NORTHERN RANGE LIMIT MOUNTAIN PINE BEETLE (DENDROCTONUS
PONDEROSAE) OUTBREAK DYNAMICS AND CLIMATE INTERACTIONS IN
MIXED SUB-BOREAL PINE FORESTS OF BRITISH COLUMBIA

by

Kathryn H. Hrinkevich
B.Sc., Radford University, 1998
M.Sc., Portland State University, 2005

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
September 2012

© Kathryn H. Hrinkevich, 2012

1+1

Library and Archives
Canada

Bibliotheque et
Archives Canada

Published Heritage
Branch

Direction du
Patrimoine de I'edition

395 Wellington Street
Ottawa ON K1A0N4
Canada

395, rue Wellington
Ottawa ON K1A 0N4
Canada
Your file Votre reference
ISBN:

978-0-494-94166-9

Our file Notre reference
ISBN:

978-0-494-94166-9

NOTICE:

AVIS:

The author has granted a non­
exclusive license allowing Library and
Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distrbute and sell theses
worldwide, for commercial or non­
commercial purposes, in microform,
paper, electronic and/or any other
formats.

L'auteur a accorde une licence non exclusive
permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans le
monde, a des fins commerciales ou autres, sur
support microforme, papier, electronique et/ou
autres formats.

The author retains copyright
ownership and moral rights in this
thesis. Neither the thesis nor
substantial extracts from it may be
printed or otherwise reproduced
without the author's permission.

L'auteur conserve la propriete du droit d'auteur
et des droits moraux qui protege cette these. Ni
la these ni des extraits substantiels de celle-ci
ne doivent etre imprimes ou autrement
reproduits sans son autorisation.

In compliance with the Canadian
Privacy Act some supporting forms
may have been removed from this
thesis.

Conform em ent a la loi canadienne sur la
protection de la vie privee, quelques
formulaires secondaires ont ete enleves de
cette these.

W hile these forms may be included
in the document page count, their
removal does not represent any loss
of content from the thesis.

Bien que ces formulaires aient inclus dans
la pagination, il n'y aura aucun contenu
manquant.

Canada

Abstract

Insect outbreaks are some of the most destructive processes o f forest change with
long-term economic and ecological effects that can be severe. For native insects that have
co-evolved with their host forests it is important to develop and implement management
strategies that are consistent with the basic ecology o f the ecosystems being affected. This
requires a sound ecological understanding o f the disturbance regime and its function, which
can only be gained through long-term stand histories appropriate to the scale of the process.
This is becoming increasingly important in light of changing climate conditions which have
already influenced disturbance regimes and the spatial distribution of insect ranges with
unknown consequences.
The objectives of this study were to develop a long-term record of mountain pine
beetle (MPB) outbreaks near the northern limit of the species’ range and to examine the
interacting factors driving that disturbance regime. Using a combination of
dendrochronological methods to reconstruct stand and outbreak history together with long­
term local and large-scale climate data, I developed an integrated model of outbreak
dynamics for north-central BC where little is known about the role o f MPB in forest stand
dynamics or the influence of climate on outbreak development. Using a 200-year outbreak
reconstruction I determined that host susceptibility and mortality patterns had distinctly
different characteristic than those described further south. Climate was more important than
forest structure in determining outbreak frequency and severity, and persistent warm
temperature anomalies, including the large-scale climate patterns driving these trends, were
the most important direct climate drivers o f outbreak development. Over short time scales,

host stress preceded outbreaks, but long-term periods of vigorous tree growth were also
important for outbreaks to develop. Differences between this study and others suggest that
outbreak risk management and predictive models must consider regional differences in forest
structure, host-climate relationships and climatic pressure on beetle populations. Given that
regional host susceptibility criteria are met, low-frequency climate patterns such as PDO and
ENSO are probably the most important outbreak drivers across the province. Recent climate
changes have already altered the historical regime, and ongoing warming will likely change
many of the interactions described here.

PERMISSION TO REPRODUCE

Chapter 2 is a revised version of:
Hrinkevich, K., and K. J. Lewis. 2012. Northern range limit mountain pine beetle outbreak
dynamics in mixed sub-boreal pine forests of British Columbia. Ecosphere:
2(10): 116.
and is reproduced with the permission of the Ecological Society o f America (permission
granted under the terms o f the Creative Commons Attribution License, 27 October, 2011)

IV

TABLE OF CONTENTS

A bstract...........................................................................................................................................ii
PERMISSION TO REPRODUCE...............................................................................................iv
TABLE OF CONTENTS.............................................................................................................. v
List o f Figures.............................................................................................................................viii
Acknowledgments.........................................................................................................................xi
1. Introduction.................................................................................................................................1
1.1 Context...................................................................................................................................1
1.2 Study objectives and rationale.............................................................................................4
1.3 Background informationon the mountain pine beetle.......................................................6
1.3.1 Life cycle and population processes............................................................................6
1.3.2 Epidemiology.................................................................................................................8
1.4 Conceptual model for describing outbreak dynamics.......................................................9
1.4.1 MPB-host interactions................................................................................................ 12
1.4.2 Climate-MPB interactions........................................................................................... 13
1.4.3 Climate-host interactions............................................................................................14
1.5 Organization of this paper................................................................................................. 15
2. Historical outbreak dynamics and forest structure at the northern range limit for mountain
pine beetle in British Columbia................................................................................................... 17
2.1 Abstract................................................................................................................................17
2.2 Introduction........................................................................................................................ 18
2.3 M ethods.............................................................................................................................. 22
2.3.1 Study A re a ....................................................................................................................22
2.3.2 Site Selection...............................................................................................................24
2.3.3 Data Collection and Analysis..................................................................................... 26
2.4 R esults................................................................................................................................ 30
2.5 Discussion...........................................................................................................................43
2.5.1 Stand history............................................................................................................... 43
2.5.2 Host selection, stand susceptibility and outbreak severity....................................... 44
2.5.3 Outbreak regime.......................................................................................................... 47
v

2.5 Conclusions......................................................................................................................48
3. Local and large-scale climate interactions with a multi-century record of mountain pine
beetle outbreaks in north-central British Columbia...................................................................50
3.1 Abstract................................................................................................................................50
3.2 Introduction........................................................................................................................ 51
3.3 M ethods.............................................................................................................................. 58
3.3.1 Study area..................................................................................................................... 58
3.3.2 Chronology development............................................................................................61
3.3.3 Climate data.................................................................................................................64
3.3.4 Climate, tree growth and outbreak relationships......................................................67
3.4 R esults................................................................................................................................ 69
3.4.1 Climate-MPB outbreak-host tree growth relationships........................................... 69
3.4.2 Climate suitability....................................................................................................... 75
3.5 Discussion........................................................................................................................... 78
3.5.1 Climate- MPB outbreak-host tree growth relationships.......................................... 78
3.5.2 Climate suitability....................................................................................................... 83
3.6. Conclusions....................................................................................................................... 85
4. Mountain pine beetle outbreak dynamics and climate interactions in sub-boreal mixed
pine forests o f north-central British Columbia: a synthesis...................................................... 87
4.1 Introduction........................................................................................................................ 87
4.2 MPB, host and climate interactions inferred from a multi-century outbreak
reconstruction........................................................................................................................... 89
4.3 Limitations and future directions...................................................................................... 95
Literature C ited............................................................................................................................ 99

List of Tables

Table 2.1. Site inventories showing the distribution and dating of live (L) and dead (D) (pine
only) trees > 10 cm diameter at breast height in each plot and the total number of dated
samples used to build the final site chronologies (C), including any additional trees sampled
outside plot boundaries. The only other species present was trembling aspen, which were
inventoried at two sites for a total o f 4 trees, none of which had cores intact enough for
dendrochronological processing................................................................................................. 33
Table 2.2. Stem density summaries by canopy class and number o f outbreaks that occurred.
Results o f the ANOVA and Kruskal-Wallis means comparisons results are listed for each
test of stem density differences between the groups defined by number of outbreaks
37
Table 3.1. Descriptive statistics computed using the program COFECHA for individual
climate sites used to compile the final chronology and for the master chronology................ 63
Table 3.2. Climate variables known to influence mountain pine beetle population dynamics
with commonly used threshold values for climate suitability modeling. Data sources for each
variable are included. The rationale for including these variables in the analysis and the
specific effects they represent in the population dynamics of MPB is described in detail in
the text...........................................................................................................................................66

List of Figures

Figure 1.1. Conceptual model of outbreak dynamics adapted from Stephens (1960) showing
the three necessary causal factors o f an insect outbreak and the interactions between factors
(b, c, d) that over longer time scales determine the outbreak regime (a) for a given area......11
Figure 2.1. Detailed site and study area locations within British Columbia relative to the
2009 cumulative mountain pine beetle outbreak extent and lodgepole pine distribution.......25
Figure 2.2. Composition and structure summarized for all trees > 10 cm DBH. Live (a) and
dead (c) diameter distributions show relative composition at 5cm size class intervals. Canopy
distribution for living trees only (b) shows compositional differences between the dominant
and codominant (main), intermediate (mid) and suppressed (sub) canopy classes. Mortality
age distribution for dead pine only (d) shows age at the time of death (at sample height) for
all pine with intact pith and outer ring dates in 10 year age classes........................................ 34
Figure 2.3. Establishment dates for all living trees greater than 10 centimeters diameter and
all dead pine from all sites combined (bottom graph). Open circles approximate the timing of
fires in my study area and are based on stand-initiation dates. Arrows indicate the timing of
mountain pine beetle outbreaks reconstructed by this study. The proportion of sites recording
each event is shown in the top graph...........................................................................................35
Figure 2.4. Stem densities at each site for all canopy levels ordered left to right by number
of outbreaks experienced in that site. Number of outbreaks was determined using the pine
mortality and survivor release data to determine whether mortality events occurred at that
site during the reconstructed outbreak periods...........................................................................36
Figure 2.5. Standardized chronologies for lodgepole pine, interior spruce and subalpine fir
from two sites exhibiting (a) synchronous pine mortality (site TS11) and (b) asynchronous
pine mortality (site NZ17). Reference lines for each chronology represent mean index
values, with sample depths drawn below each species. Dashed vertical lines indicate periods
o f two-year cycle budworm outbreaks identified by Zhang and Alfaro (2002, 2003)............39
Figure 2.6. JOLTS results from selected sites representing synchronous and asynchronous
mortality patterns. Black bars show pine mortality, dark (light) grey bars show the
percentage of non-host (pine) trees experiencing a doubling o f growth or greater for that
year. Reference lines are drawn at 40% to delineate major release events. Dashed vertical
lines indicate periods o f two-year cycle budworm outbreaks identified by Zhang and Alfaro
(2002, 2003)................................................................................................................................. 40

Figure 2.7. Landscape scale outbreak reconstruction based on composite data for all sites.
Proxy reconstruction (a) shows the timing of pine (light grey bars) and non-host (dark grey
bars) growth releases relative to (b) pine mortality events. Sample depth figure (c) shows the
number of surviving pine (hatched light grey fill) and non-host (solid
grey fill)treesused to
calculate release percentages, the number of sites at which treeswere releasing(black line),
and the number of sites recording pine mortality (black bars) annually. Ten year relative
growth increments standardized by tree size (d) represent tree vigor at the time of death for
all confirmed beetle-killed pine trees. Reference line represents the mean ten year
standardized growth increment for all living pine trees, above which I consider pine
mortality to be occurring in healthy trees................................................................................ 42
Figure 3.1. Study area map showing the locations o f each of the climate sites used to
compile the master climate chronology as well as the locations of the study sites used to
develop the mountain pine beetle outbreak reconstruction. Inset shows location of study area
in British Columbia relative to the distribution o f lodgepole pine and the 2009 cumulative
mountain pine beetle outbreak extent.......................................................................................60
Figure 3.2. Superposed epoch analysis results for variables showing significant relationships
to outbreak events. Variable names are listed across the top. The top row shows annual
relationships between the variable and outbreak events. The bottom row shows
representative examples of the lower frequency trends that were significantly related to
outbreak timing, with the trend length indicated for each variable. Event window years
represent the 8 years analyzed around the outbreak events, with year zero (indicated by an
arrow at the top of each graph) corresponding to the year the outbreak occurred and negative
(positive) numbers corresponding to years preceding (following) the event. Hashed lines
represent 95% and 99% confidence intervals and shaded bars indicate significant
relationships at that window position..........................................................................................70
Figure 3.3. Annual and 5-year running means of the tree ring chronology used in the SEA
analysis overlaid on the reconstructed outbreak periods. Distinct periods o f punctuated host
stress (below-average growth) occur at the start of each outbreak, while 5-year trends of
persistent above-average growth precede the start of each outbreak....................................... 73
Figure 3.4. Climate-tree growth correlations between mean temperatures and total
precipitation and my master residual climate chronology over a 17-month window beginning
in May of the preceding year and ending in the December of the current year for the period
1912-2004. Asterisks indicate significant correlations for that month that were assessed
using 95% confidence intervals (black lines)............................................................................. 74

Figure 3.5. Climate suitability parameters for the study area based on established threshold
values for each parameter listed in Table 3.2. White boxes indicate favourable climate years
for mountain pine beetle and black boxes indicate unfavourable climate years. Cold
mortality days (bottom graph) show the number o f days below the 50% mortality threshold
(dark grey bars) and 100% mortality threshold (black bars). Light grey vertical shading
across all plots shows the timing of the reconstructed outbreak periods................................. 77

x

Acknowledgments
My dissertation research would not have been possible without the continuous and
generous support of my supervisor, Dr. Kathy J. Lewis, who taught me much more than I
ever expected to learn in an academic endeavor. From the moment I arrived at UNBC, Kathy
provided me with an entirely new perspective on academia, encouraged and allowed me to
fumble my way through some innovative and non-traditional methods of study that from the
onset neither of us was sure were possible or even rational (especially those models I
developed in the hospital tethered to an IV), and showed me that balancing life with an
academic career was not only possible, but could be rewarding as well. I can’t thank her
enough for the confidence she entrusted in me to take this project in the direction o f my
choosing, and for the patience and countless opportunities outside the study itself she
provided or condoned that at the end of the day were just as important in completing this
process and for my development into the self-assured, well-rounded person I leave this
program as. I am honoured to have had the chance to work with Kathy and look forward to
our continued collaboration in the future, particularly when it involves fishing.
I thank the other members of my graduate committee, Drs. Scott Green, Lori Daniels,
Brad Hawkes, and Rene Alfaro for their contributions and insightful comments on earlier
drafts, as well as for the extra time many have spent outside of official committee business
teaching me about dendrochrolongy and throwing ideas around about disturbance ecology,
entomology, climate change, and just in general about where life can lead from here. I am
also grateful to Dr. Rosemary Sherriff for taking the time to serve as my external examiner; I
can’t imagine a better fit for this project look forward to future discussions about the
overlapping components of our work.
I have been fortunate to have the support o f many field and lab assistants, among
them Rurik Muenter was an amazing worker, theorist and friend through the most trying
field season of the project, without whom I don’t know that 1 would have tracked down as
many of the sites as I did, sampled so efficiently and extensively, and transported as much
‘data’ (i.e. heavy pieces of wood) from the middle o f nowhere back to the lab. I also thank
Doug Thompson for his willingness and knowledge to introduce me to the northern-interior
forests o f British Columbia and to pinch hit whenever I found myself in need o f a field
assistant to drop everything for a week to come camp in the rain with me. Doug was also a
tremendous help in the lab and during the analysis and writing phases, always willing to talk
tree rings with me. I thank all the other members of the tree ring lab who assisted, talked
shop, and simply put up with the sounds of sanding and the beeping o f the Velmex for so
long: Chris Konchalski, Kennedy Boatang and Sean Sweeny, you were a great crew to
conspire with in a lab that required detail-oriented people with excellent taste in music.
Thank you to the Takla First Nation for sharing the beauty of their territory and
allowing me to do much o f my sampling within their lands, and to Terry Teegee for helping
to facilitate access. I am also tremendously indebted to many people outside o f this project

who helped me to maintain my sanity, stay focused and calm, and remember that life is
meant to be celebrated with the people who inspire you, put up with you on your bad days,
make you smile and encourage you to fulfil your dreams. Thank you Sharon Hunter, Sonja
Ostertag, Jose Delgado-Guevara, Heidi Knudsvig, Stacey McCarthy, the Chilcotes, and all
you dendro fanatics out there I’ve known for years thanks to NADEF and AAG. A special
thank you to Margaux Shilling; I could have never finished this without your love and
support and look forward to repaying the favor with a lifetime o f the same.
Funding for this research was provided by the Government of Canada through the
Mountain Pine Beetle Program, a three-year $100 million program administered by Natural
Resources Canada, Canadian Forest Service. Additional support was provided through a
fellowship from the Pacific Institute for Climate solutions and NSERC.

1. Introduction
1.1 Context
Forest disturbances are important processes that shape the composition, structure and
functional components of the ecosystem through the damage or removal of individual trees
in a stand and the alteration of the growing space for the surviving and newly established
occupants (Oliver and Larson 1996). Disturbances can be endogenous, arising from within
the stand such as insect and pathogen outbreaks, or exogenous, originating from outside the
stand as in cases o f flooding and fire. They can be natural or human created, biotic or abiotic
and can range from stand-replacing events to minor interruptions temporarily suppressing
the growth o f select individuals. Often, it is the interaction o f multiple disturbance agents
through time that shapes the past, present and future successional trajectory o f a forest
system (Rogers 1996).
The social, economic and ecological effects of forest disturbances can be large and
long term. In North America, insect and pathogen outbreaks are the most destructive
processes of forest change, affecting nearly 20.4 million hectares per year with an economic
impact o f over 2 billion dollars per year in the United States alone, almost 6 times the cost
associated with forest fires (USDA 1997, Dale et al. 2001). In addition to timber losses
(Parkings and McKendrick 2007), outbreaks influence wildlife habitat (Chan-McLeod
2006), fire behavior (Jenkins et al. 2008), carbon storage (Kurz et al. 2008) forest structure
and stand dynamics (Shore et al. 2006), property and recreation values (McGregor and Cole
1985) and watershed quality through hydrologic impacts (Redding et al. 2008). Despite these
effects, where native insects are concerned that have co-evolved with the host forest,
outbreaks are considered normative events (Mattson 1996) because o f the important role
1

they play in the natural cycle o f forest development and regeneration (Logan and Powell
2005). For forest managers, this means it is essential that mitigation strategies are consistent
with the basic ecology of the ecosystems (Romme et al. 2006) to ensure that management
objectives do not negatively impact other natural components of the system. To do so
requires a sound ecological understanding of forest disturbance processes and their function.
A major limitation to understanding forest disturbance processes is that the temporal
scale of forest change often exceeds the scope of recorded history. Because trees can survive
for many centuries and the disturbance interactions that shape stand development and
susceptibility to future disturbance begin at the time of stand-initiation (Veblen et al. 1994),
long-term histories are necessary for developing effective sustainability goals under an
ecosystem management paradigm (Christensen et al. 1996). One of the most useful methods
for reconstructing past disturbances is dendrochronology which uses the natural record of
environmental variability stored in the annual rings of trees to develop stand histories (Speer
2010). The great value of dendrochronology lies in the practice o f crossdating (Douglas
1941) which allows the precise dating of tree ring-derived time series with an annual
resolution that can extend as far back in time as there are temporally-overlapping wood
samples available, live or dead. Using this approach, dendrochronologists have reconstructed
multi-century records o f insect outbreaks (Hadley and Veblen 1993, Zhang and Alfaro 2002,
Berg et al. 2006, Hrinkevich and Lewis 2012), fire (Baisan and Swetnam 1990, Brown and
Swetnam 1994, Stephens et al. 2003), foliar pathogens (Welsh et al. 2009) and climate
(Jacoby and D’Arrigo 1989, Diaz et al. 2001, Davi et al. 2003, Trouet and Taylor 2009),
establishing links between these components through patterns o f long-term interaction
(Veblen et al. 1994, Zhang et al. 1999, Heyerdal et al. 2002, Grissino-Mayer et al. 2004,

2

Axelson et al. 2009, Sheriff et al. 2011). The specific relationships between climate, insect
and other forest disturbances often vary depending on species life history strategies,
physiological tolerances and host ecology (Dale et al. 2001, Bentz et al. 2010), so
management strategies that incorporate as much understanding of the effects that forest stand
dynamics and climatic variation have on outbreak dynamics greatly improve our ability to
mitigate risk (Taylor et al. 2006).
The importance o f climate in the population dynamics of forest insects is well
established, however, rising concerns over the impacts of climate change on forest
disturbance processes have increased attention on defining specific climate-outbreak
relationships (Bentz et al. 2010). These concerns arise from the fact that climate can
influence the timing, frequency, intensity, extent and duration of disturbances (Dale et al.
2001), and disturbances can greatly accelerate or amplify the effects of changing temperature
and precipitation on forest composition and structure (Overpeck et al. 1990, Swetnam 1993,
Swetnam and Lynch 1993) when compared with the direct effects of a changing climate on
tree species distributions that are more likely to occur over hundreds o f years (Loehle and
LeBlanc 1996, He et al. 1998). Insects, in particular, are extremely sensitive to climatic
variation because, as poikilotherms, virtually every aspect of the life cycle is influenced by
temperature. Some consequences of rising temperatures on insect herbivores have already
been documented, including genetic adaptations to shifts in seasonal timing (Bradshaw and
Holzapfel 2006) and range expansions into formerly unsuitable habitats (Battisti et al. 2005).
Indirect climatic effects that are mediated through other organisms such as the host plant or
community associates may also alter the balance of external pressures on insect populations
that influence colonization and reproduction success (Bentz et al. 2010). In order to predict

3

the effects that climate change will have on insect disturbance regimes we need long term
data that will allow us to identify and decouple climatic effects from the stand-dynamics
components of the system to determine how these factors interact. This should result in an
integrated risk model accounting for the major drivers of the disturbance regime that can
inform forest management decisions in response to environmental changes.

