SMALL SCALE DISTURBANCES AND STAND DYNAMICS IN INONOTUS
TOMENTOSUS INFECTED AND UNINFECTED OLD-GROWTH AND PARTIAL
CUT WET, SUB-BOREAL FORESTS IN BRITISH COLUMBIA
by
Jules (Ted) Newbery
B.Sc. The University of Northern British Columbia, 1997

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE
in
NATURAL RESOURCES MANAGEMENT

©Jules (Ted) Newbery, 2001
THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA
May 2001
All rights reserved. This work may not be
reproduced in whole or in part, by photocopy
or other means, without the permission of the author.

1 ^ 1

National Library
o( C an ad a

Bibliothèque nationale
du Canada

Acquisitions and
Bibliographie Services

Acquisitions et
senrices bibliographiques

385 W*lln 0lon Strw l
Ottawa ON K1A0N4
Canada

385, rua Waüngton
Ottawa ON K1A0N4
Canada
w •» tuoMiWmnB*
Our a» «emiWMn»

The author has granted a non­
exclusive licence allowing 6e
National lib n n y (^Canada to
reproduce, loan, distribute w sell
copies o f 6 is 6esis in microfonn,
papor or electrcaiic formats.

L’auteur a accordé une licence non
«[elusive pennettant à la
Bibliodièqoe nationale du Canada de
reproduire, prêter, distribuer ou
vôidre des copies de cette thèse sous
la fmrne de microfiche/film, de
re^oductitm sur papier on sur format
électronique.

The author retains ownershÿ of the
copyright in this thesis. Neither the
thesis nor substantial extracts hom it
may be printed or otherwise
reproduced without 6e author’s
permission.

L’auteur conserve la propriété du
droit d’auteur <pri protège cette thèse.
Ni la thèse ni des extraits substantiels
de cdle-d ne doivent être inqximés
ou autranent rejnoduits sans son
autorisation.

0 612 80681-2
-

CanadS

-

APPROVAL
Name:

Jules E. Newbery

Degree:

Master of Science

Thesis Title:

SMALL SCALE DISTURBANCES AND STAND
DYNAMICS IN INONOTUS TOMENTOSUS INFECTED
AND UNINFECTED OLD-GROWTH AND PARTIAL CUT
WET, SUB-BOREAL FORESTS INJgRIHSH COLUMBIA

Examining Committee:
Clmir: Dr. Max Blouw
Vice-President (Research)
Professor, Biology Program
UNBC

Supervisor: Df. Kathy Lewis
Associate Professor, Forestry Program
UNBC

îe Member: Dr. Josef Ackerman
tate Professor, Enyironm^ataT^tudies Program
UNBC

Conunittee Mêfiibér Dr. Michael Walters
Assistant Professor, Department of Forest Resources
Michigan State University

Committee Member: Michael Jun, MSc
Manager, Aleza Lake Research Forest
UNBC

External Examiner: Richard IQg^zems, MSc
Research Scientist, Ministry of Forests
Prince George Region (Fort St. John)

Date Approved:

fsJov/c.m.b€^ 2., z o o I

ABSTRACT
Disturbance and stand dynamics were compared for old-growth and partial cut forests
with and without Inonotus tomentosus (Fr.) Teng caused root disease in wet, cool SubBoreal Spruce forests near Prince George, British Columbia, Canada. Two methodologies
(a time since death model and tree ring growth rate criteria) were developed to support
two studies - a gap level analysis and a stand level analysis. The objectives of the studies
were to; 1) Determine the spatial and temporal patterns of small-scale disturbance in oldgrowth and partial cut forests with and without the influence of I. tomentosus-, 2)
Determine how stand composition and structure differ between I. tomentosus infected and
non-infected old-growth forests; 3) Determine how I. tomentosus affects stand dynamics
in partial cut forests and how its effects differ from its effects in old growth forests.
In the gap level analysis, four forest types were sampled in 10 meter fixed radius plots
(Old-growth with I. tomentosus (GOT) (n=21), and partial cut with I. tomentosus (PCT)
(n=22), Old-growth without /. tomentosus (OGNT) (n=23) and partial cut without /.
tomentosus (PCNT) (n=23)). P. glauca x engelmannii mortality was 50% lower in partial
cut forests than in old-growth forests regardless of infection status. The functional gap
size caused by a single tree averaged 16.76 m^. Summed gap-size measures for all trees
dying in a decade indicated that between 6.9 and 8.1% of stand area was made available
to understory trees per decade. Due to high mortality rates and low recruitment rates to
the canopy for P. glauca x engelmannii, old-growth forests are shifting to a canopy
dominated by^L lasiocarpa. In the partial cut plots, higher relative P. glauca x
engelmannii recruitment and lower mortality indicate that P. glauca x engelmannii
populations may rise relative to present densities.

In old-growth forests, the spatio-temporal patterns of canopy disturbance and
canopy patch structure were quantified from 0.49ha (n=6) stem-mapped plots using
Moran’s I and Standard Normal Deviates. Canopy disturbance and canopy composition
were similar for I. tomentosus infected and uninfected stands at low levels of incidence.
For stand types pooled, average decadal canopy disturbance ranged from 5.09%-6.0%.
Gaps size averaged < 7m in diameter and were irregularly distributed across the forests.
Species patch structure analysis indicated that P. glauca x engelmannii is found in small
patches which are probably remnants of a once nearly homogenous spruce canopy. These
results show that small-scale disturbances are important successional mechanisms in oldgrowth Sub-Boreal Spruce forests because of their effects on stand structure and
dynamics.

Ill

TABLE OF CONTENTS

ABSTRACT

ii

TABLE OF CONTENTS

iv

LIST OF TABLES

ix

LIST OF FIGURES

x

ACKNOWLEDGEMENT

xii

PREFACE

I

CHAPTER 1. INTRODUCTION TO SUB-BOREAL STAND DYNAMICS;
LITERATURE REVIEW.

3

1.1 Species Associations, Range and Climate

3

1.2 Succession

5

1.3 Small and Medium Scale Disturbances in Boreal, Sub-Boreal and
Sub-Alpine Forests

7

1.4 Stand Development In Sub-Boreal, Boreal And Sub-Alpine Forests

9

1.5 Resource Dynamics in Gaps

11

1.6 Inonotus tomentosus as a Cause of Small Scale Disturbance

13

1.7 Research Questions

17

1.8 LITERATURE CITED

19

CHAPTER 2. ESTIMATING TIME SINCE DEATH IN PICEA GLAUCA x
ENGELMANNII AlAU ABIES LASIOCARPA IN A WET-COOL
SUB-BOREAL SPRUCE FOREST.
2.0 ABSTRACT

26

2.1 INTRODUCTION

28

2.2 METHODS

31

2.21 Study Area

31

IV

2.22 General Approach

31

2.23 Explanatory Variables

32

2.24 Analysis and Modeling

35

2.3 RESULTS
2.31 Model Validation

38
39

2.4 DISCUSSION

40

2.5 CONCLUSIONS

43

2.6 LITERATURE CITED

44

2.7 APPENDIX

47

CHAPTER 3. A METHODLOLOGY FOR ESTIMATING YEAR OF DEATH IN
SBS FORESTS USING TREE RING GROWTH RATES.
3.0 ABSTRACT

59

3.1 INTRODUCTION

61

3.2 METHODS

63

3.21 Study Area and Site Selection

63

3.22 Sampling Design and Plot Measurements

64

3.23 Analysis

64

3.3 RESULTS AND DISCUSSION

70

3.31 Comprehensive Analysis

70

3.32 Selective Analysis

71

3.4 CONCLUSIONS

76

3.5 LITERATURE CITED

77

CHAPTER 4. SPATIOTEMPORAL PATTERNS OF SMALL-SCALE
DISTURBANCES IN SUB-BOREAL SPRUCE FORESTS:
IMPLICATIONS FOR PARTIAL CUT HARVESTING AND
INONOTUS TOMENTOSUS ROOT DISEASE.
4.0 ABSTRACT

82

4.1 INTRODUCTION

84

4.2 METHODS

87

4.21 Study Area

87

4.22 Site Selection

88

4.23 Sampling Design and Measurements

89

4.24 Data Analysis

91

4.3 RESULTS

94

4.31Stand Composition

94

4.32 Disturbance

95

4.33 Accumulated Mortality

95

4.34 Functional Gap Size

96

4.35 Transition Probabilities

97

4.36 Simulations of Future Stand Compositions

97

4.4 DISCUSSION

99

4.5 CONCLUSIONS

105

4.6 LITERATURE CITED

106

VI

CHAPTER 5. STAND LEVEL SPATIO-TEMPORAL DISTURBANCE
PATTERNS CAUSED BY INONOTUS TOMENTOSUS AND
OTHER AGENTS IN SUB-BOREAL SPRUCE-FIR FORESTS.
5.0 ABSTRACT

115

5.1 INTRODUCTION

116

5.2 METHODS

119

5.21 Study Area

119

5.22 Site Selection

120

5.23 Sampling Design and Plot Measurements

121

5.24 Interpretation of Growth Patterns

123

5.25 Analysis for Spatial Patterns of Disturbance

123

5.3 RESULTS

126

5.31 Canopy Composition

126

5.32 Canopy Ascension by Species

126

5.33 Canopy Ascension by Decade

126

5.34 Spatial Patterns of Canopy Disturbance

127

5.35 Spatial Patterns of Species Patches

128

5.4 DISCUSSION

129

5.41 Spatial Patterns of Disturbance

129

5.42 Species Patch Structure

130

5.5 CONCLUSIONS

131

5.6 LITERATURE CITED

133

VII

CHAPTER 6. SUMMARY OF FINDINGS, MANAGEMENT
RECOMENDATIONS AND CONCLUSIONS.
6.0 STUDY RATIONALE

144

6.1 SUMMARY OF RESEARCH

149

6.11 Canopy Disturbance

149

6.12 Gap Size

150

6.13 Canopy Mortality By Species

150

6.14 Stand Composition, Canopy Replacement and Future Stand
Composition

150

6.15 Impacts of Inonotus tomentosus and Partial Cutting on
Disturbance and Stand Composition

151

6.2 MANAGEMENT APPLICATIONS

153

6.21 Prescribed Harvest Intensity for Partial Cut Harvesting in WetCool SBS Forests

153

6.22 Harvest Dispersion

153

6.23 Harvest Timing

153

6.24 Development of Methodology

154

6.25 Future Research

156

6.3 CONCLUSIONS

158

6.4 LITERATURE CITED

159

vni

LIST OF TABLES

CHAPTER 1. INTRODUCTION TO SUB-BOREAL STAND DYNAMICS;
LITERATURE REVIEW.
Table 1. Climatic comparisons for various sub-boreal, sub-alpine and
boreal forests present in British Columbia.

^3

CHAPTER 2. ESTIMATING TIME SINCE DEATH IN PICEA GLAUCA x
ENGELMANNII AND ABIES LASIOCARPA IN A WET-COOL
SUB-BOREAL SPRUCE FOREST.
Table 1. Bivariate regression statistics.

51

Table 2. Correlation matrix for spruce model and fir.

52

Table 3a. Summary of regression statistics for the spruce models.

53

Table 3b. Summary of regression statistics for the fir models.

54

Table 4. Comparison of mean years since death.

55

CHAPTER 4. SPATIOTEMPORAL PATTERNS OF SMALL-SCALE
DISTURBANCES IN SUB-BOREAL SPRUCE FORESTS:
IMPLICATIONS FOR PARTIAL CUT HARVESTING AND
INONOTUS TOMENTOSUS ROOT DISEASE.
Table 1. Summary of stand composition and mortality data.

109

Table 2. Species composition dynamic over the next 200 years.

110

CHAPTER 5. STAND LEVEL SPATIO-TEMPORAL DISTURBANCE
PATTERNS CAUSED BY INONOTUS TOMENTOSUS AND
OTHER AGENTS IN SUB-BOREAL SPRUCE-FIR FORESTS.
Table 1. Proportion of canopy accession types by treatment type, and
species.

^

Table 2. Summary of spatial autocorrelation of species association using
joins-count statistics.

137

LIST OF FIGURES

CHAPTER 1. INTRODUCTION TO SUB-BOREAL STAND DYNAMICS ;
LITERATURE REVIEW.
Figure 1. Location of Sub Boreal Spruce biogeoclimatic zone and the
Aleza Lake Research Forest.

24

Figure 2. Models of Succession.

25

CHAPTER 2. ESTIMATING TIME SINCE DEATH IN PICEA GLAUCA x
ENGELMANNII
ABIES LASIOCARPA IN A WET-COOL
SUB-BOREAL SPRUCE FOREST.
Figure 1. Regression fit for spruce model.

56

Figure 2. Regression fit for fir model.

57

Figure 3. Plotted values of dates of tree mortality.

58

CHAPTER 3. A METHODLOLOGY FOR ESTIMATING YEAR OF DEATH IN
SBS FORESTS USING TREE RING GROWTH RATES.
Figure 1. Demonstration of various incremental growth patterns used to
assess date of canopy ascension.

79

Figure 2. Box-plots showing the range in initial 10-year average growth
rate.

80

Figure 3. Decision set for determining year of canopy ascension.

81

CHAPTER 4. SPATIOTEMPORAL PATTERNS OF SMALL-SCALE
DISTURBANCES IN SUB-BOREAL SPRUCE FORESTS:
IMPLICATIONS FOR PARTIAL CUT HARVESTING AND
INONOTUS TOMENTOSUS ROOT DISEASE.
Figure 1. Disturbance chronologies for the four forest types.

111

Figure 2. Transition probability data from 1930-1950.

112

Figure 3. Transition probability data from 1950-1970.

113

Figure 4. Transition probability data from 1970-1997.

114

CHAPTER 5. STAND LEVEL SPATIO-TEMPORAL DISTURBANCE
PATTERNS CAUSED BY INONOTUS TOMENTOSUS AND
OTHER AGENTS IN SUB-BOREAL SPRUCE-FIR FORESTS.
Figure I . Canopy ascension by decade and by species for Inonotus
tomentosus infected plots.

138

Figure 2. Canopy ascension by decade and by species for Inonotus
tomentosus uninfected plots.

139

Figure 3. Locations of canopy ascension dates for trees in the three noninfected plots.

140

Figure 4. Locations of canopy ascension dates for trees in the three
Inonotus tomentosus infected plots.

141

Figure 5. Correlogram showing Moran’s I coefficients for Inonotus
tomentosus infected plots.

142

Figure 6. Correlogram showing Moran’s coefficients for Inonotus
tomentosus uninfected plots.

143

CHAPTER 6. SUMMARY OF FINDINGS, MANAGEMENT
RECOMENDATIONS AND CONCLUSIONS.
Figure 1. Diameter class distribution (dbh) of dead trees in old-growth
forests.

160

XI

ACKNOWLEDGEMENT
“If you know how long its going to take, how much money it will cost and what to
expect, isn’t research”. It took lots of time, quite a bit of money and there were several
unexpected obstacles along the way. That said there are many people to thank for all of
their time and efforts, financial contributions and guidance in this project. Alison -you
tolerated 3 Vi years of this! How will 1 every repay you? You gave me time to work and
time to play. To Mike and Kathy who trusted 1 could do and learn - you are important
mentors. Thanks to my assistants Elana, Lorraine and Christian. For their financial
contributions 1 would like to thank Forest Renewal British Columbia, The Science
Council of British Columbia, Grant Smith at Canfor, and Terry Kuzima at Carrier
Lumber - your contributions were greatly appreciated. 1 would also like to thank Richard
Kabzems, my external reviewer, Mike lull and Joe Ackerman for your careful reviews
and important perspectives. Your participation greatly improved the quality of this
research!
Thanks to Mom, Dad and Tim for many enjoyable days in the woods.

XII

To Edward and Andrina Newbery

xin

PREFACE
“The outstanding scientific discovery of the twentieth century is not television, or radio,
but rather the complexity of the land organism. Only those who know the most about it
can appreciate how little is known about it. The last word in ignorance is the man who
says o f an animal or plant: ‘what good is it?’ If the land mechanism as a whole is good,
then every part is good, whether we understand it or not. If the biota, in the course of
æons, has built something we like but do not understand, then who but a fool would
discard the seemingly useless parts? To keep every cog and wheel is the first precaution
of intelligent tinkering.” Aldo Leopold, 1966.

The practice of forest management in British Columbia is evolving from
traditional ideas of sustained yield management to ecosystem management approaches as
suggested above by Aldo Leopold. With this paradigm shift, strategies are being
developed to maintain the complexity of the land organism and thus maintain healthy
ecosystems. It has been proposed that one way of maintaining healthy ecosystems is to
use harvest patterns that mimic natural disturbance regimes so as to provide for more
natural levels o f stand and landscape level diversity.
Secondary succession in wet sub-boreal, spruce-fir forests in central British
Columbia is thought to be initiated by catastrophic fire, with return intervals ranging 2276, 250 years. Due to the relative infrequency of fire in this area compared to drier forest
ecosystems, small-scale disturbance agents that cause mortality of individual or small
groups of trees can be important processes affecting succession. In areas where small
disturbances are known to be a predominant successional mechanism, partial cut
harvesting is used to mimic the natural disturbance regime. However, little is known
about small-scale disturbance regimes and associated forest dynamics in the sub-boreal,
forests of western North America. Without precise information on these attributes, we
may not be able to develop partial cut harvesting protocols that mimic the natural smallscale disturbance regime. An important small-scale disturbance agent in wet, sub-boreal.

spruce-fir forests is Inonotus tomentosus (Fr.) Teng. The root disease caused by this
fungus can spread to adjacent trees and inoculum can remain active in stumps and roots
for decades. These characteristics may create important differences in disturbance regime
and stand development compared to other disturbance agents {e.g. bark beetles and
windthrow). Moreover, L tomentosus caused mortality may be exacerbated by partial cut
harvesting because inoculum could be transferred more readily from infected stumps to
new regeneration and residual trees. Given the need for partial cut guidelines in wet, subboreal forests and the potential interaction of /. tomentosus with partial cutting, there is a
need to quantify the disturbance regime and subsequent stand dynamics for forests
infected by /. tomentosus and those uninfected by /. tomentosus.
This thesis addresses three main questions for wet-cool sub-boreal forests located east
of Prince George, British Columbia:
1. What are the spatial and temporal patterns of small-scale disturbance for old-growth
and partial cut forests with and without the influence of I. tomentosus!
2. How does stand composition and structure differ between 7. tomentosus infeeted and
non-infected old-growth forests?
3. How does 7 tomentosus affect stand dynamics in partial cut forests and how does this
differ from its affects in old-growth forests?
These questions are addressed at both the tree and stand level by combining
information fi-om two investigative approaches that measure stand dynamics at different
spatial scales. In combination, these approaches provide a multi-scale understanding of
small-scale natural disturbance regimes, stand dynamics and succession in wet, subboreal forests. It is hoped that these studies will improve our understanding of small-scale
disturbances in sub-boreal forests, particularly the ecology of 7. tomentosus, and provide
a biological basis for 7. tomentosus management and partial cut silviculture.

CHAPTER 1. INTRODUCTION TO SUB-BOREAL STAND DYNAMICS:
LITERATURE REVIEW

1.1 Species Associations, Range and Climate
The sub-boreal forest region of North America is composed of a variety of forest
ecosystems transitional to the southern boundary of the boreal forest. The southern
boundary of the sub-boreal zone varies from east to west. In eastern North America, the
sub-boreal forest is bounded to the south by the deciduous forest region, beginning in
south-central Ontario and ranging into the northeastern United States (Rowe 1972). In
this area, sub-boreal forests are mixtures of eastern conifers (eastern white pine, red pine
and eastern hemlock), eastern hardwoods (yellow birch, sugar maple, red maple, red oak,
basswood, white elm, poplars, beech, white oak, butternut and white ash), and boreal
species (white and black spruce, balsam fir, jack pine, poplars and white birch) (Rowe
1972). In central North America, the sub-boreal forest is absent due to the abrupt
boundary between the boreal forest in the north and the grassland region of the Great
Plains to the south (Farrar 1995). In western North America the sub-boreal forest is
comprised of mixtures o f Picea glauca x engelmannii Voss x Parry (hybrid spruce,
hereafter referred to as spruce) and Picea mariana B.S.P. (black spruce), Abies
lasiocarpa Nutt, (sub-alpine fir, hereafter referred to as fir), Pinus contorta Dougl.
(lodgepole pine), Populus tremuloides Michx. (trembling aspen), Betula papyrifera
Marsh (paper birch), Populus trichocarpa Torr. & Gray (black cottonwood) and
Pseudotsuga menziesii Franco (Douglas-fir) (Pojar et al. 1982). Here, sub-boreal forests
are transitional to higher elevation forests and the boreal forest northward.
The sub-boreal climate is slightly less continental than the boreal (Table 1). It has
shorter winters, longer growing seasons, higher precipitation, and lower rates of

évapotranspiration. Generally, the average growth rates of trees are higher in sub-boreal
than boreal forests due to the slightly more favorable macroclimate (Pojar et al. 1982). In
British Columbia, large areas of the sub-boreal forest are classified as Sub-Boreal Spruce
(SBS) forests. The SBS forest region is located in central interior British Columbia
between 52° and 57° North latitude and 122° and 128° West longitude (Pojar et a/. 1982;
Meidinger and Pojar 1991) (Figure 1). This diverse zone is transitional to higher
elevation montane Douglas-fir and sub-alpine forests in the south and the boreal forests
in the north (Pojar et a/. 1982; Meidinger and Pojar 1991). Generally, the SBS is
dominated during late serai stages by spruce and fir. SBS forests typically occur at low to
medium elevations (500 to 1300 meters above sea level) on the gently rolling terrain of
the Fraser and Nechako plateaus and the Fraser basin (Pojar et a l 1982). The climate of
the SBS is broadly continental with seasonal extremes of long, cold, snowy winters and
short, warm, moist summers (DeLong et a l 1993). Mean annual temperature ranges from
1.7°C to 5°C and average temperatures are below 0°C for 4-5 months and above 10°C for
4-5 months of the year (Table 1) (Meidinger and Pojar 1991). The mean annual
precipitation is widely variable, between 440 mm and 900 mm, but extremes in
precipitation of 415 mm and 1650 mm have been recorded from short-term data (Table 1)
(Meidinger and Pojar 1991). A second sub-boreal zone also found in British Columbia is
the Sub-Boreal Pine-Spruce. This zones’ climate is slightly drier and cooler than the SBS
(Table 1) and lodgepole pine-engelmann spruce {Picea engelmannii) mixtures dominate
late serai ecosystems.

1.2 Succession
Stand Initiation
SBS forests originate by fire and the stand initiation stage is characterized by the
development of a single cohort of several tree species initially achieving high densities.
On most sites, shade intolerant species, generally faster growing and with regeneration
strategies that are fire adapted: lodgepole pine, paper birch, trembling aspen, spruce and
Douglas-fir, establish quickly while more shade tolerant species and those less fire
adapted: fir, western hemlock {Tsuga heterophylla Sarg.) and western red cedar {Thuja
plicata Dorm) fill in over time and gradually increase in abundance (Oliver and Larson
1996). Spruce, due to its moderate shade tolerance is also found in the understory of these
forests.
Stem Exclusion
Following stand initiation, the single cohort stand undergoes a period of density
dependent mortality called stem exclusion (Oliver and Larson 1996). Canopy closure and
limited resources prevent further understory initiation (germination) and vigorous
individuals out-compete weaker and smaller individuals resulting in their mortality and/or
suppression to subordinate canopy positions (Waring and Schlesinger 1985; Davis and
Johnson 1987; Oliver and Larson 1996). More shade tolerant trees, such as fir and spruce,
may survive in sub-canopy positions, however, this stage is characterized by high
mortality rates for all species regardless of shade tolerance (Kneeshaw and Burton 1997).
Understory reinitiation
Towards the end of the stem exclusion stage, mortality starts to occur in the
overstory creating growing space for a second post-fire cohort. In the SBS a variety of
shade tolerant trees: fir (especially), spruce, and occasionally western red-cedar, and

western hemlock undergo a second period of germination and/or growth from advanced
regeneration (Oliver and Larson 1996). In some cases paper birch and Douglas-fir,
relatively shade intolerant species, are able to regenerate successfully. The understory
reinitiation stage continues to influence stand structure and composition in SBS forests
due to competition induced mortality and mortality caused by small and medium scale
disturbances mainly caused by Inonotus tomentosus, spruce beetle outbreaks
{Dendroctonus rufipennis) (Humphreys and Saffanyik 1993), windthrow and snow or ice
breakage.
Old-growth
Oliver and Larson (1996), define old-growth as the complete replacement of the
initial fire-origin cohort. This is achieved by the absence of stand replacing disturbances
and continued mortality of trees, due to small and intermediate scale disturbances
(Kneeshaw and Bergeron 1998). Spruce, and fir mixtures with a wide representation of
size and age classes typically characterize old-growth SBS forests. The time frame for
old-growth development is variable and the definition of old-growth given above may
only rarely be realized in the SBS. However, once a post-fire cohort begins to replace the
fire-origin cohort in the canopy, the development of an old-growth forest has begun.
In particularly wet SBS forests, fire may be so infrequent that true old-growth
forests are common place. Hawkes et a l (1997) reports that fire return intervals in very
wet and cool SBS forests near the Rocky Mountains likely range 1200-6250 years.
DeLong and Tanner (1996) reported intervals ranging 227-345 years in slightly drier SBS
forests in the foothills of the Rocky Mountains. In between these long return intervals,
small scale disturbances caused by root rot fungi (e.g. Inonotus tomentosus), wood
feeding insects (e.g. Dendroctonus rufipennis), tree life spans, and abiotic factors such as

windthrow and snow loading, kill individual or small groups of trees. These small
disturbances create old-growth forests and facilitate the maintenance and renewal of
forest structure in forests with long fire return intervals. Furthermore, it is acknowledged
that they are becoming important processes where fire suppression or climatic change has
excluded or reduced fire’s influence on the ecosystem (Clark 1994; Frelich and Reich
1995; Andison 1996; Kneeshaw and Bergeron 1998).

1.3 Small and Medium Scaled Disturbances in Sub-Boreal, Boreal and Sub-Alpine
Forests
In wet SBS forests, and other similar forest regions such as sub-alpine and eastern
North American and Asian (sub)-boreal forests, small scale disturbance agents are
important processes to forest dynamics and succession during the inter-fire period
(Veblen 1989; Lertzman 1992; Veblen et al. 1994; Frelich and Reich 1995; Kubota 1995;
Yamamoto 1995; Kneeshaw and Burton 1997; Kneeshaw and Bergeron 1998; Yong et al.
1998; Lewis and Lindgren 1999). However, knowledge about the patterns and processes
created by these small disturbances is limited, especially in northern interior British
Columbian forests. This is partly because of the previously held belief that SBS forests
had short fire return intervals that minimized the importance of small and intermediate
sized disturbances. However, a body of literature is beginning to demonstrate the
importance of small to medium sized disturbances to stand dynamics in many sub-alpine,
boreal and sub-boreal forest types.
In a Rocky Mountain sub-alpine fir forest in southern British Columbia, Veblen
(1986) reported that spruce beetle outbreaks averaged 200 hectares in size and return
intervals were 116.5 years. In a SBS forest near Smithers, British Columbia, Kneeshaw
and Burton (1997) reported that spruce beetle caused 5 - 90% mortality over areas

ranging from 1-6, 200 hectares. Another key disturbance in sub-boreal forests is Inonotus
tomentosus which reportedly caused 0-10% mortality in patches ranging in size from < 1
hectare to 100 hectares (Van Groenewoud and Whitney 1969; Merler et al. 1988).
Kneeshaw and Bergeron (1998) characterized the percent area in recently formed
gaps as a function of stand age for southeastern boreal forests in northwestern Quebec.
For young forests (65 years), 7.1% of the study area was occupied by gaps whereas old
forests (230 years) were found to have 40.4% of stand area in gaps. Mortality was mainly
attributed to overstory tree senescence and spruce budworm.
In a sub-alpine forest in central Japan, Yamamoto (1995) reported that of the total
stand area studied, 7.4% was oecupied by gaps, gap density equaled 17.2 ha

and mean

gap size was 43.3 m^, with the data highly skewed to smaller gap sizes. Of these gaps,
65.3% were caused by single tree deaths, with about 55% of these attributed to wind. In a
sub-boreal forest in northern Japan, Kubota (1995) reported different disturbance patterns
in four similar systems dispersed widely across their study area. Some forests were
affected by continuous small disturbances mainly due to wind, killing one to a few trees
in isolated areas. Other forests were affected periodically by slightly larger disturbances
such as typhoon.
A few studies have examined disturbanee across multiple scales (Veblen et al.
1994; Frelich and Reich 1995). In a southern boreal forest in Minnesota, Frelich and
Reich (1995) demonstrated the importance of scale in suceession. Five models of
succession were tested in their study: 1) Cyclic Model, succession begins at stand A then
evolves to stand B, then C, then D, and back to A (Figure 2a); 2) Convergent Model, two
dissimilar forests suceeed towards the same stand composition (Figure 2b); 3) Divergent
Model, stand A succeeds into two different stand compositions (Figure 2c); 4) Parallel

Model, stand A and stand B maintain their characteristics even after disturbance (Figure
2d); 5) Individualistic Model, the stand begins at point A and is driven to an array of
potential points depending on a variety of mechanisms and relationships (Figure 2e). At
small spatial scales, they reported the divergence hypothesis best described successional
patterns with small groups of trees (35 m^), of the same species, age, height, etc.,
developing from a homogeneous fire origin stand. At a larger scale (1-16 hectares)
succession lead to a convergent model (Figure 2b) with one stand comprised of mixtures
of several species with no patchiness pattern evident as in the small scale study (Frelich
and Reich 1995). In both cases, succession was driven by openings in the canopy 10-30
meters in diameter caused by wind, insects and disease, but the results clearly show how
different interpretations can be made depending on the scale examined. It is important to
realize that disturbance occurs across this hierarchy of scales, and single scale
disturbance investigations do not explicitly take this approach. Thus, the combined
interpretations of different studies will reveal the true outcome of disturbances operating
at multiple scales.

