Twenty years of community dynamics in a mixedconifer – broad-leaved forest under a selectionsystem in northern Japan

Toshiya Yoshida, Mahoko Noguchi, Yukio Akibayashi, Masato Noda,Masahiko Kadomatsu, and Kaichiro Sasa

Abstract: Single-tree selection has been employed widely in northern Japanese mixed forests, but management-inducedchanges in forests are not well understood. This study examined demographic parameters of major tree species duringa 20-year study of a 68 ha stand in which single-tree selection has been conducted since 1971. Results showed thatgrowth and survival of conifers (mostly Abies sachalinensis (Fr. Schm.) Masters) was the most strongly positively af-fected by the treatment. Nevertheless, recruitment of conifers was not sufficiently improved, suggesting their decreaseddominance over the longer term. Instead, shade-intolerant broad-leaved species (mainly Betula ermanii Cham.) willgradually increase because of their higher recruitment rates after the treatment. Shade-tolerant broad-leaved species(mainly Acer mono Maxim. and Tilia japonica (Miq.) Simonkai) appeared to experience the most distinct negative ef-fects, especially on survival. These trends differed markedly from those reported in previous papers concerning partialharvesting systems, which predicted an increase in dominance of shade-tolerant species. The results shown here shouldbe generalized carefully because we have investigated only one stand without repetition of the control area. Neverthe-less, trends described in this large-scale, long-term study could provide a basis for simulating stand dynamics. We dis-cussed possible reasons for the observed patterns and provided implications for sustainable management in the region.

Résumé : Le jardinage par pied d’arbre a été largement utilisé dans les forêts mixtes du nord du Japon, mais les chan-gements induits par l’aménagement des forêts ne sont pas encore bien compris. Cette étude se penche sur les paramè-tres démographiques des principales espèces arborescentes sur une période de 20 ans dans un territoire d’étude de68 ha à l’intérieur duquel le jardinage par pied d’arbre est appliqué depuis 1971. Les résultats montrent que la crois-sance et la survie des conifères (surtout Abies sachalinensis (Fr. Schm.) Masters) ont le plus bénéficié du traitement.Néanmoins, le recrutement des conifères n’a pas été suffisamment amélioré, ce qui suggère une diminution de leur do-minance à long terme. En contrepartie, la présence de feuillus intolérants à l’ombre (surtout Betula ermanii Cham.)augmentera graduellement à cause de leur taux de recrutement plus élevé après le traitement. Les feuillus tolérants àl’ombre (surtout Acer mono Maxim. et Tilia japonica (Miq.) Simonkai) semblent connaître les effets négatifs les plusmarqués, particulièrement dans le cas du taux de survie. Ces tendances vont clairement à l’encontre de celles qui ontété publiées précédemment au sujet des systèmes de coupe partielle et qui prévoyaient plutôt une augmentation de laprésence d’espèces tolérantes à l’ombre. Les résultats devraient être généralisés avec précaution parce que les auteursn’ont étudié qu’un seul peuplement sans répétition de la zone témoin. Néanmoins, les tendances décrites dans cetteétude à long terme sur un grand territoire pourraient servir de base pour simuler la dynamique des peuplements. Ilsdiscutent des raisons qui pourraient expliquer les résultats qu’ils ont observés et faitent état des implications pourl’aménagement durable dans la région.

[Traduit par la Rédaction] Yoshida et al. 1375

1363

Can. J. For. Res. 36: 1363–1375 (2006) doi:10.1139/X06-041 © 2006 NRC Canada

Received 14 August 2005. Accepted 23 January 2006. Published on the NRC Research Press Web site at http://cjfr.nrc.ca on11 May 2006.

T. Yoshida.1 Uryu Experimental Forest, Field Science Center for Northern Biosphere, Hokkaido University, Moshiri, Horokanai074-0741, Japan.M. Noguchi. Faculty of Environmental Earth Science, Hokkaido University, Tokuda 250, Nayoro 096-0071, Japan.Y. Akibayashi and K. Sasa. Field Science Center for Northern Biosphere, Hokkaido University, Sapporo 060-0809, Japan.M. Noda. Wakayama Experimental Forest, Field Science Center for Northern Biosphere, Hokkaido University, Hirai, Kozakawa649-4563, Japan.M. Kadomatsu. Nakagawa Experimental Forest, Field Science Center for Northern Biosphere, Hokkaido University, Otoineppu098-2501, Japan.

1Corresponding author (e-mail: yoto@fsc.hokudai.ac.jp).

Introduction

Recently, the use of partial harvesting systems is raisingconcerns about managing forests both for timber productionand ecosystem functions (Oliver and Larson 1996; Kohmand Franklin 1997; von Gadow et al. 2002) because it oftenengenders stands with complex structure and composition(O’Hara 1998; Deal 2001; Deal and Tappeiner 2002; Harveyet al. 2002). In a northern Japanese mixed conifer – broad-leaved forest (including Abies, Picea, Acer, Tilia, Quercus,Betula, etc.), a typical forest type of that region (Tatewaki1958), a selection system (single-tree selection) has beenwidely employed since its exploitation began in the early20th century as a superior option to manage natural forests.Most forests have experienced at least one instance of har-vesting in the past. However, despite basic silvicultural prin-ciples (e.g., Smith et al. 1997), actual harvesting practiceswere often excessive and showed preferential extraction oflarge trees (Nagaike et al. 1999), thereby degrading regionalforests. Most previous studies of harvesting practices innorthern Japan have emphasized sustainable yields (Ohganeet al. 1988; Nigi et al. 1998); few studies have examinedbroader objectives, including maintenance of biodiversity.Contributing factors for ecosystem integrity (Hunter 1999),such as changes in stand structure and tree species composi-tion, have not been well understood.

Sustainable forestry attempts to mimic natural distur-bances through management practices (Spence 2001; Frank-lin et al. 2002; Mitchell et al. 2002). The natural disturbanceregime of northern Japanese mixed forests is a combinationof rare large-scale blowdown (>10 000 m2; note that fire isunimportant in this region) and frequent small-scale treedeath (<1000 m2: Ishikawa and Ito 1989; Kubota 1995;Nakashizuka and Iida 1995). Although little information isavailable regarding the large-scale disturbances in this re-gion, several previous studies have predicted that the estab-lishment of some major component tree species, includingBetula ermanii Cham., Quercus crispula Bl., and Abiessachalinensis (Fr. Schm.) Masters, is closely related to large-scale disturbances (Suzuki et al. 1987; Sano 1988; Ishikawaand Ito 1989; Osawa 1992; Ishizuka et al. 1998). On theother hand, as shown in many other regions (Hibbs 1982;Runkle 1985; Tanaka and Nakashizuka 1998), small-scaledisturbances have been shown to have a fundamental influ-ence on the structure of these natural forests. Natural treemortalities in terms of basal area observed in this regionwere frequently less than 1% per year (Hiura and Fujiwara1999; Kubota 2000); some major tree species (Abiessachalinensis, Acer mono Maxim., and Sorbus commixtaHedl.) regenerate in the resultant canopy gaps (Ishizuka1984; Kubota et al. 1994; Hiura et al. 1996; Kubota 2000;Takahashi et al. 2003).

