Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Relationships between stand and site factors in naturally established fire-origin lodgepole pine stands… Brisco, David James 2001

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

Item Metadata

Download

Media
831-ubc_2001-0159.pdf [ 4.53MB ]
Metadata
JSON: 831-1.0075154.json
JSON-LD: 831-1.0075154-ld.json
RDF/XML (Pretty): 831-1.0075154-rdf.xml
RDF/JSON: 831-1.0075154-rdf.json
Turtle: 831-1.0075154-turtle.txt
N-Triples: 831-1.0075154-rdf-ntriples.txt
Original Record: 831-1.0075154-source.json
Full Text
831-1.0075154-fulltext.txt
Citation
831-1.0075154.ris

Full Text

RELATIONSHIPS BETWEEN STAND AND SITE FACTORS IN NATURALLY ESTABLISHED FIRE-ORIGIN LODGEPOLE PINE STANDS IN THE UPPER FOOTHILLS OF ALBERTA by DAVID JAMES BRISCO B.Sc, The University of British Columbia, 1993 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Forest Sciences Faculty of Forestry We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA March 2001 © David James Brisco, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT 60 sample plots within unmanaged, fire-origin, immature lodgepole pine dominated stands, were established in the Upper Foothills natural subregion of western Alberta. The plots were deliberately located to represent four forested ecosites: b (bearberry/lichen), d (Labrador t e a -mesic), e (tall bilberry/arnica), and h (Labrador tea-subhygric); and were used to study the relationships between ecological measures of site quality, and soil and stand attributes. Soil nutrient regimes (SNRs) and soil moisture regimes (SMRs) were quantitatively characterized for the study area. Significant differences in soil chemical properties were found between S N R s , in particular with those properties linked to soil fertility. Direct measures of soil nutrients agreed with field-identified SNRs , justifying their use. Following climatic and water balance analysis, relative S M R s were related to actual S M R s . The 'mesic' relative S M R is equivalent to the fresh actual S M R , with no significant water deficits or surpluses during the growing season. Relationships between foliar and soil nutrients were examined and quantified. Strong positive relationships between the two groups of variables demonstrate that increases in the availability of soil nutrients are reflected in increased foliar nutrient concentrations and improved stand nutrient status. This relationship has strong implications for operational fertilization. Of all categorical measures of site quality, S M R accounted for the greatest amount of variation (63%) in stand productivity. Stand density was explained relatively equally (-30%) by all three (SNR, S M R and ecosite) categorical measures of site quality. As a combination of both soil moisture and nutrient conditions, the use of ecosites as a management tool is justified, although its relationships with various measures of forest productivity and density are not as strong as would be expected. Along with additional stand measures such as density and age, ecosites can be used as a viable framework for developing site-specific silvicultural management strategies. All stands, regardless of ecosite, undergo severe reductions in lodgepole pine site index as density increases. The degree of height repression due to increasing stand densities is influenced by site quality. Stand productivity on S M R fresh sites demonstrated a significantly different re lat ionship with inc reas ing s tand densi t ies than obse rved on either moisture def ic ient or surp lus s i tes , hav ing a s teeper regress ion s lope and exhibi t ing much greater site index va lues at low dens i t ies . Wi th in the range of s i tes condi t ions invest igated during this resea rch , d i f ferences in ava i lab le soi l moisture a p p e a r e d to expla in the majority of obse rved d i f ferences in lodgepo le pine product iv i ty and deve lopment . At densi t ies greater than 30 ,000 s tems per hectare, the charac ter is t ics of 40-year -o ld s tands remain relatively uniform with increas ing densi ty . Th i s m a y qual i fy a s a definit ion of lodgepole pine 'stagnat ion' . TABLE OF CONTENTS Abstract : ii Acknowledgments )( Table of Contents •. iv List of Tables vi List of Figures v , ( l 1.0 Introduction "I 1.1 Lodgepole pine 1 1.1.1 Silvical Characteristics 1 1.1.2 High Density Stands 3 1.2 Ecosystem Classification in Alberta 4 1.3 Research Objectives 4 2.0 The Study Area 7 2.1 Climate 7 2.2 Topography and Soils 8 2.2.1 Windfall Burn Area (Marlboro 20 polygon) 9 2.2.2 Smith Creek Burn (Berland 6 polygon) 10 2.2.3 Gregg Burn (McLeod 4 polygon) 10 3.0 Materials and Methods 11 3.1 Sampling Design '• 11 3.2 Data Collection... 11 3.3 Analysis Methodology 14 3.3.1 Soil Moisture Regime Characterization 15 3.3.2 Soil Nutrient Regime Characterization 16 3.3.3 Soil - Foliar Nutrient Relationships 16 3.3.4 Relationships between Site Factors and Stand Productivity 17 3.3.5 Relationships between Site Factors and Stand Density 18 4.0 Quantitative Characterization of Soil Moisture and Nutrient Regimes 22 4.1 Introduction 22 4.2 Soil Moisture Regimes 25 4.2.1 Results and Discussion 25 4.3 Soil Nutrient Regimes 27 4.3.1 Results and Discussion 28 4.4 Conclusions 32 5.0 Foliar Nutrients 33 5.1 Introduction 33 iv 5.2 Stand Nutrient Status 33 5.2.1 Results and Discussion 33 5.3 Relationships between Soil and Foliar Nutrients 36 5.3.1 Results and Discussion 36 5.4 Conclusions 41 6.0 Relationships Between Stand Productivity and Site Properties 43 6.1 Introduction 43 6.2 Results and Discussion 43 6.2.1 Physical Site Factors 43 6.2.2 Site Index and Categorical Measures of Site Quality 46 6.3 Conclusions 50 7.0 Relationships Between Stand Density, and other Stand and Site Properties..... 51 7.1 Introduction 51 7.1.1 Relative Densities, Stand Development, and Structural Differentiation 53 7.1.2 Ecological Influences on Stand Density 55 7.2 Results and Discussion 56 7.2.1 Overall Relationships between Stand Density and other Stand Properties 56 7.2.2 Stagnation Thresholds 61 7.2.3 Stand Density Management Diagram 61 7.2.4 Stand Density and Categorical Measures of Site Quality 64 7.2.5 Ecosite-Specific Relationships between Stand Density and Productivity 66 7.3 Conclusions 69 8.0 General Conclusions 71 9.0 Literature Cited 73 Appendix 1 Measured Variables: Definitions and Descriptions 81 Appendix 2 Plot Data Summary: Overall and by Ecosite 83 Appendix 3 Regression Slope and Elevation Testing Output 85 v, LIST OF T A B L E S Table 2.1. Cimate data from Hinton weatherstation ID: 3063. (53°24'N 117°35'W, 1021m, 1964-1980) as supplied by Environment Canada 8 Table 3.1. Statistics and significance of dummy regression coefficients and slope interaction term for B and D ecosite comparison 20 Table 3.2. Statistics and significance of dummy regression coefficients without slope interaction term for B and D ecosite comparison 21 Table 4.1. Quantitative characterization of S M R s for the Upper Foothills natural subregion, based upon annual water balance and depth of the growing-season groundwater table :...27 Table 4.2. Classification matrix illustrating where the errors in field identification of soil nutrient regimes occurred and displays the percent correctly identified according to DA of chemical nutrient properties 29 Table 4.3. Results of the discriminat analysis of three SNR groups using five selected soil parameters as variables 30 Table 4.4. Mean values for mineral soil and forest floor properties for each SNR. Italicized text is one standard error of the mean 31 Table 5.1. Mean values for foliar macro- and micronutrient properties for each SNR (VP, very poor; P, poor; M, medium) 34 Table 5.2. Correlations between variable groups and their respective canonical variates; a) soil properties, and b) foliar properties 37 Table 5.3. Correlations between variables and their respective canonical variates; a) soil variables, and b) foliar variables 41 Table 6.1. Pearson correlation coefficents and associated probability levels between site characteristics and stand productivity (site index) 44 Table 6.2. Analysis of variance output of mean site index (m) of study stands by slope position. 45 Table 6.3. Stand productivity by slope position. Mean SI values with dissimilar superscripts are significantly different (Tukey HSD Test; a=0.05) 46 VI Table 6.4. Coeffients and R 2 values for various constructed additive A N O V A models predicting mean site index (m) 47 Table 7.1. Formulas for logarithmic and exponential functions 56 Table 7.2. Estimated parameters for quantification of relationships between stand properties and density 57 Table 7.3. Characteristics of dominant trees that may identify 40-year-old 'stagnant' stands 61 Table 7.4. Coeffients and R 2 values for various constructed additive A N O V A models predicting mean site index (m) 65 Table 7.5. Summary table of site index slope and elevation comparisons between ecosites. '(*) indicates that the comparison found significant differences between the ecosites 67 Table 7.6. Statistics and significance of dummy regression coefficients with slope interaction term for grouped B+H and D+E ecosite comparison 68 LIST OF FIGURES Figure 2.1. Map of the W E L D W O O D FMA indicating general locations of lodgepole pine study sites (small circles) 9 Figure 3.1. Edatopic grid illustrating ecosites investigated in this study (shaded boxes) in relation to soil moisture and nutrient regimes. (Modified from Beckingham etal. 1996) 12 Figure 3.2. Diagrams illustrating stand measurements, a) Index of stand height differentiation (IHD), and b) Estimation of crown volume, and crown volume to height ratio 14 Figure 3.3. Scattergram illustrating reductions in site index with increasing stand density on sites of different ecological quality (B and D ecosites) 20 Figure 4.1. Mean annual water balance for a fresh site located within a wetter montane boreal climate (Hinton, AB), showing all phases of water use: surplus, utilization, deficit, and recharge 26 Figure 4.2. Ordination of sample plots by soil nutrient regime classes as produced by discriminant analysis 30 Figure 5.1. Distribution of foliar nitrogen status among study stands. Graph a) frequency of nutrient status classes among ecosites. Graph b) distribution of study stands with respect to stand density and site index 36 Figure 5.2. Plots of first canonical variates for soil and foliar variables respectively, with study stands stratified by (a) SNR and (b) ecosite 38 Figure 5.3. Plots of second canonical variates for soil and foliar variables respectively, with study stands stratified by (a) SNR and (b) ecosite 39 Figure 5.4. Plots of first canonical variates for soil and foliar variables (reduced), with study stands stratified by (a) SNR and (b) ecosite 41 Figure 6.1. Mean site index (m) of study stands in relation to slope position. Error bars indicate one standard error for the mean 45 Figure 6.2. Trophosequences illustrating the effect of moisture upon site index across soil nutrient regimes 47 vi 11 Figure 6.3. S u r f a c e plot of the relat ionships be tween soi l moisture, soi l nutr ients, and s tand productivi ty. SI su face is interpolated with the d is tance weighted least s q u a r e s smooth ing techn ique ( D W L S ) 49 F igure 6.4. S i te index contour plot relating site index to combina t ions of S M R and S N R . Con tou r l ines are 1.25 metres apart 50 F igure 7.1. T r e n d s of se lec t s tand propert ies il lustrating the patterns in lodgepo le pine s tand charac ter is t ics as they c h a n g e with increas ing densi t ies 58 F igure 7.2. Scat terp lot i l lustrating ecos i te -spec i f i c re lat ionships be tween the index of height differentiation and s tand densi ty . L inear regress ion l ines for e a c h ecos i te a re d i sp layed to illustrate ecosite-specific IHD trends 60 Figure 7.3. S t a n d densi ty m a n a g e m e n t d iagram for natural lodgepo le pine s tands . A d a p t e d f rom original ( C a n a d i a n Fores t Se rv i ce , Pac i f i c and Y u k o n Reg ion) by C . F a r n d e n and D. B r i sco . Da ta S o u r c e : T A S S genera ted m a n a g e d s tand yield table 6 3 F igure 7.4. H y g r o s e q u e n c e s illustrating the effect of moisture upon s tand dens i ty a c r o s s soi l nutrient reg imes 65 F igure 7.5. Ecos i te -spec i f i c s imple l inear regress ion relat ionships be tween es t imated site index a n d s tand dens i ty 66 F igure 7.6. Sca t te rg ram illustrating the relat ionship be tween site index and s t e m s per hectare for s igni f icant ly different (a=0.05) g rouped ecos i tes (D+E and B+H) 69 ix ACKNOWLEDGMENTS This thesis was made possible by the work and assistance of many people. I would first like to thank my supervisor, Dr. Karel Klinka, for his guidance professionally and personally. Dr. Klinka hired me as a summer assistant in 1994, thus initiating a shift in my academic focus from the anatomical to the ecological. With his mentorship, both in the classroom and in the field, Dr. Klinka has taught me more than this thesis could ever represent. His door is always open and no question is trivial. I would also like to thank my committee members, Dr. P. Marshall and Dr. G. Weetman, for their knowledge and accessibility during my years of study. Weldwood of Canada - Hinton Division Ltd. and the Forest Resource Improvement Association of Alberta (F.R.I.A.A.) have been substantial financial supporters of this study. All the Weldwood staff involved have been extremely helpful with any requests. The late D. Presslee deserves special recognition for drawing attention to the need for research in this area. Without his considerable initiative, this study would not have been undertaken. I am grateful to C. Farnden for his assistance with the digital construction of the lodgepole pine stand density management diagram. I would like to thank the Klinkoid Research Group for their assistance and technical support over the years I have been in the Forest Sciences Department. In particular, C. Chourmouzis and B. Collins for travelling to Alberta to help with the fieldwork of this project, and P. Varga who is always available for numerous questions regarding statistics and software. I have spent seven enjoyable years working on a multitude of field projects with D. New. His wide range of practical knowledge and skills were always valuable assets in the field. He spent a significant time away from his busy family life to assist with this particular project, for which I am very grateful. Finally, and most significantly, I would like to thank my family and friends for their love and support. Kathryn, my wife, and our daughter Evelyn deserve special mention because their strength and encouragement have allowed me to focus on this project for the past few years. Their love gives me insight into the beauty and wonder of God's creation. X 1.0 INTRODUCTION 1.1 Lodgepole pine Lodgepole pine (Pinus contorta Doug. ex. Loud.) is a softwood species with a large range throughout Western North America, and is the most widely distributed pine species in western Canada. There are four variants of this species, divided primarily by geography (Lotan and Critchfield 1990). Pinus contorta var. contorta is found along the Pacific coast from the Alaskan panhandle to northern California. It is commonly known as shore pine or coast pine. Pinus contorta var. bolanderi is shrub-like and found in a localized region of Medicino County of California. It is commonly called Bolander pine. Pinus contorta var. murrayana can be found from the southern Cascade Mountains, to the Sierra Nevada, to northern Baja California. This tree is often known as Sierra lodgepole pine or Tamarack pine. Pinus contorta var. latifolia is the most widely distributed of all four varieties. Commonly known as Interior pine or Rocky Mountain lodgepole pine, its range extends from boreal regions of western Alberta, throughout the Rocky Mountains, to the inland coastal areas of British Columbia. To the north, it reaches the southern Yukon, and is found as far south as northwestern Colorado (Farrar 1995). It was in stands of the Rocky Mountain variety of lodgepole pine that this study was performed. 1.1.1 Silvical Characteristics Rocky Mountain lodgepole pine distinguishes itself from the other three varieties by possessing the following combination of silvical characteristics: fast juvenile growth; low shade tolerance, straight form; low taper; prolific seed production; and serotinous cones; which in combination with its ability to grow on a wide variety of site conditions (from dry sands to wet organic soils), make it a desirable species within managed forests (Smithers 1961; Pfister and Daubenmire 1975). With some exceptions, it is typically a serai species and is best maintained as a productive forest under an even-aged management system (Tackle 1961; Alexander and Edminster 1981) due to its low shade tolerance and regeneration strategy. The broad distribution of the species and its variants across such a large array of climates 1 and site conditions gives evidence to the wide ecological amplitude of this species. The soil nutrient requirements for lodgepole pine are known to be relatively low (Tackle 1961; Duffy 1964; Miller etal. 1978; Lotan and Perry 1983; Lotan and Critchfield 1990). This characteristic facilitates the establishment of lodgepole pine stands in conditions inhospitable to other associated species. While soil nutrients may not be a site factor strongly controlling lodgepole pine distribution and productivity, soil moisture and aeration (controlled to some degree by slope position and aspect) have been strongly implicated in influencing lodgepole pine growth rates (Dumanski et al. 1973). Both excesses and deficiencies of available moisture can contribute to substantial reductions in stand productivity, yet lodgepole pine is still often found growing under these conditions, outcompeting other tree species. Lodgepole pine's abundance in western North America dictates its importance on the landscape, in terms of both timber production and wildlife habitat. In British Columbia, 18.8 x10 6 m 3 of lodgepole pine was harvested in the 1997-98 fiscal year, out of a total harvest of 70 x10 6 m 3 , making it the single most important individual timber species in this province (Anonymous 1999) on a volume basis. In Alberta, 47% of the productive forestland is coniferous, of which 4 1 % is pine. Approximately 40,000 ha of coniferous forest were harvested in the 1994-95 fiscal year (Anonymous 1996). The wide ecological amplitude of lodgepole pine can be considered a boon to silviculturists attempting to restock lands that are too wet, too dry or too poor to support the productive growth of other timber crop species. However, the accumulation of serotinous cones and lodgepole pine's ability to survive even under adverse site conditions can lead to extremely dense stands following wildfires. These stands can achieve initial densities of up to 800,000 stems per hectare (Smithers 1961; Bassman and Crane 1983), resulting in varying degrees of height growth repression or even stand stagnation. Stands can remain in high density conditions for an extremely long time. A 70 year-old stand has been observed at 220,000 stems per hectare (Johnstone 1985), and at ninety years of age, lodgepole pine stands can still be found at densities of 12,500 stems per hectare (Smithers 1961). The potential for high regeneration densities, and the large losses in stand growth associated with them, has led to the development 2 of guidelines for maximum allowable stocking following harvesting (Marshall 1990; Anonymous 2000a). This is the opposite of typical stocking guidelines, which ensure that there are an adequate number of regenerated crop trees (free-to-grow) following harvest to fully occupy the site. 1 . 1 . 2 H i g h D e n s i t y S t a n d s Forest managers throughout British Columbia and Alberta have been concerned for some time (Mitchell and Goudie 1980; Silfor 1997, 2000) about the growth, stand development, and tending of high-density, immature lodgepole pine stands that have regenerated naturally following a large-scale disturbance such as wildfire. Trees in many of these high-density stands exhibit severely retarded height and radial growth in comparison to neighbouring stands with lower densities, apparently caused by the inability of these stands to self-thin and differentiate. With increasing utilization pressures upon forest resources, and a shrinking operable landbase, areas occupied by low-productivity stands of lodgepole pine are being targeted for inclusion in allowable cut determinations. Unfortunately, these immature stands offer very little in terms of standing timber for harvest and it remains questionable if some of them ever will. Foresters want to know what cultural activities are recommended to quickly bring these stands to some level of harvestable volume, or, at the least, will release them from their apparent state of stagnation. The underlying site factors that are influencing stand density need to be determined and understood if these stands are to be culturally manipulated to produce a desirable outcome. Horton (1953) states that high-density lodgepole pine stands can occur anywhere, regardless of site conditions. More recent observations in the literature (Duffy 1964, Dumanski et al. 1973) indicate that the occurrence of very high-density lodgepole pine stands may indeed be site-specific (i.e., density varies with site quality), directly contrasting Horton's findings. The development of an ecological site classification system for Alberta (Strong and Leggat 1981; Corns and Annas 1986; Beckingham etal. 1996) has provided a site-specific basis for both research and silvicultural management in western central Alberta. 3 1.2 Ecosystem Classification in Alberta In Alberta, an ecosystem classification system has been developed (Corns 1992; Beckingham etal. 1996) that somewhat parallels the biogeoclimatic ecosystem classification (BEC) system of B.C. (Pojar etal. 1987; Klinka etal. 1991; Meidinger and Pojar 1991). Climatic data were used to delineate the province of Alberta into six broad climatic regimes termed Natural Regions. The six Natural Regions are: Boreal Forest, Canadian Shield, Rocky Mountain, Foothills, Parkland and Grassland. Each Natural Region is subdivided into Natural Subregions, which are further partitioned into ecosites, phases, and at the most localized scale, plant community types. Natural regions and subregions are stratified primarily upon climatic and physiographic features of reference sites (Strong 1992). Finer divisions in ecological units (ecosites etc.) were determined through the analysis of vegetation, site, soil, and tree productivity data (Beckingham etal. 1996). Ecosites are the standard management unit, and are equivalent to site series within the B E C system. Throughout this manuscript, ecosites will be referred to as 'sites'. For example, a B.ecosite will be referred to as a 'b-site'. 