1.4 Stand Development in Sub-Boreal, Boreal and Sub-Alpine Forests
Stand development following a disturbance is contingent upon the nature of the
disturbance, the age of the stand, the health of the stand, abiotic factors such as soil
texture, biotic factors such as herbivory, and chance. However, some simplifications of
stand characteristics are possible even with generalizations regarding disturbance regime.
Throughout the regions discussed above, Abies spp. and Picea spp. dominate late
successional boreal and sub-boreal forests. Generally, Picea spp. has greater basal area
due to a greater number of larger trees hut Abies spp., is generally found in higher total

densities (about %:\,Abies:Piced) (Veblen 1986; Veblen et al. 1994; Frelich and Reich
1995; Yamamoto 1995; Kneeshaw and Burton 1997; Kneeshaw and Bergeron 1998; Jull
and Famden unpublished data).
In Picea spp.-Abies spp. forests horizontal stand structure has been reported to be
generally clumped with groups of individuals surrounded by open patches (Lewis and
Lindgren 1999) and there is a high degree of vertical canopy stratification in old forests
(Veblen 1986; Veblen et al. 1994; Frelich and Reich 1995; Yamamoto 1995; Kneeshaw
and Burton 1997; Kneeshaw and Bergeron 1998; Jull and Famden unpublished data).
Age class and diameter distributions range from uneven (exponential decay) (Kubota
1995; Jull and Famden unpublished data) to bimodal or irregular in old forests (Johnson
etal. 1994).
Mortality appears mainly in upper canopy positions in mature forests (Veblen
1986; Frelich and Reich 1995; Kubota 1995; Kneeshaw and Burton 1997; Kneeshaw and
Bergeron 1998; Jull and Famden unpublished data), and in most cases (but see Johnson
et al. 1994), recmitment from the understory occurs reliably, replacing lost stems and
maintaining canopy cover. There appears to be agreement that these ecosystems are selfmaintaining in the absence of fire, with small scale disturbances being a key process
allowing understory species to replace lost canopy trees. Generally, various mixtures of
Abies spp. and Picea spp. maintain the canopy. In some cases, Betula spp. (Jull and
Famden unpublished data', Frelich and Reich 1995; Kubota 1995; Yamamoto 1995;
Kneeshaw and Bergeron 1998), Tsuga spp., and Thuja spp. (Frelich and Reich 1995;
Yamamoto 1995) are maintained in small numbers due to the presence of suitable
germination substrate, gap environments or chance.

10

In most cases, it is thought that Picea spp. and Abies spp. are able to coexist
indefinitely in these forests due to differences in their mortality schedules and recruitment
strategies (Kneeshaw and Bergeron 1998). However, Johnson et al. (1994) inferred from
his results that a long period of Picea glauca exclusion during stem exclusion will not
allow it to be maintained without fire. Differences in shade tolerance and seed
germination may be responsible for these apparent differences in recruitment strategies.
Fir mortality rates reported by lull and Famden {unpublished data) from long-term
permanent sample plot data, increase linearly with diameter or age of the tree. Spruce
appears to have a U-shaped mortality curve, with slightly higher mortality rates for
sapling to small pole sized trees than fir. Spruce mortality decreases from pole (7 cm
diameter) to mature trees (55 cm diameter) then increases again as trees become
susceptible to pests and reach old age (Jull and Famden unpublished data). Fir survival is
higher than spmce for young trees while the opposite relationship holds for older trees.
Furthermore, fir regenerates and subsists in the understory in higher densities than spmce
due to greater shade tolerance and more successful germination on thick humus layers
(Jull et al. 1996). These recmitment-mortality relationships appear to maintain Abies
spp.- Picea spp. coexistence in many forests throughout the world.

1.5 Resource Dynamics in Gaps
The maintenance of these forests through understory establishment and
recraitment to the canopy is due largely to the above and below ground resource
dynamics resulting from recently dead or dying canopy trees. Light, nutrition, moisture
and temperature regimes are all affected by disturbance and these resources and
environments can become more abundant or favorable to potential recmits currently
n

living in or germinating in the gaps. In this thesis, resource availability in gaps is not
measured, thus the relationships between gap formation, resource availability and tree
growth cannot be examined. Based on previous research (Walters and Reich 1997)
understory trees are assumed to be limited by multiple resources, thus the increase in
growth of understory trees is attributed to increases in several limiting resources, as
described below.
Light
Canopy gaps change many physical ecosystem components, however the most
obvious effect is on light regimes. Canham (1989), found that with the exception of
forests with extreme canopy height to crown width ratios (common in Picea-Abies
forests), single tree-fall gaps significantly increase the quantity and quality o f light an
understory tree receives. However, the actual duration of this increased light is short in
northern latitudes even in large gaps (several trees), since increasing latitude results in
decreasing overhead light duration throughout the day (Canham 1989). In addition to this
effect, the light received falls north of the gap with increasing latitude. Therefore, the
quantity and quality of light is higher a few meters north of the dead canopy tree
(Canham 1989).
As mentioned, forests with extreme canopy height to crown width ratios tend to
reduce the total impact of light regimes in gaps. Tall trees with narrow crowns
characterize Ficea-Abies forests, but due to deep live crowns typical of spruce and sub­
alpine fir, the impact of increased light should be at least partially responsible for
increased understory growth rates in these forests (Kneeshaw and Burton 1997).
A number of studies have linked increased light with increased plant productivity
in the field and in vitro (Walters and Reich 1997). Regardless of variation in species
12

shade tolerance, growth rates are higher with greater light levels. Many studies have also
shown that trees will respond with increased height and diameter growth rates following
release from shade (Canham 1989; Veblen 1989; Yong et al. 1998). Canham (1989)
suggests that even single tree gaps increase light regimes as much as two-fold and that
this additional light can trigger strong growth releases especially in shade tolerant
species. Shade tolerant species are favored in these environments because they can
regenerate under low light, closed canopy environments and out-survive less tolerant
species at any given low light level. When a gap forms, they are able to respond quickly
by increasing leaf area (Canham 1989). Shade intolerant species follow a different mode
of establishment. These species colonize larger gaps germinating on mineral substrate in
high light and grow quickly into the newly available space (Canham 1989). Thus, the size
of a disturbance may favor a particular species of tree due to the gaps’ physical and
environmental characteristics and the trees’ ecophysiological adaptations.
Nutrients, Temperature and Moisture
While increased light may be the most obvious change in resources caused by gap
formation, nutrient availability, moisture, and temperature also increase in response to
gap formation (Picket and White 1985, Pritchett and Fisher 1987, Schmidt et al. 1988,
Walters and Coates, unpublished data). Thus increased seedling growth following gap
formation in SBS forests may be due to increases in light, moisture, temperature and
nutrient availability and not just increases in light alone

1.6 Inonotus tomentosus as a Cause of Small Scale Disturbance
One important cause of gap formation in wet SBS forests is Inonotus tomentosus
(Fr.) Teng (Merler et al. 1988; Lewis and Hansen 1991a; Lewis et al. 1992; Lewis 1997).
13

This root pathogen induces gradual decline in tree vigor and eventually mortality (Lewis
and Hansen 1991b) due to dysfunctional roots, or windthrow due to weakened root
systems (Lewis and Hansen 1991b; Lewis 1997). I. tomentosus generally produces
pockets of infection and mortality that are distributed throughout the forest (Lewis et al.
1992). Due to its slow colonization processes, mortality is particularly evident in older
forests (Lewis et al. 1992). This disease has a wide host range but Picea spp. suffers the
most damage (Lewis and Hansen 1991a; Meidinger and Pojar 1991; Hunt and Unger
1994). Other tree species such as lodgepole pine and Douglas-fir are moderately affected
by the disease and Abies spp. even less. Hardwoods appear to be immune (Whitney
1993^
I. tomentosus can spread from host to host in two known ways. Root contact
provides bridges for hyphae to move from infected to healthy trees (Lewis and Hansen
1991b). Basidiospores also spread the disease, and are thought to cause new infection
centers but the mechanism of this infection process is not known (Lewis and Hansen
1991b). Once a new host is infected, possibly beginning in small feeder roots, the fungus
accesses the root heartwood (Lewis et al. 1992). Its presence in the heartwood is
indicated by reddish brown stain proceeding to darker and thicker bands of stain and
finally to pockets of advanced decay characterized by longitudinal pitting (Lewis et al.
1992). In roots smaller than 2 cm in diameter, the fungal mycelium can be located in the
bark or the cambium. This difference may be due to the thinner bark and lack of defense
mechanisms in these small roots and the lack of significant heartwood accumulation
(Lewis eta/. 1992).
In large roots, the infection gradually weakens the trees’ support making it
susceptible to windthrow (Lewis and Hansen 1991b). The disease can infect a significant
14

portion of these large roots before reductions in growth occur, since the heartwood is not
involved with the transport of resources (Hunt and Unger 1994; Lewis 1997). Mortality
of the roots occurs when the sapwood becomes significantly decayed (Lewis et al. 1992)
or when structural integrity is degraded to the point where tree fall occurs. Whitney
(1980) speculates that mortality of the tree itself occurs when root mortality reaches 80%.
However, mortality is probably more common due to weakened roots, resulting in
windthrow (Whitney 1980). Following tree mortality, colonization of the root sapwood
occurs rapidly due to the inactivity of active defense mechanisms. In the large roots, the
decay will progress radially from the heartwood to the sapwood, thereby increasing
probability that inoculum will contact a new host (Lewis et al. 1992).
Since the fungus spreads through root-to-root contact and prefers Picea spp. as a
host, the gaps it forms may spread outward from the original infection center (Lewis et al.
1992). Therefore it can be hypothesized that Picea spp. may be eliminated from these
progressive gaps and resistant species, especially sub-alpine fir in SBS forests, may be
the species that is favored in the colonization of these gaps. Therefore, the disturbance
regime and subsequent stand development could potentially be quite different compared
to gaps and forests without the disease because it may reduce mature and regenerating
hybrid spruce in wet, cool SBS forests in central British Columbia.
This hypothesized effect of /. tomentosus may have important forest health
implications in both managed and unmanaged forests. For example, the choice of harvest
systems in British Columbia is guided by the principle of ecosystem management (British
Columbia Ministry o f Forests 1995). One of the central tenants of this management
philosophy is that imitating natural disturbance patterns with harvesting patterns will aid
in achieving sustainable forest development (British Columbia Ministry of Forests 1995).
15

Due to the small scale disturbances that are common in wet SBS forests, partial eut
silviculture systems might be appropriate tools for forest management because they may
more closely mimic the natural disturbance process between long fire return intervals
(Davis and Johnson 1987; Oliver and Larson 1996; Jull 1997). However, applying partial
cut silviculture systems to sites where root disease is present may cause the disease to
spread faster, due to a hypothesized increase in inoculum volume that may occur in the
roots and stumps of harvested trees. This is known to occur with other root diseases such
as Armillaria ostoyae and Heterobasidion annosum (Koening 1969; Cruickshank et al.
1997; Garbelotto et al. 1997), but the effect of harvesting on inoculum volume for I.
tomentosus is unknown. This result was not supported by Whitney (1993) who reported
that partial cut forests of Picea glauca had a lower proportion of individuals infected with
I. tomentosus than unthinned forests. He proposed various reasons for this including
discontinuity o f spruce roots preventing root to root contact and increased due to more
abundant resources following thinning.
It can be hypothesized that a combination of several factors may lead to increased
infection and mortality of residual trees. The acceleration of disease-related mortality
might be caused by harvesting infected trees, which when cut, lose the ability to actively
limit root colonization, and therefore inter-tree spread. Furthermore larger trees are
typically favored over smaller diameter trees in most partial cut or selective systems
including diameter limit, stand improvement, seed tree, shelter wood, group selection,
and single tree selection systems (Smith 1986). Since these larger trees are more likely to
be infected with I. tomentosus (Lewis 1997) the effect of the cut may increase the
probability of residual tree or regenerating tree contact with mycelium in an infected
stump, and potentially, mortality rates and/or gap demography may be altered. Thus, it is
16

critical to understand the impacts of partial cutting on the incidence of I. tomentosus
caused spruce mortality and its implications for stand structure.

1.7 Research Questions
This research addresses two main questions associated with small scale
disturbances in old-growth and partially cut forests. First, does small-scale disturbance
and forest dynamics differ between Inonotus tomentosus infected and uninfected forests?
The answers to this question will enhance our understanding of how old-growth SBS
forests developed to their present condition and may help predict future forest conditions.
This information would aid resource managers in understanding how well old-growth
forests will continue to provide the values for which they were preserved. Furthermore,
this study will provide quantitative information necessary for forest managers to design
partial cut systems that mimic old-growth dynamics and provide information on the long­
term sustainability of small-scale forest dynamics in old-growth SBS forests. The second
question is; Does partial cutting increase the incidence of /. tomentosus compared to
uncut forests and if so, what implications does this have for differences in stand dynamics
between infected and uninfected forests? This information could provide guidelines for
partial cutting in I. tomentosus infected stands and predictions about future stand
conditions where partial cutting is applied. Therefore, the objectives of this thesis are to
quantify and compare fine and coarse spatial scale disturbance regimes and stand
dynamics in 7. tomentosus infected and non-infected stands for both partial cut and oldgrowth SBS forests.
These objectives are addressed in the following chapters. Chapter 2 describes the
development of a methodology used to help recreate a fine-scaled disturbance chronology
17

for unmanaged and partially cut forests with and without I. tomentosus. This new method,
called the Time Since Death model (TSD model) estimates when a tree died based on its
state of decomposition.
Chapter 3 is a second methodology study developed to utilize tree ring growth
rate patterns in canopy trees to determine the date a tree ascended to the canopy. Chapter
4 summarizes the gap-scale disturbance regime of wet SBS forests. This study combines
the methodologies developed in Chapters 2 and 3 to recreate the disturbance dynamics
occurring in four forest stand types (/. tomentosus infected, and uninfected old-growth
forests and I. tomentosus infected and uninfected partial cut forests). Basic stand
composition and attribute data are also sampled to compare these data across forest stand
types. Chapter 5 summarizes the stand scale study. This study uses spatial autocorrelation
analysis of tree ring data and species association to determine differences in the canopy
level disturbance regime and canopy patch structure for I. tomentosus infected and
uninfected old-growth forests using six 0.49 ha plots with 7 meter resolution. Chapter 6
integrates the results of the fixed radius plot and grid plot studies and draws conclusions
regarding disturbance regimes, stand development, and implications for forest
management.

18

1.8 LITERATURE CITED

Andison, D. 1996. Managing for landscape patterns in the sub-boreal spruce forests of
British Columbia. Ph.D. dissertation. Faculty of Forestry, University of British
Columbia. Vancouver, BC. 178 pp.
British Columbia Ministry of Forests. 1995. Biodiversity Guidebook. Crown
Publications. Victoria, BC. 116 pp.
Canham, D.D. 1989. Different responses to gaps among shade tolerant species. Ecology
69:548-550.
Clark, D.F. 1994. Post-fire succession in the sub-boreal spruce forests of the Nechako
Plateau, central British Columbia. M.Sc. thesis. Department of Biology,
University o f Victoria. 198 pp.
Cruickshank, M.G., Morrison, D.J. and Punja, Z.K. 1997. Incidence of Armillaria species
in precommercial thinning stumps and spread of Armillaria ostoyae to adjacent
Douglas-fir trees. Can. J. For. Res. 27:481-490.
Davis, L.S. and Johnson, K.N. 1987. Forest Management. Third Edition. McGraw-Hill.
Toronto, Ontario. 790 pp.
DeLong, C., Tanner, D. and Jull, M.J. 1993. A field guide for site identification and
interpretation for the southwest portion of the Prince George Forest Region.
Ministry of Forests Research Program. Victoria BC. 311 pp.
DeLong, S.C. and Tanner, D. 1996. Managing the pattern of forest harvest: lessons
from wildfire. Biodiversity and Conservation 5:1191-1205.
Farrar, J.L. 1995. Trees in Canada. Fitzhenry and Whiteside Limited and the Canadian
Forestry Service, Natural Resources Canada. Markham Ontario.
Frelich, L.E. and Reich, P.B. 1995. Spatial patterns and succession in a Minnesota
southern-boreal forest. Ecological Monographs 65:325-346.
Garbelotto, M., Slaughter, G., Popenuck, T., Cobb, F.W. and Bruns, T.D. 1997.
Secondary spread of Heterobasidion annosum in white fir root disease centers.
Can. J. For. Res. 27: 766-773.
Hawkes, B., Vasbinder, W. and DeLong, S.C. 1997. Retrospective fire study: fire
regimes in the SBSvk & ESSFwk2/wc3 biogeoclimatic units of northeastern
British Columbia. McGregor Model Forest Association. Prince George, BC. 34
pp.
Humphreys, N. and Safranyik, L. 1993. Forest pest leaflet: spruce beetle. Forestry
Canada, Forest Insect and Disease Survey. Pacific forestry Center, Victoria, B.C.
19

Hunt, R.S. and Unger, L. 1994. Forest pest leaflet: Tomentosus root disease. Forestry
Canada, Forest Insect and Disease Survey. Pacifie forestry Center, Victoria, B.C.
Johnson, E.A., Miyanishi, K. and Kleb, H. 1994. The hazards of interpretation of static
age structures as shown by stand reconstruction’s in a Pinus contorta - Picea
engelmannii forest. J. Ecol. 89:923-931.
Jull, M.C., DeLong, S.C., Eastman, A., Sagar, R.M., Stevenson, S. and DeLong, R.L.
1996. Testing silvicultural systems for the ESSF: early results of the Lucile
Mountain Project. BC Ministry of Forests, Prince George Forest Region Research
Note #PG-01, March 1996.
Jull, M.C. 1997. Interior British Columbia silvicultural systems: crumbling myth and
emerging reality. RPF Forum 4(5):5-6.
Kneeshaw, D.D. and Bergeron, Y. 1998. Canopy gap characteristics and tree
replacement in the southeastern boreal forest. Ecology 4(3):783-794.
Kneeshaw, D.D. and Burton, P.J. 1997. Canopy and age structures of some old
sub-boreal Picea forests in British Columbia. Journal of Vegetation Science
8 : 615 - 626 .

Koening, J.W. 1969. Root rot and chlorosis of released thinned western red cedar. Jour.
For. 67: 312-315.
Kubota, Y. 1995. Effects of disturbance and size structure on the regeneration process in
a sub-boreal coniferous forest, northern Japan. Ecological Research 10:135-142.
Leopold, A. 1986. A Sand County Almanac: With Essays of Conservation from
River. Ballantine Publishing Group.

Round

Lertzman, K.P. 1992. Patterns of gap-phase replacement in a sub alpine, old-growth
forest. Ecology 78(2):657-669.
Lewis, K.J. 1997. Growth reduction in spruce infected by Inonotus tomentosus in central
British Columbia. Can. J. For. Res. 27:1669-1674.
Lewis, K.J. and Hansen, E.M. 1991a. Survival of Inonotus tomentosus in stumps and
subsequent infection of young forests in north central British Columbia. Can. J.
For. Res. 21: 1049-1057
Lewis, K.J. and Hansen, E.M. 1991b. Vegetative compatibility groups and protein
electrophoresis indicate a role for basidiospores in spread on Inonotus tomentosus
in spruce forests o f British Columbia. Can. J. Bot. 69:1756-1763.
Lewis, K.J. and Lindgren, B.S. 1999. Influence of decay fungi on species composition
and size class structure in mature Picea glauca x engelmannii and Abies

20

lasiocarpa in mature sub-boreal forests of central British Columbia. Ecology and
Management 123:135-143.
Lewis, K.J., Morrison, D.M. and Hansen, E.M. 1992. Spread o ïInonotus tomentosus
from infection centers in spruce forests in British Columbia. Can. J. For. Res.
22: 68-72.
Meidinger, D. and Pojar, J. 1991. Ecosystems of British Columbia. Research Branch,
Ministry of Forests. Crown Publications Inc. 330 pp.
Merler, H.A., Shulting, P.J. and VanderKamp, B. 1988. Inonotus tomentosus (Fr.) Teng
in central British Columbia. In Proceedings of the 7th International Conference
on Root and But Rots, Vernon and Victoria , B.C., August 9-16, 1988. lUFRO
working party s2.06.01. Edited by D. Morrison. Forestry Canada, Pacific
Forestry Center, Victoria, B.C. 985 pp.
Oliver, C.D. and Larson, C.B. 1996. Forest Stand Dynamics. John Wiley and Sons.
Toronto. 527 pp.
Pickett, S.T.A. and White, P.S. (eds). 1985. The ecology of natural disturbances and
patch dynamics. Academic Press, Toronto, Ontario. 472 pp.
Pojar, J., Love, R., Meidinger, D. and Scagel, R. 1982. Some common plants of the
Sub-Boreal spruce Zone. Ministry of Forests Research Branch. Victoria, BC.
95 pp.
Pritchett, W.L. and Fisher, R.F. 1987. Properties and Management of Soils. 2"'* ed.
John Wiley and Sons, Toronto, Ontario.
Rowe, J.S. 1972. Forest Regions of Canada. Department of the Environment. Canadian
Forestry Service Publication No. 1300.
Sagar, R.M. 1993. Aleza Lake, British Columbia AES Climate Station Data 1952-1980:
data summary and users guide. Unpublished report prepared by R.M. Sagar
Consulting Ltd. for the Prince George Forest Region. BC Ministry of Forests,
Prince George, BC. 18 pp.
Schmidt, M.G., Ogden, A.E. and Lertzman, K.P. 1988. Seasonal comparison of soil
temperature and moisture in pits and mounds under vine maple gaps and conifer
canopy in a costal western hemlock forest. Can. J. Soil. Sci. 78(2):291-300.
Smith, D.M. (ed). 1986. The practice of silviculture. John Wiley and Sons, Inc. Toronto,
Ontario
Van Groenewoud, H. and Whitney, R.D. 1969. White spruce mortality in Saskatchewan
and Manitoba. Pulp Pap. Mag. Can. April, pp. 1-4.

21

Veblen, T.T. 1986. Treefalls and the co-existenee of eonifers in sub-alpine forests of the
central Rockies. Ecology 67(3):644-649.
Veblen, T.T. 1989. Tree regeneration responses to gaps along a transandean gradient.
Eeology 70:541-543.
Veblen, T.T., Hadley, K.S., Nel, E.M., Kitzberger, T., Reid, M. and Villalba, R. 1994.
Disturbance regime and disturbance interactions in a Rocky Mountain sub-alpine
forest. Journal of Ecology 82:125-135.
Walters, M.B. and Reieh, P.B. 1997. Growth o f Acer saccharum seedlings in deeply
shaded understories of northern Wisconsin: effects of nitrogen and water
availability. Can. J. For. Res. 27:237-247.
Waring, R.H. and Schlesinger, W.H. 1985. Forest ecosystems: concepts and
management. Academic Press. Toronto.
Whitney, R.D. 1980. Polyporous tomentosus root and butt rot of trees in Canada.
Proceedings of the 5th International Conference on Problems of Root and Butt
Rot in Conifers, Aug. 1978, Kassel, Germany. Edited by L. Pimitri.
Whitney, R.D. 1993. Inonotus tomentosus in Ontario, and the effects of thinning on the
disease. For. Chron, 69(4):445-449.
Yamamoto, S. 1995. Gap characteristics and gap regeneration in sub-alpine old-growth
coniferous forests, central Japan. Ecological Research 10:31-39.
Yong, B., Huacheng, X., Bergeron, Y. and Kneeshaw, D.D. 1998. Gap regeneration of
shade-intolerant Larix gmelini in old-growth boreal forests of northeastern China.
Journal of Vegetation Science 9:529-536.

22

Table 1. Climatic comparisons for various sub-boreal, sub-alpiue and boreal forests
present in British Columbia. Numbers in parenthesis denote extremes taken from short
term weather data. Derived from Meidinger and Pojar (1991) and Sagar 1993.
Forest

Mean Annual

Average

Average

Mean Annual

Percent

Zone

Temperature

Months

Months

Precipitation

Precipitation

Range (°C)

Below 0 °C

Above 10

(mm)

as Snow

°C

4 40(415)Sub-Boreal

1 .7 -5

4 -5

2 5 -5 0

2 -5
900(1650)

Spruce
Aleza Lake
682.4 Research

0 .9 -4 .9

N/A

N/A

38 %

1315.3
Forest

Sub Boreal
0 .3 -2 .7

4 -5

1 -3

335 - 580

3 0 -5 0

2.9 - 2

5 -7

2 -4

3 3 0 -5 7 0

35-55

2-2

5 -7

400 - 2000

50 - 70

PineSpruce

-

Boreal

-

Sub-Alpine

23

L '/
1

..

/V

1

Dawson

'r ? ^ r m c o
^ Rupe

CS/ %\

Aleza Lake
Research Forest

.w illiam s Ç.Î-. /
y

V. ,

V
(

mu1avi
ÿÇVtc^^ria

U .S.A

Figure 1. Location of Sub Boreal Spruce biogeoclimatic zone and the Aleza Lake
Research Forest in central interior British Columbia (Government of British Columbia,
Ministry of Environment Lands and Parks and Ministry of Forests 2001).
24

A

a)

b)

A

d)

A

B

D

\ C/
c)

B

C

B

e)

D

G
Figure 2. a) Cyclic model of succession, b) Convergent model of succession, c) Divergent
model of succession, d) Parallel model of succession, e) Individualistic model of
succession. Each arrowed line represents a disturbance and the state the stand reaches due
to succession (comes from Frelich and Reich 1995)

25

CHAPTER 2. ESTIMATING TIME SINCE DEATH IN PICEA GLAUCA x
ENGELMANNII NHO ABIES LASIOCARPA IN A WET-COOL SUB-BOREAL
SPRUCE FOREST

2.0 ABSTRACT
Age class analysis, direct gap measurements, and dendrochronology, are
retrospective methods used to quantify disturbance regimes in forest ecosystems, but each
method has limitations. This paper presents a new method that helps overcome some of
the limitations current methods have in disturbance regime studies. The method is a set of
multiple regression models that estimate the year of death for Picea glauca x
engelmannii and Abies lasiocarpa from the tree characteristics: species, position
(standing/down), decay class, proportion of decay, bark presence and integrity, primary
branch presence and integrity, and fine branch presence and integrity.
The model was developed from a sample of 183 trees with known dates of death
from 0-70 years before present (±2.5 for periods 1926-1963, 1988-1998 and ±12.5 yrs for
period 1963-1988) determined from permanent sample plot data obtained from the Aleza
Lake Research Forest, in east central British Columbia, in the wet-eool foothills of the
Rocky Mountains. Four models were developed, two for each species based on their
current position (standing or down). The P. glauca x engelmannii model explained
87.3% and 75.3%, and the A. lasiocarpa model explained 76.8% and 84.7% of the
variation in years since death for standing and down trees, respectively. On an
independent sample of dead trees (N = 48), time since death was estimated with the
parameterized model. These values were compared to year of release determined from
tree ring cores in understory trees that were subordinate to the modeled dead trees. The
two estimates were strongly related (R^= 92.4%), indicating that model estimates provide

26

acceptable estimates for year of death in the two species. Other benefits the models
provide are a quick (yet accurate) field approach for estimating year of death, and they
can be used to determine year of death for small trees that fail to cause growth increases
in understory trees and for trees that do not have subordinate individuals that release
following overstory mortality.
In this study, P. glauca x engelmannii and A. lasiocarpa decayed at similar rates
but these species were found to decay much faster than Thuja plicata in coastal
rainforests and slower than Abies (spp.) and Pinus (spp.) in the Sierra Nevada. Climate,
log size and species may be important sources of variation in these results.

