Ecological responses of natural and planted forests to thinning in southeastern Korea: a chronosequence study
Ecological responses of natural and planted forests to thinning in southeastern Korea: a chronosequence study
Journal of Ecology and Environment. 2011. Dec, 34(4): 347-355
Copyright ©2011, The Ecological Society of Korea
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
  • Received : January 19, 2011
  • Accepted : August 30, 2011
  • Published : December 01, 2011
Export by style
Cited by
About the Authors
Yong-Chan, Cho
Department of Forest Resource Conservation, Korea National Arboretum, Pocheon 487-821, Korea
Jung Hun, Pee
Department of Biology, Graduate School of Seoul Womens University, Seoul 139-774, Korea
Gyeong Soon, Kim
Department of Biology, Graduate School of Seoul Womens University, Seoul 139-774, Korea
Bon-Yoel, Koo
Department of Biology, Graduate School of Seoul Womens University, Seoul 139-774, Korea
Hyun Je, Cho
Korea Forest Ecosystems Institute, Daegu 702-894, Korea
Chang Seok, Lee
Department of Life and Environmental Engineering, Seoul Women's University, Seoul 139-774, Korea

Effects of forest thinning on community level properties have not been understood yet in Korea. We investigated regen-eration patterns and trajectories after a disturbance by applying a chronosequence approach. Light availability, litter and woody debris cover, and species composition were determined for twenty 50 m line-transect samples representing a dis-turbance duration gradient (within 11 years). Environmental factors such as light availability and coverage of woody de-bris and litter changed abruptly after thinning and then returned to the pre-disturbance state. Although species richness was gained at shrub and ground layer in a limited way in both forests, cover of various functional types revealed diversity in their responses. Notably, Alnus firma stands exhibited a larger increment of cover in woody plants. Ordination analysis revealed different regeneration trajectories between natural and planted stands. Based on ordination analysis, rehabili-tated stands showed movement to alternative states compared with natural ones, reflecting lower resilience to perturba-tion (i.e., lower stability). Our results suggest that community resilience to artificial thinning depends on properties of the dominant species. But to get more explanatory ecological information, longer-term static observations are required.
In Korea, forest tending works (thinning) for natural and planted forests have been conducted on a broad scale after International Monetary Fund (IMF) financial sup-port (Woo 2003). In this operation, canopy compositors (as the dominant species) are partially removed (by visu-ally selecting healthy or degrading individuals), and sub-canopy and shrub layer species are eliminated to improve growth of target species and environmental factors such as water retention capacity and soil physical properties. Canopy openness remained at 50-80% compared with the pre-thinned state (Korea Forest Service instruction No. 890). Thus, structural and environmental changes due to thinning are smaller than to clear cutting.
Management activities (clear cutting, thinning and so on) in forest ecosystems are broadly applied in most nations. Tremendous disturbances such as clear-cutting usually alter the structure and function of forest commu-nities. As a result, various successional stages commonly appear. Thinning also changes understory vegetation indirectly by modifying overstory openness and density and their effects on understory microclimate, light, and soil water (Dodson et al. 2008, Jung et al. 2008). In addi-tion, thinning may result in higher availability of water and mineral nutrients through the formation of root gaps (Parsons et al. 1994) and increased competitive perfor-mance of residual or invaded species. In managed forest ecosystems, residual species and rapidly establishing in-vaders are primary contributors to secondary succession after disturbances (logging) (Halpern 1988, 1989).
Forest operations result in various responses of plant communities, such as increasing (Metlen and Fiedler 2006, Dodson et al. 2008), decreasing (Metlen et al. 2004, Fulé et al. 2005) species richness or neutral response (Nel-son et al. 2008). Responses of species cover also varied among functional types, suggesting differences in toler-ance or resilience to disturbances (Halpern 1988, Metlen et al. 2004, Metlen and Fiedler 2006, Moore et al. 2006). In understory vegetation, various consequences of second-ary succession after forest management are predictable, but the emerging patterns and pathways of disturbed for-ests are not yet clarified in natural and planted forests in Korea.
Communities may be considered stable if these path-ways, after management activities, result in a return to the initial composition within an appropriate time scale (Sutherland 1981). In this perspective, understory com-munities play a central role in the dynamics and func-tioning of forest ecosystems by influencing long-term patterns of vegetation development (D’Amato et al. 2009). Understory vegetation is also a critical component of for-est ecosystems, typically supporting the vast majority of components of total ecosystem floristic diversity (Halp-ern and Spies 1995, Thomas et al. 1999, D’Amato et al. 2009).