1.3 Research Objectives Although categorical measures of site quality have been used for site description for many years in Alberta, they have been utilized in relative terms. A site under investigation might be termed 'mesic' or 'submesic' with no determination of its actual moisture availability. In the same sense, sites have been described as 'poor' or 'medium' in plant-available soil nutrient levels, but there has been no quantification of these relative classes. Is the use of these arbitrary class labels justifiable? Actual differences in available nutrients between relative nutrient regimes need to be established for their use to be constructive for site description and silvicultural planning. This research will attempt to quantify these qualitative variables for both soil moisture and nutrients within the study area. In addition, it will address the relationships between physical continuous and categorical measures of site quality, and some characteristics (density, productivity, foliar, and structural) of the study stands. There are two broad areas of investigation, with each broken into multiple specific objectives as outlined in the following list. 4 1. Re la t ionsh ips be tween soi l nutrients, fol iar nutrients and categor ica l m e a s u r e s of eco log ica l site quality. a . T o charac ter ize soi l nutrient reg imes of s tudy s tands . H 0 : The re are no signif icant d i f ferences in soi l nutrients be tween s tudy s i tes . b. T o charac ter ize soi l moisture reg imes of the study s tands . c. T o eva luate the foliar nutrient status of study s tands , and quanti fy the re lat ionships be tween soi l and foliar nutrients (determine if d i f ferences in so i l nutrients are ref lected in foliar nutrient concent ra t ions) . H 0 : The re are no re lat ionships be tween soi l nutrients and fol iar nutrients of s tudy s tands . 2. Re la t ionsh ips be tween categor ica l m e a s u r e s of site quality, s tand productivity, s tand densi ty , and other s tand character is t ics . a . T o exam ine the relat ionships be tween site index and categor ica l m e a s u r e s of site qual i ty, and determine wh ich m e a s u r e is most corre lated with site index H 0 : The re are no signif icant relat ionships be tween site index and site qual i ty. b. T o exam ine the relat ionships be tween s tand dens i ty and categor ica l m e a s u r e s of site quality, and determine wh ich m e a s u r e is most corre la ted with s tand densi ty . H 0 : The re are no signif icant relat ionships be tween s tand densi ty and site quality. c . T o e x a m i n e the s i te-speci f ic re lat ionships be tween s tand dens i ty and s tand productivi ty through l inear regress ions . H 0 : T h e relat ionship patterns be tween densi ty and s tand productivi ty is s imi lar a c r o s s ecos i tes . B e c a u s e a major e lement of this invest igat ion was to explore the re lat ionships be tween dens i ty and other s tand propert ies, s tands se lec ted for this study were those of a b o v e a v e r a g e dens i ty relative to sur round ing s tands . A s a result, m e a n densi t ies desc r ibed in this study m a y be h igher than reported e l sewhe re for s imi lar s tand a g e s and types . T h e a n a l y s e s of t hese dens i ty 5 relationships are based upon the following assumption: 'the range of initial regeneration densities is similar across ecosites; therefore, present differences in density ranges between ecosites are due to site factors and, subsequently, stand development.' The thesis is divided into eight chapters, each detailing the investigation of components of the overall study, with specific background information found at the beginning of each chapter. A description of the study area is found in chapter 2.0, and the sampling design and the soil-site sampling methodology can be found in chapter 3.0. The characterization and description of S M R s and S N R s is reported in chapter 4.0, and the investigation into the relationships between soil and foliar nutrients is presented in chapter 5.0. The relationships between stand productivity and site properties are examined in chapter 6.0. Chapter 7.0 investigates the relationship between stand density and other site and stand properties, in an overall and ecosite basis. The final summary conclusions of this thesis are found in chapter 8.0. Descriptions and summaries of measured variables and tables of statistical output are given in Appendices. 6 2.0 THE STUDY A R E A The sites sampled in this study are located within lodgepole pine dominated stands, naturally regenerated following wildfire. All study stands are approximately 40 years old; the products of three wildfires that occurred in the summer of 1956. These burned areas are situated roughly in a circle centred upon the town of Hinton; located along Highway 16, between Jasper and Edson (Figure 2.1). All three burns are found within the Forest Management Agreement (FMA) area of Weldwood of Canada Ltd., Hinton Division (WELDWOOD). 2.1 Climate Situated along the eastern slopes of the Rocky Mountains, where the study stands were located, is a component of the Foothills Natural Region (Strong 1992). The Foothills Natural Region is subdivided into the Upper and Lower Foothills natural subregions, which are distinguished elevationally (Beckingham etal. 1996). All study stands studied in this investigation are all located within the Upper Foothills natural subregion, the lower boundary of which varies from 900 m in the south to 1150 m in its northern range. The upper boundary of the Upper Foothills natural subregion varies from 1300 to 1500 m, separating it from the Montane natural subregion of the Rocky Mountain natural region. As indicated by the climatic data obtained from Environment Canada (Table 2.1), the Upper Foothills natural subregion is climatically equivalent to the Wet Cool Boreal White and Black Spruce subzone (BWBSwk) (Meidinger and Pojar 1991, Delong etal. 1990) of British Columbia. Both of these climatic units (Upper Foothills natural subregion and the BWBSwk subzone) have relatively dry winters, with the greatest precipitation occurring during the growing season, typical of the boreal climatic zone. 7 Table 2.1. Cimate data from Hinton weather station ID: 3063. (53'24,N 117°35'W, 1021m, 1964-1980) as supplied by Environment Canada. Parameters Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual Temperature Mean Daily Maximum fC) -6.1 0.3 2.8 9.3 15.2 19.9 22.1 21.8 16.5 11 -0.6 -4.3 9 Mean Daily Minimum fC) -16.9 -12 -9.8 -4.2 0.9 4.7 6.8 6.2 2 -1.6 -10.4 -15 -4.1 Mean Daily Temperature fC) -13.1 -6.5 -2.9 2.9 8.1 12.4 14.6 13.6 9.2 4.8 -5.0 -11.1 2.3 Extreme Daily Maximum fC) 10 12 13.6 18.7 23.8 27.9 29.5 30 25.7 22 12.3 8.2 31.5 Extreme Daily Minimum ('C) -35.2 -28.2 -26.2 -14.2 -4.7 -1.7 0.8 -0.5 -5.6 -11.4 -27 -32.3 -38.2 Precipitation Total Rainfall (mm) 0.9 0.2 0.2 6.8 52.6 69.1 70.6 87.5 34 20.4 0.6 0.6 343.5 Total Snowfall (mm) 21.3 16.3 17.7 20.6 3.7 0 0 0 3.4 4.1 16.7 16.7 120.5 Total Precipitation (mm) 27.2 18.1 17.8 31 41.3 78.7 99.4 61.3 38.9 21.9 16.9 20.1 472.7 Extreme Daily Rainfall (mm) 0.9 0.2 0.2 3.5 17.2 26.3 23.2 27.3 11.1 9.6 0.6 0.6 30.2 Extreme Daily Snowfall (mm) 10.7 6.2 7.6 8.6 2.2 0 0 0 2 3.5 7.6 8.1 21.6 Extreme Daily Precipitation (mm) 11.1 6.4 7.7 10 17.4 26.3 23.2 27.3 11.4 10 7.7 8.1 32.2 2.2 Topography and Soils With up to 100 kilometres separating the locations of the three burn locations, the parent materials contributing to the soils and topography of each locality can be quite dissimilar. The pH of soils in the Hinton-Edson area is typically highest (7.0-8.0) near the Athabasca River and close to the Rockies. As one moves east (and north or south) from Hinton the pH levels drop, with the values in the areas of our study stands in the range of 5.5 to 6.0 (Dumanski etal. 1972). These published values were confirmed by the chemical analysis results from the soils sampled du ring the course of this study. Humus forms are typically Morto Moders (Green et al. 1993), often with significant bryophyte (S-layer) cover. 8 Figure 2.1. Map of the WELDWOOD FMA indicating general locations of lodgepole pine study sites (small circles). 2.2.1 Windfall Burn Area (Marlboro 20 polygon) The Windfall burn is dominated by soils of the Mayberne association. This association is typically medium textured, dark yellowish brown till, and is often extremely cobbley. Several of our study stands are located at the outskirts of a gravel quarry used for road construction. Textures within this association are predominantly silt loams, sandy clay loams or sandy loams with an increasing proportion of clay particles in subsoil horizons. The local topography is determined by the formations of the underlying bedrock (Dumanski etal. 1972), with undulating plateaus and moderately sloping valleys. 9 2.2.2 S m i t h C r e e k B u r n (Ber land 6 p o l y g o n ) The soils within the Smith Creek burn are predominately of the Marlboro soil series association. This association consists chiefly of luvisolic soils developed on medium to fine-textured Marlboro till of Cordilleran source. The parent material of the Marlboro association is a friable and moderately stony olive to yellowish-brown coloured till. Generally, the till is clay to clay loam in texture, with less abundant loams and sandy loams observed. Several of our study sites uncovered partially decomposed sandstone deposits, which are noted as belonging to the Peskapoo Formation (Dumanski etal. 1972). This locality is dominated by moderately to strongly rolling terrain. Lower slopes, flats and depression areas are usually wet due to the fine textures of the soils that impede drainage. 2.2.3 G r e g g B u r n ( M c L e o d 4 p o l y g o n ) The Gregg burn is dominated by the Robb soil association. A collection of Gray Luvisols and Eutric and Dystirc Brunisols make up the majority of this association. Typically, the Robb soils have sandy loam to loam surface horizons, with clay loam to sandy clay loam subsoil horizons. Generally, the topography is more sharply and steeply sloped than that of the other burn areas, with the terrain moderately rolling to hilly. 10 3.0 MATERIALS AND METHODS 3.1 Sampling Design Sixty study sites were located across three areas affected by wildfire burns: Gregg, Smith Creek, and Windfall. All three burns occurred in 1956 producing the current even-aged 40-year-old lodgepole pine stands. These burns are found in the McCleod 4, Berland 6, and Marlboro 20 working circle polygons respectively (Figure 2.1). Five study stands for each of the four investigated ecosites were established within each burn, totalling twenty stands per burn, and fifteen stands per ecosite. Study site conditions were deliberately located to be uniform and representative of the four ecosites investigated. Potential study stand locations were determined primarily through field reconnaissance. The four ecosites investigated in this study represent approximately 55% of the area in the Upper Foothills natural subregion within the W E L D W O O D FMA (Downing 1999). These ecosites are: b (bearberry/lichen), d (Labrador tea-mesic), e (tall bilberry/arnica), and h (Labrador tea-subhygric) ecosites, with the majority of this area classified as d and e-sites. 3.2 Data Collection Once the plot perimeter (5 m x 5 m - horizontal distance) was measured and visibly marked, vegetation and environmental surveys were carried out. Vegetation was recorded in terms of plant species, percent cover and strata. A soil pit was excavated and described ( S C W G 1998) for each study stand. Mineral soil was sampled from the top 30 cm of the pit, with horizons sampled proportionately to their relative thickness within that zone. Depth to water table and depth to gleyed horizon were also measured in the soil pit. If no water table or gleyed horizons were detected, then the value was arbitrarily set to 100 cm for use in data analysis, approximately 30 cm deeper than the deepest observed rooting depth. Soil textures were estimated using a field key (Landon 1988) for the hand-texturing of mineral soils. 11 Soil nutrient regime VP P M R VR 1 b-sitc 2 n = 15-v c-site 3 4 d-site n = 15 e-site n = 15 f-site 5 MM 6 h-sito i-site j-site 7 8 k-site l-site m-site Figure 3.1. Edatopic grid illustrating ecosites investigated in this study (shaded boxes) in relation to soil moisture and nutrient regimes. (Modified from Beckingham etal. 1996). Each selected stand was clearly dominated by lodgepole pine in the canopy layer. Other tree species often found associated with the lodgepole pine were black spruce (Picea mariana (Mill) B.S.P.) , white spruce (Picea glauca (Moench) Voss) and various willows (Salix spp.), with these species found predominately in the upper shrub stratum or lower (intermediate) canopy positions. Stems within the plots were tallied for density calculations and described by three categories: 1) species; 2) live or dead; and 3) > or < 1.3 metres tall. The five tallest (dominant) and the five shortest (suppressed) live lodgepole pine trees were selected and measured for height, diameter at breast height (DBH), and length and width of live crown. Composite foliar samples were collected in late September (following first frosts and bud set) from the upper third of the crowns of these trees according to the methodology outlined by Ballard and Carter (1986). The needles were stripped from the current year's growth at each site following collection, and placed in paper bags for drying and storage until chemical analysis. Two of each five trees (dominant and suppressed) were randomly selected for tree-ring analysis. Discs were cut from the two selected trees at both the base and 1.3 metres. Stand age and radial increment were determined from these samples. Forest floor depth was measured in five random locations within the plot boundaries, and composite samples were collected. Composite mineral soil, forest floor, and foliar samples were submitted to Pacific Soil Analysis Inc. (Vancouver) for chemical analysis. All stem increment (radial) data was measured in the east and west directions and then averaged. Increment measurements were taken in the laboratory with the use of WinDendro computer software ver. 6.5 (R.I.I. 1999) and a high-resolution scanner. Average heights from the five dominant trees, and breast height age (BHA) determined by growth ring measurements were used to determine site index (SI) of the stand. SI was calculated using the formulas produced by Huang et al. (1997) and provided in Microsoft Excel® spreadsheet format by The Forestry Corp Ltd. (Edmonton, AB). The assessment of site quality (soil moisture and nutrient conditions) was performed according to the methodologies outlined by Green and Klinka (1994), with actual soil moisture and nutrient regime characterization described in more detail in chapter 4.0. Soil aeration regimes (SARs) were estimated as adequate, restricted, or deficient as characterized by the study of Wang and Klinka (1996) in the Sub-boreal Spruce zone of B.C. Along with standard stand measurements (i.e., dominant height, DBH, etc.), two additional stand properties were investigated for consideration. The index of stand height differentiation (IHD) is a new relative measure of the difference in height between the dominants and the suppressed trees within a stand (Figure 3.2a). It was intended to provide an estimate of the relative structural uniformity of a stand's canopy, and potentially distinguish productive stands from those that were 'stagnant'. Crown volume calculation (Figure 3.2b) and crown volume to height ratio (CRNHT) determination were calculated with the intention of providing a more accurate measure of crown health from which to base a stand's potential response to silvicultural interventions such as thinning. The crown volume factor used in these calculations is an ocular estimation of the proportion of a hypothetical cylinder, calculated from the live crown and crown width measurements, which would be filled by the crown of the measured tree. This would allow the crown volume measure to account for asymmetrical crown shapes and branching patterns. All investigated site and stand properties, including mineral soil, forest floor and foliar nutrients are listed and described in Appendix 1. Tables summarizing collected stand and site information of study stands are found in Appendix 2. 13 a) Index of Stand Height Differentiation b) Crown Volume Estimation Low IHD Diameter of Live Crown . Diameter of Live Crown Length of Live Crown High IHD Length of Live Crown I I I I I I I Volume Factor: O.S 0.8 Volume « Pi {r 2) * length live crown * volume factor Figure 3.2. Diagrams illustrating stand measurements, a) Index of stand height differentiation (IHD), and b) Estimation of crown volume, and crown volume to height ratio 3.3 Analysis Methodology Before any data can be used in parametric statistical analyses, it must meet certain assumptions (Zar 1984; Marshall et al. 1995). In particular, the variable data sets must be normally distributed and the individual datum must be independent (Sabin and Staford 1990; Hicks 1993; Neter et al. 1996). To meet the first criterion of normally distributed data points, transformations had to be performed upon several variables, in particular those concerning the soil and foliar chemical analyses values. All stand and chemical analysis data was examined for normality with the use of distribution histograms, probability plots, and skewness (one tail longer than the other) and kurtosis (flatness of the distribution) values. Most variable distributions that did not exhibit normality achieved a more normal distribution with a log(10) transformation. A few variables required square root transformations to achieve more normalized distributions. Where appropriate, model residuals generated from analyses were also examined graphically for normality and homoskedasticity of error variance (Sokal and Rohlf 1995). Since there are no time 14 interval measurements and all data comes from separate study stands, the second condition of data point independence appears to have been met. Unless otherwise specified, Tukey HSD multiple range tests (post hoc) were used to determine differences between investigated sample group means. Ellipses in subsequent figures are centred on the sample means of the x and y variables. The unbiased sample standard deviations of x and y determine its major axes and the sample covariance between x and y, its orientation. The size of the ellipse is specified by a probability value between 0 and 1, which was 0.7 in all instances in this manuscript. All statistical analyses were completed using S Y S T A T ver. 9.0 ( S P S S Inc. 1999a, 1999b) or the S A S STAT software ver.6.12 (SAS Institute Inc. 1990). Graphs were produced with either S Y S T A T ver. 9.0. or Sigmaplot v.4.01 3.3.1 S o i l M o i s t u r e R e g i m e Charac te r i za t ion Climatic data (Table 2.1) from the Hinton climate station was obtained from Environment Canada. These monthly climatic normals (1964-1980) were used, following the methods developed by Thornthwaite and Mather (1955, 1957), to arrive at a description of water balance on a 'zonal' site in the Hinton area (Figure 4.1) within the Upper Foothills natural subregion. The water balance is a bookkeeping method of balancing water inputs (precipitation), storage (soil water-holding capacities), and outputs (evapotranspiration and runoff). Evapotranspiration is the combined amount of water lost to the atmosphere due to evaporation from the soil surface and transpiration from vegetation. Actual evapotranspiration (AET) is the true or real rate of water loss that occurs on a site, given the real soil and moisture conditions present. Potential evapotranspiration (PET) represents water loss that could occur on a site, given the existing plant community and an unlimited supply of moisture (Klinka era/. 2000). The degree of actual evapotranspiration is dependent upon the amount of available soil moisture and temperature. A zonal (reference, or modal) site is one in which the influence of regional climate (mesoclimate) is not significantly altered by local edaphic or topographical features (Sukachev and Dylis 1964; Hills 1952). Typically, these sites have moderately deep soils of medium nutrient availability, friable consistency, loamy texture and coarse fragment contents less than 50% by volume. Zonal sites are found on gentle to moderate slopes, which are neither water shedding 15 nor water receiving (Meidinger and Pojar 1991). 3.3.2 Soil Nutrient Regime Characterization The soil nutrient availability of study sites was estimated in the field by identifying each site as belonging to one of three potential soil nutrient regimes (SNRs). SNR identification was based upon the field personnel's understanding of soil, vegetation, and topographical characteristics that can indicate the plant-available nutrients of a site. These indicators were used in combination to arrive at a final qualitative field-estimated SNR for the site. To quantify the field-estimated S N R s , composite soil and forest floor samples were collected from 60 study stands. Discriminant analysis (DA) was utilized to investigate and determine soil (mineral and forest floor) properties that may by used to characterize SNRs . DA identifies those soil properties that most easily discriminate between the three field-identified (a priori) S N R s . Nutrient values from chemical analysis were assigned to each plot, as was the field-estimated SNR. Using forward selection (F-probability level <0.15), soil variables were added to the discrimination procedure until none of the remaining variables could offer enough of an increase in discriminatory ability to be useful for analysis. Initially, this investigation included all mineral soil and forest floor variables. Many of these soil and foliar parameters were transformed (log(10)* or square root**) before analysis to meet normal distribution assumptions. For a complete listing and description of the following variables, refer to Appendix 1. The 22 soil variables explored were: pHMS, C M S * , NMS* , C N M S * , MinNMS, PMS* , KMS** , CaMS* , MgMS**, S M S * , S E C M S * , pHFF*, C F F , NFF, C N F F , MinNFF**, P F F , KFF** , C a F F * , MgFF*, S F F and S E C F F . Once the discriminating variables were selected, recommendations to 'correctly' re-label study site S N R s were derived from the Mahalanobis distance-square from group mean values and posterior probabilities for group membership. These recommendations were evaluated by re-examining vegetation, soil and topographical data. 3.3.3 Soil - Foliar Nutrient Relationships The relationships between soil (mineral and forest floor) and foliar nutrients were explored 16 with canonical correlation analysis (CCA). C C A allows the investigation of correlation between groups of variables instead of just one variable at a time. Combining the variable groups (soil and foliar) into linear combinations to be contrasted against one another, C C A describes the 'loading' (influence) of each individual soil and foliar property upon that between-group contrast. Many linear combinations are tested, with the combination possessing the greatest correlation between the variable groups labelled as the first canonical correlation coefficient. The combination producing the second highest independent correlation is the second canonical correlation and so on. By examining the output corresponding to the first few canonical coefficients, those individual soil and foliar properties having the greatest combined correlation with one another can be identified. Many of these soil and foliar parameters were transformed (log(10)* or square root**) before analysis to meet normal distribution assumptions. For a complete listing and description of the following variables, refer to Appendix 1. The 20 soil variables explored were: pHMS, C M S * , NMS* , C N M S * , MinNMS, PMS* , KMS** , CaMS* , MgMS**, S M S * , pHFF*, C F F , NFF , C N F F , M i n N F F " , P F F , KFF** , CaFF* , MgFF*, and S F F . The 15 foliar variables explored were: NeedWt*, FolN, FolP, FoICa, FolMg, FolK*, FolS, FolCu, FolZn, FolFe*, FoIMn*, FolB*, FolAI, FolNP and FolNS. A second C C A was run with a reduced number of soil and foliar variables in an attempt to increase the focus of the analysis. The soil variables were those that discriminated <0.15) between S N R s (MinNMS, NMS, C M S , C N F F , PFF) , as described in section 4.3. Important foliar variables were selected by regressing these parameters against site index (SI) as the dependent variable, using stepwise backwards selection (a <0.15), which reduced the these variables from 15 to 6 (Neter et al. 1996). The final foliar variables included in this second C C A procedure were: NeedWt, FolN, FolP, FolS, FolCu and FolB. 3.3.4 R e l a t i o n s h i p s be tween Si te F a c t o r s a n d S t a n d Produc t iv i t y Pearson correlation coefficients and associated probability levels were calculated to 17 determine general relationships between physical site factors and stand productivity. A single-factor A N O V A model was constructed to test for differences in productivity with respect to qualitative measures of slope position. To investigate the degree that synecological categorical measures (SNR, SMR, and ecosite) contribute to estimations of stand productivity, several single-factor G L M models were constructed using three S N R s , five S M R s and four ecosites. Site index was the dependent variable. Additionally, a two-factor model was made to examine the combined effect of S N R and S M R upon the variation in site index. Interactions between S M R and SNR were unable to be statistically tested because of the unbalanced nature of the data set. For example, the S N R V P sites were only found in combination with MD, SD, and F S M R s ; and the SNR M sites were only found in combination with F, and M SMRs . However, graphing the SNR x S M R cell means did not indicate any significant interactions between the different levels of these two factors (Figure 6.2), with site index following a similar pattern across S M R s for each SNR. The models were constructed with dummy variable coding and were not meant for prediction. They have not been validated statistically with an independent data set, and are meant only to explore the ability of synoptic ecological site properties to account for variation in site index/stand productivity. As with any single factor G L M model, the sum of the intercept and coefficient for any individual factor is actually the mean site index of that factor. 