27

2.1 INTRODUCTION
The study and quantification of small-scale disturbance regimes is difficult
without extensive long-term documentation of the scale and timing of these disturbances
as they occur. Due to the paucity of these real time assessments, a variety of retrospective
techniques have been developed. One technique, the interpretation of static age-class
distributions, is generally not appropriate in forests of complex age structure because
different species o f trees have different recruitment and mortality schedules. These
differences can lead to misconceptions about the disturbance regime (Johnson et al.
1994). Therefore, tree ring analysis or direct gap measurements are often used to quantify
the patterns of canopy level disturbance (Veblen 1986; Lorimer and Frelich 1988; Frelich
and Lorimer 1991; Lertzman 1992; Veblen et al. 1994; Abrams et al. 1995; Frelich and
Reich 1995; Kubota 1995; Yamamoto 1995; Cherubini et al. 1996; Kneeshaw and
Bergeron 1998; Yong et al. 1998).
Tree ring analysis uses dramatic and sustained increased radial growth patterns,
minimum growth rate thresholds that indicate gap-origin probability, and other
interpretive patterns about incremental diameter growth to indicate the temporal and
spatial patterns of tree mortality. These techniques are useful in describing small to largescale canopy level disturbances, but are limited to situations where canopy mortality
causes an understory tree to increase markedly in growth rates. Since diameter growth is
a more sensitive measure of stand competition than any other growth index (Assman
1970), these techniques are considered reliable, but they require extensive local
calibration of growth rate criteria in order to assign a year of death to tree release criteria.
Furthermore, these methods may not be applicable where large increases in understory
resources might not accompany overstory tree death. The tall, narrow crowned boreal and
28

sub-boreal forests are one such example as are stands with low density such as might be
found on water-logged or very dry sites.
For direct measures of gap size, the percent area of stand disturbance is
calculated using geometric formulae. However, the resultant value must be restricted to
recent gaps; about 30 years old (Veblen 1986; Lertzman 1992; Kubota 1995; Yamamoto
1995; Kneeshaw and Bergeron 1998; Yong et al. 1998). Another problem with applying
this method are that the sampled gap-makers are usually restricted to canopy trees.
Therefore, the percent gap area is underestimated because it does not include sub-canopy
trees, which also contribute to gap dynamics. There also is limited ability to quantify
temporal and spatial gap dynamics with this approach. Therefore, it is unknown if the rate
of gap formation is constant, variable over time, randomly distributed, or has some
underlying spatial pattern. Perhaps the main limitation with this approach, at least in subboreal forests, is that gaps are not single definable features of the canopy or any other
layer. Therefore, it is nearly, if not entirely impossible to determine where one gap starts
and another one ends.
Due to limitations of these current methodologies, a study was initiated to develop
an alternative method for reconstructing stand dynamics. The method is based on
estimating time since death from a tree’s decomposition characteristics. To do this, decay
characteristics of dead Picea glauca x engelmannii (Parry ex Engelm.) (hybrid spruce,
hereafter referred to as spruce and Abies lasiocarpa (Hook.) Nutt, (sub-alpine fir,
hereafter referred to as fir) trees of known dates of death were collected from permanent
sample plots in central British Columbia. Then, multiple regression models using decay
characteristics as predictor variables and time since death as the dependent variable were
developed. Resultant Time Since Death models (TSD models) were then validated with
29

an independent data set where time since death was estimated by tree ring release of
suppressed trees subordinate to dead canopy trees.
Specifically the objectives of this study are to:
1. Construct a set of TSD models that accurately estimate the year of death for
trees using variables that are easily measured in the field to indicate the degree
of decomposition after death.
2. Assess the capabilities and limitations of the models.
3. Validate the models with independent samples of dead trees.
4. Compare the decay characteristics of spruce and fir, and species from other
ecosystems.

30

2.2 METHODS
2.21 Study Area
To develop a multiple regression model that estimates the time since death for
trees, a known and reliable date of death is needed. The date of death data was provided
for trees from long-term growth and yield sample plots at the Aleza Lake Research Forest
(ALRF) (British Columbia Ministry of Forests, Inventory Branch). The ALRF is located
at 54° 07’ N, 122° 04’ W, about 60 kilometers east of Prince George, British Columbia,
Canada. It lies between 600 and 750 meters above sea level on the Nechako plain of the
Fraser River Basin in the Interior Plateau physiographic region (Holland 1976). The
ALRF is located in a wet, cool type of the Sub-Boreal Spruce (SBS) biogeoclimatic zone,
and is classified as the SBSwkl according to a biogeoclimatic system in common usage
in British Columbia (see Meidinger and Pojar, 1991, for details). The SBSwkl climate is
characterized by cold, snowy winters and moist, cool summers. The climate is slightly
less continental than typical for more westward SBS areas due to the orographic
influence of the Northern Rocky Mountains to the east, resulting in higher precipitation
than usual for the rest o f the zone (Meidinger and Pojar 1991).

2.22 General Approach
In 1928, a series of growth and yield plots were established at the ALRF and trees
greater than 10 cm diameter were stem-mapped and tagged for future identification.
Diameter, species, and height were measured on the trees in these plots beginning in
1928, and remeasured every 5 years until 1963. A shift in government policy prevented
further remeasurements until 1988, with the most recent occurring in 1998. Observational
data on the predictor variables were collected for model development on a sample of 192
31

trees with known dates of death determined from ALRF growth and yield records in six
Experimental Plots (106,107,112, 148,149, 150). These variables were chosen from a
review of previous studies, which found relationships between certain components of tree
morphology and decomposition rates (Henry and Swan 1974; Grier 1978; Graham and
Cromack 1982; Sollins 1982; Raphael and Morrison 1987; Carpenter et al. 1988; Kelsey
and Harmon 1989; Harmon et al. 1994; Daniels et al. 1997).
The final measurement of a tree in the permanent sample plot records at the
ALFR indicates it died somewhere during the remeasurement interval, which is inferred
to be the mid-point of the remeasurement interval. Therefore, there are records of when
trees died ± 2.5 years for the 5 year remeasurement intervals and ± 12.5 years for the one
interval between 1963 and 1988. Trees which died many years ago were difficult to
locate. Old stem-maps and metal detectors were used to locate the metal tags that were
used to identify trees in the permanent sample plots.

2.23 Explanatory Variables
Tree species: Studies have shown that species vary in decomposition rates and
this variation is related to differences in phenol concentrations (Daniels et al. 1997),
agents of mortality, and post-mortality decay organisms (Carpenter et al. 1988; Kelsey
and Harmon 1989; Harmon et al. 1994). The dominant species at the ALRF are spruce
and fir. All other species are uncommon in mature forests, usually representing less than
5% of stand composition. Therefore, data were collected only for spruce and fir. Previous
studies (Daniels et al. 1997) indicated that modeling should be species specific, so
models were developed independently for spruce and fir.

32

Tree position: Whether a tree remains standing or falls down after it dies and how
long it has been in each position effects exposure to physical weathering and access to
decay organisms, with downed trees decomposing faster (Carpenter et al. 1988). Trees
were recorded as either standing (1) or down (2). It was not known how long the tree
remained standing after it died because the permanent sample plot records have only
recently included this attribute. Thus, only the current position of the tree was recorded
and no attempt was made to project the length of time in each position.
Proportion o f Decay: Trees rot from both the bark towards the pith and from the
pith towards the bark depending on species and presence of decay organisms (Daniels et
al. 1997). However, dead sapwood is generally less resistant to decay than the heartwood
and trees do not generally decay from the outside, until they die (Kelsey and Harmon
1989). Therefore depth o f decay on the outer bole is proportional to how long sapwood
decay has been occurring. Using the methods of Daniels et al. (1997) the depth of decay
from the bark towards the pith was recorded, using an axe to chop into the wood and a
ruler to measure the depth to sound wood (cm). When the entire radius was decayed, it
was recorded as half of the diameter of the tree. This was done at several points within
the first two meters from the roots of the tree to develop an average depth o f decay.
Proportion of decay was calculated by dividing the average depth of decay by one-half of
the tree’s diameter. Depth of decay varies with tree size, therefore proportion of decay
was used instead of depth of decay because this variable incorporates the difference in
decay rate due to tree size.
Decay Class: Decay class is a qualitative, categorical index based on cumulative
decomposition of the main bole. The density of the bole was estimated at several points
within the first 10 meters from the roots of the tree as suggested by Daniels et al. (1997).
33

Decay Class (DC) I trees are solid with no deeay in the sap or heartwood. DC II trees are
solid but have some preliminary signs of decay. DC 111 trees are in stages of advaneed
decay but the wood still has structural integrity and can still support its own weight. DC
IV trees are completely rotten, able to be kicked apart and cannot support their own
weight. DC V trees are similar to DC IV trees in structure but are assimilated into the
duff and have coniferous or deciduous vegetation well established and roots ramified
throughout.
Primary Branches: Primary branches were assessed as two separate variables,
branch presence and branch integrity. Branch presence had four categories: (1) all
branches remaining; (2) partial branch remains; (3) only broken stubs and; (4) no primary
branches. Branch integrity had 3 categories: (1) branches cannot be moved in the knot;
(2) branches are loose in the knot; and (3) no primary branches.
Fine Branches and Needles: Fine branches were also assessed as two variables:
fine branch presence and fine branch integrity. Fine branch presence had three categories:
(1) trees with 75-100% o f original branches; (2) trees with 1-75%; and (3) no fine
branches remaining. Fine branch flexibility had three categories (1) pliable, (2) brittle, or
(3) none. Needle abundance had the same categories as fine branch presence.
Bark: Bark is also important for estimating recent mortality (Raphael and
Morrison 1987). Bark was assessed as two variables: Bark integrity: (1) tight, (2)
variable, (3) loose and (4) missing; and bark presence (1) present, (2) variable, and (3)
missing).

34

2.24 Analysis and Modeling
Years since death and proportion of decay were treated as continuous variables.
Tree species was treated as nominal and the remaining variables were treated as ordinal.
For the two continuous variables, a Shapiro-Wilk W test was used to test for normality.
The nominal and ordinal variables were assessed for normality using a Chi-square
likelihood ratio test based on hypothesized probabilities totaling one with equal
probability distributed across the categories. The distributions for the variables were non­
normal. Due to the large sample size; N = 69 for spruce and, N = 114 for fir, and because
regression is robust to non-normality, parametric methods were used (Lewis - Beck
1980).
Outliers were examined by assessing residual plots of each independent variable
against the predicted values of years since death and with normal quantile plots of the
studentized residuals. Leverage values were calculated to determine overly influential
data points. Nine trees out of 192 total trees were omitted from the analysis due to large
residuals (1.5 * interquartile range), which are likely due to a measurement error or errors
in the ALRF data set.
A number of variables included in the analysis had significant heteroscedasticity
(unequal variance in error terms). For both models, heteroscedastic error terms were
indicated by the Brown-Forsythe test (Fox 1991). For spruce, bark class, and bark
integrity had significant heteroscedasticity. For fir, density, branch integrity, and bark
integrity had significant heteroscedasticity (a = 0.05). For heteroscedastic variables,
attempts were made to improve the condition by creating new variable categories. None
of these attempts decreased heteroscedasticity that was due to high variation in the final
category relative to preceding categories, since the final category is applicable to infinite
35

times since death. However, including variables with heteroscedastic error terms is
justified since the goal was not to develop a parsimonious model but to develop a robust
model for predicting time since death. Eliminating variables with not statistically
significant partial slopes (resulting from the heteroscedasticity) or conducting best sub­
sets regression was not done because all explanatory variables are required to avoid
specification error and maximize the coefficient of multiple determination. For example,
eliminating all the variables in the down spruce model with not statistically significant
partial slope coefficients would lower the

from 0.753 to 0.680. An

difference test

comparing the two R^ values indicates this is a significant (p>0.05) loss in explanatory
power.
The variables were initially screened using bivariate regressions and correlation
matrices, to examine linearity and multicollinearity. Needle presence, fine branch
presence, and primary branch presence were eliminated from analysis due to very poor
correlation (R^ < 0.05) with years since death and high collinearity between them.
Interactions between the position of the tree and the remaining explanatory
variables were tested. These analyses were performed by plotting each variable for each
position and comparing the slopes using standard procedures (Fox 1991). Significant
interactions were found and justified the splitting of each species into a positional model
(standing and down).
Ordinary least squares regression was then used to develop each model. The
models were then externally validated by estimating time since death on an independent
sample (dates of death unknown) of previously dominant or co-dominant dead trees (n =
48, 27 spruce, 21 fir). The decomposition variables were collected and increment cores
were taken from one nearby understory tree for each dead tree. Estimated dates of death
36

from the TSD model were regressed on dates of understory tree release (for trees
releasing at least 50%) determined from the increment cores using ordinary least squares
regression.

37

2.3 RESULTS
For both species, all variables had a significant linear relationship to years since
death (Table 1) as determined by bivariate linear regression analysis. It appeared that
proportion of decay had a non-linear relationship with years since death, however, no
transformation improved the

value. Therefore these data were considered to be best

described by a linear model. For both species, years since death was most strongly related
to decay class (Table 1). Since position is a binomial category, linearity cannot be
evaluated, however, for both species the trend indicates that downed logs are older than
standing logs.
A correlation matrix (Table 2) shows relationships among the independent
variables. For both species, decay class is highly correlated with proportion of decay and
bark class is highly correlated with bark integrity. This is because each pair of variables is
measuring similar attributes; the degree of decomposition on a tree’s main bole and the
bark retention on the tree, respectively.
To determine the degree of influence multicollinearity had on the models, variable
inflation factor analysis was conducted. Variable inflation factors (VIF) are equal to 1/(1R^), where the R^ is equal to the coefficient of determination for one explanatory variable
regressed against the other explanatory variables. The higher the regression coefficients
between some set of explanatory variables, the higher the VIF (StatSoft 1997). Thus, a
high VIF (>9) indicates that collinearity is strongly affecting the precision of estimation
(Fox 1991). For both spruce and fir, no VIF exceeds 9 for all possible combinations,
therefore the multicollinearity is not extreme (StatSoft 1997) which justified keeping
highly correlated predictor variables in the model.

38

Interactions between position and all the remaining explanatory variables were
assessed independently for each species by plotting each predictor variable against years
since death at the two levels of position. For spruce (p = 0.009) and fir (p = 0.047) a
significant interaction between position and decay class was found indicating that for
both species, decomposition rates of the main bole are faster in down trees. These
interactions were justified by separating each species model into two models based on
position.
Thus four models, ((2) species x (2) position) were generated, each using the
following predictor variables: decay class, branch integrity, fine branch flexibility, bark
class, bark integrity, and proportion of decay. The mixed model for each species (Table
3a and 3b; Figures 1 and 2) indicates that the explanatory variables explain a significant
proportion of the variation in years since death, and range from a low of
the down spruce model to a high of

= 75.30% for

= 87.30% for the standing spruce model.

2.31 Model Validation:
Using an independent sample, the year of death for 48 dead trees (27 spruce and
21 fir) was estimated with one of the four TSD models and then death dates were
compared to year of tree ring release measured on near by understory trees with ordinary
least squares regression (Figure 3). The two estimates of tree mortality were very close
(R^ = 92.4%). Note the regression line does not intersect the x and y axis in a perfect 1:1
relationship. For older mortality the regression line falls slightly above the x-y
intersections and slightly below for recent mortality. This is likely because of continual
increasing variance in the TSD model estimate with increasing values of the tree ring
estimate as seen in the data (Figure 3).
39

2.4 DISCUSSION
The TSD models developed from dead tree characteristics accurately predicted
time since death for spruce and fir in the wet SBS forests of central interior British
Columbia. The results suggest that the TSD models can be used to estimate the year of
death for spruce and fir with a minimum diameter of 10 cm DBH up to about 70 years
since death. These estimates can be applied towards describing disturbance history in the
SBSwkl forests around Aleza Lake.
The relative importance of the explanatory variables can be attributed to their
“longevity of measurability”. For instance, fine branches fall off the tree quickly
following mortality. Thus, they are irrelevant in predicting mortality beyond the time they
disappear yet, to remove them from the model would eliminate the most sensitive
variables in estimating recent mortality. This suggests that variables which are sensitive
to recent mortality as well as older mortality are most useful in terms of their predictive
capability. With respect to the variables examined here, decay class, bark integrity, bark
class and proportion of decay would be the most widely useful predictors of years since
death because they blend the ability to be sensitive to recent as well as past mortality.
The heteroscedasticity present in these variables is caused by larger variance in
the last category o f each variable relative to the preceding categories. For example, the
condition o f no bark class had the widest variation relative to intact, variable and notintact bark class. This is because once the condition of no bark is reached the tree is
always described as a tree with no bark and therefore, a large variation in time since
death arises. The same arguments could also be made for the all the other variables. This
is a limitation in the measurement of these variables that cannot be adequately addressed
due to the finite nature o f the variables.
40

Due to the range in years since death encountered in this study, as well as the
increasing variance in the error term (heteroscedasticity) in the explanatory variables, this
model must be restricted to estimating death to about 70 years before present, which was
the approximate range in years since death encountered at the Aleza Lake Research
Forest. However even if the records at the ALRF did go back 100 years (or more), none
of the current predictor variables would be reliable for estimation since they all will
experience severe heteroscedasticity with increasing time since death. Radiocarbon
dating, bulk density and nutrient flux functions like those used by Grier (1978), Graham
and Cromack (1982), and Sollins (1982), could be used to extend the estimate.
This study’s findings coincide with a few studies which have also examined
morphological characteristics of decay and their relationship to years since death. It was
found that about 97% of the trees sampled in the present study lose 99% of the fine
branches and needles after 12 years. In the Sierra Nevada, Raphael and Morrison (1987)
reported Abies (spp.) and Pinus (spp.) lose all needles and fine branches after 5 years.
Graham and Cromack (1982) reported strong correlations between year o f death and
decay rate using a similar classification as the decay class variable in the present study
for Picea sitchensis (R^= 0.421) and Tsuga heterophylla (R^= 0.227) in Olympic National
Park in Washington, USA. The present study showed a similar strong correlation for
spruce (R^= 0.581) and fir (R^= 0.690) (Table 1) for the qualitative measure of decay
class.
This study’s classification of decay class also corresponds to the decay class
system of Daniels et al. (1997). The present study’s results show both fir and spruce
decay much quicker than Thuja plicata studied by Daniels et al. (1997) (Table 4). Both
the larger tree size of coastal Thuja plicata, the cold, relatively dry climate of the sub41

boreal forest relative to coastal ecosystems and the acknowledged resistance of Thuja
plicata to decay are likely responsible for these large discrepancies. For the present study
it was noted that spruce seems to decay faster than fir (Table 4). The documented
negative correlation for decay rate and tree size (Harmon et al. 1986) combined with the
fact that larger snags usually stand longer than small snags (Raphael and Morrison 1987)
suggests that fir should decay faster than spruce because spruce is generally a larger tree
in this system and therefore should remain standing longer. Furthermore, published
information about resistance to decay places both Picea spp. and Abies spp. as species
with low resistance to decay. Therefore, it seems plausible that the faster decay rate of
spruce could be attributed to different types of decay fungi rather than differences in the
extractives found in the wood itself.

42

2.5

CONCLUSIONS

By accurately estimating the year of death for trees, the TSD models can be used
as tools to quantify the date of tree mortality in forest stands. These data can then be used
in disturbance studies to quantify spatiotemporal patterns of disturbance in wet-cool,
SBSwkl forests. Given the limitations of the range of estimation, disturbance studies in
wet sub-boreal forests using this method alone should be restricted to 70 years before
present. Although this period will not span the time since stand establishment in old
forests, it can provide fine detail on disturbance dynamics over a short time and be used
in forests where other methodologies are relatively ineffective. Combined with indirect
techniques like those developed by Lorimer and Frelich (1989), Frelich and Lorimer
(1991), Frelich and Graumlich (1994), Abrams et al. (1995), Frelich and Reich (1995),
and Cherubini (1996), the TSD results can be used to quantify medium and fine-scaled
disturbances and their interactions.

43

2.6 LITERATURE CITED
Abrams, M.D., Orwid, D.A. and Demeo, T.E. 1995. Dendroecological analysis of
successional dynamics for a presettlement origin white-pine-mixed-oak forest in
the southern Appalachians, USA. Jour. Ecol. 83:123-133.
Assman. 1970. The principles of forest yield study. Permagon Press. Oxford England.
Carpenter, S.E., Harmon, M.E., Ingham, E.R., Kelsey, R.G., Lattin, J.D. and Schowalter,
T.D. 1988. Early patterns of heterotroph activity in conifer logs. Proc. R. Soc.
Edinb. Sect B. (Biol). 94:33-43.
Cherubini, P., Piussi, P. and Schweingrubber, F.H. 1996. Spatiotemporal growth
dynamics and disturbances in a sub-alpine spruce forest in the Alps: a
dendroecological reconstruction. Can. J. For. Res. 26:991-1001.
Daniels, L.D., Dobry, J., Klinka, K.K. and Fuller, M.C. 1997. Determining year of death
of logs and snags of Thuja plicata in southwestern coastal British Columbia.
Can. J. For. Res. 27:1132-1141.
Fox, J. 1991. Regression Diagnostics. Sage University Paper series on Quantitative
Applications in the Social Sciences, 07-050. Beverly Hills and London: Sage
Publications.
Frelich, L. E. and Graumlich, L.J. 1994. Age-class distribution and spatial patterns in
an old-growth hemlock hardwood forest. Can. J. For. Res. 24:1939-1947.
Frelich, L.E. and Lorimer, C.G. 1991. Natural disturbance regimes in hemlockhardwood forests of the Upper Great Lakes Region. Ecological Monographs 61 :
145-164.
Frelich, L.E. and Reich, P.B. 1995. Spatial patterns and succession in a Minnesota
southern-boreal forest. Ecological Monographs. 65:325-346.
Graham, R.L. and Cromack, K., Jr. 1982. Mass, nutrient content and decay rate of dead
boles in rain forests of Olympic National Park. Can. J. For. Res. 12:511-521.
Grier, C.G. 1978. A Tsuga heterophylla - Picea sitchensis ecosystem of coastal Oregon:
decomposition and nutrient balances of fallen logs. Can. J. For. Res. 8:198-206.
Harmon, M.E., Sexton, J., Cladwell, B.A. and Carpenter, S.E. 1994. Fungal sporopcarp
mediated losses of Ca, Fe, K, Mg, Mn, N, P and Zn from conifer logs in the early
stages of decomposition. Can. J. For. Res. 24:1883-1893.
Henry, J.D. and Swan, J.M.A. 1974. Reconstructing forest history from live and dead
plant material - an approach to the study of forest succession in southwest New
Hampshire. Ecology 55:722-783.
44

Holland, S.S. 1976. Landforms of British Columbia: a physiographic outline. Bulletin
48, British Columbia Department of Mines and Petroleum Resources. Victoria,
British Columbia.
Kelsey, R.G. and Harmon, M.E. 1989. Distribution and variation of extractable total
phenols and tannins in the logs of four conifers after 1 year on the ground. Can. J.
For. Res. 19:1030-1036.
Kneeshaw, D.D. and Bergeron, Y. 1998. Canopy gap characteristics and tree
replacement in the southeastern boreal forest. Ecology 4(3):783-794.
Johnson, E.A., Miyanishi, K. and Kleb, H. 1994. The hazards of interpretation of static
age structures as shown by stand reconstruction’s in a Pinus conforta - Picea
engelmannii forest. J. Ecol. 89:923-931.
Kubota, Y. 1995. Effects of disturbance and size structure on the regeneration process in
a sub-boreal coniferous forest, northern Japan. Ecological Research 10:135-142.
Lertzman, K.P. 1992. Patterns of gap-phase replacement in a sub-alpine, old-growth
forest. Ecology 78(2):657-669.
Lewis - Beck, M. 1980. Applied regression: an introduction. Sage University Paper
series on Quantitative Applications in the Social Sciences, 07-022. Beverly Hills
and London: Sage Publications.
Lorimer, C.G., Frelich, L.E. and Nordheim, E.V. 1988. Estimating gap origin
probabilities for canopy trees. Ecology 69(3):778-785.
Meidinger, D. and Pojar, J. 1991. Ecosystems of British Columbia. Research Branch,
Ministry of Forests. Crown Publications Inc. 330 pp.
Raphael, M.G. and Morrison, M L. 1987. Decay and dynamics of snags in the Sierra
Nevada, California. For. Sci. 33(3):774-783.
Sollins, P. 1982. Input and decay of coarse woody debris in coniferous forests in
western Oregon and Washington. Can. J. For. Res.12:18-28.
StatSoft, Inc. (1997). STATISTICA for Windows [Computer program manual]. Tulsa,
OK: StatSoft, Inc., 2300 East 14th Street, Tulsa, OK.
Veblen, T.T. 1986. Treefalls and the coexistence of conifers in sub-alpine forests of the
central Rockies. Ecology 67(3):644-649.
Veblen, T.T., Hadley, K.S., Nel, E.M., Kitzberger, T., Reid, M. and Villalba, R. 1994.
Disturbance regime and disturbance interactions in a Rocky Mountain sub-alpine
forest. Journal of Ecology 82:125-135.

45

Yamamoto, S. 1995. Gap characteristics and gap regeneration in sub-alpine old-growth
coniferous forests, central Japan. Ecological Research 10:31-39.
Yong, B., Huacheng, X., Bergeron, Y. and Kneeshaw, D.D. 1998. Gap regeneration of
shade-intolerant Larix gmelini in old-growth boreal forests of northeastern China.
Journal of Vegetation Seience 9:529-536.

46

2.7 APPENDIX
Data are presented here to provide the reader with the primary data collected from the
dead trees in the permanent sample plots. See section 2.2 METHODS for description of
categories.
E.P.
Plot
#
106

Dead
Tree
Num ber

Species

Position

Density

Depth
of
Decay

Primary
Branch
Integrity

Primary
Branch
Presence

Fine
Branch
Flexibility

Fine
Branch
Presence

Bark
Presence

Bark
Integrity

Year
of
Death

2

Spruce

Down

4

Fir

Down

106

261

Fir

Down

106

14

Spruce

Down

14
14
0
10

106

18

Fir

Down

106

Spruce

Down

106

22
25

Fir

Down

106

26

Spruce

Down

106

32

Spruce

Down

106

268

Fir

Down

106

70

Spruce

Down

106

498

Fir

Standing

106

72

Spruce

Down

106

71

Spruce

Down

497

Fir

Down

76

Spruce

Down

474

Fir

Standing

106

84

Spruce

Standing

106

48

Fir

Down

106

Spruce

Down

106

92
99

Spruce

Down

106

422

Fir

Standing

106

Fir

Standing

Spruce

Standing

Fir

Standing

106

423
34
30
29

Spruce

Standing

409

Fir

Standing

106

246

Spruce

Standing

106

278

Spruce

Down

106

45

Fir

Down

106

Spruce

Down

106

88
288

Spruce

Down

106

255

Fir

Standing

106

65

Fir

Down

3
1
1
3
3
3
1
1
1
2
2
1
2
2
2
2
3
2
3
3
1
3
1

4
4
2
4
4
4
4
4
2
1
4
1
4
4
1
4
1
1
4
4
4
1
1
1
2
2

106

3
3
3
3
3
3
3
3
3
3
3
2
3
3
2
3
1
2
3
3
3
1
1
2
3
3
1
3
3
3
3
3
3
3
3
2
3
2
2
3
3

1

106

4
4
3
3
3
3
4
4
3
3
3
1
4
3
1
3
1
1
3
3
3
1
1
1
2
4
1
2
2
2
2
4
4
3
3
1
4
2
2
4
3
1

3
3
3
3
3
3
3
3
3
3
3
3
1
3
3

106

3
3
1
2
2
3
3
3
2
2
2
2
1
3
2
1
2
1
1
2
1
1
1
1
1
1
3
1
1
1
1
1
3
3
2
2
1
1
3

3
3
3
3
3
3
3
3
3
3
3
1
3
3

106

5
5
1
2
4
5
4
4
2
2
3
1
3
3
1
2
1
1
3
3
2
1
1
1
2
2
1
1
2
2
1
2
1
3
3
1
3
1
1
2
3
1
1

1945

106

1

2

106
106

106

79

Fir

Down

106

15

Spruce

Standing

106

5

Spruce

Down

106

1

Spruce

Standing

106

12

Spruce

Standing

106

20

Spruce

Down

106

24

Spruce

Down

106

23

Fir

Down

106

46

Spruce

Standing

15
11
12.05
10
1.5
5.5
0
13^
16
0
1
0
0
15
12.5
9
0
0
0
1

24
0
0
1
1
1
5
1
2
12
0
25
0
0
2
2
0
0

1

1
3
3

1

3
2
1
3
3
3
3
3
2
3
3
3
3
3
3
3
3
3
3
3
3
1
3
3
3
3
3

1

1

3
3
2
1

1932
1977
1977
1932
1932
1932
1950
1960
1977
1977
1995

1955
1950
1995
1977
1995
1995
1945
1977
1955
1995
1995
1995
1977
1977

1

1995

3
3
3
3
4
2
4
4

1977

1

1995

4
1
3
4
2
2
1

1977

47

1990
1977
1990
1977
1977

1955
1950

1995
1990
1960
1977
1977
1995

106

420

Fir

Standing

106

51

Fir

Down

106

423

Fir

Down

106

59

Spruce

Down

106

63

Fir

Down

106

60

Spruce

Standing

106

306

Spruce

Standing

106

61

Spruce

Down

1
2
1
2
2
1
1
1

112

51

Spruce

Down

112

56

Fir

Standing

112

60

Fir

Standing

2
2
2
3
1
2
4
3
1
2
5
3
4
5
2
3
3
2
2

106

73

Spruce

Down

106

74

Fir

Down

112

422

Fir

Down

112

7

Fir

Down

112

18

Spruce

Standing

112

22

Spruce

Standing

112

30

Fir

Down

112

48

Spruce

Down

112

430

Fir

Down

112

131

Spruce

Down

112

95

Spruce

Down

112

46

Fir

Down

112

45

Spruce

Down

112

50

Spruce

Down

112

55

Spruce

Standing

112

44

Spruce

Down

112

15

Spruce

Standing

1

112

36

Fir

Down

112

412

Fir

Standing

112

92

Spruce

Down

112

433

Spruce

Down

112

59

Fir

Standing

112

381

Fir

Standing

112

55

Spruce

Down

112

379

Spruce

Standing

112

62

Fir

Standing

2
2
2
2
3
1
2
1
3

Standing

1

Standing

112

384

Spruce

107

1

Fir

Standing

107

27

Fir

Standing

1
1
1
1

Standing

1

Standing

1
1
1
2

112
112

107
107

399
67

28
29

Spruce
Spruce

Spruce
Spruce

Down

107

10

Fir

Standing

107

19

Spruce

Standing

107

20

Fir

Standing

107

21

Spruce

Standing

107

22

Fir

Down

107

2

Fir

Down

107

8

Fir

Down

3
2

107

10

Fir

Standing

1

107

11

Fir

Standing

107

13

Fir

Standing

107

16

Spruce

Standing

2
1
3

1
1

0
1
0
1
2
0
0
0
1
2
1
17
0
1
25.9
15
0
1
4.7
3.5
11
6
2.5
15
26
1.5
0
1
5
1
4
2
4
0
3
0
1
0
0
0
0
0
0
0
0
0
13
0
0
3.5
1.5
0

