Fate of Overstory Trees and Patterns of
Regeneration 12 Years After
Clearcutting with Reserve Trees in
Southwest Washington
Karl R. Buermeyer and Constance A. Harrington, Pacific Northwest Research Station, 362593rd Ave SW, Olympia, WA 98512-9193.
ABSTRACT: Changes in management objectives for some forestlands in the Pacific Northwest have spurred
interest in the creation of multistoried stands and the use of natural regeneration systems, but data on such
systems are lacking. We assessed the status of the overstory trees and the regeneration 12 yr after a clearcut
harvest with reserve trees in an even-aged, 145-yr-old Douglas fir stand on a moderately productive site
(site class 3) in southwest Washington. The 15 ha harvest unit was superimposed over two areas differentially
thinned 15 and 34 yr before clearcutting. The clearcut harvest retained 18 trees/ha with a mean diameter of
63 cm.
The reserved overstory trees had a 93% survival rate after 12 yr; most dead trees had been windthrown.
Diameter growth for the reserved trees averaged 3.3 cm and was greatest during the most recent 3 yr period,
which also had the highest growing-season precipitation. In a 1 ha mapped area, there were 5,854 seedlings/
ha, and more than 99% of the regeneration was Douglas fir. Most seedlings were less than 2 m tall. Seedling
density was somewhat clumped (value of 2.1 for Pielou’s index of nonrandomness), but 79% of randomly
located 4.04 m2 (mil-acre) plots and 98% of 5 x 5 m grid cells had at least one conifer seedling. There was
no obvious pattern of regeneration based on direction from the reserved trees, but both seedling density and
seedling size within the drip lines of reserved tree crowns were less than in the rest of the area. The number of
seedlings was similar on the two halves of the plot corresponding to the original thinning blocks, but seedling
size and age differed. In the half of the study plot that had been twice lightly thinned, only 14% of the seedlings
were >0.5 m tall; however, 41% of the seedlings were >0.5 m in the block that had been thinned more heavily.
There was no difference between the thinning blocks in the ages of seedlings ≤0.5 m tall (mean age of 5 yr). This
example of clearcutting with reserve trees resulted in reasonable survival of the overstory trees and adequate
stocking but slow growth rates in the naturally regenerated Douglas fir. Heavier thinning before harvest was
associated with more advance regeneration, more shrub cover, and less windthrow of the reserved trees than in
the more lightly thinned block. If an abundance of tree species other than Douglas fir was desired on this site,
interplanting would be required. West. J. Appl. For.17(2):78–85.
Key Words: Thinning, Douglas-fir, tree growth and survival.
For some forestlands in the Pacific Northwest, changes
in management objectives to emphasize values other
than wood production (Franklin et al. 1997, Curtis et al.
1998) have spurred interest in the use of new silvicultural
prescriptions. Information is limited, however, on the
long-term effects of silvicultural systems other than
clearcutting (Rose and Muir 1997, Curtis et al. 1998).
Although some Pacific Northwest investigators studied
natural reseeding after clearcutting (Bever 1954,
Lavender et al. 1956, Franklin 1963) or use of shelterwood
systems (Minore and Carkin 1978, Seidel 1979,
Williamson 1983), most forest stand management in the
Note: Constance Harrington can be reached at (360) 753-7670; Fax: (360)
956-2346;E-mail:charrington@fs.fed.us.Thisarticlewaswrittenby
U.S. Government employees and is therefore in the public domain.
78
latter half of the 20th century involved clearcutting and
replanting, so most literature addresses these methods.
The shelterwood regeneration method (the retention of
overstory trees at harvest and their later removal once the
seedlings are established) has traditionally been used on
harsh sites to modify the microenvironment by providing
partial shading or thermal cover for young seedlings.
Similar harvest prescriptions can be used to create
two-story stands if the reserved trees are not removed
(Williamson and Ruth 1976, Smith 1986, DeBell et al.
1997, Curtis et al. 1998). This concept has been referred
to as green tree retention (Franklin 1992), irregular
shelterwood (Smith 1986), or clearcutting with reserve
trees (Curtis et al. 1998). The term green tree retention has
been widely used in this region to describe an alternative
harvest treatment; however, at this point in time it
Reprinted from Wester Journal of Applied Forestry, Vol. 17 No. 2, April 2002. Not for further reproduction.
is not a silvicultural system that implies “a planned
program of silvicultural treatment during the whole life
of a stand....” (Smith 1962 p. 353). Thus, retaining green
trees at harvest is probably best considered as one of
the objectives of a silvicultural treatment or harvest
prescription, while clearcutting and shelterwood are
terms reserved for silvicultural systems. Clearcutting
with reserves differs from an irregular shelterwood
or shelterwood with reserves in that clearcutting does
not have microsite moderation as one of its objectives
(Helms 1998). Reserved or retained overstory trees may
provide one or more additional canopy layers, various
sizes and conditions of trees, as well as a continuous
source of future snags and down wood, all necessary
structural features in meeting some forest management
objectives (Franklin et al. 1997). Long-term data on the
effects of overstory retention have come primarily from
retrospective studies of regeneration resulting from
natural disturbances (Rose and Muir 1997, Zenner et al.
