INTRODUCTION
The Baekdudaegan Mountain Range begins at Mt. Baekdusan and extends to Wonsan, Geumgangsan, Odaesan, Soknisan and Deokyusan Mountains all the way to Mt.
Jirisan, making it Korea's longest mountain range. The concept of 'Baekdudegan', which means a mountain range that stretches out of Baekdusan, is a concept that is along the lines of feng shui and has been formed from the Goryeo Dynasty. However, only the Joseon Dynasty has the concept of mountain ranges and riversides which became common from the aspect of everyday living. A system of mountain maps was established after the mid 18th century in the late Joseon Dynasty (Yang, 1997).
The section between the Jilmoi Wetlands through the Donghae Observatory in Baekdudaegan is a key ecosystem axis located within the Odaesan National Park, which was designated as a national park in 1975. The nearby Jilmoi Wetlands are an international conservation area designated as a Ramsar site together with the Sohwangbyeongsan Wetlands and the Jogaedong Wetlands. Furthermore, the Daegwallyeong Samyang Pasture, which is Korea's only general organic livestock ranch, is spread wide and is located near the Marugeum of Baekdudaegan, and therefore has continuously received human interference.
The importance of managing Baekdudaegan, which is a key ecological axis as Korea's largest mountain range, began in the late 1990s when thoughtless development, such as the construction of large roads, was well underway. By sector, research has been continuously conducted regarding setting the management range, studies on the vegetation structure characters by the Marugeum sections in Baekdudaegan, plans to restore damaged areas, and an environmental analysis of Baekdudaegan. In particular, many studies were conducted regarding the vegetation structure features per section due to the characteristics of the vast 670km section.
When examining preceding studies related to Baekdudaegan for setting the management range Yoo (2002) conducted a study to establish a conceptual management range model of Baekdudaegan, and assessed the ecological, physical and socio-cultural indices to divide the area into three zones a preservation zone, buffer zone, and multi-use zone. Kwon et al. (2004) attempted to set the efficient management range of Baekdudaegan through GIS hydrographic analysis and used the watershed expanding method to analyze changes in land usage depending on the watershed degree, and in result set the 5th-8th degree watershed areas as the management range. Shin (2004) claimed that in border setting for appropriate space, it must be set based on traditional thinking and that the management range should be set considering the natural environment. Shin set the management range up to the third degree water system and claimed that the conservation area must include at least one water system.
In the Baekdudaegan Marugeum vegetation study, there are many studies on investigating the flora and plant community structure by sector. In studies on plant species, research on the status of the plant species distribution by section (Kim et al., 2003; Lim, 2003) and species appearance features by topography (Choung, 1998) have been conducted. In the plant community structure studies where research is most active, while focusing on studies on plant community structure distribution features by sectors (Choo and Kim, 2004; Cho et al., 2005; Choi et al. 2004; Kim and Choo, 2003; Lee et al., 2012), studies on vegetation distribution features according to topographic changes in altitude and slopes (Choi et al., 2003; Hwang et al., 2012; Park and Choi, 2004; Jeong and Oh, 2013) have been carried out. A considerable amount of features and succession tendencies per sector were investigated through preceding studies, and the decline of P. densiflora forests and succession to Q. mongolica forests, as well as the tendency of Q. mongolica forests to succeed to marsh deciduous broad-leaved trees or old stands (Kim et al., 2003; Kim and Baek, 1997; Hwnag et al., 2012).
In damaged area restoration plan research, there were studies on restoring the road slope for the case of Jirisan National Park (Seo et al., 1991) and vegetation restoration (Oh et al., 1998), and vegetation restoration plans for the forest fire area around Samshinbong (Lee et al., 2001). From the aspect of damaged area management and restoration, they are all on Baekdudaegan Marugeum's highland farms, ranches and road slope faces, and it was mentioned that due to the planting of exotic species during the vegetation restoration of these areas, there were problems in terms of natural scenery and nature preservation. Recently, there was also a study on environmental assessments to strengthen management for areas with various land usage and that rated the ecosystem around Baekdudaegan Marugeum (Lee and Lee, 2013).
