1554
Regeneration responses influenced by single-tree
selection harvesting in a mixed-species tree
community in northern Japan1
Mahoko Noguchi and Toshiya Yoshida
Abstract: The objective of this study was to assess the effects of fine-scale canopy disturbances induced by selection harvesting and its associated practices (artificial planting and machinery skidding) on the successful regeneration of tree species in a northern Japanese mixed forest. We set up 163 plots in a 6.7 ha permanent study stand where trees have been
partially harvested at approximately 10-year intervals since 1974. The regeneration of trees (4324 stemsha–1) occurred
more frequently under closed canopies than under canopy gaps, except for a typical shade-intolerant species, Betula ermanii Cham. In particular, small canopy openings that tend to close quickly displayed higher understory tree regeneration
densities likely due to the suppression of competition from dwarf bamboos. The surface soil disturbances that occurred
during planting and harvesting even further enhanced understory regeneration. The results shown here should be generalized carefully because we have investigated only one stand. Nevertheless, our findings clearly indicated that the creation
of small canopy gaps associated with site preparation that contains soil disturbances should be examined in management
practices to maintain the community structure in this type of mixed forest.
Résumé : L’objectif de cette étude consistait à estimer les effets de perturbations fines de la canopée induites par la coupe
de jardinage et par les traitements qui y sont associés (plantation et débardage mécanisé) sur le succès de la régénération
en espèces arborescentes dans une forêt mixte du nord du Japon. Nous avons établi 163 parcelles échantillons dans un
peuplement sous étude permanente de 6,7 ha ayant fait l’objet, depuis 1974, de coupes partielles environ à tous les 10 ans.
La régénération d’arbres (4 324 tigesha–1) s’est installée plus fréquemment sous les canopées fermées que dans les trouées,
sauf dans le cas d’une espèce typiquement intolérante à l’ombre, Betula ermanii Cham. De façon plus particulière, les petites ouvertures de la canopée, qui tendent à se refermer rapidement, sont associées à une plus forte densité de la régénération arborescente, probablement attribuable à l’élimination de la compétition du bambou nain. Les perturbations de la
surface du sol causées par la plantation et la récolte des arbres ont favorisé encore plus la régénération sous couvert. Les
résultats de cette étude devraient être généralisés avec précaution puisque nous avons étudié un seul peuplement. Néanmoins, nos résultats indiquent clairement que la création de petites trouées dans la canopée, jumelée à une préparation de
terrain incluant une perturbation du sol, devrait être considérée parmi les pratiques d’aménagement pour maintenir la structure de la communauté de ce type de forêt mixte.
[Traduit par la Rédaction]
Introduction
Recently, there has been increased interest in revising silvicultural systems for forest stands to incorporate conservation objectives into management plans (Kohm and Franklin
1997; Angelstam 1998; Hunter 1999). This revision has engendered a widespread shift from clearcut harvesting to partial cut harvesting because the residual trees contribute to
Received 15 November 2006. Accepted 28 May 2007. Published
on the NRC Research Press Web site at cjfr.nrc.ca on
17 October 2007.
M. Noguchi.2 Shikoku Research Center, Forestry and Forest
Products Research Institute, 2-915 Asakuranishimachi, Kochi
780-8077, Japan.
T. Yoshida. Uryu Experimental Forest, Field Science Center for
Northern Biosphere, Hokkaido University, Hokkaido, Japan.
1This
article is one of a selection of papers published in the
Special Forum IUFRO 1.05 Uneven-Aged Silvicultural
Research Group Conference on Natural Disturbance-Based
Silviculture: Managing for Complexity.
2Corresponding author (e-mail: [email protected]).
Can. J. For. Res. 37: 1554–1562 (2007)
the maintenance of the original structure and composition
of the forest (Deal 2001; Yoshida et al. 2005a). Consequently, special emphasis has been put on partial cut harvesting systems that are applicable to the regeneration of
mixed tree species communities with an uneven-aged structure. (Greene et al. 1999).
Fine-scale canopy openings (canopy gaps) induced by
windthrow or other wind-caused tree damage have been recognized as a major force in natural stand dynamics (Borman
and Likens 1979; Pickett and White 1985). Selection harvesting can emulate the disturbance (Deal and Tappeiner
2002; Lundquist and Beatty 2002; Perera et al. 2004). Openings generally improve resource availability in the understory, thereby enhancing tree regeneration (Runkle 1981;
Poulson and Platt 1989; Kneeshaw and Bergeron 1998).
However, in forests with rich understory cover, tree regeneration under a canopy gap is often hindered through enhanced development of the understory vegetation
(Lautenschlager 1999; Beckage et al. 2000; Royo and Carson 2006). To conserve the structure and composition of
such a forest under a management regime, it is crucial to
clarify whether harvesting (creation of canopy gaps) contrib-
doi:10.1139/X07-103
#
2007 NRC Canada
Noguchi and Yoshida
utes to the regeneration of indigenous tree species over the
long term and to understand what kind of functional traits
of tree species are important for establishment and growth
in these stands.
In addition to canopy openings, practices associated with
harvesting also play an important role in developing tree regeneration. Soil disturbances, which appear naturally as
landslides or treefall mounds (Nakashizuka 1989; Peterson
et al. 1990; Clinton and Baker 2000), can be induced by machinery skidding or site preparation practices (Heninger et
al. 2002; Buckley et al. 2003; Yoshida et al. 2005b).
Uneven-aged conifer–broadleaved mixed forests (including
Abies, Picea, Acer, Quercus, Tilia, Betula, etc.) are a typical
forest type in Hokkaido, the northernmost island of Japan.
