EVALUATING NATURAL SELECTION AS A MANAGEMENT STRATEGY
FOR RESTORING WHITEBARK PINE
Shawn Thomas McKinney
B.A., University of Colorado at Boulder, 1996
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts
This thesis for the Master of Arts
Shawn Thomas McKinney
has been approved
Diana F. Tomback
Leo P. Bruederle
McKinney, Shawn Thomas (M.A., Biology)
Evaluating Natural Selection as a Management Strategy for Restoring Whitebark Pine
Thesis directed by Professor Diana F. Tomback
Declining whitebark pine (Pinus albicaulis Engelm.) populations throughout
northwestern United States and southwestern Canada are an issue of considerable
ecological concern. Reduced basal area and opportunities for regeneration are attributed
to advanced succession and damage wrought by blister rust (Cronartium ribicold), which
kills cone-bearing branches prior to killing the tree. This thesis examined the likely
effectiveness of the natural selection stand approach to whitebark pine management, a
strategy based on Clarks nutcrackers (Nudfraga Columbiana) dispersing seeds from
resistant trees located in heavily damaged stands and caching them in forest openings
created by fire or thinning. The model predicts that genetic resistance in dispersed seeds
coupled with an absence of competing species will, in 50 years, lead to stands dominated
by rust-resistant whitebark pine trees. To determine whether increased rust damage
influences seed dispersal potential and the likelihood of restoration through natural
regeneration, I measured structural and compositional characteristics of four one-hectare
stands in Montana and Idaho and observed cone survival of 183 whitebark pine trees
(2,436 cones) found in the stands. Comparable stands varied in degree of blister rust
damage but otherwise did not differ statistically in structure and composition. Slopes of
the regression lines of cone survival for the four stands differed significandy. Stands
with higher levels of rust damage had significandy lower cone densities, cone survival,
and probabilities of seed dispersal than ecologically similar stands with lower rust
damage. Results demonstrate that increasing blister rust levels reduce seed dispersal
potential at the stand level and suggest that the natural selection stand approach to
management cannot be relied upon exclusively to restore whitebark pine. To increase
chances of dispersal into prepared sites, restoration projects should be chosen based on
the cone density and spatial configuration of the proximate seed source stand.
This abstract accurately represents the content of the candidate^f^esis. I recommend
Diana F. Tomback
I dedicate this thesis to my son Finnegan. May he come to know the wonders and
beauty of whitebark pine forests.
I would like to thank my advisor, Dr. Diana F. Tomback, for all of her support
throughout the evolution and construction of this thesis. Her tireless efforts in aiding
my progress and in helping the future of whitebark pine cannot be overestimated.
Thanks to the United States Department of Agriculture, Forest Service, Rocky Mountain
Research Station Fire Sciences Laboratory, Missoula, Montana for funding the research
that allowed for the development of this thesis. Particularly to Dr. Robert Keane,
Research Ecologist at the Fire Sciences Laboratory, for taking interest in and recognizing
the merit of this study and for logistical and logical support. I thank the Department of
Biology at the University of Colorado at Denver for tuition and travel support to help
with the realization of this project. Thank you to my committee members Dr. Leo P.
Bruederle and Dr. Greg Cronin for help in designing this study and for staying with me
through these past three years. Thanks to Dr. John Basey for all of his insight into
experimental design and scientific reasoning and for giving me a job and place to hang
my hat for five years. My deepest gratitude and appreciation go out to my wife Lisa
without whom this goal would never have been achieved. Lisa was my field crew,
graphic designer, editor, and overall supporter throughout the development of this
thesis. Finally, thanks to my father, Dr. Thomas McKinney, who always believed in me
and allowed me to find my way while subdy leading by example. He fielded coundess
questions and concerns and for this I am grateful.
1.1 Whitebark Pine Decline.................................................3
1.2 Problem Statement......................................................5
2.1 Study Sites............................................................7
2.2 Tree Measurements and Blister Rust Infection...........................9
2.3 Whitebark Pine Seed Cones..............................................9
2.4 Seed Predator Observations............................................10
2.5 Data Analyses.........................................................11
3.1 Blister Rust Effects on Initial and Final Cone Density................14
3.2 Cone Survival.........................................................18
3.3 Comparison Between 2001 and 2002...............................23
4.1 Effect of Blister Rust on Cone Density....................................25
4.2 Effect of Cone Density on Cone Survival and Seed Dispersal................26
4.3 Implications to Management of Whitebark Pine............................27
A. Upper Big Creek Study Site................................................31
2.1 Map of Montana and Idaho showing the locations of the three whitebark pine
forests sampled from in this study..........................................8
3.1 Differences in seed predator occurrence rates due to harvesting among four
whitebark pine stands .....................................................17
3.2 Cone survival in four whitebark pine stands number of cones remaining
at time-of-caching by nutcrackers in mid August relative to the number in
3.3 Multiple comparisons of regression coefficients for cone survival (number
of whitebark pine cones remaining at time of seed dispersal on the number
of cones present in mid-July) from four stands in western MT
and eastern ID.............................................................21
3.4 Relationship between the ratio of seed predators to cone density and the
proportion of cones surviving to seed dispersal among four whitebark
3.1 Topography, structure, and composition of comparable whitebark pine stands
sampled in eastern Idaho and western Montana.............................15
3.2 White pine blister rust, overall mortality, and whitebark pine seed cone
values of comparable whitebark pine stands sampled in eastern Idaho
and western Montana......................................................16
3.3 F-test for equality of slopes of regression lines of cone survival.......20
3.4 Comparison of cone survival and observations of nutcrackers pouching and
caching whitebark pine seeds.............................................23
Whitebark pine (Pirns albicaulis Engelm.), a keystone species of subalpine forests
in northwestern United States and southwestern Canada, is in danger of regional
extinction (Arno 1986; Tomback et al. 2001). For about the last seventy years, fire
suppression and an introduced fungus white pine blister rust (Cronartium ribicold) have
acted to diminish whitebark pine basal area (via successional replacement) and
opportunities for new regeneration. One proposed management solution the natural
selection stand approach predicts that observed levels of genetic resistance to blister
rust (WPBR), coupled with silvicultural treatments that reduce site competition and
increase seed caching opportunities for Clarks nutcracker (Nucifraga columbiana), will lead
to stands dominated by rust-resistant trees (Hoff et al. 1994). However, litde is known
about the ecological effect of WPBR on pre-dispersal survival of seeds, and therefore,
the likely success of this strategy.