1.2 Study objectives and rationale
The broad objectives of this study are to develop a landscape-scale disturbance
history for the mountain pine beetle (MPB) (Dendroctonus ponderosae Hopkins
(Curculionidae: Scolytinae)) in north-central British Columbia (BC) near the northern extent
of the beetle’s range and to examine the long-term interactions between MPB outbreaks,
climate and host forest stand-dynamics, including other disturbance agents influencing the
outbreak cycle. The MPB is an aggressive bark beetle native to western North America
whose populations periodically increase to outbreak levels capable of causing tree mortality
over many thousands of hectares, and it is considered the most destructive biotic disturbance
agent of mature pine forest (Fumiss and Carolin 1977, Safranyik and Carroll 2006). In BC
lodgepole pine (Pinus contorta var. latifolia Engelm.) is the primary host although most pine
and occasionally other species such as Engelmann spruce are also attacked (Fumiss and
Carolin 1977, Unger 1993). Lodgepole pine-dominated stands cover approximately 14
million hectares of forestland in BC and account for roughly 25% of the provincial timber
supply (BC Ministry o f Forests 1995) so the socio-economic and ecological impacts of MPB
epidemics can be severe. Currently western Canada is experiencing the largest and most

4

devastating MPB epidemic on record that has affected over 17.5 million hectares o f forest
land in British Columbia (BC) (BC Ministry o f Forests, Lands and Natural Resource
Operations 2011) and has recently become established on the eastern slopes of the Rocky
Mountains and in west-central Alberta (Cullingham et al. 2011).
Our current understanding o f MPB outbreak dynamics is based almost wholly on
research from climatically optimal habitats where infestations have historically been most
severe, but recent changes in climatic conditions are causing marginal environments to
become increasingly susceptible to outbreaks (Bentz et al. 2010). Some of the most dramatic
changes are expected to occur at the northern range limit where climatic susceptibility has
been historically low (Carroll et al. 2004). A major concern here is that host availability is
not limiting and climatic changes could provide MPB access into previously unaffected pine
forests to the north and the boreal forest to the east that extends across the county from
Alberta to Nova Scotia (Cullingham et al. 2011). With the exception of the Forest Insect and
Disease Survey (FIDS) records that began aerial surveys in 1959 and some written accounts
of historical outbreaks in north-central BC dating back to 1948 (Wood and Unger 1996),
there is very little empirical evidence about outbreak dynamics near the northern range limit.
It is unclear whether outbreak regimes respond similarly to the factors driving epidemics in
the better studied, climatically optimal parts of the range or if universal management
principles apply across the province. Broad structural and compositional differences between
the northern-interior and central-interior forests suggest this may not be the case, and the
interactions between stand dynamics and climate on MPB disturbances in the north is
unclear. What is certain is that if current warming trends continue, climatically marginal
areas will become increasingly important MPB habitat and serve as primary sources for

5

range expansion. Outbreak management in these areas will likely become a priority as this
occurs, which will require a sound ecological understanding of the MPB disturbance regime
and the major factors driving outbreak development. This study will provide the long-term
records necessary to describe the interactions between MPB, climate, and stand dynamics on
an ecologically appropriate time scale for the northern-interior pine forests of BC. It will also
serve as baseline information for evaluating existing and future susceptibility to MPB under
the prevailing climate change scenarios and may help in managing range expansion into
adjacent forests that share more structural similarities with my study area than with the
better-studied pine forests to the south.

1.3 Background information on the mountain pine beetle

1.3.1 Life cycle and population processes
The MPB life cycle, described in detail by Safranyik and Carroll (2006), consists of
four stages: eggs, larva, pupa and adult, all o f which occur beneath the bark of the host tree
except during the dispersal phase of mature adults. Populations are typically univoltine,
producing one generation per year that begins in late July to mid-August when dispersing
adults mate and females deposit eggs into galleries excavated in newly colonized host trees.
Eggs hatch within a week of oviposition and larvae immediately begin feeding on the
phloem tissue of the inner bark by excavating larval mines perpendicular to the parental
gallery. Larvae pass through four instars, gaining cold-hardiness as they develop (Amman
1973). MPB overwinter as late instar larvae and complete development in the spring once
temperatures begin to warm. By June they pupate, developing into teneral adults between

6

late June and early July. Tenerals mature for an additional two weeks, feeding on the inner
bark and on spores of blue stain fungi lining the pupal chambers to support flight muscle
development (Reid 1958) and to pick up the microorganisms necessary for successful
colonization and reproduction in their new hosts (Safranyik et al. 1975).
The timing of emergence is temperature-dependent and usually peaks between midJuly and mid-August (Safranyik and Carroll 2006). Following emergence, beetle flights may
be short- or long-range, although the large majority of dispersals are within-stand (Safranyik
et al. 1992). Above-canopy long-distance dispersal is probably more important once locallyavailable host trees have been depleted (Safranyik and Carroll 2006). Once a host is selected,
usually through a combination of random landings, visual cues and direct assessments of
suitability after landing (Raffa and Berryman 1982, Pureswaran and Borden 2003), the
initiation of the egg gallery by the pioneer female prompts a series o f chemical interactions
that attracts a mate and results in the production of aggregation pheromones that instigate a
mass attack (Safranyik and Carroll 2006). For colonization to succeed, the attacking beetles
must overwhelm the host tree defenses which involve the production o f constitutive and
induced secondary resins (Berryman 1972) to slow the attack and either expel the beetles
from the tree or confine and kill them in lesions of dead tissue produced during the induced
resinosis (Safranyik and Carroll 2006). MPB rely on two strategies to overcome this
response. First, the aggregation pheromones produced by early attacking females act to
rapidly concentrate attacks that can exceed the resistance capacity o f the selected tree
(Safranyik et al. 1975). Second, the fungal spores acquired during adult maturation in the
pupal chambers o f the former host are inoculated into the new host upon attack, penetrating
the living cells of the phloem and xylem quickly (Safranyik et al. 1975) and disrupting the

7

production of defensive resins (Safranyik and Carroll 2006). Successful colonization results
in the death of the host tree and the production o f a new brood for attack the following year.

1.3.2 Epidemiology
There are four distinct phases in the MPB population cycle that are primarily defined
by the strength of the interactions between population size, stand resistance, and distribution
of MPB infestations on the landscape. Endemic populations are the most common phase,
persisting at very low levels in a stand in numbers too small to successfully kill a single
large-diameter tree and restricted to low-quality hosts with little defensive capacity
(Safranyik and Carroll 2006). Generally, after stand resistance peaks, usually around the
time of crown closure for lodgepole pine (Famden 1996), stands can maintain steady
endemic MPB populations in the low-vigour individuals that are competing for resources as
the stand undergoes self-thinning (Yoda et al. 1963). The combination o f low host quality,
patchy host availability, and small population sizes usually limits endemic populations from
growing larger, but any decrease in host resistance or increase in MPB population size can
allow an incipient-epidemic infestation to develop which is defined by the capacity of the
population to successfully mass-attack a single large-diameter tree in the stand (Safranyik
and Carroll 2006). Consecutive years o f favorable climate for beetle survival combined with
a period of stress or damage to the host tree are generally cited as drivers of the sustained
increases in beetle population size out of the endemic phase (Safranyik et al. 1974, Thomson
and Shrimpton 1984, Carroll et al. 2004).
Incipient-epidemic populations attack primarily large-diameter trees and experience a
sustained yearly population growth which is reflected by infestation patterns consisting of

8

clumps of attacked trees that establish throughout individual stands (Safranyik and Carroll
2006). Once these spot infestations begin to coalesce and new infestations disperse at a
landscape scale, the epidemic phase has begun in which MPB populations become
increasingly resilient to large losses in their numbers, the size and severity of the infestation
grows annually and the large-diameter component o f the affected stands are severely
depleted (Safranyik and Carroll 2006). Epidemic populations eventually collapse due to a
depletion of suitable hosts or an unseasonably cold event during the late fall or early spring
that drastically reduces the number of beetles surviving to colonize new hosts (Safranyik and
Carroll 2006). The post-epidemic phase simply describes the decline of the residual MPB
population back to endemic levels through various pathways o f reduced brood survival
(Safranyik and Carroll 2006) that may also include some additional pine mortality caused by
secondary bark beetle species that survived the MPB collapse (Safranyik et al. 1974).

1.4 Conceptual model for describing outbreak dynamics
Disturbances, although variously defined, generally refer to discrete events involving
a particular agent disrupting the structure of an ecosystem or community that alters resource
availability and/or the physical environment (White and Pickett 1985). In contrast to a
disturbance event, such as an insect outbreak, a disturbance regime refers to the spatial and
temporal dynamics of repeated disturbances over longer time periods, including
characteristics such as disturbance frequency, return interval, size, intensity and severity
(Turner 2010). Disturbance histories such as those developed using dendrochronology
provide the long term ecological perspective appropriate for describing disturbance regimes,

9

however, this is only part of the information necessary to understand the dynamics driving
the regime. For insect disturbances, outbreak dynamics refers to the long-term interactions
between the endogenous and exogenous elements involved in rapid insect population
fluctuations, particularly the factors that trigger and terminate outbreaks and define the
outbreak regime (Rose and Harmsen 1981). In this paper, I combine a reconstructed MPB
outbreak history with an analysis of the long-term ecological interactions driving the
disturbance regime to describe the outbreak dynamics of MPB in north-central BC, focusing
primarily on the factors that trigger the reconstructed outbreaks.
In order to describe outbreak dynamics, it is useful to examine the disturbance system
using the conceptual model originally developed for plant diseases by Stephens (1960)
referred to as the disease triangle. The disease triangle represents the occurrence o f a biotic
disturbance, in this case a plant disease, as the obligate interaction between a susceptible
host, a virulent pathogen, and a favorable environment for disease development (Stephens
1960, Agrios 2005). The same concept applied to an insect disturbance states that outbreak
development absolutely depends on the relationships between a susceptible host, an insect
agent and a suitable climate to support population growth. The long-term interactions
between these three necessary causal factors define the outbreak dynamics that provide a
more complete understanding the ecological process as a whole (Figure 1.1). For MPB, these
interactions are fairly well-defined based on observations of individual outbreaks, although
less is known about the long-term dynamics, particularly at the northern extent of its range.

10

MPB

Outbreak
regime
Host

Climate

d
Figure 1.1. Conceptual model of outbreak dynamics adapted from Stephens (1960) showing
the three necessary causal factors of an insect outbreak and the interactions between factors
(b, c, d) that over longer time scales determine the outbreak regime (a) for a given area.

11

1.4.1 MPB-host interactions
There are a number of important MPB-host interactions related to the conflicting
requirements for what defines a suitable host and a susceptible host. Suitable hosts refer to
those trees with characteristics, usually size-related, that are capable of supporting MPB
population growth, while susceptibility depends on the defensive capacity of the tree and is
often age-related. MPB prefer large trees within a stand because many of the characteristics
associated with tree size influence brood production and survival. In general the thicker bark
and thicker phloem of large trees enhance survival and attack success by producing larger
broods and larger beetles with faster life cycles and greater relative emergence rates
(Safranyik and Jahren 1970, Amman and Cole 1983, Safranyik and Carroll 2006) as a result
of the greater quantity and probably nutritional quality o f the resource (Safranyik and Carroll
2006) and the additional protection from temperature extremes, natural enemies and
sapwood drying (Reid 1963, Safranyik et al. 1974). The minimum diameter of trees for
successful attack and survival of MPB in is usually around 10 cm (Safranyik and Carroll
2006), however, Safranyik et al. (1974) found that only trees larger than 25.4 cm diameter
produced enough emerging beetles to sustain an outbreak. In risk modeling, stand averages
greater than 20.5 cm diameter are generally considered susceptible (Safranyik et al. 1975,
Cole and Amman 1980), assuming this structural threshold coincides with a reduction in tree
defenses.
Large trees are normally the most vigorous and resistant to attack (Shrimpton 1973)
until approximately 60 years, at which time stand resistance begins to decline rapidly
because the normal process o f stand aging results in mature trees competing for resources
(Safranyik et al. 1974). By the time stands reach 80 to 100 years, defensive capacity is low

12

enough that they are considered highly susceptible to MPB outbreaks (Safranyik and Carroll
2006). Because stand age is a major determinant o f outbreak risk, any factors that influence
age distributions also influence stand susceptibility, which is why fire cycles are often
included in outbreak risk models (Taylor et al. 2006). There is also an upper limit for agerelated susceptibility which Taylor and Carroll (2004) constrained to 160 years when
modeling stand level outbreak risk in BC, citing this as the age at which thinning phloem
associated with senescence began to adversely affect MPB population growth. In
climatically optimal regions, the size, severity and periodicity o f outbreaks is considered
highly dependent on the age and size structure o f the host forest (Alfaro et al. 2004, Axelson
et al. 2009), but it is unclear what role forest structure plays in marginal climates.

1.4.2 Climate-MPB interactions
Virtually every aspect of the MPB life cycle is controlled by temperature, and
accordingly much attention has been given to describing the individual controls at various
stages o f development and population phases. Carroll et al. (2004) and Safranyik and Carroll
(2006) provide detailed summaries of the specific climatic thresholds used to define outbreak
suitability, most of which relate to survival probabilities based on the ability to maintain
univoltine life cycles and seasonal synchrony (Logan and Powell 2001). Annual and
seasonal heat accumulations influence developmental rates and the timing of emergence and
peak flight (Reid 1962, McCambridge 1971) which largely control the potential for winter
survival and mass attack success (Safranyik and Carroll 2006). Cold tolerance is probably
the most well-studied climate interaction, in large part because it is the single greatest cause
o f MPB mortality (Cole 1981) and as such one of the most important controls of population
13

growth and outbreak potential. Seasonally varying cold-tolerance thresholds based on larval
development rates (Wygant 1940) can profoundly influence MPB population dynamics and
winter temperatures are consistently cited as the dominant climate driver in climate-outbreak
modeling studies (e.g. Aukema et al. 2008, Fauria and Johnson 2009).

1.4.3 Climate-host interactions
Climate-host relationships are probably the least well-understood interactions
influencing MPB outbreak dynamics. These indirect effects on MPB populations that are
mediated by the host tree are almost always discussed as a decrease in host-resistance
brought on by climate-induced water stress (Carroll et al. 2004), but the mechanisms o f this
are unclear because of complications with other climate feedbacks and regional ecosystem
differences (Bentz et al. 2010). There is little consistency among climate-outbreak
interaction studies as to the role o f drought and host stress in MPB outbreak dynamics, yet
enough evidence exists supporting the effect (e.g., Shrimpton and Thompson 1982,
Thompson and Shrimpton 1984, Trzcinski and Reid 2009) that it is included as an
explanatory variable in almost all climate suitability analyses. In this paper, I challenge the
assumption that precipitation alone is a universally appropriate proxy for host stress and
present an alternative method for examining climate-host interactions in relation to MPB
outbreaks.

14

1.5 Organization of this paper
The first objective of this study is to reconstruct a landscape-scale disturbance history
for MPB in north-central British Columbia (BC) where the outbreak regime is expected to be
highly sensitive to climatic variability and likewise highly susceptible to the effects of
climate change. Based on that reconstruction my additional objectives are to examine the
long-term interactions between MPB outbreak history, climate and forest stand dynamics,
including other disturbance agents that influence MPB, in order to define the major drivers
o f outbreak dynamics in this system and to aid in the development of predictive models for
assessing and managing future risk. Based on the outbreak dynamics model described and
illustrated in Figure 1.1, this study contains four component parts: (1) the outbreak
reconstruction, (2) the MPB-host forest interactions, (3) the climate-MPB interactions, and
(4) the climate-host interactions. Because these components are highly interrelated, rather
than divide the study into four data chapters and discuss each component separately, I
present the information as two larger chapters that, from an ecological perspective, provide a
more comprehensive and cohesive representation of MPB outbreak dynamics in northcentral BC.
In Chapter 2 , 1 develop a multi-century reconstruction o f MPB outbreaks at a
landscape-scale in north-central BC (Figure 1.1a) and use the characteristics of the
disturbance regime to examine interactions between MPB and the host forest (Figure 1.1b),
particularly the stand dynamics that are associated with outbreaks. This analysis includes
information about the role o f fire and other disturbance agents operating in the system, host
selection and stand susceptibility, changes in outbreak severity, and the successional
trajectory of this forest given the historic disturbance regime and with respect to expected
15

climate changes. I compare these results with those from other studies to address regional
differences and the implications of these for management. In Chapter 3 , 1 examine the direct
climate drivers influencing the timing of the reconstructed MPB outbreaks, including both
local and large-scale climate patterns with known relationships to MPB population dynamics
(Figure 1.1c) as well as the indirect climatic effects on outbreak timing that are mediated by
the host tree (Figure 1. Id). For the indirect effects I examine climate-tree growth
relationships first to determine how local climate influences host trees, and then analyze the
relationship between the host climate response and outbreak timing to determine specifically
how climate-driven tree-growth variability influences MPB outbreak development. By
combining the two climate interactions I provide a comprehensive explanation of the major
climate influences on MPB in north-central BC that addresses the relative importance of
both direct and indirect effects, and compare those relationships to climate-outbreak models
developed in other parts o f the province to explore regional differences. Lastly, I discuss the
implications of the climate interactions for developing predictive climate-outbreak risk
models. Chapter 4 presents a synthesis of all the results, integrating the roles of host forest
dynamics and climate in a comprehensive discussion of the interrelationships driving
outbreak dynamics in north-central BC. I also discuss the overall implications of these
findings for MPB management in BC, currently and with respect to climate change and the
future of this disturbance system. Lastly I address future directions this research could take
to improve our understanding of MPB outbreak dynamics in BC that should aid in risk
management and the development of predictive disturbance models in order to mitigate the
ecological and economic effects that future epidemics may have.

16

2. Historical outbreak dynamics and forest structure at the northern range limit for
mountain pine beetle in British Columbia

2.1 Abstract

Recent widespread mountain pine beetle (Dendroctonus ponderosae) (MPB)
outbreaks in western North America have caused unprecedented tree mortality across much
of the beetle’s known historic range and into novel habitats near the range margins. Our
current understanding of outbreak dynamics and their relationship to forest structure is based
almost wholly on research from climatically optimal habitats where infestations have
historically been most severe. As range expansion continues, marginal habitats may become
increasingly susceptible to outbreaks, yet little is known about the factors driving outbreak
development in these forest types or how the disturbance regime differs. This study uses
dendroecology and forest structure analysis to reconstruct historic outbreak dynamics at the
northern MPB distribution limit in British Columbia to extend the disturbance record beyond
the scope of documented history and to evaluate the long-term relationships between stand
structure, host suitability, and regime characteristics at the range margin. I sampled 22 sites
with evidence o f past infestations, and using mortality dates, survivor growth releases, and
tree vigor analyses I reconstructed five probable outbreak periods predating the current
epidemic. At least two o f these events fit the profile o f large epidemics occurring in the
1900s-1910s and again in the 1940s-1950s, with evidence of two 19th century infestations
and chronic localized mortality beginning around the 1980s. Stand age appears to be the
most important factor limiting outbreaks prior to 1870, indicating a strong relationship with
fire history at centennial time scales. Outbreak cycles were comparable to those reported in
other locations, yet beetle preference for larger and older trees suggests that provincial
17

susceptibility models may not accurately represent northern outbreak risk, particularly in
older stands. Following stand maturation, host availability does not appear to limit outbreak
severity or recurrence. Here density-independent factors associated with additional climate
pressures are likely more important. With enhanced climate suitability, outbreak scale and
severity will likely increase dramatically in this landscape which is currently not limited by
host distribution. My reconstruction can inform management strategies as this occurs and
outbreaks move into previously uninfested forests that share many structural and ecological
similarities with my sites.

2.2 Introduction
The mountain pine beetle (Dendroctonus ponderosae Hopkins (Curculionidae:
Scolytinae)) (MPB) is a native bark beetle o f western North America. It is an aggressive
species whose populations periodically increase to outbreak levels capable o f causing tree
mortality over many thousands of hectares, and is considered the most destructive biotic
disturbance agent in mature pine forests (Fumiss and Carolin 1977, Safranyik and Carroll
2006). Most pine species are susceptible to attack, but in British Columbia (BC) the most
common host is lodgepole pine (Pinus contorta var. latifolia Engelm.) which covers
approximately 14 million hectares of forestland and accounts for roughly 25% of the
provincial timber supply (Taylor and Carroll 2004). Currently, British Columbia is
experiencing the most severe outbreak on record, affecting over 10.1 million hectares at the
peak of the infestation in 2007 (Westfall and Ebata 2007). Widespread and ongoing
mortality of pine forests is also occurring throughout the western United States in numerous

18

outbreaks that began building in the late 1990s (Gibson et al. 2008). In addition to extensive
timber loss, the combined socio-economic and ecological impacts of MPB epidemics can be
severe. Beetle-related mortality may influence fuel distribution and fire behavior,
successional trajectories, carbon storage, watershed quality, wildlife composition, and
recreational values (Safranyik et al. 1974, McGregor and Cole 1985, Jenkins et al. 2008).
Despite a co-evolved insect-host relationship based on normative outbreaks that are
central to lodgepole pine disturbance ecology (Mattson 1996), current evidence suggests the
natural cycle of disturbance and forest development is shifting towards unprecedented
infestations throughout North America, including movement into novel habitats with
potentially devastating consequences (Logan et al. 2003, Logan and Powell 2005). Some of
the most dramatic changes are expected at high elevation and high-latitude sites near the
beetle’s range limits where changing climatic conditions have enhanced the potential for
beetle survival and reproduction (Carroll et al. 2004), yet we have very little empirical
evidence describing these effects. It is unclear whether outbreak dynamics in marginal
environments respond similarly to the factors driving epidemics in the better studied,
climatically optimal parts of the range or if universal management principles apply across
the province. What is certain is that if current warming trends continue, climatically
marginal areas will become increasingly important MPB habitat and serve as primary
sources for range expansion. Outbreak management in these areas will likely become a
priority as this occurs, particularly at the northern limit which borders a nearly continuous
supply of previously uninfested lodgepole pine and jack pine (Pinus banksiana Lambert) to
the north and east across the continent.

