Species Interactions in Spruce–Fir Mixed Stands and Implications for Enrichment Planting in the Changbai Mountains, China

Abstract To identify the sapling species that can be planted in spruce–fir forest in the Changbai Mountains, China, we analyzed the associations among saplings and between saplings and adult trees of 4 tree species (spruce [Picea koraiensis Nakai], fir [Abies nephrolepis (Trautv.) Maxim.], Korean pine [Pinus koraiensis Sieb. et Zucc.], and lime [Tilia amurensis Rupr.]), and the spatial distribution of saplings under the adult trees. We observed positive associations between the saplings and adult trees of spruce, fir, Korean pine, and lime. In addition, the numbers of saplings of spruce, fir, Korean pine, and lime with distance from adult trees exhibited a positively skewed distribution. We conclude that saplings of any one or several of these 4 species can be planted in the mixed forests and the optimum distance between saplings and adults trees of the coniferous species is 2–3 m, whereas the corresponding optimal distance from the broadleaved lime is 4–5 m. Our results provide new insights into the development of reforestation techniques in spruce–fir forests in the Changbai Mountains.


Introduction
Spruce-fir forest is one of the major forest types in the Changbai Mountains in northeastern China; it is generally considered to be the climax vegetation in the region. Since the 1970s, this forest type has suffered from intensive selective cutting aimed at timber extraction and hence it has become greatly degraded (Tang et al 2011). The current stand volume of degraded spruce-fir forest is 160 m 3 ha 21 , whereas for primary spruce-fir forest in this region the stand volume is reported to be around 350-400 m 3 ha 21 (Gong 2009). Therefore, restoration of the degraded secondary forest is urgently needed, especially in the context of climate change and biodiversity conservation (Girma et al 2010;Liu et al 2011).
Given previous intensive felling of the valuable climax tree species, such as Abies nephrolepis (Trautv.) Maxim., Picea koraiensis Nakai, Pinus koraiensis Sieb. et Zucc., and Tilia amurensis Rupr., fewer mature trees remain for natural regeneration (Dai et al 2003). Furthermore, following long-term anthropogenic disturbance the forest landscape is fragmented, which hinders seed dispersal of climax tree species from nearby primary forest. Enrichment planting for climax tree species-that is, a technique for promoting artificial regeneration of forests in which seedlings of preferred timber trees are planted in the understory of existing logged-over forests and then given preferential treatment to encourage their growth (Lamprecht 1989)-is therefore essential to facilitate the restoration process, and hence accelerate positive succession.
For sound enrichment planting, species selection and the species/individual spatial distribution for both saplings and adult trees, especially the spatial arrangement in primary forest, need to be identified in advance, in order to guide the restoration process (Zhang et al 2010). Indeed, selection of species can be easily determined by analyzing the importance value of the species in the overstory of remnant old-growth forest. Assessment of the species/individual spatial arrangement can be obtained through analyzing interspecific associations and spatial patterns of main tree species.
Interspecific association, which is one of the most important quantitative and structural characteristics of communities, indicates the spatial distribution relationship and functional dependency between different species (Peng et al 1999;Zhang et al 2003;Lin et al 2005 consideration has been given to the associations among saplings and between saplings and adult trees. Spatial pattern is defined as the horizontal and vertical distribution of individuals, which is determined by a combination of historical and environmental factors (Janzen 1970;Newbery et al 1986;Kenkel 1988;Houle 1994;Camarero et al 2000;Takahashi et al 2001;Li et al 2008). Similarly, numerous studies have investigated the spatial patterns of either adult trees or saplings (Hou and Han 1997;Wang et al 2003;Li et al 2008;Liu et al 2008;Shu et al 2008;Han et al 2009). In contrast, the distribution of saplings in a circular pattern under adult trees is poorly explored.
We conducted the present study in an old-growth spruce-fir forest with the objectives of (1) identification of the species composition of dominant trees for selecting appropriate species in enrichment planting, (2) clarification of the interspecific association and spatial pattern of the main species in order to determine the species/individual arrangement, and (3) offering of detailed recommendations for enrichment planting for restoration of degraded spruce-fir forest.

Study site
The study area was located in the Jingouling Experimental Forest enterprise (43u229N, 130u109E) in the Xueling division of Laoye Ling Mountain in the Changbai Mountains, China. The altitude range of the area is 300-1200 m and the slopes mostly vary between 5u and 25u. The average annual temperature is about 3.9uC and the annual precipitation is 600-700 mm, with most rain falling in July. Frosts occur from mid-September to the end of May, with an average growing season of about 120 days. A gray-brown podzolic soil is present on the low and intermediate mountains in the area and is derived from the parent basalt rock. The study area was mainly covered with degraded secondary forest dominated by A. nephrolepis (Trautv.) Maxim. and P. koraiensis Nakai, in association with other tree species including Pin. koraiensis Sieb. et Zucc., T. amurensis Rupr., Betula costata Trautv., Acer mono Maxim., Betula platyphylla Sukaczev, and Ulmus pumila L. Primary forests, which have been under strict protection and represent the climax vegetation, are also found in this area, although their extent is limited.