As was mentioned above, the effects of forest manage-ment practices in Korea such as clear cutting and thin-ning on species diversity and composition have not been understood sufficiently (but see Cho et al. 2011). This insufficiency and lack of basic information reflects the paucity of community-level studies in natural and man-aged stands. Combined with emerging principles in the landscape and in restoration ecology of natural or man-aged ecosystems, changes in the society’s conceptions of a broader range or monotonous forest management op-tions are causing a reassessment of forestry practices. For-est and natural resource managers are making an effort to balance economical and ecological values of the forest by applying alternative models of forest management. In this viewpoint, the introduction and founding of manage-ment and operation options and policies surely require a fundamental understanding of vegetation processes and conceptual models that can predict responses to pertur-bations in target systems.
The chronosequence approach to study vegetation dy-namics has been validated (Foster and Tilman 2000), and has provided significant insights into the patterns and mechanisms of plant succession (e.g., Inouye et al. 1987, Pickett 1989, Lee et al. 2002, 2008, Lee 2006a, 2006b, Hau-go and Halpern 2007, Cho et al. 2011). The present study aims to clarify how management activity influences com-munity resistance and resilience. In order to arrive at this goal, we analyzed successional patterns and trajectories (pathways) of understory vegetation and abiotic variables (light availability, woody debris and litter fall) by applying a chronosequence approach after forest thinning in natu-ral and planted broadleaved forests.
- Study area
The study sites were located in rural areas in Pohang-si, Gyeongsangbuk-do, southeastern Korea. The altitudinal distributions of sites were between 150-305 m above sea level ( Table 1 ). The study area is classified as a warm-tem-perate deciduous broad-leaved forest (such as Quercus serrate forest) zone (Lee et al. 1996, Kim and Lee 2006). In this area, the lowland is occupied by oak and artificial ( A. firma, Robinia pseudoacacia, Pinus thunbergii, and P. rigi-da ) forests. Most plantations in this area were introduced through a large scale forest restoration project called in the Yeongil erosion control project, in the early 1970s. For natural and restored stands, forest-tending works (thin-ning) to improve resource quality in lowland forests have been practiced from 1998 to date.
Soils were shallow, sandy loams and loams derived from mudstone and shale on the tertiary deposit in the Cenozoic Era based on digital detailed soil maps pub-lished by the Rural Development Administration. Annual mean temperature and precipitation (1971-2007) ob-tained from the Pohang-si weather station, the closest to this study area, were 14.0°C and 1,166.8 mm, respectively.
We explored the study area in a vegetation survey and selected 10 stands in the natural broadleaved forest (dom-inated by Q. serrata ). After the field survey, we verified that the stands selected were young vegetation in the early 1970s through orthogonal image interpretation (taken in the early 1970s). Among them, 5 stands remained in their natural state without any treatment; the others were man-aged (by thinning). In five managed stands, we confirmed the years lapsed after the operation (one 11 year stand; three 4 year stands; one 2 year stand) by counting the age of stems sprouted from the cut stump ( Table 1 ). Natural Q. serrata stands were mainly composed of Carpinus laxi-flora in the canopy layer, Acer pseudosieboldianum and Lindera erythrocarpa in the subtree layer, Rhododendron schlippenbachii, L. obtusiloba, and Viburnum erosum in the shrub layer, and Disporum smilacinum, Carex humi-lis, Carex gifuensis, and Ainsliaea acerifolia in the herba-ceous layer.
For rehabilitated stands, a total of 10 stands were se-lected in A. firma stands, which were introduced through the Yeongil erosion control project. Management history was checked by the same processes as those used for the natural forest. Among 10 A. firma stands, 5 remained in the natural state after planting. Managed stands were composed of two 3 year stands, two 8 year stands, and one 4 year stand. Unmanaged A. firma stands were composed of Q. serrata and C. laxiflora in the canopy layer, Sorbus alnifolia A. pseudosieboldianum and L. erythrocarpa in the subtree layer, R. schlippenbachii, L. obtusiloba, and V. erosum in the shrub layer, and Oplismenus undulatifolius, Disporum smilacinum, and Parthenocissus tricuspidata
General environmental features of 20 vegetation survey sitesS, south; N, north; E, east; W, west.*Site with no management traces.
Lager Image
General environmental features of 20 vegetation survey sites S, south; N, north; E, east; W, west. *Site with no management traces.
in the herbaceous layer.