19

There are a number o f factors that can influence MPB population dynamics and
allow normally endemic populations to erupt to epidemic levels. These include cross-scale
processes (Raffa et al. 2008) that increase the availability or susceptibility of host trees
(Shore and Safranyik 1992, Taylor and Carroll 2004) or favor beetle population growth
(Safranyik and Carroll 2006). In general, the distribution of mature healthy pine trees with
thick phloem determines host availability, while climate suitability, primarily as it relates to
developmental temperatures, governs beetle success (Safranyik and Carroll 2006). Outbreak
dynamics, specifically the timing, frequency, scale and severity o f infestations, are
determined by interactions between the two (Royama 1992), and vary with geography (Cole
and Amman 1980). Other disturbance agents also commonly interact with outbreak regimes
(Rogers 1996). In BC, fire history affects the availability and susceptibility of host trees for
MPB (Parker et al. 2006) and is a key driver of the forest structure-outbreak relationship.
Because the temporal scale of these interactions can extend for centuries, managing for
major epidemics requires an examination o f long term disturbance histories to understand the
ecological significance of these processes. Currently, documented evidence o f historical
MPB activity in British Columbia is limited to ground surveys dating back to 1910,
province-wide aerial surveys beginning in 1959 (Wood and Unger 1996), and a small
number of dendroecological reconstructions conducted primarily in the southern interior
region of the province (e.g. Alfaro et al. 2004, Axelson et al. 2009). In the north, little
information exists detailing outbreak history prior to the 1940s, and it is unclear whether
outbreak models developed further south accurately describe the disturbance regime here.
Broad compositional and structural differences between BC’s northern and southern interior
lodgepole pine forests suggest that this may not be the case. Before we can begin to assess

20

the influence of climate or implications o f climate change, benchmark data describing the
outbreak cycle and its relationship to forest structure is critical for developing risk models
and management strategies suitable for marginal environments.
Dendroecology is a useful tool for reconstructing forest disturbance histories that pre­
date written records. It has been successfully used to study the disturbance dynamics of
defoliators (Swetnam and Lynch 1993, Speer et al. 2001, Zhang and Alfaro 2002) foliar
pathogens (Welsh et al. 2009), and a number of bark beetle species (Veblen et al. 1991b,
Stuart et al. 1989, Berg et al. 2006). For MPB, evidence of past outbreaks can be found
directly through the dating o f pine mortality events and indirectly in the accelerated radial
growth response o f surviving individuals in the affected stand. This indirect approach has
been used in a variety of forest types to study MPB outbreaks where mortality data is limited
(Heath and Alfaro 1990, Veblen et al. 1991a, Campbell et al. 2007, Axelson et al. 2009). A
number of additional dendroecological approaches can support of these reconstructions,
which are strongest when multiple lines o f evidence converge on a particular interpretation.
The purpose of this study was to use dendroecology to reconstruct the long term outbreak
dynamics of MPB at the northern extent of the beetle’s range in BC in relation to stand and
host susceptibility. Specifically, my objectives were to: (1) describe forest structure and
composition in the context of stand history, particularly the patterns of establishment and
disturbance influencing outbreak ecology; (2) characterize MPB-related mortality relative to
host selection (tree age, size, and vigor) and stand susceptibility and outbreak severity; and
(3) reconstruct the timing, frequency, and magnitude of historical MPB infestations beyond
the scope of written records using pine mortality, survivor growth releases, and host vigor
estimates as proxy evidence. I compared my results with studies conducted in the southern

21

interior of BC and the Rocky Mountains in the US to evaluate differences associated with
forest type and geographic location, and examined changes observed in the regime over
time. The relevance o f these results is discussed in the context o f provincial risk modeling,
outbreak management strategies and range expansion scenarios consistent with projected
climate changes.

2.3 Methods

2.3.1 Study Area
The study area is located in north-central British Columbia, at the northwest comer of
the Sub-boreal spruce (SBS) biogeoclimatic zone, with sites located in the SBSmc2 and
SBSwk3 subzones (Meidinger and Pojar 1991). It is roughly bounded by Takla Lake to the
north, Stuart Lake to the east, and Babine Lake to the west and south (Figure 2.1). This
region, northwest of Fort St. James, also represents the most active area for documented
outbreaks in the north (Safranyik et al. 1974). The Forest Insect and Disease Surveys (FIDS)
records describe at least one major outbreak predating the current one in this region between
the 1940s and the 1960s, in addition to numerous smaller infestations, with some stands
being classified as chronically infested (Wood and Unger 1996). The regional climate of the
Fort St. James area is characterized by short, warm summers (15.3°C average July
temperature) and cold, snowy winters (-11.3°C average January temperature), with moderate
average annual precipitation (487.1 mm), just under half falling as snow (Environment
Canada 2010). These climate normals are typical for SBS forests in British Columbia.

22

Upland coniferous forests dominate the landscape, with hybrid white spruce (interior
spruce) (Picea engelmannii Parry x P. glauca (Moench)) and subalpine fir (Abies lasiocarpa
(Hook.) Nutt.) leading late successional stands, and lodgepole pine and trembling aspen
(Populus tremuloides Michx.) establishing as serai associates. Other common tree species
include paper birch (Betula papyrifera Marsh.), Douglas-fir (Pseudotsuga menziesii (Mirb.)
Franco) and black spruce (Picea mariana (Mill.) B.S.P.). SBS forests occur at low to mid­
elevations (500-1300 m) across the entire study area, with patches of Engelmann Spruce Subalpine fir (ESSF) forest types occupying higher elevations. I restricted sampling to lower
elevation SBS forest types where lodgepole pine plays a more dominant role. Stand histories
in SBS forests can vary greatly because o f multiple, often concurrent, disturbance agents
acting within the same stands. Fire and mountain pine beetle are the most prominent
mortality agents, although fire in this system occurs on much longer time scales with cycles
o f at least 100 years to over 6000 years, and typically results in stand-replacing events rather
than selective thinning (Wong et al. 2004). For my specific subzones sampled, fire cycles
have been found to be 125 years in the SBSmc subzones o f the Lakes and Morice Timber
Supply Areas (Steventon 2001) and up to 500 years in the willow variant o f the wet cool
subzone (SBSwkl) (DeLong 1988). Also important in the SBS, although less likely to kill its
host, the two-year cycle budworm (Choristoneura biennis (Freeman) (Lepidoptera:
Tortricidae)) is a major insect defoliator o f interior spruce and subalpine fir that has caused
up to eight large outbreaks in my study area during the past two centuries (Zhang and Alfaro
2002, 2003).

23

2.3.2 Site Selection
Candidate stands for field sampling were identified using FIDS records (Wood and Unger
1996) and forest cover maps to select sites within known historic outbreak regions that were
conifer-dominated, 120+ year-old stands likely originating as serai lodgepole pine. During
ground surveys I screened candidate stands for signs of probable MPB outbreak history that
included the presence of standing dead or downed pine with visible pitch tubes, J-shaped
parental galleries, lateral larval galleries and/or exit holes in the remaining tree bark, as well
as indirect structural evidence including a pine-dominated mortality distribution and size
structure indicative o f a MPB-dominated disturbance history. This structural criterion
compensated for the degradation of direct evidence over time as a result o f weathering,
decomposition, and secondary beetle attack, and limited the effect o f any confounding
canopy disturbances targeting other species. Ultimately I selected 22 sites (Figure 2.1) for
the disturbance history, and 4 additional sites distributed across the study area showing no
evidence o f past disturbance to serve as a control representing the regional climate-tree
growth response of pine for my study area. Control sites were mature open-canopy lodgepole
pine-dominated stands growing on steep or well drained rocky slopes with no downed trees,
fire scars or other signs of tree damage or pathogen activity. Many o f the trees in these sites,
however, were recently dead from the ongoing MPB outbreak, estimated to have died within
1-3 years based on needle color and retention.

24

FI. S I James

■ Study H a
-fr Control wlM
Major roods

Kilometers

Figure 2.1. Detailed site and study area locations within British Columbia relative to the
2009 cumulative mountain pine beetle outbreak extent and lodgepole pine distribution.

25

2.3.3 Data Collection and Analysis

At each control site, I selected up to 20 of the largest, presumably oldest pine trees
and collected 2 increment cores at 0.3 meters height. Within the outbreak sites I established
variable-sized rectangular plots to sample all standing and downed trees greater than 10 cm
diameter at breast height (DBH). Plot size varied (300-625 m2) with stand density and dead
pine distribution to ensure suitable representations o f stand structure (minimum 20 live trees)
and mortality (minimum 5 dead pine). In cases of extensive canopy thinning due to the
current MPB outbreak, recently dead standing pine (needles present) were included to meet
the 20 live tree minimum. I cored all living trees at 0.3 meters and recorded species, DBH
and canopy position. Dead standing trees were inventoried similarly, but only pine were
cored and any evidence of mortality agent noted. At 12 sites I cored additional trees outside
plot boundaries to improve sample sizes, for a cumulative total (all sites) of 45 spruce, 9
subalpine fir, and 11 pine sampled out of plot. For all downed trees originating within the
plot I recorded species and DBH, noted cause o f death, and cut cross-sections from all pine
with wood intact enough to determine mortality dates. At 13 sites I collected additional
cross-sections (46 total) from downed pine outside of the plot to increase sample sizes. All
stems smaller than 10 cm were inventoried in 50 m2 circular subplots centered at the comers
of the rectangular plots which were averaged into standardized stems per hectare for each
site. Stems were inventoried in the following classes: 4-10 cm diameter trees, saplings that
were at least 1.3 m in height and less than 4 cm diameter, and seedlings that were less than
1.3 m tall. I tested for differences in recruitment for each class, including canopy trees
greater than 10 cm, which were associated with the number of MPB outbreaks reconstructed

26

for that site using ANOVAs or Kruskal-Wallis means comparisons tests depending on
whether the requirements for parametric statistics were met.
I mounted and sanded all cores and cross-sections following standard
dendrochronological procedures (Stokes and Smiley 1968), and used the list method
(Yamaguchi 1991) to visually crossdate each sample. Annual rings were measured using a
Velmex system and the program COFECHA (Holmes 1983) used to assist in correcting and
verifying the dating of all live and dead trees. Within sites I developed mean standardized
chronologies for each species represented by 10 or more trees using the program ARSTAN
(Cook and Holmes 1986) to fit a horizontal line through the mean o f each series to
emphasize deviations from mean growth, especially periods of sustained growth release that
could be linked to canopy disturbances (Veblen et al. 1991a).
To determine the timing and scale o f growth releases I used the program JOLTS
(Holmes 1999, unpublished, University of Arizona) to analyze each time series with a
moving windows procedure that calculates the difference between two adjacent sections of
the series at annual increments and flags any years where the differences exceed a userspecified threshold value. Release detection criteria were formulated using guidelines
established in similar studies (e.g. Rubino and McCarthy 2004, Berg et al. 2006, Axelson et
al. 2009), with some differences to account for additional disturbance signals observed in my
mean chronologies, namely growth responses to repeated two-year cycle budworm outbreaks
in spruce and fir. For these species I used sites with strong MPB signals, i.e. synchronous
pine mortality pulses, to calibrate detection parameters that identified growth releases
associated with known canopy mortality events while avoiding periods of growth recovery
following budworm defoliation. First I combined spruce and fir at each site into a single data
27

set based on high correlations between chronologies (Pearson r = 0.452 - 0.801, all p <
0.001) to strengthen the common signals I was interested in isolating (MPB and budworm)
and limit the effects of species-specific responses. I then fit a spline curve with a frequency
response period of 20 years over each series and tested 10 year windows of forward to
backward spline values for years with a doubling o f radial growth. The spline helped to
remove the effects o f sudden, short-term suppression/recovery events, while the 10 year
window and 200% release threshold ensured that flagged events would be sufficiently longer
and higher in magnitude than typical budworm recoveries (Zhang and Alfaro 2002) or
responses to inter-annual climate variability. The length of the final release chronologies was
determined by the interval containing at least 5 recording trees.
Releases in the standardized pine chronologies were smaller and less consistent than
in the spruce and fir, not surprising given lodgepole pine’s serai role and canopy-dominant
position in these stands. To improve detectability of pine releases I isolated the disturbance
signal in each series from baseline growth using the program OUTBREAK (Holmes and
Swetnam 1996, unpublished, University of Arizona) to subtract my control pine chronology
(undisturbed sites) from the pine collected at the outbreak sites. This produced corrected
index series emphasizing only deviations from baseline growth unique to the disturbed sites.
This is common practice in outbreak reconstructions for defoliators such as the two-year
cycle budworm (Zhang and Alfaro 2003), larch sawfly (Girardin et al. 2001) and western
spruce budworm (Campbell et al. 2006) which use a non-host species to remove the
influence of climate. My method differed in that I was able to use same species as a control
because canopy mortality events can be readily identified in the field. I developed the final
control chronology using ARSTAN (Cook and Holmes 1986) to fit a negative exponential

curve to 72 series (50 trees) selected for the strongest regional agreement. All series were
truncated to 1800 consistent with the recommended common signal strength (EPS > 0.85)
for chronology development (Cook and Kairiukstis 1992). I ran the OUTBREAK-corrected
pine series through JOLTS using a 10-year running mean procedure and the same release
magnitude (200% increase) used for spruce and fir. All JOLTS results were quality checked
at the individual site level against known mortality dates, then combined into a composite
data set for a landscape-scale analysis.
For all pine definitively identified as beetle-kill I calculated radial growth increments
for the 10-year period prior to death as an indication of tree vigor at the time of attack. Tree
vigor can be a useful indicator of MPB population pressure because it is positively related to
phloem thickness and host resistance (Cole 1973). Therefore, mortality in vigorous trees can
indicate population buildup into an outbreak phase. Conversely, endemic phase mortality
should be limited to weakened trees exhibiting slower radial growth. Each growth increment
was standardized by tree diameter to account for differences in tree size following the
methods of Bloomberg and Morrison (1989). Periodic growth was expressed as a percentage
of total tree size at the beginning of the period using the following equation: (x,/ r-x,) * 100
where x represents the sum of the ring widths for the interval i years and r-x* equals the
radius o f the tree at the beginning o f the interval. I used the mean 10-year periodic growth
rate for all living pine trees as my baseline for comparison.

29

2.4 Results
Species composition and stand structures varied in the relative distribution o f live
pine, spruce and subalpine fir (Table 2.1), however, the dead component of each stand was
consistently and predominantly large diameter pine (M = 29.1 cm, SD = 9.5) (Figure 2.2).
Mean DBH for positively identified beetle-killed trees (n=168) was slightly higher (M = 33.7
cm, SD=8.2). Live pine had a similar size distribution (M = 29.0 cm, SD = 8.3). Sites
experienced between 22-100% pine mortality, accounting for up to 74% of the total
mortality observed at a given site (Table 2.1). I collected a total of 963 increment cores and
199 cross-sections, including 327 dead pine. I did not find any fire scars that could be
sampled for fire dates. Mean live pine age was 212 years (SD=44) and ranged from 138-352
years. Mean pine age (at coring height) at the time of death for all samples and in confirmed
beetle-kill only was 185 years (SD=47) and 194 years (SD=43), respectively, and ranged
from 70-353 years (Figure 2.2). My earliest reconstructed mortality occurred in 1835.1
developed 54 standardized chronologies, and stand development followed a similar pathway
at each site, colonized initially by single-cohort, fast-growing pine, followed by spruce and
fir establishment soon after (Figure 2.3). Repeated suppressions consistent with the two-year
cycle budworm outbreaks described by Zhang and Alfaro (2002) were prevalent in the
spruce and fir chronologies. The majority of stands established in the mid to late 1700s with
the exception of two 17th century initiations and one in the early 1800s (Table 2.1).
Establishment dates for all live trees > 10 cm diameter (Figure 2.4) show an age
distribution ranging between 59 and 352 years old with pine and spruce dominating the older
age classes and subalpine fir dominating the more recent establishment events. The timing of
fire events, estimated by the stand establishment dates of the sites (Table 2.1) is consistent
30

with the pulses of pine establishment, the last of which occurred in the 1820s. Very low pine
recruitment continued until the 1860s, after which time no new pine trees established.
Subalpine fir and spruce establishment were continuous throughout the record, but subalpine
fir became much more successful around the mid-1800s when pine and spruce establishment
began to decline. Stem densities by site which include all the regeneration size classes
(Figure 2.5) show canopy trees, small trees that were below diameter threshold for aging,
sapling and seedling densities ranging from 425-1475 stems per hectare, 50-3300 stems per
hectare, 50-8500 stems per hectare and 300-17550 stems per hectare respectively. Subalpine
fir almost wholly dominates the stem classes smaller than those aged. The results of the
ANOVA and Kruskal-Wallis tests (Table 2.2) showed no significant differences at any
canopy level in stem density relative to the number o f reconstructed outbreak events a site
had experienced.
Within individual sites I observed synchronous and asynchronous pine mortality
(Figure 2.3). Synchronous pulses occurred throughout the 1940s-1950s, 1980s, and again in
2006-2007. Stand level growth releases generally followed within 1-4 years, but the
characteristics varied between sites and between species with spruce and fir exhibiting
consistently higher magnitude growth responses than pine. Release duration was also
variable, with periods of sustained high growth rates ranging from approximately 20 to over
100 years based on visual assessment of the site chronologies. Incremental pine mortality
over many different years was more common than large pulses within sites and relative to
these events stand-level growth releases were less clear, not surprising as plot averages can
mask the response by smaller proportions o f survivors. The JOLTS analysis identified the
strongest and most consistent responses at each site, resulting in the detection o f 1-2 major

31

release periods (>40% trees recording) at each site (Figure 2.4). Twentieth century major
releases generally corresponded to at least one mortality date and were recorded by both pine
and the non-host species, though not always at similar magnitudes. In some cases major non­
host releases coincided with known two-year cycle budworm outbreaks, but there was no
consistent response within sites to repeated defoliation events.

32

Table 2.1. Site inventories showing the distribution and dating o f live (L) and dead (D) (pine
only) trees > 10 cm diameter at breast height in each plot and the total number of dated
samples used to build the final site chronologies (C), including any additional trees sampled
outside plot boundaries. The only other species present was trembling aspen, which were
inventoried at two sites for a total o f 4 trees, none o f which had cores intact enough for
dendrochronological processing.
Lodgepole Pine

Interior (Black f ) spruce

Subalpine fir

Site

L (n)

D (n)

D (% )

C (n)

Master

L (n)

C (n)

Master

L (n)

C (n )

Master

CB3

18

5

23}

29

1779-2006

7

17

1784-2007

13

17

1884-2007

17 t

1828-2007

8

13

1814-2007

13

12

1845-2007

28

27

1869-2006

KZ16

23

18

69

34

1786-2007

SL25

5

5

21}

10

1808-2007

17 t
10

1821-2007

TS7

11

14

52

15

1807-2006

0

CB5

10

14

56

31

1780-2006

2

1846-2006

5

1901-2006

CB4

10

15

68

31

1780-2006

6

1792-2006

6

1837-2006

TE13

8

13

37

22

1684-2007

5

13

1693-2007

24

24

1684-2007

FL23

5

9

39 §

18

1824-2007

12

14

1826-2007

7

14

1839-2007

11

1766-2007

47

47

1748-2007

1795-2006

18

18

1863-2006

1678-2007

29

29

1733-2007

1809-2006

11

11

1867-2006

1800-2007

18

15

1805-2007

TK24

5

10

38

16

1743-2007

7

TS8

4

9

82

11

1783-2006

4

NA21

4

9

27}

16

1654-2007

4

CB6

3

10

48

18

1780-2006

3

PV22

5

21

45

27

1788-2007

18

11
18

TS11

2

10

50

10

1787-2006

4

10

1797-2007

51

48

1864-2006

NZ18

3

15

61

15

1753-2007

10

10

1774-2007

31

31

1820-2007

12

CB9

2

19

73

NN19

1

10

42

NZ17

1

14

37

TB14

1

22

54

1762-2006

2

1790-2006

36

35

1859-2006

1780-2003

6

1792-2007

25

23

1879-2007

20

1748-2007

8

10

1786-2007

15

16

1778-2007

21

1764-2002

3

12

1825-2007

37

36

1825-2007

1786-2007

22

20

1786-2007

1785-2006

16

16

1884-2006

29

29
486

1787-2007
1684-2007

KL15

1

28

68

28

1778-2007

9

12

TS10

0

13

68

10

1781-1973

11

11

0

34

74

28
431

1780-2007
1654-2007

0

TR20
Total

202

1678-2007

Notes: Master chronology length indicates earliest and latest dated rings, but chronologies
were only built for sample sizes >10. Blank cells indicate sites with too few dated trees for
chronology development. Pine mortality (D) is summarized as total number dead trees (n)
and as a percentage o f total plot mortality (all species). Pine led mortality in all sites except
where noted.
t Interior spruce found in all sites except KZ16 which was entirely black spruce
| Unknown (unidentified) species led total mortality
s Subalpine fir led total mortality

33

LIVE TREES
(b) Canopy position

(a) Live diam eter distribution

400

S pecies
i
I
m

200

I Unknown
I Subalpine fir
Spruce

■■Pine

c

8

200-

o

100

Main

Mid

Sub

Canopy class

DBH class
DEAD TREES
(c) Dead diam eter distribution

(d) Pine mortality ag e distribution

150
40 -

100

20 -

O

10

20

30

40

50

60

DBH class

100

200

300

Age a t death

Figure 2.2. Composition and structure summarized for all trees > 10 cm DBH. Live (a) and
dead (c) diameter distributions show relative composition at 5cm size class intervals. Canopy
distribution for living trees only (b) shows compositional differences between the dominant
and codominant (main), intermediate (mid) and suppressed (sub) canopy classes. Mortality
age distribution for dead pine only (d) shows age at the time o f death (at sample height) for
all pine with intact pith and outer ring dates in 10 year age classes.

34

100
c

0)
>
<D
O)
C
13

I

50 -

£
<f>
2
if)

i i i i ^ i i r ? | " t i mrImpi 'InpmLmmJ'im

o

4000

g>

MPB Outbreak
I Stand establishment

o

o o

t

o

Mi l

o o |

subalpine fir
spruce
H I lodgepole pine (live)
i
I lodgepole pine (dead)

3000

1

JZ
U.

&
vt

2000

E
<D
if)
1000

0 "♦'"TTTT

1600

1650

1700

1750

1800

1850

Pith Year

Figure 2.3. Establishment dates for all living trees greater than 10 centimeters diameter and
all dead pine from all sites combined (bottom graph). Open circles approximate the timing o f
fires in my study area and are based on stand-initiation dates. Arrows indicate the timing of
mountain pine beetle outbreaks reconstructed by this study. The proportion of sites recording
each event is shown in the top graph.