1998) although Isaac (1940) followed the fate of seed
trees after harvest. A harvest unit in the Wind River
Experimental Forest provided an opportunity to study in
detail the survival and growth of retained trees and the
species composition, age, size, and spatial distribution
of natural regeneration 12 yr after clearcutting with
reserve trees.
Methods
Description of Study Site and Cutting History
The study was conducted in a 15 ha harvest unit in the
Panther Creek Division of the Wind River Experimental
Forest in southwestern Washington (latitude 45°50’N,
longitude 121°50’W). The unit was part of the Spearfish
timber sale. It is on a west slope of generally less than
25%; elevation is between 670 and 730 m. Mean annual
precipitation averages 2,820 mm. Douglas-fir 50 yr site
index is 26.1 m (King 1966). The unit is at the lower
elevational end of the Pacific silver fir–salal plant
association (transitional with the western hemlock–
salal association), which is moderately productive for
tree growth and is characterized by warm site shrubs
(Brockway et al. 1983). Predominant shrubs present at
our site are vine maple (Acer circinatum), Douglas maple
(A. glabrum var. douglasii), hazel (Corylus cornuta
var. californica), salal (Gaultheria shallon), and red
huckleberry (Vaccinium parvifolium). Pacific dogwood
(Cornus nuttallii) was the most common broadleaved
species at the site. The soil is mapped as Zygore (loamyskeletal, isotic, frigid Andic Dystrudepts), which is a
very deep, well drained soil that formed in colluvium
and outwash weathered from andesite and basalt and
mixed with ash.
The
even-aged,
predominantly
Douglas-fir
(Pseudotsuga menziesii) stand originated after a
wildfire around 1840; the low level of residual woody
material present in the area suggests that one or more
additional fires may have preceded the 1840 event.
Our study area included two of the nine blocks from
a study begun in 1952 to study responses to various
densities of commercial thinning (Williamson 1966,
1982). The block assigned to light thinning in the
1952 study had 22% of the basal area removed (thinned
from 67 to 52 m2/ha) and the heavy thinning block had
35% of the basal area removed (thinned from 55 to 36
m2/ha). Williamson (1982) mentions a second thinning in
1971; no details are given, other than it “was instigated by
mortality from [bark beetles (Dendroctonus pseudotsugae)
and root rot (Phellinus weirii)].” The intention in 1971 was
to lightly rethin the light thinning block from 1952 and
to more heavily rethin the heavy thinning block. Based
on differences in tree density between these two blocks
that were visible in 1979 aerial photographs, we assume
that intention was met. Trees sampled in a preharvest
1985 stand examination were entirely Douglas-fir
except for three small western redcedars (Thuja plicata)
and one western hemlock (Tsuga heterophylla).
The prescription for the 1986 harvest called for
retaining dominant, undamaged trees with healthy
crowns and selecting against limby trees (Franklin
1984b). The intent at the time was to leave the overstory
trees through the following rotation to produce high
quality timber, create a naturally regenerated stand, and
study the effects of the retained overstory trees on the
regeneration (Franklin 1984a). Because site modification
was not a stated objective, this harvest would best be
characterized in current terminology as a “clearcut with
reserves” (Helms 1998). Before clearcutting, basal area
was 33.7 m2/ha; after harvest, it was 6.1 m2/ha (Table 1).
The 18 reserve trees/ha (all Douglas-fir) averaged 65 cm
in diameter and 41 m in height and were spaced about 24
m apart. Size of reserved trees in the two thinning blocks
was similar.
Measurements and Analyses
Heights and diameters of all 268 residual overstory
trees in the 15 ha harvest unit were measured in August
1989. All surviving overstory trees were remeasured
in fall 1998, 12 yr after harvest. Diameter and tree
condition were recorded for all overstory trees, and
height and height-to-live-crown were measured with
a sonar hypsometer on a subset of 60 trees. Height to
diameter ratios and proportion of tree height in live
crown were calculated from these data. We used 1989
data to compare characteristics of dead trees with those
of surviving trees.
Concurrent with the overstory tree measurement
in 1998, a 100 x 100 m (1 ha) plot was established to
monitor the size and distribution of the regeneration.