The objective of this study was to present implications forsustainable harvesting systems, particularly through evaluat-ing the effects of harvesting on tree species composition.Single-tree selection superficially resembles a small-scaledisturbance regime (creation of canopy gaps), but it engen-ders some fundamental differences, such as removal ofstems and the consequent reduction of fallen trees (Noguchiand Yoshida 2004). In a mixed forest, we can therefore pre-dict that single-tree selection generally enhances the estab-

lishment and growth of particular species that adapt ecologi-cally to small canopy gaps (cf. Runkle 1982; Canham andMarks 1985; Runkle and Yetter 1987; Abe et al. 1995;Messier et al. 1999). In fact, some North American studieshave reported that dominance of shade-tolerant species grad-ually increases after single-tree selection (Leak and Fillip1977; Jenkins and Parker 1999; Webster and Lorimer 2002;Shuler 2004). Nevertheless, such a trend in compositionalchange may differ among forest types. Constituent speciesprobably adapt to inherent environmental conditions andinterspecific relationships, both of which can be strongly af-fected by harvesting treatment. The northern Japanese mixedforest is characterized by a relatively large gap area in thecanopy layer, which is often more than 20% (Kubota 1995,2000; Takahashi et al. 2003). This characteristic is conspicu-ously apparent in heavy-snowfall regions; it is inferred to berelated to slow recruitment rate caused by dense undercoverof dwarf bamboo species (Sasa kurilensis (Rupr.) Makino &Shibata or Sasa senanensis (Franch. & Savat.) Rehd.)(Takahashi 1997; Umeki and Kikuzawa 1999). Small treesmay not respond directly to single-tree selection becausedwarf bamboos increasingly dominate the canopy gaps(Noguchi and Yoshida 2005; cf. Messier et al. 1998).

We addressed the following questions in this study:(1) Does single-tree selection change tree species composi-tion? (2) If change occurs, does it correspond to the expecta-tions derived from characteristics (e.g., shade-tolerance) ofconstituent species? We examined 20 years of communitydynamics of a mixed conifer – broad-leaved forest in north-ern Japan under a selection system by which single-tree se-lection at 10-year intervals had been conducted since 1971.We compared demographic parameters of major componenttree species (and species groups) between treated and un-treated areas based on a repeated census of ca. 27 000 indi-viduals in a large (68 ha) plot. We summarized speciesrecovery for that period and discussed long-term effects ofsingle-tree selection on stand structure and species composi-tion in the region.

Methods

This study was conducted in the Nakagawa ExperimentalForest of Hokkaido University (44°48′N, 142°14′E, 150 melevation; Fig. 1). The forest is located in a cool-temperateregion with heavy snowfall in winter. Typical vegetation inthe Nakagawa forest is mixed conifer – broad-leaved forest,as described by Tatewaki and Igarashi (1971). This type offorest is widely distributed in northeastern Asia as represen-tative natural vegetation (Tatewaki 1958) and is responsiblefor regional biodiversity. The mean annual temperature is5.4 °C and the mean precipitation is 1449 mm. Nearly halfof the precipitation falls in the nongrowing season(November–April). Snow cover usually extends from earlyNovember to early May, with a maximum depth of 200 cm.The soils are classified as Inceptisol (acidic brown forestsoil), and the predominant bedrock is Cretaceous sedimen-tary rock (Tatewaki and Igarashi 1971).

Since 1967, a long-term study has been conducted at a110 ha study site (Fig. 1, Table 1). Few large-scale distur-bances or outbreaks were recorded during the study period,enabling us to clearly evaluate the effects of selection har-

© 2006 NRC Canada

1364 Can. J. For. Res. Vol. 36, 2006

vesting. The primary objective of the original study was todevelop a relevant management system to extract sustainableyield (Ohgane et al. 1988; Nigi et al. 1998). The “checkmethod” presented by Knuchel (1953), which was intendedto maintain timber volume and quality in the stand, was ap-plied (see later paragraph for detailed practices).

The original stand had a multiaged structure that was sim-ilar to that of other mixed forests in the region (Suzuki et al.1987; Takahashi et al. 2003). Before plot establishment, thestand had been partially harvested at very low intensity fordomestic fuel materials. The sum of the basal area at thefirst census (i.e., preharvest basal area) was 29.9 m2·ha–1,which was typical of the region. There were 27 tall-tree spe-cies in all (4 evergreen conifers and 23 deciduous broad-leaved species), of which Abies sachalinensis shared nearlyhalf of the total basal area, with lesser areas occupied, indescending order, by Acer mono, Tilia japonica (Miq.)

Simonkai, Quercus crispula, Betula ermanii, and others (seeTable 2 for a complete list).

The forest was divided into 11 compartments (areas rang-ing from 5.6 to 17.6 ha) including a control compartment(5.9 ha). The dominant species showed no strong prefer-ences in their spatial distributions (T. Yoshida, unpublisheddata), and tree species compositions before the treatmentswere similar among the compartments. A selection systemwas employed in which single-tree selections were con-ducted at 10-year intervals (one compartment per year). Thisstudy analyzed 20 years of data (i.e., two rounds of harvest-ing) from six compartments and the control (68 ha in all),within which 27 300 trees were censused individually andidentified with numbered tags. Compartments 1–4 were notincluded in this study, because their censuses were con-ducted without individual identifications. The area, the sumof the basal area, and the harvested level of the compart-

© 2006 NRC Canada

Yoshida et al. 1365

Fig. 1. Location and topography of the study area. Compartments 5–10 (62.1 ha) and the control (5.9 ha) were studied.

Basal area (m2/ha)

Compartment* Area (ha) Initial First period Second period Census year†

5 11.1 27.8 6.8 3.2 1970, 1980, 19906 5.6 34.0 4.4 5.7 1971, 1981, 19917 8.2 22.6 3.2 4.2 1972, 1982, 19928 6.7 27.9 2.9 3.8 1974, 1983, 19939 12.9 25.3 5.3 3.4 1976, 1984, 199410 17.6 31.4 6.8 4,2 1977, 1985, 1995Control 5.9 26.6 0.0 0.0 1979, 1990

*Compartments 1–4 were not analysed in this study, because their trees were censused without individualidentification.

†The census years were in the year before harvesting.

