 97 1-3-H-N 18 22 45 55 6-1-S 4 135 3 97 1-1-H-S 64 9 88 12 6-2-N 21 217 9 91 1-2-H-S 18 7 72 28 6-2-S 8 43 16 84 1-3-H-S 66 8 89 11 6-3-N 15 201 7 93 2-1-B-N 3 80 4 96 6-3-S 95 571 14 86 2-2-B-N 2 34 6 94 7-1-S 66 172 28 72 2-3-B-N 3 153 2 98 7-2-S 24 105 19 81 2-1-H-S 57 31 65 35 7-3-S 33 83 28 72 2-1-H-N 97 268 27 73 8-1-N 6 81 7 93 2-2-H-S 113 160 41 59 8-1-S 77 137 36 64 2-2-H-N 153 58 73 27 8-2-N 26 106 20 80 2-3-H-S 70 103 40 60 8-2-S 244 466 34 66 2-3-H-N 71 123 37 63 8-3-N 46 76 38 62 3-1-N 6 95 6 94 8-3-S 31 176 15 85 104 PLOT ID Spruce Fir % Spruce % Fir PLOT ID Spruce Fir % Spruce % Fir 3-2-N 18 64 22 78 9-1-N 33 246 12 88 3-1-S 59 242 20 80 9-1-S 41 180 19 81 4-1-N 27 92 23 77 9-2-N 31 71 30 70 4-1-S 18 113 14 86 9-2-S 50 42 54 46 4-2-N 25 77 25 75 9-3-N 62 222 22 78 4-2-S 23 305 7 93 9-3-S 23 58 28 72 9 9 H. Quantity of coarse woody debris (kg/m ) per 400 m plot Plot ID CWD (kfi/m2) Plot ID CWD (kg/m2) Plot ID CWD (kg/m2) Plot ID CWD (kg/m2) 1-1-B-S 1-1-B-N 1-2-B-S 1-2-B-N 1-3-B-S 1-3-B-N 1-1-H-S 1-1-H-N 1-2-H-S 1-2-H-N 1-3-H-S 1- 3-H-N 2- 1-B-N 2-2-B-N 3.0 4.5 2.6 7.0 1.3 4.5 0.2 1.0 0.4 0.8 0.2 0.9 2.0 3.0 2-3-\u00E2\u0080\u00A2B-N 2- 1--H-N 2-2--H-N 2-3 -H-N 2-\u00E2\u0080\u00A21 -H-S 2--2 -H-S 2--3 -H-S 3-\u00E2\u0080\u00A21 -N 3--2 -N 3--1 -S 4--1 -N 4--1 -S 4--2 -N 4--2 -S 2.5 0.1 0.5 1.4 * 3.0 0.3 3.0 0.5 0.4 0.3 3.5 2.5 3.8 4- 3-N 5- 1-N 5-1-S 5-2-N 5-2-S 5-3-N 5- 3-S 6- 1-N 6-1-S 6-2-N 6-2-S 6-3-N 6- 3-S 7- 1-S 3.0 4.0 1.2 5.5 3.0 1.7 3.5 7.0 * 5.5 2.7 4.5 3.3 6.0 7-2-S 7- 3-S 8- 1-N 8-1-S 8-2-N 8-2-S 8-3-N 8- 3-S 9- 1-N 9-1-S 9-2-N 9-2-S 9-3-N 9-3-S 9.0 9.1 5.0 * 4.0 4.0 * 4.5 1.8 5.2 4.5 3.0 7.0 4.5 * indicates missing data. 105 APPENDIX III STATISTICAL TEST RESULTS FOR 13 VARIABLES (A-M) IN FIRE ORIGIN AND HARVEST ORIGIN STANDS A. ANOVA for number of live trees between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment 1 126.94 0.00 1 0.02 0.89 Location 1 0.80 0.40 Trtmt*Loc 1 0.98 0.35 B. ANOVA for dbh (cm) of live trees between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment 1 0.32 0.59 1 2.98 0.13 Location 1 2.81 0.13 Trtmt*Loc 1 1.02 0.34 C. ANOVA for abundance of live spruce trees between fire origin and harvest origin stands Source Age 1 Age 2 df F- ratio P df F-ratio P Treatment Location Trtmt*Loc 1 73.00 0.00 1 2.86 0.13 1 0.87 0.38 1 5.74 0.05 106 D. ANOVA for abundance of live fir trees between fire origin and harvest origin Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment Location Trtmt*Loc 1 8.40 0.02 1 21.36 0.00 1 11.34 0.01 1 2.35 0.20 E. ANOVA (after arcsine transformation) for percent composition of live spruce trees between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment Location Trtmt*Loc 1 38.06 0.00 1 43.34 0.00 1 0.76 0.42 1 20.98 0.00 F. ANOVA (after arcsine transformation) for percent composition of live fir trees between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment Location Trtmt*Loc 1 21.05 0.00 1 23.23 0.00 1 4.18 0.09 1 67.12 0.00 G. Kruskal-Wallis non-parametric test of variance for snags between fire origin and harvest origin stands. Age 1 Age 2 Group yV Rank Sum Group N Rank Sum Fire origin Harvest origin M-W a U-stat= 36 %2=8.61 df= 1 P=0.00 6 57 6 21 Fire origin Harvest origin M-WU-stat= 17 X2=5.12 df= 1 P=0.02 3 23 6 22 aMann-Whitney U statistic 107 H. ANOVA for abundance of seedlings between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment Location Trtmt*Loc 1 3.21 0.11 1 3.88 0.08 1 0.14 0.72 1 4.29 0.08 I. ANOVA for abundance of spruce seedlings between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment Location Trtmt*Loc 1 7.63 0.02 1 4.56 0.07 1 4.56 0.07 1 18.26 0.00 J. ANOVA for abundance of fir seedlings between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment Location Trtmt*Loc 1 0.47 0.51 1 0.07 0.80 1 32.07 0.00 1 0.73 0.44 K. ANOVA (after arcsine transformation) for percent composition of spruce seedlings between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment 1 7.69 0.02 1 7.25 0.05 Location 1 2.80 0.13 Trtmt*Loc 1 23.29 0.00 108 L. ANOVA (after arcsine transformation) for percent composition of fir seedling between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment Location Trtmt*Loc 1 4.24 0.07 1 0.48 0.51 1 16.42 0.00 1 17.23 0.01 M. ANOVA for quantity of coarse woody debris (kg/m2) between fire origin and harvest origin stands Source Age 1 Age 2 df F-ratio P df F-ratio P Treatment 1 43.31 0.00 1 14.54 0.02 Location 1 13.92 0.01 Trtmt*Loc 1 5.97 0.04 109 "@en . "Thesis/Dissertation"@en . "2003-11"@en . "10.14288/1.0075095"@en . "eng"@en . "Forestry"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Effects of natural disturbance and harvesting on the landscape and stand level structure of wet, cold Engelmann Spruce-Subalpine Fir forests of south-central British Columbia, Canada"@en . "Text"@en . "http://hdl.handle.net/2429/14291"@en .