It was positioned midslope, halfway between the
access road on the west side of the unit and the eastern
edge of the stand (minimum distance to the stand
edge was about 60 m). The plot was bisected by
the boundary of the two former thinning treatment
blocks to represent the two thinning densities equally
79
WJAF 17(2) 2002
and have both portions of the plot occupy similar slope positions. The 1 ha block was made into a 10 x 10 m square
grid to facilitate mapping. The approximate location of
all conifer regeneration, overstory trees, skid roads, down
trees, and significant shrub clumps (defined as clumps of
apparent size and density to exclude conifer regeneration)
was mapped. As the regeneration was mapped, each seedling was coded as to species and height class. The height
classes used were ≤0.5 m, >0.5 to 2 m, >2 to 4 m, and >4
m. For readability, and because the distinction between the
terms “seedling” and “sapling” is vague (Ford-Robertson
1971), individuals in the regeneration layer will be referred to as “seedlings” here, regardless of size. All visible
seedlings were recorded; the smallest were about 3 cm tall
and the tallest about 6 m.
All 19 surviving overstory trees inside the 1 ha study
plot were measured for height and height to live crown.
Increment cores were taken from 18 of these trees; growth
increment was recorded for 10 yr before the 1986 harvest
and at 3 yr intervals from 1987 through 1998. Growing
season (May 1–October 31) precipitation data for the
Carson National Fish Hatchery (the nearest recording site,
about 12 km WNW of the study site, at an elevation of
350 m) were obtained from a Web site maintained by the
Western Regional Climate Center, at the Desert Research
Institute (http:// www.wrcc.dri.edu). Data at this site were
available for 1977 through 1998. Missing or incomplete
monthly data were estimated by regression with data
from the same Web site for the next nearest recording
station, Peterson’s Ranch, 29 km to the NW and at 200
m elevation. The R2 values for these monthly regression
formulas ranged from 0.78 to 0.95.
In the office, the field mapping sheets were used to create a point layer and a polygon layer for use in ArcView (a
geographic information system program). The point layer
included all tree locations, and the polygon layer included shrub clumps and skid roads. Crown diameters for the
overstory trees in or adjacent to the 1 ha mapped area were
estimated by measuring shadow widths on a 1988 aerial
photograph (scale estimated to be 1:6250) and were added
to the database. Examination of the spatial data indicated
that ground mapping had missed portions of skid trails that
had partially overgrown; thus, the locations of the skid trails
were redigitized based on the 1988 aerial photograph.
After reviewing the initial data, we found an obvious
difference in height distribution of the seedlings between
the two halves of the plot corresponding to the two thinning treatments. To better interpret the height–class information, we felt we needed to know seedling ages. Thus,
the following spring, we counted bud scars on the stems to
determine the ages of ten Douglas-fir seedlings from each
of the four height classes in both of the thinning treatments
(except that there were only six Douglas-fir seedlings taller
than 4 m in the heavily thinned portion and none in the
lightly thinned portion). Destructive sampling and counting rings from similarly sized individuals immediately
outside the study plot verified seedling ages determined by
this method.
Seedling numbers were summarized by species and height
class for the entire 1 ha study plot and for the portions of the plot
representingthetwoearlierthinningtreatments.Pielou’s(1959)
WJAF 17(2) 2202
80
nonrandomness index was calculated to test the pattern
of seedling distribution in the 1 ha study plot. Distance to
the nearest seedling from each of 72 interior grid points
(the intersection of the 10 m squares) was measured
in ArcView and used to calculate the nonrandomness
index. To approximate the USDA Forest Service standard
stocking surveys, we also used these 72 grid points as the
center of simulated 4.05 m2 (mil-acre) plots to test for
seedling stocking percentage. In addition, to provide
another measure of stocking, all 400 of the 5 x 5 m grids
in the 1 ha plot were examined in ArcView to determine
conifer seedling presence.
Seedling densities were calculated for (1) seedlings in
skid trails versus the number on the rest of the area. and
(2) seedlings within versus those outside the drip lines
of the overstory trees (as approximated by the crown
diameters). Diameter growth of the reserved trees (in 3
yr increments) was predicted via linear regression from
time since harvest and from 3 yr mean growing-season
precipitation. The correlation between time since harvest
and growing season precipitation was also calculated.
Results and Discussion
Overstory Tree Survival
Of the 268 overstory trees retained in the 15 ha harvest
unit, 93% were surviving 12 yr after harvest; this reduced
live tree density from 18 to 17 trees/ha. Of those that died, 3
were standing dead, 1 snapped off below live crown, and
15 were uprooted dead (uprooted trees exhibit symptoms
of laminated root rot). Twice as many trees blew down in
the light thinning block as in the heavy thinning block.
The standing dead and broken trees (4 total) were split
evenly between the two treatments and had average to
below-average diameters. Epicormic branching along the
mid- and upper stem (above 10 m) was observed on most
of the overstory trees, and several trees appeared to have
sustained minor top damage.