In the above studies, various research was conducted that proposed restoration plans for damaged areas and to set the management range focusing on studies on plant community features by sections. However, there are many areas that have not been investigated due to the vast range of Baekdudaegan, and it is necessary to establish a concrete management area setting and management plans for sections that requires the harmony of conservation and use.
Therefore, this study aimed at revealing plant community features and the succession trends for the section between the Jilmoi Wetlands through the Donghae Observatory, which was not yet examined in the past, and to ultimately be used as a form of basic data for conservation and management and to establish an ecological management plan for the Daegwallyeong Samyang Ranch that has been managed as an artificial ranch for a long period of time.
MATERIALS AND METHODS
1.Research Target Area
Odaesan National park was designated as the 11th national park of Korea in 1975 and covers a total area of 326 km2. It has a northern temperate climate and its warmth index stands at 65.6(°C·month, as of 2012). The target area of this study is the area near Daegwallyeong Samyan Pasture on the eastern side of Mt. Odaesan and is the approximately 5.7 km section from the Jilmoi Wetlands to the Donghae Observatory. And an altitude distributed between 1,052 m and 1,138m (Figure 1). The investigation area was selected along the Baekdudaegan ridge and takes into consideration vegetation changes where 12 investigation plots (10 m×40 m, 20 m×20 m) were set.
2.Vegetation Structure and the Soil's Physical and Chemical Feature Examination Analysis
1)Vegetation Structure
The vegetation structure investigation was referred to in the method of Monk et al. (1969) and investigated the breast height diameter, tree height, timber height, and crown width for trees with a breast height diameter of more than 2cm for trees and understory, and the tree height, timber height, and crown width for trees with a breast height diameter of less than 2cm or a tree height of less than 2m for shrubs.
In order to compare the relative dominance of tree species based on vegetation investigation, the relative importance percentage (I.P.) (Brower and Zar, 1977) that was integrated and showed in percentage the importance value (I.V.) of Curtis and McIntosh (1951) was used to analyze by tree crown layer. I.P was calculated by (relative density + relative coverage)/2, and tree crown coverage was based on the breast height area. The mean importance percentage (M.I.P.) that granted weight value by tree crown layer considering the size of individual trees was found by {(3×tree I.P.)+(2×understory I.P.)+(1×shrub I.P.)}/6 (Yim et al., 1980; Park et al., 1987). As an indirect expression of tree age and stand movements, the distribution by the breast height diameter class (Harcomb and Marks, 1978) that can estimate the forest succession mode was analyzed. The sample trees were selected considering average dimensions among dominant species per inspection plot, and an increment borer was used 1.2m above the ground to extract a block of wood. In order to analyze the diversity of species by community, the number of species, number of individuals, Shannon's species diversity index (H') (Pielou, 1975), evenness (J'), dominance value (D), and max species diversity (H'max) were analyzed. Community classification used ordination analysis according to DCA (Hill, 1979a) and classification analysis according to TWINSPAN (Hill, 1979b), and by layer, and was classified by taking into consideration the species composition according to I.P. by layer.
2)Physical and Chemical Features of Soil
The soil physical and chemical features analyzed physical features such as soil and chemical features such as pH, organic material content (O.M.) available phosphate (Avail. -P), and exchangeable cation. Soil acidity analysis was carried out using the glass electrode method (1:1), and after agitation for one hour, it was measured repeatedly three times using a pH meter (TOA HM30V). The avail. -P was quantified according to the Bray No. 1 method (Bray and Kurtz, 1945) and among exchangeable cation, Ca2+, Mg2+ was quantified using the EDTA method and K+ using an atomic absorption spectrometry (SP-9, UNCAM Co., U.K.).