3.3.5 R e l a t i o n s h i p s be tween Si te F a c t o r s a n d S t a n d D e n s i t y Stand Density Management Diagram Construction It was intended that the study stand data from this investigation would be placed upon the standard stand density management diagram (SDMD) for naturally regenerated stands of lodgepole pine of B.C. as produced by Farnden (1996). Placing this data upon the density diagram would demonstrate where the study stands fit with respect to computer based stand development simulations. This is most appropriate since much of the stand development data from which the Tree And Stand Simulator (TASS) model (Anonymous 1998) was developed, calibrated and used for the construction of this SDMD, are from the same geographic region in 18 the Upper Foothills of western Alberta as the investigated study stands. Unfortunately, due to the high-densities and small sizes of the measured study stands, they were unable to be located upon the standard diagram. A new diagram had to be constructed that extended the Y-axis down (smaller mean tree volumes) and the X-axis further to the right (higher densities). WinTIPSY (Anonymous 2000b) software was utilized to generate the data points necessary to produce a new S D M D that would accommodate the collected data. Once the figure was constructed by hand, Craig Farnden assisted in the production of the new digital SDMD product. The mean and lower limits of the zone of imminent competition mortality (ICM) were extended into the new portion of the diagram. Mortality lines had to be adjusted slightly to more accurately reflect the data output provided by WinTIPSY. Stand Density and Categorical Measures of Site Quality The relationships between stand density and categorical measures of site quality were investigated in a similar fashion to the relationships between categorical measures of site quality and forest productivity (section 6.2.2). Similar single-factor and two-factor G L M models were constructed and tested (Table 7.4). These models were not meant for prediction, as they have not been validated statistically with an independent data set. Site-specific Patterns of Height Growth Repression Using general linear models with dummy variable coding (Freese 1964; Zar 1984; Neter et al. 1996), interaction factors were investigated for significance between the ecosite-specific regression slopes. Site index was the dependent variable and stems per hectare was the independent variable. Each pair of ecosites was contrasted independently, totalling six individual comparisons. An overall significance level of 0.05 was maintained through the multiple comparisons by using the Bonferoni procedure (Neter etal. 1996). Since there were 6 two-tailed contrasts, the slope interaction coefficient of the model must have had a two-tailed p-value of ^0.004 to be considered significant. If the slopes were not significantly different (i.e. they were parallel), then the model was re-run without the interaction factor, and the intercepts (elevation) of the regression functions were contrasted for differences. The significance level for this test was also the adjusted value of 0.004. 19 As an example of this procedure, the individual contrast between B and D ecosites is used for illustration (Figure 3.3). The output from the first analysis shows that the slope interaction term (LIVEPL*ECO) was not significant (p»0.004) , and therefore, the slopes of both the B and the D ecosites were determined to not be significantly different (Table 3.1). Table 3.1. Statistics and significance of dummy regression coefficients and slope interaction term for B and D ecosite comparison. Effect Coefficient Std Error t P(2 Tail) C O N S T A N T 17.045 1.024 16.653 <0.001 L IVEPL -0.000 0.000 -3.262 0.003 E C O -2.689 1.419 -1.895 0.069 L I V E P L * E C O 0.000 0.000 0.566 0.576 n=30 Adjusted R 2 : 0.675 Standard error of estimate: 1.477 P-value for total model: p < 0.0001 20 to CD >. O LO 15 x CD TJ _C w to TJ CU CO LU I I X d I I X ^ o \ o o E C O o o B I I o x D 0 10000 20000 30000 40000 50000 PI Density (stems per hectare) Figure 3.3. Scattergram illustrating reductions in site index with increasing stand density on sites of different ecological quality (B and D ecosites). The general linear model with one dummy variable (ECO) was re-run to determine if the levels (elevation) of the two regression lines were significantly different, with the insignificant slope interaction term removed. 20 Table 3.2. Statistics and significance of dummy regression coefficients without slope interaction term for B and D ecosite comparison. Effect Coefficient Std Error t P(2 Tail) C O N S T A N T 16.593 0.63125 26.28629 9.9E-16 L IVEPL -0.00014 0.00003 -5.22377 0.00002 E C O -1.95993 0.58677 -3.34022 0.00246 n=30 Adjusted R 2 : 0.683 P-value for total model: Standard error of estimate: 1. 458 p < 0.0001 The p-value of the dummy variable (ECO) was less than the 0.004 level required for significance (Table 3.2), indicating that the levels (elevations) of the two regression lines are significantly different. The pattern of SI reduction with increasing stand density is similar in slope (parallel) for both b and d ecosites, but they differ significantly in elevation. The b-site is located below the d-site. The five remaining comparisons were analyzed in a similar way, to determine which of the slopes and levels of the various ecosites were significantly different from one another. The full output from these tests can be found in Appendix 3, with the results summarized in Table 7.5. Ecosites that had similar regression lines were combined into groups to illustrate their overall relationships between site index and stand density. The grouped ecosites were then tested for differences. As this was only a single comparison t-test, the alpha level of significance was not adjusted with the Bonferoni procedure, remaining at 0.05. 21 4.0 QUANTITATIVE CHARACTERIZATION OF SOIL MOISTURE AND NUTRIENT REGIMES 4.1 Introduction Forests, as products of an enormous variety of climatogenic, pedogenic, and biogenic processes, are the most diversified and highly evolved entities encountered in the earth's biosphere (Krajina 1960). This makes forest classification both difficult and complex. To classify forest systems, the factors controlling forest and stand development must be understood. Solar radiation, temperature, moisture, aeration, and nutrients are primary abiotic factors (Progrebnyak 1930; Major 1963; Hills 1960, Spurr and Barnes 1997) determining the distribution patterns of vegetation and its productivity. Once these factors are identified, their continuous values need to be categorically described by unique classes. The number of classes and their limits are usually arbitrarily determined. The lower the number of classes, the more heterogeneous are the observations within a class, and the simpler the classification system; the greater the number of classes, the more similar the within-class observations, but the organization, interpretation, and management of these classes may be more complicated. To assist forest classification endeavours, Progrebnyak (1930) proposed a two coordinate (soil moisture and nutrient gradients) system for simplifying and synthesizing many site factors to characterize site quality. This system was adopted by Krajina (1969), and subsequently, his students to provide the backbone of the biogeoclimatic ecosystem classification system used in B.C. (Pojar etal. 1987, Meidinger and Pojar 1991). The two-coordinate edatopic grid provides a framework for site classification, which, when used within a climatic context, allows investigation into the interactions of site factors and their respective influence upon vegetation distribution, and productivity. The continuum of available soil nutrients has been stratified into five classes termed soil nutrient regimes (SNRs), and available soil moisture has been stratified into nine classes known as soil moisture regimes (SMRs). This produces forty-five combinations (edatopes) of SNR and S M R . In Canada, similar two-coordinate systems have been developed in Manitoba 22 (Mueller-Dombois 1964), Ontario (Hills 1952, Hills and Pierpoint 1960) and the Maritimes (Loucks 1962). A similar ecological site classification system has been developed in Great Britain (Pyatt 1995). Progrebnyak's edatopic grid provides many possible combinations of moisture and nutrient levels. Within a given climatic regime, sites of the same edatope (similar S N R and SMR) will provide comparable growing conditions for plants. In the case of tree species, patterns of stand productivity can be observed across edatopes as moisture and nutrients vary from deficient to optimal to surplus levels. This allows for preliminary explanations and comparisons of observed productivity levels, and can alert an investigator if a site is expressing lower than expected growth. Because of the high number of possible edatopes, it would be unnecessarily complicated and impractical to implement management guidelines for each edatope. Individual edatopes possessing qualities that produce relatively analogous growth conditions for plants may be combined and generalized, with the resulting groups representing similar levels of productivity. To accommodate differences in site quality and preserve a large enough number of site classes to easily identify factors influencing productivity, yet enabling the combination of edatopes providing relatively similar growth conditions, plant community information was integrated into site classification. Plant communities synthesize many ecological factors and may be used to indicate the ecological-equivalence (Bakuzis 1969) of different sites. The determination of the soil moisture and nutrient amplitudes (location on the edatopic grid) of different plant communities can indicate the locations of 'natural' boundaries between ecologically-equivalent sites. Within a similar climatic regime, the grouping of edatopes following this type of methodology has resulted in the site series groupings of the B E C system, and ecosite classes used in Alberta. Classifying sites in this manner allows the characterization of the edatopic group as being controlled by the same or similar limiting factors. The qualitative merging of many influential site factors into a synoptic few has facilitated investigations into the spatial distribution and classification of plant communities, and the processes that control them (Krajina 1969; La Roi and Hnatiuk 1980; Corns 1983, 1992; Klinka and Krajina 1986; Courtin et al. 1988). Attempting to model timber crop productivity using 23 individual morphological site properties (i.e. coarse fragment content of the soil) can be confusing due to compensating factors and interactions between the individual site properties (Pfister 1984), amounting to nothing more than 'fishing' for ecological relationships. Sites with different morphological characteristics may yield similar productivity due to comparable levels of available moisture and nutrients. These sites would be considered ecologically equivalent,.and demonstrate why the use of S N R s and S M R s is key to elucidating growth and productivity relationships within complex forest systems. The concept of site equivalence provides the basis of the B E C system that has been used to classify forest sites into a framework for resource management and scientific research. Since its introduction, the B E C system has become increasingly refined and specific within British Columbia (Pojar etal. 1987; Klinka etal. 1991). Actual soil moisture regimes (Green etal. 1984; Lloyd etal. 1990; Meidinger and Pojar 1991; Wang and Klinka 1996), soil aeration regimes (Wang and Klinka 1996), and soil nutrient regime classes (Kabzems and Klinka 1987a, 1987b; Courtin etal. 1988; Klinka etal. 1994; Wang 1997; Chen etal. 1998a; Splechtna 1999) have been characterized for many areas of the province, especially the southern coast. While plant communities and their relationships to site quality have been studied for some time in Alberta (La Roi et al. 1988; Strong etal. 1991; Corns 1992), soil moisture and nutrient regimes have not been described to the same extent as in B.C. Before relationships between various ecological factors can be firmly established, the factors in question need to be quantified. Once these basic values are substantiated and characterized, then further ecological relationships can be explored with greater confidence (Chen etal. 1998b). In addition, quantified data can be used as benchmarks for the evaluation of newly acquired data from difficult to diagnose sites or stands. The first objectives of this thesis are to objectively characterize and quantify the soil moisture and soil nutrient regimes in the study area, as estimated by field observations and samples collected from study stands. 24 4.2 Soil Moisture Regimes Soil moisture regime (SMR) represents the long-term balance between the amount of available soil water and the demand for that water by vascular plants (Klinka et al. 1989). Typically, S M R is assessed in the field using a relative scale (eg., 0 through 7 or very xeric through hydric), using soil textures, local topography and vegetation as indicators of available moisture. Despite their ease of identification in the field, relative S M R s express the actual available soil moisture inconsistently so that a xeric S M R under one climatic regime could be drier or wetter than a xeric S M R in another area influenced by a different climate. It seems inadequate to state that a site is the driest in relation to other sites within a certain climate regime without referring to the magnitude of that water deficit. Developing a classification of actual S M R s and relating these classes to relative S M R s will lead to improvements in identification and description of ecosites, as well as providing a better ecological uncerstanding of investigated ecosystems. 4.2.1 R e s u l t s a n d D i s c u s s i o n The monthly precipitation pattern (Table 2.1) shows very low values during the winter months with the bulk of the annual precipitation occurring during the growing season. This summer rainfall offsets the increased evapotranspiration pressures from the higher summer temperatures, eliminating water stress experienced by forest stands during the growing season on zonal sites. A marginal water deficit occurs from mid-May to early October (Figure 4.1), implying that zonal sites are experiencing 'slightly dry' actual moisture conditions (Table 4.1) during the growing season. In spite of this apparent deficit, 'fresh' was chosen to characterize the actual S M R of zonal sites rather than 'slightly dry' because the amount of water deficit during the growing season is relatively insignificant due to the amount of summer rainfall, and the A E T / P E T value is > 0.95. In addition, the investigated sites are, on average, 100-200 metres higher in elevation than the climate station used for this analysis. The amount of precipitation received at these higher elevations will be slightly greater than those recorded in Table 2.1, likely removing any potential water deficits during the growing season. Furthermore, it was found that the best 25 tree growth (SI) occurred on 'mesic' sites across the investigated hygrosequences (Figure 6.2). If there were significant water deficits on these zonal sites, then the tree growth would be better elsewhere (likely on the subhygric sites). 120 100 precipitation PET' AET "deficit utilization soil recharge I p T 1 I I 1 1 1 Jan Feb: Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 4.1. Mean annual water balance for a fresh site located within a wetter montane boreal climate (Hinton, AB), showing all phases of water use: surplus, utilization, deficit, and recharge. The assumed available water storage capacity of the soil is 200 mm. PET, potential evapotranspiration; AET, actual evapotranspiration. For determining S M R at specific sites, depth to water table or mottling/gleying was used to quantify moist, very moist, and wet sites. For those sites for which there was soil moisture deficits, the actual S M R assigned was determined through field-estimations of soil water holding capacity, relative abundance of indicator plant species (Klinka etal. 1989) and the correlations between relative S M R s and actual S M R s established in other studies (Klinka and Carter 1990; Klinka etal. 1996; Kayahara et al. 1997). The water balance models used by Spittlehouse and Black (1981) were not appropriate for use in this investigation. Required data, such as solar radiation, were not available for use, and the model parameters have been calibrated only for Douglas-fir stands in southern coastal B.C. 26 Table 4.1. Quantitative characterization of S M R s for the Upper Foothills natural subregion, based upon annual water balance and depth of the growing-season groundwater table (adapted from Klinka et al, 2000). The numerical relative moisture scale down the left side of the table is that used in Alberta, and is greater by one in value for equivalent moisture regimes in B.C. Current relative moisture regime Equivalent actual moisture regime Description of actual moisture regime Rooting zone groundwater absent during the growing-season: 2 Xeric Very dry (VD) Water deficit occurs >3 but < 5 months, or A E T / P E T < 0.75 but > 0.55 3 Subxeric Moderately dry (MD) Water deficit occurs >1.5 but £ 3 months, or A E T / P E T < 0.90 but > 0.75 4 Submesic Slightly dry (SD) Water deficit occurs >0.0 but < 1.5 months, or A E T / P E T > 0.90 5 Mesic Fresh (F) No water deficit (current needs for water do not exceed that stored in soil) Rooting zone groundwater present during growing-season: 6 Subhydric Moist (M) Groundwater table > 60 cm deep 7 Hygric Very moist (VM) Groundwater table > 30 but < 60cm deep 8 Subhydric Wet (W) Groundwater table >0 but £ 30 cm deep 9 Hydric Very wet (VW) Groundwater table at or above the ground surface 4.3 Soil Nutrient Regimes . Soil nutrient regimes (SNRs) represent the estimated amount of required plant-available nutrients found within the rooting zone of a site. These nutrients are typically related to N-availability and are assessed in the field using a 5-class relative scale (eg., 'A' through 'E ' or very poor (oligotrophic) through very rich (eutrophic)). Without quantification of these five classes, and ensuring that there are actually true nutritional differences between the field-estimated S N R s , their actual use is inconsistent. The number of classes may or may not be adequate to provide comprehensive characterizations of the plant available nutrients within the existing range of soil conditions. In this study however, only three S N R s were investigated (very poor, poor, and medium), occupying the 'poorer' portion of the edatopic grid. 27 4.3.1 R e s u l t s a n d D i s c u s s i o n Through the use of DA with forward selection (a=0.15), the soil variables that express the greatest discrimination between the three investigated S N R s are (in decreasing order of discriminatory ability): MinN, tN, and tC of the mineral soil, and tC:tN ratio and P of the forest floor. These soil properties are known (Armson 1977; Pritchett and Fisher 1987) to be important indicators of plant-available nutrients. Once the short list of five discriminatory variables was obtained, DA was further utilized to verify the accuracy of the field estimations of SNR (Kabzems and Klinka 1987a; Courtin era/. 1988; Wang 1997; Chen etal. 1998a; Splechtna 1999). The initial statistical run indicated that 92% (14/15), 78% (29/37) and 88% (7/8) of medium, poor and very poor sites, respectively, were identified correctly based upon discrimination by the soil nutrient data alone. By using Mahalanobis distance-square from group mean values and posterior probabilities for group membership, the procedure recommends the 'correct' SNR identification of the study plots, with some requiring new SNR labels. The majority of these suggestions occurred on poor sites, with the analysis implying that the sites were actually either slightly richer (medium) or poorer (very poor) than originally estimated in the field. Because the SNR recommendations are based upon the soil nutrient levels alone, and other site factors are not considered, the recommendations need to be evaluated before they are accepted as accurate. Upon re-examination of the original collected ecological data, several of these suggestions were accepted (for example, plot PR09, which was originally identified as 'poor', and was re-labelled as 'very poor') due to reconsideration of the vegetation and soil data. Some suggestions were rejected due to site characteristics that would prevent a site from being field-identified as belonging to the suggested S N R . For example, the suggestion that plot PM14 should be changed from an initially assessed S N R of poor to very poor was rejected because the soil possessed fine textures, soil colours were medium and the understory community did not support this change. The analysis was run a second time on the 're-classified' plots to see if there were more suggested changes. Suggestions were evaluated as before (Klinka etal. 1994) and appropriate changes were made. This procedure was repeated one more time and over the course of the entire procedure, 10 of the 60 study plots were 28 suggested for SNR re-labelling according to the linear discriminant posterior probabilities output. While four of these suggestions were rejected, six were accepted and their S N R labels were changed according to the discriminant analysis output and reconsideration of the site property information. Table 4.2. Classification matrix illustrating where the errors in field identification of soil nutrient regimes occurred and displays the percent correctly identified according to DA of chemical nutrient properties. Columns indicate field-idenitifed S N R s , while rows indicate S N R s assigned by DA. Boldface text indicates agreement between field-identified and DA assigned S N R s . M P V P % correct M 13 2 - 87 P 2 33 2 89 V P - - 8 100 Total 15 35 10 90 As shown by the classification matrix in Table 4.2, the final analysis results indicate that 87% (13/15), 89% (33/37), and 100% (8/8) of the medium, poor, and very poor study sites, respectively, were identified correctly in the field. There was no need to alter ecosite labels for any study plots, as all S N R changes were into adjacent SNRs , and the vegetation information did not require a site's placement into another ecosite. The discriminatory ability of the five selected soil and forest floor variables to differentiate study plots according to S N R is illustrated graphically in Figure 4.2. Stands are clustered according to SNR with very little overlap, expressing good dispersion, with the poor sites conspicuously located on the graph between the very poor and medium sites. This illustrates the continuum of the soil properties as they transition from lower to higher nutrient availability. 