A selection system (single-tree selection) has been widely
employed since the early 20th century to yield sustainable
timber of both conifer and broadleaved species. The forests
generally have dense understory cover dominated by dwarf
bamboos (Sasa spp.), and therefore, there is concern about
regeneration failure following the selection harvesting.
However, few studies have explicitly clarified the longterm consequences of the practices on the natural regeneration of component tree species.
The objective of this study was to assess the effects of
canopy disturbance induced by selection harvesting and its
associated practices (artificial planting and machinery skidding) on the successful regeneration of common tree species
in northern Japanese mixed forests. Long-term effects of
canopy disturbance history (i.e., continuation of open or
closed conditions of the canopy in the most recent 30 years)
were examined using large permanent study stand data
(6.7 ha area) along with spatial information of canopy trees
(M. Noguchi and T. Yoshida, unpublished). The following
two questions were addressed. (i) Are regeneration densities
under canopy gaps larger than those under closed canopies?
We predicted that canopy gaps would have a limited contribution to regeneration even in shade-tolerant species because of the exclusive dominance of dwarf bamboos in the
understory. (ii) Are regeneration densities in sites affected
by surface soil disturbances created during practices such as
artificial planting and skidding? We predicted that these
practices would considerably enhance regeneration.
Materials and methods
Study site
The present study was conducted in a conifer–broadleaved
mixed forest in the Nakagawa Experimental Forest of Hokkaido University located in Otoineppu Village, Hokkaido
Prefecture (44848’N, 142815’E). The mean annual temperature is 5.4 8C, the mean annual precipitation is 1449 mm,
and the mean maximum snow depth is 190 cm at the nearest
meteorological station (8 km from the study site). Small
ridges and valleys running southwest from a main ridge on
the northwest edge of the stand characterize the topography.
The soils are classified as Inceptisol (acidic brown forest
soil), and the predominant bedrock is Cretaceous sedimentary rock (Tatewaki and Igarashi 1971).
We selected a 6.7 ha stand for this study because its management history was completely recorded over a 30-year period at an individual tree scale (see below). The original
1555
stand (at the first census in 1974) had an uneven-aged structure with a basal area of 27.9 m2ha–1 that was typical of the
region (Takahashi et al. 2003; Yoshida et al. 2006). Before
the plot establishment, the stand had been partially harvested
for domestic fuel materials at a very low intensity. No largescale natural disturbances such as windthrows or pest outbreaks were recorded during the study period, which enabled
us to clearly evaluate the effects of selection harvesting.
The stand contained 21 tall tree species in all (three evergreen conifers and 18 deciduous broadleaved species) of
which Abies sachalinensis (Fr. Schm.) Masters comprised
nearly one third of the density followed by Acer mono
Maxim., Tilia japonica (Miq.) Simonkai, Quercus crispula
Bl., Betula ermanii Cham., and others (Table 1). The understory was densely covered with dwarf bamboo species (Sasa
kurilensis (Rupr.) Makino et Shibata and Sasa senanensis
(Franch. et Savat.) Rehd.).
Silvicultural practices
Single-tree selection harvesting has been conducted at 9to 10-year intervals (1975, 1984, and 1994) with respective
harvesting intensities of approximately 12%, 17%, and 8%
in basal area. These intensities were determined carefully in
consideration of the census data to maintain harvested volume equivalent to the growth increment. Although little
change has been recorded in basal areas over the stand because harvest and growth were at similar levels during the
interval (Ohgane et al. 1988; Yoshida et al. 2006), considerable local variation exists within the stand (M. Noguchi and
T. Yoshida, unpublished). In these harvestings, the priority
in cutting was given to trees with a lower timber value (i.e.,
with injuries, cracks, etc.) to improve the future economic
quality of the stand. Nevertheless, the harvested trees were
selected from a broad range of size classes and no preference was shown for any particular tree species, e.g., the total
number of harvested conifers and that of harvested broadleaved trees were roughly equal.
Harvesting was carried out in winter with snow cover normally 100–200 cm deep on the forest floor. Therefore, the
harvesting processes seldom damaged understory vegetation
or the soil directly. However, skid trails with 2–3 m widths
exist where considerable surface soil (O and A layers) was
removed by a tractor during skidding processes. At the current census (conducted in 2003), 9 years had passed from
the most recent trail use. On the other hand, tree seedlings
were planted artificially to fill some large nonwooded openings (aprroximately >0.2 ha in area) in the stand. In 1976
and 1986, seedlings of A. sachalinensis were planted with
initial densities of 1800–4000 stemsha–1. The preplanting
site preparations involved exposing mineral soil by hand for
each planting microsite. Subsequently, a weeding treatment,
in which aboveground parts of plants competing with seedlings were cut by hand, was applied once per year during
the initial 10 years. Canopies of these plantations had already been closed. Because the variation in environmental
conditions among the plantations seemed to be small, we
used them as a treatment in this study.
Field data collection
The uneven structure of the stand, together with a spatially dispersed distribution of harvested trees with a broad
#
2007 NRC Canada
1556
Can. J. For. Res. Vol. 37, 2007
Table 1. Tree species composition in the 6.7 ha study stand; stem densities of overstory trees (DBH ‡ 12.5 cm), saplings
(DBH < 12.5 cm and height ‡ 2 m), and seedlings (height < 2 m).
Density
Species
Abbreviation
Groupa
Seed dispesal
agent
Abies sachalinensis
Acer mono
Tilia japonica
Quercus crispula
Betula ermanii
Alnus hirsuta
Kalopanax pictus
Magnolia obovata
Sorbus commixta
Phellodendron amurense
Prunus sargentii
Salix spp.
