ArticlePDF Available

The 1912 Douglas-Fir Heredity Study: Long-Term Effects of Climatic Transfer Distance on Growth and Survival


Abstract and Figures

The 1912 Douglas-Fir Heredity Study is one of the first studies undertaken by the US Forest Service, and one of the first forest genetics studies in North America. The study considers provenance variation of 120 parent trees from 13 seed sources planted at five test sites in the Pacific Northwest. The unique, long-term nature of the study makes it valuable to revisit and consider its biological and historical significance. This analysis considers how far climatically Douglas-fir populations may be moved without incurring unacceptable declines in growth and survival. Results indicate that Douglas-fir seed sources may be moved at least 2° C cooler or warmer and still retain good long-term survival and productivity. However, projected future climate change beyond 2° C may lead to lower survival and productivity. One option to address these concerns is assisted migration; however, if seed sources are moved beyond 2–3° C to a cooler climate in anticipation of warming, or from a more continental to a maritime climate, we are likely to see increased mortality and associated losses in productivity in the near-term. Lessons from this study include: (1) pay attention to good experimental design; we were able to overcome limitations from the design by using new statistical approaches; (2) maladaptation may take time to develop; poorer survival was not evident until more than two decades after planting; and (3) long-term studies may have value for addressing new, unforeseen issues in the future.
Content may be subject to copyright.
Journal of Forestry, 2019, 1–13
Research Article - forest ecology
Received July 2, 2019; Accepted October 21, 2019
Advance Access publication December 10, 2019
Published by Oxford University Press on behalf of the Society of American Foresters 2019.
This work is written by (a) US Government employee(s) and is in the public domain in the US.
Research Article - forest ecology
The 1912 Douglas-Fir Heredity Study: Long-Term
Effects of Climatic Transfer Distance on Growth
and Survival
J.Bradley St.Clair, GlennT. Howe, and JenniferG. Kling
J. Bradley St.Clair (, USDA Forest Service, Pacific Northwest Research Station, Corvallis, OR.
Glenn T.Howe ( and Jennifer G.Kling (, Oregon State
University, Corvallis, OR.
The 1912 Douglas-Fir Heredity Study is one of the first studies undertaken by the US Forest Service,
and one of the first forest genetics studies in North America. The study considers provenance variation
of 120 parent trees from 13 seed sources planted at five test sites in the Pacific Northwest. The unique,
long-term nature of the study makes it valuable to revisit and consider its biological and historical sig-
nificance. This analysis considers how far climatically Douglas-fir populations may be moved without
incurring unacceptable declines in growth and survival. Results indicate that Douglas-fir seed sources
may be moved at least 2°C cooler or warmer and still retain good long-term survival and productivity.
However, projected future climate change beyond 2°C may lead to lower survival and productivity.
One option to address these concerns is assisted migration; however, if seed sources are moved be-
yond 2–3°C to a cooler climate in anticipation of warming, or from a more continental to a maritime
climate, we are likely to see increased mortality and associated losses in productivity in the near-term.
Lessons from this study include: (1) pay attention to good experimental design; we were able to over-
come limitations from the design by using new statistical approaches; (2) maladaptation may take
time to develop; poorer survival was not evident until more than two decades after planting; and (3)
long-term studies may have value for addressing new, unforeseen issues in the future.
Keywords: provenance test, adaptation, climate change, climatic transfer distance, forest history
The turn of the 20th century saw a great increase in
the protection and management of forests in the United
States. The Forest Reserve Act of 1891 allowed the
President to establish forest reserves from public lands;
the Organic Act of 1897 provided for the management of
those reserves; and in 1905, the administration of all fed-
eral forestry activities was united under the newly formed
US Forest Service within the Department of Agriculture.
With the new emphasis on protection and management,
there was increased interest in research and the applica-
tion of “scientic forestry.” The early 20th century also
saw the establishment of the rst forestry programs at
universities in the United States. In 1898, for example,
Cornell established a forestry program led by Bernard
Fernow, an early forestry leader educated in Germany.
One of his rst students, a Russian immigrant named
Rafael Zon, became head of the Forest Service’s Ofce
of Silvics, and a major proponent of forestry research.
After touring Germany, Austria, and France in the winter
of 1908, Zon wrote a memo to Gifford Pinchot, Chief
of the Forest Service, proposing to establish forest ex-
periment stations in the western United States similar to
those he saw in Europe. The rst experiment station was
established in the spring of 1908 at Fort Valley on the
Downloaded from by guest on 16 December 2019
2Journal of Forestry, 2019, Vol. XX, No. XX
Coconino National Forest in Arizona. Later, Zon argued
for separating all research work from administration
within the Forest Service, leading to the establishment of
the Branch of Research in1915.
Forestry research saw its beginnings in the Pacic
Northwest with the arrival in 1908 of Thornton
T.Munger, fresh from completing a masters degree in
forestry from Yale University. Munger was assigned to
the Section of Silvics at the Forest Service’s new North
Pacic District in Portland, Oregon, where he estab-
lished some of the rst forestry research plots in the
western United States. These plots were important for
demonstrating the high growth potential of Douglas-
r (Pseudotsuga menziesii var. menziesii). Prior to that,
the species was seen primarily as a resource to be mined
by the timber industry. Around the same time, res
burned through large areas of the Pacic Northwest.
The Yacolt burn of 1902 was particularly damaging,
burning 238,000 acres and killing 65 people near the
Columbia Gorge. Because natural regeneration was
largely ineffective, the Wind River Nursery, one of the
rst in the West, was established in 1909 to reforest the
burned and cutoverlands.
At the encouragement of Zon, Munger began
studying the suitability of species and seed sources
for reforestation. These studies likely stemmed from
two important inuences. First, exotic species were
receiving a great deal of attention in Europe, particu-
larly species from western North America, including
Douglas-r. Second, differences among seed sources
were recognized in early European provenance tests,
such as the rst large-scale provenance study that
was initiated in 1907 with Scots pine (Langlet 1971).
Another important inuence was the rediscovery
of Mendel’s Laws of Heredity in 1900 (Sandler and
Sandler 1986). Surely, the activities in Europe and the
new science of genetics must have contributed to the
thinking and activities undertaken by Zon and Munger.
In 1912, Munger initiated two inuential studies
on the suitability of species and seed sources for the
Pacic Northwest. The rst study, known as the Wind
River Arboretum Study, began with a few trials of
eastern hardwood species and then grew to 152 conifer
and hardwood species and varieties (Silen and Olson
1992). The Wind River Arboretum was important for
demonstrating the superiority of native species in the
Pacic Northwest, something that was not all that cer-
tain at the beginning. It appears, however, to be a one-
way street—although Douglas-r and other western
species are important worldwide, most exotic species
did not perform well in the Pacic Northwest (Silen
and Olson 1992). The Wind River Arboretum has also
been important to show the impact of tree develop-
ment and climate over time. If foresters had put too
much faith into early results, we might have seen many
failed plantations of Siberian larch in the region.
The second study became known as the Douglas-
Fir Heredity Study, one of the earliest forest genetics
studies in North America. In the fall of 1912, Munger
had a crew collect cones from 13 locations in the Coast
and Cascade Ranges that differed in latitude, elevation,
and soil type (Munger and Morris 1936). At some
sites, parent trees were selected to reect differences
in stand age, stand density, and disease infection. The
progeny from 120 maternal parent trees were grown
at the Wind River Nursery, and then outplanted in
1915 and 1916 at six locations in western Oregon and
Washington. The original objectives were to determine
the best parents to use as seed trees after logging or
for collecting seed for articial regeneration. Munger
was particularly interested in knowing whether the off-
spring of parents from different sites or classes of trees
inherited different characteristics—hence the name of
Early results by Munger and Morris (1936) found
no important differences among the progeny of parents
Management and Policy Implications
Douglas-fir seed sources may be expected to retain good growth and survival for changes in climate of 2°C warmer
or colder, whether those changes occur by seed transfer or by climate change at a site. This is within current seed zone
guidelines and within current observed levels of climate change. Thus, past reforestation practices have helped ensure
productive forests.
Climates are expected to warm more than 2°C over the next few decades, leading to lower survival and productivity of
Douglas-fir stands. Assisted migration to move seed sources from warmer to colder climates has been suggested to en-
sure adaptation of future forest stands. Seed sources should not be moved, however, more than 2–3°C to a cooler climate
in order to ensure good survival in the near-term.
Movements from the more continental climates of the Cascade Range to the more maritime climates of the Coast Range
should be avoided.
There is value in long-term field studies. Changing objectives may lead to new insights.
Downloaded from by guest on 16 December 2019
3Journal of Forestry, 2019, Vol. XX, No. XX
differing in age, stand density, fungal infection, or soil
type. However, high-elevation sources grew best at the
high-elevation test site, and the coastal seed source
grew best at the milder coastal site. Despite this evi-
dence for local adaptation, two sources from the cen-
tral Washington Cascades, near Granite Falls and
Darrington, grew well across all test sites. In contrast to
growth, they did not nd large differences in mortality
among the seed sources at any of the test sites. These
results, combined with anecdotal evidence of poor
growth of off-site plantations compared to adjacent
naturally regenerated stands, spurred the development
of the rst seed collection guidelines and seed zones
for the Douglas-r region (Isaac 1949). In 1939, the
USDA established a “Forest Seed Policy” that required
the use of seed of known origin, and recommended
that seed be planted within 100 miles and 1,000 feet in
elevation from its origin. In 1942, Munger divided the
region into nine “provenances” that he considered to
be climatically homogenous. Beginning in the 1950s,
Roy Silen of the US Forest Service Pacic Northwest
Research Station was instrumental in continuing to
measure the study over the next half century. He re-
ported on 50-year results in a Station annual report
in 1963, and at two conferences in 1964 (Silen 1963,
Silen 1965, Silen 1966). His ndings strongly inu-
enced ideas of ne-scaled local adaptation in the re-
gion. By 1962, a system to certify forest tree seed was
established for Oregon and Washington, and in 1966,
seed-zone maps were developed that continue to be
widely used. In addition to seed zones, family differ-
ences observed in the Douglas-Fir Heredity Study were
inuential in promoting tree improvement programs in
Despite the inuence of the Douglas-Fir Heredity
Study, little has been published since the reports in 1936
and the 1960s. The unique, long-term nature of the
study makes it particularly valuable to revisit and con-
sider its biological and historical signicance. Although
the study has limitations associated with the experi-
mental design, new statistical techniques allow us to
better evaluate the results. Furthermore, the increasing
concern over climate change brings renewed interest
in understanding adaptive responses of populations
to climate, and the climatic distance that populations
may be moved in anticipation of continued warming.