- Vegetation and abiotic variables sampling
Vegetation sampling was conducted from August to September 2008. In selected stands, a 50 m line transect was installed parallel with the contour and was > 30 m from all stand edges. The transect was consisted of five plots (5 m × 5 m) spaced at 5 m intervals and 10 subplots (1 m × 1 m). Within each plot (1-2 m in height) and sub-plot (< 1 m in height), coverage (%) of all vascular species was recorded. Coverage of litter and woody debris was also determined in subplots. The age of each stand was estimated by counting tree rings from annual ring cores for the biggest tree species in the plot. Mean ages of Q. serrata and A. firma forests were 36 and 34 years, respec-tively. Samples of unmanaged natural ( Q. serrata ) and re-stored ( A. firma ) stands were used as a control (or refer-ence, or pre-disturbed) state in this study.
In order to estimate light availability and canopy open-ness (%), we took digital hemispherical photographs at each subplot from a height of 1.8 m using a Nikon D80 digital camera (Sendai, Japan) with Sigma 4.5mm F2.8 EX DC Circular Fisheye HSM lens (Fukushima, Japan). Pictures were taken on uniformly overcast days to avoid direct sunlight. Total transmitted light or photosynthetic photon flux density (mol m -2 day -1 ) averaged over a year was calculated with Gap Light Analyzer 2.0 (Mansfield, CT, USA) employing the standard overcast sky model (Frazer et al. 1999, Haugo and Halpern 2007).
- Statistical analyses
Functional types such as growth form and life trait (or habitat affinity) are often used to generalize species’ per-formance (Smith et al. 1987). To compare regeneration patterns after thinning, each species recorded was classi-fied into five growth forms, i.e., tree, shrub, forb (includ-ing ferns), graminoid and vine, and mean numbers and coverage of each form to samples of focal successional stage were calculated. Average coverage of litter and wood debris (WD), and light availability were also computed for each successional stage. In addition, in order to assess species diversity, a Shannon index ( H′ ) was calculated for each successional stage (Magurran 2003).
To analyze successional pathways of stands, the impor-tance value of each life form in each subplot was calculat-ed. The values were determined by multiplying 100 by the fractional value of each life form to the summed values of all functional types in each subplot. Then, a matrix of importance values for life forms was constructed and that data were fed to ordination analysis. We used Nonmetric Multidimensional Scaling (NMS) (Kruskal 1964) ordina-tion to examine differences in successional pathways due to management activities in the understory of natural and planted stands. NMS ordination was performed us-ing Søresen’s distance measure and we selected the “slow and thorough” autopilot of PC-ORD ver. 4.0 (MjM Soft-ware, Oregon, USA) conducting 40 runs that yield one to six dimensional solutions (McCune and Mefford 1999). A Monte Carlo test was performed to compare the stress of randomized data (McCune and Grace 2002). In the re-sult of NMS ordination, the minimum stress configura-tion included three axes, but the two axes accounting for the most variability in the original data set are presented (NMS I as 22.2% and NMS III as 34.7%). SPSS ver. 15.0 (SPSS Inc., Chicago, IL, USA) and PC-ORD ver. 4.2 (Mc-Cune and Mefford 1999) were applied to univariate and multivariate analyses, respectively.
- Changes in abiotic environments
Lager Image
(a-c) Changes in mean coverage of litter and woody debris (WD) and transmitted light as a function of the lapse of time after thinning in the study area. Vertical bars in the diagram represent mean ± standard error. Q. serrata, Quercus serrata; A. firma, Alnus firma.
Canopy conditions (estimated light availability) and ground conditions (coverage of litter and WD cover) changed dramatically in the earlier stages after thinning, and as years passed, means of environmental variables gradually approached the control state ( Fig. 1 ). Coverage of litter decreased immediately after thinning and then approached a reference state with the passage of time in both Q. serrata and A. firma stands ( Fig. 1 a). Immediately after thinning, there were big changes in transmitted light (from 4.3 mol m -2 day -1 to 8 mol m -2 day -1 ) and in coverage of WD (from 8.8% to 17.7%) in Q. serrata stands; the trend was the same for A. firma stands ( Fig. 1 b and 1 c).