RESULTS AND DISCUSSION
1.Vegetation Structure
1)Classification and Ordination Analysis of Investigation Plot
Classification analysis was carried out based on TWINSPAN for 12 investigation plots. In level 1, division 1, it was divided into two groups depending on the appearance of Sorbus commixta (-), and in level 2, division, it was separated into two groups depending on the appearance of F. rhynchophylla (+). Based on the classification analysis results through TWINSPAN, it was divided into a total of three communities. Community I included investigation plots 8 and 11 (2 plots), Community II included investigation plots 1, 2, 3, 4, 5, 6 (6 plots), and Community III included investigation plots 7, 9, 10, 12 (4 plots) (Figure 2). Upon community classification, Community I showed Q. mongolica as the dominant species with differences among accessory species and Community II had a dominant species of Q. mongolica and P. densiflora. For Community Ⅲ, Q. mongolica and T. amurensis were the dominant species with different accessory species.
Among the ordination methods analyzed using classification and mutually supplementing methods, the DCA techniques were used to analyze all investigation plots. The results showed inconsistencies in communities I, II and III, and was different with the classification of TWINSPAN. In plots 3, 4 and 6 where P. densiflora was the dominant species, they were distributed continuously on the right (Figure 3). Due to the character of target areas with similar topographies and relatively complex species composition of communities, it was difficult to distinguish communities clearly through classification and ordination methods, but taking into consideration M.I.P, they were classified as Community I (P. densiflora community), Community II (Q. mongolica community), and Community III (Q. mongolica-T. amurensis community).
2)General Conditions of Investigation Plots
Table 1 shows the general conditions per community that were classified using TWINSPAN analysis and M.I.P. Community I is located between 1,079 and 1,099m above sea level with a slope of between 4-16°, and excludes the flat areas, it faces the southeast. For the canopy layer, the average tree height was between 8-12m, the average breadth height diameter was between 22-40 cm, and it had coverage of between 65-90 %. For the understory layer, the average tree height was between 4-7m, the average breadth height diameter was between 5-8 cm, and coverage was between 10-70 %. For the shrub layer, tree height was under 2.0 m and coverage was between 20-90 %. Community II is distributed between 1,052 and 1,138m above sea level with a slope of between 3-22°, and its main directions are northeast and northwest. For the canopy layer, the average tree height was between 5-9m, with an average breadth height diameter between 10-18 cm, and the coverage was between 60-95 %. For the understory layer, average tree height was between 3-6 m, average breadth height diameter was between 4-6 cm, and coverage was between 5-60 %. For the shrub layer, the tree height was under 2.0 m and coverage was between 15-40 %. Community III was distributed between 1,090 and 1,167 m above sea level with a slope of between 2-12°, and excluding the flat areas, its main direction was southeast. For the canopy layer, average tree height was between 6-11m, average breadth height diameter was between 10-25cm, and coverage was between 80-90 %.
For the understory layer, the average tree height was between 3-6 m, average breadth height diameter was between 4-6 cm, and coverage was between 5-10 %. For the shrub layer, tree height was under 0.6 m and coverage was between 40-70 %.
The research target area is the ridge area of Baekdudaegan and is featured by high altitudes and the low tree height of the canopy layer. The canopy layer has mid to large hardwood trees and is in the understory layer, it is covered with smaller trees, thus having an overall multi-layer structure.