With this in mind, the objectives of this analysis were
to explore (1) the climatic distance that Douglas-r
populations can be moved while maintaining accept-
able growth and survival, and (2) how tree growth and
survival change over time and in response to specic
climatic events. Asecondary goal is to provide the his-
torical context for this unique, long-termstudy.
Materials and Methods
In the fall of 1912, cones were collected from 120
open-pollinated Douglas-r trees from 13 seed sources
(also referred to as provenances) in western Oregon
and Washington (Table 1A; Figure 1). The number of
parent trees per source location varied between three
and 21. Three provenances (Santiam, Palmer, and Race
Track) were chosen to represent higher elevation lo-
cations, and one provenance (Lakewood) came from
an area of glacial outwash soils with poor site quality.
Different parents were selected within provenances to
contrast different age classes and open-grown versus
dense competition from neighboring trees. In addition,
trees were selected within the Wind River and Gates
provenances to explore differences between parents in-
fected versus uninfected with the red ring rot pathogen
(Phellinus pini). Munger and Morris (1936) found
little difference in heights and survival between parents
chosen from different age classes, stand densities, and
infection status within sites, and little difference be-
tween the Lakewood provenance from poor soils and
the other sites. These results are consistent with our
ndings; thus, our analysis focused on differences
among provenances as related to the climates of the
source locations.
Test plantations
Seeds from the 120 parent trees were sown at the Wind
River Nursery in the springs of 1913 and 1914, grown
for two growing seasons, and then outplanted at six
test sites in the springs of 1915 and 1916 (Table 1B).
The test sites were chosen to represent typical plan-
ting sites in the study area (Figure 1). Are in 1917
destroyed the middle elevation site in the northern
Oregon Cascades; thus, this site was abandoned (and
not included in Table 1B). The 1917 re also destroyed
part of the Upper Mt Hood site. Shortly after plan-
ting, the 1915 portion of the Stillaguamish site and the
1916 portion of the Hebo site were damaged by moun-
tain beaver and were not measured again until 1963
(Silen 1963). Additional early mortality was mostly
caused by falling snags. Trees that died before 1917
and trees that had been killed by re were excluded
from the data analyses, as were the 1915 portion of the
Stillaguamish plantation and the 1916 portion of the
Hebo plantation.
Downloaded from by guest on 16 December 2019
4Journal of Forestry, 2019, Vol. XX, No. XX
Table 1. Location and climate information* for provenances and test sites used in the Douglas-Fir Heredity Study (sorted by decreasing mean
annual temperature).
Name Latitude (°) Longitude (°) Elevation (m) MAT (°C) MCMT (°C) MWMT (°C) TD (°C) MAP (mm) MSP (mm)
A. Provenances
Portland 45.489 –122.730 90 11.0 3.5 18.8 15.2 1,150 200
Lakeview 47.176 –122.592 30 10.7 3.8 17.9 14.1 1,063 198
Benton 44.642 –123.580 215 10.5 3.8 17.7 13.8 1,716 236
Carson 45.718 –121.825 120 10.4 1.7 18.8 17.1 1,721 244
Gates 44.750 –122.417 290 9.9 2.8 17.6 14.7 1,748 312
Granite Falls 48.104 –121.917 120 9.6 2.3 17.1 14.8 1,939 413
Hazel 48.263 –121.844 275 8.9 1.6 16.7 15.0 2,340 476
Darrington 48.254 –121.592 150 8.9 0.8 17.1 16.3 2,418 403
Wind River 45.823 –121.958 410 8.9 0.2 17.9 17.7 2,444 297
Fortson 48.267 –121.725 150 8.5 0.9 16.4 15.5 2,689 507
Race Track 45.897 –121.850 790 7.2 –0.7 15.9 16.6 2,628 332
Santiam 44.661 –121.907 975 7.1 –0.6 16.0 16.6 2,041 298
Palmer 45.559 –122.001 915 6.8 –0.7 15.0 15.7 3,372 548
B. Test sites
Wind River 45.792 –121.927 353 9.0 0.3 17.9 17.7 2,626 316
Hebo 45.148 –123.756 638 8.3 2.5 14.7 12.2 2,675 393
Stillaguamish 48.075 –121.606 579 7.7 –0.5 15.8 16.4 3,250 653
Lower Mt Hood 45.268 –121.821 853 6.9 –0.8 15.4 16.2 2,014 349
Upper Mt Hood 45.263 –121.774 1,372 5.4 –2.0 14.2 16.1 1,912 330
Note: MAT, mean annual temperature; MCMT, mean coldest month temperature; MWMT, mean warmest month temperature; TD, continentality as determined by the
difference between MWMT and MCMT; MAP, mean annual precipitation; MSP, mean summer precipitation.
*Climatic information derived from ClimateNA (Wang etal. 2016).
Downloaded from by guest on 16 December 2019
5Journal of Forestry, 2019, Vol. XX, No. XX
The eld design at each test site was family row
plots with 20 trees per row in 1915, and 10 trees
per row in 1916. Because families and provenances
were replicated across years, but not within years, we
treated the 1915 and 1916 plantings as blocks during
data analysis. Trees were planted at a 2.1 m × 2.1 m
spacing, and ller trees were used when necessary to
ll out a row. In the 1915 plantings, an additional 11
families from ve of the 13 provenances were planted
in 100-tree row plots across the length of the plan-
tation (included in the analysis) with the idea of ac-
counting for microsite variation at each site. Although
ultimately not fruitful, this was an idea that was
quite forward-thinking for its time. Within each plan-
ting year, families were grouped by provenance; that
is, family row plots from the same provenance were
planted adjacent to one another. Border rows were
not planted around each plantation, although adja-
cent trees were of similar size and likely resulted from
regeneration at about the same time. The study was
not thinned at any testsites.
Although survival and heights were measured in
early years, complete records were only available be-
ginning in 1923. Survival was measured after planting
and approximately every 10years between 1923 and
1993, and in 2013. Trees that died before 1917 were
excluded from the analysis with the assumption that
much early mortality may be due to poor planting.
Diameter was measured at breast height on all trees
approximately every 10years from 1931 to 1993, and
in 2013. Height was measured on all trees in 1923,
1931, 1963, 1993, and 2013. We also calculated in-
dividual tree volume (using methods of Poudel and
Hailemariam 2016) and volume per hectare for those
years in which both height and diameter measurements
were available.
Data Analysis
We hypothesized that provenance performance was re-
lated to the climatic difference between the test site and
the seed-collection location (i.e., where the parent trees
were exposed to climatic selection pressures). The vari-
ation among provenances for each trait in each year
of evaluation was analyzed with the SAS GLIMMIX
procedure (SAS Institute 2014), using restricted max-
imum likelihood and a fully random model. The rep-
licated plantings of families from each provenance in
1915 and 1916 provided a measure of the nongenetic
variation within test sites. The statistical model used
for the across-site analyses was as described below.
A similar reduced model was used for analyses of
+SY Ps·y·p+Ff(p)+SFs·f(p)+YF
n(sypf )
where Zsypfn is the observation for the nth tree in the fth
family in the pth provenance in the yth planting year at
the sth test site; Ss is the effect of the sth test site; Yy is
the effect of the yth planting year; SYs∙y is the interaction
of the sth test site and yth planting year; Pp is the effect
of the pth provenance; SPs∙p is the interaction of the sth
test site and pth provenance; YPy∙p is the interaction of
the yth planting year and pth provenance; SYPs∙y∙p is the
interaction of the sth test site, yth planting year, and
pth provenance; Ff(p) is the effect of the fth family in
the pth provenance; SFs∙f(p) is the interaction of the sth
test site and fth family in the pth provenance; YFy∙f(p) is
the interaction of the yth planting year and fth family
in the pth provenance; SYFs∙y∙f(p) is the interaction of
the sth test site, yth planting year, and fth family in the
Figure 1. Map of provenances (orange circles) and test
sites (blue squares) in the Douglas-Fir Heredity Study.
Downloaded from by guest on 16 December 2019
6Journal of Forestry, 2019, Vol. XX, No. XX
pth provenance; and εn(sypf) is the random, independent
error, ~N(0, σ
2). The overall mean (µ) is a xed effect,
and all other effects in the model are random.
For growth measurements, we rst performed ana-
lyses for each trait–year variable at each site to identify
outliers. Observations with studentized residuals that
were greater than three were excluded from further
analyses. The intercept from the linear model for each
site was used as an estimate of the site mean. For the
combined analysis of growth traits across sites, data
were divided by the square root of the error variance
at each site to remove scale effects and standardize the
variance (White 1996). Ageneralized linear model with
a logit link function was used to analyze survival traits.