Ovulate cones of whitebark pine are indehiscent at maturity and are produced in
the upper third of the crown (Mirov 1967; McCaughey and Schmidt 1990). Whitebark
pine is functionally an obligate mutualist because it requires the Clarks nutcracker to
disperse its large and wingless seeds (Tomback and Linhart 1990). The nutcracker is a
facultative mutualist since it relies upon nutritious whitebark pine seeds but also utilizes
other sources of food, such as small vertebrates, carrion, insects, vegetable matter, and
seeds of other Pinus species (Guintoli and Mewaldt 1978; Tomback 1978; Tomback and
When whitebark pine seed coats harden (as early as August 15), nutcrackers
break into cones, gather seeds in their sublingual pouch, fly to a cache site, dig a shallow
trench, and deposit seeds in the substrate (Tomback 1978). Nutcrackers cache seeds in a
variety of locations, but tend to cache seeds in recendy disturbed open areas such as
those created by fire (e.g., Tomback 1986; Tomback 2001). By dispersing whitebark pine
seeds into recendy burned areas, nutcrackers are responsible for the early successional
status of the species (Tomback 1982; Tomback and Linhart 1990). Prior to seed coat
hardening, however, nutcrackers break into whitebark pine cones and consume unripe
seeds. This can occur as early as mid-July (Tomback 1978). During this period (mid-July
to mid-August), the nutcracker is a seed predator that reduces the whitebark pine seed
Nutcrackers are not the only threat to a stands seed crop. Red squirrels
(Tamiasciurus hudsonicus) are known to have significant impacts on the number of seeds
left for avian dispersal (Benkman et al. 1984; Benkman 1995; Samano and Tomback
2003). Benkman et al. (1984) found in forests that contained red squirrels, 80 percent of
the seed cones of limber pine (PinusJlexilis James) (a bird-dispersed pine) were harvested
prior to opening. These results were contrasted with forests where squirrels were absent
and only 30 percent of the cones were lost prior to time of dispersal. The loss of seeds
to red squirrels can be so great that it alters energy allocation within a population.
Benkman (1995) reported that limber pine allocated more than twice as much energy to
seeds in the absence of squirrels compared to areas where they were present.
1.1 Whitebark Pine Decline
Whitebark pine is declining throughout much of its range for two main reasons:
fire suppression and white pine blister rust. In the Northern Rocky Mountains,
whitebark pine is a successional species on favorable sites and relies upon periodic fire to
reduce encroachment by its competitors primarily Engelmann spruce (Picea engelmannii
Parry ex Engelm.) and subalpine fir (Abies lasiocarpa Hook.) and renew forest
communities (Arno and Hoff 1990; Arno 2001). The successful exclusion of fire from
the landscape, since around 1929, has altered this process and created conditions
unfavorable for whitebark pine. Additionally, fire suppression has reduced the
occurrence of disturbed areas suitable for nutcracker caching and whitebark pine
seedling growth, which has further exacerbated the successional replacement of
whitebark pine by spruce and fir (Keane et al. 1994).
The second leading cause of whitebark pine decline is white pine blister rust.
Cankers caused by the rust form on a tree, first killing the branches, then the crown, and
eventually the entire tree (McDonald and Hoff 2001). Most importantly, because seed
cones of whitebark pine form in the upper crown, cone production ceases and an
infected individual becomes reproductively dead well before the tree itself dies (Amo
and Hoff 1990).
Mortality induced by blister rust infection is as high as 90 percent in some stands
in regions such as the northwestern Rocky Mountains and the Olympic and Cascade
ranges (Kendall and Amo 1990; Keane et al. 1994). Throughout northwestern United
States and southwestern Canada, WPBR infection levels range from 20 to 100 percent
(Kendall and Keane 2001). Blister rust is spreading geographically within the range of
whitebark pine, intensifying within stands (e.g., Campbell and Antos 2000; Zeglen 2002),
and has the potential to result in regional extirpation of the species.
Keane et al. (1996) compared the historical distribution of whitebark pine in the
interior Colombia River Basin and Bob Marshall Wilderness Complex in Montana to its
current range and estimated that it has declined by 45 percent during the past 100 years.
Further evidence of population decline also comes from the Bob Marshall Wilderness
Complex where fewer than 10 percent of whitebark pine stands sampled had any
regeneration younger than fifty years old (Keane et al. 1994).
Within heavily infected stands, however, researchers rarely observed 100 percent
mortality from WPBR, and in some of these stands at least one to a few canker-free trees
were identified. Such observations suggested resistance to WPBR. Field observations of
phenotypically resistant trees were confirmed in the laboratory by Hoff et al. (1994) who
demonstrated that surviving whitebark pine trees from high-mortality stands possessed
higher levels of heritable resistance than individuals from stands exhibiting lower
mortality. In addition, more than 40 percent of progeny grown from seeds collected
from healthy survivors of heavily impacted stands demonstrated some resistance to the rust
(Hoff et al. 2001).
