Ecol Res (2009) 24: 281–290 DOI 10.1007/s11284-008-0505-1 O R I GI N A L A R T IC L E Jiri Dolezal Æ Jong-Suk Song Æ Jan Altman Stepan Janecek Æ Tomas Cerny Æ Miroslav Srutek Jiri Kolbek Tree growth and competition in a post-logging Quercus mongolica forest on Mt. Sobaek, South Korea Received: 22 November 2007 / Accepted: 7 April 2008 / Published online: 28 May 2008
The Ecological Society of Japan 2008 Abstract Secondary woodlands in South Korea cover most mountains from low to middle elevations. While general patterns of forest succession are well understood, little is known about mechanisms of stand recovery after disturbance. We examined the spatio-temporal variations in establishment, growth, size inequality, and mode of competition among trees in a 50-year-old post-logging Quercus mongolica-dominated stand. We further compared the growth and stem allometry of single trees, presumably of seed origin, with multi-stemmed trees resprouting from stumps. Q. mongolica formed the upper canopy 16–22 m tall, 88.3% of total stand basal area, and 36.2% of total stem density, with most trees established during the first post-logging decade (51.2% were resprouts). During the first three decades, the Q. mongolica recruits grew exponentially, and disproportionately more in diameter than few older reserved trees left after the last cutting. This substantially decreased size inequality. The reverse trend was observed from 1994 to 2004: larger trees grow more, indicating an increasing asymmetry of competition for light. Neighborhood analysis revealed that when target trees had more or larger neighbors, their exponential phase of growth was reduced and maximum J. Dolezal (&) Æ S. Janecek Department of Functional Ecology, Institute of Botany, Czech Academy of Sciences, Dukelská 135, 379 82 Třeboň, Czech Republic E-mail: jiriddolezal@gmail.com J. Dolezal Æ J. Altman Æ M. Srutek Faculty of Science, Department of Botany, University of South Bohemia, Na Zlate stoce 1, 370 05 České Budějovice, Czech Republic J.-S. Song Department of Biological Science, College of Natural Sciences, Andong National University, Andong, Gyeongbuk 760-749, South Korea T. Cerny Æ M. Srutek Æ J. Kolbek Department of Geobotany, Institute of Botany, Zámek 1, 252 43 Průhonice, Czech Republic size was decreased. After the 50 years of stand development, more than 70% of Q. mongolica showed growth decline as a result of competitive stress, and mortality was about 30%, concentrated in smaller size classes. Compared to single stems, resprouts within clones do not seem to compete less asymmetric as might be expected based on studies of clonal herbaceous plants and physiological integration within genets. As Q. mongolica was also negatively affected by competition from woody species currently prevailing in the lower tree stratum (Tilia amurensis, Acer mono, Fraxinus rhynchophylla, Acer pseudosieboldianum), we predict the stand will become increasingly dominated by these more shade-tolerant trees. Keywords Size-dependent growth Æ Quercus mongolica Æ Sprouting Æ Neighbor effects Æ Size inequality Æ Richards growth model Introduction Many temperate and tropical forests have experienced a significant rise in logging activities in the second half of the 20th century. This has increased the importance of understanding the patterns and mechanisms of stand recovery after disturbance (Halpern 1988; Pinard et al. 2000). There is a large body of empirical studies that related the pattern of forest recovery to the nature of canopy disturbance (e.g., selective vs. clearcut logging), interaction with ground-layer vegetation, and topsoil perturbation (Ishikawa and Ito 1989; Pinard et al. 2000), but also to composition of the pre-disturbance community (Yoshida et al. 2006), and the reproductive characteristics of the available species (Kammesheidt 1998; Vesk and Westoby 2004). Fewer studies have examined how the pattern of forest recovery is influenced by growth and local interactions of trees early in stand development (e.g., Xiaodong and Shugart 2005). The present study examines the spatio-temporal changes in growth and survival of Quercus mongolica in a 282 post-logging stand in South Korea in relation to mode of regeneration and competition among trees. Quercus mongolica is a canopy tree species that dominates cool-temperate forests in NE Asia (Tatewaki 1958). Its natural regeneration is often inhibited by dense bamboo understory and high rodent densities and acorn removal rates beneath oak canopies (Iida 2004). Q. mongolica can be described as a long-lived pioneer because it often requires large stand disturbances (or bamboo dieback) to regenerate successfully from seeds (Ishikawa and Ito 1989; Osawa 1992), but then gradually overgrow other species (Kato and Hayashi 2007) to form the upper, more or less even-aged canopy for 1–300 years (Sano 1997; Abrams et al. 1999). The low acorn germination rates due to animal feeding have been recently reported from mixed-oak forests in South Korea, with more than 70% seedlings originating from sprouts (Suh and Lee 1998). The intense resprouting in Q. mongolica has been linked with frequent occurrence of short, crooked, and multi-stemmed trees, which contrasts with the straight and longer trunks of the same species (syn. Q. crispula, Q. grosseserrata, Q. liaotungensis) growing in northeast China and northern Japan, where single trees of seed-origin are thought to be more common than resprouts (Wu 2000; Takahashi et al. 2003). As sprouts exhibited a higher growth potential than seedlings, Suh and Lee (1998) have predicted that the number and size of sprouts will determine the species composition and stand structure of future secondary forests after cutting. To our knowledge, these predictions remain largely unexplored. Here, we use a dataset consisting of single- and multi-stemmed spatially mapped tree individuals to assess the relative importance of different regeneration strategies in post-logging recovery of a mixed-oak forest. A number of studies have shown that resprouts early in stand development gain competitive advantage by utilizing resources of stem remnants they originated from (Bellingham and Sparrow 2000), and hence grow