35

Number of outbreaks
►
Canopy trees >10 cm DBH

2000 -i

1

;

2

m
I

4

Subalpine fir
Spruce
Lodepole pine
ITrembling aspen

1500 -

S 1000 to

500 -

| «M|.

■

I
I

Small trees <10 cp DBH

I

3000
2000
E

a?
<55 1000
0
Saplings
8000
6000
nj

§ 4000
E

I

S

to 2000

Seedlings
20000 - |

I

15000 ■
<0
§ 10000 E

&

to

5000

M ill
W ' - f f l N ’- o t n o ^ a r t i D S t ' t N f f i i f l c o r t n
CN' r “ C / j C / ) C N T" T," C N C N J a 3 C D ,”

T” C D ^ ’ C>J u j Q 3 T,* C v J ^ —

Site

Figure 2.4. Stem densities at each site for all canopy levels ordered left to right by number
of outbreaks experienced in that site. Number of outbreaks was determined using the pine
mortality and survivor release data to determine whether mortality events occurred at that
site during the reconstructed outbreak periods. All spruce were interior spruce except for site
KZ16 which was entirely black spruce.
36

Table 2.2. Stem density summaries by canopy class and number of outbreaks that occurred.
Results o f the ANOVA and Kruskal-Wallis means comparisons results are listed for each
test of stem density differences between the groups defined by number o f outbreaks.
M 1 (SD)b

M2 (SD)C

Canopy tree

M (SD ) 3
840 (274)

930 (302)

907(313)

796 (259)

Subcanopy tree

1092(811)

1125(796)

1233(673

721 (535)

Sapling

1864(1906)

2890 (3300)

1425 (722)

Seedling

5375 (4108)

4860 (2424)

4417 (2085)

Canopy class

M3 (SD)d

M4 (SD)*
708 (237)

*Fdf
0.636 3 1 a

P
0.602

1488(1365)

0.849 3 1 8

0.485

1178(1078)

2438(2010)

*1.8113

0.613

5750 (6396)

6800 (3990)

0.283 3 1 8

0.837

aMean stem density for all sites with standard deviation in parentheses
bMean stem density for sites experiencing one outbreak with standard deviation in parentheses
cMean stem density for sites experiencing two outbreaks with standard deviation in parentheses
dMean stem density for sites experiencing three outbreaks with standard deviation in parentheses
cMean stem density for sites experiencing four outbreaks with standard deviation in parentheses
*All results are from ANOVA tests except for the sapling group which used a Kruskal-Wallis test

37

In the composite data set I observed up to 5 periods predating the current outbreak
with coincident growth release, pine mortality, and above average tree vigor estimates
indicative of MPB canopy disturbances (Figure 2.5). Across all years, annual mortality
ranged from 0-24 events occurring in up to 8 sites concurrently, and release percentages
ranged from 0-43.3% of trees recording an event in either species group at a maximum o f 22
sites. The most widespread event, based on number of sites recording, occurred in 1945 and
the most severe, based on highest mortality, in 2006. Despite a declining sample depth for
reconstructed mortality further back in time, mortality events match the timing o f growth
releases in the 1830s, late 1860s-1870s, 1900s-1910s, 1940s-1950s, and the 1970-1980s.
Non-host releases are out of sync with pine releases in the 1970s and to a lesser degree in the
1870s and 1940s. I believe this to be an artifact of the detection procedure that occurs when
canopy thinning occurs in close succession with two-year cycle budworm
defoliation/recovery events. Although the moving windows procedure is calibrated to ignore
short, low magnitude growth fluctuations occurring independently, it is unable to decouple
the compounding effects of coincident responses resulting in slightly skewed timing or
magnitude of the flagged event. For this reason, the timing of pine releases probably more
accurately approximates outbreak timing.

38

(b) NZ17, Asynchronous mortality

(a) TS11, Synchronous mortality
Subalpine fir

Ic

I

?

£
|

40 -

O

n

n

40 —

"

Interior spruixt
x
2 —

%

5m
6on
10

6 —I

Lodgepole pine

3 —

3 —

20 — |

8 —i Pine mortality

2H

T
1800

1850

1900

1950

1750

2000

1800

1850

1900

1950

2000

Year

Figure 2.5. Standardized chronologies for lodgepole pine, interior spruce and subalpine fir
from two sites exhibiting (a) synchronous pine mortality (site TS11) and (b) asynchronous
pine mortality (site NZ17). Reference lines for each chronology represent mean index
values, with sample depths drawn below each species. Dashed vertical lines indicate periods
of two-year cycle budworm outbreaks identified by Zhang and Alfaro (2002, 2003).

39

0

i

,1111,.... , .... |

,

,

TS11

NN19

I I I I I I I | r"f

r"i"'i i r i i | i r
NA21

1750

1800

1850

1900

1950

2000

Year

Figure 2.6. JOLTS results from selected sites representing synchronous and asynchronous
mortality patterns. Black bars show pine mortality, dark (light) grey bars show the
percentage of non-host (pine) trees experiencing a doubling o f growth or greater for that
year. Reference lines are drawn at 40% to delineate major release events. Dashed vertical
lines indicate periods of two-year cycle budworm outbreaks identified by Zhang and Alfaro
(2002, 2003).

40

Standardized ten year periodic growth increments at the time o f death for beetle
killed pine (M = 2.31%, SD = 1.66) exceeded the mean live pine growth increment (M =
3.38%, SD = 2.29) in the 1900s, 1940s-1950s, and fairly regularly from the late 1970s
onward (Figure 2.5d), each corresponding to a release and/or mortality period identified
above. Lower than average radial growth increments were recorded relatively continuously,
but occurred exclusively during the 1920s, 1930s and the 1960s during periods of
comparatively low release and mortality. Across the length of the entire record (1800-2007),
landscape-scale outbreak events occurred on average every 29.6 years. There appeared to be
a shift towards larger events around the 1870s when release percentages more than double
for pine, though that increase was not observed in the non-host record until the turn of the
century. I observed a similar shift in recent years when mortality levels exceeded the highest
on record, however, release data is unavailable for comparison after 1997 because o f the tenyear window used in the JOLTS calculations.

41

(a) Proxy outbreak reconstruction

(b) Pine mortality

(c) Sample depth

Figure 2.7. Landscape scale outbreak reconstruction based on composite data for all sites.
Proxy reconstruction (a) shows the timing of pine (light grey bars) and non-host (dark grey
bars) growth releases relative to (b) pine mortality events. Sample depth figure (c) shows the
number of surviving pine (hatched light grey fill) and non-host (solid grey fill) trees used to
calculate release percentages, the number of sites at which trees were releasing (black line),
and the number of sites recording pine mortality (black bars) annually. Ten year relative
growth increments standardized by tree size (d) represent tree vigor at the time of death for
all confirmed beetle-killed pine trees. Reference line represents the mean ten year
standardized growth increment for all living pine trees, above which I consider pine
mortality to be occurring in healthy trees.

42

2.5 Discussion

2.5.1 Stand history
The establishment patterns and species compositions observed are typical for subboreal forests of north-central BC, consisting o f single cohort serai pine stands giving way to
shade-tolerant stands dominated by spruce and/or subalpine fir. The structure o f the tree
strata below the canopy (< 10 cm DBH) indicates that subalpine fir will ultimately dominate
the landscape in the absence of another pine or spruce establishment event. Fast early growth
rates for pine and spruce suggest these stands originated following stand-replacing fires
(Veblen et al. 1991b). In the majority of sites, fires probably occurred during the mid- to
late-18th century, with the exception of two stand initiations occurring earlier in the 1600s
and two in the early 1800s. The stand age similarities for most sites in the study area suggest
that infrequent, large severe fires are important, but over much longer time periods than
MPB outbreaks. The prevalence of fire-intolerant subalpine fir in the smaller diameter
classes and a lack of fire scar evidence suggests that fire has been absent from these stands
since initiation. Under a typical stand-replacing fire regime, long fire-free intervals such as
this are likely a prerequisite for outbreaks to develop. Age distributions o f the living trees
and stem densities for all trees, saplings and seedlings suggest that pine, and to a lesser
degree spruce, persistence in these stands is dependent on the return o f fire. Because we
specifically targeted older sites in order to reconstruct MPB outbreaks, these statements
about the relationship between fire and MPB only apply to remnant stands that have escaped
more recent fires and not to the fire cycle for the landscape as a whole. The MPB outbreaks
reconstructed following stand maturation did not appear to have any influence on
regeneration patterns, either by causing pulses of tree establishment (Figure 2.3) which are
43

commonly linked to outbreaks in the southern-interior o f BC (Axelson et al. 2009, Axelson
et al. 2010) or through the alteration of stand densities at any canopy level (Figure 2.4, Table
2.2) even through repeated outbreaks. Rather, recruitment in these mixed-species SBS stands
in the absence o f fire appears to be a function of continuous regeneration by shade-tolerant
species where composition ultimately depends on the competitive advantage o f higher
recruitment rates by subalpine fir.
With MPB, evidence for repeated two-year cycle budworm outbreaks was pervasive
in my non-host chronologies and I suspect that both disturbance agents have a strong and
widespread influence on stand development over similar time scales. Methodologically, the
persistence o f both signals in the tree-ring record introduces some problems for decoupling
outbreak history, but ecologically this interaction may be an important component of
outbreak and stand dynamics and should be explored further. Regardless o f defoliation
history, I know from the composition and structure o f dead trees that my sites experienced
selective pine mortality, eliminating the need to differentiate between mortality processes
affecting other species. Direct beetle evidence on more than half o f the dead pine sampled
suggests that for two centuries MPB has been the most influential mortality agent affecting
these stands.

2.5.2 Host selection, stand susceptibility and outbreak severity
On average, beetle killed pine trees in my study area were larger than the 25.4 cm
DBH threshold defined by Safranyik et al. (1974) as necessary to sustain an outbreak, and
comparatively large living pine today suggest that stand averages, at least for pine, have

44

exceeded the 20.5 cm DBH minimum used to define size structures at risk for an outbreak
(Safranyik et al. 1975, Cole and Amman 1980) for much of the past century. Early outbreaks
in the 1800s were probably less skewed towards larger diameter trees because the stands that
had reached susceptible ages were younger. Smaller trees may have limited these early
outbreaks from developing into larger events, because the more recent preference for larger
than average trees suggests that beetle populations here may not be able to buildup and
maintain epidemic population sizes in hosts that are considered to be in susceptible size
classes elsewhere. Furthermore, the persistence of large diameter living pine in age ranges
that closely match the age ranges o f dead pine indicates that outbreak severity has
historically been low and limited by some factor other than host availability, contrary to the
density-dependent theories proposed in climatically optimal regions (Cole and Amman 1980,
Alfaro et al. 2004). This is consistent with the asynchronous mortality pattern observed
resulting in the incremental removal of pine from the canopy in multiple, low severity events
rather than large, devastating outbreaks. Amman and Baker (1972) attribute higher
proportions of large diameter pine trees at higher elevations to increased host survival,
presumably related to the adverse effects of temperature on beetle success. The higher
latitude of my study area likely has a similar effect. Larger trees provide a number of
advantages over smaller diameter trees for northern beetle populations subject to greater
climatic pressure. Most of these can be attributed to thicker bark and thicker phloem which
have been shown to enhance survival and attack success by producing larger broods, larger
beetles, faster life cycles and greater relative emergence rates (Safranyik and Jahren 1970,
Amman and Cole 1983, Safranyik and Carroll 2006).

45

Pine mortality in this area is also skewed towards older trees than those typically
considered susceptible to beetle attack. Taylor and Carroll (2004) constrained age-related
susceptibility to between 80 and 160 years when modeling stand level outbreak risk in BC,
the upper limit representing the age at which thinning phloem associated with senescence is
thought to begin adversely affecting population growth. My mortality dates for confirmed
beetle kill averaged 194 years and up to 353 years, suggesting that these northern pine stands
remain susceptible to beetle attack long after provincial models suggest. Above average
periodic growth increments in many of the beetle killed pine indicate that these infestations
are not endemic populations restricted to weakened trees in the stand. Rather, beetle success
in vigorous trees suggests that these older pine stands can support epidemic populations. The
relatively continuous mortality observed in less vigorous trees also suggests these stands
maintain a nearly continuous supply of endemic beetle populations. The prevalence of large,
old pine trees available likely enhances the stability of these populations between outbreak
phases relative to post-outbreak stand structures dominated by smaller diameter, vigorous
young pine that is much less favorable for beetle survival and reproduction. This has
dramatic implications for climate change impacts in the region because beetle populations
here may be able to respond more quickly and successfully to increasingly favorable
conditions, with potentially greater risk for large epidemics to develop if high-latitude
endemic populations function more like incipient-endemic populations held in check by
more continuous exogenous pressures.

46

2.5.3 Outbreak regime

Between 1800 and 2007,1 reconstructed six periods of MPB disturbance occurring in
the 1830s, late 1860s-1870s, 1900-191 Os, 1940s-1950s, 1980s, and the epidemic underway
at the time of the study. Survey records confirm the timing of a widespread, severe
infestation in the 1940s-1950s, as well as numerous, chronic infestations occurring
throughout the 1980s and 1990s until the current outbreak began building in the 2000s
(Wood and Unger 1996). My proxy data extends the record 140 years describing three
additional release periods consistent with the known outbreak response, although the
degradation o f direct mortality evidence prior to 1900 requires cautious interpretation of
these events as MPB outbreaks. The two-year cycle budworm signal in the non-host
chronologies may have influenced the release analyses for these species, though conservative
thresholds and supporting evidence from other sources suggest the filtering procedure was
successful in identifying canopy disturbance events. To reduce the effect of any timing
errors, I limit my discussion to a decadal resolution and use direct pine mortality evidence
wherever possible to describe outbreak details.
The current outbreak is the most severe on record, but the scale, severity, and
mortality characteristics of the 1900-1910s and the 1940-1950s events also fit the profile of
large epidemics (Amman and Schmitz 1988, Safranyik and Carroll 2006). The 1980s saw
abundant mortality, but these were localized infestations until the 2000s. Steady population
pressure from the 1980s could have contributed to the unprecedented mortality o f the 2000s.
Consistent with the growth release data, I suspect that prior to 1870, outbreaks would have
been small-scale, low-severity disturbances that failed to coalesce across the landscape, such
as the 1830s event. Based on stand ages from all my plots, landscape-scale susceptibility

47

would have occurred around 1860-1870 (Raffa and Berryman 1982), establishing reasonable
potential for outbreaks to grow large around this time. My landscape-scale outbreak return
cycle of approximately 30 years agrees well with the 20 to 40-year cycle described by
Amman and Schmitz (1988) and the 42 and 36-year cycles observed in south-central BC
(Alfaro et al. 2004, Axelson et al. 2009), however, unlike these studies, stand age and size
structure in the northern interior is probably less important for controlling outbreak
periodicity. Rather, the mortality and stand development patterns reconstructed here suggest
that following stand maturation, density-independent factors such as climate suitability are
the dominant influence limiting outbreak development.
2.5 Conclusions
My outbreak reconstruction at the northern limit o f the MPB range served two
purposes: first, to lengthen survey records to an ecological time scale appropriate for
examining long term disturbance processes and interactions; and second, to provide an
outbreak susceptibility profile for forests that are distinctly different from those found in
climatically optimal parts of the range. Differences between my study and others highlight
the importance of geography and forest structure in modeling outbreak risk in marginal
habitats, especially in response to climate change projections for these sensitive zones.
Rising outbreak severity demonstrated in recent years suggests this area is already
experiencing an altered disturbance regime, and ongoing range expansion into structurally
similar, previously uninfested forests to the north and east calls for a better understanding of
outbreak dynamics in these types o f environments. This is particularly true with respect to
susceptibility models developed in other parts of the province which may underestimate
outbreak risk in the north, especially in older stands with a large proportion of pine
48

remaining. Although the effects of climate change can only partly be addressed by this study
as I did not include any information about past climate in the reconstruction, my results
suggest that historically unfavorable climate has been more important than stand structure in
regulating outbreak development. This is probably due, in part, to long fire-free intervals
allowing large cohorts of mature pine to persist across extensive areas. As climate becomes
more suitable, I expect outbreak severity will continue to increase throughout the northern
range margin following the distribution of available hosts, which are currently not limiting in
the landscape. My reconstruction may help to forecast outbreak potential and develop timber
management plans accordingly as this occurs.

49

3. Local and large-scale climate interactions with a multi-century record of mountain
pine beetle outbreaks in north-central British Columbia

3.1 Abstract
Phytophagus insect populations are extremely sensitive to climatic variation at
virtually every stage o f their life cycle. Recent unprecedented outbreaks o f bark beetles
across western North America and rising concerns of the effects that climate change may
have on future disturbance regimes has increased attention on the specific links between
climate and outbreak development. The objectives of this study were to examine the direct
and indirect effects of known climatic drivers on mountain pine beetle outbreaks near the
northern limit of the species range where the disturbance regime is expected to be extremely
sensitive to climatic variation, yet very little is known about specific climate responses here.
Using a 200-year outbreak history available for the study area I examined the long­
term relationships between outbreak timing and the known local and large-scale climate
drivers using superposed epoch analysis (SEA) to determine which factors were most
important in outbreak development. I examined the indirect climatic effects mediated
through the host tree-climate response by including a climate-sensitive host chronology
developed using dendrochronological methods as a climate variable in the analysis. Lastly I
looked at climate suitability for the study area as defined by existing models to visually
assess the individual climate parameters and support the results of the SEA tests.
My results indicated that persistent positive winter temperature anomalies and the
large-scale patterns o f climate variability that drive them (the Pacific Decadal Oscillation
and El Nino-Southern Oscillation) were the most important predictors of when outbreaks
will occur. Extreme cold events and the large-scale patterns associated with them (the Arctic
50

Oscillation and Pacific North American pattern) showed no skill for predicting outbreaks,
but this does not discount the importance of minimum temperatures in influencing inter­
annual population dynamics. Short-term host stress was an important precursor to outbreak
development, however, differences in geography and species responses to climatic
variability dictate that a precipitation may not be a universally acceptable proxy for host
stress. Over longer time periods, prolonged host vigor was significantly related to outbreak
development, suggesting that healthy trees may provide certain advantages in marginal
environments subject to greater climatic pressures.

3.2 Introduction
The importance of climate in the population dynamics of phytophagus insects is well
established, however, recent widespread epidemics and rising concerns over the impacts of
climate change on forest disturbance processes have increased attention on the specific links
between climate and outbreak development (Dale et al. 2001, Bentz et al. 2010). Climatic
effects can be direct or indirect and can influence insect herbivore physiology, behavior or
success at virtually every stage of the life cycle. The most well-studied direct effects are
temperature-driven, primarily because poikilotherms are extremely sensitive to thermal
conditions and, relative to other types o f climatic variation, projected increases in global
temperature are expected to have the most dramatic influence on population dynamics (Bale
et al. 2002). Some consequences of rising temperatures have already been documented,
including genetic adaptations to shifts in seasonal timing (Bradshaw and Holzapfel 2006)
and range expansions into formerly unsuitable habitats (Battisti et al. 2005). Indirect effects
include any climate-induced changes that are mediated through other organisms such as the

51

host plant or community associates. Over short time scales the most important indirect effect
is often associated with a decrease in host resistance brought on by climate-induced water
stress, but the mechanisms of this are not well understood and often complicated by
interactions with other critical feedbacks or by regional ecosystem differences (Bentz et al.
2010 ).

Forest insects capable o f attacking and reproducing in healthy trees are key drivers of
ecosystem change because of their ability to dramatically influence tree mortality rates. In
the western United States and Canada, the mountain pine beetle (MPB) (Dendroctonus
ponderosae Hopkins (Curculionidae: Scolytinae)) is a native bark beetle that experiences
periodic population eruptions responsible for landscape-scale mortality in mature pine
stands. Timber losses have important economic and social impacts (Parkins and McKendrick
2007), but outbreaks can also have significant ecological effects on wildlife habitat (ChanMcLeod 2006), fire behavior (Jenkins et al. 2008), carbon storage (Kurz et al. 2008), and
hydrology (Redding et al. 2008). Currently western Canada is experiencing the largest and
most severe MPB epidemic on record that has affected over 17.5 million hectares o f forest
land in British Columbia (BC) (BC Ministry of Forests, Lands and Natural Resource
Operations 2011) and MPB has recently become established on the eastern slopes of the
Rocky Mountains and in west-central Alberta (Cullingham et al. 2011). Situated at the
northern extent of the beetle’s range, MPB populations in BC are expected to be particularly
sensitive to climatic variation and have received much attention over the previous decade in
an effort to define the role of climate in outbreak dynamics and develop predictive models
for future susceptibility across the expansive host range to the north and east.

52

Climate suitability models originally developed by Safranyik et al. (1975) to assess
beetle habitat suitability and later adapted for spatio-temporal analyses (Carroll et al. 2004,
Moore et al. 2005) combine three categories of variables critical to the climate-outbreak
relationship to estimate outbreak risk. These categories include variables to describe heat
accumulation as it pertains to developmental rates o f individual beetles and population
synchrony, cold mortality potential during the overwintering period, and host stress which
includes any climatic factors linked to changes in host defensive capacity. The specific
variables may differ depending on the spatial and temporal scope of the model. Detailed
laboratory and field observations have identified threshold values for each parameter that are
used to quantify favourability. Heat accumulation thresholds define the temperature cues
necessary to synchronize life cycle processes with critical seasonal events in order to
maximize reproductive success (Logan and Powell 2001). Examples include the minimum
temperature requirement for flight that triggers emergence and promotes mass attack
(McCambridge 1971) or the cumulative number of degree days specified for specific periods
that initiate egg hatch and ensure developmental rates sufficient for overwinter survival and
a univoltine life cycle (Carroll et al. 2004). Cold mortality thresholds define the seasonally
varying tolerances of MPB larvae to winter temperatures (Logan et al. 1995), below which
survival rates have limited potential to sustain epidemics (Safranyik and Carroll 2006).
Winter temperature suitability can be assessed by counts of days below the temperature
curves developed by Wygant (1940) that detail the frequency and timing of extreme cold
events or by seasonal and monthly averages that represent overall winter temperature trends.
Host stress thresholds are less specific than the temperature requirements defined for MPB
development and survival. Typically host stress is defined by some measure of water deficit

53

over two or more years, usually early growing season precipitation totals that are below a
long-term average although other indices have also been used to approximate precipitation
variability (Carroll et al. 2004). A more direct indicator of host stress can be measured using
the radial growth of host trees prior to attack, where reduced growth represents a reduced
capacity for a defensive resin response and consequently lowered host resistance (Safranyik
et al. 1974).