1.2 Problem Statement
The natural selection stand approach to whitebark pine management was
proposed to ameliorate the conditions wrought by blister rust and fire exclusion by
utilizing processes occurring on the landscape natural selection favoring rust resistant
trees and seed dispersal by nutcrackers. The proposal maintained that if areas adjacent
to stands that contained at least a few rust-resistant trees were cleared of vegetation,
nutcracker-caching opportunities would be created and site competition reduced. In
about 50 years a new stand dominated by rust-resistant whitebark pine would exist (Hoff
et al. 1994).
This approach would be an attractive management solution that is low in cost
and labor if it were not for the possible effect of seed predators on an inherently
diminished cone crop. High mortality areas have low whitebark pine tree density and
likely low cone density. If there is not a corresponding reduction in the number of seed
predators in the area, the seed predator to cone density ratio will greatly increase. Such a
change in this relationship could result in a depletion of the cone crop before mid-
August when nutcrackers begin to disperse seeds.
Therefore, if WPBR reduces whitebark pine cone density, does a decrease in
density lead to lower pre-dispersal survivorship of whitebark pine seeds? 'Hus study
examined the influence of WPBR on cone density, pre-dispersal seed survival, and the
likelihood of seed dispersal across similar stands that varied in degree of WPBR
infection, crown kill, and mortality. I postulated that: (1) stands with higher levels of
WPBR will have lower cone densities than similar stands with lower rust levels, (2)
stands with lower cone densities will have a smaller proportion of their cones survive to
time of seed dispersal, and (3) stands with a lower proportion of cones surviving will be
less likely to have seeds dispersed by nutcrackers.
2.1 Study Sites
Field sampling was conducted from 15 July to 4 September 2001 and from 15
July to 10 October 2002 in west-central Montana and east-central Idaho. Study sites
were selected based on the distribution of whitebark pine and known levels of WPBR
infection. Within each study site, two stands were selected for detailed observation.
Stands were selected based on their accessibility, distance from established cone
collection sites, and density of cone-bearing whitebark pine trees.
Blister rust infection levels are high in the Bitterroot National Forest, Montana,
and the Saddle Mountain study site was selected to represent that condition. The higher
elevation stand within this site was dominated by whitebark pine. This stand, Saddle
Mountain Upper, was located on the side of Saddle Mountain (4543'N, 11459'W,
2574 m elevation). The lower elevation stand at this site was comprised of a mixed
association of species, and whitebark pine was successional. This stand is referred to as
Saddle Mountain Lower (4541 'N, 11358'W, 2451 m elevation). Both stands were
adjacent to areas burned by wildfire in 2000.
The site containing lower blister rust infection values was located within the
Salmon National Forest, Idaho, and is referred to here as Blue Nose study site. Within
this site, the higher elevation stand was characterized as whitebark pine climax and is
located on Blue Nose peak. This stand is referred to as Blue Nose Upper (4528'N,
11421 'W, 2660 m elevation). The lower mixed association stand is referred to as Blue
Nose Lower (4527'N, 11420'W, 2553 m elevation). All four stands sampled in this
study were one hectare (ca. 100 x 100 m) in area and were accessible by a short walk
from established roads.
Figure 2.1 Map of Montana and Idaho showing the locations of the three whitebark
pine forests sampled from in this study.
| | Saddle Mountain
A Blue Nose
1-4 Upper Big Creek
2.2 Tree Measurements and Blister Rust Infection
Within each stand, all trees > 1.5 m in height and > 1 cm diameter at 1.3 m
above the ground (diameter at breast height DBH) were measured for height with a
clinometer and DBH with diameter measuring tape. All individuals < 1.5 m in height or
< 1 cm DBH were tallied and recorded by species.
All whitebark pine trees within each stand were inspected for the presence of
blister rust and tree damage from rust. A tree was considered to have the rust if it
exhibited active and inactive branch or bole cankers, branch or bole blisters, and/or a
dead or partially dead crown. 1 estimated the percentage of crown kill for all whitebark
pine trees that showed symptom of the rust. Percentage crown kill is a categorical
variable used to estimate the relative damage to a trees crown. The crown of a tree was
considered to include the area from the highest point of the tree down to the lowest
branches. Dead whitebark pine trees were tallied and, if possible, their cause of death
2.3 Whitebark Pine Seed Cones
All whitebark pine trees with seed cones were counted and tagged with a specific
identification code. A trees identification code was used for all other data or
measurements obtained for that individual. Densities of cone trees in the different study
stands were similar (mean=29.2/ha, min=27, max=32).
In mid July of 2001 and 2002, I counted all seed cones present on these trees.
This value is referred to as the initial number of cones and was used to calculate stand
means and initial cone density per stand. I used a spotting scope with a lx ocular and
40x objective, and 8 x 32 binoculars to survey and count cones. Cones were tallied with
a handheld counting device. This procedure was repeated between 10 and 16 days apart
for all cone-bearing trees within all stands.
The time of seed dispersal was determined for each stand when at least one
nutcracker was seen pouching seeds from a whitebark pine tree in the immediate area.
Following such an observation, I conducted another cone count in the stand. The value
obtained from this count is referred to as the final number of cones. Therefore, I
observed nutcrackers pouching in each study site before conducting a final cone count in
die stands of that site.
2.4 Seed Predator Observations
Pine seed predators that frequented these stands consisted primarily of Clarks
nutcrackers, red squirrels, and to a lesser extent, chipmunks (Tamias spp). Stands were
monitored for animal activity between 0530-2030 (minimum=two hrs. maximum height
hrs.) throughout the summer (mid-July to September). Field data were converted into
the number of sightings per hour for each predator. Observation periods were
distributed evenly within this timeframe to reduce temporal variation in animal activity.