Field measurements
In July 2010, sixty-eight 10 3 10-m regeneration plots with 10-m buffer zones were systematically established in primary forest. In each plot, the coordinates (x and y), diameter at breast height (DBH), height (h), and the crown width of adult trees (DBH $ 5 cm) were measured. Similarly, for each sapling (h $ 30 cm and DBH # 5 cm) we measured the coordinates (x and y), ground diameter, h, and crown width. We also surveyed all adult trees and saplings in the buffer zones by recording the species and coordinates (x and y).

Data analysis
Importance value: The importance value is a comprehensive quantitative indicator used to characterize the status and role of each species in the community. The larger the importance value of a tree species, the greater the dominance of the species in the plot. The importance value, relative dominance, abundance, and frequency of each species were calculated as follows: Importance value of the tree layer (IV)( Interspecific association analysis: Interspecific associations comprise positive association, negative association, and no association (Li and Shi 1994). Positive association means that 2 species are likely to be found in the same location and negative association means that species A is less likely to occur in the same community as species B. We used x 2 statistics based on a 2 3 2 contingency table to examine interspecific associations (Zhang 2004). The formula is as follows: where n is the total number of quadrats, a is the number of quadrats with both A and B present, b indicates the number of quadrats with A only, c indicates the number of quadrats with B only, and d indicates the number of quadrats without A and B. When x 2 , 3.841, there is no interspecific association; when 3.841 # x 2 , 6.635, there are certain associations between species; and when x 2 $ 6.635, there are significant associations between species.
When ad . bc, the interspecific association is positive, and when ad , bc, the interspecific association is negative.
Nearest neighbor analysis: Average distances between saplings and adults of different tree species: given saplings of species A and adult trees of species B, the number of saplings of A is N, and the distance from a random sapling of species A to the nearest adult tree of species B is D i . We calculated the average distance between saplings of species A and nearest adult trees of species B with the following formula: We used the formula to determine the degree of dependence of saplings of species A on adult trees of species B. Assuming that a sapling and a nearest adult tree constitute a sample unit, when the sample number exceeds 50 the average distance shows an approximately normal distribution.
Thus, we can obtain the formula: where is the average distance between the saplings of species A and the nearest adult tree of species B, S is the standard variance of the sample distance, and is the expected value of , with the confidence interval Distribution of the number of saplings with the distance between saplings and adult trees: The distribution area of saplings of species A around the nearest adult tree of species B was divided into annular zones (many concentric circles were drawn with the adult tree of species B at the center and the radius equal to a (a 5 1, 2, 3, 4…); the region between 2 adjacent circles was an annular zone), and the number of saplings of species A in each annular zone was counted. The distance, for which the ratio of number of saplings of species A to all saplings is maximal, is the optimal distance for replanting saplings of species A under adult trees of species B.

Importance value of different tree species
Among the tree species studied, the importance value of fir (A. nephrolepis) was largest (27.98) and that of yew (Taxus cuspidata Sieb. et Zucc.) was smallest (0.27) ( Table 1). On the basis of cluster analysis of the importance value (Figure 1), these species were divided into 3 groups: (1) A. nephrolepis, which had the highest importance value; (2) spruce (P. koraiensis), Korean pine (Pin. koraiensis) and lime (T. amurensis), for which the importance values were markedly higher than those of the remaining species; and (3) birch (B. costata), maple (A. mono), white birch (B. platyphylla), elm (U. pumila), larch (Larix gmelinii Rupr.), Cathay poplar (Populus cathayana Rehd.), Manchurian ash (Fraxinus mandshurica Rupr.), yew, and mixed hardwood species (because the importance value of these species were markedly smaller than those of the first 2 groups, these species were classified as a single group). The first 2 groups, for which the total importance values were close to 70, were the dominant species in the mixed stands. As reflected in their lower importance value (,30), the third group was composed of nondominant species.