Table 1. Description of the study forest.

ments are shown in Table 1. The mean harvested level interms of basal area was 19% (range 10%–28%) for the firsttreatment and 16% (range 12%–18%) for the second treat-ment. These intensities were determined carefully in consid-eration of the census data, to maintain a harvested volumeequivalent to the growth (i.e., volume increment) of thecompartments. In these harvests, trees with a lower timbervalue (i.e., with injury, crack, etc.) were selected mainly toimprove the future economic quality of the stand. Neverthe-less, the harvested trees were selected from a broad range ofsize classes and showed no bias for any particular tree spe-cies; the total amounts of harvested conifers and harvestedbroad-leaved species were roughly equal.

Harvests were conducted in the winter. Trees were felledand their major branches cut immediately using a chainsaw.Then the logs were extracted using a tractor (a sled was usedonly in a few early years). Snow, normally 100–200 cmdeep, covered the forest floor. Consequently, soil disturbancesthat typically accompany treatments were minor. Althoughsupplemental plantings were conducted for some largenonwooded openings (e.g., canopy gaps larger than 400 m2),we did not consider them in this study because planted treesdid not reach the lower limit size set for the census (see laterparagraph). More detailed descriptions of the practices areavailable in Ohgane et al. (1988).

In the 68 ha area, we identified 27 300 trees with a diame-ter at breast height (DBH) ≥12.5 cm. The DBH classes (5 cmintervals) of all living trees were determined by measure-ment in the year before treatment. Checks of survival ordeath, repeated measures of the DBH class, and records ofnewly recruited trees were carried out in the second andthird censuses. The basal area of individuals was calculatedusing the median of the DBH class. We calculated the fol-lowing demographic parameters.

[1] Tree DBH class promotion rate

= p0.1

( )N N

N

+⎡

⎣⎢

⎤

⎦⎥ −1

[2] Tree mortality = 101

− −⎡⎣⎢

⎤⎦⎥

( ).

N NN

d

[3] Tree recruitment rate =( )N N

N+⎡

⎣⎢⎤⎦⎥

−r0.1

1

where N represents the initial tree number (except for har-vested trees); Np is the number of trees that are promoted tothe next DBH class; Nd is the number of naturally dead

© 2006 NRC Canada

1366 Can. J. For. Res. Vol. 36, 2006

Basal area

Species Abbreviation Group* m2/ha % Density (stems/ha) Max. DBH (cm)

Abies sachalinensis AS CF 14.68 49.1 212.9 80Acer mono AM BT 3.10 10.4 78.9 70Tiljajaponica TJ BT 3.06 10.2 50.4 75Quercus crispula QC BI 3.06 10.2 33.1 150Betula ermanii BE BI 2.17 7.2 27.1 95Α1nus hirsuta AH BΙ 0.64 2.1 12.3 75Magnolia obovata MO BI 0.56 1.9 16.2 45Ulmus davidiana UD BT 0.49 1.6 6.8 80Ulmus laciniata UL BT 0.44 1.5 6.0 65Kalopanax pictus KΡ BI 0.28 0.9 3.5 100Sorbus alnifolia SA BT 0.23 0.8 6.1 50Sorbus commixta SC BT 0.23 0.8 7.3 45Prunus sargentii PS BT 0.21 0.7 6.4 45Phellodendron amurense PA BI 0.18 0.6 4.8 50Prunus ssiori PR BT 0.12 0.4 3.2 50Betula maximowicxiana BM BI 0.10 0.3 1.0 75Picea jezoensis, PJ CF 0.08 0.3 1.0 55Taxus cuspidata TC CF 0.07 0.2 1.4 60Fraxinus mandshurica FM BT 0.05 0.2 0.9 65Salix hultenii SH BI 0.04 0.1 1.6 40Swida controversa SW BI 0.04 0.1 1.2 45Picea glehnii PG CF 0.03 0.1 0.3 55Betula platyphylla BP BI 0.02 0.1 0.5 50Acanthopanax sciadophylloides AC BT 0.01 0.04 0.4 40Juglans mandshurica JM BT 0.01 0.02 0.3 25Total 29.9 100.0 484 150Conifer 14.9 49.7 216 80Shade-tolerant broad-leaved 8.0 26.6 167 80Shade-intolerant broad-leaved 7.1 23.7 101 150

*CF, conifer; BI, shade-intolerant broad-leaved species; BT, shade-tolerant broad-leaved species. The classification was based on Koike (1988) andMasaki (2002).

Table 2. Preharvest composition of trees with DBH ≥12.5 cm in the study forest.

trees; and Nr is the number of trees recruited during the10-year period. Compartments with exceptional intervals(8 or 9 years) in the first period were calculated separately;then the weighted averages by area were used as a represen-tative value. We calculated these indices for four DBHclasses (12.5–22.4, 22.5–32.4, 32.5–42.4, and over 42.4 cm(in the case of the DBH promotion rate, 42.5–52.4 cm)) byspecies. In addition, species were grouped into conifer,shade-tolerant broad-leaved species, and shade-intolerantbroad-leaved species to summarize the results. The conifergroup consists of four shade-tolerant species, representedmostly by Abies sachalinensis. The broad-leaved specieswere classified with reference to previous studies of physio-logical traits (Koike 1988) or regeneration mode in an old-growth stand (i.e., species with abundant smaller trees areassumed to be shade tolerant: Masaki 2002) (Table 2). Coni-fers shared nearly half of the total basal area: shade-tolerant(dominated by Acer mono and Tilia japonica) and shade-intolerant broad-leaved species (dominated by Quercuscrispula and Betula ermanii) respectively shared a quarter.

Results

GrowthThe DBH class promotion rates of the treated area were

lower than those in the control (about 0.8 times) in the firstperiod, but became higher (about 1.5 times) in the secondperiod in all the DBH classes (Fig. 2). Improvement in thesecond period was particularly apparent for conifers, but notso distinct for broad-leaved species. A markedly high pro-motion rate was observed in the largest DBH class (42.5–52.4 cm) of shade-intolerant broad-leaved species in the

control, but it was not followed by that in the treated area,even in the second period. In the first period, nearly half thespecies had lower DBH promotion rates in the treated areathan in the control (Fig. 3). In the smallest DBH class (12.5–22.4 cm), shade-tolerant broad-leaved species tended toshow higher promotion rates, whereas shade-intolerantbroad-leaved species tended to show lower rates. Neverthe-less, in the larger DBH classes (≥32.5 cm), a shade-intolerantbroad-leaved species, Quercus crispula, which generallyshowed the highest promotion rate among species in thecontrol, showed a considerably lower rate. In the second pe-riod, although the apparent decrease in the treated area wasobserved only for Quercus crispula in the largest DBH class(42.5–52.4 cm), most species showed an improved promo-tion rate. A shade-tolerant conifer, Abies sachalinensis,showed the highest promotion rates among species in thetreated area in all the DBH classes in the second period.