Ulmus laciniata
Ulmus davidiana var. japonica
Sorbus alnifolia
Fraxinus mandshurica
Taxus cuspidata
Swida controversa
Prunus ssiori
Acanthopanax sciadophylloides
Picea jezoensis
Total
AS
AM
TJ
QC
BE
AH
KP
MO
SO
PA
PS
SL
UL
UD
SA
FM
TC
SW
PR
AC
PJ
CF
BT
BT
BI
BI
BI
BI
BI
BT
BI
BT
BI
BT
BT
BT
BT
CF
BI
BT
BT
CF
Wind
Wind
Wind
Animal
Wind
Wind
Bird
Bird
Bird
Bird
Bird
Wind
Wind
Wind
Bird
Wind
Bird
Bird
Bird
Bird
Wind
Overstory
stemsha–1
Sapling
stemsha–1
Seedling
stemsha–1
134
59
43
41
41
18
13
12
10
10
5
5
4
4
4
2
2
2
1
1
0
411
110 (392)
344 (809)
206 (793)
74 (231)
89 (389)
43 (250)
60 (257)
92 (371)
89 (375)
39 (135)
0 (–)
4 (42)
25 (138)
7 (84)
39 (208)
4 (42)
18 (110)
14 (103)
11 (94)
4 (42)
4 (42)
1273 (1768)
1014 (2993)
326 (1210)
319 (1880)
149 (520)
14 (168)
85 (554)
14 (168)
362 (2146)
241 (1121)
35 (251)
7 (84)
0 (–)
213 (1428)
64 (418)
7 (84)
99 (768)
78 (479)
0 (–)
21 (253)
0 (–)
0 (–)
3048 (5559)
Note: For saplings and seedlings, the averages and the standard deviations (in parentheses) of the 141 plots that have not experienced marked soil
disturbances are shown.
a
CF, conifer; BT, shade-tolerant broadleaved; BI, shade-intolerant broadleaved species. The classification is based on Koike (1988) and Masaki
(2002).
range of size classes, led to the introduction of considerable
variation in environmental conditions within the stand (M.
Noguchi and T. Yoshida, unpublished). Although we could
not make a stand replication in this study, the following intensive sampling, along with historical records, enabled us
to evaluate the effects of the local variation.
Long-term monitoring of trees has been conducted since
1974 in the study stand. All living overstory trees
(‡12.5 cm diameter at breast height (DBH)) were identified
and classed by DBH at intervals of 5 cm. The survey was
repeated at approximately 10-year intervals (1974, 1983,
1993, and 2003). During 2000–2002, we mapped all of the
living overstory trees and the identifiable dead trees (including cut stumps) in the whole stand.
We sampled the saplings (height ‡ 2 m and DBH < 12.5 cm)
and seedlings (height < 2 m, except for current-year seedlings) at 163 plots set up at 20 m intervals in the 6.7 ha
stand. In the summer of 2003, saplings were identified and
densities recorded in the 20 m2 circular subplot for each
plot. Similarly, seedlings with heights ‡ 50 cm and those
with heights < 50 cm were recorded, respectively, in the
nested 10 and 5 m2 circular subplots.
We recorded the forest floor condition of each sampling
plot: (i) intact, (ii) disturbed to create a plantation, or
(iii) disturbed by a skid trail. The regular setup (at 20 m intervals) of the plots obtained small samplings for the latter
two. Therefore, we made 10 additional plots for plantations
and eight for skid trails. Plots that had undergone a marked
natural disturbance (such as treefall mounds) were excluded;
such plots were very few, probably because of the repetitive
partial harvestings (Noguchi and Yoshida 2004).
For the plots with intact forest floors (141 plots), we also
recorded the current canopy condition (gap or closed). In
this study, the canopy gap was defined as an area without
tree crowns larger than 5 m in height and 10 m2 in area.
The plot was designated as a ‘‘canopy gap’’ when such an
opening was found within the surrounding 20 m2 area. We
then characterized the canopy condition with the past disturbance history, which was derived from the long-term tree
census and their location (coordinate) data. Within the circular area with a radius of 7.5 m (roughly corresponding to the
largest crown radius of the trees) from the plot center, we
checked the occurrence of harvesting and the natural death
of large trees (trees with DBH ‡ 32.5 cm, which are assumed to attain the canopy layer) during the total census period (30 years). The plots were then categorized into four
classes: (i) canopy gaps with a recorded disturbance, where
harvesting or natural death in the 30-year period was assumed to contribute to the current canopy condition,
(ii) canopy gaps without a recorded disturbance, where gaps
predated the first census 30 years before, (iii) closed canopies with a recorded disturbance, where the canopy briefly
opened but closed again in the period, and (iv) closed canopies without a recorded opening, where canopy closure was
maintained in the period.
Data analysis
To elucidate the effects of the canopy opening, we exam#
2007 NRC Canada
Noguchi and Yoshida
ined only the sampling plots without soil disturbances (141
plots). The w2 test was used to test the preference in frequency occurrence of saplings or seedlings of major component species for sites under canopy gaps (72 plots) compared
with those under closed canopies (69 plots). In the case
when the frequency occurrence was small, Fisher’s exact
test was applied. Regarding sites under canopy gaps and
closed canopies, respectively, similar comparisons were
made for sites with and without recorded disturbances during the most recent 30 years.
Although gap size is often used to characterize canopy
gaps, it is difficult to measure their size exactly in forests
of this region because of their continuous spatial distribution, which is attributable to a high gap rate (frequently >20%) (Kubota 2000; Takahashi et al. 2003).