Interspecific associations
Interspecific associations between saplings: The x 2 test showed that among all species pairs, 40 pairs exhibited a positive association and 15 pairs showed a negative association. A statistically significant association was found between spruce 3 Korean pine and fir 3 Korean pine. The x 2 test also indicated that the positively associated pairs with an association coefficient of .0.7 were fir 3 Korean pine, fir 3 lime, and fir 3 birch ( Figure 2).
Associations between adult trees and saplings: The association between adult trees and saplings was analyzed for the 4 species with the highest importance value. The x 2 test showed that 11 species pairs were positively associated and 5 species pairs were negatively associated. Statistically significant positive associations were observed for (saplings 3 adult trees) fir 3 fir and lime 3 fir. The x 2 analysis also indicated that 5 pairs (saplings 3 adult trees), which comprised fir 3 lime, spruce 3 spruce, spruce 3 lime, Korean pine 3 Korean pine, and lime 3 lime, showed weakly negative associations (Figure 3).

Nearest distance from adult trees to saplings for 4 dominant species
Average distance between adult trees and saplings: The saplings of 4 dominant species (fir, spruce, Korean pine, and lime) were located closest to fir adult trees, followed by Korean pine and spruce, and were located furthest from lime adult trees (Tables 2-5).
Distribution of number of saplings with distance between saplings and adult trees: We analyzed the distribution of the saplings of the 4 dominant species on the basis of the distance between the saplings and adult trees. All skewness values were greater than 0. Therefore, the distribution was positively skewed (Tables 2-5; Figure 4). Thus, the sapling number first increased with the distance from the sapling to the adult tree, and then decreased beyond a certain critical distance. The number of saplings of the 4 species reached maximum values at a distance of 2-3 m from an adult tree for fir, spruce, and Korean pine, and of 4-5 m for lime. The saplings of each species were mainly distributed in the range of 1-5 m from an adult tree for fir, 1-6 m for Korean pine and spruce, and 1-7 m for lime.

Discussion
Interspecific associations Interspecific associations between saplings of different tree species: Three major factors might cause the formation of interspecific associations (Yang et al 2007).
The first factor is the competition type of the species, which can be classified as positive competition or negative competition. Two species are negatively associated if they are competitive with each other and positively associated if they are mutually promotive (Miller 1994;Wootton 1994;Callaway 1995;Hamback et al 2000;Rousset and Lepart 2000;Callaway et al 2002;Palmer et al 2003;Moeller 2004;Graff et al 2007;Brooker et al 2008). Second, one species creates suitable conditions for another species, and thus results in a positive association. For example, pioneer trees can create a regeneration environment suitable for certain shade-tolerant species that have difficulty regenerating in intense light conditions. Third, interaction between 2 species via physical or chemical factors may result in positive or negative associations (Guo et al 1997).
Our results showed that the associations between spruce saplings 3 Korean pine saplings, and between fir saplings 3 Korean pine saplings, were significantly positive ( Figure 2). This might be because the saplings of spruce, Korean pine, and fir have similar ecological preferences and thus can share similar habitats. For example, the saplings have strong shade tolerance and can survive under a canopy in weak light, which might result in coexistence patterns among the species. Although the coexistence might induce competition and increase mortality, overall a positive interspecific association can be formed (Richard and Corine 2011).
Associations between adult trees and saplings: Significant positive associations between saplings and adult trees were observed for the species pairs fir 3 fir and lime 3 fir. The reason might be that saplings of fir and lime require a shady environment for regeneration, which is provided by adult trees of fir and Korean pine. In contrast, most of the other species pairs showed nonsignificant positive associations; however, to some extent this result might indicate that adult trees of most species create suitable conditions for saplings (Figure 3). A significant negative association was observed for a few species pairs (Korean pine adults and saplings; spruce adults and saplings). The negative association might be attributable to human interference and environmental effects. For example, previously Korean pinecones were harvested because of their high economic value, and this led to a lack of saplings, which resulted in a negative association (Liu et al 2004). Consistent with our findings, a negative association between spruce adults and saplings was reported by Prokonev and Zhai (1964) and Chen et al (2005). This phenomenon could be explained by the Janzen-Connell theory, which states that soil pathogens and predators near spruce adult trees could induce high mortality of spruce saplings (Gilbert 2002).

Spatial distributions of saplings of the 4 dominant species around adult trees
Average distances between adult trees and saplings: The ranking of the average distance between saplings and adult trees of the 4 dominant species showed that saplings of the 4 species are most likely to regenerate under the canopy of fir adult trees and are least likely to FIGURE 1 Dendrogram from cluster analysis of the importance value of all overstory tree species in uneven-aged spruce-fir stands. The horizontal axis represents the relative distance between individuals and groups, for which the maximum value is 25. The dendrogram was generated with average-linkage clustering between groups.  establish under lime adult trees, which was consistent with their interspecific associations (Tables 2-5). In contrast, some results were not consistent with the interspecific association analysis. For example, nearest neighbor analysis indicated that fir and Korean pine saplings most favored growing under the canopy of fir adult trees, which is inconsistent with the results of the interspecific association analysis. This finding might be because of the difference in analytical approach of the 2 methods.