More recently large-scale patterns of climate variability with atmospheric
teleconnections in western Canada have become a focus of MPB climate suitability studies
because of the potential for improving ecologically relevant climate predictions over larger
areas and longer time scales than local variables allow (Stenseth et al. 2003). Linking largescale climate indices to outbreak dynamics can reduce the complexity associated with
individual climate variables by providing a more generalized probability o f when suitable
conditions occur. The benefit of these relationships is that, as we improve predictions of
phase shifts, we may also increase our ability to forecast the probability of MPB outbreaks
occurring. The dominant climate indices with biologically-relevant links to MPB ecology in
BC include the Pacific Decadal Oscillation (PDO), the El Nino - Southern Oscillation
(ENSO), the Arctic Oscillation (AO), also known as the Northern Annular Mode (NAM),
and the Pacific North America pattern (PNA) (Stahl et al. 2006a, Fauria and Johnson 2009,
Thomson 2009). The PDO is a climate pattern centered in the North Pacific Ocean that
oscillates on a 20-30 year cycle and is generally expressed in western North America as
winter (October-March) temperature anomalies and spring snowpack and streamflow
anomalies, with warm (cool) phases linked to above (below) average temperature and below
(above) average precipitation and snowpack in my study area (Mantua et al. 1997). Stahl et
54

al. (2006a) linked the frequency of MPB cold mortality events in BC with winters with
strong negative PDO indices, suggesting climate favourability and potential for MPB
expansion during positive and neutral PDO years. The PDO may also indirectly influence
MPB success if host susceptibility is affected by PDO-driven climate trends, but this
relationship has not been studied. ENSO phases, which are often assessed by the Southern
Oscillation Index (SOI) to describe the warm El Nino (negative SOI) and cold La Nina
(positive SOI) conditions associated with fluctuating sea surface temperatures in the
equatorial Pacific, share similarities with PDO teleconnections in BC, but cycle over a much
shorter time scale of 2-7 years (Shabbar and Bonsai 2004). BC winters are generally warmer
(colder) and drier (wetter) following the onset o f an El Nino (La Nina) event (Shabbar and
Khandekar 1996, Shabbar et al. 1997), with the warm ENSO phase associated with
significantly higher (lower) frequencies of warm (cold) spells and extreme warm (cold)
temperature events (Shabbar and Bonsai 2004). The cold ENSO phase is generally
associated with the opposite effects.

Interactions between large-scale climate indices can also influence climatic
suitability for MPB outbreaks. For example, Bonsai et al. (2001) noted that ENSO
temperature impacts were enhanced when in-phase with the PDO, consistent with the
findings of Stahl et al. (2006a) who showed the highest probability for MPB cold mortality
was associated with negative PDO winters that occurred during La Nina or neutral ENSO
years. The PNA is a PDO-driven atmospheric circulation pattern that has been associated
with significantly warmer January minimum temperatures in its positive phase during
periods with above average NAM index values (Thomson 2009), which could also influence
overwintering survival rates. Alone, AO phases, which describe the strength and location of
55

the westerly winds around the northern polar areas, show inconsistent relationships with
winter temperatures (Zhou et al. 2001), but are important in regulating the occurrence of
Arctic air flows to the south which can influence cold spell duration and warm spell
frequency in BC (Shabbar and Bonsai 2004). Stahl et al. (2006a) attribute the highly
negative AO indices of 1984/85 to the cold mortality events that ended the Chilcotin
outbreak despite a positive PDO winter.

Much o f what we know about climate suitability for MPB is based on detailed studies
of individual climate components in a controlled setting, such as cold tolerance at various
life stages (Logan et al. 1995), or observations of a set o f climatic conditions surrounding
single outbreaks, especially the Chilcotin outbreak that began in the 1970s in south-central
BC (e.g. Thomson and Shrimpton 1984, Aukema et al. 2008). Some studies examine climate
suitability independently of outbreak data using established climate drivers to classify
outbreak risk based on historical climate profiles (e.g. Safranyik et al. 1975) or to examine
the influence o f large-scale ocean-atmosphere patterns on specific parameters o f climate
susceptibility such as winter temperature variability (e.g. Stahl et al. 2006b, Thomson 2009).
Very few studies have compared long time series of MPB outbreak data directly against
long-term climate data to investigate the consistency o f these relationships through multiple
outbreak events. This is in part because detailed outbreak histories for BC generally do not
exist before 1959 when the Canadian Forest Service, Forest Insect and Disease Survey
(FIDS) began conducting annual aerial overview surveys of these disturbances provincewide. Since this time only two major epidemics have been documented, the Chilcotin
outbreak of the 1970s and 1980s and the present outbreak which began to develop in the
1990s, although there is evidence o f another large epidemic in north-central BC that was in

56

decline around the time monitoring began and numerous smaller infestations throughout the
1980s (Wood and Unger 1996, Hrinkevich and Lewis 2011). Although the FIDS (19591996) and later BC Ministry of Forests annual forest health surveys (1999-present) provide
nearly continuous MPB outbreak data since 1959, they have limited value for investigating
low-frequency climate-outbreak relationships with patterns such as the PDO that have only
undergone one phase shift during the survey period. Fauria and Johnson (2009) caution in
their study linking PDO and AO to the aerial survey records that significant relationships
observed over these short time series do not provide enough evidence to determine causality
of the relationships observed. For PDO, this would require at least some replication of
known phase shifts which would include an outbreak record at least as long as the
instrumental PDO record (1901-present).
In this study I use a proxy MPB outbreak record covering the period 1800-2007
derived from tree-ring data (Hrinkevich and Lewis 2011) to examine long-term climateoutbreak relationships in north-central British Columbia. Specifically my objectives are: (1)
to examine the interrelationships between the known climatic drivers o f MPB population
dynamics, host tree growth and outbreak timing, (2) to analyze climate-tree growth
relationships in order to determine which climatic effects are mediated through the host
trees, and (3) to evaluate existing models of climate suitability for MPB in the study area
relative to outbreak history. My study addresses a number o f gaps in existing climateoutbreak research, including a need for increased understanding of outbreak dynamics in
climatically sensitive regions at the northern range limits where the effects o f climate change
are expected to be most severe. Furthermore, by using long-term data I provide additional
evidence about the causal links between climate and outbreak development, particularly for

57

low-frequency patterns. Lastly, my use of a host tree growth chronology allows me to
directly examine the relationship between climate-induced host stress and outbreak
development rather than using indirect climate variables that may not accurately reflect
changes in host susceptibility. Although climatic links to host stress were identified in early
investigations o f host-climate-outbreak relationships (Shrimpton and Thomson 1982,
Thomson and Shrimpton 1984), more recent work typically assumes the same climate-tree
growth relationships apply and uses a growing season precipitation variable to represent host
stress. I question whether this assumption holds in pine ecosystems other than the Chilcotin
plateau where they were developed and present an alternative method for examining the role
of host stress.

3.3 Methods

3.3.1 Study area
The study area is located in the northern interior o f British Columbia, at the
northwest comer of the Sub-boreal spmce (SBS) biogeoclimatic zone (Meidinger and Pojar
1991). It is roughly bounded by three o f BCs largest natural lakes, Takla, Stuart, and Babine
to the north, east, and southwest respectively (Figure 3.1). Historically, this region near Fort
St. James represents the most active area for documented MPB outbreaks at this latitude and
lies close to the historic northern limit of the beetle’s range in BC (Safranyik et al. 1974).
Survey records report a major epidemic in the area in the 1950s followed by a period of
chronic infestations beginning in the 1980s and building throughout the 1990s (Wood and
Unger 1996), with ongoing infestations coalescing into the present epidemic during the mid-

58

2000s. An independent tree ring-reconstruction of MPB outbreaks in the study area
(Hrinkevich and Lewis 2011) details these events as well as three additional outbreak
periods, including one large epidemic in the early 1900s and two 19th century events.
The regional climate of the Fort St. James area is characterized by short, warm
summers and cold, snowy winters (15.3°C and -11.3°C average July and January
temperatures), with moderate average annual precipitation (487.1 mm) distributed relatively
evenly throughout the year (Environment Canada 2010). These are typical climate normals
for SBS forests in British Columbia. Upland coniferous forests dominate the landscape,
characterized by late successional stands of hybrid white spruce (interior spruce) (Picea
engelmamii Parry x P. glauca (Moench)) and subalpine fir {Abies lasiocarpa (Hook.) Nutt.
Lodgepole pine {Pinus contorta Dougl. Var. latifolia Engelm.) and trembling aspen {Populus
tremuloides Michx.) primarily lead serai stands. Other tree associates include paper birch
{Betula papyrifera Marsh.), Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) and black
spruce {Picea mariana (Mill.) B.S.P.).

59

t Ft. St James

. MPB outbreak sites
★ Climate sites
■kilom eters

Figure 3.1. Study area map showing the locations o f each of the climate sites used to
compile the master climate chronology as well as the locations of the study sites used to
develop the mountain pine beetle outbreak reconstruction. Inset shows location o f study area
in British Columbia relative to the distribution of lodgepole pine and the 2009 cumulative
mountain pine beetle outbreak extent.

60

3.3.2 Chronology development
I sampled four sites distributed across the study area defined by the outbreak
reconstruction (Hrinkevich and Lewis 2011) for my climate-tree growth analysis (Figure
3.1). To maximize the climate signal and limit any stand-level influences on tree growth, I
selected only mature open-canopy, pine-dominated sites on steep or well-drained slopes that
showed no signs of past disturbance, although some of the trees were recently dead or dying
from the ongoing mountain pine beetle outbreak. Needle color and retention in these stands
suggested attack dates within three years o f the sampling date. I selected up to 20, but as few
as 4 where availability was limited, of the largest trees at each site and extracted two
increment cores from each, with the exception of site CB1 where only single cores were
collected, resulting in a combined total of 82 cores from 58 trees.
Increment cores were mounted and sanded following standard dendrochronological
methods (Stokes and Smiley 1986) until the cell structure of individual growth rings was
clearly visible under a dissecting microscope. I used the list method (Yamaguchi 1991) to
dot and visually crossdate cores from all the living trees with a known outer ring date, then
used the living-tree marker years to assign calendar years to the dead series. I measured
the annual rings of all samples using a Velmex measuring system and used the program
COFECHA (Holmes 1983) to assist in correcting and verifying the dating o f each series.
Series with anomalous radial growth patterns were truncated or discarded to produce a final
chronology with the highest series intercorrelation containing 67 cores from a total of 47
trees (Table 3.1). High intercorrelations among sites indicate a strong regional tree growth
signal that applies across the study area. I developed a standardized mean chronology using
the program ARSTAN (Cook and Holmes 1986) to first fit a negative exponential curve or a

61

linear regression to each series to remove the effects o f age on tree growth and then to
remove autocorrelation from each of the detrended series using the built-in autoregressive
modeling function. The final chronology was computed using the residuals from the
autoregressive modeling to obtain an estimate of mean growth for each year with serial
persistence removed to enhance the common climate signal (Cook 1985). I confirmed that
the master chronology maintained the recommended common signal strength (EPS > 0.85)
for the entire period of record (Cook and Kairiukstis 1992).

62

Table 3.1. Descriptive statistics computed using the program COFECHA for individual
climate sites used to compile the final chronology and for the master chronology
Site
CL1
CB1
PINC
CBHC

Na
5(4)
17(17)
36 (20)
9(9)

rb
0.650
0.684
0.653
0.710

MSC
0.262
0.299
0.249
0.315

Period*5
1850-2005
1875-2006
1800-2005
1850-2005

1800-2006
M aster
0.269
67 (47)
0.633
a Number of series in each chronology with the number of trees in brackets
b Pearson’s correlation coefficient calculated between all series in chronology
c Mean sensitivity is an indicator of the relative variation in ring width in series
d Period covered by each chronology
'Total number of years in the chronology

63

Length6
156
132
206
156
207

3.3.3 Climate data
I obtained temperature and precipitation records from the nearest meteorological
station in Fort St. James for the climate variables known to influence MPB population
dynamics (Table 3.2). Monthly variables were compiled from the Adjusted and
Homogenized Canadian Climate Data (AHCCD), a quality-controlled dataset of monthly
averages that incorporates adjustments to the original station data to address changes in
instrumentation and observation procedures over time (Environment Canada 2011).
Variables based on daily temperature thresholds were compiled from the National Climate
Data and Information Archive records available online from Environment Canada
(http://www.climate.weatheroffice.ec.gc.ca/prods_servs/index_e.html#cdcd) and I used
Wygant’s (1940) cold mortality curves to calculate the number of days below the 50% and
100% mortality curves. I also calculated 2-5 year running means for each variable to
examine relationships associated with more persistent climate conditions such as consecutive
warm winters or prolonged droughts.
For the four dominant large-scale climate modes thought to influence MPB success
in BC, I obtained data online from the National Center for Atmospheric Research, Climate
Analysis Section (SOI, http://www.cgd.ucar.edu/cas/catalog/climind/soi.html and AO,
http://www.cgd.ucar.edu/cas/jhurrell/indices.html), the Joint Institute for the Study of the
Atmosphere and Ocean (JISAO) at the University of Washington (PDO,
http://jisao.washington.edu/pdo/), and the World Data Center for Paleoclimatology, Boulder
and the NOAA Paleoclimatology Program (PNA,
http://www.ncdc.noaa.gov/paleo/metadata/noaa-recon-8697.html) (Table 3.2). Following
Redmond and Koch (1991), I focused on winter climate associations with these indices when

64

North Pacific teleconnections are strongest. I defined winter SOI by the preceding JuneNovember index values (Redmond and Koch 1991), PDO using December-February (DJF)
means (Stahl et al. 2006a), and AO using mean December-March (DJFM) index values
(Hurrell 1995). I used a reconstructed December-January (DJ) PNA index (Trouet and
Taylor 2009) because the instrumental records available were too short (1948-2010) to
coincide with a sufficient number of outbreak events for my analysis. As with the
temperature and precipitation data, I also calculated running annual means for each index to
examine lower frequency trends associated with outbreak development. I included 2-10 year
averages for PDO, PNA and AO, and 2-5 year averages for SOI because o f the shorter
periodicity between phases.

65

Table 3.2. Climate variables known to influence mountain pine beetle population dynamics
with commonly used threshold values for climate suitability modeling. Data sources for each
variable are included. The rationale for including these variables in the analysis and the
specific effects they represent in the population dynamics of MPB is described in detail in
the text.
Category3

Climate Variable

Predicted threshold11

Period11

Source data

Heat
Accumulation

August maximum
temperature

> 18.3 °C
(McCambridge 1971)

1895-2010

(Environment
Canada 2002,
2011)

Summer and annual
degree days

> 305 dd > 5.5 °C (Aug 1 to
end o f growing season)

1895-2002

(4)
(3)

> 833.3 dd > 0 °C (Aug 1
to July 31)
(Carroll et al. 2004)
Cold Mortality

Host stress

Large-scale
indices

tt it

tt tt

it tt

October-March
minimum daily
temperatures

Days above mortality curve
(Wygant 1940)

1896-2001

Winter (DJF) monthly
temperature
(mean/minimum)

DJF > average
(Fauria and Johnson 2009)

1896-2010

Early growing season
(AMJ) precipitation

2+ years < average
(Thompson and Shrimpton
1984)

1896-2010

Radial growth
(residual chronology)

Index values <1
(Shrimpton and Thompson
1982)

1800-2006
(6)

(Hrinkevich
and Lewis
2011)

PDO

DJF > 0
(Stahl et al. 2006a)

1901-2009

(Mantua et al.
1997)

June-November < 0
(Redmond and Koch 1991)

1867-2009

DJF > 0
(Fauria and Johnson 2009)

1899-2011

DJ > 0
(Thomson 2009)

1725-1999

SOI

AO (NAM)

PNA (reconstructeed)

(3)
tt tt

(4)
tt tt

(4)

(4)
(5)

(Trenberth
1984)
(Hurrell 1995)

(4)
(5)

(Trouet and
Taylor 2009)

a Major categories of climate variables thought to influence MPB outbreak dynamics
b Predicted thresholds used to define climate suitability for MPB with source
c Period indicates time covered by the data series. The number of outbreaks that occurred
during that time period which were included in the SEA analysis are in parentheses.
66

3.3.4 Climate, tree growth and outbreak relationships

I calculated correlation and response functions using DendroClim2002 (Biondi and
Waikul 2004) to test for relationships between my residual climate chronology and the
instrumental temperature and precipitation data. These procedures calculate a sequence of
coefficients between the annual tree-ring index values and a pre-defmed set of monthly
climate variables using Pearson’s product moment correlations and principle components
regression models, respectively, generating bootstrapped error estimates for significance
testing. Comparing the results of both functions helps to minimize problems associated with
collinearity in the climate predictors and the identification o f spurious relationships (Biasing
et al. 1984). I tested a 17-month window o f mean and minimum temperatures and total
precipitation beginning in May of the preceding year through the current December for the
period 1912-2004.
To examine the relationships between the climate variables and the timing o f MPB
outbreaks, I used the proxy outbreak record developed by Hrinkevich and Lewis (2011) to
identify event years that represented the first year of an outbreak in which large pulses of
mortality well exceeded the baseline mortality levels observed between outbreak periods. I
defined event years based on the best combination of data available for that period including
direct evidence found in survey records and the reconstructed mortality dates, as well as
indirect outbreak indicators such as tree vigor at the time of death and survivor growth
releases as described in detail in the reconstruction. For the period 1800-2007 the six MPB
event years included in this analysis were 1837, 1868, 1906, 1945, 1987, and 2006.1
investigated current year and lagged relationships between the start of an outbreak (event
years) and each of the climate variables of interest including the residual tree ring

67

chronology representing the host climate response (Table 3.2) using superposed epoch
analysis (SEA). SEA is a compositing technique that calculates mean conditions in a
response variable before, during, and after a set of events which have been superposed onto
the variable such that all events occur at time zero and the annual variation surrounding each
event can be summarized across the entire time series. It is widely used to study fire-climate
relationships (Baisan and Swetnam 1990, Swetnam 1993, Grissino-Mayer and Swetnam
2000, Hessl et al. 2004), but has also been used to analyze climatic and tree growth variation
associated with insect outbreaks (Nola et al. 2006, Pohl et al. 2006). I used an 8-year window
surrounding each MPB event including the 5 years preceding and 2 years following the onset
of the outbreak which allowed us to examine climate relationships that may be more
important in the years building up to or immediately following a MPB population eruption
than in the event year itself, such as a lagged response to winter temperature associated with
slow population growth. I used the program EVENT (Holmes and Swetnam 1994) to
compare my climate variables within the event window to a set o f 1000 randomly selected
event years used to calculate bootstrapped confidence intervals. Significant departures of
actual from simulated means were evaluated at the 95% and 99% levels using the normal
approximation of the means and percentile rank methods depending on the distribution of the
input variables (Mooney and Duval 1993).
I examined climate suitability in the study area using the known favourability
thresholds for MPB populations (Table 3.2) to visually assess which factors were most
limiting in this region and how climate variability over time related to outbreak history. I
plotted values for summer and annual degree days, August maximum temperature,
standardized spring precipitation and number of days below the 50% and 100% cold

68

mortality thresholds for the complete period on record (1895-2002), using the threshold
values to classify years as either suitable or unsuitable for each variable in the model.
Precipitation suitability was defined by a two-year trend based on the current and previous
year’s standardized score so that only the second year of consecutive annual precipitation
deficits was classified as favorable. I overlaid the reconstructed outbreak periods over all the
climate variables as a reference, defining the starting years by the same criteria used to
identify event years for the SEA and estimating the end of the outbreak period based on
declining trends in tree mortality. The most recent outbreak that began in the area in 2006
was not included in the analysis because daily temperature data was only available through
2002 .

3.4 Results

3.4.1 Climate-MPB outbreak-host tree growth relationships
The SEA results for the local climate variables were largely non-significant, with the
exception of mean DJF temperatures that, at 5 year running averages, were significantly
higher (p < 0.05) in the year of an outbreak (Figure 3.2a). I found no relationships over
shorter time periods (< 4 years) or for minimum DJF temperatures at any frequency. August
maximum temperature, summer and annual degree days, and April-June precipitation were
also unrelated to outbreak timing at any lag or running average. I tested May-June
temperature and May-June precipitation in addition to the original variables compiled
because o f the significant correlations between these variables and host tree growth, but

69

(a) DJF temperature

(b) Decern ber-February PDO

(c) June-November SOI

(d) Standardized ring width

_____________________2 -]

__ __________________ 0.2 -

--------------------- ----------

<n

0)

E

I

'

J T u fl

3

.

■

m i d i u .'

II I UU”

•1 -

n f l m „ iri

00

<n

E

5
0to
£
e
a
&

5 year

j

4 year

|

5 year

■ nl in n .. Tin
.

10 year

|

j

u u n ” U1-1 00 '
-0.1 -

-5 -4 -3 -2 - 1 0 1 2

-5 -4 -3 -2 - 1 0

1 2

-5 -4 -3 -2 - 1 0

1 2

-5 -4 -3 -2 -1 0

1 2

Event window year

Figure 3.2. Superposed epoch analysis results for variables showing significant relationships
to outbreak events. Variable names are listed across the top. The top row shows annual
relationships between the variable and outbreak events. The bottom row shows
representative examples of the lower frequency trends that were significantly related to
outbreak timing, with the trend length indicated for each variable. Event window years
represent the 8 years analyzed around the outbreak events, with year zero (indicated by an
arrow at the top of each graph) corresponding to the year the outbreak occurred and negative
(positive) numbers corresponding to years preceding (following) the event. Hashed lines
represent 95% and 99% confidence intervals and shaded bars indicate significant
relationships at that window position.