Likelihood ratio chi-square tests were used to determine
which trait–year combinations had signicant variance
components for sites, provenances, and site × prov-
enance interactions. To investigate relations between
provenance performance and provenance climatic
origin, we used Best Linear Unbiased Prediction (BLUP)
estimates for site × provenance interactions for trait–
year variables that were signicant in the across-site
analysis using a p-value of <0.10. Climates of test sites
and source locations were determined using 30-year
normal data (1961–90) from ClimateNA (Wang etal.
2016). Pearson correlation coefcients between site ×
provenance BLUPs and source climates were estimated
at each site. To further characterize the effects of cli-
mate variables on adaptation across a range of environ-
ments, we obtained linear and quadratic equations for
the regression of site × provenance BLUPs on climatic
transfer distances. These analyses accounted for differ-
ences in site productivity in two ways. First, because
we standardized the observations by plantation, subse-
quent analyses were based on standardized BLUPs that
accounted for differences in site productivity. Second,
although the random effects model included the main
effects of site and provenance, the pooled transfer func-
tion used only the site × provenance BLUPs—essentially,
deviations from the main effects. Site main effects may
result from differences in climate, soils, topography,
competing vegetation, and management, whereas prov-
enance main effects may result from differences in cli-
matic origin, demographic history, and sampling error.
By omitting the main effects from the pooled transfer
function, we reduced the impact of these confounding
factors, resulting in a transfer function that is broadly
applicable to the provenances and sites we studied. For
ease of interpretation, site × provenance BLUPs for sur-
vival were centered on the mean percent survival across
sites. This was obtained using the inverse link function
for the intercept from the combined logistic regression
analysis. Climatic transfer distances were estimated for
each climate variable as the test site climate minus the
source climate.
Test sites differed in growth and survival (Tables 2
and 3). In particular, trees at the three warmer sites of
Wind River, Hebo and Stillaguamish were taller and
had a greater diameter than those at the two cooler
sites of Lower Mt Hood and Upper Mt Hood (Table
2). Survival by ages 18 and 19 was high at all sites (78–
92 percent). Survival declined substantially over time,
ranging from 31 percent at Lower Mt Hood to 15 per-
cent at Stillaguamish by 2013. Some decades showed
higher rates of mortality, including the 1930s at Upper
Mt Hood, the 1940s at Stillaguamish, the 1950s at
Hebo, and the 1960s at Wind River (Figure 2).
Provenance × site interactions provide evidence
for local adaptation to climate. Aprovenance × site
interaction was found for survival for measurements
taken in 1941 and later (p< 0.10) and was particularly
strong by 2013 (p= 0.008) (Table 3). Aprovenance ×
site interaction for survival, however, was not found
in 1931. The differential survival of provenances at
different sites appeared to be driven by factors arising
Table 2. Test site means for growth and survival traits in 1931 and2013.
1931 2013
Test site
hectare (m3 ha–1)
Wind River 4.5 6.0 78 39.9 42.8 783 20
Hebo 6.3 8.4 88 33.5 47.6 523 16
Stillaguamish 5.2 6.1 84 36.0 44.0 726 15
Lower Mt Hood 1.9 2.2 92 26.0 29.3 483 31
Upper Mt Hood 1.6 1.4 86 19.0 30.4 354 23
Downloaded from by guest on 16 December 2019
7Journal of Forestry, 2019, Vol. XX, No. XX
between measurements in 1931 and 1941. Evidence
for differences in growth among provenances planted
at different sites was weak, except for diameter in
1931 (p= 0.014). Results from the 1931 measurement
are consistent with the earlier ndings of Munger and
Morris (1936). Adaptation is evident for survival, but
not for growth, at least not for those trees that sur-
vived beyond1931.
Signicant relations (p < 0.05) with climate vari-
ables were consistently found at Upper Mt Hood and
Hebo for survival from 1941 to 2013. Regressions of
survival on transfer distances across sites indicate that
the patterns of survival in 1941 and 2013 can be de-
scribed by a quadratic function primarily driven by re-
sponses at Upper Mt Hood and Hebo (Figure 3, Table
4). When provenances from locations with warmer
winters were transferred to the cooler Upper Mt Hood
site, they had lower survival than provenances trans-
ferred from cooler locations, in both 1941 and 2013
(Figure 3A, 3C). When provenances from locations
with colder, more continental climates were transferred
to the warmer, more maritime climate at Hebo, they
had lower survival than provenances transferred from
maritime climates, in both 1941 and 2013 (Figure 3B,
3D). In general, survival declined when provenances
were moved more than 2°C mean coldest month tem-
perature (MCMT) to a colder or warmer winter tem-
perature, or when provenances were moved than more
Table 3. Tests of significance (p-values) for site, provenance, and site × provenance variance components
from BLUP analysis (prob>chi-square for likelihood ratio tests).
Trait and source of variation
1923 1931 1941 1953 1963 1973 1983 1993 2013
Site 0.415 0.581 0.000 0.000 0.002
Provenance 0.094 0.105 0.312 0.254 0.157
Site × provenance 1.000 1.000 1.000 1.000 1.000
Site 0.006 0.040 0.063 0.011 0.005 0.008 0.002 0.002
Provenance 0.844 0.369 0.519 0.986 1.000 1.000 1.000 0.755
Site × provenance 0.014 1.000 1.000 1.000 1.000 1.000 1.000 1.000
Site 0.031 0.016 0.065 0.504
Provenance 0.693 1.000 1.000 0.330
Site × provenance 0.146 1.000 1.000 1.000
Site 0.051 0.031 0.195 0.280
Provenance 0.811 1.000 1.000 1.000
Site × provenance 0.590 0.423 0.480 0.092
Site 0.241 0.031 0.004 0.002 0.004 0.008 0.008 0.043 0.020
Provenance 1.000 1.000 0.371 0.598 0.348 1.000 1.000 1.000 1.000
Site × provenance 1.000 0.937 0.010 0.091 0.037 0.070 0.096 0.080 0.008
Figure 2. Survival over time at the five test sites of the
Douglas-Fir Heredity Study.
Downloaded from by guest on 16 December 2019
8Journal of Forestry, 2019, Vol. XX, No. XX
2°C in continentality (TD) from a more continental
climate (larger TD) to a more maritime climate (smaller
TD) (Figure 3). Although adjusted R2 values were low
for the quadratic models across all sites (0.07–0.20),
R2 values for the within-site regressions for the best
models at Upper Mt Hood and Hebo, which may in-
clude more than one climate variable, were quite high
(0.52–0.79). The importance of the Upper Mt Hood
and Hebo test sites for understanding adaptation and
the effects of seed transfer are reected in the correl-
ations between seed source climates and growth or
survival at each test site (Table 5). Consistent and sig-
nicant correlations (p < 0.05) between provenance
values and the climate of seed sources were found
only at the Upper Mt Hood and Hebo test sites. The
strongest correlations were found for MCMT and TD,
although May–September precipitation was also im-
portant for survival atHebo.
Discussion and Conclusions
Understanding the consequences of a changing cli-
mate, whether from moving populations or from cli-
mate change over time, requires testing climatically
diverse provenances across a wide range of climates.
Fortunately, the Douglas-Fir Heredity Study included
two climatically distinct test sites—Upper Mt Hood
and Hebo. The patterns of provenance survival were
distinctly different between Upper Mt Hood, the
colder high-elevation site, and Hebo, the coastal site
with warmer winters and cooler summers.
The seed source climate variable most closely asso-
ciated with survival at the Mt Hood site was MCMT.
Provenances transferred to sites that were more than
2°C colder suffered greater mortality. For example, if
provenances are moved from locations that are 6°C
colder MCMT, we predict that 100-year survival will
Figure 3. Transfer functions for survival in 1941 and 2013 as a function of climatic transfer distance (site minus provenance)
for mean cold month temperature (MCMT) and continentality (TD).
Downloaded from by guest on 16 December 2019
9Journal of Forestry, 2019, Vol. XX, No. XX
be 16 percent compared to 22 percent for provenances
from a local or similar climate. Thus, given an initial
planting density of 2,200 trees/hectare, low elevation
sources, such as Benton, moved to a high-elevation
site, such as Upper Mt Hood, are expected to have 351
trees/hectare compared to 483 trees/hectare for the
local source (i.e., 27 percent fewer trees). Although cold
damage was not measured during the life of the stand,
the differential survival of provenances was probably
a consequence of maladaptation to cold. Common
garden studies have found a strong relation between
seed source climate and cold damage (Benowicz etal.
2001, Bower and Aitken 2006, St.Clair 2006, Bansal
etal. 2015). Genetic clines associated with cold tem-
peratures have been found in several other Douglas-
r studies (e.g., St.Clair etal. 2005, Leites etal. 2012,
Rehfeldt et al. 2014). Leites et al. (2012), however,
found evidence that local seed sources were not best;
Table 5. Pearson correlation coefficients between provenance values and the climate of source locations for
the Upper Mt Hood and Hebo testsites.