29 4 S N R o M x P + V P First Canon ica l Variate Figure 4.2. Ordination of sample plots by soil nutrient regime classes as produced by discriminant analysis. Ell ipses indicate 0.7 probability level. (VP, very poor; P, poor; M, medium). Table 4.3. Results of the discriminat analysis of three S N R groups using five selected soil parameters as variables. * indicates that the parameter was log(10) transformed to achieve normal distributions before analysis. Variable Standardized coefficients for the first two canonical variates 1 2 t C M S * 1.313 1.237 tNMS* -0.939 -1.864 MinNMS -1.132 0.521 tC: tNFF 0.649 -0.189 P F F -0.201 0.053 Eigenvalues: Canonical Correlations: Cumulative proportion of dispersion: 3.044 0.868 0.786 0.829 0.673 1.000 The derived canonical coefficients (standardized by within variances) (Table 4.3) give an indication of which variables contribute to the data point dispersion seen in Figure 4.2. Because some of the variables were log-transformed before the analysis, the coefficients are difficult to interpret precisely. As values decrease from right to left along the first canonical variate axis, log-transformed tN and non-transformed mineralizable nitrogen in the mineral soil, and the total phosphorus in the forest floor increase, while the log-transformed values for tC in the mineral soil, and the non-transformed tC:tN ratio of the forest floor decrease. As values of the second 30 canonical variate axis decrease, so do the values of log-transformed tC and untransformed MinN of the mineral soil and percent P of the forest floor. Log transformed tN of the mineral soil and tC:tN ratio of the forest floor increase as the axis values decrease. In relative terms, axis 1 is most heavily influenced by tN and MinN in the mineral soil, and the tC:tN ratio in the forest floor, and accounts for -79% of the graphed dispersion between the study stands. The second canonical variate axis is influenced primarily by tN of the mineral soil. tC of the mineral soil acts as a discriminating factor equally strongly on both axes. Table 4.4. Mean values for mineral soil and forest floor properties for each S N R . Italicized text is one standard error of the mean. Values in the same row with dissimilar superscripts are significantly different (post hoc Tukey pairwise comparison test; a=0.05). Some parameters were transformed (*log(10), " squa re root) before analysis to meet required distributional assumptions. S .E . is the standard error of the mean. S N R Medium Poor Very Poor n= 15 n= 35 n= 10 Mean S.E . Mean S .E . Mean S .E . Mineral Soil PH (%) 5.13 0.180 4.89 0.048 4.87 0.112 tC* (%) 1.15 0.149 0.96 0.099 0.73 0.087 tN* (%) 0.07 a 0.006 0.06 a 0.003 0.04 b 0.003 tC:tN ratio* 15 .94 a 1.084 15.20 a 0.728 20.66 b 1.399 MinN (ppm) 11 .23 a 0.733 5.25 b 0.466 4.76 b 0.471 P* (ppm) 12.20 2.256 24.57 3.611 21.30 3.490 S E C * (ppm) 2220.3 a 502.8 1468.8 a b 203.2 820.3 b 143.6 Forest Floor • pH* (%) 4.45 a 0.105 4 . 1 3 b 0.056 4.35 a b 0.156 tC (%) 45.41 0.471 45.97 0.467 47.25 1.046 tN (%) 1.29 a 0.028 1.09 b 0.021 1.13 b 0.048 tC:tN ratio 35.36 a 0.794 42.56 b 0.870 42.66 b 2.251 MinN** (ppm) 764.4 a 66.17 444.5 b 34.60 4 1 4 . 6 b 59.69 P (ppm) 200.5 20.50 164.9 11.10 132.2 22.72 S E C (ppm) 5569.1 737.8 4115.2 264.6 4714.1 678.7 Most mean soil nutrient levels decrease as the nutrient regimes shift from medium to very poor, some significantly (Table 4.4). In particular, mean values for total and mineralizable nitrogen significantly decrease in both the mineral soil and forest floor between medium and very poor sites. Mean tC:tN ratios in both the mineral soil and the forest floor increase from medium to very poor sites, indicating decreased availability of plant-useable forms of nitrogen (Pritchett and 31 Fisher 1987). The sum of exchangeable cations in the mineral soil decreases from medium to very poor sites. This characterization is considered a 'first approximation' because the data set is somewhat incomplete. The sites examined only occur across the relative soil nutrient regimes very poor (VP), poor (P), and medium (M). Until additional sites located on more productive soils (rich and very rich) are added to the database and re-analyzed in conjunction with the presented data, this preliminary analysis will remain applicable to only to the 'poorer' portion of the soil population in the study area. The final quantitative characterization of each SNR, as described by the discriminant soil variables, for selected soil properties is found in Table 4.4. All subsequent data analyses in this thesis utilizing study plots categorized by S N R are based upon the values given in Table 4.4. 4.4 Conclusions The water balance model enabled the conversion of the present relative S M R s to equivalent actual S M R s . A zonal site of 'mesic' moisture levels will have available soil moisture indicated by the 'fresh' actual SMR, i.e., no substantial moisture excesses or deficiencies during the growing season. Quantified actual SMR descriptions will improve the understanding of the influence of soil moisture on vegetation within this subregion, and additionally allow the comparison of productivity between similar actual S M R s across different subregions where actual soil moisture regimes are also known. There were significant differences in available soil nutrient levels among sites of different S N R s ; therefore, the null hypothesis (1a) is rejected. Discriminant analysis techniques identified soil nutrient properties suitable for characterization of three SNRs , validating the identification of S N R s in the field. In order of decreasing ability, five soil nutrient properties that are most influential in discrimination between SNRs were: mineralizeable nitrogen, total nitrogen, and total carbon of the mineral soil, and carbon to nitrogen ratio and percent phosphorus of the forest floor. This information provides a quantitative framework for SNR classification. It also allows the comparison of the inherent soil nutrient availability of the study area to that of other regions. 32 5.0 FOLIAR NUTRIENTS 5.1 Introduction Nutrient levels found within the foliage of dominant trees in a stand can give an indication of the nutritional status of that stand, and assist in the identification of nutrients that may be limiting its growth. This knowledge can provide explanations for ecological amplitudes of tree species, and has implications upon species' response to nutrient additions through fertilization. The objectives of this chapter were to 1) determine foliar nutrient levels of study stands, 2) determine which nutrients in the foliage are related to soil nutrients, 3) estimate the strength of those relationships, and 4) establish how these nutrient relationships are associated to measures of site quality (i.e., S N R s and ecosites). 5.2 Stand Nutrient Status 5.2.1 R e s u l t s a n d D i s c u s s i o n Foliar nutrient concentrations of study stands were summarized by soil nutrient regime (Table 5.1). Significant differences between S N R s were indicated for foliar levels of N, P, S, and needle weight (Table 5.1). Needle weight and foliar S on poor and very poor sites were found to be significantly lower than on medium. Foliar concentrations of nitrogen and phosphorus are also significantly lower on poor and very poor soils than on medium soils. Mean foliar zinc and boron concentrations decreased from medium to very poor sites, but not significantly. Mean values for other macro and micronutrients do not indicate any trends across SNRs . Considering all 60 stands where foliar sampling was performed, the guidelines proposed by Brockley (1996) and recently updated (Brockley 2001) indicate the following interpretations of stand nutrient status as characterized by foliar nutrient concentrations. 33 Table 5.1. Mean values for foliar macro- and micronutrient properties for each S N R (VP, very poor; P, poor; M, medium). Italicized text is one standard error of the mean. Values in the same row with dissimilar superscripts are significantly different (Tukey HSD Test; a=0.05). S .E . is the standard error of the mean. Soil Nutrient Regimes Foliar Variables M e d ! u c m P o ° r Very Poor n=15 n=35 n=10 Mean S.E . Mean S .E . Mean S . E NeedWt (g/100) 2.366 a 0.080 2.066 b 0.038 2.031 b 0.084 FolN (%) 1.219 a 0.028 1.109 b 0.020 1.030 b 0.037 FolP (%) 0.161 a 0.004 0.142" 0.003 0 . 1 3 2 b 0.005 FoICa (%) 0.146 0.007 0.146 0.004 0.151 0.007 FolMg (%) 0.113 0.003 0.108 0.002 0.108 0.004 FolK (%) 0.531 0.015 0.547 0.009 0.507 0.009 FolS (%) 0 .125 a 0.002 0 .110 b 0.002 0 .104 b 0.003 FolCu (ppm) 6.055 0.226 5.841 0.104 5.912 0.270 FolZn (ppm) 56.47 1.208 53.14 1.155 52.02 0.890 FolFe (ppm) 47.45 2.470 56.34 4.604 40.48 1.592 FoIMn (ppm) 480.8 32.35 501.6 24.19 451.7 30.54 FolB (ppm) 20.25 1.068 19.65 0.725 17.30 0.850 FolAI (ppm) 495.1 37.96 585.2 26.65 546.0 27.39 Twelve study stands are considered severely deficient in nitrogen; 23 are moderately deficient, and 24 are slightly deficient. Only one of the 60 sampled stands is considered to have an adequate concentration of foliar nitrogen. Most stands had adequate levels of phosphorus, with eight stands considered slightly deficient. Nitrogen to phosphorus ratios indicate that phosphorus deficiencies could be induced by nitrogen fertilization in two stands, but the remainder were sufficient. Sulphur concentrations in 10 of the 60 stands may indicate slight deficiencies, with the remaining 50 stands considered to have adequate sulphur levels. All sampled stands had foliar N/S ratios less than 12, suggesting that they possess adequate sulphur to undergo nitrogen fertilization without experiencing the induction of sulphur deficiencies. There were no deficiencies detected in potassium, calcium, or magnesium concentrations. All observed foliar macronutrient concentrations are similar in magnitude to those reported by Beaton etal. (1965), although the ranges of some collected data (phosphorus and potassium in particular) are narrower. This is likely due to the relatively small study area. 34 There were no deficiencies in iron, copper or zinc micronutrients in any of the study stands. Eight stands were marginally deficient in boron, with the induction of severe deficiencies probable with the application of N-fertilizers. The remaining stands had no indications of inducible boron deficiencies. Available foliar sulphate-sulphur and active Fe levels were not evaluated. Nitrogen is considered the most growth-limiting soil macronutrient (Armson, 1977; Zasoski 1979; Courtin etal., 1988; Carter etal. 1998). Because of the importance of nitrogen, foliar nitrogen status of the sampled stands was investigated in slightly more detail. The foliar nitrogen concentration of a stand can identify the presence (or lack) of a nitrogen deficiency and its magnitude, indicating potential growth response to increased soil nitrogen availability through the addition of fertilizer (Brockley 1996, 2000). Figure 5.1 illustrates the nitrogen status distribution of study stands relative to ecosite, density, and site index. All b-sites exhibited moderate to very severe foliar nitrogen deficiencies (Figure 5.1a). As expected, stands on e-sites (which have the highest soil nitrogen availability of the four sampled ecosites) generally had only slight N-deficiencies and 'possessed' the only stand that was considered to have 'adequate' levels of foliar nitrogen. The d- and h-sites had a range of deficiency levels. With respect to site index and density (Figure 5.1b), stand nutrient status generally improved as site index increased and density decreased, as indicated by the shape of the ellipses. However, a couple of stands with severe N deficiencies had unexpectedly high levels of productivity (site index > 15 m) at lower densities (approx. 10,000 sph) relative to stands with higher concentrations of nitrogen. In addition, it is surprising that the study stand with the greatest density (52,000 sph) has only a slight to moderate nitrogen deficiency. This is unexpected because the amount of nutrients available to an individual decreases as the number of stems on a site increases. Logically, the greater the number of individuals on a site with limited resources, the greater the deficiencies experienced by those individuals. These seemingly anomalous stands suggest that stand productivity and foliar status are not controlled entirely by nutrient availability. 35 12 o c: 0) 23 CT 0) D E Ecos i te • S E V a M O D es SLI H A D E H b) 10000 20000 30000 40000 50000 6000C PI Density (stems per hectare) Figure 5.1. Distribution of foliar nitrogen status among study stands. Graph a) frequency of nutrient status classes among ecosites. Graph b) distribution of study stands with respect to stand density and site index. Ellipses indicate 0.7 probability level. Stands were stratified by nutrient status (SEV, very severely deficient; MOD, moderate to severe deficiency; SLI, slight to moderate deficiency; A D E , adequate) from Brockley 2001. 5.3 Relationships between Soil and Foliar Nutrients 5.3.1 Results and Discussion Canonical correlation analysis (CCA) was utilized to investigate the relationships between soil and foliar nutrients, and whether changes in plant available soil nutrients were reflected in foliar nutrient concentrations. The magnitude of the first canonical correlation coefficient (0.90) indicates that there were several strong linear relationships between the soil and foliar data. With a multiple correlation coefficient value of 0.81, this combination of soil and foliar variables graphically clustered the stands according to estimated SNRs , but not so strongly when identified by ecosites (Figure 5.2), which include the effects of moisture. Although there are strong relationships between the groups of variables as explored in this combination, and the variables involved appear to be associated with SNR, this combination only accounted for 12% of the total 3 6 variation within the soil domain (Table 5.2a). The five soil variables most correlated with the first soil variate were tNMS, MinNMS, KMS, C a M S , and NFF. The first foliar variate in Table 5.2b indicated slightly stronger linear relationships within this variable group than those observed within the soils, accounting for approximately 14% of the variation. The foliar variables most correlated with the first variate combination are: FolN, FolP, FolS, and FolZn. All of these variables except FolZn are considered useful indicators of a stand's nutrient status, accounting for the dispersion/clustering by SNR demonstrated in Figure 5.2a. The large amount of overlap among ecosites B, D, and H (Figure 5.2b) implies that the nutrient status of these stands may be similar, which is a logical expectation as they were all located on poor to very poor sites. Table 5.2. Correlations between variable groups and their respective canonical variates; a) soil properties, and b) foliar properties. Only the first three canonical variates are listed. The variates are standardized. Some parameters were transformed (*log(10), "square root) before analysis to meet required distributional assumptions. a) Soil Canonical Variates b) Foliar Canonical Variates Variables SOIL1 SOIL2 SOIL3 Variables FOL1 FOL2 FOL3 pHMS 0.4226 0.2457 -0.6306 N E E D W T * 0.2503 -0.2916 0.2382 tCMS* 0.3545 -0.4671 -0.016 F O L N 0.7438 0.2544 -0.1035 tNMS* 0.4955 -0.2274 0.2206 F O L P 0.7053 0.0977 -0.0731 tC:tNMS* -0.0684 -0.5197 -0.3452 F O L C A 0.1417 0.3144 -0.7278 MinNMS 0.6285 -0.3494 -0.1398 F O L M G 0.2508 0.3056 -0.3485 P M S * -0.0545 0.4471 0.5279 F O L K * 0.0736 0.022 0.1161 KMS** 0.4469 0.2813 -0.17 F O L S 0.605 -0.4819 0.0913 C A M S * 0.4616 0.2322 -0.3376 F O L C U -0.2247 -0.2479 0.2152 M G M S * * 0.3174 0.235 -0.2085 F O L Z N 0.4538 0.3818 -0.1409 S M S * -0.1965 -0.4421 0.3209 F O L F E * 0.1876 0.3693 0.1303 p H F F * 0.3544 0.2155 -0.7206 F O L M N * 0.2374 0.1511 0.5132 t C F F 0.0507 0.1772 0.2128 F O L B * 0.428 0.1781 0.1115 tNFF 0.4795 -0.0876 -0.1085 F O L A L -0.0585 0.6458 0.2518 tC:tNFF -0.3895 0.1445 0.17 F O L N P -0.1557 0.1625 -0.0243 MinNFF** 0.4227 -0.3811 0.1823 F O L N S 0.1258 0.8407 -0.2344 P F F 0.1166 -0.4968 0.586 K F F * * 0.1747 0.0925 0.2952 C A F F * 0.3087 0.2123 -0.7427 M G F F * 0.2678 0.2072 -0.7515 S F F 0.1500 -0.1945 0.4218 Variance 0.1213 0.0969 0.1758 Variance 0.1428 0.1422 0.0816 Cumulative 0.1213 0.2183 0.3941 Cumulative 0.1428 0.2850 0.3666 37 The second canonical correlation coefficient (0.891) indicated that there were strong linear correlations between soil and foliar groups of data in this variate combination as well. As the second independent variate combination most maximizes (R 2 = 0.79) the linear correlation between soil and foliar variables, graphing the canonical values (Figure 5.3) demonstrated that this combination does not 'cluster' the plots according to categorical measures of nutrient availability (SNR) or site quality (ecosite) as well as the first variate. The most influential soil variable contributors to this linear combination were: tCMS, tC:tNMS, P M S and P F F (Table 5.2a). The major foliar contributors (Table 5.2b) were: FolS, FolAI. and FolN:S. These soil and foliar parameters are not as strongly linked to SNRs or ecosite labels as those influencing the first canonical variate, thereby explaining this combination's inability to cluster the study stands by either of these categories. 38 3 3 8 co o B x Q + E A H ECO -3 o M x P + VP •3 -3 -1 1 Second Soil Canonical Variate 3 -3 -1 1 Second Soil Canonical Variate 3 Figure 5.3. Plots of second canonical variates for soil and foliar variables respectively, with study stands While this analysis sheds some light on which soil and foliar characteristics have linearly correlated relationships, there still is a lot of overlap among the study stands when identified by S N R or ecosite. Because estimates of soil nutrient availability (SNR) and stand productivity (SI) are important factors worth consideration, and to narrow the focus of the C C A , soil variables were reduced to those previously demonstrating relationships to SNR (section 4.3), and the foliar variables were reduced to those related (a=0.15; backward selection) to site index. The resulting investigated soil and foliar parameters were: (soil) MinNMS, NMS, C M S , C N F F , P F F , and (foliar) NeedWt, FolN, FolP, FolS, FolCu and FolB. Once the number of investigated variables was reduced and the analysis was re-run, the clustering of stands (as identified by SNR or ecosite) graphed by the first canonical variates greatly improved (Figure 5.4). The stands on very poor sites do not have any ellipse overlap with those identified as medium. The soil variables most prominently correlated with this linear combination are tNMS, MinNMS, tC:tNFF and P F F (Table 5.3a). The loading values show that all the investigated foliar variables except FolCu and FolB were well represented in this linear combination (Table 5.3b). Foliar S concentrations had the greatest individual influence upon this first foliar variate. The total correlation between the first canonical variates of soil and foliar parameters was 0.74 with an R 2 value of 0.54. These values are lower than those of the original stratified by (a) S N R and (b) ecosite. Ellipses indicate 0.7 probability interval. 39 analysis, likely due to the reduced number of independent variables and the limited pairing of similar soil and foliar nutrients. While the correlation coefficients were lower, the results were still significant and limited to those variables considered important in terms of site quality and stand productivity. The graphical dispersion of stands as characterized by this limited number of soil and foliar variables (Figure 5.4b) indicates that there may be differences in the uptake of soil nutrients and their incorporation into foliar tissue among ecosites B, D, and H. In general, the plotted dispersion follows what would be expected based upon the nutrient values and S N R distribution of the ecosites. B-sites had many very poor sites, whereas d- and h-sites did not. This segregates these ecosites somewhat and may indicate differences in availability of nutrients f a foliar incorporation. Because the d- and h-sites were similarly distributed within the poor S N R , the fact that they are not directly overlapping indicates that there may be ecosite-specific uptake of those nutrients associated with stand productivity, dependent upon the moisture conditions of the site in addition to the SNR. As the centre of the h-site ellipse is offset to the right and below the d-site in Figure 5.4b, one can interpret the h-sites as having slightly more soil nutrients (associated with fertility) than the d-sites, but slightly fewer foliar nutrients overall. To understand this, the moisture conditions of the respective ecosites should be considered. H-sites have more available soil moisture than do d-sites, and so can be expected to have a greater build-up of organic material (increased moisture reduces decomposition rates) relative to the drier d-sites, resulting in higher levels of organic nutrients (carbon, nitrogen etc.) in the soil. Although there may be more nutrients in the soil on the h-sites, the greater moisture levels restrict the aeration and partially reduce the ability of the trees to assimilate those nutrients and incorporate them into the foliage. 40 Table 5.3. Correlations between variables and their respective canonical variates; a) soil variables, and b) foliar variables. Only the first three canonical variates are shown. The variance is standardized. a) Soil Canonical Variates b) Foliar Canonical Variates Variables SOIL1 SOIL2 SOIL3 Variables FOL1 FOL2 F O L 3 C M S 0.6169 0.4406 0.5488 N E E D W T 0.5807 -0.2498 -0.2616 NMS 0.6733 0.6619 0.3015 F O L N 0.5128 0.3589 -0.7494 C N M S 0.1518 -0.1511 0.5615 F O L P 0.4907 0.1267 -0.6156 MINNMS 0.7063 0.2878 -0.0622 F O L S 0.9537 -0.1407 -0.1721 C N F F -0.583 0.2714 0.6233 F O L C U 0.0719 -0.4701 0.1331 P F F 0.5517 -0.4582 0.2765 F O L B 0.1218 0.7823 0.2623 Variance Cumulative 0.3333 0.3333 0.1703 0.5036 0.1961 0.6997 Variance Cumulative 0.2951 0.2951 0.1767 0.4718 0.1875 0.6593 Figure 5.4. Plots of first canonical variates for soil and foliar variables (reduced), with study stands stratified by (a) S N R and (b) ecosite. Ellipses indicate 0.7 probability level. 