1
1
1
1
3

1

1
1
2

0
25

1

1
3
1

1
1
1
2
1
1
1
2
1
2
3
1
1
3
2
3
3
2
2
2
2
1
1
1

1

1

1
1
1
1

1
1
1
1
1

1
1
1

1
4
1
3
4
1
1

2
2
2
3
3
1
3
3
4
3
1
4
3
4
4
2
3
3
3
3
3
3
2
1
2
2
3
2
1
3
3
3
2
1
2
1
1

1
2
3
2
2
3
3
2
2
2
3

2
3
1
3
3
2
3
3
3
3
3
3
3
3
3
3
3
2
3
3
3
3
3
3
3
3
3
1
1
3
1
2
3
3
3
2
3
1
1
1
1

3
2
2
1
3
3
3
2
3
3
2
3
2
3

3
1
1
3
3
3
3
3
2
3
3
3
3
3
3
3
3
3
3
3
3
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
2
3
3
3
3
2
1
3
2
3
2
3
3
3
2
3

1
3
1
2
3
1
3
3
2
2
3
3
3
3
3
3
3
1
3
3
3
3
3
3
3
3
3
1
3
3
3
3
3
2
2
1
2
1

1
2
1
2
4
1

1

1995
1977
1995
1977
1960
1990

1

1977

2
2
2
4
4
3
4
4
4
2
2
4
4
4
4
2
4
4
2
2
3
4
4
2
2
4
2
2
1
2

1977

1

1977

1
1

1977

1

1995

1
1
1

1984

1977
1955
1977
1950
1977
1955
1935
1977
1977
1990
1935
1977
1950
1945
1977
1977
1977
1960

I960
1977
1977
1977
1990
1977

1977
1977
1977
1990
1977

1
1
1
3
3
1
1
3
2

1

1990

1
2
3

1990

1

1

1984

3
3
1

4
2
2
3

1940

1

1971

4

1930

1
1

3

48

1990

1990
1990

1971
1971

1955
1971
1960

107

1

4

0

1

1

1

0

2

1

0

1

1

1

0

1

6

2

2
2
2
3
3
2

Down

2

Fir

Down

1

Spruce

Standing

1

Fir

Standing

1

549

Fir

Standing

81

Fir

Standing

15

107

273

107

107

107

288

107
107

Fir

23

3

107

18

Fir

Standing

5
3

10

3

107

536

Spruce

Standing

1

0

1

107

22

Fir

Standing

2

1

1

107

23

Fir

Down

2

1

1

107

326
323
286

Spruce

Standing

1

0

1

Fir

Down

1

0

3

2
4
3
2
3
4
2
3

Fir

Standing

]

0

1

1

107

29

Fir

Down

39

Fir

Down

107

31

Fir

Down

14
17
12
23

2

107

3
1

4
4
4
4

3
3
3
2
3
3
3
3

107
107

2

107

32

Fir

Down

5
5
3
5

107

293

Spruce

Standing

1

0

1

2

2

107

45

Fir

Down

42

Fir

Down

16
13

1

107

4
4

3
3

150

2

Fir

Standing

5
5
1

0

1

1

1

3
3

3
3
3
3
3
3
3
1

3
3
3
3
3
3
3
3
3
3
3
3
3

150

340

Fir

Standing

2

1

2

1

1

1

150

9

Fir

Down

2

1

150

19

Fir

Down

1

0

3
3

150

21

Fir

Down

1

0

1

3
3
3

150

15

Fir

Standing

1

0

2

150

22

Fir

Standing

1

0

3
3

3
3
3
3
3

150

27

Fir

Down

11

1

150

28

Fir

Down

3
4

10

1

150

45

Spruce

Standing

1

0

1

150

46

Fir

Standing

2

2

1

4
4
2
2

150

51

Fir

Standing

1

0

1

2

150

68

Spruce

Down

3

2

1

3

1

2

3
3
3
3
1
3
3
2
2
2
3
2

2

2

1

1
2

3

4

1930

1

1

1995

1

1

1995

1

1

1990

2
3
3
3

1

1995

4
2

1930

1

1995
1990

3
3
3

2
4
4
2

1

1

1990

3
3
3
3

4
4
4
4

1930

1

1

1971

3
3
3
3
3

4
4
3
2

1930

2

1977

1

1977

1

1
2
1

1

1

1995

3
3

4
4

1936

1

1

1984

1

2

1990

1

1

1990

3

1984

I

2
2
2

1

2

1984

1

1

1977

3
1

2

1990

1

1984

1

3

1971

1971
1971
1971

1930
1930
19.30

1935
1990
1984

1977
1977

19.36

150

70

Spruce

Down

2

2

150

73

Fir

Down

2

2

150

75

Fir

Down

2

2

3

3

150

82

Fir

Down

1

0

1

1

2

150

88

Fir

Down

1

0

1

1

2

150

87

Spruce

Standing

1

0

1

1

2

3
3
3
3
1
3
3
3
3
3

150

95

Fir

Standing

1

0

1

1

2

1

1

1

1990

150

106

Spruce

Standing

1

0

1

1

2

1

1990

345

Fir

Standing

]

0

1

1

1

1

1995

150

312

Fir

Down

1

0

1

1

1995

331

Fir

Down

1

0

1

3

1

3
3

2

150

4

1977

150

116

Fir

Standing

1

1

1

1

1

1995

150

119

Fir

Down

1

1.5
0

3
2
3
3
3

1

150

1

2

1990

118

Fir

Down

1

0

1

3
3

2

150

3
1

150

120

Fir

Down

1

0

1

2
2

1990

1

0

2

1

17

1

2
2
2
3

1

0

1

1

2

1
1

0

1

1

2

1

1

1

1

2
2

2

1

1

1

2

2

3

3

150

122

Fir

Down

150

125

Spruce

Down

150

131

Fir

Standing

150

134

Fir

Standing

150

132

Fir

Down

150

328

Fir

Down

1

0

150

133

Fir

Down

2

2.5

1

3
3

3
3
3
3
3
2
3

1

1960
1977

1995

1

1

1990

3

1950

1

4
1

1

1

1995

1

3
3
2

1990

49

1995

1977
1984

150

137

Fir

Down

150

138

Fir

Down

150

140

Fir

Standing

150

145

Spruce

Down

150

147

Fir

Down

150

149

Fir

Down

149

153

Fir

Standing

149

195

Spruce

Down

149

169

Fir

Down

149

198

Fir

Standing

149

285

Fir

Standing

149

255

Fir

Down

1
1
2
2
2
2
2
1
2
1
2
2

1

1

149

316

Fir

Standing

1

0

149

318

Fir

Down

8

149

409

Fir

Standing

3
1

149

385

Spruce

Standing

1

1
1
2
1
1
1
1
2
2
1
2
2
1
2
1
1
1
1

149

28

Fir

Standing

149

210

Fir

Standing

149

220

Fir

Standing

149

217

Fir

Standing

149

284

Fir

Down

1
1
3
1
3

149

258

Fir

Standing

1

149

279

Spruce

Standing

149

274

Fir

Standing

1
1

149

303

Spruce

Standing

1

149

302

Fir

Standing

]

149

372

Spruce

Standing

149

241

Fir

Standing

149

186

Fir

Standing

149

189

Fir

Down

1
1
2
2

0
0
1
1
0
1.5
0
0
1
0
1.5

1
1
1
2
1
1
1
1
1
1
1

0
0
0
0
4
0
5
0
0
0
0
0
0
0
1.5
1

2
2
3
3
3
3
1
3
3
2
3
2
1
3
2
1
2
2
2
3
3
3
1
1
1
1
1
2
2
2

3
3
3
3
3
3
1
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2

3
3
3
3
2
1
3
3
3
3
3
3
3
3
3
2
3
3
3
3
3
1
3
3
3
3
3
3
3
3

1
1
3
3
3
3
3
3
3
3
3
3
1
3
3
1
1
1

3
3
3
3
1
3
1
1
1
3
3
3

1
1
3
3
3
4
1
4
2
2
2
2
1
4
2
2
1
1
2

1977

1

1971

4
3
1
2
1
1
1
3
3
3

50

1984
1977
1960
1955
1960
1995
1971
1960
1971
1984
1984
1990
1971
1984
1971
1995
1971
1971

1971
1984
1990
1984
1990
1995
1990
1984
1971
1971

Table 1. Bivariate regression statistics for each explanatory variable regressed (bivariate
regression analysis) against years since death used in preliminary analysis to detect for
linearity for each species.
VARIABLE

Fir (n = 114)

Spruce (n = 69)
R"

P

R^

P

Decay Class

0.581

<0.001

0.690

<0.001

Proportion of Decay

0.495

<0.001

0.612

<0.001

Bark Integrity

0.516

<0.001

0.594

<0.001

Bark Class

0.469

<0.001

&582

<0.001

Branch Integrity

0.466

<0.001

0.639

<0.001

Fine Branch Flexibility

0.324

<0.001

0.424

<0.001

Position

0.179

<0.001

0.181

<&001

51

Table 2. Correlation matrix for spruce and fir models. The number of correlations
amongst explanatory variables that exceed r = 0.8 indicates the degree of
multicollinearity in the overall model. Bold values indicate a combination of variables
with significant multicollinearity.

Spruce

POSITION

DECAY BRANCH
CLASS INTEGRITY

FINE
BARK
BARK
PROPORTION
BRANCH
CLASS INTEGRITY OF DECAY
FLEXIBILITY

P osition
D ec a y C ia ss

0.650

Branch Integrity

0.072

0.264

Fine Branch
Flexibility

0.450

0.536

0.138

Bark C la ss

0.605

0.684

0.179

0.671

Bark
Integrity

0.612

0.679

0.227

0.641

0 .8 3 2

Proportion of
D ec a y

0.562

0 .8 4 9

0.293

0.516

0.676

Fir

POSITION

Decay BRANCH
Class INTEGRITY

0.689

FINE
PROPORTION
BARK
BARK
BRANCH
CLASS INTEGRITY OF DECAY
FLEXIBILITY

P osition
D e c a y C ia ss

0.324

Branch integrity

0.037

0.151

Fine Branch
Flexibility

0.305

0.489

0.157

Bark C la ss

0.417

0.671

0.107

0.454

Bark
Integrity

0.460

0.727

0.088

0.447

0 .8 4 0

Proportion of
D ec a y

0.327

0 .8 3 5

0.127

0.422

0.609

0.664

52

Table 3a. Summary of regression statistics for the spruce models. For both standing and
down trees, both models were significant (°c = 0.05), but many of the partial slope
coefficients were not significant due to the heteroscedasticity and multicollinearity.

ST A N D IN G

DOWN

R S q u are A djusted

0.873
0.800

0.753
0.588

R oot Mean S quare Error

6.147

10.414

Mean of R esp o n se

14.636

28.472

O bserv atio n s

33

36

SUMMARY OF MODEL FIT; S P R U C E
R S quare

ANALYSIS OF VARIANCE
DEGREES OF FREEDOM
Source
Model

Standing
12

Error
C Total

20
32

SUM OF SQUARES

MEAN SQUARE

F - RATIO

Down

Standing

Down

Standing

Down

Standing

Down

14
21

5188.027
755.607
5943.634

6945.666
2277.306
9222.972

432.336
37.780

496.119
108.443

11.443
<0.0001

4.575
0.0009

35

EFFECT TEST
DEGREES OF FREEDOM

SUM OF SQUARES

Source

Standing

Down

Standing

Down

Decay Class
Branch Integrity

2
2

4
2

638.881
158.324

2359.423
63.657

Fine Branch Flexibility
Bark Class
Bark Integrity
Proportion of Decay

2
2
2
1

2
2
2
1

305.934
10.943
298.367
117.380

358.263
194.782
54.536
114.371

F -RATIO

PROBABILITY > F
Down

Standing Down

Standing

8.455
2.095
4.049
0.145
3.949

5.439
0.294
1.652
0.898
0.252

0.002

0.004

0.149
0.033
0.866
0.036

0.749

3.107

1.331

0.093

0.216
0.422
0.780
0.262

53

Table 3b. Summary of regression statistics for the fir models. For both standing and down
trees, both models were significant (<%= 0.05), but many of the partial slope coefficients
were not significant due to the heteroscedasticity and multicollinearity.
SUM MARY OF MODEL FIT; FIR

S T A N D IN G

DOWN

0.768
0.691

0.847

R S q u are
R S q u are A djusted

0.800

R oot Mean S q u are E rror

7.202

9.661

Mean of R esp o n se

14.302

31.230

O b serv atio n s

53

61

ANALYSIS OF VARIANCE
DEGREES OF FREEDOM

SUM OF SQUARES

MEAN SQUARE

F - RATIO
Standing

Down

Source

Standing

Down

Standing

Down

Standing

Down

Model

13
39

14
46

6704.430
2022.734

23777.154

515.725

1693.370

9.444

18.196

Error

4293.633

51.865

93.340

<0.0001

<0.0001

C Total

52

60

8727.167

28070.787

EFFEC T TEST
DEGREES OF FREEDOM

SUM OF SQUARES

F - RATIO

PROBABILITY > F

Source

Standing

Down

Standing

Down

Standing

Down

Standing

Down

Decay Class
Brancti Integrity

3
2

4
2

1349.623
9.642

798.058
263.915

8.674
0.093

2.138
1.414

0.091
0.254

Fine Branch Flexibility
Bark Class
Bark Integrity
Proportion of Decay

2
2
2
1

2
2
2

1393.927
345.451

1015.112
208.686

378.869

271.540

13.430
3.331
2.435

5.438
1.118
0.9697

0.0002
0.911
<0.0001

1

64.156

792.044

1.237

8.486

0.040
0.080

0.008
0.336
0.415

0.273

0.006

54

Table 4. Comparison of mean years since death and Decay Class for Thuja plicata
(Cedar) reported by Daniels et al., (1997) and spruce and fir (this study).

Decay

Cedar

Fir

Spruce

Class

Standing

Down

Standing

Down

Standing

Down

1

-

3.5

12

15

10

13

2

-

50

27

23

19

28

3

276

279

67

24

23

45

4

122

1200

-

47

67

63

5

150

-

-

56

-

66

55

2020

S
CD

E
■S5
LU

20 0 0 -

CD

X3

O
1980 -

CD
CD

Q

O 1960 cô

■a
CD

ro
E

1940 -

%

LU

1920
1920

1930

1940

1950

1960

1970

1980

1990

2000

A ctual Y ear o f D eath (S a m p le Plot R eco rd s)

•
O

standing Spruce
Standing Spruce Regression
= 0.873
Down Spruce
Down Spruce Regression R =0.753

Figure 1. Regression fit for spruce model with aetual year of death taken from the Aleza
Lake Research Forest permanent sample plot data and x-axis and predicted year of death
from the TSDM.

56

2000 -|

s 1990 E

S 1980 -

0)
~oo

s

1970 1960 -

Q

0
CO

1950 0

1

1940 O

I 1930 -

0

ijy
1920 - -------1920

o
cP

-y -

1930

1940

1950

1960

1970

1980

1990

2000

A ctual Y ear of D eath (S a m p le Plot R eco rd s)

•
O

standing Fir
Standing Fir Regression
= 0.768
Down Fir
Down Fir Regression R^ = 0.847

Figure 2. Regression fit for fir model with actual year of death taken from the Aleza Lake
Research Forest permanent sample plot data and x-axis and predicted year of death from
the TSDM.

57

2000
1990
1980
-

1970

I
I 1960

1950
g

1940

8
m
E

p

1920
1920

1930

1940

1950

1960

1970

1980

1990

2000

Tree Ring Estimate (Year of Death)
Figure 3. Plotted values of dates of tree mortality from the Times Since Death Model (yaxis) and from tree ring cores whose radial growth increased by 50% following tree
mortality (x-axis). Linear regression line fitted to the data pairs indicates the two
estimates are very close R^= 0.924.

58

CHAPTER 3. A METHODLOLOGY FOR ESTIMATING YEAR OF DEATH IN SUBBOREAL FORESTS USING TREE RING GROWTH RATES
3.0 ABSTRACT
Tree ring growth rate criteria have been widely used to quantify disturbance
regimes in forests of complex age structure but the methodology requires local calibration
of growth rate parameters. This chapter summarizes the development of tree ring growth
rate criteria for understory trees used to assign a canopy ascension date to canopy trees.
Canopy ascension dates correspond to overhead mortality and this information can be
used to quantify disturbance regimes. The criteria were developed in a wet, cool SubBoreal Spruce forest near Prince George, British Columbia. Two sampling methods were
used. First in eight 10-meter radius plots, all dead and live trees (>30 cm tall) were stemmapped. Date o f death in dead trees (n = 101) was estimated using a time since death
model (Chapter 2). Increment cores or basal sections were then taken from all living trees
in the plot and tree attributes were collected (species, diameter, crown class, live crown
ratio). Year of death from the time since death model was then compared to year o f death
estimated from release patterns in nearby understory trees. When there was agreement
between the two estimates of year of death (+/-10 yrs), the understory tree was classified
as a released tree. Trees not showing release were eliminated. From the remaining trees it
was determined that only trees > 20 cm dbh when they died caused release in understory
trees. Furthermore, understory trees were typically <5 meters from the dead tree, they
were at least 5cm dbh < than the dead tree and they were of vigorous growth. Next, the
relationships described above were used as criteria for selecting an independent
population of understory trees (n = 428) subtending dead trees located at random. From
these data growth rate criteria were developed to determine canopy ascension dates. The

59

criteria are as follows: 1) Gap origin growth rates is 1.50mm/yr, 1.72mm/yr and
1.07mm/yr for Picea glauca X engelmannii, Abies lasiocarpa, and Betula papyrifera. 2)
Release criteria (for all species) is a 65% increase in growth sustained for 15 yrs
following a 15 yr period of slow growth. 3) Other criteria used to assign canopy
ascension dates were constant declining, parabolic or ambiguous tree ring patterns, which
indicated gap origin trees that were growing slower than the gap-origin growth rate
criterion. Criteria were also developed to assign release from suppression from the
interpretation of irregular growth patterns.

60

3.1 INTRODUCTION
Analyses of tree ring growth patterns have been used to reconstruct disturbance
regimes in forests o f complex age structure (Frelich and Lorimer 1989). The method
relies on the development of annual ring growth criteria that indicates the date a tree is
released from suppression, or the date a seedling establishes in a gap, when a near-by
canopy tree dies (i.e. canopy ascension dates). The location and timing of canopy
ascension dates in a stand have to be analyzed to provide information on the spatial and
temporal patterns of canopy level disturbance for eastern hardwood forests, eastern boreal
forests, and western sub-alpine forests (Lorimer and Frelich 1989; Frelich and Lorimer
1991; Frelich and Graumlich 1994; Abrams et al. 1995; Frelich and Reich 1995;
Cherubini 1996). However, the methodology has not been developed for sub-boreal
systems in western North America, which are inherently challenging for this
methodology due to the tall, narrow tree crowns which may not cause as substantial
increases in understory light regimes after death as broad crowned hardwood forests.
Furthermore, growth rate criteria need to be developed locally in order to avoid
over or under estimating disturbance intensity. For example, climatic effects, ontogenetic
patterns, and stand development factors (such as canopy thinning) could all potentially
cause release in a tree and be misinterpreted as a canopy disturbance. Thus specific
criteria need to be developed that take into account these effects as well as the inherent
variation in tree response due to species, type of disturbance and stand (species
composition, age, height, canopy structure), and site factors (nutrient availability,
moisture status, soil temperature).

61

The development of tree ring growth rate criteria in this study is assisted by
multiple linear regression models that estimate the date of death for Picea glauca x
engelmannii (Parry ex Engelm.) (hybrid spruce, hereafter referred to as spruce) and Abies
lasiocarpa (Hook.) Nutt, (sub-alpine fir, hereafter referred to as fir) (Chapter 2). These
models improved the process of developing tree ring growth rate criteria by providing an
independent estimate of time since death for canopy trees. Therefore the objectives of this
study are to determine the reliability of using tree ring release information to date tree
mortality in SBS forests and develop criteria for estimating canopy ascension date using
growth rate patterns contained within tree ring cores.

62

3.2 METHODS
3.21 Study Area and Site Selection
The research was conducted in two old-growth forests at the Aleza Lake Research
Forest which is located at 54° 07’ N, 122° 04’ W, about 60 kilometers east of Prince
George, British Columbia, Canada. Stand 1 is located on the north side of the Bear Road
approximately 1 km east of the Bear Road and Aleza Road junction. Stand Two is located
on the west side of the Aleza Road approximately 2.5kms south of the Bear Road and
Aleza Road Junction. The elevation of the research forest is between 600 and 750 meters
above sea level on the Nechako Plain of the Fraser River Basin in the Interior Plateau
physiographic region (Holland 1976). The Aleza Lake Research Forest is located in wet,
cool, sub-boreal spruce-fir forest. The region is classified as the Sub-Boreal Spruce, wetcool 1 (SBSwkl) biogeoclimatic zone according to a biogeoclimatic classification system
in common usage in British Columbia (See Meidinger and Pojar (1991) for details). The
SBS wkl climate is characterized by cold, snowy winters and moist, cool summers. The
climate is slightly less continental than typical for the SBS due to the orographic
influence of the Northern Rocky Mountains to the east, resulting in higher precipitation
than usual for the rest of the zone (Meidinger and Pojar 1991). The old-growth forests are
mixtures of Picea glauca x engelmannii (spruce) and Abies lasiocarpa (fir) with scattered
Pseudotsuga menziesii var. glauca (Douglas-fir), Pinus conforta var. latifolia (lodgepole
pine) and Betula papyrifera (birch). Old-growth forests at the Aleza Lake Research
Forest are uneven aged (Decie 1957). Sampling was conducted in three old-growth stands
located on medium to good sites with minimal variation in soils and topography.

63

3.22 Sampling Design and Plot Measurements
Two sampling approaches were used in this study. First (Comprehensive
Analysis), in eight 10 meter radius plots all dead trees (n = 1 0 1 ) ^ 0 em dbh were located
and decay characteristics used as predictor variables in the TSD model were collected
(Chapter 2). Ten cm dbh was used as a minimum diameter since preliminary
investigations revealed that trees smaller than this rarely caused release in nearby trees.
Decay data were entered into the TSD model to estimate the date of death for each tree.
All live trees >1.3 meters tall were cut down at 1.Cm and a basal section was collected or,
in larger trees, an increment core was taken (also at 1.0m). Tree ring cores were stored in
plastic straws, mounted on 1 inch thick grooved Styrofoam strips. The tree ring samples
and the basal samples were dried, sanded and scanned using a flatbed scanner. The
scanned images were analyzed using Windendro® (Regent Instruments, Blaine, Quebec)
which measures and records annual ring width growth (mm). In the second sampling
approach (Selective Analysis), dead trees were located unsystematically (n =
176).Variables required for the TSD model (Chapter 2) were collected from these trees
and a minimum of one gap-filling tree was selected for an increment core sample (based
on criteria developed in the Comprehensive Analysis, see results).

3.23 Analysis
Comprehensive Analysis
In order to determine if a release (i.e. sustained increase in growth) event occurred
following overstory mortality, ring width data for live trees was visually inspected for
growth patterns that suggested release. Trees that did not show release were eliminated

64

{i.e. trees with constant declining growth or extremely flat incremental growth patterns).
On the remaining trees, tree ring cores were inspected for growth increases that occurred
within +/-10 years of the TSD model estimate for year of death. Release events were
eliminated which did not correspond to the date of death estimated from the TSD model
(+/- 10 years) or were unlikely due to unrealistic distances from gap-maker to gap-filler,
or due to the size relationship of the gap-maker to gap-filler. On the rest of the trees, a 15year mean growth rate before and after the release date was calculated. Percent release
was then calculated for each event as: 15 year growth after release /15 year growth
before release) x 100. The 15-year average was used to eliminate the influence of short­
term variations in growth rate on the chronology. Trees releasing after 1982 could not be
averaged for a full 15 years. Releases after 1982 were only included if they were
sustained until the year of sampling (1998). Trees releasing after 1990 were not included
in the analysis because assumptions could not be made that the release would be
sustained for 15 years.
Understory trees whose ‘released’ growth rates were 25% > than pre-release
growth rates and within +!- 10 yrs of the TSD model estimates were used to examine
several relationships. Dead-tree - understory-tree size relationships were used to
determine the relative size difference necessary to cause a release and the minimum size
required to cause release in a subordinate tree of any size. The diameters of dead trees
and estimated diameter of the live trees (estimated as: present diameter - (2 x total ring
width increment since release)) at the time of death provide evidence for how large a
dead tree must be before it causes a growth increase in a given size of understory tree.
Dead-tree - understory distance relationships were also quantified. The distribution of

65

distance data provided an estimate for the range in influence mortality has on the
understory trees as measured from the base of the dead tree to the base of the released
tree.

Selective Analysis
Interpretation of Growth Patterns
In the Selective Analysis, detailed annual growth criteria (early growth rate
criteria, percent release criteria, and overall growth criteria) were developed to be used in
future studies (Chapter 4 and 5). The objectives here were to; 1) Determine a gap-origin
growth rate threshold (i.e. high rates of early growth indicate a tree was growing in a gap
created by canopy mortality when it reached the coring height). 2) Determine a release
threshold where slow growth followed by sustained high growth indicates a tree was
initially suppressed by overstory competition then released following a gap-making event
(Figure la). 3) Determine criteria to deal with trees that do not show either 1) or 2) but
may have some pattern that may indicate a tree originated in a gap or was released, even
though it did not meet the early growth rate or release criteria (Figure Ib-e). This
approach generally followed the methods of Lorimer et al. (1988) and Lorimer and
Frelich (1989) with modifications for this study’s forests where necessary.
Early Growth Rate Criteria
Trees germinating in gaps should have faster annual radial growth rates than
understory tree growth rates (Oliver and Larson 1996). Therefore, trees meeting a
minimum growth rate threshold can then be used to judge the decade of an overhead
mortality based on their total age at the coring height. Over 400 gap-filling trees were

66

classified as either gap-origin or initially suppressed in the following manner. Open
grown trees have sigmoidal increases in cumulative diameter growth that results in an
incremental annual growth pattern that peaks early on, usually around 20-30 years,
followed by a generally declining pattern. Thus, if a tree is in the canopy and the past
growth rate is equal to or greater than current growth, it ean be assumed that the tree
originated in a gap (Lorimer et al. 1988). A 10-year average growth rate was then
determined for eaeh tree for the first 10 years of growth from the pith. In total, 319 fir,
104 spruce, and five birch were sampled.
The growth rate data were then used to set the early growth rate criteria using
Equation 1 (below) obtained from Lorimer et al. (1988). The formula was used in an
iterative proeess until P^i = 95% was obtained. This assures that 95% of the time, the
growth rate threshold selected, correctly distinguishes true gap-origin trees from fastgrowing suppressed trees. However, this strict (high) criteria will also prevent slow
growing gap-origin trees from passing the criteria. Lorimer et al. (1988) suggest that it is
preferable to maintain this high confidenee in gap origin because other growth rate
pattern criteria (see below) can be developed to “recover” gap-origin trees that fail to
meet the threshold.