FIGURE 3
Half-matrix diagram of x 2 associations between saplings and adult trees of 4 dominant species in uneven-aged spruce-fir stands. Distribution of saplings with the distance between saplings and adult trees: We observed that the number of saplings of all 4 dominant species increased with increment in the distance from a mature tree until a maximum value was reached, and then decreased steadily at greater distances ( Figure 4). This finding can be illustrated to a certain extent by relying on the Janzen-Connell theory, namely: (1) the number of seedlings/seeds is negatively correlated with the distance between the seedlings/seeds and mature trees, and (2) high mortality is more likely to occur in the vicinity of the mature trees because of soil pathogens and predators near the mature trees. In accordance with (1), the number of spruce, fir, Korean pine, and lime seeds/ seedlings decreases with distance from the mature tree, whereas for (2) the optimal distance at which the maximum number of saplings is reached occurs at a certain distance from mature trees rather than in their immediate vicinity (Figure 4). This is consistent with the conclusion of Schupp and Jordano (2011). Besides, the same results were also found between saplings and heterogenic adult trees, which might be explained as follows: (1) high mortality tends to occur near the adult trees due to soil pathogens and predators according to the Janzen-Connell theory; (2) the saplings of these 4 climax tree species could successfully survive under a shady environment; and (3) the crown in the upper layer inhibits the growth of shrubs and herbs, reduces their competitiveness with the saplings, and hence leads to the aggregation of most saplings under the canopy (Jin et al 2005). The maximum number of saplings for coniferous species (fir, spruce, and Korean pine) was observed at a   distance of 2-3 m, whereas for the broadleaved species lime the corresponding distance was 4-5 m (Figure 4). This is consistent with the crown widths of coniferous and broadleaved tree species. Therefore the larger the crown width was, the longer the optimal distance where maximum saplings occurred. Our results are slightly inconsistent with the conclusion of Jin et al (2005), who argued that the maximum number of saplings was mainly distributed in the range of 1.5-3.5 m from the nearest adult tree. The reason for the discrepancy is probably that Jin et al (2005) did not classify the adult tree species into coniferous and broadleaved species when analyzing the spatial distribution of saplings. In future it might be more advisable to conduct this kind of analysis with coniferous and broadleaved species separately.

Conclusions and recommendations
We conclude that: (1) saplings of any one or several species among Korean pine, spruce, fir, and lime are the best choice for enrichment planting in an uneven-aged spruce-fir forest stand; and (2) the optimum distance of saplings from an adult tree is 2-3 m from fir, Korean pine, and spruce adult trees and 4-5 m from lime adult trees. These results provide critical information for definition of the optimum spatial arrangement of saplings of different species for enrichment planting.
In contrast to natural regeneration, which is reliant on seed dispersal mediated by wind and animals, artificial regeneration by enrichment planting is an expensive method, but has an obvious advantage in terms of the flexibility of either species combinations or spatial arrangements. For example, seedlings or saplings of Korean pine, spruce, fir, and lime can be planted at the above-mentioned optimal distance from fir, Korean pine, spruce, and lime adult trees in accordance with the findings in this study. In addition, the number of seedlings and saplings planted can be precisely controlled in accordance with the species composition of the upper canopy, stem density, and site conditions. For example, a greater number of seedlings or saplings could be planted in forest with a low density of trees in the upper canopy and in high-quality sites.
Overall, the spatial distribution of saplings of different species from adult trees and the mechanisms responsible are complex. The potential mechanisms include the ecological preferences of tree species (Xu 2001), intraspecific and interspecific competition (He and Duncan 2000;Druckenbrod et al 2005), microenvironmental factors (Sugita and Tani 2001;Mori and Komiyama 2008), site condition, and chemical interactions between tree species (Silvera et al 2003;Wright et al 2005), and hence further exploration of the mechanisms responsible is needed in order to further refine enrichment planting techniques in such a context.

ACKNOWLEDGMENTS
This study was supported by Forestry Public Welfare Industry Research Item (No. 200804027). We thank all classmates, junior brothers, and junior sisters for their help in data acquirement. We also thank Yan Gao, Qixiang Feng, Zhuohui Wang, Dongning Zhao, Bin Wang, and all Jingouling forest farmers who took part in the field measurement, and the Wangqing Forestry Bureau of Jilin Province in China for their help in data acquirement. Finally, we thank the anonymous reviewers for their useful comments.