5.4 Conclusions There were significant differences in mean fascicle weight and foliar nitrogen, phosphorous, and sulphur concentrations among stands of different soil nutrient regimes, with medium sites possessing the greatest values and very poor sites with the lowest. Almost 40% of 41 the study stands were considered moderately to severely N-deficient. Higher productivity ecosites (d- and e-sites) had greater proportional representation of stands with adequate and slight deficiencies of foliar nitrogen than low productivity sites. Increasing stand densities typically resulted in lower foliar nitrogen concentrations for stands on sites of equivalent quality. Canonical correlation analysis demonstrated the presence of significant linear relationships between soil and foliar variables; therefore, the null hypothesis (1c) is rejected. The soil and foliar properties most strongly correlated to one another in linear combinations were: (soil) tNMS, MinNMS, K M S , C a M S , and NFF; and (foliar) FolN, FolP, FolS, and FolZn. As plant-available soil nutrients increased, so did foliar nutrient concentrations. A.second analysis with a reduced variable data set revealed that as soil nutrients related to SNR increase, so do foliar nutrients related to stand productivity. 42 6.0 RELATIONSHIPS BETWEEN STAND PRODUCTIVITY AND SITE PROPERTIES 6.1 Introduction Forest ecosystems are a complex interaction of biological and physical factors that can combine in multiple dimensions (spatial and temporal) to produce a multitude of conditions for forest growth (Chen et al. 1998b). Growth conditions directly influence the potential productivity of a site for a given tree species. Site index is the most widely accepted indirect estimate of potential productivity for tree species (Mader 1963; Mogren and Dolph 1972; Tesch 1981). It is defined as the height, at a specified base age, of trees that have always been dominant or codominant and healthy (Carmean 1975). In spite of its limitations (Monserad 1984), site index acts as a synthesizer of many site characteristics, combining them into a single species-specific measure of tree growth on a given site. Investigating the relationships between site index and site characteristics can lead to the identification of influential site factors (Corns and Pluth 1984; Beland and Bergeron 1996; Kayahara etal. 1997) and determine their optimal ranges with respect to potential productivity. This information adds to the understanding of the interaction between site and stand factors, and can enable the prediction of potential stand productivity. In this chapter, I will test the relationships between stand productivity, and continuous and categorical measures of site quality, identifying those measures most correlated with site index. 6.2 Results and Discussion 6.2.1 Physical Site Factors Continuous physical site characteristics investigated were: elevation (m), aspect (°), slope (%), thickness of forest floor (cm), coarse fragment content of the soil (%), rooting depth of the stand (cm), depth to water table (cm), depth to gleyed soil horizon (cm). Of all the investigated site factors, only slope and depth to water table showed significant positive correlation with site index. Elevation and aspect did not play significant roles in forest 43 productivity within the areas studied, likely due to the relative gentleness of the terrain. This is comparable to the observations of Duffy (1964) who investigated pine productivity in a location slightly south of the study area and found no correlation between productivity and aspect. All study sites fall within a narrow range of 1,150 to 1,385 metres in elevation. The terrain may not have sufficient relief to influence the mesoclimate enough to counteract variations in soil properties that may confound the effect of aspect. Therefore, aspect does not appear play a significant role in influencing forest productivity within the areas studied as it does elsewhere (Hills 1952). Similar ranges of forest floor thickness, rooting depth, and depth to gleyed soil horizons were found in all study stands and did not appear significantly correlated to stand productivity. The fact that both slope and water table depth are significantly (a=0.05) and positively correlated to site index (Table 6.1) suggests that soil moisture is one of the primary determinants of forest productivity within the study area. Steeper slopes impart good drainage regardless of soil texture, whereas the presence of a water table indicates poor or non-existent drainage, retarding root aeration and development. This positive correlation between slope percent and tree growth confirms observations published in other studies performed in western Alberta (Duffy 1964; Dumanski etal. 1973, Corns and Pluth 1984). Table 6.1. Pearson correlation coefficents and associated probability levels between site characteristics and stand productivity (site index). Significance: (*) p<0.05; (**) p<0.01; (***) p<0.001. Site Index Pearson correlation coefficient P-value ELEVAT ION -0.217 0.096 A S P E C T -0.230 0.129 S L O P E 0.261 0.044* F O R E S T F L O O R 0.102 0.439 C O A R S E F R A G -0.195 0.136 R O O T D E P T H -0.091 0.491 H20 T A B L E 0.372 0.003** G L E Y D E P T H 0.089 0.500 The important role of moisture upon forest productivity can be further confirmed by investigating the effect of slope position (categorical variable). Figure 6.1 illustrates how slope position and consequently water movement affect tree growth. Stand productivity on mid-slopes 44 is significantly different from crest and depression positions (Table 6.2 and Table 6.3). Crests are water-shedding sites and depressions are water-collecting sites. Upper and lower slopes are marginally water-shedding and water-collecting respectively. Mid-slopes and flats are neither water-shedding nor collecting. The free movement and drainage of water appears essential for good forest growth in this region, if there is no deficiency in supply (i.e. crest position). E Slope Position Figure 6.1. Mean site index (m) of study stands in relation to slope position. Error bars indicate one standard error for the mean. Table 6.2. Analysis of variance of mean site index (m) by slope position. Source df Sum-of-Squares Mean-Square F-ratio P Slope Position 5 195.082 39.016 7.015 <0.001 Residual Error 54 300.359 5.562 R 2 = 0.394 S E E = 2.36 45 Table 6.3. Stand productivity by slope position. Mean SI values with dissimilar superscripts are significantly different (Tukey HSD Test; a=0.05). S E is the standard error of the mean. Slope position n Mean SI S E Depression 1 7.1 a 2.358 Crest 11 10.191 a 0.711 Lower 9 12.711 a b 0.786 Flat 10 12.91 a b 0.746 Upper 8 13.35 a b 0.834 Middle 21 14.848 b 0.515 6.2.2 S i te Index a n d C a t e g o r i c a l M e a s u r e s of S i te Qual i ty The relative influence of the categorical variables SNR, S M R , and ecosite upon SI is quantified by the respective R 2 values of their A N O V A models (Table 6.4). Because these four models were all run upon the same data set, the R 2 values can be directly compared. S N R s explained 32% of the variance in site index. S M R s and ecosites were also investigated individually, with R 2 values of 0.63 and 0.54, respectively. In combination, SNR and S M R accounted for 69% of the variation in site index; greater than any other investigated single measure of site quality. For the range of sites examined in this study, S M R was the single synoptic variable that accounted for the greatest amount of variation in site index, corroborating the observations of Dumanski et al. (1973). S N R s accounted for only about half the variation in site index as S M R s did. Because the investigated soil nutrient range only encompasses three out of a possible five S N R s for this particular data set, the ability of this categorical variable to account for stand productivity is somewhat restricted. It is likely that substantially more variation in site index could have been explained by this synoptic variable if the sampling had included the full range of S N R s (very poor through very rich). 46 Table 6.4. Coeffients and R 2 values for various constructed additive G L M models predicting mean site index (m). Abreviations include: (Mi, medium, P, poor, V P , very poor, MD, moderately dry; S D , slightly dry; F, fresh; M 2 , moist; V M , very moist, B, b ecosite, D, d ecosite, E, e ecosite, and H, h ecosite. Root mean square error of the model (RMSE). Significance of model: (*) p<0.05; (**) p<0.01; (***). p<0.001. Factor Intercept Coefficients R2 (RMSE) Significance level S N R 10.34 V P 0.00 P 2.31 Mi 5.14 0.323 (2.42) *** S M R MD SD F . M 2 V M 0.630 *** 9.41 1.02 2.47 5.97 3.40 0.00 (1.83) S N R V P P Mi MD SD F M 2 V M 0.688 *** & S M R 7.84 0.00 1.34 2.75 2.03 2.69 5.61 3.50 0.00 (2.03) Ecosite 11.42 B -0.72 D 2.53 E 4.6 H 0.00 0.539 (2.02) *** SMR Figure 6.2. Trophosequences illustrating the effect of moisture upon site index across soil nutrient regimes. Error bars are one standard error of the mean. S M R (MD, moderately dry; S D , slightly dry; F, fresh; M, moist; V M , very moist); S N R (M, medium, P, poor, V P , very poor). 47 The relationships among soil moisture, soil nutrients, and stand productivity are illustrated graphically with trophosequences of the three investigated S N R s (Figure 6.2). Regardless of the SNR, similar patterns were followed across the S M R s , with the highest SI values expressed at the fresh S M R , where there were no water deficits or excesses during the growing season. As actual soil moisture levels decreased or increased, there were significant reductions in the productivity of the stand. On wetter sites, poor soil aeration due to high water tables can restrict the rooting activity of a stand, and on drier sites, the lack of soil moisture reduces a stands' ability to take up of nutrients, becoming a productivity-limiting factor (Carter and Klinka 1990). The data points displayed on the rear facets of Figure 6.3 illustrate independently how SI is related to S M R s and S N R s . On average, SI values increased as sites improved in nutrient regime. Considering that lodgepole pine is not typically known as a nutrient-requiring species, there were significant improvements in stand productivity as available soil nutrients increased. The mesh surface demonstrates again (as in Figure 6.2) how the most productive stands occur on fresh sites regardless of SNR, with site index declining as moisture levels become either relatively deficient or excessive. Contours found across the bottom facet of Figure 6.3 are graphed separately (Figure 6.4) for easier viewing. The boxed black outline circumscribes the SNR and S M R edatopic cells that were sampled during the course of this study. Isolines located outside of this outline are approximations, calculated by distance weighted least squares (DWLS). Considering that ecosites are generalized combinations of both S N R and S M R , their relatively poor ability to account for variation in site index was unexpected. S M R alone has a greater R 2 value. This may indicate that ecosites, as they are presently characterized, are not strongly associated with lodgepole pine productivity. 48 Figure 6.3. Surface plot of the relationships between soil moisture, soil nutrients, and stand productivity. SI suface is interpolated with the distance weighted least squares smoothing technique (DWLS). Tension = 0.5. 49 VP P M Soil Nutrient Regime Figure 6.4. Contour plot illustrating the relationships among of soil moisture, nutrients, and stand productivity. Contour lines are 1.25 metres apart. Black line circumscribes the S N R and S M R combinations of study stands. Abbreviations include: V P , very poor; P, poor; M, medium; R, rich; VR, very rich; MD, moderately dry; SD, slightly dry; F, fresh; M, moist; V M , very moist. Contours were interpolated with the DWLS technique. Tension = 0.5. 6.3 Conclusions All three investigated categorical components of site quality were significantly correlated to site index; therefore, the null hypothesis (2a) is rejected. Soil moisture regime was determined to be the single measure most strongly correlated with site index, and can be used to estimate stand productivity slightly more reliably than ecosites, accounting for 63% of the observed variation in site index. As availability of soil moisture can be controlled by physical site characteristics such as slope position and gradient, these two properties were also positively and significantly correlated with site index. With other site characteristics remaining equal, site index increased with increasing SNR. Maximum site index for any SNR straddles the fresh S M R and declines as moisture becomes increasingly deficient or excessive. 50 7.0 RELATIONSHIPS BETWEEN STAND DENSITY, AND OTHER STAND AND SITE PROPERTIES 7.1 Introduction All tree species demonstrate changes in stand characteristics as stand density increases. This is a universal observation due to the inherent carrying capacity of a site to produce a certain volume of wood. On a given site with limited moisture and nutrients, as stand density increases, there is less soil moisture and nutrients available per tree and so mortality is incurred due to competition, as described by self-thinning theory (Reinecke 1933). For most tree species, top height is the one stand property that has been considered least influenced by stand density. This is different from many other stand characteristics such as DBH or height to live crown, which are known to dramatically shift with changes in density. The height of dominant trees in a stand at a reference age is a commonly used species-specific index of stand productivity. It is the removal of the confounding effects of density that makes site index such a simple, robust, and widely utilized measure of stand productivity. Lodgepole pine appears to be the exception to this common assumption, with height growth being significantly affected by stand density, even at 'moderate' densities (Smithers 1956, 1961). Of course, site index will be adversely affected by stand densities that are 'extreme' in character regardless of tree species, but the boundary between 'moderate' densities and those that may be considered 'extreme' is not well established. It is believed that height growth repression begins at 10,000 stems per hectare (Farnden 1996), but at what age does this limit apply? How does this boundary compare to the densities commonly observed in naturally established lodgepole pine stands, and does it depend upon site quality? Foresters' interest in the management of high-density lodgepole pine stands has been ongoing for many years. However, over the last decade forest managers in western Canada have grown increasingly concerned over large areas occupied by high-density lodgepole pine stands. As productive, mature stands continue to be harvested, the future wood supply is becoming more 51 dependent upon the eventual wood supply from areas that are today exhibiting stagnant or repressed growth behaviour, primarily due to excessively high initial regeneration densities. Stagnating lodgepole pine stands may reach rotation age and yet still not possess timber of merchantable size. The future wood supply could be compromised if these high-density stands are not treated to improve their growth. The interest in this problem is reflected in the high attendance numbers at conferences devoted to the topics of density management and strategic planning (Bamsey 1998, 1999). Reductions in growth due to excessive densities have been widely noted (Smithers 1961; Johnstone 1981a, 1981b, 1985; Reid 1983) for lodgepole pine stands naturally established after wildfire. Experiments involving various silvicultural treatments of high-density stands have been conducted to determine and quantify the benefits of spacing or thinning (with or without subsequent fertilization) to reduce initial densities (Johnstone 1981a, 1981b). If spaced at a young age, stagnated stands are capable of achieving normal height growth (Mitchell and Goudie 1980), with height, diameter and volume response more pronounced on lower fertility sites (Johnstone 1981a). Height response to thinning is not generally observed except in very young stands (Dahms 1971). The physiological characteristics of lodgepole pine that may make these stands more susceptible to high density regeneration patterns, contributing to the development of repressed/stagnating growth, have been investigated by many researchers, including Bassman and Crane (1983), Keane and Weetman (1987), and Lieffers and Titus (1989). Lieffers and Titus (1989) determined that pine seedlings grown at low density had lower root to shoot ratios than those grown at high densities, indicating less resource allocation to stemwood production at high densities. Also, lodgepole pine seedlings grown at high densities did not express the same amount of differentiation in stem size as seen in white spruce, but did increase differentiation in root size. This apparent inability to differentiate in size may be a characteristic of lodgepole pine that predisposes this species to stagnation. Keane and Weetman (1987) observed that, as stand density increased from low (5,000 sph) to high (100,000 sph), the leaf area of dominant trees decreased to just 6.5% of that found in low-density conditions. In addition, the ratio of leaf area to sapwood area decreased substantially (0.30 to 0.15 m2cm"2) from low to high-density stands. 52 These findings suggest that lodgepole pine stands may remain in stagnated high-density conditions because the trees are unable to access required moisture (even when periodically available) due to severely reduced (relative) amounts of foliage and stemwood hydraulic conductivity. It seems that lodgepole pine is caught in a stagnation loop, with the presence of larger trees necessary to take advantage of periodic moisture (increased leaf area and sapwood area) yet stands appear to be unable to differentiate substantially relative to other species in similar conditions. Stand density and differentiation are characteristics that have been under investigation since the beginning of the 20 t h century (if not earlier) by foresters around the world. The impact of density upon stand characteristics has been an important consideration in terms of forest management of which stand stagnation is but a small species-specific component. The following section briefly discusses some historical stand density investigations and their present state and use in forest management. 7.1.1 Re la t ive D e n s i t i e s , S t a n d D e v e l o p m e n t , a n d St ructura l Di f ferent iat ion Wilson's spacing factor (Wilson 1946; 1979) is measure of relative density of a stand in comparison to its mean height [eq.1]. It allows the comparison of relative densities among stands of different heights to contrast stocking levels and ascertain deficiencies. A commonly accepted optimal range for spacing factor in managed lodgepole pine stands is 17 to 2 1 % (Dave Presslee pers. comm.). Wilson estimated a critical threshold density for the onset of height growth repression to occur at a spacing factor of 11%. Studies by Vyse (1985) later corroborated this value. The stands measured in the present study are predominantly below this threshold level of relative density, with a mean value of 8.59 ± 1.98 (1 standard deviation), indicating that they are all likely experiencing repressed height growth. S P A C F = [SQRT (10,000/stems per hectare) / stand he ightp 00 [eq. 1 ] One tool that can help foresters understand the relationships between stand density, tree size and volumes is the Stand Density Management Diagram (SDMD). SDMD theory is built upon 53 maximum stand yield and density relationships that have been under investigation for many years (Reinecke 1933; Drew and Flewelling 1977). SDMDs are composed of a log-log scale with mean tree volume on the y-axis and stems per hectare on the x-axis, based upon the 3/2 power law theory of self-thinning proposed by Yoda etal. (1963), which suggests that individuals of a population will reach a maximum mean size (biomass) and no larger (for a given site and density). It substantiates the existence of a maximal carrying capacity for a site in terms of biomass that cannot be exceeded. The higher the density, the lower the mean size of individuals within that population. SDMDs for commercial tree species have been calibrated through numerous computer simulations with stand development data from hundreds of permanent sample plots. They can be utilized to investigate and illustrate the development dynamics between density and tree size (dominant height, average diameter and volume) of even-aged single-species stands, including the prediction of respective outcomes from silvicultural activities. The diagrams are species-specific, and when used in combination with the relevant site index curves, they provide the ability to anticipate how even-aged single-species stands should develop and enable the forecasting of timeframes to meet stand objectives (merchantable diameters, specific volumes etc.). There have been SDMDs available for interior lodgepole pine for over 15 years (Flewelling and Drew 1984; McCarter and Long 1986). Most recently, SDMDs are available for both unmanaged (natural) and managed lodgepole pine stands (Farnden 1996). In this context, Drew and Flewelling (1979) have defined relative density (RD) as the ratio of actual stand density to the maximum stand density attainable for a given mean tree volume. This is a different measure from Wilson's spacing factor. As outlined in Nyland (1996), the maximum mean tree volume attainable for a given stand density is illustrated on the S D M D , and occurs at a RD of 1.0 which is represented by the outer edge of the diagram on the upper right-hand side. The current SDMDs in use in B.C. and Alberta (Farnden 1996) for lodgepole pine do not use the maximum size-density line (maximum asymptotic density - Yoda et al. 1963) as the upper right limit to the diagram. Instead, the upper right indicates the mean asymptote for stand survivorship curves. Crown closure is expected to occur at a RD value of approximately 0.15, and 54 competition induced mortality occurs at RDs of greater than 0.55. These stages in stand development can be illustrated on the SDMD as diagonal lines, typically, as a stand develops, it tracks upward on the S D M D and increases ih RD. As the stand approaches a RD value of approximately 0.55, competition-induced mortality begins to occur and stand density decreases. The stand continues to track up and to the left on the diagram, eventually reaching an asymptote parallel to the mean stand survivorship curve. Maximum stand growth rates are predicted to occur just below the 0.6 (RD) level where losses in gross yield due to competition are minimized, and the stand remains in a fully stocked (B-level) state (Langsaeter 1941). Following large-scale disturbance, forest stands typically move through several structural and serai stages as they develop over time. Oliver and Larson (1996) identify four stages: stand initiation, stem exclusion, understory reinitiation, and old-growth. In natural even-aged stands, variability in the size of individuals is considered an important factor in the self-thinning process (Weiner 1985). To further investigate the structural development of the study stands, a new stand measure was created and termed the 'index of stand differentiation' (IHD), as introduced in section 3.2. By indicating the relative difference in height between the smallest (shortest) live suppressed trees in the understory and the largest (tallest) dominants of the canopy, this index was intended to give a simple measure of the structural uniformity of a stand. The greater the value of the index, the greater the relative difference in height between the dominant and suppressed individuals of the stand. Alternately, the lower the value of the index, the more structurally uniform a stand is. 7.1.2 E c o l o g i c a l In f luences o n S t a n d D e n s i t y Both Horton (1953) and Smithers (1956) stated that the density of lodgepole pine stands is controlled by the interaction of many historical factors (wildfire severity, seed source and availability, conditions of germination etc.) but is not influenced by site characteristics. However, subsequent research has noted (Smithers 1961; Duffy 1964) that high-density lodgepole pine stands are often found in locations with soils dominated by an abundance of coarse fragments. In addition, soil moisture levels have been shown to have a strong correlation with productivity (Duffy 1964, Dumanki 1973, section 6.0). These findings imply that site quality is influencing the 55 development of lodgepole pine, and its subsequent susceptibility to repression or stagnation. In addition, Weiner (1985) discovered that fertilization of herbaceous plants promoted an increase in the size variability of the studied population. Recent observations by Farnden (1998) also suggest that site quality can dramatically affect stand development. Through the remeasurement of a spacing trial established near Fish Lake, located 40 kilometres southeast of Prince George, B.C., Farnden found that unspaced, high-density lodgepole pine stands that received N-fertilizer 18 years after wildfire disturbance were 'released' from their state of stagnation, appearing to progress more quickly through development than equivalent control stands. The application of fertilizer may be interpreted as a localized, short-term improvement in site quality, implying that stand density may be highest in stands with the poorest site quality regardless of the historical events affecting regeneration etc., and vice versa. Within this context, the relationships between stand characteristics and density were investigated. Additionally, whether the patterns of these relationships were site-specific was also considered. 