Upper Mt Hood
Diameter 1931 –.82 –.80 –.70 .49 .64 .51
Survival 1931 –.45 –.28 –.62 –.18 .58 .64
Survival 1941 –.77 –.88 –.53 .77 .67 .47
Survival 1953 –.68 –.78 –.46 .68 .50 .38
Survival 1963 –.70 –.77 –.49 .64 .51 .39
Survival 1973 –.67 –.77 –.42 .70 .48 .31
Survival 1983 –.68 –.78 –.43 .72 .49 .30
Survival 1993 –.67 –.80 –.37 .79 .48 .22
Survival 2013 –.57 –.75 –.24 .85 .41 .14
Volume/hectare 2013 –.56 –.73 –.23 .83 .37 .05
Diameter 1931 .33 .39 .14 –.43 –.07 –.06
Survival 1931 .37 .36 .27 –.26 –.27 –.12
Survival 1941 .16 .40 –.16 –.75 –.07 .28
Survival 1953 .15 .40 –.18 –.76 –.03 .35
Survival 1963 .13 .31 –.18 –.63 .18 .57
Survival 1973 .15 .35 –.16 –.67 .14 .50
Survival 1983 .16 .36 –.16 –.69 .15 .49
Survival 1993 .23 .43 –.09 –.71 .09 .40
Survival 2013 .12 .30 –.17 –.62 .21 .47
Volume/hectare 2013 .08 .26 –.15 –.54 .17 .36
Note: |r|>.55 is statistically signicant at p=0.05 (N=13)|. MAP, mean annual precipitation; MAT, mean annual temperature;
MCMT, mean coldest month temperature; MSP, mean summer precipitation; MWMT, mean warmest month temperature; TD,
continentality as determined by the difference between MWMT and MCMT.
Table 4. Regression equations for survival (Yij) of the ith provenance at the jth site as a function of transfer
distance (X) for mean cold month temperature (MCMT) and continentality (TD) using the quadratic model
2P>|t| for β
2Multiple R2
1941 78.31 –.70 –.31 0.0017 .15
2013 21.69 –.41 –.22 0.0014 .17
1941 78.28 –.30 –.23 0.0296 .07
2013 22.02 –.50 –.26 0.0003 .20
Downloaded from by guest on 16 December 2019
10 Journal of Forestry, 2019, Vol. XX, No. XX
seedlings from warmer environments had better height
growth than local sources when grown in cold environ-
ments. The differences with our study, however, may be
attributed to evidence for adaptation as measured by
survival over decades as compared to seedling growth.
The climate variable that was most closely as-
sociated with survival at the Hebo plantation was
continentality (TD). Provenances transferred between
locations differing by more than 2°C in continentality
are expected to suffer higher mortality than proven-
ances transferred from a local or similar climate (Figure
3). This difference is equivalent to moving provenances
from the east end of the Columbia Gorge to the Hebo
site in the Coast Range. For example, we predict that
moving provenances 5.5°C in continentality to a more
maritime climate, such as from Wind River to Hebo,
would lead to 17 percent survival compared to 22
percent for the local sources after 100 years. This is
equivalent to 373 trees/hectare compared to 483 trees/
hectare (i.e., 23 percent fewer trees).
A clue as to the cause of differential mortality at Hebo
is found in a Forest Service report from 1942 (Munger
and Morris 1942). In the report, it was noted that
Rhabdocline needle disease was found on a large propor-
tion of the trees at the Hebo plantation in 1938. The au-
thors concluded that trees from some seed sources were
more susceptible to the disease, although they did not in-
dicate which seed sources. The presence of Rhabdocline
needle disease in 1938 is consistent with our results that
the provenance × site interaction for survival becomes
large between 1931 and 1941. A more recent study
found that moving Douglas-r seed sources more than
3°C in continentality, from a continental to a maritime
climate, increased the probability of Rhabdocline infec-
tion by more than 25 percent (Wilhelmi et al. 2017).
Thus, the effects of Rhabdocline may explain the associ-
ations between survival and TD. Wilhelmi etal. (2017)
also found that Rhabdocline disease increased when seed
sources were moved to locations with more summer
precipitation and warmer winter temperatures, which
is consistent with our ndings (Figure 3). In general,
Rhabdocline disease has not been a problem in Douglas-
r plantations, probably because the use of seed zones
has limited long-distance seed transfers. Wilhelmi etal.
(2017) concluded that Rhabdocline disease was pri-
marily associated with long-distance transfers from the
continental California Sierra and Klamath Mountains to
the maritime areas of western Oregon and Washington.
Our long-term results suggest that transfers from the
Cascades to the Coast Range, that is, transfers as short
as 160 km, may also be a concern.
Early results from the Douglas-Fir Heredity Study,
and anecdotal evidence of maladapted plantations,
prompted the development of seed transfer guidelines
and seed zones in the Pacic Northwest. In particular,
the seed zones developed for Oregon and Washington
in 1966 have been widely accepted and used (Randall
1996). The Oregon/Washington seed zones are de-
lineated as 152-m elevation bands intersected with a
geographic delineation. The elevation bands are a crit-
ical component of seed zones because of the strong
relation between elevation and cold temperatures.
Using GIS, we characterized the climatic width of
the Oregon/Washington seed zones to compare with
the transfer functions from the Douglas-Fir Heredity
Study. The average climatic width for winter temper-
atures (MCMT) of the 673 seed zones/elevation bands
in western Oregon and Washington (dened as west
of the Cascade crest) is 2.0°C. However, there is con-
siderable variation among seed zones, with 5 percent
of them exceeding a climatic width of 4.1° C. Thus,
based on experience from seed zones, managers may
feel condent that they can move seed sources up to
2° C MCMT, and perhaps as much as 4° C. This is
consistent with our ndings. When considering move-
ments from a continental to a maritime climate, the
average climatic width of TD within seed zones in
western Oregon and Washington is 2.0° C, which,
again, is close to the acceptable transfer distance that
we observed at the Hebo site. Managers have been
using these seed zones for more than a half century,
and the consensus is that they have been effective in
ensuring adapted, healthy, and productive plantations.
However, as climates start to warm beyond 2°C, the
adage that “local is best” may no longer be true, and
resource managers are beginning to reconsider the use
of local seedzones.
One conclusion by early researchers of both the
Douglas-Fir Heredity Study and the Wind River
Arboretum is that evidence for maladaptation may
take time. Our results support that conclusion be-
cause maladaptation, as measured by differences
in survival, was not evident until two decades after
planting. Local adaptation, as indicated by differen-
tial survival of provenances, rst became apparent
in 1941 at ages 28 and 29 (Table 3). Between 1931
and 1941, survival dropped from 86 percent to 62
percent at the Upper Mt Hood plantation (Figure 2).
Cold events in the winters of 1930 and 1937 may
have contributed to this mortality and the differences
among provenances. For example, MCMT values in
those years were –9.4° C and –8.8° C, respectively,
Downloaded from by guest on 16 December 2019
11Journal of Forestry, 2019, Vol. XX, No. XX
compared to an average for the 1930s of –2.2° C
(data derived from ClimateNA). In western Oregon
and Washington, a particularly well-documented cold
event occurred in November 1955, when there was a
week-long cold wave after an unusually mild October
(Dufeld 1956). Mortality over the next few years
was high across much of the region, mostly from
cambial damage (Childs 1961). The Hebo plantation
seemed to be particularly affected (Figure 2)—mor-
tality increased between 1953 and 1963, and frost
cracks were observed in the dead trees (Roy Silen,
pers. obs.). The Hebo site may have been particularly
affected by the cold wave because the mild environ-
ment at the site would have resulted in lesser cold ac-
climation (see Bansal etal. 2015). However, the 1955
cold event appears to have affected all provenances
equally since the provenance × site interaction, and
the relative rankings for survival were unchanged.
Instead, differences among provenances in maladap-
tation at the Hebo site may have arisen earlier as trees
gradually died because of Rhabdocline needle disease.
Although maladaptation may not become ap-
parent for decades in provenance tests, short-term
genecology studies may shed light on patterns of vari-
ation in adaptive traits such as phenology, cold hardi-
ness, drought resistance, and growth potential (e.g.,
St.Clair etal. 2005, Rehfeldt et al. 2014, and studies
cited therein). Such short-term studies may be valu-
able for delineating seed zones and breeding zones, but
long-term eld studies are needed to inform managers
about the long-term consequences of different transfer
limits and evidence for local adaptation.
The Douglas-Fir Heredity Study is one of the
longest-running forestry studies in the North America
and perhaps the world. What have we learned over
the past 100 years? First, experimental design and
statistics have evolved over the last century. Since the
establishment of the study, eld tests routinely use ran-
domization and replication with blocking. More re-
cently, new analytical approaches—mixed models and
BLUPs—allow better estimates of different sources
of variation of interest to the researcher. These new
approaches allow us to revisit older studies to better
evaluate variation because of provenances, test sites,
and their interactions. This study also highlights the
value of sampling across a large climatic range for both
plantations and provenances. We would have learned
very little from this study without the inclusion of the
warmest and coldest test sites. Provenance tests should
emphasize a wide sampling of test sites and proven-
ances across climatic gradients of adaptive signicance
(e.g., cold temperatures, aridity, and continentality).
Awell-designed study with respect to climate is more
efcient and allows a better determination of response
and transfer functions (Wang etal. 2010).
Second, the use of conservative seed zones, breeding
zones, and seed transfer guidelines has probably in-
creased plantation survival and productivity compared
to plantations established during the rst half of the
20th century. These seed zones and breeding zones
probably kept operational seed transfers within the
transfer limits we identied. However, clinal variation
and large provenance variation around the transfer
function, combined with large genetic variation within
populations, may indicate some lost opportunities to
select and deploy genetic material over a larger area
with accompanying economies ofscale.