Therefore, to examine the comparisons of regeneration densities among classes, we used regression analysis, which incorporates the basal area of surrounding overstory trees
(roughly corresponding to the gap area) as explanatory variables. Because of the total low densities and frequent absence of regenerated trees in the particular plots, a logistic
regression model was employed. The occurrence probability
of saplings or seedlings in a plot (poc) was modeled as
½1
logðpoc =ð1 poc ÞÞ ¼ 1 þ 2 CFBA þ 3
BLBA þ 4 CGBA
where CFBA and BLBA, respectively, denote the sums of
basal area in the surrounding area (7.5 m radius from the
plot center) in 1974 of coniferous and broadleaved canopy
trees (DBH ‡ 32.5 cm). In addition, CGBA is the change in
the sum of basal area of the surrounding canopy trees during
1974–1993 that represents the intensity of disturbances (harvesting and natural death) in its balance of basal area. Explanatory variables were selected using a stepwise
procedure.
In addition, we compared the frequency occurrence in
plots that were disturbed by associated practices (21 plantations and 19 skid trail sites) with intact plots. We used the R
2.2.1 statistical package (R Development Core Team 2005)
for these analyses.
Results
In terms of stem density, the most dominant species in the
overstory (DBH ‡ 12.5 cm) was A. sachalinensis followed
by A. mono, T. japonica, Q. crispula, and B. ermanii
(Table 1). Among the saplings (height ‡ 2 m and DBH <
12.5 cm, 1276 stemsha–1), A. sachalinensis occurred in a
considerably lower proportion compared with the overstory,
and A. mono and T. japonica were dominant. The proportion
of seedlings (height < 2 m, 3048 stemsha–1) of A. sachalinensis was similar to that in the overstory but was considerably lower for B. ermanii and Kalopanax pictus (Thunb.)
Nakai. In contrast, Sorbus commixta Hedl. and Ulmus laciniata (Trautv.) Mayr occurred in high proportions despite
their only slight dominance in the overstory.
Among the saplings, B. ermanii showed a significant preference for sites under canopy gaps, but A. sachalinensis,
A. mono, and Magnolia obovata Thunb. occurred significantly more frequently in sites under closed canopies
(Fig. 1a). In seedlings, A. sachalinensis again showed a sig-
1557
nificant preference for closed canopies, similarly to T. japonica, S. commixta, and Q. crispula (Fig. 1b). In contrast,
seedlings of Ulmus davidiana Planch. var. japonica (Rehd.)
Nakai and Fraxinus mandshurica Rupr. var. japonica
Maxim. occurred primarily in canopy gap sites.
Regarding the canopy disturbance history of the most recent 30 years, a significant preference for canopy gaps with
a recorded disturbance was found only in saplings of B. ermanii (Fig. 2a) and seedlings of M. obovata (Fig. 2b). In
contrast, seedlings of A. mono in gaps were more frequent
at sites without a recorded disturbance. Those of U. davidiana and F. mandshurica occurred only in canopy gaps without a recorded disturbance.
Under the current closed canopy, saplings of A. sachalinensis and Q. crispula were more frequent at sites with a recorded disturbance (Fig. 2c); the opposite preference was
not found. Among seedlings, many species (A. sachalinensis, T. japonica, S. commixta, and M. obovata) showed a
preference for closed canopies with a recorded disturbance
(Fig. 2d), but few species showed the opposite preference.
Logistic regression analyses suggested that the change in
the surrounding basal area had a positive influence on regeneration densities: the coefficient was significant for saplings
and seedlings of A. sachalinensis (marginally significant for
saplings of A. mono and seedlings of A. mono, Q. crispula,
and S. commixta). In addition, the basal area of surrounding
conifers appeared to be influential for many species; a significant positive coefficient was found for saplings of
Q. crispula, A. sachalinensis, and K. pictus and for seedlings
of A. sachalinensis (marginally significant for saplings of
S. commixta and seedlings of A. mono). The basal area of
the surrounding broadleaved trees had a significant negative
coefficient for saplings of S. commixta.
Saplings of B. ermanii, Phellodendron amurense Rupr.,
and K. pictus appeared more frequently in plantations,
whereas those of A. mono and T. japonica were more frequent in intact sites (Fig. 3a). Alnus hirsuta Turcz was the
only species whose saplings showed a significant preference
for skid trails (Fig. 3c); it did not occur at all in plantations.
In seedlings, a significant preference for disturbed sites was
common, particularly in A. sachalinensis but also in
A. mono, S. commixta, P. amurense, and Q. crispula
(Figs. 3b and 3d). Similarly to its saplings, seedlings of
A. hirsuta did not occur at all in plantations.
Discussion
Effects of canopy opening
The total regeneration density (sum of saplings and seedlings) of this stand was 4324 stemsha–1 (Table 1), which
was low in comparison with similar mixed stands with less
dwarf bamboo dominance (Ishida et al. 1991; Nagaike et al.
1999). In all, the regeneration of tall tree species occurred
more frequently under closed canopies than under canopy
gaps in this mixed forest. This trend was found not only for
shade-tolerant species but also for shade-intolerant species
in this study; only a few species showed an occurrence preference for sites under canopy gaps (Fig. 1). Regression
analyses also indicated positive effects of the change in surrounding basal area (indicating the negative effect of canopy
opening) in many cases (Table 2). Particularly under closed
#
2007 NRC Canada
1558
Can. J. For. Res. Vol. 37, 2007
Fig. 1. Densities of (a) saplings and (b) seedlings with respect to the current condition; mean densities under the canopy gap were plotted
against those under the closed canopy. The equivalent line is shown. Asterisks indicate the significant preference for the particular site
(w2 test or Fisher’s exact test at p < 0.05). Abbreviations of species name are listed in Table 1.
Fig. 2. Densities of (a and c) saplings and (b and d) seedlings with respect to the current condition (gap or closed) together with the past
disturbance history; mean densities in sites with recorded disturbance were plotted against those in sites without disturbance (see text for
details). The equivalent line is shown. Asterisks indicate the significant preference for the particular site (w2 test or Fisher’s exact test at p <
0.05). Abbreviations of species name are listed in Table 1.