7.2 Results and Discussion 7.2.1 Overall Relationships between Stand Density and other Stand Properties Selected stand characteristics were modelled and depicted (Figure 7.1) with the following functions (Table 7.1). Models a) through e) were estimated with the natural logarithmic function; model f was estimated with an exponential function (Table 7.2). These functions were graphed to indicate general trends of the stand data as densities increased. Table 7.1. Formulas for logarithmic and exponential functions. Function Formula Exponential Logarithmic Y = a+b\n(x) Y = a{\-e-"x) 56 Table 7.2. Estimated parameters for quantification of relationships between stand properties and density. Significance of non-linear model: (*) 0.05>p>0.01; (**) 0.01>p>0.001; (***) p<0.001. , . Dependent Estimated Parameters D 2 o ; „ n ; « „ ^ » M o d e l Variable (Y) a b ~ R Significance a) SI 49.1552 -3.7231 0.522 b) DBH 40.1739 -3.1414 0.548 c) 10PRI 25.2302 -1.7033 0.138 d) C R N V O L 7.6904 -0.6910 0.336 e) C R N H T 50.5973 -4.2385 0.250 f) IHD 0.6441 0.0001 0.407 Most fundamental stand properties declined appreciably as stand density increased, with the most substantial reductions occurring with site index, DBH, and crown volume (Figure 7.1). Other stand variables exhibited an amount of scatter around the regression lines, especially at the 'lower' densities (10,000 stems per hectare (sph)). As mean stand properties declined with increasing density, eventually the values flattened out and stabilized. Site index dropped from values of 16-17 m at densities less than 9,000 sph down to less than 11 metres at densities of 35,000 sph. For the same ranges in stand density, DBH values dropped from 12-13 to less than 8 cm, and crown volumes of dominant trees fell from approximately 1.7 to less than 0.4 m 3 . 57 a) 2 2 -j ib.h.) 2 0 -8 50 yrs 16 • Jex (m < 14 -tz 12 -ited Sits 10 -8 -w LU 6 -4 -FT = 0.522 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 Stems per hectare 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 , 5 0 0 0 0 6 0 0 0 0 Stems per hectare 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 Stems per hectare d) % E 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 5 0 0 0 0 6 0 0 0 0 Stems per hectare 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 5 0 0 0 0 6 0 0 0 0 . Stems per hectare 2 0 0 0 0 3 0 0 0 0 4 0 0 0 0 Stems per hectare Figure 7.1. Trends of select stand properties illustrating the patterns in lodgepole pine stand characteristics as they change with increasing densities. Take note of 'threshold' density levels between 25 and 30 thousand stems per hectare. 58 The index of height differentiation (IHD) of the study stands followed an opposite trend (Figure 7.1 f); increasing with density until reaching an upper boundary of approximately 0.7. In 40-year-old stands, the IHD values showed that stands appeared relatively uniform (relative differences in height between the tallest and shortest trees were lowest) at the lowest densities. With increasing density, structural complexity of the stands increased, reaching a maximum value at 25,000 sph. At densities above this level, most stands remained somewhat differentiated, although the lower IHD values of some stands indicated increasing homogeneity of the canopy. This is the opposite trend of what was expected. Lower density stands were expected to have greater levels of differentiation (high IHD values), while excessively dense stands were expected to have low levels of differentiation (low IHD values) implying a stagnant state of development. On a relative scale (IHD), those 'uniform' dense stands did not appear as homogeneous as initially believed. When the absolute differential between dominant and suppressed trees is only 1-2 metres, it can make the stand appear uniform. However, if the stand height is only 4-5 metres, then the relative difference is much greater than on a taller stand with a similar absolute differential. Those high-density stands that do have very low IHD values (uniform crown structure) are likely products of true stagnation, with differentiation occurring at very low rates. The IHD values imply the rate of development occurring in the stand. The differentiation rates can be grouped by stand density. Below 10,000 sph, IHD values were below 0.4, indicating that these stands have already moved through the majority of their self-thinning and differentiation stages, resulting in few truly suppressed trees remaining alive. They have died out due to normal stand development. Stands that were between 10,000 and 20,000 sph can be considered 'intermediate' in terms of differentiation rates. The majority of these stands had IHD values between 0.4 and 0.65 indicating that the stands were at all levels of structural heterogeneity. Some still exhibited very small (relatively) suppressed trees while others appeared on the way to thinning out and achieving a uniform canopy. Stands with densities greater than 20,000 sph had IHD values between 0.55 and 0.7, and possessed very small, suppressed stems (relatively). These stands tended to have very low (non-existent) rates of mortality and differentiation, persisting in a very juvenile stage of development where waves of mortality have 59 not yet occurred. These structural trends illustrate different stages of stand development across similarly aged stands on different sites. The IHD graph was re-plotted with the stands identified by their ecosite to examine how ecosite (as a surrogate measure of site quality) was related to the structural differentiation of the study stands (Figure 7.2). This figure illustrates that there are between-ecosite differences. Assuming similar initial regeneration densities, the highest productivity sites (e-sites) have already developed to a stage where the smallest suppressed trees died off and the shortest trees are of the intermediate or codominant canopy classes. B and d-sites have IHD values that are scattered over the upper half of the graph. Although d-sites did not have densities as high as those of the b-sites, the poor nutrient status of these soils appears to restrict the rate of stand development (low mortality rates of suppressed trees) in those stands. H-sites had lower IHD values than did b-sites in similarly dense stands, indicating that suppressed trees are relatively taller on these moist sites. At lower densities, the h-sites follow the trend of the e-sites, becoming increasingly uniform. The greatest structural differentiation of the h-sites can be found at approximately 25,000 sph. 0 100002000030000400005000060000 PI Density (stems per hectare) Figure 7.2. Scatterplot illustrating ecosite-specific relationships between the index of height differentiation and stand density. Linear regression lines for each ecosite are displayed to illustrate ecosite-specific IHD trends. 60 7.2.2 S t a g n a t i o n T h r e s h o l d s Increasing stand density was related to declining values of several major stand properties. At a certain density level, these properties began to level off, with some variability around a mean value. This levelling off suggests the idea of a 'threshold density', beyond which stand properties do not continue to decline. It appears that this threshold may occur between 25.and 30 thousand sph for the study stands (Figure 7.1). Bassman and Crane (1983) suggest that a threshold density inducing repressed growth occurs at approximately 50,000 stems per hectare. They do not state at what age this density effect impacts stand growth, nor do they clearly define the term 'repression'. From the context of their text, they imply a state of stagnation, although they use the two terms interchangeably. Approximate values for properties identifying stagnant stands are presented in Table 7.3. Apparent reductions in stand productivity and growth appear to occur almost immediately with increases in stand density, even at 'low' levels (9000 sph). This suggests that site index values for lodgepole pine may be strongly influenced by density in all naturally established stands. Table 7.3. Characteristics of dominant trees that may identify 40-year-old 'stagnant' stands Variable SI DBH 10 Rl C R N V O L C R N V O L H T Stagnant Stand < 10 m < 6 cm < 8 mm < 0.5 m 3 < 6 7.2.3 S t a n d D e n s i t y M a n a g e m e n t D i a g r a m Study stands were plotted on the SDMD and indicated with different symbols to identify each stand by its ecosite (Figure 7.3). Because actual volume per tree was not known, the stands were located on the S D M D using the height curves and density. As was expected, most of these high-density stands are located at or above the lower line of ICM. Also as expected, the e-site stands (most productive) are found just under the mean ICM line near the middle of the diagram and are the furthest to the upper left of the study stands, indicating that they are the furthest along in stand development. As one moves down and to the right (higher densities), the e-sites overlap with many d-sites. At the farthest lower right corner of the diagram, there is a mix of primarily b and h sites. Stands from these two sites occupy a similar range on the SDMD, with some moving 61 further to the left at lower densities. Overall, as the stands track from higher to lower densities, they transition from b and h sites to d-sites and finally e-sites. One observation regarding the SDMD is that stands from lower productivity ecosites (b and h) are generally found below (shorter heights) those of higher productivity (d and e) at equivalent densities. Conventional wisdom teaches that stands of similar regeneration densities should track through the SDMD identically, with the only differences being the rate at which self-thinning occurs (dependent upon stand productivity). There could be several explanations for the observed pattern. Perhaps the lower productivity stands established at lower initial densities and are still tracking upwards, not having achieved the greater heights yet (they are not yet experiencing significant ICM). Alternatively, these lower productivity sites (b and h) may track towards the upper left at a slightly lower (parallel?) asymptote along the SDMD than do the d and e sites. Since most stands have very high densities at which height repression is likely occurring, it is possible that the degree of height repression is site-specific. The site may be influencing the degree of height repression exhibited by a stand. If this is the case, it might demonstrate that the more productive sites have a greater inherent carrying capacity, accommodating stands of a greater height than those of lower productivity sites at equivalent (high) densities. Logically, with more available nutrients, richer sites should be able sustain higher levels of biomass production If the zone of ICM is split in half, producing upper and lower portions, the discrepancy in positioning between the lower and higher productivity ecosites on the SDMD can be more closely observed. Of the 30 d- and e-site plots, only four fall into the lower portion of the zone of ICM. In the same way, only two of the 30 b- and h-site plots are located in the upper portion of the zone of ICM. This may be indicating that stands on lower productivity sites may never actually achieve 'full' stocking as defined by relative density (RD=1.0). As briefly suggested earlier, it is possible that these 'lower' stands may just not have reached relative densities near 1.0 because of their slow growth, but if the TASS-predicted mortality curves are to be accepted as reliable, then it is unlikely that these stands will ever achieve much higher RD values. This interesting trend should be followed up with actual volume per tree data, allowing more precision when plotting the location of the study plots on the SDMD. 62 3 d LODGEPOLE PINE STAND DENSITY Js>*L oo Cffl. - : : : ' :':^-->;' ui s 3 -J o > ui ui a MANAGEMENT DIAGRAM (NATURAL STANDS) - special edition -6 m Quadratic Mean Diameter - . \ Top Height ... TASS-Predicted Mortality Curves ^ '. Mean and Lower Limits of the Zone of Imminent Competition-Mortality A^1 \ A * 3 \ % - b ecosite # - d ecosite @ - e ecosite X - h ecosite % \ # \ 15^ * j 400 700 1000 2000 4000 7000 TREES PER HECTARE 10.000 20,000 40,000 60,000 Figure 7.3. Stand density management diagram for natural lodgepole pine stands. Adapted from original (Canadian Forest Service, Pacific and Yukon Region) by C. Farnden and D. Brisco. Data Source: T A S S generated managed stand yield tables contained in the computer program WINTIPSY 3.0 (B.C. Ministry of Forests, Forest Productivity and Decision Support Section) 63 7.2.4 S t a n d D e n s i t y a n d C a t e g o r i c a l M e a s u r e s of S i te Qual i ty The relative influence of the<investigated categorical variables upon stand density can be seen by their respective R 2 values (Table 7.4). All three categorical measures of site quality demonstrated significant correlation with stand density. The R 2 values indicated that both S N R and S M R accounted for similar amounts (-34%) of the variation in stems per hectare of the study stands. In combination, SNR and SMR factors accounted for 44% of the variation in stand density. As mentioned in section 3.3.4, the unbalanced nature of the data set precludes the statistical testing of interactions between S N R and S M R , but graphical analysis of the data gave no indication that these interactions would be significant (Figure 7.4) and so they were not included in the A N O V A model. Ecosites 'explained' slightly less (-31%) of the variation in stand density than either SNR or SMR alone within the range of sites sampled. Although this difference is unlikely to be significantly lower than the 34% of variation accounted for by SNR or S M R , it is unexpected since ecosites are a combination of both SMR and SNR classes. One would expect this two-factor measure to account for more variation in stand density than SNR or S M R individually, much like the combined SNR and SMR model. This result implies that ecosites may not be a fine enough resolution to facilitate the determination of suitable silvicultural interventions for density management. Additional site or stand characteristics (age, density, site index etc.) may have to be considered before making such decisions. Considering the many random factors that can contribute to initial regeneration densities (intensity and severity of stand destroying fire, available seed source and viability, germination conditions, etc.), the correlation between various measures of site quality and present stand density is higher than might otherwise be expected. One can imagine that if the initial stocking densities had been controlled to similar levels across the four ecosites, the relationships between stand density and site quality would have been stronger. 64 Table 7.4. Coeffients and R 2 values for various constructed additive A N O V A models predicting mean site index (m). Abreviations include: (Mi, medium, P, poor, V P , very poor, MD, moderately dry; S D , slightly dry; F, fresh; M2, moist; V M , very moist, B, b ecosite, D, d ecosite, E, e ecosite, and H, h ecosite. Root mean square error of the model (RMSE). Significance of model: (*) p<0.05; (**) p<0.01; (***) p<0.001. Factor Intercept Coefficients R2 (RMSE) Significance level S N R 32044 V P 0.0 P -12511 Mi -20898 0.345 (9057) *** S M R 22800 MD 6700 SD -600 F -9784 M 2 -3160 V M 0.0 0.336 (9279) *** S N R & S M R 32885 V P 0.0 P -9201 Mi -14501 MD 449 SD F M 2 V M -2133 -8575 -3513 0.0 0.436 (8712) *** Ecosite 21093 B 6213 D -2720 E -10613 H 0.0 0.307 (9399) *** 60000 % 50000 -1—* o CD 4 0 0 0 0 h CD Q_ CO § 30000 -*—< CO ~ 20000 CO c CD a , 10000 0 S N R M D S D F M V M Soi l Mo is tu re R e g i m e M P V P Figure 7.4. Trophosequences illustrating the effect of moisture upon stand density across soil nutrient regimes. Error bars are one standard error of the mean. S M R (MD, moderately dry; S D , slightly dry; F, fresh; M, moist; V M , very moist); S N R (M, medium, P, poor, V P , very poor). 65 7.2.5 E c o s i t e - S p e c i f i c R e l a t i o n s h i p s be tween S t a n d D e n s i t y a n d P r o d u c t i v i t y Site index of the study stands ranged from almost 20 metres in lower density stands, to less than 7 metres at the highest densities (Figure 7.5). This graph is the same as seen in Figure 7.1a, except that the data points are identified by ecosite. Individual simple linear regression lines are provided to illustrate ecosite-specific relationships between site index and stand density, and demonstrate the different ranges in site index and density expressed by each ecosite. E-site data points are located primarily in the upper left corner of the graph (lowest densities and the highest site index values), while b and h-sites are in the lower half (lowest site index values and highest stand densities), and d-sites are mainly in the central portion of the graph. This figure gives an indication that the declining trend in site index may not be the same for each of the four study ecosites. These linear regression lines were tested for similarity in slope and/or level to determine which ecosites shared similarly patterned relationships, and which did not. H E C O 10000 20000 30000 40000 50000 60000 PI Density (stems per hectare) Figure 7.5. Ecosite-specific simple linear regression relationships between estimated site index and stand density. 66 E-sites were the most productive, followed by d-sites, with b and h-site having similar ranges in site index. E-sites appeared to have the steepest slope (negative) with increasing density, followed by d-sites, with b and h-sites appearing almost identical in both slope and elevation. Multiple contrasts determined that no two ecosites were different in slope, but there were significant differences in regression line elevation. B and h-sites were the same, and d and e-sites were the same (a=0.05)(Table 7.5). The full output from these tests are in Appendix 3. Table 7.5. Summary table of site index slope and elevation comparisons between ecosites. (*) indicates that the comparison found significant differences between the ecosites. Ecosite Comparison p-value for significantly different slopes (a=0.004) p-value for significantly different levels (a=0.004) B vs D 0.567 0.002* B vs E 0.195 0.001* B vs H 0.369 0.859 D vs E 0.163 0.262 D vs H 0.686 0.0001* E vs H 0.277 0.0002* By grouping the ecosites with similar regression line elevations (DE and BH), additional contrast testing detected significant (cx=0.055) differences in regression line slope between the groups (Table 7.6). Each ecosite grouping followed a different pattern in productivity reduction with increasing stand density (increases in stems per hectare did not reduce the site index of the two groups to the same degree). The site index of the DE-site group dropped more rapidly than the BH-site group, although at higher densities (within the ranges sampled), the levels of productivity for the two groups began to converge. To determine why there might be differences in slope and/or elevation between these grouped ecosites, ecological characteristics of the sites should be considered. Both d and e-sites have a S M R of fresh (Chapter 4), while b-sites are typically moisture deficient during the growing season, and h-sites experience moisture surpluses. The primary difference between the d- and e-sites is that e-sites are richer (medium SNR) than d-sites (poor SNR). Neither the b- or h-sites are optimal in terms of moisture availability for lodgepole pine growth. This testing demonstrates that an excess or deficiency in soil moisture availability during the growing season can dramatically 67 reduce the productivity and development of lodgepole pine stands, and that this reduction follows the same linear pattern. Nigh (1997) determined that moderately dense lodgepole pine stands in B.C. growing on dry or wet sites exhibit similar patterns of height growth, justifying the use of the same function to model height growth on these different sites. In this context, the apparent similarity in growth patterns on dry and wet sites suggests that the pattern of site-specific height growth repression identified in my study may be applicable to the entire lodgepole pine population, and is not a phenomenon of the study area. Table 7.6. Statistics and significance of dummy regression coefficients with slope interaction term for grouped B+H and D+E ecosite comparison. Effect Coefficient Std Error t P(2 Tail) C O N S T A N T 14.345 0.5880 24.397 <0.001 LIVEPL -0.0013 0.00002 -5.946 <0.001 E C O 3.7205 0.8388 4.436 <0.001 L I V E P L * E C O -0.00009 0.00004 -1.972 0.054 n=59 Adjusted R 2 : 0. 7461 P-value for total model: Standard error of estimate: 1.409 p < 0.0001 Since the DE-site group has a significantly (a=0.055) steeper slope that the BH-site group, stands on d- and e-sites achieve greater levels of productivity at lower densities than those found on b- or h-sites. The steepness of the site index decline with increasing density on the d- and e-sites implies that productivity on these sites is less limited by site factors such as soil moisture. Without these site factors limiting lodgepole pine growth, near-optimal levels of site index are achieved at lower densities, and true potential productivity levels unaltered by the effects of stand density are more apparent. From the illustrated relationships with density (Figure 7.6), the potential productivity levels attainable with proper density management can be estimated from the regression lines by noting their intercept values. Although the increases in productivity will likely not remain linear at densities below 3 to 4 thousand sph (probably the lower limit of density-induced height growth repression), for d- and e-sites, the regression intercepts the y-axis at 18 m (±1.7 m). This could be a potential maximum achievable productivity level, although unlikely. The true attainable 68 productivity levels (average) would be somewhere between this intercept (18 m) and the highest levels indicated by the regression line (17 m). In the same way, b- and h-sites have an intercept value of approximately 14.3 m (±1.2 m). If the repression of height growth due to density was removed (i.e. initial stand densities were reduced to levels that did not influence height growth), it is likely that most of the study stands could achieve site index values just slightly below this value. Since e-sites have more available nutrients for growth than do d-sites (all other site factors being equal), the e-sites will likely achieve a site index closer to 18 m, while the d-sites should achieve a site index closer to 17 m. Both b and h-sites have similarly low levels of available nutrients and have additional soil moisture constraints, and so their potentially attainable site index values will be similar, and substantially less than that of the D+E grouping. 