During observation periods, time was split between surveying from vantage points with
spotting scope and binoculars or by moving throughout the stand. Observations of
squirrels and chipmunks were divided into two categories: 1) seed/cone harvesting or 2)
presence in the stand but not harvesting. Nutcracker observations were separated into
three categories: 1) seed harvesting 2) seed pouching or 3) within stand but not on cones.
The overall seed predator rate (SP) for each stand was calculated by the following:
SP = (CINu + ReSq + Chip) 1000
CINu = total of nutcracker seed harvesting observations/total hours of observation
ReSq = total of red squirrel seed/cone harvesting observations/total hours of observation
Chip = total of chipmunk seed harvesting observations/total hours of observation
2.5 Data Analyses
Data were plotted and tested for normality by looking at median, mode, and
mean values, histograms, and values of kurtosis and skewness. Descriptive statistics
were calculated for cone, blister rust infection level, whitebark pine tree, and basal area
data. Levenes test was used to test the null hypothesis of homogeneity of error variance
for cone (initial and final number), tree height, and tree diameter data.
Two-sample t-tests were used to compare population mean tree height and
diameter at breast height between paired stands. Two-sample Z-tests for proportions
were used to compare the proportion of whitebark pine relative to all other tree species,
proportion of whitebark pine infected with blister rust, proportion of infected trees with
greater than or equal to 50 percent crown kill (blister rust damage), and the proportion
of dead whitebark pine from all causes relative to the number living among paired
stands. All analyses were interpreted at a 95 percent confidence level (i.e., a = 0.05).
Data from 117 whitebark pine trees that produced 2,008 cones were analyzed
from 2001. Cone survival was determined by analyzing the relationship between the
number of final cones and the number of initial cones for whitebark pine trees in each
stand by simple linear regression. A spreadsheet was used to create a table of basic
comparisons in regression for all cone data. With these results I calculated an F-test for
homogeneity of slopes of the four regression lines (four stands) in order to test the null
hypothesis that seed survival is equal across all stands. I used Tukey-Kramer Multiple-
Comparison Procedure (Dunnett 1980) to calculate minimum significant differences
between pairs of regression coefficients (b,) with 95 percent confidence. Gabriels
Approximate Method (Gabriel 1978) was used to calculate 95 percent comparison
intervals among all bx's to determine specifically which stands were significandy different,
in terms of bu from one another.
The proportion of cones surviving to time-of-caching by nutcrackers was
calculated as follows: no. of final cones /no. of initial cones. This value was determined
for each whitebark pine cone tree and for stand means. The mean proportion surviving
was regressed against the ratio of SP to initial cone density for all stands to test the
strength of the relationship of predator abundance and initial cone density against cone
In 2002, the same 117 whitebark pine trees were surveyed that had been
surveyed in 2001. Far fewer cones were produced in 2002 and therefore analyses were
confined to testing for a year effect. Paired t-tests were used to compare the population
mean difference between 2001 and 2002 in terms of 1) number of cone-producing
whitebark pine trees, 2) number of initial cones, and 3) number of final cones. All
analyses were interpreted at a 95 percent confidence level.
3.1 Blister Rust Effects on Initial and
Final Cone Density
Differences in topography, stand structure, and relative abundance of whitebark
pine were controlled among comparable stands (Table 3.1). The comparable low
elevation stands, Blue Nose Lower and Saddle Mountain Lower, and high elevation
stands, Blue Nose Upper and Saddle Mountain Upper, were not statistically different in
the population proportion of whitebark pine relative to all tree species (lower stands
height of whitebark pine (lower stands tc1.89
but did differ in population mean height of whitebark pine (lower stands tc3.74>t530052.01,
upper stands tc4.68>t6OOO52.00) (Table 3.1).
Results indicated that WPBR severity as apparent from high levels of infection,
crown kill, and mortality was associated with a reduction in cone density among similar
stands. Blister rust infection levels and damage were greatest and cone density was
lowest in the Saddle Mountain stands (Table 3.2). 'There was a significantly higher
proportion of whitebark pine infected with blister rust (lower ZC10.62>Z0 051.96, upper
Zc6.34>Z0051.96), dead from all causes (lower ZC2.81>Z0051.96, upper ZC2.56>Z0051.96),
and with at least 50 percent crown kill (lower ZC13.10>Z()051.96, upper ZC3.05>Z0051.96)
in the Saddle Mountain stands compared to the Blue Nose stands. In addition, the
population mean number of initial whitebark pine cones (mid-July) was significandy
greater in the Blue Nose stands than in the Saddle Mountain stands (lower
tc5.81>t530052.01, upper tc7.67>t60ao52.00).
Table 3.1 Topography, structure, and composition of comparable whitebark pine
stands sampled in eastern Idaho and western Montana.
BN Lower SM Lower BN Upper SM Upper
Elevation (m) 2,553 2,451 2,660 2,574
Aspect South South South Southeast
Area (ha) 1 1 1 1
Slope (degrees) 5 7 28 22
No. of cone trees 28 27 30 32
Mean height (m) 12.4a 14.6b 7.4C 8.5d
Mean DBH (cm) SO oo n 21. V 15.2f 13.7f
Proportion WBP 0.45s 0.45s 0.91h 0.96h
Note: comparable stands with the same letter are not different with 95% confidence (to.os, Zo os). Values
are for whitebark pine trees only.