- Regeneration trajectory
Ordination results explained 56.9 % of the variation in the raw data (final stress = 17.3, final instability = 0.0297). NMS ordination of subplots was expressed in mean (± standard error) of successional stages ( Fig. 2 ). Variation in vegetation among successional stages was explained by NMS I (22.2%), which represented a gradient from sites with higher coverage of vine species in the negative pro-portion of the NMS I to those with higher scores of tree species in the positive portion. The arrangement of sam-ples along NMS III, which explained 34.7 % of the varia-tion, ranged from sites with higher coverage of graminoids to sites with higher cover of tree species. NMS ordination revealed different successional trajectories depending on forest type. Understory composition of natural stands ( Q. serrata ) gradually approached reference values (N0), but pathways in planted stands ( A. firma ) differed from natu-ral ones as P8 samples were distributed far away from P0 samples.
- Changes in species diversity and abundance
Lager Image
Nonmetric Multidimensional Scaling (NMS) ordination based on ground cover of functional types (growth form) of natural (Quercus ser-rata) and planted (Alnus firma) stands. NMS I and NMS III accounted for 22.2% and 34.7% of the variation in composition of functional types, re-spectively, resulting in a cumulative 56.9% variation in composition. Num-bers after letters (N and P) indicate years since management operation. Q. serrata, Quercus serrata; A. firma, Alnus firma N, natural; P, planted.
In the shrub layer, the increase in overall richness was limited after thinning (see scale of y axes, Figs. 3 a, 3 c, and 3 e). The change in total coverage was affected largely by
Lager Image
(a-f) Changes in mean species richness and coverage of shrub layer species with the lapse of time after thinning in natural (Quercus serrata) and planted (Alnus firma) stands.
Lager Image
(a-j) Changes in mean species richness and coverage of ground layer species over time after thinning in natural (Quercus serrata) and planted (Alnus firma) stands.
the gradual increase of tree species and the decrease of shrubs ( Figs. 3 b, 3 d, and 3 f). A limited gain in species rich-ness after thinning was observed in ground vegetation ( Figs. 4 a, 4 c, 4 e, 4 g, and 4 i). Vine species were excluded from analysis because of too little richness and coverage. In Q. serrata stands, the increase in total richness after thinning was mainly due to an increase in forb species. In A. firma stands, woody (tree and shrub) species slightly increased in cover after thinning. Changes in the diversity index ( H′ ) in natural and planted forests were similar to changes in total species richness and there were limited changes in diversity that ranged from 1.4 to 1.7 in the Q. serrata forest and from 0.8 to 1.1 in the A. firma forest ( Fig. 5 ).
Overall, total coverage of ground vegetation increased after thinning ( Figs. 4 b, 4 d, 4 f, 4 h, and 4 j). Peak coverage of ground communities appeared during earlier stages in both Q. serrata and A. firma stands, after then decreased
Lager Image
Changes in the Shannon index (H′) with the lapse of time after thinning in natural (Quercus serrata) and planted (Alnus firma) stands.
and gained again. In Q. serrata forests, the increase in total coverage was mainly due to forb species and, in A. firma stands, tree species usually contributed to increases in the coverage of ground vegetation.
We applied a chronosequence approach to illustrate community resilience and resistance to a thinning distur-bance by showing the patterns (changes in species rich-ness and abundance) and pathways of Q. serrata and A. firma stands. Halpern and Spies (1995) noted that forest management effects can be grouped into two broad cat-egories: 1) initial effects of disturbances associated with logging and site preparation, and 2) effects of subsequent activities (e.g., brush cutting and mowing) that directly or indirectly alter rates or patterns of succession. In our study, subsequent forest activities after forest operation were not observed. Thus, our results showed only earlier sequences in understory regeneration in lowland natu-ral and planted broadleaved forests in Korea. To acquire more explanatory ecological information on various for-ests in different successional stages, longer-term static observations are strongly recommended.
Tending works brought various changes in forest struc-ture and ground environments. Release of tree density resulted in alteration of light and nutrient products such as decrease of litter fall (Jung et al. 2008). Coverage of ground litter decreased in earlier stages and more than 10 years were required to restore the normal state in both forests ( Fig. 1 ). Increases in ground woody debris in earlier stages are likely due to tree felling operations. Decreases in woody debris in later stages were due to decomposition and reduction of debris production. In addition, loss of rainfall interception because of thinning (Jung et al. 2008) may increase runoff and consequently, by wash away, can affect the amount of litter and woody debris in forests.