Natural mortalityIn the second period, natural mortality for the larger DBH

classes (≥32.5 cm) in the treated area was considerably lowerthan in the control when considering all species (Fig. 4).This trend was largely reflected by conifers, and the respec-tive mortalities of broad-leaved (both shade-tolerant andshade-intolerant) species in the treated area were ratherhigher than of those in the control. Among shade-tolerantbroad-leaved species, the increase in mortality in the small-est DBH class (12.5–22.4 cm) was significantly large.

The species’ trends (Fig. 5) demonstrated that the range ofnatural mortality among species was generally smaller in thetreated area than in the control: species with higher mortalityin the control tended to show a lower mortality rate, and vice

© 2006 NRC Canada

Yoshida et al. 1367

All

0

0.01

0.02

0.03

0.04

0.05

First period

Second period

Control Conifer

0

0.01

0.02

0.03

0.04

0.05

Shade-tolerant

broad-leaved

0

0.01

0.02

0.03

0.04

0.05Shade-intolerant

broad-leaved

0

0.01

0.02

0.03

0.04

0.05

DBH class (cm)

22.5-32.412.5-22.4 32.5-42.4 42.5-52.4 22.5-32.412.5-22.4 32.5-42.4 42.5-52.4

DB

Hcla

ss

pro

motion

rate

s(y

ear-1

)

Fig. 2. The DBH class promotion rates in the treated area (first and second periods) and the control.

© 2006 NRC Canada

1368 Can. J. For. Res. Vol. 36, 2006

versa. Trees in the larger DBH classes (≥ 32.5 cm) exhibitedgenerally lower mortality in the second period. In contrast,for trees in the smaller DBH classes (<32.5 cm), the changebetween the periods was small; it showed no general trendamong species.

RecruitmentFor the three species groups, recruitment rates in the

treated area in the first period were similar to that in thecontrol and double that in the second period (Fig. 6). The re-cruitment rate of shade-intolerant broad-leaved species washighest: in the second period, recruitment was roughly twoand three times higher than that of shade-tolerant broad-leaved species and conifers, respectively. Although shade-tolerant broad-leaved species had the lowest recruitmentrates among the groups in the control, they showed the

0.0

0.1

0.0

0.1

(a) First period

12.5 � DBH < 22.5

(c) First period

22.5 DBH < 32.5

KPAH

AS

UD

QC

TJ

AM

PRPASA

MO

BE SB

SC

AS

TJ

QC

AH

UD

AMMO

0.0

0.1

0.0

0.1

(b) Second period

(d) Second period

AS

KP

AH

QC

UD

TJ

AM

PRPA

MO

SBSC

SA

BE

AS

QCUD

TJ

AM

AH

MO

conifer shade-tolerant broad leaved shade-intolerant broad-leaved

0.0

0.1

0.0

0.1

0.0 0.1

(e) First period

(g) First period

DBH class promotion rates (year ) in control–1

AS

QC

TJ

UD

AM

AS

TJ

QC

0.0

0.1

0.0

0.1

0.0 0.1

AM

(f) Second period

(h) Second period

AS

QC

TJ

UD

QC

TJ

AS

�

12.5 � DBH < 22.5

22.5 DBH < 32.5�

32..5 DBH < 42.5�

42..5 DBH < 52.5�

32.5 DBH < 42.5�

42.5 DBH < 52.5�

DB

Hcla

ss

pro

mo

tio

nra

tes

(ye

ar

)in

the

tre

ate

da

rea

–1

-

Fig. 3. DBH class promotion rates for major tree species in the treated area (first and second periods) plotted against those in the con-trol. Species abbreviations are defined in Table 2.

greatest improvement. In contrast, the improvement wassmallest for conifers.

Five species (four shade-intolerant broad-leaved and oneconifer) had lower recruitment rates in the treated area(Fig. 7). However, in the second period, almost all the sub-jected species were over the equivalence line (i.e., thetreated area showed higher rates than the control), with con-siderable improvement of shade-intolerant broad-leaved spe-cies (such as Betula ermanii and Salix sp.). In contrast to thecontrol, Abies sachalinensis showed the lowest rate amongall species in the treated area. Shade-tolerant broad-leavedspecies did not show a clear trend among species.

Change in basal areaTable 3 summarizes changes in basal area of species (and

species groups) during the 20-year study period. In total,87% of the harvested basal area was recovered in that pe-riod; conifers showed considerably higher recovery (112%).Three minor conifers (Picea jezoensis (Sieb. & Zucc.) Carr.),Picea glehnii (Fr. Schm.) Masters, and Taxus cuspidate Sieb.et Zucc.) showed a similar trend to that of Abies sachalinensis.These high recoveries were mainly attributable to highgrowth, and less so to recruitment.

In contrast, the recovery of shade-tolerant broad-leavedspecies remained at only 48%, reflecting both low growthand high mortality. In particular, species with high deadbasal area (Acer mono, Sorbus alnifolia (Sieb. & Zucc.) C.Koch, and Prunus sargentii Rehd.) exhibited the lowest re-covery. Shade-intolerant broad-leaved species showed inter-mediate recovery (85%), with high a contribution ofrecruitment. The species with lowest recoveries (Magnoliaobovata Thunb., Alnus hirsuta Turcz., and Betula maximo-

wicziana Regel) in this group also showed high dead basalarea.

Discussion

Demographic parameters of tree species and speciesgroups differ widely between the treated area and the con-trol; the response to single-tree selection changed consider-ably between the first and second harvests. Such a temporalchange in the response may be attributable to a delay in theresponse of trees (e.g., Oliver and Larson 1996; Youngbloodand Ferguson 2003) and a cumulative effect of repetitiveharvests. General expectations of selection harvesting are in-creased growth, recruitment, and survival of trees, all posi-tive effects, to compensate the loss of harvested trees.Actually, opposite and negative effects were dominant in thefirst period.