disproportionaly more than stems emerging from seeds (Cao and Peters 1998). However, later in stand development, the resprouts can grow disproportionaly less than single stems as they experience more intense competition within dense polycormons (Peterson and Jones 1997). To quantify these temporal size-dependent growth variations, we used the size-growth analysis (Hara 1988) that enables to delineate between linear, size-proportional and non-linear, size-disproportional increase in growth rate, and to infer an alternative mode of plant interference (symmetric vs. asymmetric competition, Weiner 1990). We assumed that if trees grow in proportion to their sizes, thereby depleting limited resources without any individual obtaining a predominance, competition among them is weak and symmetric, whereas if larger individuals grow disproportionately more than others, thereby preemptying resources at the expense of smaller plants, competition is intense and size-asymmetric (Newton and Jolliffe 1998; Hou et al. 2006). One major drawback to ‘size-growth’ approach is that the growth of individuals is taken as a series of separate intervals, with no assumptions about the temporal pattern of plant growth beyond the single growth interval analyzed (Damgaard et al. 2002). Recently, the modelling of the individual growth has been applied in analyzing and interpreting plant interaction as an alternative approach to the size-growth analysis (Weiner 1995; Nord-Larsen et al. 2006). This method is based on fitting individual cumulative growth curves representing general size–age relationship with explicit growth equations such as the Richards function (Richards 1959), which is the generalization of the classical curves of logistic growth (Damgaard et al. 2002), and looking at the distribution of growth parameters among individuals, and at the effects of different treatments or factors such as stem origin or different intensity of neighbor competition on these distributions. Mountain ranges in South Korea are covered by secondary woodlands from low to middle elevations (Lee et al. 1989). While general patterns of forest succession and stand productivity are well understood (e.g., Kim and Yang 1996; Son et al. 2004), little is known about factors that affect tree growth and size-structural dynamics. We studied a young Q. mongolica population 50 years after logging by combining (1) size-growth approach described above, with (2) the modelling of individual growth using the Richards equation, and (3) the assessment of neighbor influence on radial increments (Phillips et al. 2003). Growth variations were determined from tree rings through dendroecological techniques (Schweingruber 1996) and related to stem origin, size, age, and local crowding intensity. Methods Study site Research was conducted on the northwestern slope of Mt. Sobaek (1,439 m a.s.l.) in the Sobaek Range, ca. 21 km northwest of the Yeongju city (36 52¢N, 128 31¢E). The climate is temperate to continental, with dry cold winters and hot humid summer. According to climatic data from Yeongju (1970–2000), mean annual temperature is 11.2C, with mean for January
3.2C (monthly minimum temperature
16.5C), and for July 23.9C (monthly maximum temperature 33.9C). The summer monsoon brings abundant moisture from the ocean and produces heavy rainfall. Mean (±SD) annual precipitation is 1236.9 ± 272.4 mm. Seasonal distribution of precipitation is 7.9% in winter, 30.1% in spring, 53.2% in summer, and 8.8% in autumn. January receives the lowest (19 mm) and July the highest (259 mm) amount of rainfall. The soil in our study site is mesotrophic forest cambisol derived from granitic/pegmatitic gneisses. Mean (±SD) values of soil chemical properties: pHwater 4.72 ± 0.18, pHKCL 3.96 ± 0.18, total nitrogen 283 0.72 ± 0.16%, total phosphorus 5.94 ± 1.85%, exchangeable magnesium 131.6 ± 11.9 mg/kg and calcium 1527.8 ± 167.1 mg/kg. The mean values are from three composite soil samples (0–5 cm) collected with a hand-auger in our plots (5–10 cores evenly distributed on the whole plot area), after the upper roots were discarded if present. The studied secondary forest occurred at 900–1,000 m elevation and established naturally after clearing an oldgrowth stand in mid 1950s and belongs to the Lindero obtusilobae–Quercion mongolicae alliance (Krestov et al. 2006). The studied stand had a well-developed herb and shrub stratum (Fig. 1). Eighty-six vascular plant species were found in the herb-layer (<0.5 m, data from 12 vegetation relevés 10 · 10 m each) including 73 forbs, 14 tree seedling species, 12 shrub seedling species, four lianas, four grasses, one sedge, and three ferns. The most common were Carex siderosticta (average coverabundance of 11.8%), Astilbe koreana (7.5%), Ainsliaea acerifolia (4.3%), Osmunda cinnamomea (3.3%) and Melampyrum roseum (2.3%). Dwarf-bamboo Sasa borealis, one of the common species of mixed-oak forests, occurred only locally. The low bamboo cover may partly explain a high species richness of understory vegetation (e.g., Kim and Yang 1996). In the shrub layer (0.5–5 m), 20 woody species occurred, most commonly Lindera obtusiloba (5.3%), Tripterygium regellii (2.9%), Stephanandra incisa (2.8%), and Acer pseudosieboldianum (1.9%). Data collection and analysis The tree measurement was conducted during the summer of 2005 in three 20 · 20-m permanent plots randomly located within the core zone of secondary forest approximately 30–60 apart. Each live and dead tree >1.3 m tall within the plot was identified and marked, and its position (x,y coordinates), breast height diameter (DBH, at 1.3 m), tree height, and crown base height recorded. In total, 406 trees >1.3 m tall were recorded within a 0.12-ha area. Additionally, to avoid edge effects, trees occurring up to 5 m of the plot border were measured and used as neighbors but not as target trees in subsequent analysis of neighbor effect on tree growth. Since there were no significant differences in species composition and size and age structure among plots (K–W test, P > 0.05), the data were pooled together for further analyses. Whether the stems originated from seeds or from sprouts was determined by examining stem base connections and