20, Ui \ CO CD o LO X CD "O _c Q) bo 2 fo E 0 ) LU O i i i I I -X X ^ l x x x - X x X x*-V X X X X X I I I I I EC01 o D+E x B+H 10000 20000 30000 40000 50000 60000 PI Density (stems per hectare) Figure 7.6. Scattergram illustrating the relationship between site index and stems per hectare for significantly different (a=0.055) grouped ecosites (D+E and B+H). 7.3 Conclusions At all observed densities (> 7,000 sph), the development of 40-year-old lodgepole pine stands was considered repressed, and the degree of repression increased as density increased. Many stand properties decreased with increasing stand density. The most significant of these were: site index, average DBH, and average crown volume. Stand densities between 25 and 30 69 thousand stems per hectare seemed to mark a threshold; beyond which stand characteristics did not vary substantially despite continued increasing density (up to 55,000 stems per hectare). Investigated categorical measures of site quality accounted for similar and significant amounts of variation in stand density; therefore, the null hypothesis (2b) is rejected. The effects of density upon stand development and productivity were found to be site-specific; therefore, the null hypothesis (2c) is rejected. The locations of study stands on the S D M D implied that higher productivity stands track differently (slightly higher) through development in comparison to those of lower productivity. In addition, it was determined that d-and e-sites followed a different linear pattern of producti\^trepression in response to increasing stand density than did b- and h-sites. 70 8.0 G E N E R A L CONCLUSIONS This study was initiated to answer some basic questions regarding the ecological relationships in lodgepole pine stands in west-central Alberta, particularly those between measures of site quality, and stand productivity and density. At a fundamental level, this study attempted to determine the usefulness of ecological measures of site quality as tools to assist lodgepole pine management. Field-estimated soil nutrient and moisture regimes were quantitatively characterized. Nitrogen-related soil properties discriminated between field-identified S N R s , justifying their use as categorical measures of soil nutrient availability. The actual soil moisture regime of zonal sites within the Upper Foothills subregion was determined to be fresh, having neither moisture deficiencies nor excesses during the growing season. Strong linear relationships were detected between soil and foliar nutrients. All investigated categorical measures of site quality were significantly correlated to both stand productivity and density. Soil moisture regimes accounted for the largest amount of variation (63%) in site index, whereas soil nutrient regimes accounted for the greatest (35%) variation in stand density and thereby, the rate of stand development. Even though they are generalized combinations of both soil moisture and nutrient regimes, ecosites poorly accounted for variation in both site index and stand density. Soil moisture strongly influenced stand productivity, was significantly correlated with stand development and density, and appeared to influence the degree of height repression. Differences in available soil moisture appeared to explain the majority of observed differences in lodgepole pine productivity. Assuming the initial regeneration densities of the study stands were similar, this investigation demonstrated the importance of site quality upon the rate of stand development, mortality, and stand productivity. The index of height differentiation was explored as a rudimentary measure that could indicate the structural uniformity of a young stand. Patterns of vertical structure across ecosites revealed that there were differences in index values between ecosites, and that more productive, low-density stands were relatively more uniform in canopy structure than were less productive, 71 high-density stands. A density between 25 and 30 thousand stems per hectare marked a threshold; beyond which study stand characteristics did not vary substantially despite continued increasing density (up to 55,000 stems per hectare). In summary, the relationships between site and stand factors are important for a better understanding of stand development. As these relationships are quantified, they allow forest stands to be manipulated for the production of desired products. This is significant as high-density immature lodgepole pine stands make up such a large component of forested land in both Alberta and B.C. 72 9.0 LITERATURE CITED Alexander, R.R. and C.B. Edminster. 1981. Management of Lodgepole Pine in Even-Aged Stands in the Central Rocky Mountains. U.S.D.A. Forest Service Research Paper RM-229, 10 p. Rocky Mountain Forest and Range Experiment Station, Ft. Collins, C O . Anonymous 2000a. Establishment to Free Growing Guidebook, Prince George Forest Region. Forest Practices Code. B.C. Ministry of Forests. Revised Edition May 2000. Available: http://www.for.gov.bc.ca/tasb/legsregs/fpc/FPCGUIDE/Guidetoc.htm (Accessed: February 13,2001). Anonymous. 2000b. Table Interpolation Program for Stand Yields (WinTIPSY - for Windows™). Available: http://www.for.gov.bc.ca/research/gymodels/TIPSY (Accessed: January 7, 2001). Anonymous 1999. British Columbia Ministry of Forests Annual Report 1997/98. Ministry of Forests, Victoria, B.C. 11 Op. Anonymous. 1998. Tree And Stand Simulator software. Information can be found at the following website: http://www.for.gov.bc.ca/research/gymodelsArASS - accessed on January 7, 2001. Anonymous 1996. The Status of Alberta's Timber Supply. Canadian Council of Forest Ministers Timber Supply Conference, November 16-18, 1994, Kananaskis, Alberta. Revised for publication on the Alberta government web site May 1996. Available: http://wwwgov.ab.ca/env/forests/fmd/timber/timbersupply.htm (Accessed: January 15, 2001). Armson, K.A. 1977. Forest soils: properties and processes. Univ. Toronto Press, Toronto. 390 p. Bakuzis, E.V. 1969. Forestry viewed in an ecosystem perspective, p. 189-258 in G .M. Van Dyne (ed.), The ecosystem concept in natural resource management. Academic Press, New York. Ballard, T. M. and R. E. Carter. 1986. Evaluating forest stand nutrient status. Land Management Handbook No. 20. British Columbia Ministry of Forests. Victoria, B.C. 60 p. Bamsey, C . 1999. (ed.) Stand density management: using the planning tools. Conference proceedings, Nov 23 and 24, 1998. Clear Lake Ltd. Edmonton, A B . Bamsey, C. 1998. (ed.) Stand density management: planning and implementation. Conference proceedings, Nov 6 and 7, 1997. Clear Lake Ltd. Edmonton, A B . Bassman, J .H . and J.L. Crane Jr. 1983 Growth Repression in Lodgepole Pine: An Ecophysiological Characterization. Northern Silviculture Committee Meeting. Williams Lake, BC . July 27-28, 1983. Beaton, J .D. , A. Moss, I. MacRae, J.W. Konkin, W.P.T. McGhee, and N. Kosick. 1965. Observations on foliage nutrient content of several coniferous tree species in British Columbia. For. Chron. 41(2): 222-236. Beckingham, J . D., I. G . W. Corns, J . H. and Archibald. 1996. Field Guide to Ecosites of West-Central Alberta. Special Report 9. Canadian Forest Service, Northwest Region. Edmonton, AB . 73 Beland, M. and Y Bergeron. 1996. Height growth of jack pine (Pinus banksiana) in relation to site types in boreal forests of Abitibi, Quebec. Canadian Journal of Forest Research 26: 2170-2179. Brockley, R.P. 2001. Foliar sampling guidelines and nutrient interpretive criteria for lodgepole pine. B.C. Min. For. Victoria. Extension Note (in press). Brockley, R.P. 2000. Using Foliar Nutrient Levels to Predict Lodgepole Pine Fertilization Response. B.C. Min. For. Victoria. Extension Note 44. 7 p. Brockley, R.P. 1996. Lodgepole Pine Nutrition and Fertilization: a Summary of B.C. Ministry of Forest Research Results. B.C. Ministry of Forests and Forestry Canada, Victoria, B.C. FRDA Rep. 266. Carmean, W.H. 1975. Forest site quality evaluation in the United States. Adv. Agron. 27: 209-267. Carter, R.E., and K. Klinka. 1990. Relationships between growing season water deficit, mineralizable soil nitrogen, and site index of coastal Douglas-fir. For. Ecol. Manage. 30: 301-311. Carter, R.E., E .R.G. McWilliams, and K. Klinka. 1998. Predicting response of coastal Douglas-fir to fertilizer treatments. For. Ecol. Manage. 107: 275-289. Chen, H.Y.H., K. Klinka, J . Fons, and P.V. Krestov. 1998a. Characterization of nutrient regimes in some continental subalpine boreal forest soils. Can. J. Soil Sci. 78: 467-475. Chen, H.Y.H., K. Klinka, and R.D. Kabzems. 1998b. Site index, site quality, and foliar nutrients of trembling aspen: relationships and predictions. Can. J. For. Res. 28: 1743-1755. Corns, I.G.W. 1992. Forest site classification in Alberta: its evolution and present status. For. Chron. 68: 85-93. Corns, I.G.W. 1983. Forest community types of west-central Alberta in relation to selected environmental factors. Can. J. For. Res. 13: 995-1010. Corns, I.G.W. and R.M. Annas. 1986. Field guide to forest ecosystems of west-central Alberta. Northern Forest Centre. Canadian Forest Service. Edmonton, AB . 251 p Corns, I.G.W. and D.J. Pluth. 1984. Vegetational indicators as independent variables in forest growth prediction in west-central Alberta, Canada. For. Ecol. Manage., 9: 13-25. Courtin, P.J . , K. Klinka, M.C. Feller, and J .P . Demaerschalk. 1988. An approach to quantitative classification of nutrient regimes of forest soils. Can. J. Bot. 66:2640-2653 Dahms, W . G . 1971. Growth response in lodgepole pine following pre-commercial thinning, p. 14-18. in D.M. Baumgartner (ed.) Proceedings of: Pre-commercial thinning of coastal and intermountain forests in the Pacific North West Symposium. Co-op extension. Serv., Washington State Univ., Pullman, WA. (as cited in Reid 1983). DeLong, C , A. MacKinnon, and L. Jang. 1990. Field Guide for site identification and interpretation of ecosystems of the northeast portion of Prince George Forest Region. Land Management Handbook No. 22. British Columbia Ministry of Forests, Victoria, B.C. Downing, D. 1999. W E L D W O O D Forest Management Area Biophysical Overview Project: Methods and Results - Summary Report. Timberline Forest Inventory Consultants. 32 p. Drew, T .J . and J.W. Flewelling 1979. Stand Density management: an Alternative Approach and Its Application to Douglas-fir Plantations. For. Sci. 25(3): 518-532. 74 Drew, T .J . and J.W. Flewelling 1977. Some Recent Japanese Theories of Yield-Density Relationships and Their Application to Monterey Pine Plantations. For. Sci. 23(4): 517-534. Dumanski, J . , J . C . Wright and J.D. Lindsay. 1973. Evaluating the productivity of pine forests in the Hinton-Edson are, Alberta, from soil survey maps. Can. J. Soil Sci. 53: 405-419. Dumanski, J . , T .M. Macyk, C F . Veauvy, and J.D. Lindsay. 1972 Soil Survey and Land Evaluation of the Hinton-Edson Area, AB. Alberta Institute of Pedology. Report No. S-72-31. 119 p. Duffy, P.J.B. 1964. Relationships between site factors and growth of lodgepole pine (Pinus contorta Dougl. var. latifolia Engelm.) in the Foothills section of Alberta. Can Dep. For. Publ. 1065. 60 p. Farnden, C. 1998. Treating density-related height suppression in lodgepole pine: 19-year results from thinning and fertilization trials near Prince George, BC. Contract report for Weldwood of Canada Ltd. , Hinton Division. Prince George, BC. 37 p. Farnden, C. 1996. Stand density management diagrams for lodgepole pine, white spruce and interior Douglas-fir. Information Report BC-X-360. Pacific Forestry Centre. Canadian Forest Service. 37 p. Farrar, J . L. 1995. Trees in Canada. Canadian Forest Service and Canada Communication Group. Ottawa. 502 p. Flewelling, J .W. and T .J . Drew. 1985. A Stand Density Management Diagram for Lodgepole Pine. In Lodgepole Pine: the species and its management. (Eds: D.M. Baumgartner, R.G Krebill, J.T. Arnott and G.F. Weetman.) Symposium Proceedings. Washington State University. Pullman, p. 239-244. Freese, F. 1964. Linear Regression Methods for Forest Research. U.S. Forest Service Research Paper, FPL #17. Forest Products Laboratory. U.S. Dept. Agriculture. 137 p. Green, R.N. and K. Klinka. 1994. A Field Guide for Site Identification and Interpretation for the Vancouver Forest Region. Land Management Handbook No. 28. B.C. Ministry of Forests, Victoria, B.C. Green, R.N., R.L Trowbridge, and K. Klinka. 1993. Towards ataxonomic classification of humus forms. For. Sci. Monogr. 29:1-48. Green, R.N. P.J . Courtin, K. Klinka, R.J. Slaco, and C.A. Ray. 1984. Site diagnosis, tree species selection, and slashburning guidelines for the Vancouver Forest Region. Land Management Handbook No. 8. B.C. Ministry of Forests, Victoria, B.C. Hicks, C R . 1993. Fundamental Concepts in the Design of Experiments, Fourth Edition. Saunders College Publishing. Toronto. 509 p. Hills, G.A. 1952. The classification and evaluation of site for forestry. Ontario Dept. Lands and Forests. Research Div. Research Report No. 24. 41 p. Hills, G.A. and G. Pierpoint. 1960. Forest Site Evaluation in Ontario. Ontario Department of Lands and Forests. Research Branch. Research Report No. 42. 64 p. Horton, K.W. 1953. Variation in the stocking of lodgepole pine regeneration following fire. Canada Dept. Northern Affaires and National Resources, Forestry Branch, For. Res. Div., Silv. Leaft. No. 95. (as cited in Smithers 1956) Huang, S., S . J . Titus, and G. Klappstein. 1997. Subregion-based compatible height and site index models for young and mature stands in Alberta: revisions and summaries (Part I). Forest 75 Management Research Note No. 9. Forest Management Division, Land and Forest Services, Alberta Environmental Protection. Pub. No. T/389. Johnstone, W.D. 1985. Thinning lodgepole pine. p. 253-262. in Baumgartner, D.M., R.G. Krebill, R.G. Arnott and G.F. Weetman (eds.) Proceedings of: Lodgepole Pine: the species and its management. Washington State Univ. Cooperative Extension, Pullman. Johnstone, W.D. 1981a. Effects of spacing 7-year-old lodgepole pine in west-central Alberta. Environ. Can. , Can. For. Serv., North. For. Res. Cent. Edmonton, AB . Inf. Rep NOR-X-236. Johnstone, W.D. 1981b. Precommercial thinning speeds growth and development of lodgepole pine: 25-year results. Environ. Can., Can. For. Serv., North. For. Res. Cent. Edmonton, A B . Inf. Rep NOR-X- 237. Kabzems, R.D. and K. Klinka. 1987a. Initial quantitative characterization of soil nutrient regimes. I. Soil properties. Can. J. For. Res. 17: 1557-1564. Kabzems, R.D. and K. Klinka. 1987b. Initial quantitative characterization of soil nutrient regimes. II. Relationships among soils, vegetation and site index. Can. J. For. Res. 17: 1565-1571. Kayahara, G .J . , K. K Klinka, and A .C . Schroff. 1997. The Relationship of Site Index to Synoptic Estimates of Soil Moisture and Nutrients for Western Redcedar (Thuja plicata) in Southern Coastal British Columbia. Northwest Sci. 71(3): 167-173. Keane, M.G. and G.F. Weetman. 1987. Leaf area - sapwood cross-sectional area relationships in repressed stands of lodgepole pine. Can. J. For. Res. 17: 205-209. Klinka, K., J . Worrall, L. Skoda, and P. Varga. 2000. The Distribution and Synopsis of Ecological and Silvical Characteristics of Tree Species of British Columbia's Forests. Canadian Cartographies Ltd., Coquitlam, B.C. 180 p. Klinka, K., Q. Wang, R.E. Carter, and H.Y.H. Chen. 1996. Height growth-elevation relationships in subalpine forests of interior British Columbia. For. Chron. 72:193-198 Klinka, K., Q. Wang, and G .J . Kayahara, 1994. Quantitative characterization of nutrient regimes in some boreal forest soils. Can. J. Soil Sci. 74: 29-38. Klinka, K., J . Pojar, and D.V. Meidinger. 1991. Revision of Biogeoclimatic Units of Coastal British Columbia. Northwest Sci. 65(1):32-47. Klinka, K. and R.E. Carter. 1990. Relationships between site index and synoptic environmental factors in immature coastal Douglas-fir stands. For. Sci. 36: 815-830. Klinka, K., V . J . Krajina, A. Ceska, and A .M. Scagel. 1989. Indicator Plants of Coastal British Columbia. UBC Press. Vancouver, B.C. 288 p. Klinka, K. and V . J . Krajina. 1986. Ecosystems of the University of British Columbia Research Forest, Haney, B.C. The University of B.C., Faculty of Forestry. 123 p. Krajina, V . J . 1960. Can we find a common platform for the different schools of forest type classification. In Symposium on Forest Types and Forest Ecosystems. The IX International Botanical Congress. Montreal. Silva fennica 105: 50-55. Krajina, V . J . 1969. Ecology of forest trees ih British Columbia. Ecol. West. N. Am. 2: 1-146. Landon, J .R. 1988. Towards a standard field assessment of soil texture for mineral soils. Soil Surv. Land Eval. 8: 161-165. Langsaeter, A. 1941. Om tynning I enaldret gran - og furskog. Meddel f d Norske Skogforsoksven 76 8: 131-216. (as cited in Drew and Flewelling 1979) La Roi, G.H. , W.L. Strong, and D.J. Pluth. 1988. Understory plnat community classifications as predictors of forest site quality for lodgepole pin and white spruce in west-central Alberta. Can. J. For. Res. 18: 875-887 La Roi, G .H . and R.J. Hnatiuk. 1980. The pinus contorta forests of Banff and Jasper national parks: a study in comparative synecology and syntaxonomy. Ecol. Mono. 50(1): 1-29. Leiffers V . J . and S .J . Titus. 1989. The effects of stem density and nutrient status on size inequality and resource allocation in lodgepole pine and white spruce seedlings. Can J. Bot. 67: 2900-2903. Lloyd, D., K. Angove, G . Hope, and C. Thompson. 1990. A Guide for Site Identification and Interpretation of the Kamloops Forest Region. 2 vol. Land Management Handbook No. 23. B.C. Ministry of Forests, Victoria, B.C. Lotan, J .E . and W.B. Critchfield. 1990. Pinus contorta Dougl. ex. Loud. Lodgepole pine, p.302-315. in R. M. Burns, and B.H. Honkala (eds.) Silvics of North America: Vol. 1. Conifers. USDA Forest Service, Agriculture Handbook 654, Washington, D . C , USA. 675 p. Lotan, J .E . , J.K. Brown, and L.F. Neuenschwander. 1985. Role of Fire in Lodgepole Pine Forest, p. 133-152. in Baumgartner, D.M., R.G. Krebill, R.G. Arnott and G.F. Weetman (eds.) Proceedings of: Lodgepole Pine: the species and its management. Washington State Univ. Cooperative Extension, Pullman. Lotan, J .E . and D.A. Perry. 1983. Ecology and Regeneration of Lodgepole pine. USDA For. Serv. Agriculture Handbook No. 606. 51 p. Loucks, O.L. 1962. Ordinating forest communities by means of environmental scalars and phytosociological indices. Ecol. Monogr. 32: 137-166. Mader, D.L. 1963. Volume growth measurement - an analysis of function and characteristics in site evaluation. J. For. 61:193-198. Major, J . 1963. A climatic index to vascular plant activity. Ecology. 32:485-498. Marshall, P.L. 1990. The Issue of Maximum Density in Lodgepole Pine: A Critique of the Maximum Density in Lodgepole Pine Submissions. B.C. Ministry of Forests, Silviculture Branch, Minor Services Contract Report No. 42850. 46 p. Marshall, P., T. Szikszai, V. LeMay, and A. Kozak. 1995. Testing the distributional assumptions of least squares linear regression. For. Chron. 71(2): 213-218. McCarter, J .B. and J .N . Long. 1986. A lodgepole pine density management diagram. W. J. Appl. For. 1(1 ):6-11. Meidinger, D. and J . Pojar (Editors). 1991. Ecosystems of British Columbia. Special Report Series No. 6. British Columbia Ministry of Forests, Victoria, B.C. 330 p. Miller, H.G., J .M. Cooper, J.D. Miller and O.J.L. Pauline. 1978. Nutrient cycles in pine and their adaptation to poor soils. Can. J. For. Res. 9: 19-26. Mitchell, K.J. and J.W. Goudie. 1980. Stagnant Lodgepole Pine. Progress Report. E.P.850.02. British Columbia Ministry of Forests. Mogren, E.W. and K.P.Dolph. 1972. Prediction of Site Index of Lodgepole Pine from Selected Environmental Factors. For. Sci. 18: 314-316. 77 Monserud, R.A. 1984. Problems with site indexes: an opinionated review, p. 16-24. in J . Bockheim (ed.)Forest land proc, Dept. Soil Sci . , Univ. Wisconsin, Madison. Mueller-Dombois, D. 1964. The Forest Habitat Types in Southeastern Manitoba and Their Application to Forest Management. Can. J. Bot. 42: 1417-1444. Neter, J . , Kutner, M.H., Nachtsheim, C .J . , and Wasserman, W.. 1996. Applied Linear Statistical Models. 4 t h edition. Irwin, Toronto. 1408 p. Nigh, G . 1997. Early Height Growth and Site Index of Lodgepole Pine Under Wet and Dry Soil Moisture Regimes in British Columbia. West. J. Appl. For. 12(1): 5-8. Nyland, R.D. 1996. Silviculture: Concepts and Applications. McGrw-Hill Inc. Toronto. 633 p. Oliver C D . and B.C. Lawson. 1996. Forest Stand Dynamics, Update Edition. John Wiley and Sons, Inc. Toronto. 520 p. Pfister, R.D. 1984. Development and use of an ecological classification system, p. 16-24 in M. Murray (ed.) Forest classification at high latitudes as an aid to regeneration. USDA For. Serv. Gen. Tech. Rep. PNW-177. Pfister, R.D. and R. Daubenmire. 1975. Ecology of Lodgepole Pine Pinus contorta Doug. p. 27-46. in D.M. Baumgartner (ed.) Proceedings of: Management of Lodgepole Pine Ecosystems. Washington State University. October 9-11, 1973. Pogrebnyak, P.S. 1930. Uberdie Methodik von Standortuntersuchungen in Verbindung mit Waldtypen. p. 455-471 in Ver. II Int. Congr. Forstl. Versuchasanstalten, 1929, Stolkholm. [English Translation: The Methodology of Habitat Studies in Association with Forest Types]. Pojar, J . , K. Klinka and D.V. Meidinger. 1987. Biogeoclimatic Ecosystem Classification in British Columbia. For. Eco. Man. 22: 119-154. Pritchett, W.L. and R.F. Fisher. 1987. Properties and Management of Forest Soils. 2 n d edition. Wiley and Sons. New York. 494 p. Pyatt, Graham. 1995. An Ecological Site Classification for Forestry in Great Britain. Research Information Note 260. Research Division of the Forestry Authority. Forestry Commission, Edinburgh. 6 p. Regent Instruments Inc. (R.I.I.) 1999. W inDENDRO version 6.5 Reference. Blaine, Quebec. 40 p. Reid, Collins Nurseries Ltd. 1983. The management and silviculture of lodgepole pine in B.C.: An interpretative review. Forest Research Council of B.C., Occasional Paper No. 1. Reinecke, L.H. 1933. Perfecting a stand-density index for even-aged forests. J. Agric. Res. 46: 627-638. Sabin, T .E. and S . G . Stafford. 1990. Assessing the Need for Transformation of Response Variables. Special Publication No. 20. College of Forestry. Oregon State University. 31 p. S A S Institute Inc. 1990. SAS/STAT User's Guide. Release 6.12 edition ed. Cary, N . C 1028 p. Silfor, 2000. Accelerated Stand Development of Dense, Fire-origin Lodgepole Pine by Fertilization: Establishment and one-year and three-year progress report for 1996-1999. Contract report for Weldwood of Canada, Hinton Division, Hinton AB . Silfor, 1997. Stand Density Management for Lodgepole Pine - Strategic Options: a component of the Enhanced Forest Management Program of W E L D W O O D of Canada Ltd., Hinton 78 Division. Contract report for Weldwood of Canada, Hinton Division, Hinton AB. Smithers, L.A. 1961. Lodgepole pine in Alberta. Canada, Department of Forestry, Bulletin 127, 153 p. Smithers, L.A. 1956. Assessment of Site Productivity in Dense Lodgepole Pine Stands. Canada, Dept. Northern Affairs and National Resources. Forestry Branch, For. Res. Div., Technical Note No. 30. 20 p. Soil Classification Working Group (SCWG). 1998. The Canadian System of Soil Classification. 3 r d ed. Agric. And Agri-Food Can. Publ. No.1646 (revised). Ottawa. 187 p. Sokal, R.R. and F.J. Rohlf. 