Many previous studies have suggested that partial harvest-ing systems can maintain structural complexities of the stand(Deal 2001; Deal and Tappeiner 2002). We can expect thatthe multiaged structure of this stand was generally main-tained, because harvested trees were selected from a broadrange of size classes. Nevertheless, our results showedclearly that selection harvesting changes tree species compo-sition, even if basal area levels are nearly maintained. Al-though longer term trends do not necessarily follow theobserved characteristics of demography, we could appliedthose results for predicting stand dynamics. In the short term(i.e., several decades), Abies sachalinensis will become moredominant relative to broad-leaved species because its growthand survival were the most strongly positively affected byselection harvesting. Such a positive response to small-scalecanopy disturbances has already been reported in natural for-

© 2006 NRC Canada

Yoshida et al. 1369

All

0.000

0.005

0.010First period

Second period

Control

Conifer

0.000

0.005

0.010

Shade-tolerant

broad-leaved

0.000

0.005

0.010

Shade-intolerant

broad-leaved

0.000

0.005

0.010

DBH class (cm)

22.5-32.412.5-22.4 32.5-42.4 42.5- 22.5-32.412.5-22.4 32.5-42.4 42.5-

Tre

em

ort

alit

y(y

ear-1

)

Fig. 4. Tree mortality in the treated area (first and second periods) and the control.

ests (Hiura et al. 1996; cf. Veblen 1986), probably becauseof the physiological ability of Abies spp. to rapidly respondto increased light availability (Noguchi et al. 2003; cf.Kneeshaw et al. 1998; Youngblood and Ferguson 2003).Nevertheless, we found that recruitment of this species wasnot sufficiently improved after the treatment: it was the low-est among major tree species. This trend should engender a

lack of small-sized trees and decreased dominance of thisspecies over the long term (i.e., more than 100 years).

Past studies have shown two different regeneration modesof Abies sachalinensis: continuous age distribution, suggest-ing dependence of its regeneration on small-scale distur-bances (Ishizuka 1984; Takahashi et al. 2003), anddiscontinuous distribution, suggesting importance of large-

© 2006 NRC Canada

1370 Can. J. For. Res. Vol. 36, 2006

0.00

0.02(b) Second period

PS

BE

SH

SC

AH

AM

TJKP

MO

AS

UD

QC

PA

SA

0.00

0.02

(a) First period

12.5 � DBH < 22.5

ASMO

QCKP

SA

PA

TJ

AM

AH

UD

SC

SH

BE

PS

0.00

0.02

0.00 0.02

(c) First period

QC

TJ

AS

AM MO

UD

AH

0.00

0.02

0.00 0.02

(d) Second period

QC

TJ

AS

AM

MO

UD

AH

0.00

0.01(f) Second period

AM

QC

UD

AS

TJ

0.00

0.01

0.00 0.01

(h) Second period

AS

TJQC

0.00

0.01(e) First period

AS

TJ

AM

UD

QC

Mo

rta

lity

(ye

ar-1

)in

the

tre

ate

da

rea

conifer shade-tolerant broad-leaved shade-intolerant broad-leaved

Mortality (year-1

) in the control

0.00

0.01

0.00 0.01

(g) First period

QC

TJ

AS

22.5 � DBH < 32.5

32.5 � DBH < 42.5

42.5 � DBH 42.5 � DBH

32.5 � DBH < 42.5

22.5 � DBH < 32.5

12.5 � DBH < 22.5

Fig. 5. Tree mortality for major tree species in the treated area (first and second periods) plotted against those in the control. Speciesabbreviations are defined in Table 2.

scale disturbances (Suzuki et al. 1987; Ishikawa and Ito1989). The present results are, at least, not consistent withthe former. We presume that the limited recruitment can beexplained as follows: (1) contrary to growth improvement(see previous paragraph), photosynthesis of Abies sachalin-ensis was found to decrease in response to severe increasedlight availability, especially in the dry summer (Noguchi etal. 2003; cf. Kneeshaw et al. 2002; Bourgeois et al. 2004);(2) because seedlings and saplings of this species are fre-quently distributed around the stems of overstory trees(Suzuki et al. 1987; Takahashi 1997), they may be affectedgreatly by physical damage accompanying the treatments(Harvey and Bergeron 1989). In general, conifers are pre-dicted to be more vulnerable to such physical damage thanare broad-leaved species because of the general inability ofconifers to sprout (Greene et al. 1999). Furthermore, becauseconifers base their establishment largely on fallen logs as apreferable substrate (Hiura et al. 1996), the possible de-crease in fallen logs resulting from repetitive selection har-vestings (Noguchi and Yoshida 2004) may accelerate thatdecreasing trend over a longer term. This tendency shouldbe more apparent in minor conifer species in this forest (but

common in the region), Picea jezoensis and Picea glehnii,which show a particular dependency upon fallen logs(Kubota et al. 1994; Takahashi 1997).

Despite increased light availability after the selection har-vesting, growth improvement of shade-intolerant broad-leaved species was less than that of Abies sachalinensis.Even in the second period, large-sized trees in the treatedarea had considerably lower growth than those in the con-trol. These facts seem to greatly reflect the trend observedfor Quercus crispula, but its small-sized trees showed an im-proved response (also see Yoshida and Kamitani 1998). Agreat difference might exist in response to canopy openingbetween size classes. Quercus crispula had discontinuousage distribution in natural mixed stands (Sano 1988;Takahashi et al. 2003), suggesting the lower contribution ofsmall-scale disturbances. Our results are consistent with thishypothesis. Regarding other shade-intolerant broad-leavedspecies, we found improved growth in small-sized Betulaermanii and Alnus firma and decreased mortality in mid-sized Magnolia obovata and Kalopanax pictus (Thunb.)Nakai. These are expected results in consideration of re-sponses to high light availability (Ishikawa and Ito 1989).

© 2006 NRC Canada

Yoshida et al. 1371

0

0.01

0.02

0.03

Tre

ere

cru

itm

en

tra

te(y

ea

r)

–1 First period

Second period

Control

All Conifer Shade-tolerant

broad-leaved

Shade-intolerant

broad-leaved

Fig. 6. Tree recruitment rate in the treated area (first and second periods) and the control.

conifer shade-tolerant broad-leaved shade-intolerant broad-leaved

0.00

0.06

0.00 0.06

0.00

0.06

0.00 0.06

(a) First period (b) Second period

Recruitment rate (year–1

) in the control

AS

TJ

PSSA

MO

AM

PR

SC

SH

UD

QC

BE AH

KP

PAKP

PA

BE

AH

QC

PS

SC

SH

UD

MO

PR

AS

TJ

AM

SA

Recru

itm

ent

rate

(year

)–1

inth

etr

eate

dare

a

Fig. 7. Tree recruitment rate for major tree species in the treated area (first and second periods) plotted against those in the control.Species abbreviations are defined in Table 2.

These species also showed higher recruitment rates after thetreatment, which may contribute considerably to their recov-ery. They may gain a relative advantage because their poten-tial high growth rate occasionally enables them to escapefaster from the understory to a lighter environment. We canexpect these species to gradually increase in density afterseveral decades.