After thinning, species richness and diversity increased in earlier stages (within 4 years) but the magnitude was relatively limited in both Q. serrata and A. firma stands ( Figs. 3 - 5 ). Because the disturbance intensity was low (maintaining canopy openness at > 50 %), increments of resources (such as light and water) required to cause species invasion was likely limited. Thinning operations at least do not reduce species diversity because residual species persistence and performance are better (Dodson et al. 2008). A gradual increase of tree cover in the shrub layer and a limited invasion of alien species such as Bidens frondosa, support our interpretation. Recorded invasive and ruderal species entry due to increments in available resources such as light, water and nutrients was rarely observed in either natural or planted forests. B. frondosa, a ruderal exotic and an annual species, and Oplismenus undulatifolius, a ruderal and perennial species, were re-corded in thinned Q. serrata stands.
In herbaceous vegetation, forb species, which contrib-uted largely to increases in richness and total cover in natural stands, are likely able to take advantage of distur-bances in earlier stages. Different patterns of richness and coverage between natural and restored stands are largely explained by domino effects of thinning on A. firma indi-viduals as early successional species planted during the Yeoung-il restoration project period (1972-1977). In a field survey of the managed A. firma stands, uncut A. firma individuals were frequently observed in snapped-off or standing dead states (although we did not check quanti-tatively). This result would be due to short longevity of A. firma. Such a decline in alder species in the canopy layer after thinning can cause additional disturbances on un-derstory composition and structure, light and nutrient availability, soil water and so on.
Community stability (or persistence) may simply re-flect the ability of resident species to resist change, or if altered by disturbances, their ability to readjust or re-cover (Halpern 1988). Different regeneration trajectories of functional type composition between stands reflect the variety of species group responses to disturbances. In addition, resistance and resilience varied between plant communities, and were likely better for natural stands and lower for plantation. The recovery pattern of the un-derstory of Q. serrata stands, as they returned towards the pre-disturbance state, shows higher community stabil-ity or resilience to thinning disturbances ( Fig. 2 ). As was mentioned above, vulnerability of early successional spe-cies, A. firma to disturbance affected compositional prop-erties of understory of plantation, as travel toward alter-native stable state.
This research was partly funded by Grant No. S1210-10L020100 of Korea Forest Service.
Cho YC , Kim JS , Lee CS , Cho HJ , Lee HY , Bae KH 2011 Early successional change of vegetation composition after clear cutting in Pinus densiflora stands in Southern Gangwon Province. J Korean For Soc (in Korean with English abstract) 100 240 - 245
D’Amato AW , Orwig DA , Foster DR 2009 Understory vegeta-tion in old-growth and second-growth Tsuga canadensis forests in western Massachusetts. For Ecol Manage 257 1043 - 1052    DOI : 10.1016/j.foreco.2008.11.003
Dodson EK , Peterson DW , Harrod RJ 2008 Understory veg-etation response to thinning and burning restoration treatments in dry conifer forests of the eastern Ca-scades, USA. For Ecol Manage 255 3130 - 3140    DOI : 10.1016/j.foreco.2008.01.026
Foster BL , Tilman D 2000 Dynamic and static views of suc-cession: testing the descriptive power of the chronose-quence approach. Plant Ecol 146 1 - 10    DOI : 10.1023/A:1009895103017
Frazer GW , Canham CD , Lertzman KP 1999 Gap Light Ana-lyzer (GLA), Version 2.0: Imaging Software to Extract Canopy Structure and Gap Light Transmission Indices from True-Colour Fisheye Photographs, Users Manual and Program Documentation. Simon Frazer Univer-sity, Burnaby, BC & Institute of Ecosystem Studies Millbrook, NY.