#
2007 NRC Canada
Noguchi and Yoshida
1559
Fig. 3. Densities of (a and c) saplings and (b and d) seedlings with respect to the practices associated with harvesting (plantation or skid
trail); mean densities in plantation or on skid trail were plotted against those on intact forest floor. The equivalent line is shown. Asterisks
indicate the significant preference for the particular site (w2 test or Fisher’s exact test at p < 0.05). Abbreviations of species name are listed
in Table 1.
canopies, the sites that had once experienced canopy disturbance (i.e., opened briefly but again closed within the
last 30 years) had higher regeneration densities, especially
in seedlings (Figs. 2c and 2d). On the other hand, under current canopy gaps, canopy disturbance in the last 30 years did
not seem to contribute to the regeneration of many of these
tree species (Figs. 2a and 2b). These results strongly suggested that canopy gaps in the forest can enhance the establishment and growth of tall tree species only when the gap
size is sufficiently small to allow its rapid closure. The potential positive effects of canopy openings on regeneration
seemed to be neutralized, and negative effects were rather
apparently attributable to the enhanced growth of understory
dwarf bamboos (Sasa spp.). The mean density and height of
dwarf bamboos, respectively, reached 14.7 culmsm–2 (SD =
11.3) and 176.7 cm (SD = 43.5) and the latter showed a significant increase with canopy openings in this stand (Noguchi and Yoshida 2005). Several researchers have reported
the negative effects of the abundance of dwarf bamboos on
tree regeneration as well as on plant species diversity (Hiura
et al. 1996; Nagaike et al. 1999; Noguchi and Yoshida
2004).
Tree species composition differed considerably among
overstory trees (DBH ‡ 12.5 cm), saplings (height ‡ 2 m
and DBH < 12.5 cm), and seedlings (height < 2 m). Proportions of A. sachalinensis, a shade-tolerant conifer, were
much lower in saplings at all densities than in the overstory
or seedlings: the distinct lower density of saplings suggested
limited recruitment from the seedling stage. The trait of this
species to survive under shaded conditions by reducing its
extension growth (Yajima 1982; a similar trait was reported
for other Abies species in Hara et al. (1991) and Messier et
al. (1999)) might cause limited recruitment because seedlings under a canopy gap must compete with dense dwarf
bamboo (the foliage crown of which is actually densest at
50–150 cm in height) during their recruitment process. Saplings and seedlings of A. sachalinensis showed a clear preference for sites under closed canopies, especially where the
canopy had been opened within the last 30 years. This fact
strongly suggests the importance of small-scale disturbances
in this type of forest. Results of regression analyses, where
the surrounding basal area (especially that of conifers) as
well as the change in basal area were selected as positive
coefficients (Table 2), also supported this tendency. The
lower mortality of overstory trees, which has frequently
been observed in natural stands in this region (<1%year–1:
Hiura and Fujiwara 1999; Kubota 2000), might contribute
to maintaining the population of this species.
#
2007 NRC Canada
1560
Can. J. For. Res. Vol. 37, 2007
Table 2. Results of the stepwise logistic regression analyses for the occurrence of understory trees or saplings of major tree species.
Species
No. of plots
with occurrence
AIC
Selected variables in the
best-fit model
Saplings
Acer mono
Tilia japonica
Quercus crispula
Abies sachalinensis
Kalopanax pictus
Sorbus commixta
51
16
16
15
12
11
of
of
of
of
of
of
141
141
141
141
141
141
185.1
99.7
96.7
90.1
82.3
78.4
+CFBA, +CGBAa
–BLBAa
+CFBAc
+CFBA,b +CGBAb
+CFBAb
+CFBA,a –BLBA,b +CGBA
Seedlings
Abies sachalinensis
Acer mono
Quercus crispula
Sorbus commixta
34
16
11
11
of
of
of
of
141
141
141
141
153.0
100.8
77.6
77.3
+CFBA,b –BLBA, +CGBAb
+CFBA,a +CGBAa
+CGBAa
+CGBAa
Note: Number of quadrats with the occurrence, the Akaike information criterion (AIC) for the
best-fit model, and selected variables in the model are shown. CFBA and BLBA are the sum of
basal area in the surrounding area in 1974 of canopy conifers and broadleaved trees, respectively.
CGBA is the change in basal area during 1974–1993. ‘‘+’’ and ‘‘–’’ indicate the positive and negative coefficient, respectively. No variables were selected for saplings of B. ermanii, M. obovata,
P. amurense, and S. alnifolia and seedlings of T. japonica.
a
p < 0.10.
b
p < 0.05.
c
p < 0.01.
Seedlings of shade-intolerant broadleaved species
(Q. crispula, B. ermanii, A. hirsuta, and K. pictus) occurred
less frequently (Table 1) in accordance with the general expectation that they have high mortality in shaded conditions.
Nevertheless, the density of their saplings was not low
(Table 1). Probably, once such trees exceed the height of
the dense dwarf bamboo cover, they grow rapidly (Hiura et
al. 1996). Betula ermanii, which showed the most distinct
early successional traits among these species (Koike 1988),
was the only species whose saplings occurred more frequently under a canopy gap (Fig. 1a). The rapid growth response seemed to be a crucial advantage in competing with
dwarf bamboos in canopy gaps.