In contrast, shade-tolerant broad-leaved species appearedto incur the most distinct negative effects from the treatment.Recovery of the basal area (percentage of the increment tothat lost) in the 20-year period was only 48%; in particular,recovery rates of Acer mono, Sorbus alnifolia, and Prunussargentii were less than 40%. These low recoveries weremainly attributable to high mortality, presumably resultingfrom increased risks of wind or frost damage through can-opy exposure (Peltola et al. 1999). In addition, we found nodistinct positive effects of the treatment on growth of thesespecies. Despite some exceptions (e.g., small-sized Ulmusdavidiana Planch. var. japonica (Rehd.) Nakai and Sorbusalnifolia), growth rates of dominant species (Acer mono andTilia japonica) were unchanged or decreased in the treatedarea. This trend is consistent with a previous finding that

these species showed low plasticity of growth against highlight availability (Umeki 2001). In contrast, recruitment ofthis species group increased in the second period, showingthe greatest improvement, compared with the control, amongthe three species groups. However, the absolute amount ofrecruitment was still half that of shade-intolerant species.

Our study demonstrated complicated responses of themixed-species community under the selection system. Theobserved trends of change in species composition differmarkedly from those reported in North America (Leak andFillip 1977; Jenkins and Parker 1999; Webster and Lorimer2002; Shuler 2004). We suspect that this difference is attrib-utable to large canopy gap areas in the overstory and domi-nance of dwarf bamboo species in this forest’s understory.Selection harvesting, and its consequent creation of canopygaps, increases dwarf bamboos (Noguchi and Yoshida 2005),which thereby suppresses tree regeneration (Nagaike 1999;Umeki and Kikuzawa 1999). Those two distinct characteris-tics are closely correlated. We presume that the results areinformative, even for management of stands with rich under-story vegetation (e.g., Harvey and Bergeron 1989; Messier et al.1998; Beckage et al. 2000; Bourgeois et al. 2004), where

© 2006 NRC Canada

1372 Can. J. For. Res. Vol. 36, 2006

Basal area (m2/ha)

Species Abbreviation Group*Harvested(H)

Grown(G)

Recruited(R)

Naturallydead (D)

Recovered(G+R+D)/H

Abies sachalinensis AS CF 4.454 5.189 0.286 0.526 1.11Acer mono AM BT 1.299 0.521 0.136 0.277 0.29Tilia japonica TJ BT 0.842 0.546 0.114 0.127 0.63Quercus crispula QC BI 0.689 0.738 0.101 0.118 1.05Betula ermanii BE BI 0.803 0.365 0.137 0.052 0.56Alnus hirsuta AH BI 0.245 0.157 0.056 0.071 0.58Magnolia obovata MO BI 0.248 0.083 0.046 0.034 0.38Ulmus davidiana UD BT 0.204 0.143 0.036 0.006 0.85Ulmus laciniata UL BT 0.017 0.103 0.009 0.085 1.59Kalopanax pictus KP BI 0.049 0.102 0.069 0.006 3.34Sorbus alnifolia SA BT 0.094 0.037 0.012 0.012 0.39Sorbus commixta SC BT 0.117 0.044 0.032 0.009 0.57Prunus sargentii PS BT 0.097 0.026 0.013 0.016 0.24Phellodendron amurense PA BI 0.043 0.062 0.093 0.010 3.32Prunus ssiori PR BT 0.024 0.026 0.009 0.012 1.00Betula maximowicziana BM BI 0.035 0.014 0.013 0.012 0.44Picea jezoensis PJ CF 0.004 0.039 0.003 0.003 9.79Taxus cuspidata TC CF 0.013 0.010 0.003 0.002 0.87Fraxinus mandshurica FM BT 0.004 0.024 0.016 0.005 9.37Salix hultenii SH BI 0.010 0.021 0.022 0.005 3.70Swida controversa SW BI 0.014 0.010 0.024 0.001 2.36Picea glehnii PG CF 0.001 0.012 0.001 0.000 11.92Betula platyphylla BP BI 0.002 0.017 0.010 0.004 10.43Acanthopanax sciadophylloides AC BT 0.005 0.003 0.002 0.002 0.40Juglans mandshurica JM BT 0.002 0.005 0.002 0.001 3.69Total 9.315 8.296 1.245 1.396 0.87Conifer 4.471 5.250 0.292 0.532 1.12Shade-tolerant broad-leaved 2.705 1.477 0.381 0.552 0.48Shade-intolerant broad-leaved 2.139 1.569 0.571 0.313 0.85

*CF, conifer; BI, shade-intolerant broad-leaved species; BT, shade-tolerant broad-leaved species. The classification was based on Koike (1988) andMasaki (2002).

Table 3. Changes in basal area during the 20-year study period.

such an indirect negative effect of harvesting on light avail-ability seems to occur.

The present results should be generalized carefully be-cause we have investigated only one 68 ha stand withoutrepetition of the control area. Nevertheless, trends describedin this large-scale, long-term study present implications forimproving current harvesting practices. Attention should bedevoted to the possible change in species composition to-ward shade-intolerant broad-leaved species, which might al-ter the ecosystem integrity of these mixed forests.Management plans must mitigate the negative effects of se-lection harvesting on growth, survival, and recruitment ofparticular species. For conifer species, supplemental plantingor site preparation for natural regeneration (Wurtz andZasada 2001; Yoshida et al. 2005) should be included to en-hance their recruitment. At the same time, while consideringthe distinct negative effects, harvesting level of shade-tolerant broad-leaved species should be limited to preservetheir relative dominance. We need more detailed mecha-nisms by which observed patterns and processes are pro-duced. In particular, we note that the amount of recruitmentin this study might depend largely on preharvesting accumu-lation of a seedling bank because most of the recruited treesseemed to be older than 20 years (H. Miya, unpublisheddata). Longer term studies or those examining dynamics ofseedlings and saplings (M. Noguchi and T. Yoshida, in prep-aration) are needed. In addition, we should consider sitevariations in the stand. Further studies that include responsesof individual trees with spatial consideration (M. Noguchiand T. Yoshida, in preparation) are expected to yield addi-tional important information. Together with the results of thecurrent study, these extension studies can form an appropri-ate basis for adaptive management in this region.

Acknowledgements

We would like to thank anonymous reviewers for theirhelpful comments and suggestions. We gratefully acknowl-edge S. Taniguchi, E. Ohgane, T. Nigi, Y. Hishinuma, K.Fujiwara, K. Koshika, and K. Kanuma for their combined ef-forts in maintaining this long-term study. We thank the staffof the Nakagawa Experimental Forest for their proper man-agement of fieldwork, harvesting practices, and the data col-lection. Thanks are also extended to the members of thelaboratory of forest management, Hokkaido University, fortheir assistance in the fieldwork.

References

Abe, S., Masaki, T., and Nakashizuka, T. 1995. Factors influencingsapling composition in canopy gaps of a temperate deciduousforest. Vegetatio, 120: 21–32.

Beckage, B., Clark, J.S., Clinton, B.D., and Haines, B.L. 2000. Along-term study of tree seedling recruitment in southern Appala-chian forests: the effects of canopy gaps and shrub understories.Can. J. For. Res. 30: 1617–1631.