searching for stump remnants, but in general it is difficult to clearly distinguish sprouts from the individuals of seed origin (Roth and Hepting 1943). Finally, all multi-stemmed individuals were considered of sprout origin, while single stems of seed origin. Spatial coordinates of trees and stem/crown heights were measured using the ultrasonic Haglöf Vertex III Hypsometer (height resolution: 0.1 m, distance resolution: 0.01 m) equipped with the 360 transponder. Ninety-seven wood cores were extracted from 72 randomly selected trees (20, 24, and 28 trees in plots 1, 2, and 3, respectively) at the 0.3–0.6 m above ground surface with steel borer (Mora, Sweden) to age the trees and to reconstruct their growth histories. In asymmetric stems, two cores were extracted on opposite sides, while one core was taken in symmetric, cylindrical trunks. The cores were dried, mounted, sanded, and inspected for injuries, reaction wood and other aberrant features. Rings were counted from pith to bark and their widths measured to the nearest 0.01 mm using the TimeTable measuring device and PAST32 software (http:// www.sciem.com); the ring-sequences were cross-dated visually using the pattern of wide and narrow rings, and verified using the program PAST32. In ca. 27% of cores in which the pith was missed (as easily happens in hardwoods with asymmetric stem), the number of missing rings was estimated from the diameter of the innermost tree-ring and the average width of the five following rings. Size- and age-dependent growth: The temporal sizedependent growth responses were examined by relating radial increments (absolute growth rate, AGR) for a specified time-period (e.g., 1999–2004) to stem size at the beginning of this period using the log-log linear regression: logðAGR1999
2004 Þ ¼ a þ q1 logðstem diameter1999 Þ þ c ð1Þ Fig. 1 A 50-year-old post-logging Quercus mongolica-dominated stand in Sobaek Mts., South Korea with Astilbe koreana and Osmunda cinnamomea as dominant understory species where a and q are intercept and slope parameters specific to given growth interval, and c the error term. The size inequality at the end of each growing interval was expressed by coefficient of variation (CV), based on the variance of cumulative increments (stem diameters) obtained from tree-ring analysis. The growth was further examined as a function of both size and age of a tree using multiple regression: 284 ð2Þ where the partial regression coefficient q2 significantly different from zero indicates that trees of the same size have different growth rates depending on their age. The growth-size relations were first determined for all Q. mongolica trees, and then separately for the singleand multi-stemmed trees. ANCOVA was used to test for differences in the slope parameter estimates between trees of different origin. Individual growth curves: Whether target trees are still growing actively and thus have no apparent asymptote, or if they show a decline in radial growth, was tested by fitting individual cumulative growth curves with the simple linear model and non-linear Richards model (Richards 1959): h i1=ð1
dÞ Dt ¼ A 1 þ ðd
1Þe
Kðt
yÞ ð3Þ where Dt is stem diameter at age t, A is the asymptotic or maximum diameter, K defines ‘‘speed’’ of growth, d determines more or less sigmoid shape of the curve (the shape parameter is the rate at which diameter approaches the asymptote at inflection point), and y is the age of growth inflection. The Richards growth model is the generalization of the classical growth curves: Gompertz (d = 1), logistic (d = 2), von Bertalanffy (d = 2/3) and can be compared against these simpler models (Damgaard et al. 2002). We determined the parameters in the individual-based growth models and then tested if trees with different growth pattern differ in stem origin, size, age, and intensity of crowding they experienced. Neighborhood analysis: To assess the influence of neighbors on target tree growth and stem allometry (relative growth rates calculated as RGR ¼ ½lnðyiþ1 Þ
lnðy1 Þ=yr, where yi+1 is final diameter, y1 is initial diameter, and year is the length of the growth period in years), several indices of local competition (CI, or index of crowding intensity) were calculated for every tree to account for both density- and size-related neighbor effects. We used the set of linear regressions to relate tree growth to density-dependent measures calculated as (1) the number of all live neighbors growing closer than 3 m from the target stem or (2) the number of all live neighbors taller than the target tree. Sizedependent measures were calculated as (3) the sum of individual basal areas or stem volumes of all live neighbors divided by their distances from the target tree within a circle of 3-m radius around the target stem, or (4) the sum of individual stem basal areas or volumes of neighbors taller than the target stem. By discounting the effects of smaller neighbors on target tree growth, the asymmetry of competitive effect on growth was examined, i.e., if neighbors larger than target tree have more than size-proportional negative effect on growth (Thomas and Weiner 1989). To describe the intraspecific interference in the species-mixed stand, the indices were calculated using only conspecifics as neighbors, and to describe the effect of interspecific competition, CI were calculated using neighbors of all species or heterospecific neighbors. The neighborhood analysis was restricted for the last 5 years, during which the arrangement of stems was likely to be the same for the entire period. A neighbor was defined as any tree within 3 m of a subject individual in order to include only the nearest neighbors that can directly interact with a target tree through the overlapping crowns and root systems (Weiner 1984). The CI values were square-root-transformed to achieve normality of the residuals and homogeneity of variances. Significant negative correlation between growth and neighborhood indices was considered as an indirect evidence of competition. Results Composition and stand structure The stand was dominated by Q. mongolica (88.3% of total stand basal area) that forms the main canopy 16– 22 m tall, with Tilia amurensis (3.9%), Fraxinus