Douglas-fir seed trees left after logging in the 1920s
were monitored at eight locations for 11 to 15 yr (Isaac
1940); windthrow at these locations ranged from 2.8
to 50%. No information on stand age or history was
provided, but given the time period we would assume the
stands were all fairly old and had not been previously
thinned. The stand with the lowest loss to windthrow
(2.8%) was on the Willamette National Forest in Oregon;
it was described as “a somewhat open stand where the
trees were naturally rather wind-firm....” The high losses
to the seed trees at some locations in Isaac’s study (four
of the eight locations lost >40% of the trees to windthrow)
were greater than we would expect from residual trees left
after harvest of most managed stands where stand density
prior to harvest would generally be much lower and
trees would be younger and with less defect than in the
unmanaged, presumably old-growth stands of natural
origin sampled in Isaac’s study. Stands on some sites,
however, particularly those on poorly drained soils where
root systems are shallow, will be more susceptible to
windthrow than those on better drained soils regardless
of stand history. Isaac (1940) also noted that trees on
ridgetops and dry slopes were the most windfirm.
Overstory Tree Growth
Mean diameter of the surviving trees in 1998 was 68
cm, height was 44 m (Table 1), the height-to-diameter
ratio was 65 cm/cm, and the proportion of live crown
was 43%. Diameter growth of the overstory trees during
the 12 yr since harvest averaged 3.3 cm; tree core data
from the 18 tree subsample and diameter measurements
of all the trees yielded similar results. Periodic diameter
growth was lowest in the first growth period after harvest
and increased in each of the successive growth periods
(Figure 1). The initial decline could have been related
to a release shock effect or to the concurrent decline in
growing-season precipitation. Given the previous thinning
history in this stand, the 40% live crown on the reserved
trees, and the fact the reserve trees were selected for
their future growth potential, we would not have expected
much of a shock effect to be visible in the growth rates of
these trees; however, the presence of epicormic branches
indicates that clearcutting resulted in a major change in
the light environment and physiological balance of the
reserved trees (Kramer and Kozlowski 1960, p. 387).
Growing season precipitation was 31 to 61% less
than the 22 yr average (from 1977 to 1998) in 6 of the
7 yr after clearcutting. Postharvest growth was positively
related with both time since harvest ( R 2 = 0.66, for
regression P = 0.12) and growing season precipitation (R2
= 0.80, for regression P = 0.07). A visual examination
of the data provided some indication of curvilinearity;
however, given the small sample size (n = 4) and no clear
theoretical basis for a curvilinear relationship, we are
only reporting the results from linear models. Because
growing season precipitation and time since harvest
were correlated with each other (R2= 0.96, P = 0.01), we
cannot separate the two factors. However, because both
growth rate and precipitation were at lower levels after
harvest than before harvest (both increased incrementally
after harvest), postharvest precipitation rates were quite
low, and the P value associated with the regression
between growth rate and precipitation was smaller than
the P value for the regression with time since harvest, we
suspect that precipitation was the more important factor
influencing growth of the residual trees. Because response
to thinning in older stands can occur following a time
lag (c.f., Tappeiner and Latham 1999), however, we
cannot rule out future increases in growth rates that are
independent of growing season precipitation.
Stocking and Distribution of Regeneration
In 1998, there were 5,854 seedlings on the 1 ha study
plot (Figure 2). Based on Pielou’s nonrandomness index
of 2.10, the seedling distribution would be considered
clumped (values <1 are considered to be random, a value
of 1 is uniform, and values >1 are considered to be
clumped). A study of spatial patterns in seeded loblolly
pine stands reported nonrandomness indices ranging from
0.66 to 9.93 (Daniels 1978); thus, our index value of 2.10
is well below what can occur. Daniels (1978) concluded
that microsite and other site factors are the major factors
influencing spatial distribution of seedlings; thus, the
clumped nature of the regeneration in our study probably
reflects the clumped nature of favorable microsites.
Seventy-nine percent of the simulated 4.05 m2 (mil-acre)
plots were stocked with at least one conifer seedling
and 98% of the 5 x 5 m grid cells used to map the study
plot contained at least one established seedling. Thus,
from the standpoint of a statistical test, the regeneration
would be considered clumped; however, from a practical
standpoint, because the nonstocked areas were small, the
area would be considered adequately stocked.
There was no apparent pattern of regeneration based on
direction from the reserved trees. The lack of regeneration
patterns attributable to seed dispersal was expected
as Douglas-fir seed can travel over 400 m from either
a stand edge or a point seed source (Isaac 1943, 1949;
Lavender et al. 1956), and at least half of it will travel
more than 60 m (Isaac 1943, 1949). Because trees in this
study plot averaged only 23 m apart, seeding patterns from
individual trees would more than sufficiently overlap to
obscure any patterns due to individual trees. The nearest
stand edges were about 60 m to the east and 250 m to
the southwest from the edges of the study plot; based on
prevailing wind direction and distance, these would not
be expected to add large numbers of seeds compared
to the reserved trees. Well-distributed Douglas-fir seed
trees, such as those present in this study, are considered
to be more effective in regenerating a unit because of
their overlapping seed distribution patterns than are
timber edges (Bever 1954, Franklin 1963) or blocks of
trees within a unit (Bever 1954).