Third, although the early researchers established the
Douglas-Fir Heredity Study with different objectives in
mind, the study has proven useful for evaluating new
objectives associated with reforestation and climate
change. Initial results did not shed much light on ques-
tions of which trees should be left as leave trees or from
which trees to collect for reforestation. The study did,
however, point to concerns about moving seed sources
between very different elevations, contributing to the
development of seed zones and seed movement guide-
lines. The study also proved valuable for promoting
early tree improvement programs by demonstrating
differences among parent trees within provenances
(Roy Silen, pers. commun.). This current analysis of
the results of the study points to implications for the
adaptation of native stands to climate change, and
possible management options for responding to con-
cerns. Results indicate that Douglas-r populations are
adapted to the local climates that they have experienced
over the past century. This suggests that climate change
may not be a big problem if the amount of climate
change is within 2°C, and extreme climatic events do
not occur with frequency. To date, climate change has
not exceeded 2°C for most of North America; how-
ever, the climate is expected to be strikingly warmer by
midcentury. This study considered climatic transfers to
warmer climates only up to about 3°C (MCMT) ac-
companied by a transfer from a continental to a more
maritime climate, resulting in decreased survival and
productivity. New studies may be required to explore
greater climatic transfer distances with and without a
concomitant change in continentality. If climate change
exceeds a 2–3°C increase, moving populations from
warmer to cooler locations to be adapted to future cli-
mates would hold promise for responding to concerns
Downloaded from by guest on 16 December 2019
12 Journal of Forestry, 2019, Vol. XX, No. XX
of maladaptation. However, managers should not
move populations beyond 2–3° C to cooler climates
to avoid risk of cold damage in the near-term. Finally,
considerable variation may be found among and within
populations. This means that there is some potential
for natural selection (e.g., Warwell and Shaw 2019)
(as well as human selection within tree-improvement
programs), but that depends on generation turnover
and the establishment of new stands. One management
option to take advantage of genetic variation is to use
mixtures of provenances to allow for natural selection
and human selection by thinning.
An early publication from 1917 describing the es-
tablishment of the Douglas-Fir Heredity Study pro-
moted the value of such long-term studies (Kraebel
1917). The author states: “The imagination refuses to
venture concerning the methods of study at so distant a
time. The largeness of the idea is at once gratifying and
disturbing, for one feels both the importance of the
work and the responsibility of doing rightly the early
steps in that work, lest the initial errors and omissions
grow in magnitude with the advancing years. We can
imagine that the early researchers never envisioned the
results from this study informing management of for-
ests in response to climate change. This article is a tes-
tament to their long-term vision.
As might be expected after a hundred years, many individ-
uals contributed to the conception, establishment, measure-
ments, and maintenance of this study. In particular, Thornton
T.Munger conceived and initiated the study, and published
the rst comprehensive results. Many helped with seed col-
lection and seedling production, including C.P. Willis, C.J.
Kraebel, A.A. Grifn, E.Hanzlik, and C.R. Tollotson. Willis
and Kraebel published early ndings on cone collection and
seed characteristics in 1915 and 1917. J.V. Hofmann was re-
sponsible for establishment and early measurements as dir-
ector of the Wind River Experiment Station from 1913 to
1924. Others contributed as well, including W.G. Morris,
L.A. Isaac, H.V. Brown, E.L. Kolbe, and A.G. Simpson.
Special acknowledgment goes to Roy Silen, who ensured the
study was maintained and measured for much of the last half
century. We thank Jeff Riddle for overseeing measurements
in 1993 and 2013, and Jeff Riddle and Nancy Mandel for
data management and earlier statistical analyses.
Bansal, A., J.B. St.Clair, C.A. Harrington, and P.J.Gould.
2015. Impact of climate change on cold hardiness of
Douglas-r (Pseudotsuga menziesii): Environmental
and genetic considerations. Glob. Change Biol.
Benowicz,A., S.J.L’Hirondelle, and Y.A.El-Kassaby. 2001.
Patterns of genetic variation in mountain hemlock (Tsuga
mertensiana (Bong.) Carr.) with respect to height growth
and frost hardiness. For. Ecol. Manag. 154:23–33.
Bower, A.D. and S.N. Aitken. 2006. Geographic and sea-
sonal variation in cold hardiness of whitebark pine. Can.
J.For. Res. 36:1842–1850.
Childs,T.W. 1961. Delayed killing of young-growth Douglas-
r by sudden cold. J. For. 59(6):452–453.
Dufeld, J.W. 1956. Damage to western Washington for-
ests from November 1955 cold wave. Research Notes.
USDA Forest Service Pacic Northwest Forest and Range
Experiment Station Research Note 129, Portland, OR. 8 p.
Isaac,L.A. 1949. Better Douglas Fir forests from better seed.
University of Washington Press, Seattle, WA. 65 p.
Kraebel,C.J. 1917. Choosing the best tree seeds. J. Hered.
Langlet, O. 1971. Two hundred years genecology. Taxon
Leites,L.P., A.P. Robinson, G.E.Rehfeldt, J.D.Marshall, and
N.L.Crookston. 2012. Height-growth response to climatic
changes differs among populations of Douglas-r: Anovel
analysis of historic data. Ecol. Appl. 22(1):154–165.
Munger,T.T., and W.G.Morris. 1936. Growth of Douglas-
r trees of known seed source. USDA Technical Bulletin
No. 537, Washington, DC. 40 p.
Munger,T.T., and W.G.Morris. 1942. Further data on the
growth of Douglas-r trees of known seed source. RS-
NW Genetics Racial Variation Douglas-Fir, USDA Forest
Service Pacic Northwest Forest and Range Experiment
Station, Portland, OR. 12 p.
Poudel, K.P. and T. Hailemariam. 2016. Calibration of
volume and component biomass equations for Douglas-
r and lodgepole pine in Western Oregon forests. Forest.
Chron. 92(2):172–182.
Randall, W.K. 1996. Forest tree seed zones for western
Oregon. Oregon Department of Forestry, Salem, OR.
Rehfeldt, G.E., L.P. Leites, J.B. St.Clair, B.C. Jaquish,
C.Sáenz-Romero, J.López-Upton, and D.G.Joyce. 2014.
Comparative genetic responses to climate in the varieties
of Pinus ponderosa and Pseudotsuga menziesii: Clines in
growth potential. For. Ecol. Manag. 324:138–146.
Sandler,I., and L.Sandler. 1986. On the origin of Mendelian
genetics. Am. Zool. 26:753–768.
SAS Institute Inc. 2014. SAS/STAT® 13.2 User’s guide. SAS
Institute Inc., Cary, NC.
Silen, R.R. 1963. The 1912 Douglas-r heredity study.
USDA Forest Service Pacic Northwest Forest and Range
Experiment Station Annual Report, Portland, OR.
Silen, R.R. 1965. Regeneration aspects of the 50-year-old
Douglas-r heredity study. Proceedings 1964 Annual
Meeting of the Western Reforestation Coordinating
Committee, Spokane, WA.
Downloaded from by guest on 16 December 2019
13Journal of Forestry, 2019, Vol. XX, No. XX
Silen, R.R. 1966. A 50-year racial study of Douglas-r in
western Oregon and Washington. Proceedings of the 1965
Annual Meeting of the Western Forest Genetics Association,
Olympia, WA.
Silen,R.R., and D.L.Olson. 1992. A pioneering exotic tree
search for the Douglas-r region. USDA Forest Service
Pacic Northwest Research Station Gen. Tech. Rep.
PNW-GTR-298, 44p.
St.Clair,J.B. 2006. Genetic variation in fall cold hardiness in
coastal Douglas-r in western Oregon and Washington.
Can. J.Bot. 84:1110–1121.
St.Clair, J.B., N.L. Mandel, and K. Vance-Borland. 2005.
Genecology of Douglas r in western Oregon and
Washington. Ann. Bot. 96:1199–1214.
Wang.T., A.Hamann, D.Spittlehouse, and C.Carroll. 2016.
Locally downscaled and spatially customizable climate
data for historical and future periods for North America.
PLoS One 11(6):e0156720.
Wang,T., G.A.O’Neill, and S.N.Aitken. 2010. Integrating
environmental and genetic effects to predict responses of
tree populations to climate. Ecol. Appl. 20:153–163.
Warwell,M.V., and R.G.Shaw. 2019. Phenotypic selection
on ponderosa pine seed and seedling traits in the eld
under three experimentally manipulated drought treat-
ments. Evol. Appl. 12:159–174.
White,T.L. 1996. Genetic parameter estimates and breeding
value predictions: Issues and implications in tree improve-
ment programs. P. 110–117 in Tree improvement for sus-
tainable tropical forestry, Dieters,M.J., A.C.Matheson,
D.G. Nikles, C.E. Harwood, and S.M. Walker (eds.).
QFRI IUFRO Conference, Caloundra, Queensland,
Wilhelmi, N.P., D.C.Shaw, C.A. Harrington, J.B. St.Clair,
and L.M. Ganio. 2017. Climate of seed source affects
susceptibility of coastal Douglas-r to foliage diseases.
Ecosphere 8(12):e02011.
Downloaded from by guest on 16 December 2019
... at the Earth's surface than any preceding decade since 1850 (IPCC 2013). One might therefore expect to see a growing evolutionary lag in provenance trials established in the early versus late 20th century. Indeed, survival transfer functions from the Douglas-fir Heredity Study established in 1912 showed negligible evolutionary lag (see figure 3 in St. Clair et al. 2020), whereas those for white spruce established between 1972 and 1976 show a distinct lag (Lu et al. 2014), although results in recent trials vary considerably with response trait and test site location within the species distribution (Savolainen et al. 2007;O'Neill et al. 2008a;Berlin et al. 2016;Leites et al. 2012a). As climates change, p ...
... ese four climate variables, which include three temperature and one precipitation variable, were selected because they have been repeatedly identified as representative drivers of population differentiation in North American tree genecology analyses (Parker and van Niejenhuis 1996;Andalo et al. 2005;Ukrainetz et al. 2011;Russell and Krakowski 2012;St. Clair et al. 2020) and because they are relatively uncorrelated. ...