70

found no relationship with outbreak timing for either. O f the large scale indices tested, PDO
and SOI showed significant relationships with outbreak timing, while PNA and the AO were
unrelated at any lag or running average. PDO was consistently higher (p < 0.05) in the 3
years preceding an outbreak for short-term trends (1-3 year averages), with outbreaks
occurring in years associated with 4-10 year trends o f higher winter PDO (Figure 3.2b). SOI
showed a pattern similar to that of the PDO, except with consistently lower index values
associated with outbreaks. Below average annual SOI values occurred 4 years prior to an
outbreak, with more prolonged negative trends occurring successively closer to the event
year. Five year trends of SOI were significantly lower in the year of an outbreak (Figure
3.2c).
SEA results for the standardized tree ring chronology were more complex than those
of the climate indices. Unlike the consistent associations observed for PDO and SOI at all
running averages, the relationship between tree growth and outbreak timing reversed
between the higher and lower frequency trends o f the running means. In the year o f the
outbreak I observed significantly lower tree growth, with this effect persisting in the highfrequency ring-width signal (1-3 year running means) for two years after the start of the
outbreak (Figure 3.2d). At lower frequencies (4-10 year running means) higher than average
tree growth was consistently associated with the onset of an outbreak, beginning three years
before the event and moving closer to the event year with progressively longer trends.
Above-average tree growth over a period of four years preceded an outbreak by three years,
while a 10-year trend preceded an outbreak by one year. These relationships are wellrepresented when annual and 5-year running means of the standardized tree ring index are
plotted against the timing of outbreak periods, with a distinct punctuated period o f low

71

growth occurring at the time o f the outbreak start and longer-term persistent above-average
tree growth occurring in the years preceding (Figure 3.3).
The correlation and response function results were comparable in the direction and
magnitude of the relationships between monthly climate and tree growth. For ease of
interpretation I present only the correlation results, noting any differences in the results of
the two approaches. Tree growth was significantly positively related to winter temperature
for November and March (Pearson r = 0.262 and 0.211, p < 0.05), and to seasonal averages
for November-March and December-February (r = 0.253 and 0.222, p < 0.05) (Figure 3.3).
Warm spring temperatures in May and June negatively influence tree growth (r = -0.261, 0.379, p < 0.05), with some evidence for a similar effect in September, although this
relationship is not supported by the response function. The strongest climate-tree growth
relationship is for spring precipitation which shows significant positive correlations with ring
width for May and June (r = 0.399 and 0.521, p < 0.05).

72

Outbreak periods
- Annual host growth

.5 year growth trends

X

0

T3
C

i
O
)
c
£
08-

Year

Figure 3.3. Annual and 5-year running means of the tree ring chronology used in the SEA
analysis overlaid on the reconstructed outbreak periods. Distinct periods of punctuated host
stress (below-average growth) occur at the start of each outbreak, while 5-year trends of
persistent above-average growth precede the start o f each outbreak.

73

0 .5

— T em perature

Preciptiation

L_

c

Q>
5
S0=)
8 o.o
C

o

£
o
o
k_

-0.5

M

J

S

N

J

Previous growing
season

M

M

J

S

N

M

Current growing
season

J

S

N

J

Previous growing
season

M

M

J

S

N

Current growing
season

Figure 3.4. Climate-tree growth correlations between mean temperatures and total
precipitation and my master residual climate chronology over a 17-month window beginning
in May of the preceding year and ending in the December of the current year for the period
1912-2004. Asterisks indicate significant correlations for that month that were assessed
using 95% confidence intervals (black lines).

74

3.4.2 Climate suitability

The climate suitability results are summarized for each variable based on the timing
of suitable years relative to the reconstructed outbreak periods. Summer degree day totals
were variable with respect to the heat accumulation threshold prior to the 1920s, but after the
mid-1920s all but three years were above the minimum requirement. The extended period of
unsuitable summer heat accumulation corresponded with the start of the 1900s outbreak, but
the latter two outbreaks occurred during periods of extended suitability for this variable.
Annual degree day totals were above the minimum suitability thresholds for all the years on
record (1896-2002) and showed no consistent pattern o f above- or below-average heat
accumulation relative to outbreak timing (Figure 3.4). August maximum temperatures were
also favorable throughout most of the period on record with the exception of two years in the
early 1900s and a group of years beginning in 1955 that did not meet the maximum
temperature threshold. Unfavorable August temperatures generally occurred between
outbreaks and never coincided with the onset of one o f these periods, although there was an
extended period of August maximum temperature below 18.3°C that coincides with the
estimated timing o f the 1940s outbreak collapse. Precipitation was more variable around the
suitability threshold than the heat accumulation variables and had a higher percentage o f
unfavorable years, but there was no consistent pattern relative to outbreak timing. Only the
earliest outbreak began during consecutive years of below-average precipitation, although
extended dry periods did occur during two o f the reconstructed events. O f the six highly
suitable periods defined by 3 or more consecutive dry years, half occurred between outbreak
events and no dry periods persisted for longer than one year after 1983. Cold mortality days
with temperatures below the 50% threshold were common throughout the record, ranging
between 0-23 days/year (M=5.3, SD=5.2), with 19 years free from cold spells. Temperatures
75

below the 100% mortality threshold were less common, occurring between 0-13 days/year
(M=1.5, SD=2.6) in approximately half o f the years on record. Two of the three MPB events
began in years with no cold mortality days. The earliest event, however, began in a year with
4(1) days below the 50% (100%) threshold which are below average for the number of
annual cold mortality events calculated over entire the time period. Above average numbers
o f cold mortality events occurred during each o f the reconstructed outbreak periods.

76

x

*

I"

1900

i

|

1 910

■»

i"

1 9 20

i

|

1 930

i

|

1 940

r

|

1950

i

f

i

| 11111......| .......i.....i

1960 1 97 0

1 980

|

1 990 2000

Figure 3.5. Climate suitability parameters for the study area based on established threshold
values for each parameter listed in Table 3.2. White boxes indicate favourable climate years
for mountain pine beetle and black boxes indicate unfavourable climate years. Cold
mortality days (bottom graph) show the number of days below the 50% mortality threshold
(dark grey bars) and 100% mortality threshold (black bars). Light grey vertical shading
across all plots shows the timing o f the reconstructed outbreak periods.

77

3.5 Discussion

3.5,1 Climate- MPB outbreak-host tree growth relationships

The links I observed between climate, MPB outbreaks, and host tree growth
emphasized the importance of long term patterns of climatic variability in the development
of large infestations. This is not surprising considering that population buildup over multiple
years is a necessary precursor for local incipient-epidemic populations to coalesce into
landscape-scale epidemics (Safranyik and Carroll 2006). Winter mean temperature was the
only direct climate driver associated with the start o f an outbreak, and only for the longest
trends (5 years) of warmer than average conditions. Although inter-annual winter
temperature variability has been shown to influence local population dynamics and the finescale spatiotemporal distributions of MPB outbreaks (Aukema et al. 2008, Fauria and
Johnson 2009), the resolution o f my MPB data prevents us from examining these detailed
population fluctuations. I do show, however, that over centennial time scales, only lowfrequency warm winter trends are directly related to the initial MPB population eruption
associated with the onset of large outbreaks.
The importance of persistent positive winter temperature anomalies is also evident in
the significant association between outbreak timing and the two large-scale climate patterns
with known teleconnections to winter temperature in BC. PDO in particular has consistently
been cited as an important driver of MPB outbreaks because of its influence on winter
temperature over 20-30 year phases (Fauria and Johnson 2009, Thomson 2009). My results
extend these findings by showing that the relationship between MPB population eruptions
and positive PDO index values holds over multiple phase shifts and four distinct outbreak

78

periods. The significant associations for all PDO trends tested (1-10 years) indicate there
may be a predictive advantage of PDO over direct winter temperature measurements that
only relate to outbreak timing for persistent trends (>4 years), whereas even annual PDO
conditions appear to be useful for modeling outbreak risk. Specifically, my data showed that
outbreaks occur as early as 3 years following a positive PDO winter, with lag time
decreasing with successively longer annual positive PDO anomalies.
The negative SOI relationships to outbreak timing are consistent with the known
links between El Nino and positive winter temperature anomalies in Canada (Shabbar and
Khandekar 1996), which is likely the mechanism driving MPB population eruptions during
negative phases o f SOI. As with PDO, the consistent association at all running means tested
(1 -5 years) reflects the lower frequency variability of the SOI relative to annual winter
temperatures and the relative predictive advantage o f using SOI phases to model outbreak
risk. My results showed outbreaks erupting 4 years following a negative SOI, with the
probability of an outbreak occurring sooner increasing with successive years of persistent El
Nino conditions. Five year negative trends were significantly associated with an outbreak
occurring in that year which, relative to the average ENSO cycle of 2-7 years, indicates a
high probability of an outbreak occurring during prolonged negative anomalies. Although
not explicitly analyzed here, my data suggests the interaction of positive PDO and negative
SOI phases represent the highest probability for outbreak development, consistent with the
findings of Stahl et al. (2006a) who showed that the most unsuitable winter temperatures for
MPB occurred during negative PDO and La Nina or neutral ENSO years.
I did not find any relationship between outbreak timing and the AO or PNA indices,
unlike a number of other studies which have linked these patterns to MPB either directly or
79

indirectly through their effect on winter minimum temperatures. Fauria and Johnson (2009)
found a low-frequency relationship between AO and total MPB area affected, which they
attribute to the influence o f AO on the probability of extreme cold events occurring. They
also found widespread correlations between winter minimum temperatures and total MPB
area affected, but were unable to establish a direct link between AO and winter minimums.
Thomson (2009), however, linked winter minimum temperatures to the AO through an
interaction with the PNA. Although I found no direct relationships either between outbreak
timing and winter temperature extremes (DJF minimum temperature or days below cold
mortality curves), there is little doubt that extreme cold events are important to inter-annual
MPB population dynamics (Aukema et al. 2008, Fauria and Johnson 2009) and that the AO
and PNA may be involved (Fauria and Johnson 2009, Thompson 2009). However, my
results indicate that, over longer time scales, extreme cold events are poor predictors of the
MPB population eruptions leading to the onset of outbreaks, probably because extreme
events are less predictable and less likely to exhibit regular trends that suppress population
buildup in this area.
The relationship between the standardized tree-ring indices and outbreak timing
varied depending on the frequency o f the signal tested, indicating a more complex
association between host stress and outbreak development than I observed for the climate
variables. Host stress does appear to be important over short time scales in the year an
outbreak develops, consistent with the results of Shrimpton and Thomson (1982) who also
observed reduced growth in the year the Chilcotin outbreak began which they later attributed
to extreme dry conditions (Thomson and Shrimpton 1984). In contrast Trzcinski and Reid
(2009) found no indication of low precipitation at the time of attack for the 1930s outbreak

80

in the Kootenay Valley, but observed a long term average precipitation deficit over the
preceding decade which they propose was responsible for increasing drought-induced host
susceptibility. Like Fauria and Johnson (2009), my results show no evidence of a spring
drought influence, long- or short-term, indicating that precipitation alone is not a sufficient
proxy for host stress in this study area. These contrasting results emphasize the importance
of understanding local climate-tree growth relationships when attempting to characterize
climatically-induced stress, which may be driven by more than one factor.
My climate-tree growth model showed that host stress is a function o f three climate
components. Although the strongest correlations are with precipitation and do indicate a
contributing role o f spring precipitation deficits in inducing host stress, the effects of winter
and spring temperature should not be discounted, particularly because these climate
parameters are also directly important to MPB population dynamics. The negative
relationship between high spring temperatures and tree growth suggest that this factor may
benefit MPB populations both directly through the benefits associated with increased brood
development rates shown to promote seasonally synchronized emergence (Reid 1962) and
indirectly as a driver of host stress. Conversely, the positive relationship between tree growth
and winter temperatures suggest that the climatic benefit of warm winters for MPB, namely
higher larval survival rates (Safranyik and Carroll 2006), may also increase host defensive
capacity (Shrimpton 1973) to the disadvantage o f attacking beetles, particularly if
temperatures drop below biological tolerances. Although more work is needed to determine
which combination of factors is most favorable for outbreak development, my data suggests
that the optimal relationship between climate, host and MPB involves a trade-off between
the benefits of warm winters for beetle populations and the possible defensive advantages

81

gained by the host, coupled with warm, dry spring conditions that clearly favor outbreak
development from a biological (MPB) and an environmental (host) perspective. Based on the
SEA results, none o f these factors on their own appear to drive outbreak development, at
least over short time scales when host stress is a significant component of the outbreak risk
model tested.
When examined over longer time scales (4-10 year growth trends) the reversal o f the
host tree growth-outbreak relationship I observed supports the hypothesis that sustained tree
vigor over multiple years may be as important for outbreak development as is short-term
stress-induced susceptibility. There are a number of advantages that prolonged periods of
above-average growth in the decade preceding an outbreak may provide in terms of habitat
suitability and food quality. Cole (1973) linked higher-than-average growth rates over 5 year
periods to increased phloem thickness which has been shown to enhance brood and beetle
size as well as developmental and emergence rates (Amman 1972, Amman and Pace 1976,
Amman and Cole 1983). Reid and Robb (1999) also found positive effects o f long-term
growth rate on brood productivity even when phloem thickness was unrelated to vigor,
emphasizing the importance of the nutritional quality o f the phloem over phloem quantity
alone. The advantages associated with larger broods and faster development may be
particularly important to MPB in northern environments that are subject to greater climatic
pressures by promoting seasonal synchrony and maximizing population growth (Safranyik
and Carroll 2006). This may be one mechanism by which low-frequency climate patterns
such as PDO and SOI influence the inter-annual population dynamics o f MPB (e.g. Aukema
et al. 2008, Fauria and Johnson 2009) by causing long-term persistent effects in the host tree
that encourage outbreak development.

82

3.5.2 Climate suitability

Annual variability around the threshold values for each of the climate suitability
parameters supports the SEA results and provides some additional information about which
climate factors are important in the study area. It is not surprising that summer and annual
degree day accumulations were unrelated to outbreak timing given that they were generally
well above the requirements defined by Carroll et al. (2004). A period o f unsuitable summer
heat accumulation during the time of the 1900s outbreak suggests that the lower rates o f egg
gallery construction, egg laying, and egg density associated with lower temperatures during
this time (Amman 1972a) did not limit brood production enough to prevent the outbreak
from occurring. This climate variable may not be very important when considered on its
own, particularly when paired with the suitable conditions across the other heat
accumulation parameters that control development rates and synchronize emergence and
mass attack. August maximum temperatures, despite some variability around the threshold,
do not drop below 18.3°C regularly enough to have much value for predicting when
outbreaks will occur. However, the extended period of low August temperatures in the mid1950s may have had some role in the collapse of the 1940s outbreak, suggesting this variable
may be more useful for predicting fine scale inter-annual MPB population dynamics as
shown by Aukema et al. (2008) than for predicting long-term outbreak cycles in my study
area. Precipitation was highly variable throughout the record, but only one of six distinct dry
spells was associated with the start of an outbreak, demonstrating why this variable is
unrelated on outbreak timing when considered on its own. A possible explanation for the
inconsistent relationships between precipitation and MPB outbreaks found by Thomson and
83

Shrimpton (1984), Trzcinski and Reid (2009), Fauria and Johnson (2009) and this study is
that variance around the mean does not reflect climatic differences between study areas,
particularly when precipitation is not the only factor limiting tree growth. For example,
temperature is often the dominant signal found in the tree ring record o f high elevation sites
in the southern interior (Wilson and Luckman 2002) and my mid-elevation site in the
northern interior responds to a combination o f temperature and precipitation, suggesting that
a climate-host stress model developed in the Chilcotin where precipitation is likely the most
limiting factor is not applicable across the entire MPB range.
The number of cold mortality days is the only climate parameter that showed a
consistent trend in the suitability analysis, with all outbreaks beginning in years with below
average or no cold mortality. Despite this relationship, years with little or no cold mortality
that do not correspond with outbreaks also occur fairly regularly in the record indicating that
these variables are not necessarily good predictors o f when the MPB population eruptions
will occur. As previously discussed, this does not discount the demonstrated influence that
extreme cold events can have on inter-annual population dynamics (e.g. Aukema et al. 2008,
Fauria and Johnson 2009). The prevalence of unsuitable days that occur during the
reconstructed outbreak periods additionally suggests that duration of cold spells may be
more important than the frequency of mortality events on MPB population dynamics. This is
not surprising given the various insulating factors associated with snow depth, tree diameter,
bark thickness and attack density that can delay significant MPB mortality for several
consecutive days of lethal air temperatures until subcortical temperatures drop below larval
tolerances (Safranyik and Carroll 2006). Although beyond the scope of this study,
calculating the number o f consecutive days below the mortality thresholds would be useful

84

to emphasize the effect of longer cold spells that are potentially the most damaging,
particularly in areas with large trees that have greater insulating effects.

3.6. Conclusions
Overall, my results support much of the existing knowledge about climate-MPB
outbreak relationships and offer some important new observations gleaned from the use of
long-term outbreak data and climate-tree growth models. Persistent positive winter
temperature anomalies and the large-scale patterns o f climate variability that drive them
(PDO and ENSO) appear to be the most important determinants o f when outbreaks will
occur. The occurrence of extreme cold events and the large-scale patterns associated with
them (AO and PNA) show no long-term predictive skill for assessing outbreak return. It is
important to note that the advantages o f using a long-term MPB outbreak reconstruction to
test these relationships over a time scale appropriate for low-frequency oscillations also
come with the disadvantage o f greatly reduced resolution which prevents us from examining
interannual climate-outbreak relationships known to be important to short-term MPB
population dynamics over the entire reconstruction period, such as the influence of extreme
cold events. Regardless, my results show that long-term outbreak risk assessments will
benefit most from incorporating PDO and ENSO predictions into the climate models.
I also showed that the relationship between climate, MPB and host stress is more
complicated than existing models recognize. First, the use of a universal climate proxy to
represent host stress, such as spring precipitation deficit, does not reflect regional variance in
the climate-tree growth response and cannot necessarily be simplified to a single variable.
85

For my study area it appears that a combination of climatic conditions drives the hostoutbreak relationship, emphasizing the importance of understanding the specific drivers of
host stress. Furthermore, while host stress is important in the short term, long-term host
vigor may be as important for enhancing habitat quality, especially in northern environments
where environmental pressures are greater. These findings have implications for improving
outbreak risk models through the use of regional climate-tree growth information and for
outbreak risk management with respect to understanding the influence o f long term forest
health on forest disturbance processes.

86

4. Mountain pine beetle outbreak dynamics and climate interactions in sub-boreal
mixed pine forests of north-central British Columbia: a synthesis
4.1 Introduction

The composition, structure and function o f forest ecosystems is largely dependent on
the forest disturbance processes that shape them, and for disturbance agents capable of
causing widespread damage and tree mortality, the social, economic, and ecological
consequences of these events can be severe. Many disturbances, however, are natural
components of the system that play an important role in the cycle of forest development and
regeneration (Logan and Powell 2005), so it is important to balance management objectives
with strategies that are consistent with the basic ecology of the forest ecosystem (Romme et
al. 2006). This requires a sound understanding of the disturbance processes and their
function. Because the interactions driving stand dynamics can extend for centuries, long­
term histories are necessary to establish the links between forest development and forest
disturbance. This is becoming increasingly important in light o f changing climate conditions
which have already influenced the timing, frequency, magnitude, severity and duration of
disturbances (Dale et al. 2001), greatly accelerating the effects of climate change on forest
composition and structure (Overpeck et al. 1990).
The objectives of this study were to develop a long-term record of mountain pine
beetle (MPB) (Dendroctonus ponderosae Hopkins (Curculionidae: Scolytinae)) outbreaks
near the northern limit of the species’ range and to examine the interacting factors driving
that disturbance regime. In British Columbia (BC) MPB is considered the most destructive
biotic disturbance agent of mature pine forests and in recent years has experienced an
outbreak that is unprecedented in size and severity relative to any historical outbreaks on
87

record. Although well-studied in terms of species biology and environmental tolerances,
very little information exists about long-term outbreak dynamics at the appropriate
ecological timescales. Conceptually, insect outbreaks result from the interactions between
three obligate causal factors: a host, a disturbance agent, and a suitable climate for insect
population growth (Figure 1.1). Over longer time scales, these interactions collectively
describe the outbreak dynamics of the disturbance regime driving forest development. Using
a combination of dendrochronological methods to reconstruct stand and outbreak history
together with long-term local and large-scale climate data, I developed an integrated model
of outbreak dynamics for north-central BC where little is known about the role of MPB in
forest stand dynamics or the influence o f climate on outbreak development. This study
addresses a number of gaps in existing outbreak dynamics research, particularly the need for
a greater understanding o f MPB in climatically sensitive regions near the northern range
limits where the effects o f climate change are expected to be most dramatic (Bentz et al.
2010 ).

In this chapter, I synthesize the information presented in Chapters 2 and 3 about the
individual components of outbreak dynamics in this region to discuss the system from a
comprehensive ecological perspective that addresses some of the causal links between the
outbreak regime, the host forest and the local and large-scale climate drivers. I discuss the
relevance o f my results with respect to similar research conducted in other areas and
consider the overall implications o f my results with respect to MPB management in BC,
emphasizing the potential impacts of climate change on this disturbance process. Lastly I
address some of the limitations of the study and future directions this research could take to
improve our understanding of MPB outbreak dynamics and our ability to manage, predict,

88

and mitigate the ecological and economic effects future epidemics may have in western
Canada.