Seedling density was slightly lower within the overstory
tree drip lines (4,801 seedlings/ha) than outside the drip
lines (5,919 seedlings/ha), and no seedlings >2 m tall were
found within the drip lines. Inhibitory effects of residual
trees on establishment and growth of western conifer
seedlings have been previously reported (Emmingham
pers. comm., McDonald 1976, Tesch and Mann 1991).
Negative effects of overstory trees also are documented
for 50- to 120-yr-old trees growing under or with oldgrowth remnant trees (Rose and Muir 1997, Zenner et
al. 1998). Negative effects could be due to allelopathy,
shading, interception of precipitation, or belowground
competition for moisture or nutrients. Our data on
seedling size (by height class) are not sensitive enough to
test for differences in seedling growth due to location of
81
WJAF 17(2) 2002
reserve trees. Given the dryness of the study site, however,
we would expect the positive effects from reserved trees
(such as partial day shading and mycorrhizal inoculation)
to be less important over time than the negative effects
(such as water and nutrient uptake and interception of
precipitation).
Regeneration was more common on skid roads
(8,693 seedlings/ha) than on the rest of the area
(5,197 seedlings/ha outside these areas; see also
Figure 2). This relation was expected as it is generally
accepted that exposed mineral soil provides the best
environment for Douglas-fir seedling establishment
(Isaac 1943, Williamson 1973, Stewart 1978). Of 1,198
seedlings growing in the skid trails, however, only 9
WJAF 17(2) 2202
82
were taller than 2 m, and 1 was taller than 4 m. The low
number of seedlings in the larger height classes on the
skid trails may reflect poorer growing conditions, such
as compacted soil, on the trails. Alternatively, seedlings
may have become established on the skid trails after the
thinning but were destroyed or damaged when the trails
were reused.
Shrub cover was not high in the area, but it was
noticeable that tree seedlings were rarely found
within the shrub clumps. It was somewhat surprising
to observe the slow successional development of the
area as evidenced by the relatively low shrub cover
and presence of exposed mineral soil 12 yr postharvest
activities. The low growing season precipitation
in the years immediately following clearcutting and
browsing by elk and deer are likely the main factors
preventing more rapid vegetation development.
Seedling Size, Age, and Species Composition
Seedling size ranged from about 3 cm to 6 m with
most seedlings less than 2 m tall. There are many very
small seedlings on the site, which indicates that seedling
establishment is still occurring. Seedlings were distributed
by height class as follows: ≤0.5 m, 72%; >0.5 to 2 m, 26%;
>2 to 4 m, 2%; and >4 m, <0.1%.
Seedling ages on the 66 sampled trees ranged from 3
to 35 yr old (Figure 3). Seedlings ≤0.5 m tall averaged 5.5
yr old. The distribution of seedlings <12 yr indicates that
regeneration established during most of the postharvest
years. Seedlings ≥12 yr became established in the understory of the previous stand; these account for 41% of the
aged seedlings. Equal numbers of seedlings from each
height class were aged; hence after taking into account
the total number of seedlings in each height class (Figure
4), we estimated that 15% of the seedlings were residuals
from the preharvest stand.
Conifer regeneration was >99% Douglas-fir; other species present were western hemlock (Tsuga heterophylla),
western white pine (Pinus monticola), grand fir (Abies
grandis), Pacific silver fir (A. amabilis), and noble fir
(A. procera). Western hemlock, normally a prolific
seeder (Packee 1990), was present at the site at harvest
but only as a few small trees. Larger hemlocks, the best
seed producers, probably would have been discriminated
against in the previous thinnings. The management
guidelines for this plant association suggest the most
suitable species for planting are Douglas-fir and noble fir,
whereas western hemlock may be planted, or may seed
in naturally, in moister plant associations in the Pacific
silver fir type (Brockway et al. 1983). Douglas-fir is more
heat tolerant and slightly more frost tolerant than western
hemlock or western redcedar (Minore 1979). More
shade-tolerant trees will probably not become established
in greater numbers until the shade from the developing
stand moderates the high light and low soil-moisture
environment; this could take many years for this plant
association. Interplanting the area with species other than
Douglas-fir could have supplemented the mix of species,
but the survival of species other than noble fir or western
white pine might not have been high. The currently established individuals of species other than Douglas-fir could be
favored in future thinning entries. If all 50 non-Douglas-fir
seedlings in the 1 ha study plot survived, at a hypothetical
future stand density of 250 trees/ha, 20% of the stems
would be species other than Douglas-fir.
Effects of Past Thinning Regimes
The 1952 and 1971 thinning regimes apparently influenced the conifer regeneration present. Total stocking of
seedlings in the portions of the plot that correspond to the
two different thinning treatments are similar: 5,678 trees/
ha in the lighter thinning vs. 6,030 trees/ha in the more
heavily thinned area. The size distribution of the resulting
regeneration, however, is markedly different (Figure 4).