... we selected distances of 2.3°C for MAT, 3.0°C for TD, and 50% for both MAP and DD5, values interpreted from published transfer functions for these, or closely correlated, climate variables for a variety of western North American conifer species (Rehfeldt et al. 1999(Rehfeldt et al. , 2003O'Neill et al. 2008bO'Neill et al. , 2014Leites et al. 2012b;St. Clair et al. 2020). ...
Full-text available
As climate change accelerates, foresters are looking to ever warmer climates to secure sources of climatically adapted tree seed with which to establish healthy and productive plantations. However, as seed procurement areas approach jurisdictional boundaries (states, provinces, nations), across which seed and seed transfer systems are not typically shared, innovative approaches are required to identify those plantation areas for which suitable domestic provenances will be lacking, and areas in neighbouring jurisdictions with matching warmer, future climates that could fill domestic seed supply gaps. We describe a straightforward, climate envelope approach to locate these areas, using British Columbia (BC), Canada, and the Pacific Northwest (PNW) USA to illustrate the analysis. We find that 21% of BC’s ecosystems (seed zones) will be at moderate or high risk of lacking adapted domestic provenances for plantation establishment by 2040. Importantly, however, we find large areas in the PNW that should be able to fill most of BC’s domestic seed supply gaps. Spatial analyses of this type will inform seed suppliers, managers and policymakers where alternative seed procurement arrangements are needed and underscore the operational and policy barriers to acquiring seed from warmer jurisdictions. More broadly they also highlight the need for inter-jurisdictional cooperation in matters pertaining to resource management.
... As a result, local seed sources are usually recommended for reforestation and restoration. This has typically been accomplished using geographically defined zones or seed transfer rules that limit the geographic or climatic distance that populations from specific locations (also called seed sources) may be moved (Johnson et al., 2004;St.Clair et al., 2020). We now have overwhelming evidence, however, that the climate is warming with an increase in global surface temperature to date of 0.82 C above the 20th century average and a projected future warming of 0.18 C per decade (NOAA, 2021). ...
... In forest trees, provenance tests with many populations sampling a wide range of environments may approximate reciprocal-transplant studies and allow evaluation of local adaptation as a function of the climatic or geographic transfer distance between seed sources and planting sites. Although results from long-term tests using adequate numbers of populations and test sites tested over a large climatic range are few, studies indicate that forest trees are, in general, locally adapted, except perhaps at the northern margins of the species ranges where they may show adaptive lag (Rehfeldt et al., 1999;Savolainen et al., 2007;St.Clair et al., 2020;Wang et al., 2006). ...
... Seedlings are particularly vulnerable to stresses from heat, moisture deficit, and cold. Multiple, repeated stresses over the life of a stand may lead to the same vulnerabilities at later ages, as well as impacts from diseases and insects (St.Clair et al., 2020). What is certain is that populations must be adapted to the near-term climate to survive to be adapted to the long-term climate. ...
Full-text available
The Seedlot Selection Tool and Climate‐Smart Restoration Tool are web‐based tools designed to match seedlots with planting sites assuming that seedlots are adapted to the past climates in which they evolved, primarily with respect to temperature and aridity. The tools map the climatic match of seedlots with the past or projected climates of planting sites. The challenge is that future climates are a moving target, which means that seedlots must be adapted to the near‐term climates as well as the climates of the mid‐ to late‐21st century. Because climate projections are uncertain, the prudent approach is to aim for the warmest climate that may be expected while ensuring that seedlots moved from warmer to colder locales are not moved so far that they risk cold damage. Uncertainty in climate projections may be mitigated by ensuring genetic diversity through mixing seed sources and having collections from many parents per seed source. Three examples illustrate how to effectively use the web tools: (1) choosing seedlots targeting different future climates for a mid‐elevation Douglas‐fir site in the Washington Cascades, (2) finding current and future seed sources for restoration of big sagebrush after fires in the Great Basin and Snake River Plain, and (3) planning to ensure that a Douglas‐fir seed inventory includes seedlots suitable for future climates in western Oregon and Washington.
... version 6.22 [57]. Mean coldest month temperature was selected because it is frequently identified as being among the most important of climate variables in accounting for population differentiation in Picea glauca × engelmannii [51] and other North American temperate conifer species [58,59]. ...
... These analyses corroborate a growing body of literature suggesting that assisted migration can help forestall some of the negative impacts of climate change on forest productivity [47,59]. In lodgepole pine (Pinus contorta), where the results of long-term provenance trials have been widely examined, assisted migration is also expected to have a positive impact on productivity [46]. ...
... Thus, at least for the case of this shade-tolerant Abies religiosa species, assisted migration must be protected assisted migration, by using nurse plants as a companion for the target migrated species [65]. This finding is consistent with that of St. Clair, Howe, and Kling [59] who found that Douglas-fir seed sources moved to sites 2 • C cooler than their origin still retained acceptable long-term survival and productivity. ...
Full-text available
Assisted migration of forest tree populations through reforestation and restoration is a climate change adaptation strategy under consideration in many jurisdictions. Matching climates in which seed sources evolved with near future climates projected for plantation sites should help reduce maladaptation and increase plantation health and productivity. For threatened tree species, assisted migration outside of the species range could help avert extinction. Here, we examine lessons, limitations, and challenges of assisted migration through the lens of three assisted migration field trials of conifers in Canada and Mexico: Pinus albicaulis Engelm., an endangered subalpine tree species in the mountains of western North America; the Picea glauca (Moench) Voss × P. engelmannii Parry ex Engelm hybrid complex, of great economic and ecological importance in western Canada, and Abies religiosa (Kunth) Schltdl. & Cham., a tree species that provides overwintering sites for the monarch butterfly. We conclude that: (a) negative impacts of climate change on productivity of Picea glauca × P. engelmannii may be mitigated by planting seed sources from locations that are 3 °C mean coldest month temperature warmer than the plantation; (b) it is possible to establish Pinus albicaulis outside of its current natural distribution at sites that have climates that are within the species’ modelled historic climatic niche, although developing disease-resistant trees through selective breeding is a higher priority in the short term; (c) Abies religiosa performs well when moved 400 m upward in elevation and local shrubs (such as Baccharis conferta Kunth) are used as nurse plants; (d) new assisted migration field trials that contain populations from a wide range of climates tested in multiple disparate climates are needed, despite the costs; and (e) where naturalization of a migrated tree species in recipient ecosystem is viewed as undesirable, the invasive potential of the tree species should be assessed prior to large scale establishment, and stands should be monitored regularly following establishment.
... s, which sometimes account for test site effects in mixed model analyses, as here, and sometimes by standardizing population means as percentages of test site means (Andalo et al., 2005;Carter, 1996;Leites et al., 2012;Matyas & Yeatman, 1992;O'Neill & Nigh, 2011;Rehfeldt et al., 2003;Rehfeldt et al.,1999;Sáenz-Romero et al., 2017;Schmidtling, 1994;St. Clair et al., 2020;Thomson et al., 2009). As stated earlier, the focus of interest in such studies has usually been the models themselves and what they imply about the opportunity to increase forest productivity by planting non-local seedlings or the potential to mitigate climatic change through assisted migration. Here, instead, we wish to draw attention ...
... esults are qualitatively valid for 50+ years as there was little evidence of imminent mortality at the time of measurement. To our knowledge, the only longer study in which survival was modelled as a function of climatic displacement is a 100-year test of Douglas-fir populations moved to several test sites in the Pacific Northwest of North America (St. Clair et al., 2020). In that study, displacements (−2.2°C to +5.6°C MAT) accounted for no more than ~6% mortality over study-wide background levels. As in the current study, climatic displacement caused marginal increases in mortality, but all populations survived in reduced numbers. These studies provide experimental evidence that trees are much more capa ...
Full-text available
Aim Based upon species distribution models (SDMs), many studies have predicted that climate change will cause regional extinctions of tree species within the next 50–100 years. SDM-based predictions have been challenged on procedural and theoretical grounds, but for tree species they are largely beyond the practical reach of direct experimental validation. Here we report the experimental consequences of moving seedlings from ~50 natural populations of each of two ash (Fraxinus) species to experimental sites spanning a range of 10°C colder to 10°C warmer (mean annual temperature) than home environments. Location Eastern North America. Methods We measured population-by-test-site survival percentages and mean trunk diameters at an average age of 35 years. We then used linear, mixed-effect models to develop transfer functions for each species and predict survival and mean annual growth as functions of fixed and random effects including, especially, the climatic distance (CD) between test site and home environment. Results Survival and growth were highest at CD ≈ 0 and declined as populations were moved to warmer or colder environments, indicating that survival and growth were optimal when populations were in home-like climates. However, variance around the model fit was substantial, and we could not statistically detect, even at α = .50, elevated mortality following displacements into environments 3.5°C (white ash) and 4.1°C (green ash) warmer in mean annual temperature. Survival rates of 80%–100% were common even within populations subjected to warming conditions greater than those predicted to cause meso-scale extinctions in this century. We show that within-population genetic variance, phenotypic plasticity and idiosyncratic aspects of the non-climatic environment and its interaction with genotype each likely contributed to these unexpected responses to climatic displacement. Main conclusions Results emphasize the uncertainty that underlies predictions of climate-induced extinctions of long-lived woody plants over time frames of 50 to perhaps 100 years into the future.
... or a transformation of timber markets that reward other tree species besides Douglas-fir, there will be little gain from extensive margin adaptation across forest types in the timber sector. One active area of adaptation research on the intensive margin for Douglas-fir is assisted relocation, where seeds are actively moved across climate gradients (St. Clair et al., 2020). ...