Seedlings of Q. crispula frequently occurred under closed
canopies (Fig. 1b). Seed limitation might be apparently important for this species because of the limited seed (acorn)
supply in large openings and the increased seed predation in
canopy gaps from the enhanced activities of rodents (Wada
1993). Actually, our supplemental analysis (w2 test at p <
0.05) showed that its seedlings were significantly frequent
in sites with conspecific overstory trees in the surrounding
area (7.5 m radius). This frequency might partly explain the
higher density under closed canopies. Similar preferences
for the sites with conspecific overstory trees were apparent
for other common species (saplings of A. sachalinensis,
T. japonica, and B. ermanii and seedlings of A. sachalinensis, S. commixta, and A. hirsuta). This needs further consideration for spatial information (e.g., the location and
distance from seed trees, and prevailing wind directions) to
clarify the detailed effects of seed dispersal on regeneration
densities.
Shade-tolerant broadleaved species (such as A. mono,
T. japonica, and S. commixta), which can be expected to
survive and maintain growth under shaded understories
(e.g., Canham 1988; Kobe and Coates 1997), showed high
dominance among both saplings and seedlings (Table 1).
The most dominant sapling species, A. mono, did not generally respond to a canopy disturbance history. Nevertheless,
considering only seedlings in canopy gaps (Fig. 2b), the occurrence was rather frequent in gaps without a recorded canopy disturbance (i.e., sites where gap conditions were
maintained for at least 30 years). A similar trend was also
found in two other shade-tolerant species, U. davidiana and
F. mandschurica. Although their densities were not necessarily high, these wind-dispersed shade-tolerant species could
be major successors in large openings.
Responses through sprouting were frequently found only
in T. japonica (data not shown), which was most frequent
at sites under closed canopies with a recorded disturbance
(Figs. 2c and 2d). Many saplings (approximately 60%) and
seedlings (approximately 30%) were recognized as having a
sprouting origin. The partial removal of stem(s) from a
multiple-stem individual seemed to enhance the growth of
the residual small stems.
One point we have missed in this study is the impact of
time; we did not consider the effect of the individual harvesting 9, 19, and 28 years before the survey. The duration
after the latest disturbance should be examined together with
the differences in growth rates among species. In this study
stand, to attain 12.5 cm DBH, shade-intolerant B. ermanii
took 38 years (SD = 8, n = 22) on average, whereas the
shade-tolerant A. sachalinensis and A. mono took 49 years
(SD = 20, n = 61) and 50 years (SD = 12; n = 42), respectively (H. Miya et al., unpublished). Long-term observations
of saplings and seedlings in relation to overstory dynamics
are imperatively needed for more accurate evaluations.
Effects of associated practices
The associated practices of harvesting (artificial planting
and machinery skidding) significantly increased the regener#
2007 NRC Canada
Noguchi and Yoshida
ation densities of various species. In a natural process, canopy gap creation is associated with the supply of mound-pit
topography as well as fallen logs, which often enhance opportunities for seedling establishment (Nakashizuka 1989;
Peterson et al. 1990; Clinton and Baker 2000). However, as
is widely recognized, repeated selection harvesting might
degrade these important structures (Noguchi and Yoshida
2004). Therefore, the contribution of mound-pits or fallen
logs to regeneration densities in this forest was very small
(only 1% of recorded regeneration).
Regarding saplings, the associated practices were more
preferable for shade-intolerant species (cf. Frey et al. 2003).
Nevertheless, in seedlings, positive effects were particularly
apparent in shade-tolerant species (in particular, A. sachalinensis and A. mono) (Figs. 3b and 3d). Such a contrast
seemed to result from the different growth rates of the two
groups. It has been reported that topsoil removal can contribute to the establishment of many species, especially
when the canopy is not completely opened (Resco de Dios
et al. 2005; Yoshida et al. 2005b). Alnus hirsuta was dominant both as saplings and as seedlings on skid trails but was
not found at all in plantations. This was presumably because
of its sensitivity to weeding treatments, which are generally
conducted during the initial 10 years after planting. Its
sprouting ability might be reduced after repeated aboveground cuttings (e.g., Hasegawa 1998).
Conclusion
Results of this study demonstrate that the effects of canopy
opening on the continuous regeneration of tall tree species
community were neutral or negative in this forest. The negative effect was rather common, especially when the canopy
does not close for a long period. For the selection harvesting
system in this region, more attention is required regarding the
spatial arrangement of residual trees to avoid large openings.
This would involve considering gap size and species composition surrounding the gap, which would influence the rate of
gap closure. On the other hand, we observed higher regeneration densities on skid trails and plantations, both of which
were not intended for natural regeneration but had undergone
soil disturbance. Regarding the small regeneration density in
the whole stand, artificial soil disturbance should be introduced to emulate the role of mound-pit topography. Surface
soil displacement using civil engineering machinery (soil
scarification), which has been utilized in the region for silvicultural site preparation, could be a convincing option. The
application of this treatment on sites under closed canopies
or small gaps appeared to enhance the regeneration of not
only shade-intolerant species but also shade-tolerant species
(Resco de Dios et al. 2005; Yoshida et al. 2005b).
We should note that the results of this study are limited
due to lack of replication of stands. Further studies are required to extrapolate our findings to a larger spatial scale,
and such trials would engender a sustainable forest management regime with continuous tree regeneration in this type
of mixed forest.
Acknowledgements
We would like to thank four anonymous reviewers for
their helpful comments and suggestions. We thank the staff
of the Nakagawa Experimental Forest and the members of
1561
the Laboratory of Forest Management, Hokkaido University,
for their combined efforts in maintaining this long-term
study. We are grateful to Takayoshi Koike, Shigeru Uemura,
Tsutomu Hiura, and Yasuhiro Kubota for helpful comments
on an earlier version of this manuscript. We also thank
Yoko Iga and Mio Nagai for their assistance with the fieldwork. This study was partly supported by the research project fund of the Ministry of Education, Culture, Sports,
Science, and Technology of Japan (No. 14760095) and by a
Sasakawa Scientific Research Grant from the Japan Science
Society.