4.2 MPB, host and climate interactions inferred from a multi-century outbreak
reconstruction

Between 1800 and 2007 I reconstructed six periods o f MPB disturbance occurring in
the 1830s, late 1860s-1870s, 1900-1910s, 1940s-1950s, 1980s and the epidemic that was
underway at the time of the study (Figure 2.5). The current outbreak was the most severe on
record, but the 1900-1910s and the 1940s-1950s events also fit the profile o f large epidemics
(Amman and Schrmitz 1988, Safranyik and Carroll 2006). The 1830s and 1870s events were
smaller, and while the 1980s saw high levels o f mortality, the infestations remained local
until the major population eruption in the 2000s. The reconstructed outbreak cycle of
approximately 30 years is fairly consistent with those reported elsewhere ranging between 20
and 53 years (Roe and Amman 1970, Amman and Schmitz 1988, Alfaro et al. 2004, Taylor
et al. 2006, Axelson et al. 2009). The timing of my outbreaks is also consistent with MPB
outbreak histories from other parts of western North America, especially between 1900 and
the present day when synchronous outbreaks occurred in the northern Rocky Mountains of
Wyoming and Idaho (Roe and Amman 1970, Amman and Schmitz 1988), Kootenay and
Banff National Parks (Taylor et al. 2006), the southern-interior o f BC (Axelson et al. 2009)
and the Chilcotin Plateau in west-central BC (Alfaro et al. 2004). This suggests that largescale processes, particularly those driving climate variability, can have a tremendous
influence on insect disturbance regimes despite vast regional differences in forest structure

89

and stand dynamics. The nature of the current MPB outbreak in BC is also only one o f
several unprecedented bark beetle outbreaks occurring throughout western North America,
including the recent spruce beetle outbreaks in Alaska (Berg et al. 2006, Sherriff et al. 2011),
and the commonality for all these events that began building in the late 1980s and early
1990s appears to be the increasing temperature trend that continues to present day (Logan
and Powell 2005). This probably explains the steady population buildup and tree mortality I
observed in my study area beginning in the 1980s (Figure 2.5) that almost certainly
contributed to the extremely high mortality continuing in the mid-2000s.
While these insights from paleoecology and current events reveal the disruptive
influence that climate change appears to be having on insect disturbance regimes at a
tremendously large scale, formulating a response that can lead to effective management
requires a more complete understanding of contributing mechanisms driving the disturbance
regime. My results for MPB outbreaks in north-central BC suggest that we may need to take
a more regional approach that incorporates forest-specific stand dynamics together with
climate interactions into management models. Chapter 2 showed that MPB and fire are the
two major forest disturbances driving forest structure and composition in my study area, and
the relationship between them largely influences when outbreaks occur. It may also be
important to recognize the strong presence o f the two-year cycle budworm in these stands.
Although it does not appear to be a major mortality agent nor directly influence hostavailability for MPB, the degree of synchronicity between budworm and MPB outbreaks
could indicate an important interaction that has not yet been investigated.
Successional development in these stands is typical for sub-boreal spruce forests.
Infrequent, large-scale stand-replacing fires operate over much longer time-scales than MPB,

90

establishing single-cohort serai lodgepole pine stands that gradually give way to shadetolerant spruce/fir stands at a rate governed by MPB outbreak frequency and severity. The
long fire-free interval in this study area has contributed to the nearly 200 year history of
MPB outbreaks dominating this system, beginning around the time the oldest stands reached
maturity and become progressively more destructive once landscape scale age-susceptibility
was reached, approximately 80 years after the most recent stand-initiation fire event (Raffa
and Berryman 1982). Changes in the fire regime, specifically an increased frequency of
stand-replacing events, whether through climate warming or management practices, will
have a profound effect on the spatial and temporal scales of MPB susceptibility in these
forests (Taylor and Carroll 2004). Similarly, because there is no pine regeneration after the
initial establishment event (Figure 2.3), changes in the frequency and severity of MPB
outbreaks will also influence susceptibility by controlling how long pine remains a
component of these stands. Since 1830, my sites experienced between 1 and 4 outbreak
events and, with the exception of two sites, all had living pine present in the stand at the time
of sampling, indicating a historically low-severity regime. MPB events do not appear to have
any influence on regeneration patterns of the other species present in the stand, particularly
subalpine fir which, in the absence of fire, will eventually dominate this landscape (Figure
2.4)
Characteristics about host susceptibility and pine mortality rates allowed me to
examine the influence that forest structure has on outbreak severity and frequency. Beetlekilled trees in my study area were larger and older (Figure 2.2) than those typically defined
for outbreak susceptibility (e.g., Safranyik et al. 1974, 1975, Cole and Amman 1980, Taylor
and Carroll 2004), which has dramatic implications for risk modelling and indicates some

91

important differences in outbreak dynamics between marginal environments and climatically
optimal regions. A preference for larger trees suggests that size-related susceptibility in the
north may be skewed by greater climatic pressures that inhibit beetle survival and attack
success without larger hosts that allow for larger broods, larger beetles, faster life cycles and
greater relative emergence rates (Safranyik and Jahren 1970, Amman and Cole 1983,
Safranyik and Carroll 2006). Higher age-related susceptibility thresholds also indicate that
provincial risk models based on age largely underestimate risk in the north-central interior.
Furthermore, the persistence of large diameter pine in most sites throughout multiple
outbreaks supports the interpretation of a historically low-severity outbreak regime that,
following stand maturation, is most likely limited by climate (Amman and Baker 1972)
rather than host availability. This theory directly contrasts with the host-limiting models used
to explain outbreak severity and cycles further south (Cole and Amman 1980, Alfaro et al.
2004). Also, the continuous supply o f endemic populations observed between outbreak
periods (Figure 2.5d), combined with a higher survival rate o f large trees during outbreaks,
likely enhances the stability of these populations between outbreak phases. This has major
implications for climate change because MPB populations in my study area may be able to
respond more quickly to increasingly favorable climatic conditions and experience rapid
regime shifts similar to those proposed by Bale et al. (2002) for other insect herbivores.
In Chapter 3 ,1 explored specific local and large-scale climate interactions associated
with the MPB outbreaks reconstructed in Chapter 2 in order to determine which climate
factors were driving outbreak development and frequency. The links I observed between
climate, MPB outbreaks, and host tree growth emphasized the importance o f long-term
patterns of climate variability in the development of large infestations. Persistent positive

92

winter temperature anomalies were the only significant direct climatic driver, which is not
surprising considering that population buildup over multiple years associated with higher
winter survival rates would greatly support epidemic development. Both large-scale climate
patterns with teleconnections to winter temperatures in BC were important predictors of
outbreak development. Outbreaks occurred consistently following multi-year trends of
positive PDO index values and negative SOI values or El Nino events, which is consistent
with the results of other studies that looked at the relationship between MPB and these
indices (Stahl et al. 2006a, Fauria and Johnson 2009). I did not find any significant
relationships between my outbreak periods and the other climatic variables known to directly
affect MPB populations. This can be explained in part because regional summer heat
accumulation was generally well within the limits for successful MPB development and
seasonal synchrony (Figure 3.4), and precipitation, although quite variable and including
several extended periods of drought, did not seem to cause a strong enough host stress
response necessary to initiate an epidemic. The lack o f findings for an effect o f cold
mortality events based on winter minimum temperatures I attribute to the fact that extreme
events, while of known importance to inter-annual MPB population dynamics (Aukema et al.
2008, Fauria and Johnson 2009), are not useful predictors of when MPB population
eruptions will occur, in part because they are themselves less predictable. This explanation,
which I address later in relation to limitations of this study, may also apply to the other
variables for which I found no significant effect because the climatic drivers of outbreak
development may be very different than the climate relationships than influence MPB
population dynamics during a particular outbreak event.

93

The examination of indirect climatic effects mediated through the host trees provided
some useful information with respect to the use o f a host-stress variable in climate-suitability
models. Host stress does appear to be an important precursor to outbreak development,
however, differences in geography and species responses to climatic variability dictate that a
universal climatic proxy such as precipitation is not appropriate for modeling host stress
between different locations. My climate-tree growth relationships indicated that a
combination o f climatic variables was driving the host stress response associated with
outbreaks. This highlights the importance of understanding climate-host relationships for a
given area, and explains much o f the inconsistency in the literature as to the effects of
precipitation on outbreak development (e.g. Thomson and Shrimpton 1984, Trzcinski and
Reid 2009, Fauria and Johnson 2009). I also found that over longer time periods,
climatically-induced and prolonged host vigor was significantly related to outbreak
development. For marginal environments subject to greater climatic pressures, the
advantages associated with sustained vigor in host trees such as increased brood
productivity, larger beetles, and faster development and emergence rates (Amman 1972,
Amman and Pace 1976, Amman and Cole 1983) may be particularly important to MPB
success.
The results of the climate interactions study have some important implications for
MPB management in BC, particularly in relation to climate change. In terms of climate
suitability modelling, host stress and vigor is important over both short and longer time
scales, but the climatic proxy used to represent this variable should be regionally appropriate
and may require some climate-tree growth analyses before model development. For outbreak
risk models, forest structural assessments should be integrated into climate models to

94

account for the varied influence that host susceptibility has on outbreak cycles which will
likely differ across climatic zones and forest types. Still, the degree o f synchrony between
outbreaks across western North America indicates that, given available hosts, large-scale
climate patterns are the most important driver of broad regional and possibly sub-continental
scale outbreak dynamics. Fortunately, these large-scale climate modes such as PDO and
ENSO are also becoming increasingly predictable, creating the potential for improving
ecologically-relevant climate-related risk predictions (Stenseth et al. 2003). Unfortunately,
climate change will likely influence every interaction represented by the outbreak dynamics
triangle model (Figure 1.1), either directly through changes in climate suitability for the
MPB or indirectly through effects on the host and the host forest stand dynamics. Managing
for these changes will require continued research to define these interactions, particularly at
species range margins where improved understanding of these dynamics may help to
mitigate the effects of range expansion and regime changes.

4.3 Limitations and future directions

As with many ecological studies, long-term data is essential to understanding
processes that occur over centuries or sometimes even millennia. In this study I
reconstructed over 200 years o f forest stand dynamics to develop a MPB outbreak
reconstruction that served two purposes: ( 1) to lengthen the survey records to an appropriate
time scale for studying outbreak dynamics and (2) to examine the interacting components o f
this outbreak regime in a system that has not been well-studied, but as a northern range limit
habitat is extremely susceptible to climatic changes and will likely become a focus of future

95

research as regime and range changes occur. A major limitation to any paleoecological
reconstruction is that by relying on proxy data to extend a record of events without direct
evidence, in my case MPB-killed trees that could be used to precisely date the timing of
mortality, you necessarily introduce some degree of uncertainty to the study. To limit this,
my approach in the reconstruction was to combine as many lines o f evidence as possible to
produce an overlapping record of supporting information that allows for a reasonably
undisputable interpretation of outbreak history. I am confident that the methods I used
produced as accurate a record as possible and through cautious interpretation of the results I
was able to use the reconstruction for its designated purpose.
A major difficulty in the outbreak reconstruction was the strong and persistent
presence o f the two-year cycle budworm signal in the non-host tree-ring record that seemed
to occur in very close succession with the MPB release signal in the same chronologies.
Although I was successful in decoupling the signals using methods explained in Chapter 2,
this co-occurrence of disturbances affecting different hosts creates questions about possible
interactions between agents that have not been addressed yet. While I suspect climatic
similarities explain some of this relationship, there are questions about ecological
interactions between agents that arise from the host vigor response preceding MPB
outbreaks. It is possible that defoliation caused by the two-year cycle budworm in mixed
pine, spruce and fir forests provides a nutrient pulse that may contribute to the increase in
pine vigor contributing to MPB population growth. This is certainly an interesting question
worth pursing in future research.
In Chapter 3, limitations of the reconstruction that could only provide dates for the
initial population eruption associated with the start of an outbreak prevented me from

96

examining climatic effects associated with inter-annual population dynamics, particularly
those that influence outbreak spread and severity during an outbreak. These have been wellstudied with the available aerial survey records dating back to 1959 and, as discussed, have
confirmed the importance of extreme cold events and in some cases the effect of summer
temperature on spatial patterns (e.g. Aukema et al. 2008, Fauria and Johnson 2009). It would
be interesting to examine the consistency of this effect over multiple outbreaks spanning a
longer time period, particularly with respect to any control these events may have on
outbreak duration. To do so would require a much more detailed profile o f tree mortality
than I had available in my study area because of the decay rates of downed trees. It may be
possible in drier climates.
Lastly, a major component o f outbreak dynamics I excluded from this study was
post-outbreak recovery which is a major concern considering the magnitude o f the current
outbreak and its effect on the future timber supply. Little is known about natural recovery
following MPB outbreaks or what the implications for management are compared to more
conventional methods of salvage logging and replanting (Griesbauer and Green 2006),
especially over a longer time periods. My study forms the basis for a future study examining
natural post-outbreak recovery by providing a time-series of outbreak events from which a
chronosequence o f regeneration and successional processes based on different outbreak dates
could be analyzed. This would be especially interesting because for mixed-species stands
that are unlikely to experience recruitment of new pine trees (Figure 2.2), understanding
advanced regeneration and the ongoing establishment of shade-tolerant vegetation could be
useful for supplementing mid-term timber supplies in the absence of pine.

97

98

Literature Cited

Agrios, G. N. 2005. Plant Pathology (5th edition). Elsevier-Academic Press. San Diego,
California, USA.
Alfaro, R. I., R. Campbell, P. Vera, B. Hawkes, and T. L Shore. 2004. Dendroecological
reconstruction of mountain pine beetle outbreaks in the Chilcotin Plateau of British
Columbia. Pages 245-256 in T.L Shore, J. E. Brooks, and J. E. Stone, editors.
Mountain Pine Beetle Symposium: Challenges and Solutions, Kelowna, British
Columbia, October 30-31, 2003. Natural Resources Canada, Canadian Forest
Service, Pacific Forestry Centre, Victoria, BC. Information Report BC-X-399.
Amman, G. D. 1972a. Some factors affecting oviposition behavior o f the mountain pine
beetle. Environmental Entomology 1:691-695.
Amman, G. D. 1972b. Mountain pine beetle brood production in relation to thickness of
lodgepole pine phloem. Journal o f Economic Entomology 65:138-140.
Amman, G. D. 1973. Population changes of the mountain pine beetle in relation to elevation.
Environmental Entomology 2:541-547.
Amman, G. D., and B. H. Baker. 1972. Mountain pine beetle influence on lodgepole pine
stand structure. Journal of Forestry 70:204-209.
Amman, G. D., and W. E. Cole. 1983. Mountain pine beetle dynamics in lodgepole pine
forests. Part II: Population dynamics. General Technical Report INT-145. USDA
Forest Service, Intermountain Forest and Range Experiment Station, Ogden, Utah,
USA.
Amman, G. D., and V. E. Pace. 1976. Optimum egg gallery densities for the mountain pine
beetle in relation to lodgepole pine phloem thickness. Research Note INT-209.
USDA Forest Service, Intermountain Forest and Range Experiment Station, Ogden,
UT, USA.
Amman, G. D., and R. F. Schmitz. 1988. Mountain pine beetle-lodgepole pine interactions
and strategies for reducing tree losses. Ambio 17: 62-68.
Aukema, B. H., A. L. Carroll, Y. Zheng, J. Zhu, K. F. Raffa, R. D. Moore, K. Stahl, and S.
W. Taylor. 2008. Movement of outbreak populations of mountain pine beetle:
influences of spatiotemporal patterns and climate. Ecography 31:348-358.
Axelson, J. N., R. I. Alfaro, and B. C. Hawkes. 2009. Influence of fire and mountain pine
beetle on the dynamics of lodgepole pine stands in British Columbia, Canada. Forest
Ecology and Management 257: 1874-1882.
Axelson, J. N., R. I. Alfaro, and B. C. Hawkes. 2010. Changes in stand structure in unevenaged lodgepole pine stands impacted by mountain pine beetle epidemics and fires in
central British Columbia. The Forestry Chronicle 86:87-99.
99

Baisan, C H. and T. W. Swetnam. 1990. Fire history on a desert mountain range: Rincon
Mountain Wilderness, Arizona, U.S.A. Canadian Journal of Forest Research
20:1559-1569.
Bale, J. S., G. J. Masters, I. D. Hodkinson, C. Awmack, T. M. Bezemer, V. K. Brown, J.
Butterfield, A. Buse, J. C. Coulson, J. Farrar, J. E. G. Good, R. Harrington, S.
Gartley, T. H. Jones, R. L. Lindroth, M. C. Press, I. Symmioudis, A. D. Watt, and J.
B. Whittaker. Herbivory in global climate change research: direct effects o f rising
temperature on insect herbivores. 2002. Global Change Biology 8:1-16.
Battisti, A., M. Stastny, S. Netherer, C. Robinet, A. Schopf, A. Roques, and S. Larsson.
2005. Expansion o f geographic range in the pine precessionary moth caused by
increased winter temperatures. Ecological Applications 15:2084-2096.
BC Ministry of Forests. 1995. 1994 forest, recreation, and range resource analysis. British
Columbia Ministry of Forests, Public Affairs Branch, Victoria, BC. 308p.
BC Ministry o f Forests, Lands and Natural Resource Operations. 2011. Facts about B.C.’s
mountain pine beetle, http://www.for.gov.bc.ca/hfp/mountainjpine_beetle/facts.htm
Bentz, B. J., J. Regniere, C. J. Fettif, E. M. Hansen, J. L. Hayes, J. A. Hicke, R. G. Kelsey, J.
F. Negron, and S. J. Seybold. 2010. Climate change and bark beetles of the western
United States and Canada: direct and indirect effects. Bioscience 60:602-613.
Berg, E. E., J. D. Henry, C. L Fastie, A. D. De Voider, and S. M. Matsuoka. 2006. Spruce
beetle outbreaks on the Kenai Peninsula, Alaska, and Kluane National Park and
Reserve, Yukon Territory: relationship to summer temperatures and regional
differences in disturbance regimes. Forest Ecology and Management 227: 219-232.
Berryman, A. A. 1972. Resistance o f conifers to invasion by bark beetle-fungal associations.
BioScience 22:598-602.
Biondi, F., and K. Waikul. 2004. DENDROCLIM2002: A C++ program for statistical
calibration of climate signals in tree-ring chronologies. Computers & Geosciences
30:303-311.
Biasing, T. J., A. M. Solomon, and D. N. Duvick. 1984. Response functions revisited. Treering Bulletin 44:1-15.
Bloomberg, W. J., and D. J. Morrison. 1989. Relationship of growth reduction in Douglas-fir
to infection by Armillaria root disease in southeastern British Columbia.
Phytopathology 79:482-487.
Bonsai, B. R., X. Zhang, A. Vincent, and W. D. Hogg. 2001. Characteristics of daily and
extreme temperatures over Canada. Journal o f Climate 14:1959-1976.
Bradshaw, W. E., and C. M. Holzapfel. 2006. Evolutionary response to rapid climate change.
Science:312:1477-1478.

100

Brown, P. M., and T. W. Swetnam. 1994. A cross-dated fire history from coast redwood near
Redwood National Park, California. Canadian Journal of Forest Research 24:21-31.
Campbell, E.M., R. I. Alfaro, and B. Hawkes. 2007. Spatial distribution of mountain pine
beetle outbreaks in relation to climate and stand characteristics: A dendroecological
analysis. Journal of Integrative Plant Biology 49:168-178.
Campbell, R., D.J. Smith, and A. Arsenault. 2006. Multicentruy history of western spruce
budworm outbreaks in interior Douglas-fir forests near Kamloops, British Columbia.
Canadian Journal of Forest Research 36:1758-1769.
Carroll, A. L., S. W. Taylor, J. Regniere, and L. Safranyik. 2004. Effects of climate change
on range expansion by the mountain pine beetle in British Columbia. Pages 223-232
in T.L Shore, J. E. Brooks, and J. E. Stone, editors. Mountain Pine Beetle
Symposium: Challenges and Solutions, Kelowna, British Columbia, October 30-31,
2003. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre,
Victoria, BC. Information Report BC-X-399.
Chan-McLeod A. C. A. 2006. A review and synthesis of the effects of unsalvaged mountainpine-beetle-attacked stands on wildlife and implications for forest management. BC
Journal of Ecosystems and Management 7:119-132
Christensen, N. L., A. M. Bartuska, J. H. Brown, S. Carpenter, C. D’Antonio, R. Francis, J.
F. Franklin, J. A. MacMahon, R. F. Noss, D. J. Parsons, C. H. Peterson, M. G.
Turner, and R. G. Woodmansee. 1996. The report of the ecological society of
America committee on the scientific basis for ecosystem management. Ecological
Applications 6:665-691.
Cole, W. E. 1973. Interaction between mountain pine beetle and dynamics o f lodgepole pine
stands. Research Note INT-170. USDA Forest Service, Intermountain Forest and
Range Experiment Station, Ogden, Utah, USA.
Cole, W. E. 1981. Some risks and causes of mortality in mountain pine beetle populations: a
long-term analysis. Researches on Population Ecology 23:116-144.
Cole, W. E., and G. D. Amman. 1980. Mountain pine beetle dynamics in lodgepole pine
forests. Part I: Course of an infestation. General Technical Report INT-89. USDA
Forest Service, Intermountain Forest and Range Experiment Station, Ogden, Utah,
USA.
Cook E. R. 1985. A time series analysis approach to tree-ring standardization. Dissertation,
University o f Arizona.
Cook, E. R., and R. L. Holmes. 1986. User’s manual for program ARSTAN. Pages 50-65 in
R. L. Holmes, R. K. Adams, and H.C. Fritts, editors. Tree-ring chronologies of
western North America: California, eastern Oregon and northern Great Basin.
Laboratory of Tree Ring Research, The University of Arizona, USA.
101

Cook, E. R., and L. A. Kairiukstis. 1992. Methods o f dendrochronology: applications in the
environmental sciences. Klewer Academic Publishers, Dordrecht, The Netherlands.
Cullingham, C. I., J. E. K. Cooke, S. Dang, C. S. Davis, B. J. Cooke, and D. W. Coltman.
2011. Mountain pine beetle host-range expansion threatens the boreal forest.
Molecular Ecology 20:2157-2171.
Dale, V. H., L. A. Joyce, S. McNulty, R. P. Neilson, M. P. Ayres, M. D. Flannigan, P. J.
Hanson, L. C. Irland, A. E. Lugo, C. J. Pererson, D. Simberloff, F. J. Swanson, F. J.
Stocks, and B. M. Wotton. Climate change and forest disturbances. 2001. Bioscience
51:723-734.
Davi, N. K., G. C. Jacoby, and G. C. Wiles. 2003. Boreal temperature variability inferred
from maximum latewood density and tree-ring width data, Wrangell Mountain
region, Alaska. Quaternary Research 60:252-262.
Delong, S.C. 1998. Natural disturbance rate and patch size distribution of forests in northern
British Columbia: Implications for forest management. Northwest Science 72:35-48.
Diaz, S. C., R. Touchan, and T. W. Swetnam. 2001. A tree-ring reconstruction o f past
precipitation for Baja California Sur, Mexico. International Journal of Climatology
21:1007-1019.
Douglas, A. E. 1941. Crossdating in dendrochronology. Journal of Forestry 39:825-831.
Environment Canada. 2010. Canadian Climate Normals (1971-2000). Fort St. James, British
Columbia, <http://www.climate.weatheroffice.gc.ca/climate_-normals/index_e.html>
Environment Canada 2011. Adjusted and Homogenized Canadian Climate Data (AHCCD).
Climate Research Branch, Meteorological Service, Canada, <http://ec.gc.ca/dcchaahccd/>.
Famden, C. 1996. Stand density management diagrams for lodgepole pine, white spruce and
interior Douglas-fir. Natural Resources Canada, Canadian Forest Service, Pacific
Forestry Centre, Victoria, BC. Information Report BC-X-360. 41 p.
Fauria, M. M., and E. A. Johnson. 2009. Large-scale climatic patterns and area affected by
mountain pine beetle in British Columbia, Canada. Journal o f Geophysical Research
114:G01012.
Fumiss, R. L., and V. M. Carolin. 1977. Western forest insects. Miscellaneous Publication
1339. USDA Forest Service.
Gibson, K., K. Skov, S. Kegley, C. Jorgensen, S. Smith, and J. Witcosky. 2008. Mountain
pine beetle impacts in high-elevation five-needle pines: current trends and
challenges. Report R1-08-020. USDA Forest Service, Forest Health Protection,
Missoula, Montana, USA.