Probability of suppression (Lorimer et al. 1988):

Equal,on I

Px,

=

j

Where:
Pxi= probability of suppression for a sapling of size elass i with growth rate x,
Sxi = proportion of suppressed (understory) trees of size elass i exeeeding growth rate x,
Gxi ==proportion of gap saplings in size class i exceeding growth rate x,
Qsi = proportion of all saplings of size class i that are suppressed

67

Qgi = proportion of all saplings of size class i that are growing in gaps, such that Qsi+Qgi
1.0

=

Percent Release Criteria
Abrupt and sustained increases in incremental diameter growth (Figure la) may
indicate that a tree was released from suppression by canopy mortality. In order to
determine the growth response of understory trees to overhead tree mortality and evaluate
these responses for release criteria, mortality dates of trees were estimated using a time
since death model (Chapter 2). These dates of mortality were then compared to dates of
release preserved in increment cores from the gap-filling trees. If the dates of release
coincided within 10 years of the time since death model estimate, the tree was classified
as a gap-filler and a percent release value was calculated from radial increment ((15-yr
growth after release /15-year growth rate before release)x 100). The release criteria were
then developed as follows.
To avoid classifying crown thinning responses and adjacent mortality as overhead
canopy mortality a minimum release duration was implemented. Since trees already in
the canopy fill in gaps through lateral expansion of the crown quickly, any growth
increase in canopy trees will be short lived (Lorimer and Frelich 1991). Furthermore,
since the gap will be quickly colonized, adjacent understory trees will also only receive
short-term benefits. Therefore a 15-year sustained release criteria was selected.
To avoid classifying an increase in growth due to a period of slow growth caused
by drought from being classified as a release, a minimum period o f slow-growth before
release was selected. Lorimer and Frelich (1989) found that the most severe drought of
the 20'*’ century only moderately affected diameter growth on the most sensitive sites in

68

eastern mixed forests. Decreases in growth due to drought occurred over 2-12 years and
averaged 5.0 years. Therefore, a “15 year slow growth before release criteria” was
proposed. The minimum slow growth criterion was used in this study to screen climatic
variations, which are normally short lived from interpretation as release.
To avoid classifying crown thinning {i.e. trees in the canopy responding to the
death of other trees in the canopy) and adjacent mortality as overhead mortality, two
approaches were taken. First, threshold diameter limits, beyond which, trees would be too
large to be candidates for understory release were established for each species. In this
study < 5% of spruce were overtopped at greater than 40 cm dbh, while for fir and birch,
< 5% were overtopped at greater than 30 cm dbh. Thus, trees greater than these diameters
at the time of release were not counted as canopy ascensions even if they met the percent
release criteria because a growth increase is likely due to canopy thinning and not a
release from suppression. Secondly a relatively high percent release value was chosen
(25%-quantile), since adjacent mortality usually cause weaker growth increases than
overhead releases (Lorimer and Frelich 1989).

69

3.3 RESULTS AND DISCUSSION
3.31 Comprehensive Analysis
Of lOI dead trees in the pilot study 51 release events occurred within 10 years of
the time since death model estimate. However, 29 dead trees were less than 20 cm DBH
when they died, and only two of these 29 trees had an associated release event. The total
number of release events associated with mortality of the 72 trees greater than 20 cm was
49 (68%). The rest of the mortalities either did not have any understory trees (n = II) or
the understory trees did not respond for reasons that could not be determined (n - 12).
Since dead trees smaller than 20 cm DBH did not cause consistent or substantial
increases in the growth of nearby understory trees, time since death model estimates of
death provide the only reliable estimate available for smaller trees. Therefore when
sampling potential gap-fillers for dead trees in the Selective Analysis, and in further
studies only dead trees greater than 20 cm DBH should be expected to cause release in
understory trees.
The mean percent release for the 49 release events was 80% with 95% of the
releases falling between 66% and 94%. The maximum release was 246% and the
minimum was 13%. Only 6% of release events were less than 25%. The time since death
model increased the certainty that each release event was associated with tree mortality
therefore, a fairly liberal threshold of 50% for release was established as the release
criterion for the Comprehensive Analysis. Trees showing <50% release were eliminated
and were not analyzed in the remainder of the relationships.
The mean distance from gap-maker to gap-filler was 2.84 meters, the minimum
was 0.51 meters and the maximum was 6.67 meters. Ninety-five percent of the release

70

events occurred between 0.54 and 6.4 meters from the dead tree. There were no strong
relationships, using linear or other functions between percent release (response) and
distance from dead tree (p>0.05 in all cases). There was a weak trend for trees close to
the dead tree (<1 meter) and far away (>5 meters) to exhibit less release than trees in the
2 - 4 meter range. There was also no difference in percent release when I. tomentosus
caused mortality (gradual mortality) was compared to other - typically more punctuated
gap-making events (p = 0.20).
There was no clear relationship between the percent release and the dead tree live tree diameter ratio, or the diameter of the dead tree, although in all cases the dead
tree was at least 5 cm (dbh) larger than the understory tree. This suggests that understory
trees which are at least 5 cm dbh smaller than the dead tree and within 5 meters of its
base, have the potential to release. Thus, distance appears to be more important than dead
tree: live tree size relationships as a sampling guideline. Therefore, sampling criteria for
gap-filling trees replacing dead canopy trees was based on the following guidelines: less
than 5 meters from the dead canopy tree, at least 5 cm less than the diameter of the dead
tree at time of release and if possible in the 5-15 cm DBH range, in intermediate or
suppressed canopy positions.

3.32 Selective Analysis
Gap-Origin Criteria
For birch the minimum gap-origin growth rate was 1.072mm/yr (n = 3) and the
maximum suppressed growth rate was 0.552 mm/yr (n = 2) (Figure 2). Since the growth
rate distributions were non-overlapping, the selection of the minimum growth rate

71

indicating gap-origin status for birch was 1.072 mm/yr. Although this value is based on a
very small sample, birch is typically associated with gaps in sub-boreal spruce-fir forests
since it is a relatively shade intolerant species (Archibold 1980; Krajina et al. 1982) thus
the gap-origin growth rate criteria can be relatively liberal.
For fir, the distribution of ring widths overlapped considerably between
suppressed and gap-origin trees (Figure 2) (n = 319). Since there were two other growth
criteria (Figure 3) used to independently estimate disturbance date, an early growth rate
threshold was selected which yielded a high confidence (95%) of gap-origin or
alternatively, a low probability of mistakenly identifying a fast-growing suppressed
sapling as a gap origin tree. Using a formula to determine gap-origin thresholds with 95%
confidence level given in Lorimer et al. (1988), the threshold for fir was set at 1.72
mm/yr. The same guidelines were followed for spruce and the threshold was set at 1.5
mm/yr (n = 104).

Percent Release Criteria
Fifty percent o f the observations for percent release fell between 65% and 129%
with the median being 92% and there was no significant difference in percent release
between species (p = 0.194). The distribution of percent release was not normal, but a
median test also did not detect significant differences in percent release between species
(Chi Square = 1.34,p=0.50). Therefore, the same release criteria were used for all species.
The percent release criterion was based on supporting evidence from other studies and on
the percent release data distribution in the present study. In most studies it was been
shown that the typical release from suppression for Picea spp. and Abies spp. is less than

72

100%. For example, seventy-one year old Picea glauca in eastern Canada released about
67% (in diameter growth ) following thinning (Fraser 1962). In a 16-year-old plantation
Picea glauca diameter release was 37% following herbicide release (Balvinder et al.
2001). Sutton (1995) reported a 50-100% increase in diameter growth in a 30 year old
Picea glauca stand following a weeding treatment and Biring et al. (1999) reported an
85% increase in growth in Picea glauca anà Abies balsamifera over twelve years
following herbicide treatments. Based on the evidence provided from these studies, a
65% increase in growth (which corresponds to the 25%-quantile in the percent release
data distribution) was chosen. In comparison, Veblen (1986) used 150% as a release
criterion for Picea glauca x engelmannii and Abies lasiocarpa, but only imposed a fiveyear sustainability requirement. In hemlock hardwood forests Lorimer and Frelich
(1989), and Frelich and Graumlich (1994) used a minimum 50% growth increase for Acer
saccharum and Betula allagheniensis sustained for a minimum of 10 years. Thus the
criteria selected here is reasonable. The trees that do not pass the release threshold or the
early growth rate threshold are recovered with overall growth rate interpretations (below)
(Figure 3).

Overall Growth Rate Pattern Criteria
The relatively high thresholds for gap-origin and percent release were established
in order to have high confidence in attributing a growth response to mortality and this
inevitably results in some trees not passing the criteria. For these trees, lifetime
incremental patterns were assigned to several groups and assumptions about these group
patterns were used to assign the tree a date of canopy ascension (Figures lb,c,d,e). The

73

first group was overall growth patterns that started out relatively high and remained
consistent or had an increasing growth pattern followed by a flat or declining pattern
(Figure lb) (Frelich and Graumlich 1994). These patterns of growth are characteristic of
dominant trees in even-aged stands (Assman 1970; Oliver and Larson 1996; Frelich and
Graumlich 1994) and accordingly it is reasonable to assume that a similar pattern would
result in gap origin trees (Lorimer and Frelich 1989). Therefore, the date of canopy
ascension for these trees is equal to the year that the tree reached the coring height of
1.0m.
Another type of lifetime growth pattern used is a parabolic shape where tree
growth increases gradually for a few decades, reaches a peak then gradually declines
(Figure Ic). These patterns were considered indicative of gap-origin trees, growing at less
than the gap-origin threshold if the peak growth rate was achieved within 25 years of the
first ring in the chronology (Lorimer and Frelich 1989).
Ambiguous patterns are those that have a period of increasing growth lasting > 25
years but do not pass the percent release criteria (Figure Id). An ambiguous zone was
established beginning with the first ring of the increasing period to where 80% of the
maximum growth rate was achieved (Lorimer and Frelich 1989). The decade of canopy
ascension was determined to be the decade where the median growth rate was achieved
during the ambiguous period. This interpretation is based on the assumption that
ascension to the canopy in these trees would not coincide with a growth increase of less
than 25% (Frelich and Lorimer 1989). Trees not meeting this criteria were treated as
irregular patterns and the interpretations are described below.

74

Trees that have alternating high and low growth cycles or other random patterns,
but do not meet the growth criteria described so far are considered to have irregular
growth patterns (Figure le). Trees that have a declining arrangement of successive peaks
with the first peak having the largest growth rate achieved within the first 25 years of
growth are interpreted as gap-origin. Trees that have a peak growth later than 25 years
tfom the earliest ring are treated as ambiguous patterns as described earlier (Lorimer and
Frelich 1989).

75

3.4

CONCLUSIONS

Tree ring growth rate criteria can be used to indicate the year of canopy ascension
for trees in western sub-boreal forests. This method had been previously untested in these
forests. The method developed here shows that the technique can be used in humid
forests in northerly latitudes and in forests dominated by trees with tall narrow crowns.
This study provides a guide for tree ring growth rate criteria in studies in similar
environments and provides criteria locally developed and tested that can be applied to
disturbance regime investigations.

76

3.5 LITERATURE CITED
Abrams, M.D., Orwid, D.A. and Demeo, T.E. 1995. Dendroecological analysis of
successional dynamics for a presettlement origin white-pine-mixed-oak forest in
the southern Appalachians, USA. Jour. Ecol. 83:123-133.
Archibold, W.O. 1980. Seed input into a postfire forest site in northern Saskatchewan.
Can. J. For. Res. 10: 129-134.
Assman. 1970. The principles of forest yield study. Permagon Press. Oxford England.
Balvinder, S.B., Yearsley, H.K. and Hays-Byl, W.J. 2001. Ten-year responses of White
spruce and associate vegetation after Glyphosate treatment at Tsilcoh River.
British Columbia Ministry of Forests. Victoria, British Columbia.
Biring, B.S., Hays-Byl, W.J. and Hoyles, H.E. 1999. Twelve-year conifer and vegetation
responses to discing and glyphosate treatments on a BWBSmw backlog site.
British Columbia Ministry of Forests, Research Branch. Victoria, British
Columbia. Working Paper 43.
Cherubini, P., Piussi, P. and Schweingrubber, F.H. 1996. Spatiotemporal growth
dynamics and disturbances in a sub-alpine spruce forest in the Alps: a
dendroecological reconstruction. Can. J. For. Res. 26:991-1001.
Decie, T. 1957. Working plan for the forest experiment station Aleza Lake: for the period
April U*, 1957 to March 3U‘, 1967. British Columbia Forest Service. Prince
George, BC.
Fraser, D. A. 1962. Apical and radial growth of white spruce {Picea glauca (Moench)
Voss) at Chalk River, Ontario, Canada. Canadian Journal of Botany. 40:659-668
Frelich, L. E. and Graumlich, L.J. 1994. Age-class distribution and spatial patterns in
an old-growth hemlock hardwood forest. Can. J. For. Res. 24:1939-1947.
Frelich, L.E. and Lorimer, C.G. 1991. Natural disturbance regimes in hemlockhardwood forests of the Upper Great Lakes Region. Ecological Monographs 61 :
145-164.
Frelich, L.E. and Reich, P.B. 1995. Spatial patterns and succession in a Minnesota
southern-boreal forest. Ecological Monographs. 65:325-346.
Holland, S.S. 1976. Landforms of British Columbia: a physiographic outline. Bulletin
48, British Columbia Department of Mines and Petroleum Resources. Victoria,
British Columbia.

77

Krajina, V.J., Klinka, K. and Worrall, J. 1982. Distribution and ecological
characteristics of trees and shrubs of British Columbia. University of British
Columbia Faculty of Forestry, Vancouver, British Columbia.
Lorimer, C.G. and Frelich, L.E. 1989. A methodology for estimating canopy disturbance
frequency and intensity in dense temperate forests. Can. Jour. for. Res. 19:651663 .

Lorimer, C.G., Frelich, L.E. and Nordheim, E.V. 1988. Estimating gap origin
probabilities for canopy trees. Ecology 69(3):778-785.
Meidinger, D. and Pojar, J. 1991. Ecosystems of British Columbia. Research Branch,
Ministry of Forests. Crown Publications Inc. 330 pp.
Oliver, C D. and Larson, C.B. 1996. Forest Stand Dynamics. John Wiley and Sons.
Toronto. 527 pp.
Sutton, R.F. 1995. White spruce establishment: initial fertilization, weed control, and
irrigation evaluated after three decades. New Forests 9:123-133
Veblen, T.T. 1986. Treefalls and the coexistence of conifers in sub-alpine forests of the
central Rockies. Ecology 67(3):644-649.

78

A b r u pt r e l e a s e

T

-H -H 4- +

il I I! I I I! I!

C o n s t a n t

c

H 444

H -H -H

D e c l i n i n g

P a r a b o i i c

5
4
-- 3
- - -

=F=4- ■4 1 1 1 1- 4

A m

1-

• i

1 4

2
1

0

b i g u o u s

i r r e g u l a r

Figure 1. Demonstration of various incremental growth patterns used to assess date of
canopy ascension. Y-Axis = tree growth in mm/yr. X-axis = year the growth rate was
recorded.
►Indicates the year o f canopy ascension. ► Indicates the year o f the
ambiguous zone.

79

4

3

0
x:

I

CD

2

>s

1
0

0

Su pressed

Gap

Suppressed

Gap

Suppressed

Gap

Betula

Betula

Alyes

Abies

Picea

Picea

Figure 2. Box-plots showing the range (mean, standard deviation, minimum, maximum,
and quartiles) in initial 10-year average growth rate for: gap-origin and suppressed Betula
papyrifera (n = 3,2), Abies lasiocarpa (n = 104,215), and Picea glauca x engelmannii (n
= 56,48).

80

1.) H igh Early Growth Rate A nalysis:
Indication:

Tree O riginated in a Gap

Criteria:

Abies lasiocarpa = 1.72m m /yr
Picea glauca x engelmannii = 1.5m m /yr
Betula papyrifera = 1.072m m /yr
2 .) Rapid, Sustained Increases in Radial G row th R ates (R elea se) A n a ly sis
Indication:

Tree w as suppressed for a period then, due to overhead m ortality w as released.

Criteria:
a)

160% R elease

b)

Sustained for 2 0 years

c)

< 4 0 em dbh for A bies lasiocarpa and < 3 0 em for Picea glauca X engelmannii

3.) O verall Growth Pattern A n alysis
Indieation:

1) Tree O riginated in a Gap

Criteria:
a)

C onsistently high or increasing grow th fo llo w e d b y a decline.

b)

Parabolic grow th pattern w ith peak grow th oecurring w ithin first 25-years.

c)

D eclin in g peaks w ith first peak indicating h igh est grow th w ithin first 2 5 years.

II) Tree w as released from suppression
Criteria:
a)

Inereasing grow th lasting > 2 5 -y ea rs but peak grow th ach ieved after 25
years.

b)

Peak grow th later than first 2 5 -years.

Figure 3. Decision set for determining year of canopy ascension. Each tree was assessed
using each of the criteria. More than one year of canopy ascension per tree is possible.

81

CHAPTER 4. SPATIOTEMPORAL PATTERNS OF SMALL-SCALE
DISTURBANCES IN SUB-BOREAL SPRUCE FORESTS: IMPLICATIONS FOR
PARTIAL CUT HARVESTING AND INONOTUS TOMENTOSUS ROOT DISEASE
4.0 ABSTRACT

In a wet-cool sub-boreal forest east of Prince George, British Columbia, Canada,
fine-scale, 70-year disturbance chronologies were compared for four forest types: oldgrowth and partially cut forests with and without Inonotus tomentosus Fr. (Teng)
infection and mortality. In both old-growth and partially cut forests 46, 10 meter radius
plots (92 total) were centered on dead or cut formerly dominant or codominant Picea
glauca X engelmannii trees. Twenty-three plots for both old-growth and partial cut forests
were centered on dead trees that showed evidence of past Inonotus tomentosus infection
and 23 were centered on uninfected dead trees. Total P. glauca x engelmannii mortality
was approximately 50% lower in the partial cut category since the decade of harvest
(1950-1960), regardless of infection status. In all four forest types decadal mortality
increased from the 1930’s to the 1970’s and then decreased since the 1980’s. The
functional gap size, averaged 16.76 m^ and was independent of dead tree diameter,
species, crown class or agent of mortality. Summed gap-size measures for all trees dying
in a decade indicated that between 6.9 and 8.1% of stand area was made available to
understory trees per decade but was also highly variable among decades. Mean deeadal
mortality was similar to estimates for similar forest ecosystems influenced by small-scale
disturbances.
Due to high mortality rates of large individuals and low recruitment rates to the
canopy for P. glauca x engelmannii, the population structure of the old-growth forests

82

appears to be shifting from a P. glauca x engelmannii dominated canopy to an A.
lasiocarpa dominated canopy, but Inonotus tomentosus does not appear to be the cause.
Rather, higher^, lasiocarpa densities in the understory and more frequent A. lasiocarpa
recruitment to the canopy combined with high rates o f f . glauca x engelmannii mortality
explain this shift. In partial cut plots, higher relative P. glauca x engelmannii recruitment
and lower P. glauca x engelmannii mortality indicate that P. glauca x engelmannii
populations may rise relative to its’ present density.

83

4.1 INTRODUCTION
The Sub-Boreal Spruce (SBS) forest region of central interior British Columbia is
a diverse forest region whose climax forests are mainly mixtures o f Picea glauca x
engelmannii (Parry ex Engelm.) (hybrid spruce, hereafter referred to as spruce) m d Abies
lasiocarpa (Hook.) Nutt, (sub-alpine fir, hereafter referred to as fir) (Meidinger and Pojar
1991). Mean fire return interval estimates in dry and wet SBS forests range from 150-250
years, and 227-6250 years, respectively (Parminter 1992; Hawkes et al. 1997). Long fire
return intervals in wet areas of the SBS enable small-scale (i.e. gap) disturbances to
become a predominant stand dynamics mechanism in these forests. Furthermore, smallscale disturbances may become increasingly important since fire suppression has
excluded or reduced fire’s influence on the ecosystem (Clark 1994; Frelich and Reich
1995; Andison 1996; Kneeshaw and Bergeron 1998).
Quantification of the patterns of small-scale disturbances and the stand dynamics
they cause is important in both managed and unmanaged landscapes. The increasing
scarcity of old-growth forests has inspired the preservation of their remnants (Welles et
al. 1998) yet the long-range implications of gap dynamics on stand structure and
composition are unknown. Secondly, in managed SBS landscapes, landscape ecology
principles that suggest biodiversity can be maintained by mimicking natural disturbance
patterns with harvesting patterns (DeLong and Tanner 1996). If this goal is to be met,
partial cut harvesting systems that approximate small-scale natural disturbance patterns
need to be designed for forests where small-scale disturbances are predominant.
Ecologists are just beginning to realize that small-scale disturbance regimes are important

84

ecological processes in wet SBS forests and quantitative information of these regimes and
how forests respond to them are lacking.
Agents of gap formation in SBS forests include root and stem rot fungi, phloem
feeding insects (e.g. Dendroctonus rufipennis), tree life spans, and abiotic factors such as
windthrow and snow loading (Kneeshaw and Burton 1997; Kneeshaw and Bergeron
1998; Lewis and Lindgren 1999). Among these is the root rot pathogen Inonotus
tomentosus (Fr.) Teng, which is common in most Picea spp. dominated forests world­
wide (Whitney 1980; Merler et al. 1988; Lewis and Hansen 1991a; Lewis et al. 1992;
Lewis 1997). In the SBS forests of central interior British Columbia, /. tomentosus
primarily attacks spruce (Hunt and Unger 1994) and infected trees are subject to chronic
declines in vigor. Eventually root dysfunction and structural weakening lead to standing
mortality or windthrow of individual or small groups of trees (Lewis and Hansen 1991b;
Lewis 1997). Following tree mortality, root colonization by the fungus increases due to
the absence of active defense mechanisms in the sapwood and the fungal mycelium
moves outward from the heartwood to the sapwood (Lewis et al. 1992).
The mycelium in the dead sapwood enables I. tomentosus to spread through rootto-root contact (Lewis et al. 1992). In this way the disease may spread from tree-to-tree
(Lewis et al. 1992) potentially reducing spruce density in disease pockets. Due to the host
preference of I. tomentosus and the ability to remain infectious for up to 40 years, gaps in
the canopy formed by this pathogen may result in higher rates of spruce mortality and
lower spruce populations in the gaps. Thus relative to some other small-scale disturbance
agents I. tomentosus could lead to compositional shifts, due both to decreased density of

85

spruce in the overstory surrounding gaps, and a decreased probability that spruce recruits
will fill the gaps.
It can be hypothesized that compared to the effects of /. tomentosus in old-growth
forests, partial-cut silvicultural systems may exacerbate the effects of I. tomentosus on
gap dynamics, and forest structure. Partial cutting may increase I. tomentosus virulence
and this may increase spruce mortality as inoculum volumes are increased by harvesting
relatively healthy infected spruce which, if left uncut, would be able to confine the fungus
to the heartwood for a much longer period. This could mean increased contact frequency
with new hosts and ultimately, increased mortality rates for spruce in partial cut
silviculture applications.
This research addresses four specific questions about small-scale disturbance regimes
and associated stand dynamics in wet SBS forests.
1. What is the temporal pattern of small-scale disturbance in I. tomentosus infected
and uninfected old-growth forests and how do these compare to infected and
uninfected partially cut forests?
2. Does the rate of spruce mortality differ between /. tomentosus infected and
uninfected old-growth forests and how does partial cutting affect these rates?
3. How has current stand composition been affected by harvesting and/or infeetion
status?
4. Is the future stand composition likely to change due to infection and/or harvesting
status?

86

4.2 METHODS
4.21 Study Area
The research was conducted in two old-growth and two partially cut forests at the
Aleza Lake Research Forest, located at 54° 07’ N, 122° 04’ W, about 60 kilometers east
of Prince George, British Columbia, Canada. It lies between 600 and 750 meters above
sea level on the Nechako Plain of the Fraser River Basin in the Interior Plateau
physiographic region (Holland 1976). For the old growth forests, one stand was sampled
on the north side of the Bear Road approximately 1 km east of the Bear Road and Aleza
Road junction. A second stand was sampled on the west side of the Aleza Road
approximately 2.5kms south of the Bear Road and Aleza Road Junction. For the partial
cut forests one stand was sampled on the west side of the Aleza Road near the junction of
the Aleza Road and the Upper Fraser Road. A second stand was sampled on the east side
of the Aleza Road adjacent to the junction of the Aleza Road and the West Branch Road.
The ALRF is located in the sub-boreal spruce, wet-cool 1 (SBSwkl) biogeoclimatic zone
(See Meidinger and Pojar (1991) for details). The SBSwkl climate is characterized by
cold, snowy winters and moist, cool summers. The climate is slightly wetter and more
moderate than central SBS sub-zones due to the orographic influence of the Northern
Rocky Mountains to the east, resulting in higher precipitation than usual for the rest of
the zone (Meidinger and Pojar 1991). The old-growth forests are mixtures of spruce and
fir with scattered Pseudotsuga menziesii var. glauca (Douglas-fir), Finns contorta var.
latifolia (lodgepole pine) and Betula papyrifera (paper birch). Old-growth forests at the
Aleza Lake Research Forest are thought to be uneven aged (Decie 1957). The dominant
spruce, perhaps members of the initial fire origin cohort, are possibly 300 years old or

87

more. The oldest fir are about 200 - 250 years and are probably members of a post-fire
establishment cohort. The scattered Douglas-fir component can be as many as 500 years
old (Decie 1957) and may be survivors of the last fire. A well-developed understory layer
is mainly comprised of fir (80%) and spruce (20%). Spot fires, spruce beetle
(Dendroctonus rufipennis), various diseases including I. tomentosus and timber
harvesting in some stands have been the main disturbance agents at the Aleza Lake
Research Forest since the last wildfire (Decie 1957).
The partial cut forests at the Aleza Lake Research Forest have the same site
characteristics as those described for old-growth forests. Partial cutting was designed to
improve spruce regeneration success and overall stand structure and quality through
selective harvests conducted during the winters of 1952-1956 (TSX 42765, 774118).
These systems implemented forest management concepts typical of uneven aged
selection systems (e.g. management of species composition, stand structure, and residual
growing-stock) (Jull and Famden unpublished data). The following guidelines were
followed by trained crews prior to harvesting: defective trees were identified for felling;
larger trees were removed where possible; vigorous trees were left; uniform spacing of
residual spruce was attempted with removal of fir where possible; and sufficient volume
was removed to insure an economic operation (DeGrace et al. 1952). The residual basal
area was approximately 50% of initial basal area.

4.22 Site Selection
Forest cover and contour maps were initially used to locate 30 candidate stands
(15 old-growth and 15 partial cut) that were relatively uniform in composition on

88

medium to good sites with minimal variation in soils and topography. The old-growth
stands were undisturbed by human activities and the partial eut sites were selectively
logged between 1952 and 1954. Each of the 30 candidate stands were ground checked for
appropriate characteristics and a number were eliminated because they didn’t meet the
initial criteria. Each stand was also checked for the presence of I. tomentosus using
standard field techniques (Finck et al. 1989). From the original 30 stands, four were
selected (two old-growth and two partial cut) that were most similar amongst themselves,
and best met the criteria above. The stands selected also had low to moderate (5-10%)
incidence of I. tomentosus infection.

4.23 Sampling Design and Measurements
Plots were established in the four selected stands systematically with a 5 0 x 5 0
meter spaced grid. Four forest types were sampled. Old-growth without tomentosus
(OGNT), and partial cut without tomentosus (PCNT) plots were free of any evidence of I.
tomentosus mortality or infection in living or dead trees. OGNT and old-growth with I.
tomentosus (OGT) plots were centered on the nearest dead tree (dead for approximately
40 years) that was a canopy dominant or codominant tree at the time of its death. This
was done to provide similar mortality rate estimates between the old-growth plots and the
partial cut plots which were logged about 40 years before sampling. PCNT plots were
centered on the nearest cut stump to the grid point with no evidence of I. tomentosus
infection or mortality with in the plot boundary. OGT plots were centered on a canopy
tree that died approximately 40 years ago and had I. tomentosus. PCT plots were
established on a cut stump that had evidence of past infection by I. tomentosus in the
roots or the stump surface. In the OGT and PCT forest types all living trees were

89

inspected for I. tomentosus infection and this information was used to determine the
incidence of infection in these forest types. For each forest type, 23, 10 meter radius
(0.314 ha), plots were established.
To quantify the disturbance regime, the location of all dead trees in the plot (> 10
cm dbh), the tree decomposition characteristics required as input for the time since death
model (TSD model) (Chapter 2) and, if possible, the cause of death were collected. For
most disturbance agents, post-hoc classification of mortality is difficult beyond a few
years. However, /. tomentosus causes diagnostic pitting in large roots and stem bases that
can be identified with confidence for at least 40 years after mortality occurs (Lewis and
Hansen 1991a). For several trees it was impossible to determine either the species or the
cause of death. In these cases “unknown” was recorded and time since death was
determined using canopy ascension date (Chapter 3).
Two potential gap-fillers were selected from among the trees subtending each
dead canopy tree and one increment core was taken from each at 1.0 m. These trees were
stem-mapped, assessed for crown class, live crown ratio, diameter, and species. Tree ring
cores were stored in plastic straws, and later mounted on 1 inch thick grooved Styrofoam
strips, dried, sanded and scanned using a flatbed scanner. The scanned images were
analyzed using Windendro® (Regent Instruments, Blaine, Quebec) which measures and
records annual ring width growth (mm). Canopy ascension criteria and the TSD model
were used in combination to estimate a date of death for each dead tree (below).

90

4.24 Data Analysis
Disturbance Chronology
Disturbance chronologies (mortality by decade) were prepared for each forest
type. Chronologies are only presented since 1930, which is the limit of the TSD model
capability and is approximately the limit for including smaller trees due to tree
decomposition. The disturbance chronologies were developed by averaging year of death
estimates from the TSD model (Chapter 2) and canopy ascension dates (Chapter 3), in
potential gap-fillers. When there were no adequate gap-filling trees for the dead canopy
tree only the TSD model was used. When the species of the dead tree could not be
determined or was of a species that could not be modeled by the TSD model, only tree
ring cores were used to date the mortality. Disturbance chronologies were divided into
10-year intervals from the beginning to end of each decade (Frelich and Graumlich
1994). For the partial cut plots, all cut stems were included in the disturbance chronology
and mortality dates for these coincide with the 1950-1960 decade.