References
Angelstam, P.K. 1998. Maintaining and restoring biodiversity in
European boreal forests by developing natural disturbance regimes. J. Veg. Sci. 9: 593–602. doi:10.2307/3237275.
Beckage, B., Clark, J.S., Clinton, B.D., and Haines, B.L. 2000. A
long-term study of tree seedling recruitment in southern Appalachian forests: the effects of canopy gaps and shrub understories.
Can. J. For. Res. 30: 1617–1631. doi:10.1139/cjfr-30-10-1617.
Borman, F.H., and Likens, G.E. 1979. Pattern and process in a
forested ecosystem. Springer-Verlag, New York.
Buckley, D.S., Crow, T.R., Nauertz, E.A., and Schulz, K.E. 2003.
Influence of skid trails and haul roads on understory plant richness and composition in managed forest landscapes in Upper
Michigan, USA. For. Ecol. Manag. 175: 509–520. doi:10.1016/
S0378-1127(02)00185-8.
Canham, C.D. 1988. Growth and canopy architecture of shadetolerant trees: response to canopy gaps. Ecology, 69: 786–795.
doi:10.2307/1941027.
Clinton, B.D., and Baker, C.R. 2000. Catastrophic windthrow in the
southern Appalachians: characteristics of pits and mounds and
initial vegetation responses. For. Ecol. Manag. 126: 51–60.
doi:10.1016/S0378-1127(99)00082-1.
Deal, R.L. 2001. The effects of partial cutting on forest plant communities of western hemlock – Sitka spruce stands in southeast
Alaska. Can. J. For. Res. 31: 2067–2079. doi:10.1139/cjfr-3112-2067.
Deal, R.L., and Tappeiner, J.C. 2002. The effects of partial cutting
on stand structure and growth of western hemlock – Sitka spruce
stands in southeast Alaska. For. Ecol. Manag. 159: 173–186.
doi:10.1016/S0378-1127(00)00727-1.
Frey, B.R., Lieffers, V.J., Munson, A.D., and Blenis, P.V. 2003.
The influence of partial harvesting and forest floor disturbance
on nutrient availability and understory vegetation in boreal
mixed woods. Can. J. For. Res. 33: 1180–1188. doi:10.1139/
x03-042.
Greene, D.F., Zasada, J.C., Sirois, L., Kneeshaw, D., Morin, H.,
Charron, I., and Simard, M.J. 1999. A review of the regeneration dynamics of North American boreal forest tree species.
Can. J. For. Res. 29: 824–839. doi:10.1139/cjfr-29-6-824.
Hara, T., Kimura, M., and Kikuzawa, K. 1991. Growth patterns of
tree height and stem diameter in populations of Abies veitchii,
A. mariesii and Betula ermanii. J. Ecol. 79: 1085–1098.
Hasegawa, M. 1998. Effect of weeding and salvage cutting on
Monarch birch (Betula maximowicziana Regal) growing in Japanese cedar (Criptmeria japonica D. Don) plantations in snowy
regions. J. Jpn. For. Soc. 80: 223–228. [In Japanese with English
summary]
Heninger, R., Scott, W., Dobkowski, A., Miller, R., and Anderson,
H. 2002. Soil disturbance and 10-year growth response of coast
Douglas-fir on nontilled and tilled skid trails in the Oregon Cascades. Can. J. For. Res. 32: 233–246. doi:10.1139/x01-195.
#
2007 NRC Canada
1562
Hiura, T., and Fujiwara, K. 1999. Density-dependence and coexistence of conifer and broad-leaved trees in a Japanese
northern mixed forest. J. Veg. Sci. 10: 843–850. doi:10.2307/
3237309.
Hiura, T., Sano, J., and Konno, Y. 1996. Age structure and response to fine-scale disturbance of Abies sachalinensis, Picea jezoensis, Picea glehnii, and Betula ermanii growing under the
influence of a dwarf bamboo understory in northern Japan. Can.
J. For. Res. 26: 289–297.
Hunter, M.L., Jr. 1999. Maintaining biodiversity in forest ecosystems. Cambridge University Press, Cambridge, U.K.
Ishida, M., Kikuzawa, K., Asai, T., Mizui, N., and Seiwa, K. 1991.
Distribution and height growth of young trees in the canopy gap
and in the closed stand; the case of coniferous-broadleaved
mixed forest in Hokkaido. J. Jpn. For. Soc. 73: 145–150. [In Japanese with English summary]
Kneeshaw, D.D., and Bergeron, Y. 1998. Canopy gap characteristics and tree replacement in the southeastern boreal forest. Ecology, 79: 783–794.
Kobe, R.K., and Coates, K.D. 1997. Models of sapling mortality as
a function of growth to characterize interspecific variation in
shade tolerance of eight tree species of northwestern British Columbia. Can. J. For. Res. 27: 227–236. doi:10.1139/cjfr-27-2227.
Kohm, K.A., and Franklin, J.F. 1997. Creating a forestry for the
21st century: the science of ecosystem management. Island
Press, Washington, D.C.
Koike, T. 1988. Leaf structure and photosynthetic performance as
related to the forest succession of deciduous broad-leaved trees.
Plant Species Biol. 3: 77–87.
Kubota, Y. 2000. Spatial dynamics of regeneration in a conifer/
broad-leaved forest in northern Japan. J. Veg. Sci. 11: 633–640.
doi:10.2307/3236570.
Lautenschlager, R.A. 1999. Environmental resource interactions affect red raspberry growth and its completion with white spruce.