3)Importance Percentage
When examining I.P. analysis results by community (Table 2), Community I (P. densiflora community) included three investigation plots (3, 4, 6) and in the canopy layer, P. densiflora (I.P.: 65.7 %) was dominant along with Q. mongolica (I.P.: 15.6 %). In the understory layer, there were F. rhynchophylla (I.P.: 34.5 %), P. densiflora (I.P.: 20.7 %), Q. mongolica (I.P.: 17.9 %), and S. alnifolia (I.P.: 8.6 %). In the shrub layer, S. incisa (I.P.: 18.7 %) and R. schlippenbachii (I.P.: 12.3 %) had high I.P., and Q. mongolica (I.P.: 9.1 %), A. pseudosieboldianum (I.P.: 8.8 %), and T. regelii (I.P.: 7.1 %) also appeared. P. densiflora was dominant in Community I, in the canopy layer, but in the understory layer, F. rhynchophylla, P. densiflora, and Q. mongolica competed and their succession was judged to be underway. In the shrub layer, S. incisa and R. schlippenbachii showed relative dominance. Lee et al. (2012) reported that the P. densiflora community that was distributed in isolation in the upper part of the mountain along the main ridge of Baekdudaegan from Daetje to Baekbongbyeong had a pure P. densiflora colony for the canopy layer, but there was competition between Q. mongolica and P. densiflora in the understory layer, while in the lower level, Q. mongolica, C. cordata, and F. rhynchophylla were dominant, and it was predicted that it would succeed these tree species in the future. In a study on vegetation structures in the ridge area of Baekdudegan between Pijae and Doraegijae, Oh and Park (2002) classified the Q. mongolica-P. densiflora community as the typical community of low ridge areas. Also, in a study on the features of Q. mongolica in Baekdudaegan according to altitudes between Hyangnobong and Gitdaegibong of Baekdudaegan, Jeong and Oh (2013) said that for the Q. mongolica community in areas 900 to 1,100m above sea level, P. densiflora grew together with Q. mongolica and that in the shrub layer, R. schlippenbachii had a rather large community. It is judged that the reason for the decreasing I.P. for P. densiflora in the understory layer was compared to the canopy layer because a community mixed with Q. mongolica appeared as P. densiflora declined with the passage of time at around 1,000m above sea level. In a study on the vegetation structure of Baekdudaegan between Nogodan and Goribong, Kim and Choo (2003) classified the Q. mongolica-F. rhynchophylla community, thus there is the possibility for succession to the Q. mongolica-F. rhynchophylla community in the future.
Community II (Q. mongolica community) included 6 investigation plots (1, 2, 5, 8, 11, 12). In the canopy layer, Q. mongolica (I.P.: 69.2 %) was the dominant species, and P. densiflora (I.P.: 11.4 %) and S. alnifolia (I.P.: 6.0 %) also appeared. In the understory layer, Q. mongolica (I.P.: 32.5 %), A. pseudo-sieboldianum (I.P.: 20.2 %), T. amurensis (I.P.: 13.2 %), and S. alnifolia (I.P.: 11.4 %) also appeared. In the shrub layer, R. schlippenbachii (I.P.: 26.2 %) had a high dominance value, while S. chinensis for. pilosa (I.P.: 9.0 %), W. florida (I.P.: 7.6 %), R. crataegifolius (I.P.: 8.0 %), and A. pseudo-sieboldianum (I.P.: 6.6 %) also appeared. For Community II, in the canopy layer, Q. mongolica was dominant and in the understory layer, Q. mongolica was dominant, but it was predicted that it would compete with T. amurensis to form the next generation. When considering the prediction by Hwang et al. (2012) that the Q. mongolica community would reduce its proportion through the succsession process and shift to broad-leaved trees such as F. rhynchophylla, T. amurensis, and T. mandshurica through Baekdudaegan natural forests in the Gangwon area community classification and succession trend predictions, there is the possibility to succeed T. amurensis and this thus requires continuous monitoring.
Community III (Q. mongolica-T. amurensis community) included three investigation plots (7, 9, 10). In the canopy layer, Q. mongolica (I.P.: 63.2 %) and T. amurensis (I.P.: 30.0%) were the dominant species, and B. costata (I.P.: 2.6 %) appeared there as well, In the understory layer, A. pseudo-sieboldianum (I.P.: 41.3 %) and Q. mongolica (I.P.: 30.5 % ) were dom inant and com peted with T. amurensis (I.P.: 25.8 %). In the shrub layer, R. schlippenbachii (I.P.: 27.4 %), W. florida (I.P.: 13.2 %), R. crataegifolius (I.P.: 9.5 %), and T. amurensis (I.P.: 4.6 %) also appeared. For Community III, in the canopy layer, Q. mongolica and T. amurensis were dominant, and in the understory layer, A. pseudo-sieboldianum, Q. mongolica, and T. amurensis competed. For the understory, A. palmatum showed relative dominance, but competed with Q. mongolica and T. amurensis in the canopy, and are thus likely to form the next generation. In the shrub layer, R. schlippenbachii and W. florida displayed relative dominance, and R. crataegifolius, T. amurensis, and S. incisa showed even distribution. Lee et al. (2012) reported that in case of the dominant Q. mongolica community in the Baekdudaegan Daetjae-Baekbongnyeong section, if there were no competing tree species to replace the lower layer, then there would be the possibility to develop into old stands (Cho, 1994; Lee et al., 1996; Choi, 2002), but if competing tree species appearred in the lower level, as shown in Cho and Choi's (2002) research results, it may develop from P. densiflora to Q. mongolica and then to broad-leaved trees such as F. rhynchophylla and T. amurensis.