In the portion thinned lightly, only 14% of the seedlings
were taller than 0.5 m, and in the heavily thinned area,
41% of the seedlings were taller than 0.5 m. The heavier
thinning regime may have opened the stand sufficiently to
promote more seedfall and germination (Reukema 1961,
Williamson 1983) or, more likely, to favor the survival
and growth of advance regeneration up to the time of the
reserve-tree harvest. The age data support our contention
that advanced regeneration was a factor because the taller
seedlings from the heavily thinned block were generally
older than those from the lightly thinned portion of the
unit. There was no difference in the ages of seedlings
0.5 m tall; they averaged 5.5 yr old with no difference
between thinning blocks. However, seedlings >0.5 m
tall were consistently older in the heavily thinned area
(Figure 5). Mean age of the seedlings in the 0.5 to 2 m
height class was greater in the heavily thinned block (15
yr) than the lightly thinned block (9 yr). The 2 to 4 m
height class averaged 17 and 12 yr old in the heavily
and lightly thinned blocks, respectively. Other than three
broken-topped remnants of the previous stand, there
were no understory trees greater than 4 m tall in the light
thinning treatment. Of the 20 seedlings taller than 0.5 m
sampled in the lightly thinned portion, 7 (35%) originated
before the 1986 harvest (≥12 yr), compared with 24 out of
27 (89%) in the heavily thinned area. The heavily thinned
portion had fewer seedlings less than 0.5 m tall, possibly
83
WJAF 17(2) 2002
released. Thus, although the heavier thinning before
harvest apparently increased shrub cover, the overall
effect on seedling stocking was quite small.
Management Implications
because more growing sites were already occupied by
advanced regeneration at the time of the 1986 harvest.
Worthington and Heebner (1964) also report that a
light thinning regime in Douglas-fir (they left 78–84%
of the existing basal area) did not significantly stimulate
regeneration. It is interesting that only four of the aged
trees in our study predated the second thinning entry
(1971). If regeneration established following the 1952
thinning entry, it may have been shaded out by crown
closure before the 1971 entry. Douglas-fir is less tolerant
of shade as it matures (Hermann and Lavender 1990); this
may have limited its age in the understory. Other possible
explanations for the lack of older seedlings are that they
were killed by mechanical damage in the 1971 thinning
operation, or older, less flexible seedlings may have been
preferentially damaged in the 1986 harvest.
There were more shrub clumps present in the 1
ha study plot under the heavily thinned regime (13
clumps covering 89 m2 or 1.8% of the area) than under
the lightly thinned regime (two clumps totaling 9 m2 or
0.2% of the area). Although the shrub clumps precluded
regeneration, they covered too small an area to have a
major influence on the amount of regeneration present in
each block. The age distribution of the conifer seedlings
also suggests that many seedlings were established in
the heavily thinned block before the shrub layer was
WJAF 17(2) 2202
84
The clearcut-with-reserve-tree harvest system met the
original objectives of creating a two-layered stand with
adequate natural regeneration. Survival of the reserved
trees was 93% after 12 yr; if this level of attrition were not
acceptable, more trees could be retained when designing
future harvests. Based on personal observations of other
areas where green trees were retained in the last decade,
as well as the eight areas reported on by Isaac (1940), it
appears that losses of reserved trees can vary substantially
among stands. The higher windthrow experienced in the
lightly thinned block reinforces previous recommendations
(Cremer et al. 1982, Somerville 1989) that increasing
growing space per tree will improve future windfirmness.
In addition, losses to windthrow will be less on sites with
deep well-drained soils or on ridge tops where long-term
exposure to wind has resulted in windfirm trees. On our
study area, we would expect annual losses to windthrow
in the next few decades to be smaller than the losses since
harvest because currently standing trees have proven to
be windfirm.
Diameter growth in our study only averaged 3.3 cm in
12 yr. We suspect that 12 yr growth rates will be greater in
future periods if growing season precipitation is closer
to the long-term historical average. Current stand growth
models do not have data on growth of reserve trees
following clearcutting, so growth projections should be
conservative until more information is available.
Although the regeneration was somewhat clumped,
most portions of the 1 ha plot were covered with Douglasfir seedlings and would be considered adequately stocked.
Spatial patterns can be evaluated at different scales
depending on the interest of the scientist or forester. For
example, the locations of the seedlings in Figure 2 are
clearly not distributed uniformly, and that is reflected in
Pielou’s nonrandomness index, which is based on point
samples. Many managers, however, would be more
interested in knowing how many regeneration plots or
5 m x 5 m blocks were not stocked, and that is reflected
in the statistics about those larger sampling units (21 and
2% respectively). Thus, the clumping of the regeneration
becomes less important as the size or spatial scale of
interest increases from points to 4.05 m2 plots to 25 m2
blocks.