Table 3.2 White pine blister rust, overall mortality, and whitebark pine seed
cone values of comparable whitebark pine stands sampled in eastern Idaho and
BN Lower SM Lower BN Upper SM Upper
Proportion infected 0.24a 0.97b 0.54c 0.96d
Mean % crown kill 3 79 13 38
Proportion >50% damage 0.02e 0.65f 0.048 0.28h
Proportion dead 0.02' 0.29' 0.03k 0.18'
Initial cone density (no./ha) 644 422 723 227
Mean initial cone (no./tree) 23.0m 15.6" 24. lp 7.0r
Mean final cone (no./tree) 18.6s 1.5r 12.0s 1.4*
Note: comparable stands with the same letter are not different with 95% confidence (to.05, Zoos)- Values
are for whitebark pine trees only.
Stands with lower cone densities did not have correspondingly fewer seed
predators (number of sightings per hour) relative to higher cone density stands (Figure
3.1). The absolute occurrence rates of seed predators failed to show a trend with
increasing WPBR severity. However, when occurrence rates of seed predators were
viewed in relation to initial cone density, a clear trend emerged: as levels of WPBR
damage and infection increased, the numbers of seed predators per cone also increased.
Idius, the Saddle Mountain stands, where WPBR infection and damage were significantly
greater, had the two highest occurrence rates of seed predators per initial cone (Figure 3.1).
Figure 3.1 Differences in seed predator occurrence rates due to harvesting among
four whitebark pine stands. Stands are paired with their ecological counterpart (BN
I,ower with SM Lower and BN Upper with SM Upper). The Saddle Mountain area had
significandy greater infection and damage from blister rust than the Blue Nose area.
Red squirrel + Chipmunk
Seed predators/cone density *
SM Lower BN Upper
Note: Seed predators/cone density values are combined rates of nutcrackers, red squirrels, and
chipmunks harvesting seeds or cones divided by the initial density of cones x 1000.
3.2 Cone Survival
Regression analyses of cone counts from the Blue Nose study stands, where
blister rust infection and damage were significantly lower, indicated that the initial
number of cones on a tree was a strong predictor of the number of cones available at the
time of nutcracker seed dispersal (lower stand R2=0.95, sy|x=5.76 cones, n=28 trees,
upper stand R2=0.72, sy|x=7.67 cones, n=30 trees). At the Saddle Mountain site, where
blister rust infection and damage were significandy greater, initial cone number was a
weak predictor of cone availability at the time of seed dispersal (lower stand R2=0.10,
sy|x=2.94 cones, n=27 trees, upper stand R2=0.29, sy|x=2.49 cones, n=32 trees) (Figure 3.2).
An F-test comparing the regression coefficients of the four stands indicated
significant differences among their regression slopes (Fc8.68>F36g0052.75) (Table 3.3).
These results indicated that cone survival the relationship between the number of
cones remaining at time of seed dispersal by nutcrackers and the number available in mid
July differed significandy depending on stand location with 95 percent confidence.
Tukey-Kramer Multiple-Comparison Procedure revealed that in the lower
elevation stand at Blue Nose, where WPBR infection and damage were the lowest of all
stands sampled, the slope of the regression line of cone survival was significandy steeper
than the slope for all the other stands. In addition, the slope of the regression line for
the upper stand at Blue Nose was significandy greater than the Saddle Mountain lower
elevation stand but not the ecologically similar Saddle Mountain upper stand. The two
Saddle Mountain stands did not differ significandy from each other in terms of cone
survival with 95 percent confidence (Figure 3.3).
The ratio of seed predator occurrence rate (SP) to initial cone density had a
significant and negative linear relationship with the proportion of cones surviving to time
of seed dispersal (R2=0.95, sy|x=0.34 cone proportion, n=4 stands, p<0.05) (Figure 3.4).
There were significandy more cones available for nutcrackers at time of seed dispersal
per tree at the Blue Nose stands than at the Saddle Mountain stands (lower stands
tc8.91>t530O52.01, upper stands tc6.32>t6O0O52.00) (Table 3.2), and a higher proportion of
cones surviving favored the likelihood of nutcrackers pouching and or caching seeds in a
study stand (fable 3.4).
Figure 3.2 Cone survival in four whitebark pine stands number of cones remaining
at time-of-caching by nutcrackers in mid August relative to the number in mid-July.
There was a significant difference among the slopes of the four regression lines
(Fc8.68>F36g00S2.75). Y-intercept (bg), slope (Â£,), and R2 values are given for each stand.
Table 3.3 F-test for equality of slopes of regression lines of cone survival. Four
whitebark pine stands were sampled in western MT and eastern ID.
Sources of Variation df SS MS Fs
Among bs (variation among regressions) 3 1120.6 373.5 8.68*
Weighted average of deviations from regression 68 2920.7 42.9
(average variation within regressions)
Note: indicates that there were significant differences with 95% confidence among the four stands
when testing for homogeneity of the regression line.
Figure 3.3 Multiple comparisons of regression coefficients for cone survival (number
of whitebark pine cones remaining at time of seed dispersal on the number of cones
present in mid-July) from four stands in western MT and eastern ID. Two coefficients
are significantly different, with 95% confidence, if their intervals do not overlap.
BN Lower SM Lower BN Upper SM Upper
Note: Confidence intervals calculated using Tukey-Kramer Multiple-Comparison Procedure. Comparison
intervals plotted by Gabriels Approximate Method. Middle points represent b\, the regression coefficient
of a stand, upper bars the upper limit, and lower bars the lower limit of comparison intervals with 95%
Figure 3.4 Relationship between the ratio of seed predators to cone density and the
proportion of cones surviving to seed dispersal among four whitebark pine stands
(R2=0.95, s ,x=0.08 cone proportion, p<0.05).
* SM Upper
1 SM L ower
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00
(Cone harvesting observations x 1000) / initial cone density
Note: X-axis values are the sum of observations of nutcrackers, red squirrels, and chipmunks harvesting
seeds and cones per hour of observation multiplied by 1000, and divided by the density of cones in a
stand. Bars represent standard error.