Bourgeois, L., Messier C., and Brais, S. 2004. Mountain maple andbalsam fir early response to partial and clear-cut harvesting un-der aspen stands of northern Quebec. Can. J. For. Res. 34:2049–2059.

Canham, C.D., and Marks, P.L. 1985. The response of woodyplants to disturbance: patterns of establishment and growth. InThe ecology of natural disturbance and patch dynamics. Editedby S.T.A. Pickett and P.S. White. Academic Press, New York.pp. 197–216.

Deal, R.L. 2001. The effects of partial cutting on forest plant com-munities of western hemlock – Sitka spruce stands in southeastAlaska. Can. J. For. Res. 31: 2067–2079.

Deal, R.L., and Tappeiner, J.C. 2002. The effects of partial cuttingon stand structure and growth of western hemlock – Sitka sprucestands in southeast Alaska. For. Ecol. Manage. 159: 173–186.

Franklin, J.F., Spies, T.A., Pelt, R.V., Carey, A.B., Thornburgh,D.A., Berg, D.R., Lindenmayer D.B., Harmon, M.E., Keeton,W.S., Shaw, D.C., Bible, K., and Chen, J. 2002. Disturbancesand structural development of natural forest ecosystems withsilvicultural implications, using Douglas-fir forests as an exam-ple. For. Ecol. Manage. 155: 399–423.

Greene, D.F., Zasada, J.C., Sirois, L., Kneeshaw, D., Morin, H.,Charron, I., and Simard, M.J. 1999. A review of the regenera-tion dynamics of North American boreal forest tree species.Can. J. For. Res. 29: 824–839.

Harvey, B.D., and Bergeron, Y. 1989. Site patterns of natural re-generation following clear-cutting in northwestern Quebec. Can.J. For. Res. 19: 1458–1469.

Harvey, B.D., Leduc, A., Gauthier, S., and Bergeron, Y. 2002.Stand–landscape integration in natural disturbance-based man-agement of the southern boreal forest. For. Ecol. Manage. 155:369–385.

Hibbs, D.E. 1982. Gap dynamics in a hemlock–hardwood forest.Can. J. For. Res. 12: 522–527.

Hiura, T., and Fujiwara, K. 1999. Density-dependence and co-existence of conifer and broad-leaved trees in a Japanese north-ern mixed forest. J. Veg. Sci. 10: 843–850.

Hiura, T., Sano, J., and Konno, Y. 1996. Age structure and re-sponse to fine-scale disturbance of Abies sachalinensis, Piceaglehnii, and Betula ermanii growing under the influence of adwarf bamboo understory in northern Japan. Can. J. For. Res.26: 289–297.

Hunter Jr., M.L. 1999. Maintaining biodiversity in forest ecosys-tems. Cambridge University Press, Cambridge.

Ishikawa, Y., and Ito, K. 1989. The regeneration process in a mixedforest in central Hokkaido, Japan. Vegetatio, 79: 75–84.

Ishizuka, M. 1984. Spatial pattern of trees and their crowns in nat-ural mixed forests. Jpn. J. Ecol. 34: 421–430.

Ishizuka, M., Toyooka, H., Osawa, A., Kushima, H., Kanazawa, Y.,and Sato, A. 1998. Secondary succession following catastrophicwindthrow in a boreal forest in Hokkaido, Japan: the timing oftree establishment. J. Sustain. For. 6: 367–388.

Jenkins, M.A., and Parker, G.R. 1998. Composition and diversityof woody vegetation in silvicultural openings of southern Indi-ana forests. For. Ecol. Manage. 109: 57–74.

Kneeshaw, D.D., Bergeron, Y., and De Grandpre, L. 1998. Earlyresponse of Abies balsamea seedlings to artificially createdopenings. J. Veg. Sci. 9: 543–550.

Kneeshaw, D.D., Williams, H., Nikinmaa, E., and Messier, C.2002. Pattern of above- and below-ground response ofunderstory conifer release 6 years after partial cutting. Can. J.For. Res. 32: 255–265.

Knuchel, H. 1953. Planning and control in the managed forest. Oli-ver & Boyd, Edinburgh.

Kohm, K.A., and Franklin, J.F. 1997. Creating a forestry for the21st century: the science of ecosystem management. IslandPress, Washington D.C.

© 2006 NRC Canada

Yoshida et al. 1373

Koike, T. 1988. Leaf structure and photosynthetic performance asrelated to the forest succession of deciduous broad-leaved trees.Plant Species Biol. 3: 77–87.

Kubota, Y. 1995. Effects of disturbance and size structure on theregeneration process in a sub-boreal coniferous forest, northernJapan. Ecol. Res. 10: 135–142.

Kubota, Y. 2000. Spatial dynamics of regeneration in a conifer/broad-leaved forest in northern Japan. J. Veg. Sci. 11: 633–640.

Kubota, Y., Konno, Y., and Hiura, T. 1994. Stand structure andgrowth patterns of understory trees in a coniferous forest,Taisetsuzan National Park, northern Japan. Ecol. Res. 9: 333–341.

Leak, W.B., and Fillip, S.M. 1977. Thirty-eight years of group se-lection in New England northern hardwoods. J. For. 75: 641–643.

Masaki, T. 2002. Structure and dynamics. In Diversity and interac-tion in a temperate forest community: Ogawa forest reserve ofJapan. Edited by T. Nakashizuka and Y. Matsumoto. Springer-Verlag, Tokyo. pp. 53–65.

Messier, C., Sylvain, P., and Bergeron, Y. 1998. Effects ofoverstory and understory vegetation on the understory light en-vironment in mixed boreal forest. J. Veg. Sci. 9: 511–520.

Messier, C., Doucet, R., Ruel, J., Claveau, Y., Kelly, C., andLechowicz, M.J. Y. 1999. Functional ecology of advance regen-eration in relation to light in boreal forests. Can. J. For. Res. 29:812–823.

Mitchell, R.J., Palik, B.J., and Hunter, M.L., Jr. 2002. Natural dis-turbance as a guide to silviculture. For. Ecol. Manage. 155: 315–317.

Nakashizuka, T., and Iida, S. 1995. Composition, dynamics anddisturbance regime of temperate deciduous forests in MonsoonAsia. Vegetatio, 121: 23–30.

Nagaike, T., Kubota, Y., and Watanabe, N. 1999. The effects of se-lective logging on stand structure and the regeneration of sub-boreal forests in Hokkaido, northern Japan. J. For. Res. 4: 41–45.

Nigi, T., Koshika, K., Kanuma, K., Natsume, S., Yone, Y., Morita,H., Fujito, E., and Hojo, H. 1998. Managerial progress and re-sults of an experimental forest practiced under the ControlMethod. Res. Bull. Coll. Exp. For. Hokkaido Univ. 55: 274–308.[In Japanese with English summary.]