Full-text available
This study develops a method to estimate the welfare impacts of climate change on landowners using a discrete-choice econometric model of land management. We apply the method to forest management in the Pacific states of the U.S. and estimate welfare effects on the region that holds the largest current commercial value – western Oregon and Washington. We find evidence that a warmer and drier climate will induce an approximate 39% loss in the economic value of timberland by 2050, though there is heterogeneity across space. The discrete-choice approach allows us to determine that the welfare losses are primarily driven by estimated losses to Douglas-fir, the most commercially valuable species. An alternative approach to welfare analysis from climate change is the Ricardian method, which gives conceptually similar estimates to the discrete-choice method. While we find similar empirical findings between the discrete-choice and Ricardian approaches, the discrete-choice approach provides more heterogeneity and somewhat larger negative welfare impacts. Our analysis is notable for providing the first empirical evidence that climate change can induce welfare losses to timberland owners, even while accounting for optimal adaptation.
... the term was coined.The first seed transfer zones in North America were developed for Douglas-fir (Psuedotsuga menziesii) from a heredity study started in 1912. Variable growth observed from different populations spurred the US Forest Service to create general seed transfer guidelines in 1939 and specific seed transfer zones for Douglas-fir in 1942(St. Clair et al., 2020). These guidelines were established prior to the broader recognition of climate change. Today, most commercially important tree species (e.g.,Rehfeldt et al., 2014), many ecologically important shrubs(Richardson & Chaney, 2018), grasses ...
Full-text available
The rate of global climate change is projected to outpace the ability of many natural populations and species to adapt. Assisted migration (AM), which is defined as the managed movement of climate-adapted individuals within or outside the species ranges, is a conservation option to improve species’ adaptive capacity and facilitate persistence. Although conservation biologists have long been using genetic tools to increase or maintain diversity of natural populations, genomics techniques could add extra benefit in AM that include selectively neutral and adaptive regions of the genome. In this review, we first propose a framework along with detailed procedures to aid collaboration among scientists, agencies, local and regional managers during the decision-making process of genomics-guided AM. We then summarize the genomic approaches for applying AM, followed by a literature search of existing incorporation of genomics in AM across taxa. Our literature search initially identified 729 publications, but after filtering returned only 50 empirical studies that were either directly applied or considered genomics in AM related to climate change across taxa of plants, terrestrial animals, and aquatic animals; 42 studies were in plants. This demonstrated limited application of genomic methods in AM in organisms other than plants, so we provide further case studies as two examples to demonstrate the negative impact of climate change on non-model species and how genomics could be applied in AM. With the rapidly developing sequencing technology and accumulating genomic data, we expect to see more successful applications of genomics in AM, and more broadly, in the conservation of biodiversity.
... or a transformation of timber markets that reward other tree species besides Douglas-fir, there will be little gain from extensive margin adaptation across forest types in the timber sector. One active area of adaptation research on the intensive margin for Douglas-fir is assisted relocation, where seeds are actively moved across climate gradients (St. Clair et al., 2020). ...
... Due to the importance of Df for forestry, the genetic variation in this species has been investigated in numerous provenance trials in North America, Europe and elsewhere. Having evolved in highly heterogeneous environments, Df exhibits large intraspecific variation in many phenotypic traits (St Clair and Howe 2007;Krakowski and Stoehr 2009;Leites et al. 2012;Rehfeldt et al. 2014;Bansal et al. 2015;St Clair et al. 2020). Thus, the response to specific climatic conditions, when moved to new regions, may vary among populations. ...
Full-text available
The contribution of Douglas-fir (Df) to European forests is likely to increase as the species is a potential adaptation option to climate change. In this study, we investigated growth and survival of Df seed sources to fill a knowledge gap regarding recommendations for the future use of Df provenances in Poland. Our experimental test site represents the most continental climate among all Df trials installed in the IUFRO 1966-67 test series in Europe. At this unique single site, we evaluated the performance of 46 Df provenances from North America, and nine local landraces of unknown origin. Repeated measurements of tree diameter, height, and volume were analysed, to age 48, representing integrated responses to geographic and climatic conditions. Significant variation in survival and productivity-related traits were found, with the interior Df provenances performing best, in contrast to previous European reports. The higher survivability and volume of the interior provenances resulted from their superior frost resistance. The low precipitation seasonality at the location of seed origin provided an additional advantage to the trees at the test site. Geographic and climatic factors of seed origin explained most of the variation in productivity (77 and 64%, respectively). The tested landraces exhibited diverse performance, implying that naturalized local seed sources in Poland need improvement and perhaps enrichment with new genetic material from North America, while considering geography and climate. Assisted migration programs should consider the limitations imposed by both frost and drought events in guiding future Df selections for continental climates. Further field testing, early greenhouse screening and DNA testing are also recommended.
The private forestry sector faces enormous challenges from complex environmental issues concurrent with societal concerns about intensive forest management and application of silvicultural tools. At the same time, research and development spending and the scientific workforce in the forestry sector has declined. However, the forestry sector has a long history of strategic deployment of science and technology to manage timberlands for many uses, including ecological services. To address science and societal needs from future forests, we describe past successes and potential future innovations of the forestry sector across (1) genetics, (2) silviculture and productivity, (3) harvesting and logging technology, (4) environmental sustainability, and (5) remote sensing and unmanned aerial vehicles. Developing technology is only one component, however; progressing towards sector goals of diversifying the workforce, explicitly valuing research collaborations, and integrating big data across ownerships to guide management decisions will hasten innovation. With the growing complexity of social-ecological-economic challenges, the private forestry sector must enact a tactical approach to addressing stressors with evolving research structures. We offer a vision of a vibrant private forestry sector poised to integrate technological innovation to continue meeting society’s needs through the intersecting effects of climate change and other challenges. Study Implications The complexity of challenges facing private forest managers is growing as stressors from climate change interact with social and economic pressures on forests. Further, spending on forestry research and development and student enrollment in forestry degree programs has declined. In this article, we describe significant technological improvements in regenerating, growing, and sustainably harvesting intensively managed forest through a changing research landscape. We review key past examples of production and sustainability improvements and describe future innovation space across the intensively managed conifer stands through the life cycle of seed-to-saw. We envision that technological improvements in forest genetics, remotely operated harvesting equipment, high-resolution information about vegetation and the earth’s surface, and analysis of big data will increase the private sector’s ability to make precise management decisions. A more developed and broader technological tool set will ensure the private forest sector is poised to supply ecosystem services and demand for wood products. To realize the gains from future innovation, the forest sector needs to support initiatives that explicitly value an inclusive and diverse workforce, cross-disciplinary problem-solving, and improved communication with external stakeholders.
It is particularly important to select the provenance with better adaptability in the context of the counteracting potential adverse effects of climate change. However, the growth environment of trees and the length of breeding cycle determine the uncertainty of early selection. Thereby, it is advisable to provide information for provenance selection by analyzing the relationship between young, middle and mature forests. Chinese fir(Cunninghamia lanceolata (Lamb.) Hook.) is an evergreen coniferous tree species with rapid growth and high-quality timber, and it is an important reforestation and commercial tree species in South China. The provenance experimental forests of Chinese fir in Jiangxi and Guangxi were taken as the test object, the growth survey went by three stages of growth: young, middle and mature ages. The linear mixed model was used to evaluate the differentiation of different forest stage provenances in terms of growth. K-medoids clustering algorithm was used to group provenances, the geographical distribution characteristics of Chinese fir were evaluated based on the major groups, and the risk of early selection and the genetic gain of excellent provenances were calculated. The highest total variability was explained by the genotype effect (G) in the young, middle-aged and mature forest stage, the influence of genotype increased through time, while the influence of site was reduced and the G×E interaction changed in a small range. The differences in heritability were similar between young and middle-aged forests, but the heritability of growth traits was higher at Jiangxi than Guangxi in the mature forest stage. The provenances with excellent productivity Chinese fir showed aggregation distribution at the mature forest stage, and the provenances with worst growth were mainly distributed in the northern marginal production area.
Full-text available
Drought‐related selection during seedling emergence and early development may play a strong role in adaptation. Yet this process is poorly understood and particularly so in relation to ongoing climate change. To evaluate drought‐induced differences in selection during early life stages, a total of 50 maternal families sampled from three climatically disparate ponderosa pine (Pinus ponderosa Doug.) populations were grown from seed in two common garden field experiments at a location that was warmer and drier than seed origins. Three drought treatments were imposed experimentally. Phenotypic selection was assessed by relating plant fitness measured as survival or unconditional expected height at age 3 to seed density (mass per unit volume), date of emergence, and timing of shoot elongation. In the year of emergence from seed, differential mortality was particularly strong and clearly indicated selection. In contrast, selection in subsequent years was far less pronounced. Phenotypes with high seed density, an intermediate but relatively early emergence date and high 2nd year early‐season shoot elongation exhibited the greatest estimated fitness under drought. The form of selection varied among seed sources in relation to drought treatment. Selection was generally more acute in the cases of greatest difference between drought treatment and climatic patterns of precipitation at the site of seed origin. These results suggest that populations of ponderosa pine are differentially adapted to drought patterns associated with the climate of their origin. To the extent that the phenotypic traits examined are heritable or correlated with heritable traits, our results provide insight into how tree populations may evolve in response to drought. This article is protected by copyright. All rights reserved.