Functional Gap Size
Functional gap size is an estimate of the growing space (m^) that a canopy tree’s
mortality {i.e. gap-maker) makes available to the understory trees most likely to attain
canopy status after it dies {i.e. gap-filler). The mean of the distance from gap-maker to
gap-filler data was used as a proxy for the radius of a circle to calculate the average area a
dead tree makes available to understory trees when it dies. Gap size estimates for each
tree dying in a given decade were summed to determine the area made available to
understory trees per decade. This data was then transformed to a percentage of stand area
converted to gaps per decade (sensu Frelich and Graumlich 1994; Yamamoto 1995).

91

Accumulated mortality (stems/ha and basal area)
The total mortality since 1930 for all plots in each forest type was used to
compare differences in accumulated mortality and spruce mortality. Plot level
accumulated mortality was converted to aecumulated basal area mortality because basal
area comparisons may show differences in stand mortality dynamics beyond the
population level (i.e. growth effects). The proportion of spruce basal area mortality over
total basal area mortality and the proportion of spruce mortality caused by I. tomentosus
in infected forests were also determined and compared using the Tukey-Kramer,
Honestly Significant Difference test (Toothaker 1993) (a = 0.05).

Stand Composition
The species tally by diameter class was used to compare compositional
differences in regeneration (trees < 10 cm diameter at breast height), and canopy layers
(trees > 10 cm diameter at breast height). Density by species were calculated as both a
proportion of and as a total stems/ha and basal area (m^/ha). Mean treatment (i.e. OGT,
OGNT, PCT, PCNT) values were compared using the Tukey-Kramer HSD test.

Transition Probabilities
Potential gap-fillers are defined as the two chosen trees subtending each dead
canopy tree that showed the best potential of replacing the dead tree in the canopy
relative to the total pool of regenerating trees (Lertzman 1992). Criteria for selection
included a marked release in annual radial growth, vigorous with a good live crown,
subordinate to the dead tree in the canopy at the time of death, and within close proximity
to the dead tree (<5m) (Chapter 3). One most likely gap-filler was selected from the two

92

potential gap-fillers based on their ability to assert dominance over potential gap-fillers
due to their vigor, growth response and proximity to the dead tree. In cases where no
single tree showed clear dominance over other potential gap-fillers, both were taken.
The sum of the gap-maker/gap-filler transitions was used to develop transition
probabilities for each species in the four forest types. Dead trees without any gap-fillers
(e.g. recent mortality or microsite limitations) were not included in the matrix
(approximately 9% of all dead trees). Gap-fillers associated with more than one gapmaker were assigned a fractional probability of transition based on the number of gapmakers it was replacing (Lertzman 1992). Gap-makers that could not be identified to
species were recorded as ‘other’. Transition probabilities were determined from 19301950, 1950-1970, and 1970-1997 to examine variation in transition probabilities with
time and due to harvest operations.

Simulation of Future Stand Composition
Predicted future stand composition was modeled using the 1970-1997 transition
matrix and the average annual rates of mortality for each species for the same period. A
constant annual mortality rate, constant transition probability, constant stand density, and
the absence of catastrophic or wide-spread canopy level disturbance were assumed for the
model. The current species composition for each forest type for trees >10 cm DBH was
taken, then the annual mortality from the current density for each species was subtracted.
Each annual mortality was then multiplied by the transition probability for each gapfilling species. These results were added to a running total of density for each species to
provide estimates o f population dynamics for a period of 200 years.

93

4.3 RESULTS
4.31 Stand Composition
Eighty-five % of the I. tomentosus infected forest types had less than 6% infection
incidence in living trees. The remaining 15% had between 6% and 15% infection
incidence. Average understory density (trees >30 cm tall and <10em dbh) for the four
forest types was 3305 stems per hectare (p>0.05) (Table 1). Spruce density was
significantly lower in both old-growth forest types than the PCNT forest type and in the
OGT forest type than the PCT forest (Tukey-Kramer HSD, p = 0.0006) (Table 1). In the
canopy layer (trees >10cm dbh), average stand density was 765, 813, 641 and 760
stems/ha for the OGNT, OGT, PCNT, and PCT forest types, respectively (Table 1). The
canopy density for the PCNT forest type was significantly lower than all other forest
types (p = 0.011) but there were no other significant differences between any other forest
type. This may indicate more intense partial cutting in this forest type relative to the other
PCT forest type. Spruce density in the canopy layer was significantly lower in both the
OGT, OGNT and PCT forest types relative to the PCNT forest type (p = 0.027) (Table 1),
suggesting that spruce density may be increased in partial cut forest types, but only when
/. tomentosus is not present.
Average basal area was similar for the four forest types (p = 0.825) (Table 1).
However, the percent spruce basal area of total basal area was highest in the OGNT forest
type compared to the OGT and PCT forest types (p = 0.0132) (Table 1). The PCT forest
type also had a significantly lower percent spruce basal area than in the OGT forest type.
These results indicate that spruce basal area is higher in forest types not infected by I.

94

tomentosus and is higher in old-growth than partial cut forests. The latter result is likely
due to the removal of large diameter-high basal area trees in partial cut stands.

4.32 Disturbance
Disturbance Chronology
Since 1930, mortality has occurred in all decades in all four forest types (Figure
1). Average decadal mortality by decade since 1930 for the OGT, OGNT, PCT and
PCNT forests was 48.07, 44.40, 32.91, 35.25 stems>10cm dbh/ha/decade (not shown).
Average decadal spruce mortality in the 1960’s was 30.12, 28.73, 11.22, and 11.42
stems/decade for the OGT, OGNT, PCT and PCNT forests, respectively. Thus spruce
mortality was more than halved following the harvest in both uninfected and infected
partial cut forests relative to the old-growth forests. In contrast, fir mortality increased
since the harvest in the partial cut forest types but not in the old-growth forest types
(Figure 1). For the partial cut sites and the old-growth sites, spruce mortality, not
including cut trees, was highest in all forests during the 1970’s. Fir mortality was highest
during the 1960’s for the old-growth forests and highest during the 1980’s for the partial
cut forests. All forest types were characterized by increasing total mortality from 1930 to
1970. In the 1980’s and 1990’s the old-growth forests had declining total mortality rates
and the same trend was evident during the 1990’s for the partial cut forests.

4.33 Accumulated Mortality
Mean accumulated mortality (less cut trees) was significantly higher in the OGT
forest type compared to both partial cut types (p = 0.0068) (Table 1). Mean accumulated
spruce mortality was also significantly higher for both old-growth forest types compared

95

to both partial cut forest types (p<0.0001). Average accumulated basal area mortality was
similarly significantly higher for both old-growth forest types compared to both partial
cut forest types (p<0.0001), as was the proportion of spruce mortality of total basal area
mortality (p<0.0001). The proportion of spruce basal area mortality killed by/.
tomentosus was 0.312, and 0.150 in the OGT and PCT forest types, respectively (p =
0.025). Total mortality, total spruce mortality, total basal area mortality, and total and
proportional spruce basal area mortality were all therefore generally reduced by partial
cutting but there is no evidence to suggest that I. tomentosus caused significantly higher
spruce mortality in this study.

4.34 Functional Gap Size
Most likely gap fillers averaged 2.35 m from dead spruce (N = 380) and 2.29 m
from dead fir (N = 490) and species differences were only marginally significant (p
=0.052). Since gap fillers respond to the two gap making species similarly, analyses
presented hereafter are for species pooled. The maximum distance from a gap-maker to a
gap-filler was 5.6 meters and the minimum was 0.05 meters. Ninety-five % of the
releases occurred between 0.55m and 4.44m. There was no significant relationship
between the distance of gap-maker to gap-filler and the diameter of the gap-maker (p =
0 .068 ).

The average gap size was 16.76m^. The 97.5*’’ quantile (r = 4.44m) gap area was
61.90m^ indicating that only 2.5% of the single tree gaps have a functional gap size larger
than 62 m^. The product of mortality rate estimates and mean gap size, indicates that the

96

percentage of stand area made available to new recruits per decade is 8.1%, 7.4%, 7%
and 6.9% for the OGT, OGNT, PCT, and PCNT forest types, respectively.

4.35 Transition Probabilities
In general, transition probabilities indicate that fir gap-fillers outnumber all other
gap-filler species by about 2:1 except in cases where the sample size of a particular gap
maker is small. In old-growth forest types, spruce recruitment does not appear to be
strongly affected by infection status since, in some cases spruce recruitment was greater
in the OGT forest types than in the OGNT forest types (Figures 2,3,4). Furthermore there
were no temporal changes in transition probability over the last 70 years in either oldgrowth or partial cut forest types. In general partial cut forest types had higher spruce
recruitment than old-growth forest types in all cases but the fir to spruce transition in
1970-1997. Furthermore, the spruce recruitment was not predictably higher in the PCNT
forest type relative to the PCT forest type. Therefore spruce recruitment was increased by
partial cutting but was unaffected by I. tomentosus. No interaction o f partial cutting and 7.
tomentosus was found.

4.36 Simulations of Future Stand Compositions
Two hundred year stand composition projections indicate that fir density may
increase from its current proportion of 65-70% to about 80% while spruce density may
decline from 28-30% to about 15% (Table 2). These projections also suggest that birch,
Douglas-fir and western hemlock will continue to be maintained in low densities in these
forests.

97

Assuming no further harvests in the partial cut forest types in the next 200 years
in the PCNT forest type (Table 2), spruce may increase from current relative density of
approximately 33% to 38% and fir may decline from 63% to 57%. Other species should
also be maintained as minor components. In the PCT treatment spruce density will not
rise as it did in the PCNT forest types but will be 7-10% higher than in the old-growth
forest types.

98

4.4 DISCUSSION
Disturbance Chronology
Disturbance chronologies indicate that old-growth, sub-boreal forests are
influenced by continuous low intensity disturbances. It also appears that I. tomentosus
infected and uninfected forests have similar disturbance dynamics. Therefore, at the low
infection levels observed in these stands (typically <6%) (British Columbia Ministry of
Forests 1995), I. tomentosus appears to contribute to the general pattern of low intensity
disturbance by working in concert with other agents to cause single tree to small group
tree mortality. However, it does not increase spruce mortality compared to uninfected
forests. Partial cut forests generally have lower rates of disturbance than old-growth
forests and spruce mortality has been reduced. The effect of higher disease incidence, and
therefore more inoculum, was not tested in this study. However, spruce mortality due to /.
tomentosus could increase with moderate to high infection incidence which may explain
the apparent species shifts observed in stands with higher infection rates (Lewis and
Trammer 2000).
Estimates of disturbance intensity (-7-8% of stand area) by deeade in these oldgrowth sprace-fir forests are similar to several other forest types dominated by small
disturbances. In an old-growth hemlock-hardwood forest in western upper Michigan,
Frelich and Graumlich (1994) reported mean decadal disturbance rates of 5.4%. They
noted many small gaps created in each decade contributed to the low disturbance rate,
which is indicative of individual to small group tree mortality. Yamamoto (1995)
reported that current stand area in gaps for an old-growth sub-alpine coniferous forest in

99

central Japan was 7.3%. Frelich and Lorimer (1991) reported 5.7 to 6.9% mortality in a
hemlock-hardwood forest moderated by light intensity disturbances.
Disturbance rates were highly variable over time ranging from virtually no
disturbance to nearly 15% of stand area. For example, the 1970’s and 1980’s were the
decades of peak disturbance for all forest types. During these decades disturbance was
approximately 13.5%, 14.2%, 7.8% and 9.7% for the OGT, OGNT, PCT, and PCNT
forests respectively. Even these upper estimates of disturbance intensity for the study area
do not approach the levels of disturbance reported for medium to heavy intensity
disturbances caused by wide-spread canopy level mortality caused by insects or
catastrophic windthrow (Frelich and Graumlich 1994; Veblen et al. 1994; Kneeshaw and
Bergeron 1998). Therefore the timing and intensity of gap formation at the Aleza Lake
Research Forest corresponds to continuous, small-scale gap formation consisting of
individual trees or small groups o f trees.

Accumulated Mortality
Total accumulated mortality, total spruce mortality and the proportion of spruce
mortality were all significantly lower in partial cut forest types compared to old-growth
forest types. Thus, the partial cut harvest probably captured potential mortality that would
have occurred if the stands were not cut and also has subsequently improved resource
availability for residual trees and improved survivorship. If spmce mortality was lower in
OGT forest types compared to PCT or PCT and PCNT forest types, there may have been
evidence for an interaction with partial cutting and I. tomentosus. The results here are
opposite, with spruce mortality lower in PCT forest types relative to OGT forest types.

100

and no difference between PCT and PCNT forest types. Therefore it can be concluded
that partial cutting in I. tomentosus infected areas does not increase the spread rate of the
root disease. Further, the evidence shows that spruce mortality can actually he lowered
(-15%) by partial cutting in infected forest types (Table 1). The results of thinning in a
Picea glauca plantation affected by I. tomentosus by Whitney (1993) indicated lower
mortality of Picea glauca in the thinned plantations relative to unthinned controls.
Whitney suggested that thinning increased discontinuity of roots leading to fewer
subsequent infections, and to increased vigor in residual trees that would prolong a trees’
life even if infected by I. tomentosus. In a detailed study on changes in inoculum volume
following harvest, Lewis (unpublished data) did not find a significant increase in total
inoculum volume with age of stump although the root disease did appear to move radially
from the heartwood into the sapwood following harvest.

Stand Composition
Documented understory tree recruitment to the canopy began as early as 1930, but
many gap-filling trees indicated releases occurring much earlier than this. The evidence
of recruitment in this study supports DeGrace (1952), who using permanent sample plot
data, reported that the forests at the Aleza Lake Research Forest were changing from
even to uneven aged during the 1940’s and 1950’s. In a similar spruce-fir forest type near
Smithers and Houston, British Columbia, Kneeshaw and Burton (1997) reported that
older forests also appeared to be successfully maintaining themselves through understory
recruitment in the absence of fire.

101

Stand composition analyses indicate that spruce populations in the understory are
similar for the two old-growth forest types and for the two partial cut forest types (Table
1). These comparisons suggest that I. tomentosus does not affect understory spruce
populations in either old-growth or partial cut stands. In general, partial cutting increased
spruce populations by creating an understory environment more suitable for spruce
colonization and survivorship. This more suitable environment appears to benefit spruce
germination even in I. tomentosus infected forest types.
In the canopy layer, it was evident that I. tomentosus had no effect on spruce
populations for old-growth stands. However, when spruce basal area was considered it
was noted that the OGT forest types had lower spruce basal area than the OGNT forest
types. Similarly, spruce densities in OGT and PCT forest types were not significantly
different but spruce basal area was higher in OGT forest types compared to PCT forest
types (Table 1). Since mortality rates were not substantially different between these
forests and populations were similar, the result may be due to decreased growth rates in
residual P. glauca x engelmannii. Thus there does not appear to be an impact of applying
partial cutting to infected stands on spruce populations but there may be an effect on
spruce volumes. Other studies have reported decreased growth rates of Picea spp. due to
I. tomentosus infections (e.g. Whitney 1980, Merler 1984, Lewis 1997).

Transition Probability and Simulations of Future Stand Composition
Based on the assumptions in the model used here, in the old-growth forests,
spruce density may decline from its current abundance of about 28-30% to around 10%
after 200 years (Table 2). This is due to lower current mortality rates for fir and its high

102

success as a transition species. This projection corresponds with Kneeshaw and Burton
(1997) who noted mixed spruce-fir forests may beeome dominated by fir. In these forests
as well as the forests of the present study, a period of low spruce regeneration, high
spruce mortality and redueed spruce recruitment to the canopy contribute to the eventual
dominance of fir. However since spruce is a longer lived species than fir, the long-term
decline in spruce populations may not be as significant as modeled here. Furthermore
other undefined terms in the model may effect the predictions presented. For example,
higher than modeled sub-alpine fir mortality due to abiotic or biotic factors would open
the canopy and potentially inerease spruee regeneration and reeruitment to the canopy.
However the model does provide a base-line estimate of stand development based on a 70
year average of mortality/recruitment dynamics and is useful as a point-of-eomparison.
In the old-growth forest types recent spruce mortality has been higher than fir
mortality. Given this result and the high rate of fir recruitment over time (Figures 2,3,4) it
is reasonable to assume that fir density has been on the rise over the last 70 years in the
old-growth forests and spruce density has been decreasing. These stand dynamics may
characterize the current serai stage of these old-growth forests where the fire-origin
cohort of spruce has experieneed high mortality rates as they approach the end of their
life-span and are replaced by prolific fir regeneration.
In the partial cut forests the combination of lower spruce mortality, following the
harvest, and generally higher spruce canopy recruitment rates relative to the old-growth
forests should mean that current spruce density is markedly higher in the partial cut
forests relative to the old-growth forests. This is clearly the case for the regeneration

103

layer (Table 1) but not for the canopy layer where spruce density was similar for both
old-growth and partial cut forest types.
However, since the stand improvement harvest specifications were to cut large
diameter stems, spruce density in the canopy would have been dramatically reduced
immediately following the harvest. Even moderate annual diameter growth rates of 2.0 to
3.0 mm/yr (Newbery, unpublished data), in advanced spruce regeneration would take
between 20-25 yrs to reach 10 cm dbh if they were 5 cm dbh at the time of harvest. If the
current old-growth stand composition is indicative of pre-harvest partial cut stand
composition, relative regeneration layer spruce density would only have been around 813% (Table 3). Since current relative regeneration spruce density in the partial cut forests
is 18-21% (Table 3) many of the existing spruce in these forests have germinated since
the harvest. At the same moderate growth rate mentioned above they would take at least
50 years to reach 10 cm dbh and thus were still not counted in the canopy layer tally
conducted in 1997 (about 42 years after harvest). Therefore, there is biological evidence
that suggests spruce density in this layer will also rise because of the harvest over the
long term as modeled by the simulations.

104

4.5 CONCLUSIONS

Overall, the disturbance regime in old-growth forests is one of low intensity
single tree to small group tree mortality which does not differ between /. tomentosus
infected and uninfected gaps. The application of partial cutting to mixed spruce-fir forests
has decreased mortality rates in general and of particular importance, spruce mortality
rates have been decreased in partial cut forests regardless of infection status. However
growth rates of spruce may be reduced in I. tomentosus infected forests. Transition
matrices indicate increasing dominance of fir at the expense of spruce assuming
disturbance regimes and transition success remain consistent in the old-growth forests. In
partial cut forests, spruce populations will generally increase in density.

105

4.6 LITERATURE CITED
Andison, D. 1996. Managing for landscape patterns in the sub-boreal spruce forests of
British Columbia. Ph.D. dissertation. Faculty of Forestry, University of British
Columbia. Vancouver, BC. 178 pp.
British Columbia Ministry of Forests. 1995. Root Disease Management Guidebook.
British Columbia Ministry of Forests and British Columbia Ministry of
Environment. Victoria, British Columbia.
Clark, D.F. 1994. Post-fire succession in the sub-boreal spruce forests if the Nechako
Plateau, central British Columbia. M.Sc. thesis. Department of Biology,
University of Victoria. 198 pp.
Decie, T. 1957. Working plan for the forest experiment station Aleza Lake: for the period
April U*, 1957 to March 3U‘, 1967. British Columbia Forest Service. Prince
George, BC.
DeGrace, L.A., Robinson, E.W. and Smith, J.H.G. 1952. Marking of spruce in the Fort
George Forest District. British Columbia Forest Service, Dept, of Lands and
Forests, Victoria, B.C. Research Note No. 20. 13 pp.
DeLong, S.C. and Tanner, D. 1996. Managing the pattern of forest harvest: lesson from
wildfire. Biodiversity and Conservation 51:191-1205.
Finck, K.E., Humphreys, P. and Hawkins, G.V. 1989. Field guide to pests of managed
forests in British Columbia. Forestry Canada, Pacific and Yukon Region.
Victoria, BC.
Frelich, L.E. and Lorimer, C.G. 1991. Natural disturbance regimes in hemlockhardwood forests of the Upper Great Lakes Region. Ecological Monographs 61 :
145-164.
Frelich, L. E. and Graumlich, L.J. 1994. Age-class distribution and spatial patterns in
an old-growth hemlock hardwood forest. Can. J. For. Res. 24:1939-1947.
Frelich, L.E. and Reich, P.B. 1995. Spatial patterns and succession in a Minnesota
southern-boreal forest. Ecological Monographs. 65:325-346.
Hawkes, B., Vasbinder, W. and DeLong. 1997. Retrospective fire study: fire regimes
in the SBSvk & ESSFwk2/wc3 biogeoclimatic units of northeastern British
Columbia. McGregor Model Forest Association. Prince George, BC. 34 pp.
Holland, S.S. 1976. Landforms of British Columbia: a physiographic outline. Bulletin
48, British Columbia Department of Mines and Petroleum Resources. Victoria,
British Columbia.

106

Hunt, R.S. and Unger, L. 1994. Forest pest leaflet: Tomentosus root disease. Forestry
Canada, Forest Insect and Disease Survey. Pacific forestry Center, Victoria, B.C.
Kneeshaw, D.D. and Burton, P.J. 1997. Canopy and age structures of some old
sub-boreal Picea forests in British Columbia. Journal of Vegetation Science
8:615-626.
Kneeshaw, D.D. and Bergeron, Y. 1998. Canopy gap characteristics and tree
replacement in the southeastern boreal forest. Ecology 4(3):783-794.
Lertzman, K.P. 1992. Patterns of gap-phase replacement in a sub alpine, old-growth
forest. Ecology 78(2):657-669.
Lewis, K.J. and Hansen, E.M. 1991a. Survival of Inonotus tomentosus in stumps and
subsequent infection of young forests in north central British Columbia. Can.
Jour.For. Res. 21:1049-1057.
Lewis, K.J. and Hansen, E.M. 1991b. Vegetative compatibility groups and protein
electrophoresis indicate a role for basidiospores in spread on Inonotus tomentosus
in spruce forests of British Columbia. Can. J. Bot. 69:1756-1763.
Lewis, K.J., Morrison, D.J. and Hansen, E.M. 1992. Spread o f Inonotus tomentosus from
infection centers in spruce forests in British Columbia. Can. J. For. Res. 22:68-72.
Lewis, K.J. 1997. Growth reduction in spruce infected by Inonotus tomentosus in central
British Columbia. Can. J. For. Res. 27:1669-1674.
Lewis, K.J. and Lindgren, B.S. 1999. Influence of decay fungi on species composition
and size class structure in mature Picea glauca x engelmannii and Abies
lasiocarpa in mature sub-boreal forests of central British Columbia. Ecology and
Management 123:135-143.
Meidinger, D. and Pojar, J. 1991. Ecosystems of British Columbia. Research Branch,
Ministry of Forests. Crown Publications Inc. 330 pp.
Merler, H.A. 1984. Tomentosus root rot of white spruce in central B.C. Masters Thesis,
University o f British Columbia, Vancouver.

Inonotustomentosus

Merler, H.A., Shulting, P.J, and VanderKamp, B. 1988.
(Fr.) Teng
in central British Columbia. In Proceedings of the 7th International Conference
on Root and But Rots, Vernon and V ictoria, B.C., August 9-16, 1988. lUFRO
working party s2.G6.Gl. Edited hy D. Morrison. Forestry Canada, Pacific
Forestry Center, Victoria, B.C. 985 pp.

107

Parminter, J.V. 1992. Typical historic patterns of wildfire disturbance by biogeoclimatic
zone. Protection Branch, BC, Ministry of Forests.
Toothaker, L.E. 1993. Multiple Comparison Procedures. Sage University Papers.
London, England.
Veblen, T.T., Hadley, K.S., Nel, E.M., Kitzberger, T., Reid, M., and Villalba, R. 1994.
Disturbance regime and disturbance interactions in a Rocky Mountain sub-alpine
forest. Journal of Ecology 82:125-135.
Welles, R.W., Lertzman, K.P. and Saunders, S.C.. 1998. Old-growth definitions for the
forests o f British Columbia, Canada. Natural Areas Journal 18:279-292.
Whitney, R.D. 1980. Polyporous tomentosus root and butt rot of trees in Canada.
Proceedings of the 5th International Conference on Problems of Root and Butt
Rot in Conifers, Aug. 1978, Kassel, Germany. Edited by L. Pimitri.
Whitney, R.D. 1993. Inonotus tomentosus in Ontario, and the effects of thinning on the
disease. For. Chron, 69(4):445-449.
Yamamoto, S. 1995. Gap characteristics and gap regeneration in sub-alpine old-growth
coniferous forests, central Japan. Ecological Research 10:31-39.

108

Table 1. Summary of stand composition and mortality data. For the partial cut forest
types, the mortality data do not include cut stems.
Layer

OG T (n =
21 )

OGNT (n
= 23)

PCT (n =
22)

PC N T (n
= 23)

3704/
7 8 2 .5 1

3236/
9 4 3 .9 8

3317/
8 5 9 .1 0

2997/
1 1 4 1 .8 6

P ercen t S p ru ce
Density: M ean/SD

12/
5 .8 4

17/
6 .9 3

21/
1 1 .4 2

23/
1 0 .0 4

P ercen t Fir Density:
M ean/SD

92/
6 .4 9

86/
7 .9 3

78/
1 2 .0 2

77/
1 0 .7 8

A v era g e D en sity
(stem s/h a ): M ean/SD

813/
2 1 0 .4 9

765/
1 4 7 .5 7

760/
1 9 5 .4 7

641/
1 3 1 .8 7

P ercen t S p ru ce
D ensity: M ean/SD

26/
1 1 .1 4

3 0/
8 .2 6

23/
1 1 .2 7

33/
1 3 .8 5

A v er a g e B a sa l Area
(m^/ha): M ean/SD

3 6 .2 3 /
2.21

3 6 .5 3 /
2 .1 0

3 8 .2 5 /
2 .1 0

3 8 .5 1 /
2 .0 6

P ercen t S p ru ce B asal
Area: M ean/SD

47/
0 .1 6

57/
0 .0 6

41
/0 .1 6

50/
0 .1 8

A v er a g e Total Mortality
(stem s/h a ): M ean/SD

3 5 5 .9 /
1 1 4 .7

3 4 0 .7 0 /
9 6 .7 2

2 5 3 .4 1 /
1 0 0 .4 8

2 9 2 .4 4 /
1 1 7 .3 9

A v era g e S p ru ce
Mortality (stem s/h a ):
M ean/SD

1 4 8 .1 7 /
7 2 .8 9

1 4 2 .6 3 /
7 1 .8 7

5 9 .5 5 /
5 0 .1 8

5 5 .0 1 /
4 0 .7 4

A v era g e B a sa l Area
Mortality (m 2/ha):
M ean/SD

2 6 .5 2 /
9 .8 0

2 9 .3 3 /
9 .8 0

1 7 .1 6 /
8 .9 9

1 8 .2 8 /
9 .1 8

P ercen t S p ru ce
Mortality of Total
Mortality: M ean/SD

51/
0 .1 9

56/
0 .1 9

31/
0 .2 6

27/
0 .2 9

31/
0 .2 8

N/A

15/
0 .1 9

N/A

Stand Attribute

A vera g e D en sity
(ste m s/h a ; M ean/SD )

R egen eration Layer

C an op y Layer

P ercen t S p ru ce
Mortality C a u se d by

Inonotus tomentosus:
M ean/SD

109

Table 2. Species composition dynamic over the next 200 years for each forest type based
on the average annual mortality rates for each species since 1930 and the transition
probability for each gap-maker obtained from the transition matrices since 1970. The
illustrations assume no catastrophic or high intensity disturbance occurs during the
modeling horizon. OGNT = old-growth without 7. tomentosus caused mortality, OGT =
old with 7. tomentosus caused mortality, PCNT = partial cut forests without 7. tomentosus
caused mortality, PCT = partial cut forests with 7. tomentosus caused mortality.
Treat­
m ent
OGNT
OGT
PCNT
PCT

Current Composition
Spruce
0.30
0.26

0J3
0.23

Fir
0.65
0.70
0.63
0.72

Other
0,05
0.04
0.04
0.05

+ 100 years

+ 50 years
Spruce
0.27

Fir

0.23

0.73
0.62
0.73

0.34

0.23

0.68

Other
0.05
0.04
0.04
0.04

Spruce
0.24
0.20
0.36
0.23

Fir
0.71
0.76
0.60
0.73

+ 200 years
Other
0.05
0.04
0.04
0.04

Spruce
0.17
0.14

0.38
0.24

Fir
0.77
0.81
0.57
0.74

Other
0.06
0.05
0.05
0.02

110

OGT

OGNT

14 0

120
100

S'

80

1930 1940 1950 1960 1970 1980 1990 2000

1930 1940 1950 1960 1970 1980 1990 2000

D e c a d e o f M o r t a lit y

D e c a d e o f M o r t a lit y

Abies
Picea

PCT

PCNT

S-

80

1930 1940 1950 1960 1970 1980 1990 2000

1930 1940 1950 1960 1970 1980 1990 2000

D e c a d e o f M o r ta lity

D e c a d e o f M o r ta lity

Figure 1. Disturbance chronologies for the four forest types. Each bar represents the
average mortality (stems/ha) occurring in each decade for spruce and fir species and total
mortality which includes other species (never exceeds more than 22 stems/ha).