4)Distribution by breast Height Diameter
The method for predicting vegetation succession process through distribution by breast height diameter class was used in past studies (Lee et al., 1990) and distribution results per breast height diameter for each of the three communities are as follows (Table 3). In case of Community I (P. densiflora community), P. densiflora was evenly distributed with a breast height diameter of between 7-52 cm, and there were 17 large trees having a breast height diameter of more than 32 cm. Q. mongolica showed the majority distribution between a breast height diameter of between 2-27 cm, and there were three large trees with a breast height diameter exceeding 32 cm. F. rhynchophylla also had majority distribution in the breast height diameter of between 2-27 cm, and though P. densiflora is currently dominant, it does not appear to have a breast height diameter of less than 7 cm, and the appearance frequency of Q. mongolica and F. rhynchophylla is concentrated in small to mid-size trees, so it is judged that in the future, Q. mongolica and F. rhynchophylla will compete for dominance.
Community II (Q. mongolica community) showed an even distribution of Q. mongolica with a breast height diameter of between 2-37 cm, and the appearance frequency was concentrated in small to mid-size trees with a breast height diameter of between 7-27 cm. For canopy, S. alnifolia, T. amurensis, and F. rhynchophylla showed an even distribution in a breast height diameter of between 2-27 cm, and in the understory S. commixta showed a high appearance frequency at a breast height diameter of between 2-27 cm and an A. pseudo-sieboldianum at a breast height diameter of between 2-17 cm. Q. mongolica had a relatively high appearance frequency in mid-sized trees and it is therefore judged that the Q. mongolica community will be maintained for the time being. Also, S. alnifolia, T. amurensis, and F. rhynchophylla had a high appearance frequency in shrubs and it was judged that its dominance in small to mid-size trees will continue to grow.
In Community III (Q. mongolica-T. amurensis community), Q. mongolica was widely distributed at a breast height diameter of between 2-52 cm, and showed a high appearance frequency in a breast height diameter of between 7-22 cm. There were three large trees with breast height diameter exceeding 32 cm. T. amurensis showed high appearance frequency at a breast height diameter of between 7-22 cm, and there were also three large trees with breast height diameter exceeding 32 cm. In the understory, A. pseudo-sieboldianum showed high appearance frequency at breast height diameter of between 2-17 cm. Currently, the community is dominated by Q. mongolica and T. amurensis, and it is judged that Q. mongolica and T. amurensis will continuously compete for dominance. Choi (2002) stated in the vegetation structure study of the ridge area in the Cheongoksan area of Baekdudaegan that the Q. mongolica community is more likely to grow into Q. mongolica old stands instead of succession. However, in the understory, in the Q. mongolica community where A. pseudo-sieboldianum is dominant, it was judged through a breast height diameter class distribution analysis that the number of T. amurensis is growing among small to mid-size trees, and therefore, the possibility of succession cannot be excluded.