rhynchophylla (3.2%), Acer pseudosieboldianum (1.6%) and Acer mono (1.2%) occurring from low to moderate density in the lower canopy stratum (Fig. 2). The remaining woody species had each less than one percent of total stand basal area (Table 1). Quercus mongolica, Acer pseudo-sieboldianum, Lindera obtusiloba and Tilia amurensis dominated in order of abundance. Thirty-one percent of Q. mongolica trees were standing dead stems (of which 36% was multi-stemmed trees). Except for liana Tripterygium regellii, all tree species had some multi-stemmed trees (Table 1). The highest proportions were found in Acer pseudosieboldianum (98.4%), Fraxinus rhynchophylla (80%) and Acer mono (70%). Q. mongolica had 51.3% stems growing in polycormons (68% clones contained two, 26% three, and 6% four stems). Mean number of conspecific neighbors (±SD) 40 Quercus mongolica Live 35 Quercus mongolica Dead Acer pseudo-sieboldianum 30 Frequency logðAGR1995
2000 Þ ¼ a þ q1 logðstem diameter1955 Þ þ q2 ðage1995 Þ þ c Tilia amurensis 25 20 15 10 5 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Tree height (m) Fig. 2 Height distributions for abundant tree species 285 The results of univariate regressions relating radial increment of Q. mongolica during a 5-year period to the stem diameter at the beginning of this period show that in seven out of nine 5-year periods analyzed, the slope parameter estimates for relationship between AGR and size on a log–log scale were significantly greater than zero but less than one (Table 2; Fig. 3). This means that the relationship between size and relative growth rate (RGR) was negative, i.e., the relative increment of smaller trees was higher than that of larger trees, and hence size inequality, expressed by coefficient of variations of stem diameters, decreased with stand age. In the multiple regression model where size was held constant, the tree age had a significant negative effect on radial growth rate in periods from 1964 to 1984, i.e., younger trees grew faster than older trees of the same size. Hence, during the early phase of stand development, the stand was composed of many small Q, mongolica trees and a few older reserved trees, most of which were single stems left after the last cutting. The younger trees were growing faster than older and larger trees during the 1959– 1984 (Table 2), and hence, the relative size differences due to differences in age decreased rapidly during that period. During the 1984–1994 period, the stand became increasingly saturated: average radial increments declined and tree size variability decreased to remain almost constant (‘period of weak, symmetric competition’). Table 1 Tree species composition, total number of stems >1.3 m tall per 0.12 ha (number of individuals that grow in polycormons are given in bracket), their basal area recalculated per ha, mean and maximum breast height diameter (DBH), stem height, and relative crown length (RCL, crown:height ratio) and stem slenderness index (SSI, stem height:DBH ratio) within a circle of 3-m around Q. mongolica was 2.4 ± 1.3 individuals when all neighbors were counted and 1.2 ± 1.3 individuals when taller neighbors were considered. As expected, the multi-stemmed Q. mongolica had more conspecific neighbors (2.9 ± 1.3) than single trees (1.8 ± 1.2) to a distance of 3 m. Mean number of heterospecific neighbors (±SD) within a circle of 3 m around Q. mongolica was 5.6 ± 3.9 individuals when all neighbors were considered and 0.2 ± 0.5 individuals when only taller neighbors were counted. Size-dependent growth responses Live trees Acer mono Acer pseudo-sieboldianum Corylus sieboldiana Fraxinus rhynchophylla Lindera obtusiloba Maackia amurensis Morus bombycis Quercus mongolica Symplocos chinensis Tilia amurensis Tripterygium regelii Ulmus davidiana Weigela florida Dead trees Basal area 2 DBH (cm) Height (m) Count % Count (m ) % Mean Max Mean Max 10 63 5 10 46 12 11 119 10 36 1 3 4 3.0 19.1 1.5 3.0 13.9 3.6 3.3 36.1 3.0 10.9 0.3 0.9 1.2 0 0 0 0 0 4 (2) 0 55 (20) 0 2 (1) 0 1 0 1.38 1.89 0.02 3.73 0.41 1.05 0.20 102.16 0.15 4.50 0.03 0.20 0.01 1.2 1.6 0.0 3.2 0.4 0.9 0.2 88.3 0.1 3.9 0.0 0.2 0.0 7.1 3.1 1.4 11.9 2.0 4.5 2.9 19.5 2.4 6.9 4.2 5.3 1.1 12.1 8.1 2.2 25.1 3.5 20.7 5.2 49.6 6.0 18.1 8.9 3.5 1.9 10.5 2.2 4.1 3.3 16.1 2.4 6.7 18 9.2 1.9 17 8.5 2.7 18 4 10 5 23 3 15 (7) (62) (3) (8) (41) (6) (2) (61) (3) (10) (2) (3) 8.9 1.4 18.5 2.2 RCL SSI 0.49 0.39 0.19 0.44 0.31 0.24 0.39 0.54 0.26 0.46 1 0.42 0.43 1.33 1.11 1.42 1.02 1.12 1.17 1.24 0.91 1.18 0.97 4.28 1.57 1.77 Table 2 Changes in intercept and slope parameter estimates obtained from regressions of log AGR (radial growth increments) on log size (stem diameter) and tree age of Quercus mongolica trees with explained variance and significance levels, and size inequality of stem diameters (measured as coefficient of variation, CVd) Period Intercept Log size Age r2adj CVd CVAGR AGR (mm) Size*origin 1959–1964 1964–1969 1969–1974 1974–1979 1979–1984 1984–1989 1989–1994 1994–1999 1999–2004 3.09 1.86 1.41 1.34 1.17 0.05
0.40
1.92
3.48 0.07* 0.50*** 0.59*** 0.59*** 0.59*** 0.82*** 0.87** 1.21*** 1.58*** –0.007 ns
0.012**
0.017**
0.011***
0.009***
0.006ns
0.004ns
0.004ns
0.006ns 0.14 0.40 0.33 0.31 0.21 0.24 0.16 0.27 0.31 0.91 0.71 0.53 0.43 0.38 0.36 0.35 0.35 0.36 0.31/0.60 0.34/0.38 0.32/0.39 0.26/0.31 0.29/0.34 0.39/0.42 0.45/0.58 0.51/0.64 0.56/0.76 18.9/15.3* 27.3/22.2a 27.5/21.4* 33.5/24.5*** 27.0/21.8* 20.3/16.7a 15.6/13.1ns 12.6/10.1ns 11.5/10.0ns 0.02/0.19 ns 0.32/0.14*** 0.12/0.14ns 0.13/0.15ns 0.21/0.21ns 0.39/0.86*** 0.42/1.25*** 0.85/1.39*** 1.11/1.89*** Shown is also spatial variability in radial increments (CVAGR) for single-stemmed and multi-stemmed trees (seed- vs. sprout-origin), their average absolute growth increments (AGR), and slope parameter estimates obtained from separate log AGR-log size regressions (size · origin). ANCOVA was used to test for significant size · origin interactions NS non-significant, aP < 0.1, *P < 0.05, **P < 0.01, ***P < 0.001 286 60 30 AGR stem 1999-2004 (mm) AGR stem 1974-1979 (mm) Fig. 3 Least-squares regressions of Q. mongolica stem diameter increment during a 5-year growth period on stem size at the beginning of this