Table 3.4 Comparison of cone survival and observations of nutcrackers pouching
and caching whitebark pine seeds. BN Lower and Upper had significantly less infection
and damage from blister rust compared to SM Lower and Upper. Indicates seed
caching observed in study site area, but not within study stand.
Variable BN Lower SM Lower BN Upper SM Upper
Cone survival3 0.81 0.10 0.47 0.21
Pouching*5 0.41 0.07 0.36 0
Caching0 yes no* yes no*
a: Proportion of cones surviving to time of seed dispersal relative to the initial mid-July number.
b: Observations of nutcrackers pouching whitebark pine seeds per hour of observation.
c: Whether or not nutcrackers were observed caching seeds in a study stand.
3.3 Comparison Between 2001 and 2002
In 2002, only 205 seed cones were produced in the same four stands compared
to 2,008 cones produced in 2001. There was a significant difference in the number of
whitebark pine trees producing cones in the four study stands between 2001 and 2002
(tc3.68>t30052.35). Consequently, in 2002, the initial and final numbers of cones were
significantly reduced relative to 2001 (initial cone number tc4.29>t30052.35, final cone
number tc2.49>t30052.35). These results are indicative of a region-wide low cone crop
for whitebark pine in 2002.
An 84 percent reduction in the number of initial cones translated into a 94
percent decline in the number of cones available for nutcrackers at time of seed dispersal
in 2002 compared to 2001. As a consequence, in the five stands monitored in 2002,1
failed to observe nutcrackers pouch seeds in all of them except one, Blue Nose upper.
The likelihood of whitebark pine seed dispersal by Clarks nutcrackers varied
greatly among four one-hectare forest stands in western MT and eastern ID. This can be
explained by differences in stand conditions resulting from infection and tree damage
from white pine blister rust and the proportion of dead whitebark pine trees (Table 3.2)
and not by differences in topography, whitebark pine relative abundance or average diameter
4.1 Effect of Blister Rust on Cone Density
WPBR infection of whitebark pine acts to reduce cone density within a stand by
killing cone-bearing branches and mature trees. Stands with significantly higher degrees
of infection, levels of crown damage, and mortality, had fewer cones per hectare and
significantly fewer initial (mid-July) cones per tree (Table 3.2).
Reduced cone densities of whitebark pine in high infection stands, however, did
not influence the frequency of seed predator observations (Figure 3.1). Populations of
forest seed predators may be experiencing a lag-time effect in terms of declining
whitebark pine populations, or they may be sustained by alternative food sources such as
the cones of other conifers. Whatever the reason, it is clear that when the combined
occurrence rates of seed predators were standardized to initial cone density, the number
of seed predators was much greater in low cone density stands than in stands with higher
4.2 Effect of Cone Density on
Cone Survival and Seed Dispersal
Reduced initial cone density due to blister-rust effects equated to significantly
fewer cones available at time of seed dispersal when higher impacted stands were
compared to lower impacted but ecologically similar stands fTable 3.2). However, a
more fundamental result of reduced cone density was evidenced by differences in the
relationship between the final number and initial number of cones, i.e., cone survival
(Figures 3.2 and 3.3). For example, the regression models for cone survival predict that
if each stand had 100 initial cones, the number of cones available at time of seed
dispersal would be 87 and 49 at the Blue Nose lower and upper stands, but only 8 and 21
at the Saddle Mountain lower and upper stands, respectively (Figure 3.2). Furthermore,
the significant difference between the regression slopes for cone survival (Table 3.3)
indicated that whitebark pine trees at Saddle Mountain and Blue Nose do not come from
the same population in regards to cone survival and, therefore, the likelihood of cone
survival is dependent upon stand location and thus blister rust condition.
When decreased cone densities are not met with corresponding reductions in
seed predator occurrence, the proportion of cones surviving to seed dispersal decreases
in a linear fashion (Figure 3.4). Within the Saddle Mountain stands the cone densities
were simply not high enough to withstand the pressure exerted upon the cone crops by
nutcracker foraging and red squirrel harvesting. Mortality of cone-bearing branches -
whether by crown kill or tree death from blister rust or mountain pine beede
(Dendroctonusponderosae) while not affecting seed predator densities, appears to be the
most serious threat to cone survival in the whitebark pine stands studied. In other
words, canopy kill reduces seed production although predator densities are not affected.
Ultimately, initial cone density and thus cone survival, influences the likelihood
of whitebark pine seed dispersal by Clarks nutcrackers. In this study, a lower proportion
of cones surviving to time of seed dispersal meant that there was a much lower
probability of observing a nutcracker pouch and disperse seeds from a stand (Table 3.4).
Therefore, the likelihood that a Clarks nutcracker will disperse seeds from a given area is
negatively related to WPBR infection and damage at the stand level.
4.3 Implications to Management of Whitebark Pine
Three criterion are necessary for the natural selection stand approach to be an
effective tool for whitebark pine restoration: 1) resistant trees survive exposure to blister
rust and are located in high mortality stands adjacent to forest openings, 2) resistant
individuals have their cones survive through the summer and their seeds dispersed by
nutcrackers, and 3) resistance in the offspring generation is higher in frequency than the
parental generation. If these criteria are met, in fifty years, there will be a stand
dominated by rust-resistant trees. Management efforts, such as thinning and burning
which reduce the density of competing species and create disturbed areas for nutcracker
caching, only foster this process. Without the three criteria met, silvicultural treatments
will be to no avail.
