Classification of the Vegetation of Pinus densiflora Forests Distributed in Baekdudaegan (From Hyangrobong to Cheonwangbong), South Korea

Abstract

Pinus densiflora and Quercus mongolica are representative forest vegetation communities in Baekdudaegan, South Korea. Recently, signs of deterioration, such as natural succession, disease, and insect pests, have been detected. Therefore, this study aims to classify the vegetation types and elucidate the vegetation structure across the entire South Korean section of the Baekdudaegan, from Hyangrobong to Cheonwangbong, while also proposing strategies for vegetation conservation and management. A vegetation survey was conducted in 341 plots investigated from 2016 to 2020. Cluster analysis revealed nine community types, with a single indicator species, Rhododendron schlippenbachii, in Community 1 (C1); two, Fraxinus sieboldiana and Calamagrostis arundinacea, in C2; six, including Carex humilis var. nana, Polygonatum odoratum var. pluriflorum, and Quercus variabilis, in C3; three, Sasa borealis, Q. mongolica, and Erigeron annuus, in C4; two, Rhododendron mucronulatum and Vaccinium koreanum, in C5; twelve, including Lespedeza maximowiczii, Tripterygium regelii, and Fraxinus rhynchophylla, in C6; two, Toxicodendron trichocarpum and P. densiflora, in C7; twenty, including Acer pseudosieboldianum, Acer pictum var. mono, Staphylea bumalda, and Carex pediformis, in C8; and thirteen species, including Oplismenus undulatifolius, Castanea crenata, and Smilax china, in C9. Our findings highlight the need for management plans that consider each vegetation type’s community structural characteristics.

1. Introduction

Pinus densiflora is distributed on the Korean Peninsula, the Shāndōng and Amnok River areas of China, and southern parts of Hokkai-dō, Japan [1,2]. Its natural habitats are rocky ridges, rock faces, sand beds, valleys, low mountain ranges, hills, and riversides; P. densiflora trees can germinate and grow across South Korea. Additionally, Pinus densiflora is vertically distributed at elevations below 1700 m and thrives in acidic soils as well as dry, sandy soils [3]. P. densiflora forests account for 25% of the total forest area in South Korea. They are the representative plant community constituting the major forest landscape in South Korea, in addition to trees of the genus Quercus [4,5]. Despite their high proportion in the forests of South Korea, owing to climate change, P. densiflora communities are predicted to exhibit a northward transition or reduced scale in high mountain ranges [6,7]. Additionally, there is a need for systematic management to minimize the damage caused by pine gall midge, pine wilt disease, and forest fires [8,9,10,11].

P. densiflora communities and deciduous Quercus trees, including Quercus mongolica communities, are representative forest vegetation types in Baekdudaegan and the ridges of South Korea [12,13,14]. Nonetheless, P. densiflora forests in the Baekdudaegan ridge areas have shown signs of deterioration due to the increasing number of visitors, increased frequency of disease and insect pest incidence, and natural succession [15,16]. Climate change has also highlighted the conservational value of P. densiflora forests in Baekdudaegan, a high mountain region, and its ridges [17].

Baekdudaegan refers to a ‘mountain stream’, continuously connected from Mt. Baekdusan, located in North Korea, to Mt. Jirisan in South Korea, without ever crossing a river even once; this mountain range runs 1400 km within the Korean Peninsula, with approximately 701 km from Hyangrobong to Cheonwangbong in Mt. Jirisan, South Korea, and the remainder through North Korea. Baekdudaegan is also known as the backbone of the forest or natural ecosystem in the peninsula, sheltering a wide biodiversity that includes alpine and subalpine plants, animals, and microorganisms, as well as comprising a natural geographical axis.

In South Korea, Baekdudaegan is divided into five distinct regions [18], namely Seoraksan (from Hyangrobong to Guryongryeong), Taebaeksan (from Guryongryeong to Gitdaebaegibong), Sokrisan (from Gitdaebaegibong to Cheonghwasan), Deokyusan (from Cheonghwasan to Namdeokyusan), and Jirisan (from Namdeokyusan to Cheonwangbong). The ecosystem of Baekdudaegan includes diverse vegetation communities that exist across multiple areas in South Korea due to regional differences and climate and environmental variability [19]. Even the P. densiflora communities that share the same physiognomic forest stand exhibit varying community structures and ecological characteristics, thus necessitating different approaches to design the best strategies and techniques for vegetation conservation and management. Therefore, there is a need to investigate the specific features of the vegetation communities distributed in each region [20].