period (see Table 2) for the single- (shaded circles, thin lines) and multi-stemmed (filled circles, thick line) trees 50 40 30 20 10 0 50 100 150 200 25 20 15 10 5 0 250 0 Stem diameter in 1974 (mm) Individual growth curves Tree growth generally showed a sigmoidal pattern, with early exponential phase followed by levelling off later in ontogeny. The growth curves of all single stems Q. mongolica trees were better fitted by the Richards model (r2 > 0.99) than the linear model (Fig. 5). The Richards model, which includes the additional shape parameter (d), gave equal or better fit than the single logistic or Gompertz model. Among the multi-stemmed trees, 67% of individuals had cumulative growth curves better fitted by the Richards model, while the remaining individuals were still growing actively in stem diameter, 200 300 400 8 Single-stemmed trees Multi-stemmed trees 7 Anual tree-ring width (mm) Only during the last decade of stand development, larger trees grow disproportionately more than smaller trees (the slope of >1 in Eq. 1), increasing stem size inequality (Table 2). Tests comparing the single- and multi-stemmed trees in terms of absolute growth rate indicate many significant patterns. First, single stems had higher mean radial increments than multi-stemmed trees, significantly in four out of nine 5-year periods analyzed within the first three decades (Table 2; Fig. 4). Second, tests including stem origin into size-growth regression model show that the slope for log AGR-log size relationship was steeper in single-stemmed than multi-stemmed trees during the 1964–1969 (significant size · origin interaction, Table 2), while the reverse pattern was found significant during the 1984–2004. The steeper slope in single stems during the 1964–1969 resulted from higher relative increments of a few older reserved trees compared to post-logging recruits. In multi-stemmed trees the slopes of >1 during the 1984–2004 period indicate asymmetric competition, i.e., larger stems within clones grow disproportionately more than smaller siblings. The differences between single- and multi-stemmed trees in the slope parameter estimates during the 1994–2004 period were no longer significant when only the post-logging trees of similar age were compared (in single stems, excluding older reserved trees led to the slope >1 during the 1994–2004 period). 100 Stem diameter in 1999 (mm) 6 5 4 3 2 1 0 1950 1956 1962 1968 1974 1980 1986 1992 1998 2004 Year Fig. 4 Mean curves of individual ring width series for single- and multi-stemmed Q. mongolica trees and were better fitted with the linear than the Richards model. These actively growing trees with no sign of growth decline (no apparent asymptote) were dominant taller trees within clones (Fig. 5), while the asymptotically growing siblings were about 3 m shorter with a half stem volume, but of similar age as dominant stems within polycormons (Table 3). Based on the estimated parameters in the Richards model, inferior and dominant individuals could also be distinguished among single trees. The dominant individuals (>50 years old) were all the reserved trees, about 15–20 years older than the post-logging recruits (Fig. 5). The dominant individuals had larger size at which radial growth began to deviate from exponential, and hence larger the maximum size (asymptotic diameter) that could be achieved (Fig. 6). The inferior trees (<50 years old) were significantly younger, but not necessarily shorter (non-significant differences in stem heights; ANOVA, P > 0.05) than dominant trees; the inferior trees had significantly slenderer stem, and thus smaller stem volume. The comparison of asymptotically growing singleand multi-stemmed post-logging trees showed significant differences in all estimated parameters of the Richards model (ANOVA, P < 0.05), albeit there were non- 287 400 400 A 350 B 350 300 stem diameter [mm] stem diameter [mm] Fig. 5 Cumulative stemdiameter growth curves and age structure of stems taller than 1.3 m for single- (a) and multi-stemmed (b) Q. mongolica 250 200 150 100 300 250 200 150 100 50 50 0 1935 1958 1981 0 1935 2004 1958 Year Table 3 Mean values of parameter estimates (A: asymptotic diameter, mm; K, growth coefficient, year
1; d, the shape parameter; y: the age of growth inflection) in the individual-based Richards growth model (Eq. 3), and mean age (years), stem height (m), 1981 2004 Year crown length (CL in m), relative crown length (RCL), index of stem slenderness (SSI), stem volume (StVol, m3) for different groups of Q. mongoliga trees Tree status A K d y Age Height CL RCL SSI StVol D_H.t. D_Q.m. V_H.t. V_Q.m. Single-stemmed Multi-stemmed Single-stemmed (>50 years) Single-stemmed (<50 years) Multi-stemmed (<50 years) Multi-stemmed (linear) 224.9 149.1 323.7 195.1 146.3 0.11 0.13 0.09 0.12 0.13 1.57 1.88 0.76 1.82 1.94 22.3 16.5 37.4 17.6 15.6 48.5 45 65.1 43.1 43.8 46.8 0.65 0.51 0.68 0.62 0.47 0.61 0.85 0.42 1.56 0.56 0.27 0.59 0.4/0.6 0.2/0.7 0.5/0.8 0.4/0.5 0.2/0.8 0.5/0.5 17.6 14.6 18.6 17.3 14.3 17.7 11.4 7.8 12.9 10.8 7.4 10.4 0.86 0.94 0.65 0.93 0.98 0.93 4.6/0.1 5.1/0.4 5.2/0 4./0.1 5.5/0.4 8.8/0 1/0.6 1.5/1.7 0.8/0.4 1.1/0.8 1.5/1.9 2.3/0.7 0.07/0.08 0.08/0.1 0.01/0 0.1/0.1 0.09/0.1 0.04/0 Shown is also average number (D) and sum of stem volume (V in m3) of heterospecific (H.t.) and conspecific (Q.m.) smaller/larger neighbors at distances of 3 m from target oak trees 480 Asymptotic diameter (mm) 480 Asymptotic diameter (mm) Fig. 6 Relationships between parameter estimates of Richards growth model for the asymptotically growing single(shaded circles, thin lines) and multi-stemmed trees (filled circles, thick line) of Q. mongolica 400 320 240 160 80 0 0 10 20 30 40 50 400 320 240 160 80 0 0 Age of growth inflection (yr) significant age differences. The single-stemmed trees were able to grow longer exponentially and to attain larger stem diameter, were significantly taller with a greater stem volume and longer crowns, and experienced less crowding. Hence, it was mainly competition among neighbors that caused individual growth decline before any age-related decrease. Tests comparing the distributions of growth and size parameters suggest that the dominant status in