Can. J. For. Res. 29: 906–916. doi:10.1139/cjfr-29-7-906.
Lundquist, J.E., and Beatty, J.S. 2002. A method for characterizing
and mimicking forest canopy gaps caused by different disturbances. For. Sci. 48: 582–594.
Masaki, T. 2002. Structure and dynamics. In Diversity and interaction in a temperate forest community: Ogawa forest reserve of
Japan. Edited by T. Nakashizuka and Y. Matsumoto. SpringerVerlag, Tokyo. pp. 53–65.
Messier, C., Doucet, R., Ruel, J., Claveau, Y., Kelly, C., and
Lechowicz, J. 1999. Functional ecology of advance regeneration
in relation to light in boreal forests. Can. J. For. Res. 29:
812–823. doi:10.1139/cjfr-29-6-812.
Nagaike, T., Kubota, Y., and Watanabe, N. 1999. The effects of
selective logging on stand structure and the regeneration of
subboreal forests in Hokkaido, northern Japan. J. For. Res. 4:
41–45.
Nakashizuka, T. 1989. Role of uprooting in composition and
dynamics of an old-growth forest in Japan. Ecology, 70:
1273–1278. doi:10.2307/1938186.
Noguchi, M., and Yoshida, T. 2004. Tree regeneration in partially
cut conifer–hardwood mixed forests in northern Japan: roles of
establishment substrate and dwarf bamboo. For. Ecol. Manag.
190: 335–344. doi:10.1016/j.foreco.2003.10.024.
Noguchi, M., and Yoshida, T. 2005. Factors influencing the distribution of two co-occurring dwarf bamboo species (Sasa kurilensis and Sasa senanensis) in a conifer–broadleaved mixed stand
in northern Hokkaido. Ecol. Res. 20: 25–30. doi:10.1007/
s11284-004-0009-6.
Can. J. For. Res. Vol. 37, 2007
Ohgane, E., Nigi, T., Hishinuma, Y., Koshika, K., and Fukui, T.
1988. Managerial progress and results of an experimental forest
practiced by the Control Method. Res. Bull. College Exp. For.
Hokkaido Univ. 45: 61–113. [In Japanese with English summary]
Perera, A.H., Buse, L.J., and Weber, M.J. 2004. Emulating natural
forest landscape disturbances: concepts and applications. Columbia University Press, New York.
Peterson, C.J., Carson, W.P., McCarthy, B.C., and Pickett, S.T.A.
1990. Microsite variation and soil dynamics within newly created treefall pits and mounds. Oikos, 58: 39–46. doi:10.2307/
3565358.
Pickett, S.T.A., and White, P.S. 1985. The ecology of natural disturbance and patch dynamics. Academic Press, New York.
Poulson, T.L., and Platt, W.J. 1989. Gap light regimes influence
canopy tree diversity. Ecology, 70: 553–555. doi:10.2307/
1940202.
R Development Core Team. 2005. R: a language and environment
for statistical computing [online]. R Foundation for Statistical
Computing, Vienna, Austria. ISBN 3-900051-07-0, URL. Available from www.R-project.org.
Resco de Dios, V., Yoshida, T., and Iga, Y. 2005. Effects of topsoil
removal by soil-scarification on regeneration dynamics of mixed
forests in Hokkaido, northern Japan. For. Ecol. Manag. 215:
138–148.
Royo, A.A., and Carson, W.P. 2006. On the formation of dense understory layers in forests worldwide: consequences and implications for forest dynamic, biodiversity, and succession. Can. J.
For. Res. 36: 1345–1362. doi:10.1139/X06-025.
Runkle, J.R. 1981. Gap regeneration in some old-growth forests of
the eastern United States. Ecology, 62: 1041–1051. doi:10.2307/
1937003.
Takahashi, K., Mitsuishi, D., Uemura, S., Suzuki, J., and Hara, T.
2003. Stand structure and dynamics during a 16-year period in a
sub-boreal conifer–hardwood mixed forest, northern Japan. For.
Ecol. Manag. 174: 39–50. doi:10.1016/S0378-1127(02)00018-X.
Tatewaki, M., and Igarashi, T. 1971. Forest vegetation in the
Teshio and the Nakagawa District experimental forests of
Hokkaido University, Prov. Teshio, N. Hokkaido, Japan. Res.
Bull. College Exp. For. Hokkaido Univ. 28: 1–192. [In Japanese
with English summary]
Wada, N. 1993. Dwarf bamboos affect the regeneration of zoochorous
trees by providing habitats to acorn-feeding rodents. Oecologia,
94: 403–407. doi:10.1007/BF00317116.
Yajima, T. 1982. Study on the growth of main tree species in the
mixed forest of needle-leaved and broad-leaved trees. Res. Bull.
College Exp. For. Hokkaido Univ. 39: 1–54. [In Japanese with
English summary]
Yoshida, T., Hasegawa, M., Taira, H., and Noguchi, M. 2005a.
Stand structure and composition of a 60-year-old larch (Larix
kaempferi) plantation with retained hardwoods. J. For. Res. 10:
351–358. doi:10.1007/s10310-005-0154-y.
Yoshida, T., Iga, Y., Ozawa, M., Noguchi, M., and Shibata, H.
2005b. Factors influencing early vegetation establishment following soil scarification in a mixed forest in northern Japan.
Can. J. For. Res. 35: 175–188. doi:10.1139/x04-156.
Yoshida, T., Noguchi, M., Akibayashi, Y., Noda, M., Kadomatsu,
M., and Sasa, K. 2006. Twenty years of community dynamics
in a mixed conifer–broadleaved forest under a selection system
in northern Japan. Can. J. For. Res. 36: 1363–1375. doi:10.
1139/X06-041.
#
2007 NRC Canada