Ill

1930-1950

Figure 2. Transition probability data from 1930-1950 for: OGNT = old-growth without /. tomentosus caused mortality, OGT = old
with I. tomentosus caused mortality, PCNT = partial cut forests without I. tomentosus caused mortality, PCT = partial cut forests with
/. tomentosus caused mortality. In all cases the transitions are in the form of gap-maker : gap-filler. For example the Spruce:Fir
transition indicates a Spruce mortality being replaced by an Fir gap-filler. Data are presented in triads such that each gap-making
species has three possible transition outcomes and there are three possible gap-making species. Thus, for each forest type, transition
probabilities total one for each species of gap-maker. (N < 5 for Spruce gap makers in the PCNT forest type).

112

1950-1970

8
2
Q.
(/)
8
2
Q.
CO

Ll
8
3
Q.
CO

0)
5
O
<D
Ü
2
CL
CO

Q.

(/)

<D

Figure 3. Transition probability data from 1950-1970 for: OGNT = old-growth without/.
caused mortality, OGT = old
with I. tomentosus caused mortality, PCNT = partial cut forests without /. tomentosus eaused mortality, PCT = partial cut forests with
I. tomentosus caused mortality. In all cases the transitions are in the form of gap-maker : gap-filler. For example the Spruce:Fir
transition indicates a Spruce mortality being replaced by an Fir gap-filler. Data are presented in triads such that each gap-making
species has three possible transition outcomes and there are three possible gap-making species. Thus, for each forest type, transition
probabilities total one for each species of gap-maker. (N < 5 for Other gap makers in the PCNT forest type).

113

1970-1997

lÏÏT
0)
8
2
Û.
CO

LL

8
2
Q.
CO

CO

8
13_

Cl

03
.C
O
0)
o
2
Q.
CO

q

Q.
CO

ÜL

0)
Ü
2

Q.
(O

x:

m
.c
O

Figure 4. Transition probability data from 1970-1997 for: OGNT = old-growth without I. tomentosus caused mortality, OGT = old
with I. tomentosus caused mortality, PCNT = partial cut forests without I. tomentosus caused mortality, PCT = partial cut forests with
/. tomentosus caused mortality. In all cases the transitions are in the form of gap-maker : gap-filler. For example the SpruceiFir
transition indicates a Spruce mortality being replaced by an Fir gap-filler. Data are presented in triads such that each gap-making
species has three possible transition outcomes and there are three possible gap-making species. Thus, for each forest type, transition
probabilities total one for each species of gap-maker. (N < 5 for Other gap makers in all forest types).

114

CHAPTER 5. STAND LEVEL SPATIO-TEMPORAL DISTURBANCE
PATTERNS CAUSED BY INONOTUS TOMENTOSUS AND OTHER AGENTS IN
SUB-BOREAL SPRUCE-FIR FORESTS
5.0 ABSTRACT
The small-scale canopy level disturbance regime in old-growth sub-boreal forests
was compared for three stands with Inonotus tomentosus (Fr.) Teng caused mortality (510% infection incidence of infection) and three stands without any evidence of the root
pathogen. The spatial and temporal patterns of canopy disturbance and the patch structure
of trees were quantified from 0.49ha stem-mapped plots using spatial autocorrelation
analysis (Moran’s 1 and Standard Normal Deviates). Spatio-temporal patterns of canopy
disturbance and canopy composition were similar for I. tomentosus infected and
uninfected stands. For infected and uninfected stand types pooled, canopy disturbance per
decade ranged from a maximum of 6.0% to a minimum of 5.09%. Canopy gaps averaged
<7m in diameter but larger gaps up to about 28m in diameter also were found. Gaps were
irregularly distributed throughout the plots over the last 250 years. Species patch structure
analysis indicated that Picea glauca x engelmannii (Parry ex Engelm.) was more likely to
be spatially associated with itself than with other species, whereas Abies lasiocarpa
(hook.) Nutt, had neither positive nor negative spatial associations with itself or with
Picea glauca x engelmannii. These results show that small-scale disturbances are
important successional mechanisms in sub-boreal spruce forests and that small scaledisturbance characteristics do not differ between forests with and without I. tomentosus
infection.

115

5.1

INTRODUCTION

Across the landscape of British Columbia’s central interior, sub-boreal forest
regions were historically characterized by mature forests dominated by Picea glauca x
engelmannii (Parry ex Engelm.) (hybrid spruce, hereafter referred to as spruce), Abies
lasiocarpa (Hook.) Nutt, (sub-alpine fir hereafter referred to as fir), and Pinus contorta
(Hawkes et al 1997). The prevailing disturbance regime notion is that catastrophic fire
came at intervals about equal to tree longevity (80-200 years) in this forest type, resetting
succession over large areas before small disturbances had much effect on stand structure
(Johnson 1992). This view of historical disturbance frequency maintains that forests were
dominated by even aged, single storied stands, and had few uneven-aged, old-growth
forests (Oliver and Larson 1996).
However, the sub-boreal forest region in central British Columbia is climatically
and topographically diverse and this diversity strongly affects disturbance regimes. For
example, in wet montane, sub-boreal and sub-alpine spruce-fir forests, catastrophic stand
replacing disturbances occur between 227-6250 years whereas in dry forests of the same
types, mean fire return intervals are about 200 years (Hawkes et al. 1997; DeLong and
Tanner 1996). Small-scale disturbances may be major factors influencing succession and
stand development in areas with long fire return intervals.
The contemporary dominance of clear-cut silviculture systems in both dry and
wet spruce-fir forests in British Columbia has caused an increase in early serai, even aged
stands across the landscape over the last 60 years. In many managed landscapes, these
young forest types are overwhelmingly abundant relative to their pre-settlement
occurrence (Alverson et al. 1988; Mladenoff et al. 1993; Parminter and Daigle 1997).

116

Concern over this alteration of natural landscape characteristics and stand level structure
has lead to the demand for silvicultural approaches that maintain the natural landscape
mosaic of stand types and structures whilst maintaining timber supplies in resource
dependent communities. Therefore ecologists have been challenged to characterize the
small-scale disturbance regime as a prerequisite for the development of appropriate
silvicultural methods.

In sub-boreal spruce forests, several agents are responsible for small to medium­
sized disturbances including: root diseases, phloem-feeding insects, wind, snow, ice and
edaphic conditions. Inonotus tomentosus (Fr.) Teng is a common and economically
significant root disease pathogen in these forests (Lewis 1997). I. tomentosus primarily
attacks spruce, and infected trees are subject to chronic declines in vigor (Hunt and Unger
1994). Eventually tree mortality occurs due to dysfunctional roots, or windthrow due to
weakened root systems (Lewis and Hansen 1991; Lewis 1997).

Due to its ability to spread from tree to tree via root-to-root contact, its preference
for spruce as a host, and a high rate of blow-down in trees infected with /. tomentosus, the
disease may cause aggregated disturbance patterns rather than randomly dispersed
patterns in a stand. This aggregated pattern of disturbance may have implications for gap
arrangement and demography (Qi and Wu 1996; Bellehumeur et al. 1997). For example,
aggregation of I. tomentosus disease pockets may allow species of trees to occupy the
gaps that are resistant to infection {i.e. fir). Thus the spatial pattern of I. tomentosus
caused disturbance may lead to clusters of fir forming the canopy in I. tomentosus
infection centers, and spruce persisting only outside the centers. Over time as the

117

infection centers eoalesce, spruce populations may be diminished significantly compared
to uninfected forests.

In a companion study (Chapter 4), 10 meter fixed radius plots were established in
both infected and uninfected old-growth forests in order to assess fine-scale patterns of
tree mortality and subsequent tree replacement tendencies. Disturbance affected an
average o f approximately 8.1% and 7.4% of stand area per decade for/, tomentosus
infected and uninfected old-growth stands, respectively. This study found that I.
tomentosus, at low (<6%) incidence, did not affect spruce mortality rates or subsequent
gap-maker - gap-filler transition probabilities relative to the remaining array of
disturbance agents affecting the stands at Aleza Lake. However, this gap-level study did
not address the spatial pattern {i.e. gap dispersion) of the gaps themselves or subsequent
species patch structure at larger scales. Therefore the objectives of this paper are to
describe and compare the spatial and temporal patterns of natural disturbance and species
patch structure for three, 0.49 hectare /. tomentosus infected stands and three uninfected
old-growth spruce-fir stands located in wet-cool, sub-boreal forests in east central British
Columbia. These objectives are addressed by developing canopy tree stem-maps and then
analyzing spatio-temporal patterns of disturbance and species spatial patterns with
Moran’s I and Standard Normal Deviates. This approach focuses on the cumulative
effects of gap formation at the stand level. Therefore this approach provides information
on the small-scale disturbance regime at a higher scale than in gap-phase studies but at
lower scale than in landscape studies.

118

5.2 METHODS

5.21 Study Area
The research was conducted in old-growth stands at the 10 235 hectare Aleza
Lake Research Forest, located at 54° 07’ N, 122° 04’ W, about 60 kilometers east of
Prince George, British Columbia, Canada. It lies between 600 and 750 meters above sea
level on the Nechako Plain of the Fraser River basin in the Interior Plateau physiographic
region (Holland 1976). The Aleza Lake Research Forest is located in a wet, cool sub­
zone of the sub-boreal forest and is classified as the Sub-Boreal Spruce, wet-cool sub­
zone (SBSwkl) according to a biogeoclimatic system commonly used in British
Columbia (Meidinger and Pojar 1991). The sub-zone is characterized by cold, snowy
winters and moist, cool summers. The climate is slightly wetter and more moderate than
typical for other sub-boreal spruce sub-zones due to the orographic influence of the
Northern Rocky Mountains to the east, resulting in higher precipitation than usual for the
rest of the zone (Meidinger and Pojar 1991).
The old forests are mixtures of spruce and fir with scattered Pseudotsuga
menziesii var. glauca, Pinus contorta var. latifolia and Betula papyri/era. Old forests at
the Aleza Lake Research Forest are thought to be uneven aged (Decie 1957). The
dominant spruce, perhaps members of an initial fire origin cohort, are as many as 300
years old. The oldest fir are about 200 - 250 years old and are probably members of a
post-fire establishment cohort. The scattered P. menziesii component can be as many as
500 years old (Decie 1957) and may be survivors of the last fire as well as some younger
individuals which successfully regenerated in the understory. The well developed

119

understory layer is mainly comprised of fir (80%) and spruce (20%) (Chapter 4). Spot
fires, insects {Dendroctonus rufipennis), various diseases, (including I. tomentosus),
timber harvesting, windthrow, and snow or ice damage have been the main disturbance
agents at the Aleza Lake Research Forest since the last wildfire (Decie 1957).

5.22 Site Selection
Forest cover stratification: Forest cover maps were initially used to locate 12
stands that were large {i.e. > Vi hectare) and relatively homogeneous mixtures of spruce
and fir on medium to good sites, undisturbed by human activities, with minimal variation
in soils and topography. Each of the 12 stands was ground checked for appropriate
characteristics and a number were dropped because they didn’t meet the initial criteria.
Each stand was checked for the presence of I. tomentosus using standard field techniques
(Finck et al. 1989). In living trees, the primary diagnostic of I. tomentosus are large,
longitudinal pits formed in roots with advanced decay and the presence of dark reddishbrown stain in the roots with relatively recent infection (Finck et al. 1989). After a tree
dies, the longitudinal pits can be used to confirm I. tomentosus for 30-40 years before
decomposition makes diagnosis less reliable.

Three stands were selected with relatively abundant I. tomentosus infection (510%) and mortality throughout and three stands were selected without any apparent I.
tomentosus infection. Plot 1 was located in a stand on the west side of the Aleza Road
approximately 2.5kms south of the junction of the Bear Road and the Aleza Road. Plot 2
was located in a stand on the north side of the Bear Road approximately 1km east of the
junction of the Bear Road and the Aleza Road. Plot 3 was located on the south side of the

120

East Loop Road approximately 500m east of the junction of the Old Ranger Road and the
East Loop Road. Plot 4 was located in a stand on the East side of the Aleza Road about
l.Skms south of the switchback approaching Camp Creek. Plot 5 was located in a stand
on the south east side of the Bear Road (near an secondary access road) about 3.5kms
south-west of the Aleza Road and Bear Road junction. Plot 6 was located in a stand on
the east side of the Aleza Road about 700m south of the Aleza Road and Upper Fraser
Road junction.

In all cases, the plots were established on flat areas, with well-drained loamy
soils. Indicator plant species present on the plots suggested that the sites were broadly
mesic in moisture status and medium to rich in nutrient status (DeLong 1996). Soil
moisture and nutrient status were not determined quantitatively because qualitative
assessments of edaphic characteristics provide acceptable means of characterizing
relative soil moisture and nutrient status within the same biogeoclimatic sub-zone (Klinka
et al. 1989).

5.23 Sampling design and plot measurements
Within each of the six identified stands, one 70 x 70 meter plot (0.49 ha) was
established. This plot size was selected because it was large enough to include many
gaps. Each plot was sub-divided into a grid network with 7 x 7 meter spacing along each
axis using a laser-surveying instrument (Criterion 400, Laser Technology Inc.). The
seven meter spacing was chosen to sample at least 100 trees in each plot. There were 11
lines along each axis for a total of 121 grid points located at the intersection of each line.

121

At each grid point a wooden stake was labeled according to its location on the grid {i.e.
x,y coordinates) and driven into the ground.
The canopy tree (defined as any tree whose crown receives direct sunlight from
above) whose stem was closest to each grid point was then mapped relative to the grid
point and was coded according to species, diameter, and live crown ratio (Frelich and
Graumlich 1994). This canopy criteria was used because trees whose crowns receive
direct light from above may have been released due to a gap-making event. Trees whose
crowns do not receive direct light from above are assumed to be suppressed by the crown
of neighboring trees. Selection of the nearest stem, rather than using the nearest crown
results in the best estimate of original gap area before lateral crown expansion of
bordering trees into the gap (Frelich and Martin 1988; Lorimer and Frelich 1989). From
each canopy tree, one increment core was taken at 1.0 m to avoid, as much as possible,
losses in core information due to butt rot. Cores terminating within 2 cm of the estimated
location of the pith were considered complete and the total age was extrapolated from the
earliest five years growth to the pith (Frelich and Graumlich 1994; Frelich and Reich
1995). If a complete core could not be taken from a tree after several attempts, the next
closest tree was selected. The cores were labeled according to the grid point location,
stored in plastic straws, dried, mounted on grooved rigid foam strips, sanded, and
scanned. Ring width increments were digitally scanned and analyzed with Windendro®
(Regent Instruments, Blaine, Quebec). Increment data was graphed for visual inspection
of radial growth patterns.

122

5.24 Interpretation o f Growth Patterns
Three general criteria were used to relate growth rate patterns to canopy mortality
(Chapter 3). The tree criteria were: 1) High rates of early growth (1.5mm/yr for spruce,
1.7mm/yr for fir, and 1.07mm/yr for birch) indicate a tree was growing in a gap (created
by a disturbance) when it was very young. 2) Tree release (>65%, sustained for 15 years,
preceded by 15yrs of slow growth, for spruce <40cm dbh and for fir and birch <3 0cm
dbh) indicates a tree was suppressed in the understory and then released by an overhead
mortality. 3) Interpretations of overall growth patterns (see Chapter 3) indicating either
gap origin status or release were used to determine the decade of canopy ascension for
canopy trees not meeting gap-origin or release criteria.

5.25 Analysis for Spatial Patterns of Disturbance
The spatial patterns of canopy ascension dates were used to describe the spatial
patterns of small-scale disturbance in the 70x70 meter grid plots. This approach does not
require knowledge of locations of canopy trees when they died, but rather assumes that
canopy ascension dates coincide with overhead mortalities in the vicinity. Spatial
autocorrelation analysis using Moran’s 1 is used to test for spatial independence in canopy
ascension date at one locality relative to adjacent localities. Adjacency is specified for the
analysis depending on the inter-tree distance and the lag distance set for the analysis
(Sokal and Oden 1978a; Sokal and Oden 1978b; Legendre and Fortin 1989; Frelich et al.
1993; Frelich and Reich 1995). Thus, trees within the set distance lag are considered
adjacent and those outside the lag are ignored. For this analysis a cumulative distance lag
is employed and the analysis determines the Moran’s 1 statistic for the decade of canopy

123

ascension for any tree within 7 meters of another. Each additional lag includes the
previous lag distance plus 7 meters.

Interpreting Moran’s I is similar to interpretations for the standard correlation
coefficient (Sokal and Oden 1978a; Sokal and Rohlf 1995; Frelich et al. 1993; Frelich
and Reich 1995). This statistic represents the strength of spatial association for canopy
ascension dates for a defined distance lag (Frelich and Graumlich 1994). Moran’s I was
determined for each distance lag (i.e. 0-7m 0-14m, 0-2Im, etc.) and a correlogram was
construeted for each plot up to 70 meters. Each correlogram plots the Moran’s I statistic
for each distance lag on the ordinate axis and the distance lag on the abscissa. High
positive values of Moran’s I at short distances indicate that canopy ascension date is
strongly correlated with canopy ascension dates in neighboring trees. Values of Moran’s I
will gradually decrease as distance increases in uneven-aged forests because date of
release will not be reliably predicted by a neighbor’s date of release. Significance testing
of the correlograms was done first at the global level by determining if at least one pvalue for individual Moran’s I coefficients was significant at the global level. This was
done by using the Bonferroni correction method for multiple tests (e.g. a = 0.05/number
of distance lags (11)) (Sokal and Oden 1978a). If the global test is has at least one
significant Moran’s I coefficient then the point at which Moran’s I becomes no longer
significant (a = 0.05) can be interpreted as the average patch size diameter created by
disturbance (Frelich et al. 1993; Frelich and Reich 1995). Globally insignificant

correlograms are interpreted that average patch size diameter is less than the distance lag.

124

spatial autocorrelation analysis was also used to determine the tendency towards
association or disassociation between pairs of unlike or like species. (Sokal and Oden
1978a; Frelich et al. 1993; Frelich and Reich 1995). With species data, a join-counts
statistic was calculated for spruce-fir joins, spruce - spruce joins, and fir-fir joins. The
join-counts calculation is done between one sampled canopy tree at a given location and
each remaining canopy tree immediately surrounding it on the grid in all directions within
7 meters. This connection scheme is a queens connection matrix (Sokal and Oden 1978a).
Significance testing (a = 0.05) is performed by calculating a standard normal deviate
(S.N.D) for each join possibility. Significant positive spatial autocorrelation for a joinpair is indicated by S.N.D’s > 2.0. Significant negative spatial autocorrelation for ajoinpair is indicated by S.N.D’s < -2.0. No significant autocorrelation (p>0.05) of species is
indicated by S.N.D’s between -2.0 and 2.0 (Moran 1948; Sokal and Oden 1978a; Reed
and Burkhart 1985; Frelich and Reich 1995).

125

5.30 RESULTS

5.31 Canopy Composition
Spruce canopy composition averaged 29.9% for the three I. tomentosus infected
plots (n = 345), while fir canopy composition was 66.1%. I. tomentosus incidence was
estimated at between 5-10%. For the three uninfected plots (n = 344), spruee density was
32.6%, while fir density was 58.7%. No significant difference

(1,634) = 0.0957) (a =

0.05) was found for the proportion of spruce in the canopy between the two forest types.

5.32 Canopy Ascension Type by Species
From the 334 canopy trees sampled in the I. tomentosus plots, there were 369
canopy ascensions and for the uninfected plots there were 401 canopy ascensions
recorded from 336 increment cores (Table 1). No significant difference (%^ (1,243) =
0.1523) was found between infeeted and uninfected forest types in the proportion of gap
origin verses release events for spruce or fir (Table 1).

5.33 Canopy Ascension by Decade
In each forest type, canopy ascensions were recorded as early as 1675 (Figure 1
and 2). For plot one (Figure 1) the species ascending to the canopy between 1670 and
1680 was a spruce and for plot four (Figure 2) the species was a Douglas-fir. Very little
mortality was recorded before 1795, which is likely because trees which would have
responded to the mortality prior to that have died and because the criteria used for canopy
ascension are vigorous enough to avoid detecting canopy thinning events in when the
stand was younger.

126

With the exception of plot 2 and 6 (Figure 1), the highest rate of canopy ascension
occurred between 1960 and 1970. For plots and 6 the highest rate of canopy ascension
occurred between 1970 and 1980. The maximum canopy ascension in these decades
ranged from a high of 52.50% of stand area in plot 3 to a low of 14.05% in plot one.
Since the canopy ascensions recorded before 1800 were sporadic, average estimates of
stand area converted to gaps in each decade (canopy turnover) were calculated from 1800
to 1998. Average canopy turnover rates were fairly consistent for all six plots, ranging
from a maximum of 6.00% in plot 3 to a minimum of 5.09% in plots 2,5 and 6.

5.34 Spatial Patterns of Canopy Disturbance
Figures 3 and 4 show some degree of clumping in ascension cohorts. The spatial
pattern of canopy ascension since about 1750 is generally characterized by small gaps
created in a given decade or couple of decades interspersed with canopy ascension dates
separated by a few decades or more. A global significance test for the correlograms using
the Bonferroni correction method indicted that at least one p-value for a Moran’s I
coefficient in a correlogram must be <0.005. Only Plot 3 (Figure 6) had any significant
Moran’s I coefficients at the global level. However, the significant Moran’s I coefficients
were close to zero which indicate there was no similarity in disturbance date. Therefore,
all six correlograms indicate that average gap size caused by disturbance is less than 7
meters in diameter. No differences in correlogram structure between I. tomentosus plots
and uninfected plots were evident indicating that the pattern of disturbance in /.
tomentosus stands is similar to uninfected stands.

127

5.35 Spatial Patterns of Species Patches
Standard normal deviates (S.N.D’s) (Table 2) indicate that: 1) positive spatial
autocorrelation {i.e. clumping) generally exists for spruce; 2) Fir-spruce spatial patterns
are random and; 3) there is a weak trend for fir to be negatively associated with itself.
There does not appear to be an effect of I. tomentosus on the species patch structure since
fir has a random arrangement, as evidenced by the not statistically significant (a>0.05),
low-order negative values of the indicated S.N.D’s (Table 2).

128

5.4 DISCUSSION
5.41 Spatial Patterns of Disturbance
Fir was the most frequent species entering the canopy in all six plots since canopy
replacement began at about 1750. However, fir was associated with canopy ascensions
occurring in more recent decades than in earlier decades where spruce was more
common. The same trend was reported in Chapter 4 where transition probability data
suggested that fir has become the dominant species of understory tree currently replacing
canopy trees (regardless of species) for the last 50 years. It is evident then that canopy
composition of fir has been increasing in recent decades. It can be speculated that
ultimately a shift from a spruce and fir mixed forest to a fir dominated forest is likely.
Since the incidence of infection in these stands was low to moderate (5-10%), small
patches of disturbance irregularly distributed were noted. Given a higher (>15%)
incidence o f infection it is possible that gaps created would be larger due to coalescence
of infection centers (Lewis et al. 1992).
Gap size averages <7m in diameter but varies from small (<7m diameter) to fairly
large (28m diameter) openings created in the canopy (Figures 3, 4, 5, and 6). The
openings are irregularly distributed over the stands. This type of spatio-temporal pattern
of disturbance suggests that windthrow or windsnap, ice and snow damage, root diseases,
senescence and endemic populations of spruce bark beetle have been the predominant
disturbances in these stands since the last catastrophic disturbance. The result is a patch­
work pattern of disturbance which has enabled understory recruitment to the canopy.
Note that the locations of canopy ascension dates (Figures 3 and 4) clearly show larger
gaps than the Moran’s I analysis indicates. Moran’s I analysis only interprets average gap

129

size. Thus, the larger patches of disturbance especially during the 1960-1980 (Figures 3
and 4) period represent variation in the disturbance regime, but the higher frequency of
smaller gaps than larger gaps in these forests bring average gap-size diameter below 7m.
The discrepancy between correlogram structures (Figures 5 and 6) and the physical
representation of the data (Figures 3 and 4) suggests that larger scale disturbances,
possibly caused by more wide spread bark beetle, snow, ice or windthrow occasionally
influenced stands but overall the disturbance regime is one of single tree mortality
dispersed irregularly throughout the forests.

5.42 Species Patch Structure
Spruce was 28.4% and 32.6% of canopy density (stems/hectare) for the infeeted
and uninfected stands respectively, was significantly positively correlated with itself.
Therefore spruce is more likely to be associated with itself than other species. Fir:Fir
joins were not statistically significant as were Spruce:Fir joins. Several ecological
explanations are plausible for the statistical grouping of spruce as shown. Spruce
grouping may be linked to a combination of the following: 1) the establishment of spruce
regeneration on nurse logs, 2) exposed mineral soil caused by root tip-ups due to
windthrow, or 3) the grouping of a residual old spruce cohort.

130

5.5

CONCLUSIONS

Based on spatial-autocorrelation analysis of release dates in understory trees,
canopy gaps are small and irregularly located in the wet Sub-Boreal Spruce forests at
Aleza Lake. These forests are affected by a variety of disturbance agents. While I.
tomentosus is a significant cause of tree mortality in some stands, its pattern of
disturbance was not found to be unlike other disturbance patterns based on this analysis.
The combined impacts of all the disturbances have enabled understory
recruitment to the canopy as early as the late 1600’s. In recent decades significant
mortality (as high as 52.5%) has occurred and although this rate happened in only one
plot, disturbances over 10% of stand area occurred quite often. This rate of canopy
mortality suggests that the forests may be changing from a multi-aged old-growth stand
to an uneven-aged old-growth forest, defined here as a forest in which all of the original
cohort has died and has since been replaced by understory species.
Current fir canopy composition is 66% in /. tomentosus infected stands versus
59% in uninfected stands. However, much of the replacement that has been occurring
over the last 50 years, consists of fir. Therefore spruce composition in the canopy has
been declining and will likely continue to decline due to the small gap disturbance regime
which suites fir shade tolerance and seed germination strategies better than spruce.
However greater life-spans for spruce will help to maintain it in the canopy (Lewis and
Lindgren 1999), particularly in these stands with low I. tomentosus incidence. Species
patch analysis indicated that spruce was positively associated with itself which is
probably due to clumping of spruce on nurse logs or exposed mineral soil cause by root
tip-ups and the grouping of a residual old spruce cohort. The decline of spruce due to a

131

variety of disturbance mechanisms suggests that over time the old-growth forests will be
dominated by fir in canopy composition and spruce may decline to about 10% of stand
composition (Chapter 4).

132

5.6 LITERATURE CITED
Alverson, K.K., Wallin, D.M. and Solheim, S.L. 1988. Forest too deer; edge effects in
northern Wisconsin. Conservation Biology 2: 348-358.
Bellehumeur, C., Legendre, P. and Marcotte, D. 1997. Variance and spatial scales in a
tropical rain forest: changing the size of sampling units. Plant Ecology. 130:
89-98.
Decie, T. 1957. Working plan for the forest experiment station Aleza Lake: for the period
April U*, 1957 to March 3 L', 1967. British Columbia Forest Service. Prince
George, BC.
DeLong, S.C. 1996. A field guide for identification and interpretation of ecosystems:
draft field guide insert for site identification and interpretation for the southeast
portion of the Prince George Forest Region. British Columbia Ministry of
Forests, Victoria, BC.
DeLong, S.C. and Tanner, D. 1996. Managing the pattern of forest harvest: lessons
from wildfire. Biodiversity and Conservation 51191-1205.
Finck, K.E., Humphreys, P. and Hawkins, G.V. 1989. Field guide to pests of managed
forests in British Columbia. Forestry Canada, Pacific and Yukon Region.
Victoria, BC.
Frelich, L.E. and Martin, G.L. 1988. Effects of crown expansion into gaps on
evaluation of disturbance intensity in northern hardwood forests. Forest Science
34: 530-536.
Frelich, L.E., Calcote, R.R., Davis, M.B. and Pastor, J. 1993. Patch formation and
maintenance in an old-growth hemlock-hardwood forest. Ecology 74(2): 513527.
Frelich, L. E. and Graumlich, L.J. 1994. Age-class distribution and spatial patterns in
an old-growth hemlock hardwood forest. Can. J. For. Res. 24: 1939-1947.
Frelich, L.E. and Reich, P.B. 1995. Spatial patterns and succession in a Minnesota
southern-boreal forest. Ecological Monographs. 65: 325-346.
Hawkes, B., Vasbinder, W. and DeLong, C.G. 1997. Retrospective fire study: fire
regimes in the SBSvk & ESSFwk2/wc3 biogeoclimatic units of northeastern
British Columbia. McGregor Model Forest Association. Prince George, BC. 34
pp.
Holland, S.S. 1976. Landforms o f British Columbia: a physiographic outline. Bulletin
48, British Columbia Department of Mines and Petroleum Resources. Victoria,
British Columbia.