Q. mongolica individuals decreases in the sequence: single-stemmed >50 years old, multistemmed <50 years old with no sign of growth decline, single-stemmed <50 years old asymptotically growing multi-stemmed trees <50 years old. Multiple compari- 1 2 3 Max. RGR sons revealed non-significant differences between the second and third group in all size parameters estimated (ANOVA followed by Tukey HSD post-hoc test, P > 0.05). Neighbor effects on tree growth The diameter increment (log AGR) of Q. mongolica from the 1999 to 2004 period was positively correlated with log stem diameter in 1999, uncorrelated with tree age in 1999, and negatively correlated with competition indices (Tables 2, 4). In the multiple regression model, stem size 288 Table 4 The beta coefficients from univariate regressions of Q. mongolica tree variables (log AGR1999–2004, n = 63; RCL and SSI, n = 119) on various indices of local crowding intensity, based on density, basal areas and stem volumes of neighboring trees, weighted (w) or not (un) by their distances from the target tree Target Neighbor DCA Competition indices based on Density Q.m. AGR Q.m. live H.t. live A.t. live Q.m. RCL Q.m. live A.t. live Q.m. SSL Q.m. live A.t. live 0 1 0 1 0 1 0 1 0 1 0 1 0 1 Basal area Volume w un w un w un – – – – – –
0.23
0.47 0.23
0.52 0.23
0.60
0.40
0.55
0.11
0.51 0.29 0.50 0.05 0.49
0.09
0.35
0.23
0.41
0.10
0.42
0.05
0.21 0.03
0.19 0.05 0.16
0.06 0.16 0.01
0.35
0.28
0.40
0.15
0.47
0.31
0.45
0.23
0.40 0.18 0.43 0.11 0.42
0.03
0.41
0.33
0.53
0.08
0.50
0.14
0.37
0.03
0.34 0.17 0.37 0.10 0.36 0.06
0.39
0.37
0.52
0.13
0.51
0.31
0.45
0.23
0.40 0.28 0.51 0.25 0.50 DCA degree of competitive asymmetry (0 all individuals used to define a local neighborhood, 1 only larger individuals considered as neighbors). Significant results (P < 0.05) are in bold. Abbreviations as in Table 3 Q.m. Quercus mongolicam, A.t. all trees, H.t. heterospecific trees and competition together accounted for 37% of the variation in radial growth rate from 1999 to 2004. The log AGR was further positively and significantly correlated with final tree height (r = 0.40), stem volume (r = 0.36), crown length (r = 0.39) and negatively correlated with stem slenderness (r = –0.34). As expected, the asymptotically growing trees had significantly smaller diameter increments from the 1999 to 2004 than trees whose growth curves were better fitted with the linear model. Finally, tree age of Q. mongolica in 2004 was positively correlated with stem volume (r = 0.64, P < 0.05). Relative crown height (crown length:tree height ratio) and stem slenderness (tree height:stem diameter ratio) of Q. mongolica were significantly correlated with our measures of local crowding when included larger than target trees as neighbors, i.e., trees grew taller than wider and had smaller crowns when surrounded by taller neighbors. The competition indices that were significantly negatively correlated with the diameter increments from the 1999 to 2004 incorporated both intra- and interspecific effects and the asymmetry value of 1, i.e., when the effects of smaller neighbors on target tree growth were excluded (Table 4). Significant r2 of regressions discounting the effects of smaller neighbor suggests that asymmetric competition was occurring among Q. mongolica stems from the 1999 to 2004. Explained variance (adjusted r2) from the univariate regressions of RGR of Q. mongolica from the 1999 to 2004 on various competition indices ranged from 3.7 to 14.3%, and it increased when (1) larger individuals were used to define a local neighborhood, (2) competition indices were calculated from stem volumes than basal areas, (3) neighboring trees were not weighted by their distances to a target tree, and (4) when heterospecific individuals were included. However, the competition index that accounted for most of the variation was based solely on density of all neighboring individuals larger than target trees (adjusted r2 = 0.143, P = 0.01). Discussion Successional stages similar to our stand were observed at other sites in South Korea, where Q. mongolica is consistently the early dominant of post-logging areas, followed by the reinvasion of more shade-tolerant Tilia amurensis, Acer mono, and Acer pseudosieboldianum, which prevail in the sapling and lower tree stratum. Some previous studies have reported this replacement sequence from intolerant light-demanding woody species to mid-tolerant to tolerant species (Suh and Lee 1998), while other studies have shown that the secondary succession is the result of patterns of differential growth, not relay floristics (Osawa 1992). According to Osawa (1992), all the trees in post-disturbance stand in Hokkaido, northern Japan, were of similar ages, and therefore have separated themselves via differential growth, not by seeding in at different times. Q. mongolica does not have necessarily greater shade-tolerance and longer life-span than other woody species, but grows faster and taller, gradually overtopping other species and creating stratified mixed-species stands. The lack of trees <8 m tall indicates that Q. mongolica is not continuing to establish in our stand, and hence its density was determined by the relative success of seedling establishment and stump sprouting during the early stage of stand development. Q. mongolica had 53% stems of sprout-origin, while other hardwoods had 70–90% resprouts (Table 1). It seems that the closed understory vegetation and compact litter layer at our site (Fig. 1) may have discouraged establishment from 289 seeds. As sprouting hardwoods utilize resources of stem remnants they originated from (Bellingham and Sparrow 2000), they become more resistant to adverse effects of competition from understory vegetation and shading from canopy trees (Kammesheidt 1998). Since the sprouts begin growth immediately after disturbance and their pre-established food resources and root systems allow growth to be concentrated in the stem (Oliver and Larson 1990), we expected multi-stemmed resprouts to grow more rapidly than single trees. However, the opposite was true because single stems displayed higher radial increments than resprouts at early stages of stand development (Table 