102

Girardin, M., J. Tardif, and Y. Bergeron. 2001. Radial growth analysis of Larix laricina from
the Lake Duparquet area, Quebec, in relation to climate and larch sawfly outbreaks.
Ecoscience 8:127-138.

Grissino-Mayer, H. D., W. H. Romme, M. L. Floyd, and D. D. Hanna. 2004. Climatic and
human influences on fire regimes of the southern San Juan Mountains, Colorado,
USA. Ecology 85:1708-1724.
Grissino-Mayer, H. D., and T. W. Swetnam. 2000. Century-scale climate forcing of fire
regimes in the American Southwest. The Holocene 10:213-220.
Hadley, K. S., and T. T. Veblen. 1993. Stand response to western spruce budworm and
Douglas-fir bark beetle outbreaks, Colorado front range. Canadian Journal of Forest
Research 23:479-491.
He, H. S., D. J. Mladenoff, and T. R. Crow. 1999. Linking an ecosystem model and a
landscape model to study forest species response to climate warming. Ecological
Modelling 114:213-233.
Heath, R., and R. Alfaro. 1990. Growth response o f a Douglas-fir/lodgepole pine stand after
thinning o f lodgepole pine by the mountain pine beetle. Journal of the Entomological
Society of British Columbia 87:16-21.
Hessl, A. E., D. McKenzie, and R. Schellhaas. 2004. Drought and Pacific Decadal
Oscillation linked to fire occurrence in the inland Pacific Northwest. Ecological
Applications 14:425-442.
Heyerdahl, E. K., L. B. Brubaker, and J. K. Agee. 2002. Annual and decadal climate forcing
of historical fire regimes in the interior Pacific Northwest, USA. The Holocene
12:597-604.
Holmes, R. 1983. Computer assisted quality control in tree-ring dating and measurment.
Tree-Ring Bulletin 43:69-78.
Holmes, R. L., and T. W. Swetnam. 1994. Program EVENT user’s manual. Laboratory of
Tree-Ring Research, University of Arizona, Tucson, Arizona USA.
Hrinkevich, K., and K. Lewis. 2011. Northern range limit mountain pine beetle outbreak
dynamics in mixed sub-boral pine forests of British Columbia. Ecosphere 10:116.
Hurrell, J. W. 1995. Decadal trends in the North Atlantic Oscillation: regional temperatures
and precipitation. 269:676-679.
Jacoby, G. C., and R. D’Arrigo. 1989. Reconstructed northern hemisphere annual
temperature since 1671 based on high-latitude tree-ring data from North America.
Climatic Change 14:39-59.

103

Jenkins, M. J., E. Hebertson, W. Page, and C. A. Jorgensen. 2008. Bark beetles, fuels, fires
and implications for forest management in the Intermountain West. Forest Ecology
and Management 254:16-34.
Kurz, W. A., C. C. Dymond, G. Stinson, G. J. Rampley, E. T. Neilson, A. L. Carroll, T.
Ebata, and L. Safranyik. 2008. Mountain pine beetle and forest carbon feedback to
climate change. Nature 452:987-990.
Logan, J. A., P. V. Bolstad, B. J. Bentz, and D. L. Perkins. 1995. Assessing the effects of
changing climate on mountain pine beetle dynamics. Pages 92-105 in R. W. Tinus,
ed. Proceedings o f the Interior west global change workshop, April 25-27, 1995, Fort
Collins, Colorado, USDA Forest Service, Rocky Mountain Forest and Range
Experiment Station, General Technical Report RM-GTR-262.
Logan, J. A., and J. A. Powell. 2001. Global warming and the mountain pine beetle
(Coleoptera: Scolytidea). American Entomologist 47:160-172.
Logan, J. A., and J. A. Powell. 2005. Ecological consequences of climate change altered
forest insect disturbance regimes, in F. H. Wagner, editor. Climate change in western
North America: evidence and environmental effects. Allen Press.
<http://www.usu.edu/beetle/documents/Logan-Powell2005.pdf>
Logan, J. A., J. Regniere, and J. A. Powell. 2003. Assessing the impacts o f global climate
change on forest pests. Frontiers in Ecology and the Environment 1:130-137.
Loehle, C., and D. LeBlanc. 1996. Model-based assessments of climate change effectds on
forests: a critical review. Ecological Modelling 90:1 -31.
Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis. 1997. A pacific
interdecadal climate oscillation with impacts on salmon production. Bulletin of the
American Meteorological Society 78:1069-1079.
Mattson, W. J. 1996. Escalating anthropogenic stresses on forest ecosystems: forcing benign
plant-insect interactions into new interaction trajectories. Pages 338-342 in E.
Korpilahti, H. Mikkela, and T. Salonen, editors. Caring for the forest: research in a
changing world. Congress Report Vol. 2. IUFRO World Congress Organizing
Committee. Finland.
McCambridge, F.W. 1971. Temperature limits of flight o f the mountain pine beetle,
Dendroctonus ponderosae. Annals of the Entomological Society o f America
64:534-535.
McGregor, M. D. and D. M. Cole 1985. Integrating management strategies for the mountain
pine beetle with multiple-resource management o f lodgepole pine forests. General
Technical Report INT-174. USDA Forest Service, Intermountain Forest and Range
Experiment Station, Ogden, Utah, USA.

104

Meidinger, D. and J. Pojar (Eds.). 1991. Ecosystems o f British Columbia. Special Report
Series No. 6. British Columbia Ministry o f Forests, Research Branch, Victoria, BC,
Canada.
Mooney, C. Z., and R. D. Duval. 1993. Bootstrapping: A nonparametric approach to
statistical inference. Sage Publications, Newbury Park, California, USA.
Moore, R. D., I. G. McKendry, and K. Stahl. 2005. Mountain pine beetle outbreaks in
western Canada: coupled influences o f climate variability and stand development.
Final Report for Climate Change Action Fund Project A676, Natural Resources
Canada, Ottawa, ON.
Nola, P., M. Morales, R. Motta, and R. Villalba. 2006. The role o f larch budmoth
(Zeiraphera diniana Gn.) on forest succession in a larch (Larix decidua Mill.) and
Swiss stone pine (Pinus cembra L.) stand in the Susa Valley (Piedmont, Italy). Trees
20:371-382.
Oliver, C. D., and B. C. Larson. 1996. Forest Stand Dynamics. John Wiley and Sons, Inc.
New York, USA.
Overpeck, J. T., D. Rind, and R. Goldberg. 1990. Climate-induced changes in forest
disturbance and vegetation. Nature 343:51-53.
Parker, T. J., K. M. Clancy, and R. L. Mathiasen. 2006. Interactions among fire, insects and
pathogens in coniferous forests o f the interior western United States and Canada.
Agricultural and Forest Entomology 8:167-189.
Parkins, J. R., and N. A. McKendrick. 2007. Assessing community vulnerability: a study of
the mountain pine beetle outbreak in British Columbia, Canada. Global
Environmental Change 17:460-471.
Pohl, K. A., K. S. Hadley, and K. B. Arabas. 2006. Decoupling tree-ring signatures of
climate variation, fire, and insect outbreaks in central Oregon. Tree-Ring Research
62:37-50.
Pureswaran, D.S. and J. H. Borden. 2003. Test of semiochemical mediated host
specificity in four species o f tree killing bark beetles (Coleoptera: Scolytidae).
Environmental Entomology 32:963-969.
Raffa, K. F., and A. A. Berryman. 1982. Physiological differences between lodgepole pines
resistant and susceptible to the mountain pine beetle and associated microorganisms.
Environmental Entomology 11:486-492.
Raffa, K.F., Aukema, B.H., Bentz, B.J., Carroll, A.L., Hicke, J.A., Turner, M.G., and
Romme, W.H. 2008. Cross-scale drivers of natural disturbances prone to
anthropogenic amplification: the dynamics of bark beetle eruptions. BioScience.
58:501-517.

105

Redding, T., R. Winkler, D. Spittlehouse, R. D. Morre, A. Wei, and P. Teti. 2008. Mountain
pine beetle and watershed hydrology: a synthesis focused on the Okanagan basin.
One Watershed - One Water Conference Proceedings. Kelowna, October 21-23,
2008.
Redmond, K. T., and R. W. Koch. 1991. Surface climate and streamflow variability in the
western United States and their relationship to large-scale circulation indices. Water
Resources Research 27:2381-2399.
Reid, R. W. 1958. Internal changes in the female mountain pine beetle, Dendroctonus
monticolae Hopk., associated with egg laying and flight,. The Canadian Entomologist
90:464-468.
Reid, R. W. 1962. Biology of the mountain pine beetle, Dendroctonus monticolae Hopkins,
in the East Kootenay region of British Columbia I. Life cycle, brood development,
and flight periods. The Canadian Entomologist 94:531-538.
Reid, R. W. 1963. Biology of the mountain pine beetle, Dendroctonus monticolae Hopkins,
in the East Kootenay region of British Columbia III. Interaction between the beetle
and its host, with emphasis on brood mortality and survival. The Canadian
Entomologist 95:225-238.
Reid, M. L., and T. Robb. 1999. Death of vigorous trees benefits bark beetles. Oecologia
120:555-562.
Rogers, P. 1996. Disturbane ecology and forest management: a review o f the literature.
General Technical Report INT-GTR-336. USDA Forest Service, Intermountain
Research Station, Ogden, Utah, USA.
Romme, W. H., J. Clement, J. Hicke, D. Kulakowski, L. H. MacDonald, T. L. Schoennagel,
and T. T. Veblen. 2006. Recent forest insect outbreaks and fire risk in Colorado
forests: A brief synthesis of relevant research. Colorado Forest Restoration Institute,
Fort Collins, Colorado, USA
Rose, M. R., and R. Harmsen. 1981. Ecological outbreak dynamics and the cusp catastrophe.
Acta Biotheoretica 30:229-253.
Royama, T. 1992. Analytical population dynamics. Chapman and Hall, New York, New
York, USA.
Rubino, D. L., and B. C. McCarthy. 2004. Comparative analysis of dendroecological
methods used to assess disturbance events. Dendrochronologia 21:97-115.
Safranyik, L., and A. L. Carroll. 2006. The biology and epidemiology of the mountain pine
beetle in lodgepole pine forests. Pages 3-66 in L. Safranyik and B. Wilson, editors.
The mountain pine beetle: a synthesis o f biology, management, and impacts on
lodgepole pine. Natural Resources Canada, Canadian Forest Service, Pacific Forestry
Centre, Victoria, BC, Canada.
106

Safranyik, L., and R. Jahren. 1970. Emergence patterns of the mountain pine beetle from
lodgepole pine. Canadian Forest Service, Bi-monthly Research Notes 26:11-19.
Safranyik, L., D. A. Linton, R. Silversides, and L. H. McMullen. 1992. Dispersal o f released
mountain pine beetles under the canopy of a mature lodepole pine stand. Journal of
Applied Entomology 113:441-450.

Safranyik, L., D. M. Shrimpton, and H. S. Whitney. 1974. Management o f lodgepole pine to
reduce losses from the mountain pine beetle. Forestry Technical Report No. 1.
Environment Canada, Canadian Forest Service, Pacific Forest Research Centre,
Victoria, BC.
Safranyik, L., D. M. Shrimpton,, H. S. Whitney, and D. M. Baumgartner. 1975. An
interpretation of the interaction between lodgepole pine, the mountain pine beetle and
its associated blue stain fungi in western Canada. Pages 406-428 in D. M.
Baumgartner, editor. Management of lodgepole pine ecosystems. Washington State
University Cooperative Extension Service, Pullman, WA.
Shabbar, A., and B. Bonsai. 2004. Associations between low frequency variability modes
and winter temperature extremes in Canada. Atmosphere-Ocean 42:127-140.
Shabbar, A., B. Bonsai, and M. Khandekar. 1997. Canadian precipitation patterns associated
with the Southern Oscillation. Journal o f Climate 10:3016-3027.
Shabbar, A., and M. Khandekar. 1996. The impact of El Nino-Southern Oscillation on the
temperature field over Canada. Atmosphere-Ocean:34-401-416.
Sherriff, R. L., E. E. Berg, and A. E. Miller. 2011. Climate variability and spruce beetle
(Dendroctonus rufipennis) outbreaks in sourth-central and southwest Alaska.
Ecology 92:1459-1470.
Shore, T. L., and L. Safranyik. 1992. Susceptibility and risk rating systems for the mountain
pine beetle in lodgepole pine stands. Information Report BC-X-336. Forestry
Canada, Pacific and Yukon Region.
Shrimpton, D. M. 1973. Age- and size-related response of lodgepole pine to inoculation with
Europhium clavigerum. Canadian Journal of Botany 51:1155-1160.
Shrimpton, D. M., and A. J. Thomson. 1982. Growth characteristics of lodgepole pine
associated with the start o f mountain pine beetle outbreaks. Canadian Journal of
Forest Research 13:137-144.
Speer, J. H. 2010. Fundamentals o f tree-ring research. The University of Arizona Press,
Tucson, Arizona, USA.
Speer, J. H., T. W. Swetnam, B. E. Wickman, and A. Youngblood. 2001. Changes in
Pandora Moth Outbreak Dynamics during the Past 622 Years. Ecology 82:679-697.
107

Stahl, K., R. D. Moore, and I. G. McKendry. 2006a. Climatology of winter cold spells in
relation to mountain pine beetle mortality in British Columbia, Canada. Climate
Research 32:13-23.
Stahl, K., R. D. Moore, and I. G. McKendry. 2006b. The role of synoptic-scale circulation in
the linkage between large-scale ocean-atmosphere indices and winter surface climate
in British Columbia, Canada. International Journal of Climatology 26:541-560.
Stenseth, N. C., G. Ottersen, J. W. Hurrell, A. Mysterud, M. Lima, K. Chan, N. G. Yoccoz,
and B. Adlandsvik. 2003. Studying climate effects on ecology through the use of
climate indicves: the North Atlantic Oscillation, El Nino Southern Oscillation and
beyond. Proceedings o f the Royal Society of London Biological Sciences 270:20872096.
Stevens, R.B. 1960. Plant Pathology, an Advanced Treatise, Vol. 3. J.G. Horsfall and A.E.
Dimond, editors. Academic Press, New York, USA.
Stephens, S. L., C. N. Skinner, and S. J. Gill. Dendrochronology-based fire history of Jeffrey
pine-mixed conifer forests in the Sierra San Pedro Martir, Mexico. 2003. Canadian
Journal of Forest Research 33:1090-1101.
Steventon, J.D. 2001. Historic disturbance regimes of the Morice and Lakes Timber Supply
Areas. Northern Interior Forest Region, B.C. Ministry of Forests, Smithers, B.C.
Draft Research Working Paper.
Stokes, M. A., and T. L. Smiley. 1968. An introduction to tree-ring dating. University of
Arizona Press, Tucson, Arizona, USA.
Stuart, J. D., J. K. Agee, and R. I. Gara. 1989. Lodgepole pine regeneration in an old, selfperpetuating forest in south central Oregon. Canadian Journal o f Forest Research
19:1096-1104.
Swetnam, T. W. 1993. Fire history and climate change in giant sequoia groves. Science
262:885-889.
Swetnam, T. W., and A. M. Lynch. 1993. Multicentury, regional-scale patterns o f western
spruce budworm outbreaks. Ecological Monographs 63:399-424.
Taylor, S. W, and A. L. Carroll. 2004. Disturbance, forest age, and mountain pine beetle
outbreak dynamics in BC: A historical perspective. Pages 41-51 in T.L Shore, J. E.
Brooks, and J. E. Stone, editors. Mountain Pine Beetle Symposium: Challenges and
Solutions, Kelowna, British Columbia, October 30-31, 2003. Natural Resources
Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, BC. Information
Report BC-X-399.
Taylor, S. W., A. L. Carroll, R. I. Alfaro, and L. Safranyik. 2006. Forest, climate and
mountain pine beetle outbreak dynamics in western Canada. Pages 67-94 in L.
Safranyik and B. Wilson, editors. The mountain pine beetle: a synthesis of biology,

108

management, and impacts on lodgepole pine. Natural Resources Canada, Canadian
Forest Service, Pacific Forestry Centre, Victoria, BC, Canada.
Thomson, A. J. 2009. Climate indices and mountain pine beetle-killing temperatures. The
Forestry Chronicle 85:105-109.
Thomson, A. J., and D. M. Shrimpton. 1984. Weather associated with the start of mountain
pine beetle outbreaks. Canadian Journal of Forest Research 14:255-258.
Trouet, V., and A. H. Taylor. 2009. Multi-century variability in the Pacific North American
circulation pattern reconstructed from tree rings. Climate Dynamics 35:953-963.
Trzcinski, M. K., and M. L. Reid. 2009. Intrinsic and extrinsic determinants o f mountain
pine beetle population growth. Agricultural and Forest Entomology 11:185-196.
Turner, M. G. 2010. Disturbance and landscape dynamics in a changing world. Ecology
91:2833-2849.
Unger, L. 1993. Mountain pine beetle. Forest Pest Leaflet 76. Forestry Canada, Pacific
Forestry Centre, Victoria, BC. 7p.
US Department o f Agriculture. 1997. Forest insect and disease conditions in the United
States 1996. Washington DC: U.S. Department of Agriculture, Forest Service, Forest
Health Protection. 87 p.
Veblen, T. T., K. S. Hadley, E. M. Nel, T. Kitzberger, M. Reid, and R. Villalba. 1994.
Disturbance regime and disturbance interactions in a Rocky Mountain subalpine
forest. Journal of Ecology 82:125-135
Veblen, T. T., K. S. Hadley, M. S. Reid, and A. J. Rebertus. 1991a. Methods of detecting
past spruce beetle outbreaks in Rocky Mountain subalpine forests. Canadian Journal
of Forest Research 21:242-254.
Veblen, T. T., K. S. Hadley, M. S. Reid, and A. J. Rebertus. 19916. The response of
subalpine forests to spruce beetle outbreaks in Colorado. Ecology 72:213-231.
Welsh, C., K. Lewis, and A. Woods. 2009. The outbreak history of Dothistroma needle
blight: an emerging forest disease in northwestern British Columbia, Canada.
Canadian Journal o f Forest Research 39:2505-2519.
Westfall, J., and T. Ebata. 2007. Summary of forest health conditions in British Columbia.
Pest Management Report No. 15. British Columbia Ministry of Forests and Range,
Victoria, BC.
White, P. S., and S. T. A. Pickett. 1985. Natural disturbance and patch dynamics: an
introduction. Pages 3-13 in S. T. A. Pickett, and P. S. White, editors. The ecology of
natural disturbance and patch dynamics. Academic Press, New York, USA.

109

Wilson, R. J. S., and B. H. Luckman. 2002. Tree-ring reconstruction o f maximum and
minimum temperatures and the diurnal temperature range in British Columbia,
Canada. Dendrochronologia 20:257-268.
Wong, C., H. Sandmann, and B. Domer. 2004. Historical variability o f natural disturbances
in British Columbia: A literature review. FORREX Series 12. FORREX-Forest
Research Extension Partnership, Kamloops, British Columbia, Canada.
Wood, C. S., and L. Unger. 1996. Mountain pine beetle: A history of outbreaks in pine
forests in British Columbia, 1910 to 1995. Forest Health Network. Natural Resources
Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria , BC, Canada.
Wygant, N. D. 1940. Effects of low temperature on the Black Hills Beetle (Dendroctonus
Ponderosae Hopk.). Doctoral Thesis. Syracuse Univeristy.
Yamaguchi, D. K. 1991. A simple method for cross-dating increment cores from living trees.
Canadian Journal o f Forest Research 21:414-416.
Yoda, K., T. Kira, H. Ogawa, and K. Hozumi. 1963. Self-thinning in overcrowded pure
stands under cultivated and natural conditions. Journal o f Biology 14:107-129.
Zhang, Q. B., and R. I. Alfaro. 2002. Periodicity o f two-year cycle spruce budworm
outbreaks in central British Columbia: A dendro-ecological analysis. Forest Science
48:722-731.
Zhang, Q. B., and R. I. Alfaro. 2003. Spatial synchrony o f the two-year cycle budworm
outbreaks in central British Columbia, Canada. Oikos 102:146-154.
Zhang, Q., R. I. Alfaro, and R. J. Hebda. 1999. Dendroecological studies of tree growth,
climate and spruce beetle outbreaks in Central British Columbia, Canada. Forest
Ecology and Management 121:215-225.
Zhou, Y., R. W. Higgins, and H. K. Kim. 2001. Relationships between El Nino-Southern
Oscillation and the Arctic Oscillation: A climate-weather link. NCEP/Climate
Prediction Center. Available at
http://www.cpc.ncep.noaa.g0v/research_papers/ncep_cpc_atlas/8/.

110