The growth rates of both the advance regeneration
and the postharvest regeneration have been slow. As
a comparison, 9-yr-old Douglas-fir planted across
the road from the study site average 2.8 m tall, while
most of the advance regeneration in our study plot
is 0.5 to 2 m tall and much of the postharvest seeded
regeneration is <0.5 m tall. Thus, it will take the natural
regeneration several years longer than planted stock to
reach a target height (such as 1.5 m, which is required in
some areas before an adjacent stand can be harvested).
The presence of advance regeneration provides a height
advantage; however, the advance regeneration will
not respond quickly to release if it has been suppressed
for many years.
If managers are interested in maintaining or increasing
tree-species diversity, they should consider the potential
long-term effects of favoring species during intermediate
or final harvest, reserving no-cut or no-entry spots within
stands, and designing unit size and shape to favor regeneration of desired species. Although not always practical,
removing the overstory immediately following a major
seed crop for desired tree species should also help in increasing tree-species diversity.
The combination of almost no diversity in the overstory at time of harvest and several years of below-average growing-season precipitation resulted in a naturally
regenerated stand with very little tree-species diversity
12 yr after harvest. If higher tree-species diversity were
desired, interplanting could have been used to increase
the mix of species; however, some species will probably
not do well until temperature extremes are moderated by
shade and thermal protection from a more closed canopy.
Thus, our results should serve as a reminder to forest
managers not to assume that naturally regenerated stands
will be more diverse in tree species than plantations.
Literature Cited
BEVER, D.N. 1954. Evaluation of factors affecting natural reproduction of forest trees in central western Oregon. Res. Bull. 3, Oregon State Board of
Forestry, Salem. 49 p.
BROCKWAY, D.G., C. TOPIK, M.A. HEMSTROM, AND W.H. EMMINGHAM. 1983. Plant association and management guide for the Pacific silver fir zone: Gifford Pinchot
National Forest. USDA For. Serv. R6-Ecol-130a. 122 p.
CREMER, K.W., C.J. BOROUGH, F.H. MCKINNELL, AND P.R. CARTER. 1982. Effects of stocking
and thinning on wind damage in plantations. N.Z. J. For. Sci. 12(2):244–
268.
CURTIS, R.O., D.S. DEBELL, C.A. HARRINGTON, D.P. LAVENDER, J.B. ST. CLAIR, J.C. TAPPEINER,
AND J.D. WALSTAD. 1998. Silviculture for multiple objectives in the Douglas-fir
region. USDA For. Serv. Gen. Tech. Rep. PNW-435. 123 p.
DANIELS, R.F. 1978. Spatial patterns and distance distributions in young seeded
loblolly pine stands. For. Sci. 24(2):260–266.
DEBELL, D.S., R.O. CURTIS, J.S. DEBELL, AND R.J. MCGAUGHEY. 1997. Comparing options
for managing Douglas-fir forests. P. 590–607 in Proc. of the IUFRO interdisciplinary uneven-aged management symposium, Emmingham, W.H.
(comp.). Oregon St. Univ., Corvallis.
FORD-ROBINSON F.C. (ED.). 1971. Terminology of forest science, technology, practice and products. Soc. of Am. For., Bethesda, MD. 349 p.
FRANKLIN, J.F. 1963. Natural regeneration of Douglas-fir and associated species
using modified clear-cutting systems in the Oregon Cascades. USDA For.
Serv. Res. Pap. PNW-3. 14 p.
FRANKLIN, J.F.1984a. Internal Forest Service memo, September 26, 1984, in
Spearfish files at the Forestry Sciences Laboratory, Olympia, WA.
FRANKLIN, J.F.1984b. Internal Forest Service memo, November 28, 1984, in Spearfish files at the Forestry Sciences Laboratory, Olympia, WA.
FRANKLIN, J.F.1992. Scientific basis for new perspectives in forests and streams.
P. 25–72 in Watershed management: Balancing sustainability and environmental change, Naiman, R.J. (ed.). Springer Verlag, New York.
FRANKLIN, J.F. D.R. BERG, D.A. THORNBURGH, AND J.C. TAPPEINER. 1997. Alternative silvicultural approaches to timber harvesting: variable retention harvest systems. P. 111–139 in Creating a forestry for the 21st century—the science of
ecosystem management, Kohm, K.A., and J.F. Franklin. (eds.). Island Press,
Washington, DC.
HELMS, J.A. (ED.). 1998. The dictionary of forestry. Soc. of Am. For., Bethesda,
MD. 210 p.
HERMANN, R.K., AND D.P. LAVENDER. 1990. Pseudotsuga menziesii (Mirb.) Franco:
Douglas-fir. P. 527–540 in Silvics of North America. Vol. 1: Conifers. USDA
Agric. Handb. 654. Washington, DC.
ISAAC, L.A. 1940. P. 11–12 in Mortality of Douglas-fir seed trees on cut-over lands.