2; Fig. 4). The number of sprouts and the size of the mother tree may have influenced the growth rate of individual sprouts. Intense coppicing could create dense clones where intense competition among ramets reduced individual growth (Cao and Peters 1998). Another possible explanation for slower growth is rot that may enter the sprout through the heartwood of the parent stump, especially in oaks (Roth and Hepting 1943), restricting radial growth. During the first three decades of stand development Q. mongolica recruits were growing exponentially, and disproportionately more in diameter than older reserved trees left after the last cutting. This substantially decreased size inequality that arose due to age differences. The decreasing size variability means that older and taller individuals had no initial advantage in competition and that negative interference, if there was any, was symmetric or its effect was greater on older than younger trees (Stoll et al. 1994). A tendency to converge to a population of similar-sized individuals is usually ascribed to the lower density of natural establishment leaving much space for individual growth (Stoll et al. 1994), greater relative importance of below-ground competition at nutrient-poor sites (Dolezal et al. 2004), physiological integration in clonal woody plants (Peterson and Jones 1997), or to preference for heightgrowth that is common for shade-intolerant pioneer species. The observed inequality decrease in our stand was likely because sprouting, which in most woody species prevailed over seed regeneration, was not abundant enough to result in dense populations. Lower stem density left much space for individual growth and led to weak competition. Only during the last decade of stand development did size variability increase as larger trees grew more than smaller trees, indicating an increasing asymmetry of competition for light (Weiner 1995). With increasing canopy closure, the taller individuals probably started to disproportionately preempt the incoming solar radiation, thereby suppressing the growth of smaller trees (Hara 1988; Weiner 1990). The influences of asymmetric competition on growth and stem allocation patterns in Q. mongolica at the later stages of stand development were further demonstrated by neighborhood analysis. We found negative effects of neighbors on absolute stem diameter increment from the 1999 to 2004. The neighborhood model fitted best when the effect of smaller neighbors on target stem growth was removed and when the neighboring heterospecific trees of larger sizes were included (see improved regressions incorporating the effects of a taller Fraxinus rhynchophylla, Acer mono, and Tilia amurensis), supporting the assumption of nonequivalent neighbor effects (Zhao et al. 2006). This is rather surprising because most heterospecific neighbors were of smaller sizes. Hence, although Q. mongolica dominated the post-logging forest with almost 90% of total stand basal area, its growth and population size was negatively influenced by other woody species progressively reinvading the stand. The observed growth patterns and analysis of neighbor effects in Q. mongolica support the existing generalization that the mode of plant interactions change during stand development, from early stages of weak competition presumably for soil resources to later stages of asymmetric competition for light (Dolezal et al. 2004). Competition led both to growth suppression and stem exclusion. After the 50 years of stand development, more than 70% of Q. mongolica trees showed growth decline as a result of competitive stress and mortality was about 30%, concentrated in smaller size classes. Thinning occurred in both single- and multi-stemmed trees, suggesting that vegetative origin does not necessarily result in higher survivorship. Thinning is common among clonal trees and shrubs, compared to their herbaceous counterparts, which are of smaller sizes and usually hold ramet density below levels at which interramet interference occurs (Suzuki and Hutchings 1997). Larger sizes of clonal woody species lead to greater size differences and shading among ramets which stimulate intra-clone competition, leading to severe suppression of some ramets by their siblings and greater incidence of thinning (Peterson and Jones 1997). In conclusion, this study combines several recent innovations in the modeling of stand development and plant interactions (Damgaard et al. 2002). The estimations of individual growth parameters with the Richards equation appear to be useful tool in analyzing and interpreting plant interactions when combined with the individual-based spatially explicit neighborhood analysis (Zhao et al. 2006). Our data show that when target trees had more or larger neighbors, their exponential phase of growth was reduced and the maximum size was decreased. In particular, the multi-stemmed Quercus mongolica trees competed asymmetrically, which led to enhanced size differences among resprouts within clones. While the dominant trees within clone were still growing actively in stem diameter, and thus had no apparent asymptote, the suppressed siblings were no longer able to grow exponentially and to achieve stem sizes comparable with single trees of similar age. Hence, our results do not support the hypothesis that resprouts compete less asymmetrically and grow faster due to physical and physiological integration than single trees. We predict that if trees continue to compete asymmetrically the stand will become more uneven-sized over time, and more open as a result of self-thinning, and will become increasingly dominated by woody species like 290 Tilia amurensis, Acer pseudo-sieboldianum, A. mono, and Fraxinus rhynchophylla since they currently occur at higher density in the sapling and shrub stratum. In contrast, Q. mongolica, which are infrequent in the lower tree strata, will decline over time, probably requiring more intense stand disturbance to successfully regenerate. Acknowledgments This study was supported by the Czech Science Foundation (GAČR 206/05/0119), and the Korea Research Foundation Grant (MOEHRD, Basic Research Promotion Fund, KRF-2006-312-C00419). References Abrams