In 2002, field sampling in the Flathead National Forest, Montana, allowed for
further evaluation of the potential effectiveness of the natural selection stand model
(Appendix A). The situation in this forest is representative of the first condition
necessary for a natural restoration approach whitebark pine mortality is greater than 80
percent and phenotypically resistant mature trees are located in stands adjacent to
clearcuts. However, cone inventories and animal observations failed to confirm the
existence of the second necessary condition. The proportion of cones surviving to seed
dispersal was only 0.23 and nutcracker seed dispersal was not witnessed. Sampling of
whitebark pine seedlings and saplings also failed to verify the third condition increased
resistance in the next generation. Transect surveys in the adjacent clearcut revealed that
91 percent of all tree sites and 92 percent of all stems were infected with WPBR.
Results from the Bitterroot and Salmon National Forests show that as damage
from blister rust increases, cone density and the probability of seed dispersal decrease at
the stand level. Therefore, stands with very high levels of mortality, damage, and
infection (>80%) may rarely have seeds dispersed. This concern was supported by the
Flathead N.F. study. These cumulative results suggest that we cannot assume whitebark
pine seeds from heavily impacted stands will be dispersed into areas prepared for
Factors influencing the pre-dispersal survival of whitebark pine seeds are more
complex than this study addressed. Across the landscape matrix there will be areas with
high and low values of cone survival that are not solely due to their levels of blister rust
infection. Reduced cone densities from mountain pine beetle, fire, and logging activities
will likely lead to similar results.
A given whitebark pine stands spatial relationship to other stands in the broader
landscape, and the condition of those other stands, will also undoubtedly influence cone
survival. It is therefore imperative that during the site selection process of a restoration
project, site-specific and landscape conditions are considered. Two potential seed source
stands that are similar with regard to the presence of rust-resistant trees and magnitude
of cone density may be very different in terms of their likelihood of cone survival and
seed dispersal based on their spatial configuration and neighboring conditions on the
Cone survival, and thus the chance of seed dispersal from a given stand, will also
vary with time. In the short term, the magnitude of the cone crop, and in the long-term,
the effects of blister rust, pine beetle, and fire on the broader landscape, will likely
continue to shape the probability of seed dispersal and the chances of natural
regeneration resulting from a given area. Public land managers assigned the task of
restoring whitebark pine habitat will have to consider both the characteristics of a seed
source stand and its spatial relationship on the landscape in order to select the
restoration approach that ensures the greatest chance for success at the least expense.
With white pine blister rust spreading and infection levels and mortality
increasing in infected stands each year, the only way that whitebark pine will be
maintained on the landscape in large portions of its range will be by hands-on
restoration. Given the dynamics outlined in this thesis, the natural selection stand model
will probably not work under conditions typical of most heavily infected areas, arguing
for more active intervention, such as burning, thinning, and planting rust-resistant seeds
or seedlings. However, stands not as heavily damaged still have potential as natural seed
sources. We are now challenged to take the next step to determine the various factors
on multiple scales that affect the chance of seed dispersal from a given stand.
A. Upper Big Creek Study Site
Upper Big Creek study site was located within the Flathead National Forest,
Montana (between 4831 'N and 4832'N at 11423'W, 1696 m to 1805 m elevation).
This area consisted of a valley that was clearcut about thirty years prior to this study with
the upper slopes and ridgeline that enclose the valley left unlogged. Located within this
uncut area are many standing dead snags and fallen logs of once mature whitebark pine
trees that have succumbed to WPBR disease and mountain pine beetle. Whitebark pine
mortality in this area exceeds 80 percent. Within the opening adjacent to the uncut
stands are many whitebark pine seedlings (< 1.5 m) and saplings (> 1.5 m). This site is
one of the areas where Dr. Raymond Hoff observed phenotypically resistant mature
whitebark pine trees and high densities of whitebark pine seedlings. I sampled at this site
because it afforded an opportunity to direcdy evaluate the effectiveness of the natural
selection stand hypothesis.
Due to high levels of mortality at Upper Big Creek, mature cone-bearing trees
are rare and sparsely arranged on the landscape. Eleven cone-bearing whitebark pine
trees along a five km transect were surveyed for cones. Height, DBH, degree of blister
rust damage, and indication of seed predators were also recorded for these trees.
One important prediction that logically arises from the natural selection stand
hypothesis is that the next generation of trees will show lower frequencies of WPBR. To
test this, on September 22, 2002,1 surveyed two 50 m by 10 m transects that were
approximately one km apart in the clearing adjacent to the whitebark pine stands. All
whitebark pine seedlings and saplings were inspected for symptoms of the rust, and had
their number of stems, heights, and diameters recorded.
I sampled in the Upper Big Creek area in 2002 in order to test assumptions of
the natural selection stand model. The first criterion of the model, that resistant trees
will survive exposure to blister rust was supported. Five of the 11 cone trees surveyed
showed no indication of blister rust, and of the six infected the mean crown kill was only
16%. These trees were located in stands with mortality levels greater than 80%.
The second criterion of the model, that resistant individuals will have their seeds
dispersed by nutcrackers was not supported. A 0.23 proportion of cones survived to
time of seed dispersal among the 11 trees sampled, and I failed to observe nutcrackers
pouching and dispersing seeds. These results indicate that this second condition had a
low probability of realization in 2002.
Finally, I failed to confirm the presence of the condition for the third criterion;
that resistance in the post-rust exposed generation will be higher in frequency than the
pre-rust exposed generation. Results from the blister rust transects showed that 92
percent of all stems and 91 percent of all tree sites sampled were infected, and 22 percent
of all tree sites were dead from WPBR.
Amo, S.F. 1986. Whitebark pine cone crops A diminishing source of wildlife food?
Western journal of Applied Forestry 1:92-94.
Amo, S.F. 2001. Community types and natural disturbance processes. Pp. 74-88 in D.F.
Tomback, S.F. Amo, and R.E. Keane, editors. Whitebark pine communities: ecology and
restoration. Island Press, Washington, D.C.