1995. Biometry: The Principles and Practice of Statistics in Biological Research, 3 r d Edition. W.H. Freeman and Company. New York. 887 p. Spittlehouse, D.L. and T.A. Black. 1981. A growing season water balance model applied to two Douglas-fir stands. Water Resources Research. 17:1651-1656 Splechtna, B. 1999. Quantitative Characterization of Field-estimated Soil Nutrient Regimes in the Subalpine Coastal Forest [online]. Scientia Silvica Extension Series, No. 21. Available: http://www.forestry.ubc.ca/klinka/ssespubs/sses21.htm (Accessed: October 15, 2000). Spurr, S .H. and B.V. Barnes. 1997. Forest Ecology. 3 r d Edition. John Wiley and Sons, New York. 774 p. S P S S Inc. 1999a. S Y S T A T 9:Statistics I. S P S S Inc. Chicago IL. 660 p. S P S S Inc. 1999b. S Y S T A T 9:Graphics. S P S S Inc. Chicago IL. 459 p. Strong, W.L. 1992. Ecoregions and ecodistricts of Alberta. Alberta Forestry, Lands and Wildlife. Edmonton, AB . Publication T/244. 77 p. Strong, W.L, D.J. Pluth, G.H. La Roi, and I.G.W. Corns. 1991. Forest understory plants as predictors of lodgepole pine and white spruce site quality in west-central Alberta. Can. J. For. Res. 21: 1675-1683 Strong, W.L. and K.R. Leggat. 1981. Ecoregions of Alberta. Alberta Energy and Natural Resources, Edmonton. Technical Report No. T/4. Sukachev V., and N. Dylis. 1964. Fundamentals of Forest Biogeocoenology. Oliver and Boyd. London. 672 p. [Translated to English by Dr. J . M. Maclennan]. Tackle, D. 1961. Silvics of Lodgepole Pine. Intermountain Forest and Range Experiment Station. U.S.D.A. Misc. Pub. No. 19. 24 p. Tesch, S.D. 1981. The evolution of forest yield determination and site classification. For. Ecol. Manage. 3: 169-182. Thornthwaite, C:W. and J.R. Mather. 1955. The water balance. Drexel Inst. Of Technol., Publ. In Climatology 8(1): 1-104 Thornthwaite, C.W. and J.R. Mather. 1957. Instructions and tables for computing potential evapotranspiration and the water balance. Drexel Inst. Of Technol., Publ. in Climatology 10(3): 183-311. Vyse, A. 1985. Spacing lodgepole pine - How much space is enough? Notes prepared for the Cariboo Section of the Northern Silviculture Committee following talk at the spring meeting. 8 p. + graphs (as cited in Marshall 1990) Wang, G . G . 1997. Soil nutrient regime classification for white spruce stands in the Sub-boreal 79 Spruce Zone of British Columbia. Can. J. For. Res. 27(5): 679-685. Wang, G . G . and K. Klinka. 1996. Classification of moisture and aeration regimes in sub-boreal forest soils. Env. Monit. Assess. 39:451-469. Weiner, J. 1985. Size hierarchies in experimental populations of annual plants. Ecol. 66: 743-752 Wilson, F.G. 1979. Thinning as an orderly discipline: A graphic spacing schedule for Red Pine. J. For. 77: 483-486 Wilson, F.G. 1946. Numerical expression of stocking in terms of height. J. For. 44: 756-761. Yoda, K., T. Kira, H. Ogawas and K. Hozkumi. 1963. Intraspecific competition among higher plants. XI. Self thinning in over-crowded pure stands under cultivated and natural conditions. J. Biol. 14: 107-129. Zar J .H . 1984. Biostatistical analysis. Second edition. Prentice-Hall, Englewood Cliffs, New Jersey. 718 p. Zasoski, R.J . 1979. Principles of soil fertility, p. 1-9 in Gessel , S.P. , R.M. Kenady, and W.A. Atkinson, (eds.) Proceedings of: Forest Fertilization Conference, Contribution No. 40, College of Forest Resources, University of Washington, Seattle, Washington. 80 APPENDIX 1 M E A S U R E D VARIABLES: DEFINITIONS AND DESCRIPTIONS 5PRI Mean 5 year radial increment of 5 cut and measured dominant stand trees (mm) 10PRI Mean 10 year radial increment of 5 cut and measured dominant stand trees (mm) A S P Aspect of slope (°) BHA Breast-height age: number of rings counted on disks cut from tree at height of 1.3 metres B S A Base age: number of rings counted on disks cut from base of dominant tree. C A F F Available calcium in forest floor (ppm) C A M S Available calcium in mineral soil (ppm) C F Percentage of coarse fragments (>2 mm) within rooting zone C R V O L H T Ratio of mean crown volume to height for 5 cut and measured dominant stand trees. D C R V O L Mean crown volume of 5 cut and measured dominant stand trees (m3) DDBH Mean diameter at breast height of 5 cut and measured dominant stand trees (cm) DHT Mean height of 5 cut and measured dominant stand trees (m) DIFFHT Absolute difference in mean height between 5 dominant and 5 smallest suppressed lodgepole pine in stand (m), DHT-SHT DLC Mean diameter live crown of 5 cut and measured dominant stand trees (m) D L E A D Mean 1999 leader growth of 5 cut and measured dominant stand trees (cm) E C O Ecosite label E L E Elevation: metres above sea level (masl) E S T Stand establishment period: BaseA minus SbaseA. FF Thickness of forest floor (cm) F L U C Evidence of fluctuating water table: Yes/No FolAI Foliar aluminium (ppm) FolB Foliar boron (ppm) FoICA Foliar calcium (ppm) FolCu Foliar copper (ppm) FolFe Foliar iron (ppm) FolK Foliar potassium (ppm) FolMg Foliar magnesium (ppm) FoIMn Foliar manganese (ppm) FolN Foliar nitrogen (%) FolNP Ratio of foliar nitrogen to phosphorus FolNS Ratio of foliar nitrogen to sulphur FolP Foliar phosphorus (%) FolS Foliar sulphur (ppm) FolZn Foliar zinc (ppm) G L E Y Depth to gleyed soil horizon (cm). Value of 100 cm if none detected H20 Depth to water table (cm). Value of 100 cm if none detected IHD Index of stand height differentiation: relative difference in height between 5 dominant and 5 smallest suppressed trees in stand; (DHT-SHT)/DHT K F F Available potassium in forest floor (ppm) K M S Available potassium in mineral soil (ppm) 81 L L C Mean length of live crown of 5 cut and measured dominant stand trees (m) M G F F Available magnesium in forest floor (ppm) M G M S Available magnesium in mineral soil (ppm) MinNFF Available mineralizeable nitrogen in forest floor (ppm) MinNMS Available mineralizeable nitrogen in mineral soil (ppm) NeedWt Needle weight: mean value of 100 dried needle fascicles, repeated 3 times (g) N O N P L Number of non-lodgepole pine stems per hectare (> 1.3 m tall) P F F Available phosphorus in forest floor (ppm) pHFF pH of forest floor pHMS pH of mineral soil PL Number of lodgepole pine stems per hectare (> 1.3 m tall) P L C Mean percentage live crown of 5 cut and measured dominant stand trees (m) P M S Available phosphorus in mineral soil (ppm) R O O T Rooting depth of stand (cm) S A R Soil aeration regime: Good; Adequate, Restricted, Deficient S B S A Base age of suppressed trees S C R V O L Mean crown volume of 5 cut and measured suppressed stand trees (m3) S D B H Mean diameter at breast height of 5 cut and measured suppressed stand trees (cm) S E C F F Sum of available bases (Ca, Mg, and K) in forest floor (ppm) S E C M S Sum of available bases (Ca, Mg, and K) in mineral soil (ppm) S F F Available sulphur in forest floor (ppm) SHT Mean height of 5 cut and measured suppressed stand trees (m) SI Site index: height of dominant trees at a reference BHA. Reference age is commonly 50 years S L E N D Slendemess coefficient: mean height to diameter ratio of 5 cut and measured dominant stand trees S L P Percent slope (%) S L P O S Slope position: Crest; Upper slope; Mid-slope; Lower slope; Toe slope; Flat; Depression S M R Soil moisture regime: Moderately dry; Slightly dry; Fresh; Wet; Very wet S M S Available sulphur in mineral soil (ppm) S N R Soil nutrient regime: Medium, Poor, Very poor S P A C F Spacing factor: SQRT(10,000/PL)/DHT S T E M S Total stems (all species) per hectare (> 1.3 m tall) tC: tNFF Total carbon to nitrogen ratio of forest floor tC: tNMS Total carbon to nitrogen ratio of mineral soil t C F F Total carbon in forest floor (%) t C M S Total carbon in mineral soil (%) tNFF Total nitrogen in forest floor (%) tNMS Total nitrogen in mineral soil (%) T O P Microtopography of site: Smooth; Slightly mounded; Moderately mounded; Strongly mounded. YTBH Years to breast-height: BaseA minus BHA 82 APPENDIX 2 PLOT DATA SUMMARY: O V E R A L L AND BY ECOSITE Means and standard deviations of measured site and stand properties; overall and by ecosite. ALL PLOTS B ECOSITE D ECOSITE E ECOSITE H ECOSITE VAR N Mean St. Dev n Mean St. Dev n Mean St. Dev n Mean St. Dev n Mean St. Dev BSA 60 39.85 1.17 15 39.80 0.70 15 40.22 0.97 15 39.70 1.10 15 39.68 1.72 BHA 60 32.72 2.91 15 31.05 3.96 15 33.35 2.26 15 33.68 2.39 15 32.78 2.18 YTBH 60 7.13 2.78 15 8.75 3.88 15 6.87 2.62 15 6.02 1.90 15 6.90 1.68 SBSA 60 29.95 3.81 15 28.44 3.78 15 29.73 4.44 15 31.17 3.24 15 30.47 3.53 EST 60 9.90 3.70 15 11.36 3.76 15 10.48 4.16 15 8.53 3.06 15 9.22 3.43 SI 60 13.02 2.90 15 10.70 2.37 15 13.95 1.62 15 16.02 1.51 15 11.42 2.41 ELE 60 1264.17 33.26 15 1266.33 37.15 15 1256.33 17.06 15 1259.33 29.39 15 1274.67 43.73 SLP 60 10.72 12.32 15 12.00 11.49 15 5.07 5.62 15 20.80 16.48 15 5.00 4.96 FF 60 8.58 3.57 15 6.03 2.42 15 8.17 2.70 15 8.90 3.39 15 11.23 3.77 CF 60 22.56 28.82 15 48.30 37.40 15 17.20 27.93 15 16.33 14.88 15 8.40 10.19 ROOT 60 37.08 12.84 15 42.63 13.10 15 38.47 12.16 15 34.00 14.03 15 33.20 10.84 PL 60 19313 10998 15 27307 11896 15 18373 7637 15 10480 3077 15 21093 12002 NONPL 60 2900 4325 15 2827 4803 15 1173 1680 .15 320 483 15 7280 4673 STEMS 60 22060 12279 15 30133 13928 15 19547 7508 15 10800 3153 15 27760 11309 DHT 60 9.61 2.53 15 7.49 1.98 15 10.43 1.59 15 12.18 1.44 15 8.36 2.02 DDBH 60 9.69 2.38 15 8.01 2.08 15 10.37 2.04 15 11.85 1.42 15 8.51 1.87 DCRVOL 60 0.98 0.67 15 . 0.70 0.55 15 1.08 0.71 15 1.45 0.69 15 0.71 0.45 DLEAD 60 26.73 8.28 15 21.13 7.15 15 27.59 7.93 15 32.57 7.08 15 25.65 7.23 LLC 60 4.13 1.03 15 3.59 1.04 15 4.42 0.88 15 4.78 0.75 15 3.72 1.01 PLC 60 0.44 0.07 15 0.48 0.05 15 0.42 0.04 15 0.40 0.07 15 0.45 0.06 DLC 60 1.08 0.24 15 0.97 0.21 15 1.13 0.22 15 1.27 0.23 15 0.97 0.16 5PRI 60 3.85 1.30 15 3.25 1.27 15 3.46 0.97 15 4.68 1.14 15 4.01 1.39 10PRI 60 8.70 2.58 15 7.62 2.30 15 7.78 1.71 15 10.74 2.49 15 8.67 2.65 SHT 60 4.50 2.32 15 2.72 0.90 15 4.48 1.18 15 7.06 2.65 15 3.75 1.59 SDBH 60 2.62 1.42 15 1.58 0.86 15 2.58 0.78 15 4.01 1.47 15 2.33 1.28 SCRVOL 60 0.012 0.017 15 0.004 0.006 15 0.006 0.004 15 0.026 0.024 15 0.011 0.015 DIFFHT 60 5.11 1.50 15 4.76 1.47 15 5.95 1.43 15 5.12 1.55 15 4.61 1.29 IHD 60 0.55 0.14 15 0.63 0.08 15 0.57 0.10 15 0.43 0.17 15 0.56 0.12 SLEND 60 99.00 9.16 15 93.28 5.79 15 101.55 8.05 15 103.12 8.50 15 98.03 11.04 SPACF 60 8.59 1.98 15 9.19 2.42 15 7.53 1.16 15 8.29 0.77 15 9.37 2.52 CRNHT 60 9.55 4.97 15 8.38 4.80 15 9.80 5.08 15 11.84 5.18 15 8.19 4.41 83 Means and standard deviations for mineral soil, forest floor, and foliar nutrient values; overall and by ecosite. ALL PLOTS B ECOSITE D ECOSITE E ECOSITE H ECOSITE VAR N Mean St. Dev n Mean St. Dev n Mean St. Dev n Mean St. Dev n Mean St. Dev pHMS 60 4.94 0.44 15 4.92 0.36 15 4.94 0.25 15 5.15 0.68 15 4.77 0.26 tCMS 60 0.97 0.55 15 0.66 0.25 15 0.76 0.25 15 1.16 0.65 15 1.29 0.67 tNMS 60 0.06 0.02 15 0.04 0.01 15 0.06 0.02 15 0.07 0.02 15 0.07 0.02 tCitNMS 60 16.29 4.67 15 17.53 5.61 15 14.03 2.63 15 16.57 5.20 15 17.05 4.33 MinNMS 60 6.66 3.71 15 3.96 1.85 15 4.93 2.04 15 10.22 3.78 15 7.55 3.25 PMS 60 20.93 18.08 15 22.40 12.74 15 32.07 27.98 15 12.87 8.74 15 16.40 11.36 KMS 60 82.65 35.70 15 71.40 33.10 15 79.33 23.21 15 99.17 38.80 15 80.70 42.30 CAMS 60 1257.0 1252.3 15 837.7 548.7 15 1086.0 761.8 15 2160.0 1990.9 15 944.3 752.0 MGMS 60 209.0 163.5 15 179.0 127.9 15 191.0 138.7 15 289.7 233.7 15 176.2 114.0 SMS 60 1.67 0.82 15 1.70 0.90 15 1.80 1.13 15 1.52 0.62 15 1.67 0.57 SECMS 60 1548.6 1404.4 15 1088.1 696.6 15 1356.3 905.2 15 2548.8 2184.8 15 1201.3 875.2 pHFF 60 4.25 0.40 15 4.34 0.45 15 4.11 0.23 15 4.45 0.41 15 4.11 0.39 tCFF 60 46.04 2.68 15 46.03 2.69 15 47.92 2.40 15 44.73 2.31 15 45.49 2.46 tNFF 60 1.15 0.15 15 1.11 0.15 15 1.12 0.11 15 1.29 0.11 15 1.07 0.13 tC:tNFF 60 40.78 5.93 15 42.11 7.05 15 43.01 4.19 15 34.97 2.87 15 43.02 4.89 MinNFF 60 519.5 256.2 15 361.3 164.1 15 502.4 210.6 15 763.4 253.1 15 451.0 213.0 PFF 60 168.4 72.5 15 124.1 73.8 15 178.9 60.4 15 206.7 79.3 15 163.9 54.3 KFF 60 1009.4 268.4 15 870.2 295.6 15 1092.0 204.2 15 1123.6 301.8 15 951.8 194.5 CAFF 60 3119.0 1965.0 15 3358.3 1742.2 15 2475.0 719.9 15 3917.9 2763.6 15 2725.0 1930.3 MGFF 60 450.0 177.5 15 472.6 198.9 15 395.1 112.9 15 520.6 191.9 15 411.9 181.8 SFF 60 67.00 25.20 15 52.98 26.31 15 73.98 23.13 15 72.21 28.21 15 68.82 18.93 SECFF 60 4578.5 2105.2 15 4701.2 1956.3 15 3962.1 867.6 15 5562.0 2863.9 15 4088.7 2044.2 NeedWt 60 2.13 0.28 15 2.05 0.29 15 2.11 0.14 15 2.36 0.27 15 2.02 0.30 FolN 60 1.12 0.13 15 1.03 0.10 15 1.16 0.12 15 1.21 0.10 15 1.09 0.13 FolP 60 0.15 0.02 15 0.13 0.01 15 0.15 0.02 15 0.16 0.01 15 0.14 0.02 FoICa 60 0.15 0.02 15 0.15 0.02 15 0.15 0.02 15 0.15 0.03 15 0.14 0.02 FolMg 60 • 0.11 0.01 15 0.11 0.01 15 0.11 0.01 15 0.11 0.01 15 0.11 0.01 FolK 60 . 0.54 0.05 15 0.51 0.04 15 0.54 0.06 15 0.54 0.06 15 0.55 0.05 FolS 60 0.11 0.01 15 0.10 0.01 15 0.11 0.01 15 0.13 0.01 15 0.11 0.01 FolCu 60 5.91 0.72 15 5.92 0.66 15 5.84 0.75 15 6.00 0.86 15 5.87 0.67 FolZn 60 53.78 5.99 15 53.38 3.78 15 56.04 6.70 15 56.33 4.25 15 49.38 6.39 FolFe 60 51.47 22.18 15 46.11 13.49 15 60.00. 33.82 15 45.89 8.13 15 53.89 23.07 FoIMn 60 488.1 131.5 15 429.7 82.4 15 592.0 146.3 15 487.3 106.5 15 443.1 126.4 FolB 60 19.41 4.09 15 19.57 5.25 15 18.91 3.06 15 19.81 3.42 15 19.35 4.63 FolAI 60 556.1 148.6 15 585.5 106.2 15 639.5 179.8 15 473.7 121.9 15 525.9 133.8 FolNP 60 7.80 0.78 15 7.93 0.80 15 7.73 0.72 15 7.49 0.76 15 8.03 0.82 FolNS 60 10.03 1.03 15 10.20 1.18 15 10.24 1.11 15 9.61 0.90 15 10.09 0.89 84 APPENDIX 3 REGRESSION SLOPE AND ELEVATION TESTING OUTPUT >REM THIS TESTS THE SLOPE INTERACTIONS BETWEEN B AND D ECOSITES Dep V a r : S I N: 30 M u l t i p l e R: 0.84164 S q u a r e d m u l t i p l e R: 0.70836 A d j u s t e d s q u a r e d m u l t i p l e R: 0.67471 S t a n d a r d e r r o r o f e s t i m a t e : 1.47719 E f f e c t C o e f f i c i e n t S t d E r r o r S t d Coef T o l e r a n c e t P(2 T a i l ) CONSTANT 17.04541 1 02356 0.00000 16 65304 2 2E-15 L I V E P L -0.00017 0 00005 -0.70470 0.24040 -3 26237 0 00309 ECOl -2.68922 1 41929 -0.52803 0.14443 -1 89476 0 06930 LIVEPL*EC01 0.00003 0 00006. 0.21689 0.07634 0 56584 0 57636 S o u r c e R e g r e s s i o n R e s i d u a l A n a l y s i s of V a r i a n c e Sum-of-Squares d f Mean-Square 137.79969 3 45.93323 56.73398 26 ' 2.18208 F - r a t i o 21.05024 3. 94507E-07 Durbin-Watson-D S t a t i s t i c F i r s t O r d e r A u t o c o r r e l a t i o n 1. 812 0.091 >REM THIS TESTS THE LEVELS BETWEEN B AND D ECOSITES Dep V a r : S I N: 30 M u l t i p l e R: 0.83950 S q u a r e d m u l t i p l e R: 0.70477 A d j u s t e d s q u a r e d m u l t i p l e R: 0.68290 S t a n d a r d e r r o r o f e s t i m a t e : 1.45847 S t d E r r o r S t d Coef T o l e r a n c e t P(2 T a i l ) E f f e c t CONSTANT L I V E P L ECOl C o e f f i c i e n t 16.59313 -0.00014 -1.95993 0 . 63125 0 . 00003 0 . 58677 0 . 0 0 0 0 0 - 0 . 6 0 1 8 4 - 0 . 3 8 4 8 3 26 0.82377 -5 0.82377 -3. 28629 9.9E-16 22377 0.00002 34022 0.00246 S o u r c e R e g r e s s i o n R e s i d u a l A n a l y s i s o f V a r i a n c e Sum-of-Squares d f Mean-Square 1 3 7 . 1 0 1 0 5 5 7 . 4 3 2 6 1 2 27 6 8 . 5 5 0 5 3 2 .12713 F - r a t i o P 32.22671 7.03415E-08 D u r b i n - W a t s o n D S t a t i s t i c 1.792 F i r s t O r d e r A u t o c o r r e l a t i o n 0.099 ECO 0 10000 20000 30000 40000 50000 PI Density (stems per hectare) 85 >REM THIS TESTS THE SLOPE INTERACTIONS BETWEEN B AND E ECOSITES Dep Var: SI N: 30 M u l t i p l e R: 0.89932 Squared m u l t i p l e R: 0.80877 Adjusted squared m u l t i p l e R: 0.78671 Standard e r r o r of estimate: 1.54110 E f f e c t C o e f f i c i e n t Std E r r o r Std Coef Tolerance t P (2 T a i l ) CONSTANT 19 34908 1 45819 0. 00000 13 26920 4 3E-13 LIVEPL -0 00032 0 00013 -1. 15072 0 03128 -2 37308 ' 0 02532 ECOl -4 99288 1 78283 -0. 76092 0 09963 -2 80053 0 00950 LIVEPL*EC01 0 00018 0 00014 0. 88997 0 01640 1 32909 0 .19536 A n a l y s i s of Variance Source Sum-of-Squares df Mean-Square F - r a t i o P Regression 261.16217 3 87.05406 36.65444 1.73520E-09 Res i d u a l 61.74983 26 2.37499 Durbin-Watson D S t a t i s t i c 1.137 F i r s t Order A u t o c o r r e l a t i o n 0.404 >REM THIS TESTS THE THE LEVELS BETWEEN B AND E ECOSITES Dep Var: SI N: 30 M u l t i p l e R: 0.89206 Squared m u l t i p l e R: 0.79578 Adjusted squared m u l t i p l e R: 0.78065 Standard e r r o r of estimate: 1.56282 E f f e c t CONSTANT LIVEPL ECOl Source Regression Residual C o e f f i c i e n t 17.54397 -0.00015 -2.87312 Std E r r o r 0.53827 0.00003 0.80796 Std Coef Tolerance P(2 T a i l ) 0.00000 -0.52677 -0.43787 A n a l y s i s of Variance Sum-of-Squares. df Mean-Square 256.96680 65.94520 2 27 128.48340 2 .44241 32.59341 9.9E-16 0.49885 -4.27797 0.00021 0.49885 -3.55600 0.00141 F - r a t i o P 52.60507 4.85665E-10 Durbin-Watson D S t a t i s t i c ' 1.205 F i r s t Order A u t o c o r r e l a t i o n •• 0.379 sz JD CO CO cu >. o LO X CD •a _c a> CO "D a> E 20 15 10 X I \ X x — y X \ — x X ... \ x KJ ">\_^ O -o G I I I o I E C O o B x E 0 10000 20000 30000 40000 50000 PI Density (stems per hectare) 86 >REM THIS TESTS THE SLOPE INTERACTIONS BETWEEN B AND H ECOSITES Dep Var: SI • N: 28 M u l t i p l e R: 0.75479 Squared m u l t i p l e R: 0.56971 Adjusted squared m u l t i p l e R: 0.51593 Standard e r r o r of estimate: 1.59854 E f f e c t C o e f f i c i e n t Std E r r o r Std Coef Tolerance t P(2 T a i l ) CONSTANT 15 47380 1 05581 0. 00000 14 65584 1 8E-13 LIVEPL -0 00019 0 00005 -0. 94110 0 28645 -3 76176 0 00096 ECOl -1 11761 1 49893 -0. 24705 0 16331 -0 74560 0 46315 LIVEPL*EC01 0 00006 0 00006 0. 40370 0 09222 0 91560 0 36898 A n a l y s i s of Variance Source Sum-of-Squares df Mean-Square F - r a t i o . P Regression 81.20048 3 27.06683 10.59227 0.00013 Residual 61.32809 24 2.55534 Durbin-Watson D S t a t i s t i c F i r s t Order A u t o c o r r e l a t i o n 1.055 0.451 >REM THIS TESTS THE LEVELS BETWEEN B AND H ECOSITES Dep Var: SI N: 28 M u l t i p l e R: 0.74477 Squared m u l t i p l e R: 0.55468 Adjusted squared m u l t i p l e R: 0.51906 Standard e r r o r of estimate: 1.59336 E f f e c t C o e f f i c i e n t Std E r r o r Std Coef Tolerance P(2 T a i l ) CONSTANT LIVEPL ECOl 14.75778 0.70705 0.00000 . 20.87227 9.9E-16 -0.00015 0.00003 -0.75416 0.85781 -5.23352 0.00002 0.11729 0.65190 0.02593 0.85781 0.17992 0.85867 Source Regression R e s i d u a l A n a l y s i s of Variance Sum-of-Squares df Mean-Square 7 9 . 0 5 8 2 8 63 .47030 2 25 39.52914 2 . 53881 F - r a t i o 15.56994 0 . 0 0 0 0 4 Durbin-Watson D S t a t i s t i c F i r s t Order A u t o c o r r e l a t i o n 1.250 0 . 3 5 9 OS CO 05 >* O LO X CD "O c 05 CO "D 05 « E 05 LU 16 15 14 13 12 11 10 9 8 7 6 X I I X ° X -x ^ v x c: o -X Q - Q • \ I I I X O | -ECO o B x H 10000 20000 30000 40000 50000 PI Density (stems per hectare) 87 >REM THIS TESTS THE SLOPE INTERACTIONS BETWEEN 0 AND E ECOSITES Dep Var: SI N: 30 M u l t i p l e R: 0.82655 Squared m u l t i p l e R: 0.68319 Adjusted squared m u l t i p l e R: 0.64663 Standard e r r o r of estimate: 1.10806 E f f e c t C o e f f i c i e n t Std E r r o r Std Coef Tolerance t P(2 T a i l ) CONSTANT 19.34908 1 04845 0.00000 18 45498 9 9E-16 LIVEPL -0 . 00032 0 00010 -1.19095 0.09358 -3 30051 0 00281 ECOl -2.30367 1 29952 -0.62850 0.09694 -1 77271 0 08799 LIVEPL*EC01 0.00015 0 00010 0.85895 0.03406 1 43600 0 16293 A n a l y s i s of Variance Source Sum-of-Squares df Mean-Square F - r a t i o P Regression 68.83908 3 22.94636 18.68913 1.13861E-06 Res i d u a l 31.92259 26 1.22779 Durbin-Watson D S t a t i s t i c F i r s t Order A u t o c o r r e l a t i o n 1. 666 0.150 >REM THIS TESTS THE LEVELS BETWEEN D AND E ECOSITES Dep Var: SI N: 30 M u l t i p l e R: 0.81121 Squared m u l t i p l e R: 0.65806 Adjusted squared m u l t i p l e R: 0.63273 Standard e r r o r of estimate: 1.12964 E f f e c t CONSTANT LIVEPL ECOl Source Regression R e s i d u a l C o e f f i c i e n t Std E r r o r Std Coef Tolerance P(2 T a i l ) 18.00560 0.48245 0.00000 . 37.32134 9.9E-16 -0.00019 0.00004 -0.71033 0.67007 -5.16685 0.00002 -0.57782 0.50391 -0.15764 0.67007 -1.14667 0.26158 A n a l y s i s of Variance Sum-of-Squares df Mean-Square 66.30725 34.45442 2 27 33.15362 1.27609 F - r a t i o P 25.98064 5.10880E-07 Durbin-Watson D S t a t i s t i c 1.745 F i r s t Order A u t o c o r r e l a t i o n • 0.100 20 to CD >, o UO X CD £ co •o £ To E x 1 1 I - x > — x V J x>k o A * n x » v X > v _ x n \ ^ o -i i 1 E C O o D x E 10000 20000 30000 40000 PI Density (stems per hectare) 88 >REM THIS TESTS THE SLOPE INTERACTIONS BETWEEN B AND E ECOSITES Dep Var: SI N: 28 M u l t i p l e R: 0.85996 Squared m u l t i p l e R: 0.73954 Adjusted squared m u l t i p l e R: 0.70698 Standard e r r o r of estimate: 1.14566 E f f e c t C o e f f i c i e n t Std E r r o r Std Coef Tolerance t P(2 T a i l ) CONSTANT 15 47380 0 75669 0 . 00000 20 44934 9 9E-16 LIVEPL -0 00019 0 00004 -0. 73845 0 54829 -5 24880 0 00002 ECOl 1 57161 1 09671 0. 37713 0 15669 1 43303 0 16475 LIVEPL*EC01 0 00002 0 00005 0. 11331 0 14148 0 40912 0 .68608 A n a l y s i s of Variance Source Sum-of-Squares df Mean-Square F - r a t i o P Regression 89.44023 3 29.81341 22.71437 3.42362E-07 Residual 31.50085 24 1.31254 Durbin-Watson D S t a t i s t i c 1.270 F i r s t Order A u t o c o r r e l a t i o n 0.3 61 >REM THIS TESTS THE LEVELS BETWEEN B AND E ECOSITES Dep Var: SI N: 28 M u l t i p l e R: 0.85891 Squared m u l t i p l e R: 0.73772 Adjusted squared m u l t i p l e R: 0.71674 Standard e r r o r of estimate: 1.12642 E f f e c t CONSTANT LIVEPL ECOl C o e f f i c i e n t 15.28509 -0.00018 1.98360 Std E r r o r 0.58977 0.00003 0.42706 Std Coef Tolerance P{2 T a i l ) 0.00000 . 25.91691 9.9E-16 -0.69979 0.99896 -6.82858 3.7E-07 0.47600 0.99896 4.64481 0.00009 Source Regression R e s i d u a l A n a l y s i s of Variance Sum-of-Squares df Mean-Square 89.22054 31.72053 2 25 44.61027 1.26882 F - r a t i o P 35.15883 5.42729E-08 Durbin-Watson D S t a t i s t i c 1.393 F i r s t Order A u t o c o r r e l a t i o n 0.301 ECO o D x H 0 10000 20000 30000 40000 50000 PI Density (stems per hectare) 89 >REM THIS TESTS THE SLOPE INTERACTIONS BETWEEN E AND H ECOSITES Dep Var: SI N: 28 M u l t i p l e R: 0.90707 Squared m u l t i p l e R:' 0.82277 Adjusted squared m u l t i p l e R: 0.80061 Standard e r r o r of estimate: 1.23350 E f f e c t C o e f f i c i e n t Std E r r o r Std Coef Tolerance t P (2 T a i l ) CONSTANT 15 47380 0. 81471 0. 00000 18 99304 9 '9E-16 LIVEPL -0 00019 0 00004 -0. 53479 0 613 65 -4 87501 0 00006 ECOl ' 3 87528 1 42337 0. 71247 0 10784 2 72261 0 01187 LIVEPL*EC01 -0 00013 0 00011 -0. 26477 0 13031 -1 11220 0 .27707 Source Regression Residual A n a l y s i s of Variance Sum-of-Squares df Mean-Square 169.52187 36.51670 3 24 56.50729 1. 52153 F - r a t i o P 37.13849 3.53967E-09 Durbin-Watson D S t a t i s t i c F i r s t Order A u t o c o r r e l a t i o n 1.407 0.250 >REM THIS TESTS THE LEVELS BETWEEN E AND H ECOSITES Dep Var: SI N: 28 M u l t i p l e R: 0.90202 Squared m u l t i p l e R: 0.81363 Adjusted squared m u l t i p l e R: 0.79872 Standard e r r o r of estimate: 1.23934 E f f e c t CONSTANT LIVEPL ECOl Source Regression R e s i d u a l C o e f f i c i e n t Std E r r o r Std Coef Tolerance P(2 T a i l ) 15.75596 0.77786 0.00000 . 20.25545 9.9E-16 -0.00021- 0.00004 -0.57665 0.69554 -5.57001 0.00001 2.42010 0.56311 0.44494 0.69554 4.29776 0.00023 A n a l y s i s of Variance Sum-of-Squares df Mean-Square 167.63976 38.39881 2 25 83.81988 1.53595 F - r a t i o P 54.57192 7.57907E-10 Durbin-Watson D S t a t i s t i c F i r s t Order A u t o c o r r e l a t i o n 1.466 0.236 SZ .ci co i _ CO CD >^ o LO X CD X> _c a> CO CD "CO E 00 LU 20 10 o I I I I -x ^ X X x \ N \ I I I X I ECO o E x H 0 10000 20000 30000 40000 50000 PI Density (stems per hectare) 90 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0075154/manifest

Comment

Related Items