Previous studies on the P. densiflora communities in Baekdudaegan include the following: a study that monitored the growth of P. densiflora in restoration areas [21], a phenological study [4], a study on the relationship between climate and P. densiflora annual ring growth in some of the ridge areas [19], and a study on the vegetation types and lifeforms of the P. densiflora forests in the ridge areas [15]. However, reports on the vegetation of the P. densiflora communities across various regions connected to Baekdudaegan are non-existent.

Therefore, this study aims to classify the Pinus densiflora communities in the Baekdudaegan region (Hyangrobong–Cheonwangbong section) using a phytosociological community classification approach. Through this classification, the study seeks to elucidate the vegetation structure of each vegetation unit by analyzing importance value, species diversity, and coincidence analysis. Furthermore, based on the characteristics of Pinus densiflora community types, this study aims to develop appropriate management strategies.

2. Materials and Methods

2.1. Overview of the Study Sites

Baekdudaegan stretches along the eastern Korean coastline from Mt. Baekdusan toward the south, before turning westward at Mt. Taebaeksan to connect with Mt. Jirisan in the southern inland region [22]. The length of Baekdudaegan in South Korea was determined to be ~684 km, from Hyangrobong to Cheonwangbong, at Mt. Jirisan; however, it was recalculated as ~701 km with the addition of areas close to the military demarcation line in 2013 [23]. Baekdudaegan, due to its unique combination of boreal and temperate plants, is considered a key biodiversity area in the Korean Peninsula [24]. Various plant species adapted to the unique environment, including Korean endemic, rare, and endangered plants, naturally grow in Baekdudaegan [25]. From Hyangrobong to Cheonwangbong, the minimum and maximum elevations are approximately 200 and 1915 m, respectively, with the largest area between 800 and 1100 m above sea level. Among the forests in the Korea Forest Stock Map, deciduous forests exhibit the largest distribution, followed by oak and mixed forests.

2.2. Vegetation Survey and Analysis

From 2016 to 2020, 2821 vegetation surveys were conducted in plots at 500 m intervals, located 50 m on either side of Marugeum (the ridge line) of Baekdudaegan (from Hyangrobong to Cheonwangbong). The final analysis was based on the data collected from 314 plots (20 × 20 m) of P. densiflora communities (Figure 1).

In the vegetation survey, the Zürich-Montpellier School (Z.-M.) phytosociological method [26,27,28] was used to measure the dominance and sociability of the detected species. Forest layers were categorized as tree, subtree, shrub, and herbaceous. Site conditions, including elevation above sea level (m), aspect (°), slope (°), bare rock (%), and litter depth (cm), were determined at each study site. Species identification was based on the Coloured Flora of Korea [29] and the Field Guide to Trees and Shrubs [30]; in turn, the nomenclature was based on the Korea Plant Names Index [31] and Korea Biodiversity Information System [32].

For community classification, a cluster analysis was performed using Ward’s method to minimize data loss during classification [33], based on the fundamental theory that dispersion across communities should be minimized when analyzing each community [34,35]. To determine the suitable number of communities, indicator species analysis (ISA) was performed according to Dufrêne and Legendre [36], and the Monte Carlo method was used [37] to test the statistical significance of the indicator value of each species. The vegetation classification in this study, conducted using cluster analysis, represents a representative quantitative community analysis method that classifies vegetation types at the community level based on species composition and dominance. Additionally, the naming of vegetation units was determined through indicator species analysis.

The correlations between community types and environmental factors were analyzed using the Coincidence method [38] based on six non-biological environmental factors (topography, elevation, bare rock, litter depth, and aspect).

The importance value (IV) for the occupation rate of the main species in each forest layer based on the species detected at each study site was analyzed according to Curtis and McIntosh [39] by incorporating the relative frequency and relative cover. For species diversity, the Shannon–Wiener method, which is suitable for vegetation data collected from a random sample plot in a forest community spread across a large area, was used to estimate species diversity (H′) [40]; additionally, the maximum species diversity (H′max), evenness (J′), and dominance (1 − J′) indices were also estimated [41].

3. Results and Discussion

3.1. Community Classification

Based on cluster analysis and the vegetation data collected from 341 P. densiflora communities in Baekdudaegan, the selected study sites were classified into 2–10 clusters (Figure 2). The basis on which a suitable number of clusters was determined was the lowest mean p-value of the indicator value, the highest number of species that satisfy the significance level, or both [42]. In this study, the number of communities that resulted in the lowest mean p-value (0.418) was 9, whereas the number of indicator species satisfying the significance level (p < 0.05) was 61; hence, the Baekdudaegan P. densiflora communities were classified into nine community types (Figure 3).