Fulé PZ , Laughlin DC , Covington WW 2005 Pine-oak forest dynamics five years after ecological restoration treat-ments, Arizona, USA. Forest Ecol Manage 218 129 - 145    DOI : 10.1016/j.foreco.2005.07.005
Halpern CB 1988 Early successional pathways and the resistance and re-silience of forest communities. Ecology 69 1703 - 1715    DOI : 10.2307/1941148
Halpern CB 1989 Early successional patterns of forest spe-cies: interactions of life history traits and disturbance. Ecology 70 704 - 720    DOI : 10.2307/1940221
Halpern CB , Spies TA 1995 Plant species diversity in natural and managed forests of the Pacific Northwest. Ecol Appl 5 913 - 934    DOI : 10.2307/2269343
Haugo RD , Halpern CB 2007 Vegetation responses to co-nifer encroachment in a western Cascade meadow: a chronosequence approach. Can J Bot 85 285 - 298    DOI : 10.1139/B07-024
Inouye RS , Huntly NJ , Tilman D , Tester JR , Stillwell M , Zin-nel KC 1987 Old-field succession on a Minnesota sand plain. Ecology 68 12 - 26    DOI : 10.2307/1938801
Jung MH , Lee DK , Um TW , Kim YS , Kwon KC , Jung KH 2008 Effects of thinning on nutrient input by rainfall and lit-terfall in natural hardwood forest at Mt. Joongwang, Gangwon-do. Korean J Soil Sci Fertil 41 1 - 8
Kim JW , Lee YK 2006 Classification and Assessment of Plant Communities. Worldscience Seoul. (in Korean)
Kruskal JB 1964 Nonmetric multidimensional scaling: a nu-merical method. Psychometrika 29 115 - 129    DOI : 10.1007/BF02289694
Lee CS , You YH , Robinson GR 2002 Secondary succession and natural habitat restoration in abandoned rice fields of central Korea. Restor Ecol 10 306 - 314    DOI : 10.1046/j.1526-100X.2002.00003.x
Lee KS 2006a Changes of species diversity and development of vegetation structure during abandoned field succes-sion after shifting cultivation in Korea. J Ecol Field Biol (in Korean with English abstract) 29 227 - 235    DOI : 10.5141/JEFB.2006.29.3.227
Lee KS 2006b Changes of biomass, net primary productivity and P/B ratio during abandoned field succession after shifting cultivation in Korea. J Ecol Field Biol (in Korean with English abstract) 29 237 - 245    DOI : 10.5141/JEFB.2006.29.3.237
Lee KS , Lee BC , Shin JH 1996 Classification of forest vegeta-tion zone over southern part of Korean Peninsula using geographic information systems. Korean J Ecol (in Korean with English abstract) 19 465 - 476
Lee SM , Cho YC , Shin HC , Oh WS , Seol ES , Park SA , Lee CS 2008 Successional changes in seed banks in abandoned rice fields in Gwangneung Central Korea. J Ecol Field Biol 32 269 - 276
Magurran AE 2003 Measuring Biological Diversity. Wiley-Blackwell New York.
McCune B , Grace JB 2002 Analysis of Ecological Communi-ties. MjM Software Design Gleneden Beach, OR.
McCune B , Mefford MJ 1999 PC-ORD. Version 4.0. Multi-variate Analysis of Ecological Data. MjM Software De-sign Gleneden Beach, OR.
Metlen KL , Fiedler CE 2006 Restoration treatment effects on the understory of ponderosa pine/Douglas-fir forests in western Montana, USA. For Ecol Manage 222 355 - 369    DOI : 10.1016/j.foreco.2005.10.037
Metlen KL , Fiedler CE , Youngblood A 2004 Understory re-sponse to fuel reduction treatments in the Blue Moun-tains of northeastern Oregon. Northwest Sci 78 175 - 185
Moore MM , Casey CA , Bakker JD , Springer JD , Fulé PZ , Cov-ington WW , Laughlin DC 2006 Herbaceous vegetation responses (1992-2004) to restoration treatments in pon-derosa pine forest. Rangeland Ecol Manage 59 135 - 144    DOI : 10.2111/05-051R2.1
Nelson CR , Halpern CB , Agee JK 2008 Thinning and burn-ing result in low-level invasion of nonnative plants but neutral effects on natives. Ecol Appl 18 762 - 770    DOI : 10.1890/07-0474.1
Parsons WFJ , Knight DH , Miller SL 1994 Root gap dynam-ics in lodgepole pine forest: nitrogen transformations in gaps of different size. Ecol Appl 4 354 - 362    DOI : 10.2307/1941939
Pickett STA , Likens GE 1989 Space for time substitution as an alterna-tive to long-term studies. In: Long-term Studies in Ecol-ogy: Approaches and Alternatives Wiley & Sons Chichester 71 - 88
Smith TM , Shugart HH , Woodward FI 1987 Plant Func-tional Types: Their Relevance to Ecosystem Properties and Global Change. Cambridge University Press Cam-bridge.
Sutherland JP 1981 The fouling community at Beaufort, North Carolina: a study in stability. Am Nat 118 499 - 519    DOI : 10.1086/283844
Thomas SC , Halpern CB , Falk DA , Liguori DA , Austin KA 1999 Plant diversity in managed forests: understory re-sponses to thinning and fertilization. Ecol Appl 9 864 - 879    DOI : 10.1890/1051-0761(1999)009[0864:PDIMFU]2.0.CO;2
Woo BM 2003 Achievements of the forest-tending works by national movement of the forest for life under the IMF Structural Adjustment Program in Korea. J Korean For Soc 92 145 - 151