133

Hunt, R.S. and Unger, L. 1994. Forest pest leaflet: Tomentosus root disease. Forestry
Canada, Forest Insect and Disease Survey. Pacific forestry Center, Victoria, B.C.
Johnson, E.A. 1992. Fire and vegetation dynamics. Cambridge University Press.
Cambridge, England.
Klinka, K., Krajina, V.J., Ceska, A. and Scagel, A.M. 1989. Indicator plants of coastal
British Columbia. University of British Columbia Press. Vancouver, British
Columbia.
Legendre, P. and Fortin, M.J. 1989. Spatial pattern and ecological analysis. Vegetatio 80:
107-138.
Lewis, K.J. and Hansen, E.M. 1991. Vegetative compatibility groups and protein
electrophoresis indicate a role for basidiospores in spread on Inonotus tomentosus
in spruce forests of British Columbia. Can. J. Bot. 69:1756-1763.
Lewis, K.J., Morrison, D.J. and Hansen, E.M. 1992. Spread o ï Inonotus tomentosus from
infection centers in spruce forests in British Columbia. Can. J. For. Res. 22:68-72.
Lewis, K.J. and Lindgren, B.S. 1999. Influence of decay fungi on species composition
and size class structure in mature Picea glauca x engelmannii and Abies
lasiocarpa in mature sub-boreal forests of central British Columbia. Ecology and
Management 123:135-143.
Lewis, K.J. 1997. Growth reduction in spruce infected hy Inonotus tomentosus in central
British Columbia. Can. J. For. Res. 27: 1669-1674.
Lorimer, C.G. and Frelich, L.E. 1989. A methodology for estimating canopy disturbance
frequency and intensity in dense temperate forests. Can. Jour. for. Res. 19: 651663.
Moran, P.A.P. 1948. The interpretation of statistical maps. Journal of the Royal
Statistical Society, Serb. 10:243-251.
Meidinger, D. and Pojar, J. 1991. Ecosystems of British Columbia. Research Branch,
Ministry of Forests. Crown Publications Inc. 330 pp.
Mlandenoff, D.J., White, M.A., Pastor, J. and Crow, T.R. 1993. Comparing spatial
patterns in an unaltered old-growth and disturbed forest landscape. Ecological
Applications 3: 294-306.
Oliver, C D. and Larson, C.B. 1996. Forest Stand Dynamics. John Wiley and Sons.
Toronto. 527 pp.

134

Parminter, J. and Daigle, P. 1997. Landscape Ecology and Natural Disturbances:
Relationships to Biodiversity. Ministry of Forests, Victoria British Columbia.
Research Extension Note #10
Qi, Y. and Wu, J. 1996. Effects of changing spatial resolution on the results of
landscape pattern analysis using spatial autocorrelation indices. Landscape
Ecology. 11(1): 39-49.
Reed, D.D., and Burkhart, H.E. 1985. Spatial autocorrelation of individual tree
characteristics in loblolly pine stands. Forest Science. 31(3):575-587.
Sokal, R.R. and Oden, N.L. 1978a. Spatial autocorrelation in biology: 1 methodology.
Biological Journal of the Linnean Society. 10: 199-228.
Sokal, R.R. and Oden, N.L. 1978b. Spatial autocorrelation in biology: 2 some biological
implications and four applications of evolutionary and ecological interest.
Biological Journal of the Linnean Society. 10: 229-249.
Sokal, R.R. and Rohlf, F.J. 1995. Biometry: the principles and practice of statistics in
biological research. Third edition. W.H. Freeman and Company. New York.

135

Table 1. Proportion o f canopy ascension types by treatment type, and species.
T yp e o f a s c e n s io n
T reatm ent

S p e c ie s

R e le a s e

G ap

Total

0 .5 6
0.41
0 .6 2

113
242

0 .0 0
1.0 0
0 .4 7

0
1

Infected
Infected
Infected

S p ru ce
Fir

0 .4 4
0 .5 9

Birch

0 .3 8

Infected
Infected
U ninfected
U ninfected

H em lock
D ouglas-fir

0 .0 0
0 .0 0

S p ru ce
Fir

0 .5 3
0 .5 4

U ninfected

Birch

0 .5 7

0 .4 3

U ninfected

H em lock
D ouglas-fir

0 .5 5

0 .4 5
0 .6 7
356

U ninfected

Total

0 .3 3
414

0 .4 6

13

131
235
21
11
3
770

136

Table 2. Summary of spatial autocorrelation of species association using joins-count
statistics. Standard Normal Deviates greater than 2.0 indicate a higher than expected
number of like joins which is interpreted as significant spatial autocorrelation.
T reatm ent Plot

Fir-Spruce
Joins

S.N.D.

Fir-Fir
Joins

S.N.D.

SpruceSpruce
Joins

S.N.D.

Infected

1

59

1 .0 5 2 0

59

-1 .2 3 1 0

12

1 .7 8 6 0

Infected

2

1 .1 1 3 7

13

6

0 .7 4 5 0

60
74

-1 .3 2 2 7

Infected

61
43

-1 .0 7 0 3

13

2 .0 2 5 6
2 .9 0 0 2

Not
Infected

3

42

0 .5 9 4 9

46

-1 .0 8 7 9

21

3 .2 1 4 0

Not
Infected

4

56

0 .8 4 3 3

36

-1 .3 0 8 1

24

2 .7 7 7 5

Not
Infected

5

49

1 .0 0 0 9

78

-1 .2 2 4 8

10

2 .4 5 5 7

137

S p r u c e a Fir O O th e r
Plot 1

20
(Q
O

10
0

“T—

05
O

I—

I— I— I— r

N

CO

O

N Nl
W
O O
O

-> l

O

CO

o

o

<JT

0>
O

CO*

CO*

CO*

CO

CO*

CO*

CO*

00

CO

CO

CO

CO

cn

CO
N

CO
CO

CO*

(0 *

CO*

CO*

CO*

00
CO

CO

o

CD
CO

CO
C J1

CD

o

o

o

o

CO*

CO

CO*

CO*

CO*

CO

CO
CD

CO

o

c;i
o

00
N

00
CO

o

o

CO*

CO*

CO*

CO*

o

CO

o

o

o

o

P lo t 2

20
(Q
0)

10

0

,H

,H

o>
N

O

o

CO

-««j

'n J
CO

00

o

cn
o

<X)

CO

-»■

CO

o

cn
o

CO

CO*

CO*

CO*

CO*

CO*

CO*

CO

CO*

CO*

CO*

CO*

CO

00

00

-N
o

00

o

o

CO*

CO*

CO*

O

-4

o

o

o

00

CO

o

o

CO
CO

P lo t 6

20
(0

I 10
0
O)
O
N

CD
O

o

CO*

CO*

CO*

CO

N
O

N

■vj

CO

C fl

o

o

CO*

CO*

CO*

o

o

C
O cn
o
o

CO*

CO*

CO*

CO

CO*

CO

CO

o

cn
o

CO

00
CO

CO

CO*

"N
o

o
CO*

Figure 1. Canopy ascension by decade and by species for Inonotus tomentosus
infected plots (Plots 1,2, and 6). The ordinal axis shows the percentage of stand area
affected by disturbance in each decade.

138

■ S p ru c e

mFir □ O th e r
P lo ts

I

50
40
30

^ 20

10
0
o

00

00

s

CO

o

00

00

cn

CO

o

00

CO

CO

o

CO

g

CO

CO
CO

g

P lo t 4

20
<0

I

10

0
G)

<y>

N

CO

O
co"

's i

N

o

o

CO

CO*

■'>1

CO

o

cn
o

o

CO*

CO*

CO*

N
CO

00

o

w
o

cn
o

CO
N

o

CO
CO

CO

o

CO

CO

o

o

CO*

CO*

CO*

CO*

CO*

CO*

CO*

CO
GO

CO

cn
o

CO

o

CO
CO

CO*

CO*

CO*

CO

N
o

o

P lo ts

20
-I

<
o0

1 -1

10

0

—

r........1........r.......,......... ,.......,........ ,........1........ ,........■......................................p
(7 )
-s |

O )
CO

(A

crT

o

o

o

c/f

C
O g
o
crT
CO

's i

o
CO

. ....... M
CO

CO

o

o

CO

CO

- - I

nn

1

CO
CO

CO

o

cn
o

N
o

CO

CO

CO

CO

CO
CD
O

CO

o

o

CO
CO*

CO*

CO
CO

CO

cn
o

CO

o

"
S
j
o

CO
CO

CO*

CO*

CO*

CO*

o

Figure 2. Canopy ascension by decade and by species for Inonotus tomentosus uninfected
plots (Plots 3,4, and 5).

139

Plot 2

Plot 1

0

7

14

21

56 63

28 35 42

70 77

0

7

14 21 28 35 42 49 56 63 70

77

Plot 6
>1800
1 8 0 0 -1 8 2 0
1 8 2 0 -1 8 4 0
1 8 4 0 -1 8 6 0
1 8 6 0 -1 8 8 0
1 8 8 0 -1 9 0 0
1 9 0 0 -1 9 2 0
1 9 2 0 -1 9 4 0
1 9 4 0 -1 9 6 0
1 9 6 0 -1 9 8 0
1980+

0

7

14 21

28 35 42 49 56

63 70

77

Figure 3. Locations o f canopy ascension dates for trees in the three infected plots.
Numbers on the axis are in meters. Blank squares indicate that no canopy tree was found
within seven meters o f the sampling location on the 7 by 7 meter grid.

140

P lo ts

0

7

14 21

Plot 4

28 35 42 49 58 63 70 77

0

7

14 21

28 35 42 49 56 63 70

Plot 5

>1800
1800-1820
1820-1840
1840-1860
1860-1880
1880-1900
1900-1920
1920-1940
1940-1960
1960-1980
1980+
0

7

14 21 28 35 42 49 56 63 70 77

Figure 4. Locations o f canopy ascension dates for trees in the three uninfected plots.
Numbers on the axis are in meters. Blank squares indicate that no canopy tree was found
within seven meters o f the sampling location on the 7 by 7 meter grid.

141

77

P2

P6

0.2
0.1 5

0 .0 5

-0 .0 5
-

0.1

-0.15
-

0.2

Distance (m)

Figure 5. Correlogram showing Moran’s I statistic on the ordinate axis and distance lag
on the abscissa for plots 1,2 and 6 (plots with Inonotus tomentosus). Open symbols
represent statistically significant Moran’s I coefficients.

142

P3

P5

P4

0.2
0 .1 5

0 .0 5

-0.05
-

0.1

-0 .1 5
-

0.2

Distance (m)

Figure 6. Correlogram showing Moran’s I statistic on the ordinate axis and distance lag
on the abscissa for plots 3,4, and 5 (plots free of Inonotus tomentosus). Open symbols
represent statistically significant Moran’s 1 coefficients.

143

CHAPTER 6. SUMMARY OF FINDINGS, MANAGEMENT RECOMMENDATIONS
AND CONCLUSIONS

6.0

STUDY RATIONALE

Ecosystems cross a hierarchy of scales. A downed log on the forest floor is an
ecosystem with its own biophysical attributes. This log is part of a larger ecosystem,
perhaps an old spruce-fir stand. The spruce-fir stand is part of a watershed or a landscape
ecosystem, and the landscape part of a regional ecosystem. The large regional ecosystem
fits into an even larger biome such as the boreal forest in this example. At each
successive level in the hierarchy, more species are involved; there is more variation in
physical characteristics; and, the interactions, function, structure, and dynamics become
more complex and variable. At each stage in the hierarchy forest ecosystems have unique
structure, specialized functions, complex interactions and are continually changing.
Individual ecosystems vary in their own complexity over time and the level of
complexity between ecosystems also varies. However, the resiliency and stability of
ecosystems are greatest when their complexity is maximized (Kimmins 1997).
Human influence on ecosystems has generally caused a decrease in complexity
and therefore a reduction in their resilience and stability. Forests have been cleared
throughout the world and are replaced with agricultural ecosystems. The orehard, com or
wheat field is managed for one species and increasingly genetically identical individuals.
Wetlands have been drained to support agriculture or urban development. Natural forests
are being cleared to make way for managed forests, converting a landscape with
astonishing biological diversity to a landscape where only a few species are managed to
optimize timber produetion on short economic rotations.

144

Knowledge about ecosystem structure, function, complexity, interactions and
change is crucial in order to understand the impact of human influence on the
environment. This knowledge may be critical for the survival of our species because
accumulated effects of ecosystem degradation may affect the entire biosphere, the
atmosphere and the hydrosphere. Together these define the biophysical characteristics of
the ultimate ecosystem. Earth. Evidence of these global consequences of ecosystem
mismanagement are already evident: global warming, desertification, extinctions, soil
erosion, and stream siltation.
As natural ecosystems are lost and the integrity of remaining ones compromised,
the concept of ecosystem management has crept into the dialogue of land and resource
management. Ecosystem management adopts principles that are entrenched in ‘landscape
ecology’ and ‘natural disturbance ecology’ (Parminter and Daigle 1997). One of the
concepts of landscape ecology in the forestry context is that if forest management mimics
natural disturbances, inherent ecosystem attributes will be preserved and important
ecological processes will be maintained (Parminter and Daigle 1997). It follows from this
principle that if natural disturbances are large and catastrophic then large clearcuts of the
same size, pattern and concentration would be the appropriate harvesting method to
mimic this disturbance. Similarly if small-scale disturbances are predominant then, partial
cut harvesting methods would be most appropriate. Most often landscapes are not
dominated by one disturbance pattern or type. Landscapes are shaped due to a variety of
disturbances operating at a variety of scales and intensities (White 1987; Engelmark et al.
1993). Forest management should therefore employ variation when planning land-use
activities to reflect the variation in pattern caused by natural processes.

145

In boreal forests considerable effort has been directed towards studying landscape
level disturbance patterns and stand dynamics. Much of the work has concentrated on
characterizing the fire return interval, patch size of fire disturbances and the resultant
landscape level age-class mosaic (Bergeron et al. 1993). For example, much of the subboreal, montane forests in central interior of British Columbia are composed of fireorigin forests whose climax species composition on upland sites are dominated by spruce
(Picea glauca, Picea engelmannii or Picea glauca x engelmannii), Abies lasiocarpa,
Pinus conforta, Populus tremuloides, Pseudotsuga menziesii, and Betula papyrifera.
Although dryer parts of the region were burned by wildfire at intervals ranging from 150250 years (British Columbia Ministry of Forests and B.C. Ministry of Environment,
Lands, and Parks 1995), wetter parts, particularly those in the foot-hills and sub-alpine
elevations historically had fire return intervals ranging from 227-6250 years (Hawkes et.
al. 1997). Within these wet, mountainous areas, information is lacking about the spatiotemporal pattern of small-scale disturbances and on subsequent stand dynamics
(succession).
Small-scale disturbances have been shown to be important processes in many
forested ecosystems including boreal forest ecosystems. However, studying the patterns
and processes of small-scale disturbances have typically been neglected in boreal forests
simply because fire was conventionally thought to be the main disturbance agent in this
forest type (Bergeron et. al. 1993). In wet boreal and sub-boreal forests, fire return
intervals have been shown to be quite long, often approaching intervals that are
characteristic of more humid environments such as west-coast temperate rainforests, and
sub-alpine forests. Given the long fire return intervals that are possible in these wet, cool

146

boreal forests, small-scale disturbances have been identified as a factor in modifying the
forest community (population structure and demographics) and the physical environment
(light regimes, temperature, moisture regimes and nutrient availability and dynamics)
(Bergeron et al. 1993). The extent of their influence is related to the length of time
between catastrophic disturbances.
Windthrow, insect attacks, and root diseases, are important types of small-scale
disturbance agents in these forests (Hofgaard 1993; Kneeshaw and Burton; 1997;
Kneeshaw and Bergeron 1998). One of the most important small-scale disturbances in
boreal forests is Inonotus tomentosus. This root disease pathogen is ubiquitous in Picea
spp. forests and is responsible for significant mortality especially in mature forest
ecosystems dominated by Picea spp. (Lewis 1997). Therefore understanding the role of I.
tomentosus in stand disturbances is necessary in order to understand its impact on Picea
spp. forest communities which make up a significant component of boreal and sub-boreal
forest landscapes. Secondly, the patterns of disturbance in general are useful in designing
forest harvesting activities that mimic natural disturbance patterns. And, lastly in order to
confidently prescribe partial cut harvesting systems, potentially adverse or undesirable
impacts on stand composition and structure that occur because of potential interactions
between partial cutting and I. tomentosus need to be understood or ruled out.
In order to contribute to the knowledge and understanding of disturbance ecology
and stand dynamics of wet, natural and unmanaged sub-boreal spruce forests in central
British Columbia, this thesis raised three basic questions:
1. What are the spatial and temporal patterns of disturbance for old-growth forests with
and without the influence of /. tomentosus!

147

2. How does stand composition and structure differ between I. tomentosus infected and
non-infected old-growth forests?
3. How does partial cutting influence stand development in I. tomentosus infected and
non-infected forests?

The remainder of this chapter summarizes the findings of the thesis research, discusses
potential applications of the methodology developed for this research and discusses
future research needs.

148

6.1 SUMMARY OF RESEARCH
6.11 Canopy Disturbance
1) Disturbance History: Mortality of canopy trees was indicated by tree ring
information sporadically as early as the 1670’s and was consistently evident since the late
1700’s. The oldest individuals in the stands were well over 300 years old and did not
appear to have fire scars. Therefore, these indirect results show it has been at least 300
years since the last stand-replacing disturbance at Aleza Lake. If the last large scale fire
occurred roughly 300 years ago, then mortality evident in the late 1700’s began
approximately 100 after the fire. This would correspond to the timing of the understory
re-initiation stage where mortality of canopy trees creates canopy gaps that allow
understory species to regenerate and recruit to canopy positions.
2) Canopy Disturbance Rates: Two independent studies (Chapter 4 and Chapter
5) indicated similar levels o f canopy mortality in natural forests (un-harvested forests) as
expressed by the percentage of canopy disturbance per decade. In a fine scale study
(fixed radius plots. Chapter 4), average percent canopy disturbance was between 6.9%
and 8.1% per decade. In a coarser-scale study (grid plots. Chapter 5), canopy disturbance
averaged 5.09% and 6.0% per decade. Therefore based on widespread sampling in four
different stands, using two independent approaches, canopy disturbance at Aleza Lake
ranges on average from 5% to 8% per decade.

149

6.12 Gap Size
Gap size was also determined using two independent and distinct methodologies
developed in Chapter 4 and Chapter 5. In Chapter 4, the estimate of average singe-tree
gap-size was 16.76m^. In Chapter 5, it was determined that patch diameter averaged <7m
which corresponds to an gap area of about 38.5m^. Note that the 38.5m^ gap size estimate
was limited to the lag distance specified in the spatial autocorrelation analysis (7m).
Based on these two approaches, it is evident that average gap size is typically <38.5m^
indicating that the vast majority of gaps are due to single tree mortality. Larger gaps do
form in these forests but with lower frequency than single tree gaps.

6.13 Canopy Mortality by Species
In Chapter 3 it was determined that spruce accounted for 42% of total
accumulated mortality and (because of its generally larger size) over 50% of accumulated
basal area mortality. This high percentage of total mortality in old forests is likely due to
the fact that a uniform cohort predominately composed of spruce, has been suffering high
rates of mortality due to a variety of mechanisms.

6.14 Stand Composition, Canopy Replacement and Future Stand Composition
Results from Chapter 4 and Chapter 5 indicate that canopy composition is
dominated by fir at about 65-70% of canopy by stems per hectare for all size classes
combined. Spruce represents about 25-30% of stand density and about 47-57% of basal
area. Understory composition was also dominated by fir at 86%-92% of stand density.

150

Due to the high density of fir in the understory, canopy tree replacement by this
species outnumbers all other understory species by at least 2:1. Due to high rates of
spruce mortality over the last 50 years or more, and high proportion of fir replacing dead
canopy trees of any species, the spruce component of the canopy has been on the decline
for the past several decades. Fir composition has been increasing over the same period.
Given the current combination of high canopy recruitment rates of fir, high spruce
mortality and the overwhelming abundance of fir in the understory layer, fir populations
will continue to increase over the next 200 years.

6.15 Impacts of Inonotus tomentosus and Partial Cutting on Disturbance and Stand
Composition
In old-growth forests with low to moderate infection incidence (<I5%), specific
differences in disturbance regime caused by Inonotus tomentosus are not apparent.
Disturbance history, disturbance rates, gap size, spatial patterns of disturbance and
accumulated spruce mortality are not significantly different in old-growth stands with I.
tomentosus compared to old-growth stands without the root disease.
Partial cutting has increased the density o f spruce in the understory, and decreased
subsequent mortality rates in both infected and uninfected stands relative to old-growth
forests. Since most of the understory spruce has regenerated post-harvest they have not
yet moved into the canopy layer. Given time, canopy layer spruce populations should
also rise. Reductions in mortality are most likely due to captured mortality of older and
larger individuals and upgrading the vigor of remaining spruce.

151

Although spruce populations were not significantly lower in I. tomentosus
infected stands compared to uninfected old-growth and partially cut stands, spruce
volumes are significantly lower (6-10%).

152

6.2 MANAGEMENT APPLICATIONS
6.21 Prescribed Harvest Intensity for Partial Cut Harvesting in Wet-Cool SBS Forests
Disturbance chronologies presented in this thesis indicated that over a period of
70 years (Chapter 4) and 200+ years (Chapter 5), average canopy disturbance averaged
between 7.4 and 8.1% per decade. From the accumulated mortality data (Chapter 4), this
level of canopy disturbance corresponds to approximately 44-48 mortalities/hectare/
decade. Diameter class distributions of dead trees (Figure 1) indicates that the average
diameter of trees dying is 29.9cm dbh. Species composition of dead trees is 41-42%
spruce (stems/ha) and 51-56% (basal area). All of this evidence suggests that partial-cut
harvesting if conducted every 10 years in these old spruce-fir forests should remove 4448 stems/ha/decade with an average diameter of 29.9cm and approximately 51-56% of
the basal area should be spruce.

6.22 Harvest Dispersion
Harvests should create a variety of gap sizes but the average gap diameter should
be 7m or less. This suggests that single tree selection silviculture systems combined with
small group tree selection systems would be the most appropriate combination of
silviculture systems for these forests to maintain natural forest structure. The gaps
themselves should be irregularly distributed across a landscape or stand.

6.23 Harvest Timing
Due to the continuous low intensity disturbance that prevails in these forests, it
would not be advisable to “bank” harvests for 3 or more decades and then triple or more

153

harvest intensity, especially if the objective is to maintain “old-forest” values {i.e. more
intense harvests may increase spruce regeneration - as it did in the partial cut forests which is contrary to natural stand-dynamics where fir recruits to the canopy). However if
the objectives are to maintain spruce as a component in the stands for timber supply,
more intense harvests on less frequent intervals would provide conditions necessary for
spruce recruitment and survival.

6.24 Development of Methodology
This research has advanced two important retrospective methodologies that have
made possible the quantification of disturbance regimes for this area: The Time Since
Death model (TSD model) (Chapter 2), and the Growth Rate/Percent Release criteria
(Chapter 3). Both methodologies could be used in similar studies so long as regional
climate and latitude are similar. The methods should be tested and calibrated for local
conditions before studies begin.
The TSD model probably has the most potential for widespread immediate use
with only minor calibrations necessary so long as the species are limited to spruce and fir.
This model allows for very quick and accurate estimates of year of mortality which
allows for large areas to be sampled and many tree deaths to be calculated. Climate has
been shown to play a major role in tree decomposition, therefore calibration of this model
should be performed before studies begin.
The simplest and most efficient means of calibrating the TSD model would be to
compare death estimates from the TSD model to year of death estimates calculated from
increment cores in understory trees. An error factor built into the time since death

154

prediction from the TSD model developed in the present study could be either added (for
slower rates of decomposition) or subtracted (for faster rates of decomposition) by
comparing the two dates. For example, if the TSD model consistently indicates mortality
dates 10 years earlier than tree ring core estimates then an error factor of + 10 years
would need to be added to the TSD model estimate.
The TSD model allows research of the mortality dynamics for forests of any age
as long as tree diameters exceed 10cm dbh. Furthermore, because it is efficient and
accurate the disturbance regime over a very large area or in many areas can be
determined. The main disadvantage of the TSD model is that it is only accurate in
predicting time since death up to about 70 years. Therefore when a longer term in
disturbance regime study is needed, tree growth pattern analysis is the still the best
method.
Until now, growth/release criteria of understory trees that enable the interpretation
of the disturbance regime has been lacking in the sub-boreal and boreal forests of British
Columbia. This thesis developed release and growth rate criteria for understory spruce,
fir, and B. papyrifera responding to the death of an overstory tree. Although this
information may not be precisely applied to other ecosystems in the boreal and subboreal forests, it should provide a guide for ecosystems with similar physical (latitude,
soils, temperature, and moisture) and biological characteristics (species composition,
crown shape/form, and tree height).

155

6.25 Future Research
Research concerned with the role of small-scale disturbances in boreal and subboreal forests has only just begun. Furthermore, disturbances operate at many different
scales, they interact with physical and biological ecosystem components and with other
disturbance agents, and disturbances are stochastic. All this variation combined with the
developing understanding means that research into small scale disturbance regimes must
continue over large areas, and in many different stands before adequate information is
gained. Thus, an efficient yet comprehensive approach needs to be adopted in order to
characterize the natural disturbance regime before too much of the landscape is managed
using harvest patterns that are not compatible with the natural disturbance regime.
A general approach is suggested here to expedite the investigation:
1) Stratify the land base: The boreal and sub-boreal forests in British Columbia should
be stratified into units that tend to have the longest stand replacing disturbance
interval. This work has already been done by the biogeoclimatic ecosystem
classification (Meidinger and Pojar 1991) and the natural disturbance type
classification system (British Columbia Ministry of Forests and British Columbia
Ministry of Environment, Lands and Parks 1995).
2) Prioritize the land base: Determine where partial cut harvesting systems should be
implemented based on ecological, or social values. Much of this work has also been
done in the Land and Resource Management Planning Process or the Timber Supply
Review.
3) Develop methodology and sample the land base: Using the TSD model developed in
this thesis, calibrating the model for climactic conditions, and linking time since death

156

estimates to the wildlife tree class system (British Columbia Ministry of Forests and
British Columbia Ministry of Environment, Lands, and Parks 1995) would provide
for a easily developed, widely applicable and easily used tool for determining time
since death. Sampling could be done in conjunction with operational timber cruising.
This could make the sampling quite inexpensive and cover many stands.
4) Compile information and develop harvesting patterns: The data from the sampling
should then complied and compared. Broadly similar disturbance regimes should be
identified at this point and incorporated into specific management guidelines.
5) Apply and monitor: once the guidelines are in place and operational harvests are
using the guidelines, a monitoring program should be followed. Monitoring should
include an assessment of whether the system is providing similar stand structure, and
composition and dynamics compared to natural stands.

157

6.3

CONCLUSIONS

This thesis has indicated that small-scale disturbances are important ecological
processes in sub-boreal forest ecosystems. The information presented in this thesis can
now be added to a small but growing body of literature concerned with investigating the
patterns and processes of small-scale disturbance across the circumpolar boreal forest. In
such a large biome, and in a relatively new discipline much work still needs to be done. It
is hoped that this thesis will play a small but significant role in supporting related
research and in advancing management applications related to disturbance ecology.

158

6.4 LITERATURE CITED
Bergeron, Y., Engelmark, O., Harvey, B., Hubert, B. and Sirois, L. 1993. Key issues in
disturbance dynamics in boreal forests: Introduction. Journal of Vegetation
Science 4:464-467.
British Columbia Ministry of Forests and British Columbia Ministry of Environment,
Lands and Parks. 1995. Biodiversity Guidebook. British Columbia Forest
Practices Code. Victoria, B.C. 99pp.
Engelmark, O., Bradshaw, R. and Bergeron, Y. 1993. Disturbance dynamics in boreal
forest: Introduction. Journal of Vegetation Science 4:730-732.
Hawkes, B., Vasbinder, W. and DeLong, C.G. 1997. Retrospective fire study: fire
regimes in the SBSvk & ESSFwk2/wc3 biogeoclimatic units of northeastern
British Columbia. McGregor Model Forest Association. Prince George, BC. 34
pp.
Hofgaard, A. 1993. Structure and regeneration patterns in a Picea abies forest in
northern Sweden. Journal of Vegetation Science. 4:601-608.
Kimmins, J.P. 1997. Forest Ecology: A Foundation for Sustainable Management.
Prentice Hall, New Jersey.
Kneeshaw, D.D., and Burton, P.J. 1997. Canopy and age structures of some old
sub-boreal Picea stands in British Columbia. Journal of Vegetation Science
8:615-626.
Kneeshaw, D.D., and Bergeron, Y. 1998. Canopy gap characteristics and tree
replacement in the southeastern boreal forest. Ecology 4(3):783-794.
Lewis, K.J. 1997. Growth reduction in spruce infected hy Inonotus tomentosus in central
British Columbia. Can. J. For. Res. 27: 1669-1674.
Meidinger, D. and Pojar, J. 1991. Ecosystems of British Columbia. Research Branch,
Ministry of Forests. Crown Publications Inc. 330 pp.
Parminter, J. and Daigle, P. 1997. Landscape Ecology and Natural Disturbances:
Relationships to Biodiversity. British Columbia Ministry of Forests Research
Program. Victoria, British Columbia.
White, P.S. 1987. Natural disturbance, patch dynamics, and landscape pattern in natural
areas. Natural Areas Journal 7(1): 14-22.

159

■ Hemlock
□ Birch
■ Unknown
□ Spruce
■ Fir

D ia m e te r C la s s

Figure 1. Diameter class distribution (dbh) of dead trees in old-growth forests ( n = 509).
Mean dbh = 29.9cm.

160