Noguchi, M., and Yoshida, T. 2004. Tree regeneration in partiallycut conifer–hardwood mixed forests in northern Japan: roles ofestablishment substrate and dwarf bamboo. For. Ecol. Manage.190: 335–344.

Noguchi, M., and Yoshida, T. 2005. Factors influencing the distri-bution of two co-occurring dwarf bamboo species (Sasakurilensis and S. senanensis) in a conifer–broadleaved mixedstand in northern Hokkaido. Ecol. Res. 20: 15–20.

Noguchi, M., Kayama, M., Yoshida, T., and Koike, T. 2003.Photosynthetic traits of seedling of Sakhalin fir grown on the se-lective cutting forest. Transactions of the meeting in Hokkaidobranch of the Japanese forestry society. 51: 36–38. [In Japa-nese.]

O’Hara, K.L. 1998. Silviculture for structural diversity: a new lookat multiaged systems. J. For. 96: 4–10.

Ohgane, E., Nigi, T., Hishinuma, Y., Koshika, K., and Fukui, T.1988. Managerial progress and results of an experimental forestpracticed by the Control Method. Res. Bull. Coll. Exp. For.Hokkaido Univ. 45: 61–113. [In Japanese with English sum-mary.]

Oliver, C.D., and Larson B.C. 1996. Forest stand dynamics. JohnWiley & Sons, New York.

Osawa, A. 1992. A development of a mixed conifer forest inHokkaido, northern Japan, following a catastrophic windstorm:

a “parallel” model of plant succession. In The ecology andsilviculture of mixed-species forests. Edited by M.J. Kelty, B.C.Larson, and C.D. Oliver. Kluwer Academic Publishers,Dordrecht, Netherlands. pp. 29–52.

Peltola, H., Kellomaki, S., Valsanen, H., and Ikonen, V.P. 1999. Amechanistic model for assessing the risk of wind and snow dam-age to single trees and stands of Scots pine, Norway spruce, andbirch. Can. J. For. Res. 29: 647–661.

Runkle, J.R. 1982. Patterns of disturbance in some old-growth me-sic forests of eastern North America. Ecology, 63: 1533–1546.

Runkle, J.R. 1985. Disturbance regimes in temperate forests. InThe ecology of natural disturbance and patch dynamics. Editedby S.T.A. Pickett and P.S. White. Academic Press, New York.pp. 17–33.

Runkle, J.R., and Yetter, T.C. 1987. Treefalls revisited: gap dynam-ics in the southern Appalachians. Ecology, 68: 417–424.

Sano, J. 1988. Studies on regeneration patterns in natural oak for-ests through community structure analysis. Res. Bull. Coll. Exp.For. Hokkaido Univ. 45: 221–266. [In Japanese with Englishsummary.]

Shuler, T. 2004. Fifty years of partial harvesting in a mixedmesophytic forest: composition and productivity. Can. J. For.Res. 34: 985–997.

Smith, D.M., Larson, B.C., Kelty, M.J., and Ashton, P.M.S. 1997.The practice of silviculture: applied forest ecology. 9th ed. JohnWiley & Sons, New York.

Spence, J.R. 2001. The new boreal forestry: adjusting timber man-agement to accommodate biodiversity. Trends Ecol. Evol. 16:591–593.

Suzuki, E., Ota, K., and Igarashi, T. 1987. Regeneration process ofconiferous forests in northern Hokkaido 1. Abies sachalinensisforest and Picea glehnii forest. Ecol. Res. 2: 61–75.

Takahashi, K. 1997. Regeneration and coexistence of two subal-pine conifer species in relation to dwarf bamboo in theunderstorey. J. Veg. Sci. 8: 529–536.

Takahashi, M., Sakai, Y., Ootomo, R., and Shiozaki, M. 2000. Es-tablishment of tree seedlings and water-soluble nutrients incoarse woody debris in an old-growth Picea–Abies forest inHokkaido, northern Japan. Can. J. For. Res. 30: 1148–1155

Takahashi, K., Mitsuishi, D., Uemura, S., Suzuki, J., and Hara, T.2003. Stand structure and dynamics during a 16-year period in asub-boreal conifer–hardwood mixed forest, northern Japan. For.Ecol. Manage. 174: 39–50.

Tanaka, H., and Nakashizuka, T. 1998. Fifteen years of canopy dy-namics analyzed by aerial photographs in a temperate deciduousforest, Japan. Ecology, 78: 612–620.

Tatewaki, M. 1958. Forest ecology of the islands of the north Pa-cific Ocean. J. Fac. Agric. Hokkaido Univ. 50: 371–486.

Tatewaki, M., and Igarashi, T. 1971. Forest vegetation in the Teshioand the Nakagawa District experimental forests of HokkaidoUniversity, Prov. Teshio, N. Hokkaido, Japan. Res. Bull. Coll.Exp. For. Hokkaido Univ. 28: 1–192. [In Japanese with Englishsummary.]

Umeki, K. 2001. Growth characteristics of six tree species onHokkaido Island, northern Japan. Ecol. Res. 16: 435–450.

Umeki, K., and Kikuzawa, K. 1999. Long-term growth dynamicsof natural forests in Hokkaido, northern Japan. J. Veg. Sci. 10:815–824.

Veblen, T.T. 1986. Treefalls and coexistence of conifers in sub-alpine forests of the central Rockies. Ecology, 67: 644–649.

von Gadow, K., Nagel, J., and Saborowski, J. 2002. Continuouscover forestry. Kluwer Academic Publishers, Dordrecht, Nether-lands.

© 2006 NRC Canada

1374 Can. J. For. Res. Vol. 36, 2006

Webster, C.R., and Lorimer, C.G. 2002. Single-tree versus groupselection in hemlock–hardwood forests: Are smaller openingsless productive? Can. J. For. Res. 32: 591–604.

Wurtz, T.L., and Zasada, J.C. 2001. An alternative to clear-cuttingin the boreal forest of Alaska: 27-year study of regeneration af-ter shelterwood harvesting. Can. J. For. Res. 31: 999–1011.

Yoshida, T., and Kamitani, T. 1998. Effects of crown release onbasal area growth rates of some broad-leaved tree species withdifferent shade tolerance. J. For. Res. 3: 181–184.

Yoshida, T., Iga, Y., Ozawa, M., Noguchi, M., and Shibata, H.2005. Factors influencing early vegetation establishment follow-ing soil scarification in a mixed forest in northern Japan. Can. J.For. Res. 35: 175–188.

Youngblood, A., and Ferguson, D.E. 2003. Changes in needle mor-phology of shade-tolerant seedlings after partial overstory can-opy removal. Can. J. For. Res. 33: 1315–1322.

© 2006 NRC Canada

Yoshida et al. 1375