Amo, S.F., and R.J. Hoff. 1990. Pinus albicaulis Engelm. Whitebark pine. Pp. 268-279 in
R.M. Bums and B.H. Honkala, technical coordinators. Silvics of North America, vol. 1.
Conifers. USDA Forest Service, Agriculture Handbook 654, Washington D.C.
Benkman, C.W. 1995. The impact of tree squirrels (Tamiasciurus) on limber pine seed
dispersal adaptations. Evolution 49:585-592.
Benkman, C.W., R.P. Baida, and C.C. Smith 1984. Adaptations for seed dispersal and
the compromises due to seed predation in limber pine. Ecology 65:632-642.
Campbell, E.A., and J. A. Antos. 2000. Distribution and severity of white pine blister
rust and mountain pine beetle on whitebark pine in British Columbia. Canadian journal of
Forest Research 30:1051-1059.
Dunnett, C.W. 1980. Pairwise multiple comparisons in the homogeneous variance,
unequal sample size case, j Abner. Stat. Assn. 75:789-795.
Gabriel, K.R. 1978. A simple method of multiple comparisons of means. ].Amer. Stat.
Giuntoli, M., and L.R. Mewaldt. 1978. Stomach contents of Clark's nutcrackers collected
in western Montana. Auk 95:595-598.
Hoff, R.J., S.K. Hagle, and R.G. Krebill. 1994. Genetic consequences and research
challenges of blister mst in whitebark pine forests. Pp. 118-126 in W.C. Schmidt and
F.K. Holtmeier, compilers. ProceedingsInternational Workshop on Subalpine Stone Pines and
Their Environment: The Status of Our Knowledge. USDA Forest Service, Intermountain
Research Station, General Technical Report INT-GTR-309, Ogden, Utah.
Hoff, R.J., D.E. Ferguson, G.I. McDonald, and R.E. Keane. 2001. Strategies for
managing whitebark pine in the presence of white pine blister rust. Pp. 346-366 in D.F.
Tomback, S.F. Amo, and R.E. Keane, editors. Whitebark pine communities: ecology and
restoration. Island Press, Washington, D.C.
Keane, R.E., P. Morgan, and J.P. Menakis. 1994. Landscape assessment of the decline of
whitebark pine (Pinus albicaulis) in the Bob Marshall Wilderness Complex, Montana,
USA. Northwest Science 68:213-229.
Keane, R.E., J.P. Menakis, and W.J. Hann. 1996. Coarse-scale restoration planning and
design in Interior Colombia River Basin Ecosystems: An example for restoring declining
whitebark pine forests. Pp. 14-19 in C.C. Hardy and S.F. Amo, editors. The use of fire in
forest restoration. USDA Forest Service, Intermountain Research Station, General
Technical Report INT-Gtr-341, Ogden, Utah.
Kendall, K.C., and S.F. Amo. 1990. Whitebark pineAn important but endangered
wildlife resource. Pp. 264-273 in W.C. Schmidt and K. J. McDonald, compilers.
ProceedingsSymposium on Whitebark Pine Ecosystems: Ecology and Management of a High-
Mountain Resource. USDA Forest Service, Intermountain Research Station, General
Technical Report INT-270, Ogden, Utah.
Kendall, K.C., and R.E. Keane. 2001. Whitebark pine decline: infection, mortality, and
population trends. Pp. 221-242 in D.F. Tomback, S. F. Amo, and R. E. Keane, editors.
Whitebark pine communities: ecology and restoration. Island Press, Washington, D.C.
McCaughey, W.W., and W.C. Schmidt. 1990. Autecology of whitebark pine. Pp. 85-96
in W.C. Schmidt and K.J. McDonald, compilers. ProceedingsSymposium on Whitebark Pine
Ecosystems: Ecology and Management of a High-Mountain Resource. USDA Forest Service,
Intermountain Research Station, General Technical Report INT-270, Ogden, Utah.
McDonald, G.I., and R.J. Hoff. 2001. Blister rust: an introduced plague. Pp. 193-220 in
D. F. Tomback, S. F. Arno, and R. E. Keane, editors. Whitebark pine communities: ecology
and restoration. Island Press, Washington, D.C.
Mirov, N.T. 1967. The Genus Pinus. The Ronald Press, New York.
Samano, S., and D.F. Tomback. 2003. Cone opening phenology, seed dispersal, and
seed predation in southwestern white pine (Pinus strobiformis) in southern Colorado.
Tomback, D.F. 1978. Foraging strategies of Clark's nutcracker. The hiving Bird 16:123-161.
Tomback, D.F. 1982. Dispersal of whitebark pine seeds by Clark's nutcracker: a
mutualism hypothesis. Journal of Animal Ecology 51:451-467.
Tomback, D.F. 1986. Post-fire regeneration of krummholz whitebark pine: a
consequence of nutcracker seed caching. Madrono 33:100-110.
Tomback, D.F., and Y.B. Linhart. 1990. The evolution of bird-dispersed pines.
Evolutionary Ecology 4:185-219.
Tomback, D.F. 2001. Clarks nutcracker: Agent of regeneration. Pp. 89-104 in D.F.
Tomback, S.F. Amo, and R.E. Keane, editors. Whitebark pine communities: ecology and
restoration. Island Press, Washington, D.C.
Tomback, D.F., S.F. Amo, and R.E. Keane. 2001. The compelling case for management
intervention. Pp. 3-25 in D.F. Tomback, S.F. Amo, and R.E. Keane, editors. Whitebark
pine communities: ecology and restoration. Island Press, Washington, D.C.
Zeglen, S. 2002. Whitebark pine and white pine blister rust in British Columbia, Canada.
Canadian Journal of Forest Research 32:1265-1274.