5)Sample Tree Growth
Upon analyzing the tree age and growth of sample trees per community, Community I's (P. densiflora) P. densiflora was from '47 and growth decreased with the passage of time. Q. mongolica was from '51, and growth decreased, but annual ring growth remained consistent. In Community II (Q. mongolica community), Q. mongolica was from '48 and its initial growth was strong, but the annual ring growth continuously decreased, and P. densiflora ('51) showed a sharp drop in annual ring growth. P. densiflora had a competitive relationship with Q. mongolica, and it is therefore presumed that annual ring growth will slow. In the case of Community III (Q. mongolica-T. amurensis community), Q. mongolica ('91) large caliber trees maintained a slightly slower initial growth, and after 1955, with the growth of T. amurensis ('61), it currently has a competing relationship. The growth of Tilia amurensis was stronger than Q. mongolica. It was judged that this was in a development stage from Q. mongolica to T. amurensis (Figure 4). In a vegetation structure study of the ridge area of Cheongoksan in Baekdudaegan, Choi (2002) analyzed the growth of the Q. mongolica community and concluded that an age of the 23cm breast height diameter Q. mongolica's was between '33 and the 34.5cm Q. mongolica was from '79. Considering the fact that in the research target area, the 23.5cm and 24cm breast height diameters of Q. mongolica were '48 and '51, respectively, and the 34.5cm breast height diameter Q. mongolica was from '93, it is evident that the growth of Q. mongolica in this target area has slowed down.
6)Species Diversity
The species diversity analysis results of the three communities in the Jilmoi Wetlands through the Donghae Observatory in Baekdudaegan are as shown in Table 4. Community I (P. densiflora community) showed Shannon species diversity (H') for each 400m2 investigation plot to have measured between 0.9578 and 1.1862, and it was similar or slightly higher than the Baekdudaegan Baekbongnyeong- Daetjae P. densiflora community species diversity of 0.9990 (Lee et al, 2012).
The species diversity per the investigation plot of Community II (Q. mongolica community) was 0.7904- 1.2286, and it was higher than the Baekdudaegan Nogodan- Goribong section Q. mongolica community of 0.9274 (Kim and Choo, 2003) and that it was examined at 500m2 and the Baekbongbyeong-Daetjae section of 0.8046 (Lee et al., 2012), but it was lower than the overall species diversity of the Q. mongolica community in Baekdudaegan in the Gangwon region of between 1.884-2.707 (Hwang et al., 2012). Jeong and Oh (2013) analyzed the average max species diversity of the Q. mongolica community at between 900-1,100m above the sea level of the Daegwallyeong area to be 0.8620~0.9773 in a study on Baekdudaegan Hyangnobong-Gitdaegibong, and this was lower than the target area of this study's max species diversity of between 1.1761-1.13424.
The species diversity per investigation plot of Community III (Q. mongolica-T. amurensis community) was between 0.8701-1.0323, and it was lower than the species diversity of the Baekdudaegan Baekbongnyeon- Daetjae section's Q. mongolica-T. amurensis community of 1.1090 (Lee et al., 2012).
2.Physical and Chemical Features of Soil
The soil of the P. densiflora community (Ⅰ) was loam and sandy loam with a relatively high content of sand (between 50%-70 %), and had pH of between 5.08-5.11, being slightly higher than the 4.80 of non-cultivated mountain soil (Kim et al., 1993). The O.M. was between 6.58-8.51 %, being lower than the 6.40% of mountain soil, and the Avail. -P was between 0.87-2.59 mg/kg, also lower than the average non-cultivated mountain soil of 5.60 mg/kg. Cation content was considerably lower in Ca2+ and Mg2+ contents compared to non-cultivated mountain soil. The Q. mongolica community (Ⅱ) had sandy loam with relatively high sand content and a pH of between 4.78-5.21, being slightly higher than 4.80 of non-cultivated mountain soil. O.M. was between 7.32-10.55 %, slightly higher than the 6.40 % of mountain soil and both Avail. -P (between 0.88-4.15 mg/kg) and cation (Ca2+, Mg2+, K+) content was also considerably lower compared to non-cultivated mountain soil. Q. mongolica-T. amurensis community (Ⅲ) also included sandy loam with relatively high sand content (between 70 %-90 %). Hydrogen ion concentration (pH) was between 4.72-5.03, displaying acidified values compared to a pH 4.80 of non-cultivated mountain soil. O.M. (between 7.49-8.34 %) was higher than average non-cultivated mountain soil, while both Avail. -P (between 1.66-2.29 mg/kg) and cation content (K+, Ca2+, Mg2+) being lower than non-cultivated mountain soil (Table 5).