MD, Copenheaver CA, Terazawa K, Umeki K, Takiya M, Akashi N (1999) A 370-year dendroecological history of an old-growth Abies–Acer–Quercus forest in Hokkaido, northern Japan. Can J For Res 29:1891–1899 Bellingham PJ, Sparrow AD (2000) Resprouting as a life history strategy in woody plant communities. Oikos 89:409–416 Cao K, Peters R (1998) Structure and stem growth of multi-stemmed trees of Fagus engleriana in China. Plant Ecol 139:211–220 Damgaard C, Weiner J, Nagashima H (2002) Modelling individual growth and competition in plant populations: growth curves of Chenopodium album at two densities. J Ecol 90:666–671 Dolezal J, Ishii H, Vetrova VP, Sumida A, Hara T (2004) Tree growth and competition in a Betula platyphylla–Larix cajanderi post-fire forest in central Kamchatka. Ann Bot 94:333–343 Halpern CB (1988) Early successional pathways and the resistance and resilience of forest communities. Ecology 69:1703–1715 Hara T (1988) Dynamics of size structure in plant populations. Trends Ecol Evol 3:129–133 Hou JH, Mi XC, Liu CR, Ma KP (2006) Tree competition and species coexistence in a Quercus–Betula forest in the Dongling Mountains in northern China. Acta Oecologica 30:117–125 Iida S (2004) Indirect negative influence of dwarf bamboo on survival of Quercus acorn by hoarding behavior of wood mice. For Ecol Manage 202:257–263 Ishikawa Y, Ito K (1989) The regeneration process in a mixed forest in Central Hokkaido, Japan. Vegetation 79:75–84 Kammesheidt L (1998) The role of tree sprouts in the restoration of stand structure and species diversity in tropical moist forest after slash-and-burn agriculture in eastern Paraguay. Plant Ecol 139:155–165 Kato J, Hayashi I (2007) Quantitative analysis of a stand of Pinus densiflora undergoing succession to Quercus mongolica ssp crispula: II. Growth and population dynamics of Q. mongolica ssp. crispula under the P. densiflora canopy. Ecol Res 22:527– 533 Kim JH, Yang HM (1996) The study of stand composition in the natural deciduous forest with Sasa borealis undergrowth. Res Bull Exp For Kangwon Nat Univ 16:26–41 Krestov PV, Song JS, Nakamura Y, Verkholat VP (2006) A phytosociological survey of the deciduous temperate forests of mainland Northeast Asia. Phytocoenologia 36:77–150 Lee DK, Lee KJ, Suh MH, Woo SY, Kim TW (1989) Studies on the development and utilization of Korean Oak Resources (II). Ministry Sci Technol, Seoul, pp 141–208 Newton PF, Jolliffe PA (1998) Temporal size-dependent growth responses within density-stressed black spruce stands: competition processes and bud-worm effects. For Ecol Manage 111:1–13 Nord-Larsen T, Damgaard C, Weiner J (2006) Quantifying sizeasymmetric growth among individual beech trees (Fagus sylvatica). Can J For Res 34:418–425 Oliver CD, Larson BC (1990) Forest stand dynamics. McGrawHill, New York Osawa A (1992) Development of a mixed-conifer forest in Hokkaido, northern Japan, following a catastrophic windstorm: a ‘‘parallel’’ model of plant succession. In: Kelty MJ, Larson BC, Oliver CD (eds) Ecology and silviculture of mixed-species forests. Kluwer, Dordrecht, pp 29–52 Peterson CJ, Jones RH (1997) Clonality in woody plants: a review and comparison with clonal herbs. In: de Kroon H, van Groenendael J (eds) The ecology and evolution of clonal plants. Backhuys Publishers, Leiden, pp 263–289 Phillips PD, Brash TE, Yasman I, Subagyo P, van Gardingen PR (2003) An individual-based spatially explicit tree growth model for forests in East Kalimantan (Indonesian Borneo). Ecol Model 159:1–26 Pinard MA, Barker MG, Tay J (2000) Soil disturbance and postlogging forest recovery on bulldozer paths in Sabah, Malaysia. For Ecol Manage 130:213–225 Richards FJ (1959) A flexible growth function for empirical use. J Exp Bot 10:290–300 Roth ER, Hepting GH (1943) Origin and development of oak stump sprouts as affecting their likelihood to decay. J For 41:27–36 Sano J (1997) Age and size distribution in a long-term forest dynamics. For Ecol Manage 92:39–44 Schweingruber FH (1996) Tree rings and environment: dendroecology. Haupt, Berne Son Y, Park IH, Yi MJ, Jin HO, Kim DY, Kim RH, Hwang JO (2004) Biomass, production and nutrient distribution of a natural oak forest in central Korea. Ecol Res 19:21–28 Stoll P, Weiner J, Schmid B (1994) Growth variation in a naturally established population of Pinus sylvestris. Ecology 75:660–670 Suh MH, Lee DK (1998) Stand structure and regeneration of Quercus mongolica forests in Korea. For Ecol Manage 106:27– 34 Suzuki J, Hutchings MJ (1997) Interactions between shoots in clonal plants and the effects of stored resources on the structure of shoot populations. In: de Kroon H, van Groenendael J (eds) The ecology and evolution of clonal plants. Backhuys Publishers, Leiden, pp 311–329 Takahashi K, Mitsuishi D, Uemura S, Suzuki JI, Hara T (2003) Stand structure and dynamics during a 16-year period in a subboreal conifer-hardwood mixed forest, northern Japan. For Ecol Manage 174:39–50 Tatewaki M (1958) Forest ecology of the islands of the north Pacific Ocean. J Fac Agric Hokkaido Univ 50:371–486 Thomas SC, Weiner J (1989) Including competitive asymmetry in measures of local interference in plant populations. Oecologia 80:349–355 Vesk PA, Westoby M (2004) Sprouting ability across diverse disturbances and vegetation types worldwide. J Ecol 92:310–320 Weiner J (1984) Neighbourhood interference among Pinus rigida individuals. J Ecol 72:183–195 Weiner J (1990) Asymmetric competition in plant populations. Trends Ecol Evol 5:360–364 Weiner J (1995) Following the growth of individuals in crowded plant populations. Trends Ecol Evol 10:389–390 Wu Z (2000) Forests of China, vols I–III. China Forestry Press, Beijing Xiaodong Y, Shugart HH (2005) FAREAST: a forest gap model to simulate dynamics and patterns of eastern Eurasian forests. J Biogeogr 32:1641–1658 Yoshida T, Noguchi M, Akibayashi Y, Noda M, Kadomatsu M, Sasa K (2006) Twenty years of community dynamics in a mixed conifer—broad-leaved forest under a selection system in northern Japan. Can J For Res 36:1363–1375 Zhao D, Borders B, Wilson M, Rathbunc S (2006) Modeling neighborhood effects on the growth and survival of individual trees in a natural temperate species-rich forest. Ecol Model 196:90–102