Full-text available
Seed-source movement trials using common garden experiments are needed to understand climate, tree (host), and pathogen interactions. Douglas-fir (Pseudotsuga menziesii var menziesii) is an important tree species native to western North America influenced by the foliar fungi Phaeocryptopus gaeumannii, a biotroph and causal agent of Swiss needle cast (SNC), and Rhabdocline species, necrotrophs that cause Rhabdocline needle cast. We used the Douglas-fir Seed-Source Movement Trial, a large provenance study of Douglas-fir that consists of populations and test sites chosen to represent the range of climate conditions experienced by Douglas-fir west of the Cascade and northern Sierra Nevada Mountains, USA, to assess disease severity and symptom expression in Douglas-fir in relation to climatic differences between test sites and population sources. Using generalized linear mixed models, probability of disease severity/expression was modeled with respect to the climate variables May through September precipitation (MSP), mean winter temperature (MWT), and continentality. Stark differences in disease expression were observed in trees from different regions, especially in relation to resistance to Rhabdocline spp. and tolerance to P. gaeumannii. There were no major differences across seed-source regions at any particular site in infection levels of P. gaeumannii assessed by fruiting body abundance, yet disease tolerance followed similar geographic patterns as resistance to Rhabdocline spp. Transfers of populations from low to high MSP, and/or cool to warm MWT, increased the probability of moderate to severe Rhabdocline spp. infection and SNC disease symptoms. Our results suggest that local seed sources are adapted to local climate and pathogen pressures and that seed sources from regions with high foliage disease pressure are most resistant/tolerant to those foliage diseases. We also confirm that temperature and precipitation are important epidemiological factors in forest disease and that assisted migration must take into account trophic interactions of trees. Movement of seed sources from dry spring and summer and/or cool winter conditions to mild, mesic environments is likely to lead to increased probability of losses due to these foliage diseases.
Full-text available
Using data from destructively sampled Douglas-fir and lodgepole pine trees, we evaluated the performance of regional volume and component biomass equations in terms of bias and RMSE. The volume and component biomass equations were calibrated using three different adjustment methods that used: (a) a correction factor based on ordinary least square regression through origin (OLS-RTO method); (b) a correction factor based on OLS with intercept (OLS-WI method); and, (c) an inverse approach. The regional volume equations performed fairly well and produced similar results as the locally fitted volume equations of the same form but the regional predicted component biomass estimates were highly biased. All adjustment methods improved the performance of regional equations for the calibration dataset. Based on leave-one-out cross validation, the calibration based on OLS-RTO and OLS-WI methods reduced the RMSE for all species-component combinations. The inverse approach improved the performance of the regional equations for Douglas-fir but it did not improve lodgepole pine regional biomass equations. The decreasing trend of RMSE in component biomass estimation by using randomly selected trees to calibrate regional equations slowed down considerably after five trees.
Full-text available
Large volumes of gridded climate data have become available in recent years including interpolated historical data from weather stations and future predictions from general circulation models. These datasets, however, are at various spatial resolutions that need to be converted to scales meaningful for applications such as climate change risk and impact assessments or sample-based ecological research. Extracting climate data for specific locations from large datasets is not a trivial task and typically requires advanced GIS and data management skills. In this study, we developed a software package, ClimateNA, that facilitates this task and provides a user-friendly interface suitable for resource managers and decision makers as well as scientists. The software locally downscales historical and future monthly climate data layers into scale-free point estimates of climate values for the entire North American continent. The software also calculates a large number of biologically relevant climate variables that are usually derived from daily weather data. ClimateNA covers 1) 104 years of historical data (1901–2014) in monthly, annual, decadal and 30-year time steps; 2) three paleoclimatic periods (Last Glacial Maximum, Mid Holocene and Last Millennium); 3) three future periods (2020s, 2050s and 2080s); and 4) annual time-series of model projections for 2011–2100. Multiple general circulation models (GCMs) were included for both paleo and future periods, and two representative concentration pathways (RCP4.5 and 8.5) were chosen for future climate data.
Full-text available
The success of conifers over much of the world's terrestrial surface is largely attributable to their tolerance to cold stress (i.e.,cold hardiness). Due to an increase in climate variability, climate change may reduce conifer cold hardiness, which in turn could impact ecosystem functioning and productivity in conifer-dominated forests. The expression of cold hardiness is a product of environmental cues (E), genetic differentiation (G), and their interaction (G×E), although few studies have considered all components together. To better understand and manage for the impacts of climate change on conifer cold hardiness, we conducted a common garden experiment replicated in three test environments (cool, moderate, warm) using 35 populations of coast Douglas-fir (Pseudotsuga menziesii var. menziesii) to test the hypotheses: 1)cool-temperature cues in fall are necessary to trigger cold hardening, 2)there is large genetic variation among populations in cold hardiness that can be predicted from seed-source climate variables, 3)observed differences among populations in cold hardiness in situ are dependent on effective environmental cues, 4)movement of seed-sources from warmer to cooler climates will increase risk to cold injury. During fall 2012, we visually assessed cold damage of bud, needle and stem tissues following artificial freeze tests. Cool-temperature cues (e.g., degree-hours below 2°C) at the test sites were associated with cold hardening, which were minimal at the moderate test site owing to mild fall temperatures. Populations differed 3-fold in cold hardiness, with winter minimum temperatures and fall frost dates as strong seed-source climate predictors of cold hardiness, and with summer temperatures and aridity as secondary predictors. Seed-source movement resulted in only modest increases in cold damage. Our findings indicate that increased fall temperatures delay cold hardening, warmer/drier summers confer a degree of cold hardiness, and seed-source movement from warmer to cooler climates may be a viable option for adapting coniferous forest to future climate. This article is protected by copyright. All rights reserved. This article is protected by copyright. All rights reserved.
Full-text available
Impacts of climate change on the climatic niche of the sub-specific varieties of Pinus ponderosa and Pseudotsuga menziesii and on the adaptedness of their populations are considered from the viewpoint of reforestation. In using climate projections from an ensemble of 17 general circulation models targeting the decade surrounding 2060, our analyses suggest that a portion of the lands occupied today primarily by coastal varieties of each species contain genotypes that should remain suitable for the future climate. A much larger portion, particularly for varieties occupying inland sites, should require either introduction of better suited species or conversion to better adapted genotypes. Regeneration strategies are considered with the goal of matching growth potential of contemporary populations to the future climate where that potential can be realized. For some lands, natural reproduction should be suitable, but most lands will require forest renewal to maintain forest health, growth, and productivity. Projected impacts also illustrate the urgent need for conservation programs for P. menziesii in Mexico.
Full-text available
Artificial freeze-testing utilizing the electrolyte-leakage method was used to test the cold hardiness of 2-year-old whitebark pine (Pinus albicaulis Engelm.) seedlings growing in a common garden. Testing across all seasons was used to determine the annual pattern of cold hardiness, and more intensive sampling in the fall and spring was used to assess genetic variation in cold injury among geographic regions spanning the range of the species. Mean hardiness varied widely from –9 °C in early summer to below –70 °C in the winter. Trees from interior and northern regions were the most hardy in the fall, while trees from California were the least hardy. Geographic patterns of hardiness in the spring were reversed. Significant differences in cold injury among regions were detected on all dates except during the winter. Heritability was low to moderate for both the spring (h2 = 0.18) and the fall (h2 = 0.28), and genetic correlation was weak (rA = 0.18). Only spring cold injury was genetically correlated with date of needle flush (rA = 0.34). Mean cold injury in the fall was most closely correlated with mean temperature of the coldest month in the parental environment (r = 0.81). Whitebark pine is well adapted to the low temperatures of the harsh environments where it is found; however, regional variation indicates that moving seed for restoration purposes from areas with higher winter temperatures to colder environments may increase the chance of fall cold injury.
Height growth data were assembled from 10 Pinus ponderosa and 17 Pseudotsuga menziesii provenance tests. Data from the disparate studies were scaled according to climate similarities of the provenances to provide single datasets for 781 P. ponderosa and 1193 P. menziesii populations. Mixed effects models were used for two sub-specific varieties of each species to describe clines in growth potential associated with provenance climate while accounting for study effects not eliminated by scaling. Variables related to winter temperatures controlled genetic variation within the varieties of both species. Clines were converted to climatypes by classifying genetic variation, using variation within provenances in relation to the slope of the cline to determine climatype breadth. Climatypes were broader in varieties of P. ponderosa than in P. menziesii and were broader for varieties inhabiting coastal regions of both species than for varieties from interior regions. Projected impacts of climate change on adaptedness used output from an ensemble of 17 general circulation models. Impacts were dependent on cline steepness and climatype breadth but implied that maintaining adaptedness of populations to future climates will require a redistribution of genotypes across forested landscapes.
Genetic variation in fall cold damage in coastal Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var. menziesii) was measured by exposing excised branches of seedlings from 666 source locations grown in a common garden to freezing temperatures in a programmable freezer. Considerable variation was found among populations in fall cold hardiness of stems, needles, and buds compared with bud burst, bud set, and biomass growth after 2 years. Variation in fall cold hardiness was strongly correlated (r = 0.67) with cold-season temperatures of the source environment. Large population differences corresponding with environmental gradients are evidence that natural selection has been important in determining genetic variation in fall cold hardiness, much more so than in traits of bud burst (a surrogate for spring cold hardiness), bud set, and growth. Seed movement guidelines and breeding zones may be more restrictive when considering genetic variation in fall cold hardiness compared with growth, phenology, or spring cold hardiness. A regional stratification system based on ecoregions with latitudinal and elevational divisions, and roughly corresponding with breeding zones used in Oregon and Washington, appeared to be adequate for minimizing population differences within regions for growth and phenology, but perhaps not fall cold hardiness. Although cold hardiness varied among populations, within-population and within-region variation is sufficiently large that responses to natural or artificial selection may be readily achieved.