USDA For. Serv. PNW Res. Note 31.
ISAAC, L.A. 1943. Reproductive habits of Douglas-fir. Charles Lathrop Pack Forestry Foundation, Washington, DC. 107 p.
ISAAC, L.A. 1949. Better Douglas-fir forests from better seed. Univ. of Washington Press, Seattle. 64 p.
KING, J.E. 1966. Site index curves for Douglas-fir in the Pacific Northwest. Weyerhaeuser Forestry Paper 8. Weyerhaeuser For. Res. Center, Centralia,
WA. 49 p.
KRAMER, P.J., AND T.T. KOZLOWSKI. 1960. Physiology of trees. McGraw-Hill, New York.
642 p.
LAVENDER, D.A., M.H. BERGMAN, AND L.D. CALVIN. 1956. Natural regeneration on staggered settings. Res. Bull. 10, Oregon State Board of Forestry, Salem. 36 p.
MCDONALD, P.M. 1976. Inhibiting effect of ponderosa pine seed trees on seedling
growth. J. For. 74 (4):220–224.
MINORE, D. 1979. Comparative antecological characteristics of northwestern
tree species—a literature review. USDA For. Serv. Gen. Tech. Rep. PNW87. 72 p.
MINORE, D., AND R.E. CARKIN. 1978. Vegetative indicators, soils, overstory canopy
and natural regeneration after partial cutting on the Dead Indian Plateau
of southwestern Oregon. USDA For. Serv. Res. Note PNW-316.9 p.
PACKEE, E.C. 1990. Western hemlock. P. 613–622 in Silvics of North America,
Volume 1: Conifers. USDA Agric. Handb. 654. Washington, DC.
PIELOU, E.C. 1959. The use of point-to-plant distances in the study of the pattern
of plant populations. J. Ecol. 47:607–613.
REUKEMA, D.L. 1961. Seed production of Douglas-fir increased by thinning. USDA
For. Serv. Res. Note PNW-210. 5 p.
ROSE, C.A., AND P.S. MUIR. 1997. Green-tree retention: Consequences for timber production in forests of the western Cascades, Oregon. Ecol. Appl.
7(1):209–217.
SEIDEL, K.W. 1979. Natural regeneration after shelterwood cutting in a grand
fir–Shasta red fir stand in central Oregon. USDA For. Serv. Res. Pap. PNW259. 23 p.
SMITH, D.A. 1986. The practice of silviculture. Wiley, New York. 527 p.
SOMERVILLE, A. 1989. Tree wind stability and forest management practices. P.
38–58 in Workshop on wind damage in New Zealand exotic forests. For.
Res. Inst. Bull. 146, Somerville, A., S. Wakelin, and L. Whitehouse (eds.).
Ministry of Forests, Rotorua, New Zealand.
STEWART, R.E. 1978. Site preparation. P. 100–129 in Regenerating Oregon’s forests, Cleary, B.D, R.D. Greaves, and R.K. Herman (eds.). Oregon St. Univ.
Ext. Serv., Corvallis.
TAPPEINER, J.C., AND P. LATHAM. 1999. Thinning to increase vigor of old-growth trees.
P. 23 in Cooperative forest ecosystem research: Annual report 1999, Littlefield, B. (ed.). Oregon St. Univ., Corvallis.
TESCH, S.D., AND J.W. MANN. 1991. Clearcut and shelterwood reproduction methods
for regenerating southwest Oregon forests. Res. Bull. 72. For. Res. Lab.
Oregon St. Univ., Corvallis. 43 p.
WILLIAMSON, R.L. 1966. Thinning response in 110-year-old Douglas-fir. USDA For.
Serv. Res. Note PNW-36. 7 p.
WILLIAMSON, R.L. 1973. Results of shelterwood harvesting of Douglas-fir in the
Cascades of western Oregon. USDA For. Serv. Res. Pap. PNW-161. 13 p.
WILLIAMSON, R.L. 1982. Response to commercial thinning in a 110-year-old Douglas-fir stand. USDA For. Serv. Res. Pap. PNW-296. 16 p.
WILLIAMSON, R.L. 1983. Seedfall under coastal Douglas-fir shelterwood stands.
Northwest Sci. 53 (3):205–211.
WILLIAMSON, R.L., AND R.H. RUTH. 1976. Results of shelterwood cutting in western
hemlock. USDA For. Serv. Res. Pap. PNW-201. 25 p.
WORTHINGTON, N.P., AND C.F. HEEBNER. 1964. Light thinning of Douglas-fir does not
stimulate regeneration. USDA For. Serv. Res. Note PNW-16.5 p.
ZENNER, E.K., S.A. ACKER, AND W.H. EMMINGHAM. 1998. Growth reduction in harvestage, coniferous forests with residual trees in the western central Cascade
Range of Oregon. For. Ecol. Manage. 102:75–88.
85
WJAF 17(2) 2002