An analysis of the indicator species of each community type classified using cluster analysis showed that Community 1 (C1) comprised a single indicator species, i.e., Rhododendron schlippenbachii. C2 comprised two indicator species, namely, Fraxinus sieboldiana and Calamagrostis arundinacea. In turn, C3 comprised six indicator species: Carex humilis var. nana, Polygonatum odoratum var. pluriflorum, Q. variabilis, Artemisia keiskeana, Q. acutissima, and Fraxinus mandschurica. C4 comprised three indicator species: Sasa borealis, Q. mongolica, and Erigeron annuus; C5 comprised two species, Rhododendron mucronulatum and Vaccinium koreanum. Meanwhile, C6 comprised twelve species, including Lespedeza maximowiczii, Tripterygium regelii, Fraxinus rhynchophylla, and A. stolonifera. C7 comprised two species, Toxicodendron trichocarpum and P. densiflora. In the cases of C6 and C7, the species with high indicator values, T. vernicifluum, L. maximowiczii, and T. regelii, were pioneer species that grew in secondary forests during the stand regeneration period [43]. As for C8, this group comprised twenty indicator species, including Acer pseudosieboldianum, Acer pictum var. mono, Staphylea bumalda, and Carex pediformis. Lastly, C9 comprised thirteen indicator species, including Oplismenus undulatifolius, Castanea crenata, and Smilax china (

Table 1. Species exhibiting significant associations based on indicator species analysis (ISA). * Indicator species are typed in bold letters.

Community Type	Species	Indicator Value	Mean	Standard Deviation	p
	Scientific Name
1	Rhododendron schlippenbachii *	50.6	10.1	2.32	0.0002
2	Fraxinus sieboldiana	38.4	10.4	2.41	0.0002
	Calamagrostis arundinacea	34.3	8	2.92	0.0002
3	Carex humilis var. nana	37	10.7	2.39	0.0002
	Polygonatum odoratum var. pluriflorum	25.4	11.7	3.77	0.0074
	Quercus variabilis	22.2	7.5	2.55	0.0012
	Artemisia keiskeana	21.7	8.2	3.48	0.0056
	Quercus acutissima	11.5	4.1	2.44	0.0206
	Fraxinus mandschurica	8.2	3.2	2.02	0.0382
4	Sasa borealis	83.6	5.2	2.28	0.0002
	Quercus mongolica	16.1	12.8	1.54	0.0356
	Erigeron annuus	9.1	2.9	1.79	0.012
5	Rhododendron mucronulatum	50.2	10.3	2.28	0.0002
	Vaccinium hirtum var. koreanum	13.3	5.6	2.67	0.0192
6	Lespedeza maximowiczii	48.6	8.6	2.79	0.0002
	Tripterygium regelii	35.6	6.6	2.87	0.0002
	Fraxinus rhynchophylla	24.5	10.1	3.39	0.0048
	Artemisia stolonifera	23.1	5	2.63	0.001
	Stephanandra incisa	12.4	5.1	2.52	0.018
	Disporum viridescens	10.5	4.7	2.79	0.044
	Syneilesis palmata	9.5	4.6	2.56	0.0488
	Eupatorium japonicum	9.4	3.1	1.93	0.0154
	Asarum sieboldii	8.6	4	2.07	0.0346
	Carex lasiolepis	8.2	2.9	1.77	0.0256
	Euonymus oxyphyllus	8.0	3.4	2.14	0.0428
	Valeriana fauriei	6.6	2.9	1.77	0.0434
7	Toxicodendron trichocarpum	15.2	9.8	2.55	0.0426
	Pinus densiflora	14.2	12.3	0.46	0.0012
8	Acer pseudosieboldianum	59.9	9.1	2.72	0.0002
	Acer pictum var. mono	21.7	4.7	2.77	0.001
	Staphylea bumalda	20	3.3	2.33	0.0006
	Carex pediformis	19.8	3.7	2.3	0.0002
	Callicarpa japonica	18.8	4.2	2.51	0.0014
	Smilax nipponica	16.4	8.9	3.49	0.0406
	Viola mandshurica	16.1	4.4	2.52	0.0034
	Styrax obassia	14.4	7.5	3.2	0.0436
	Securinega suffruticosa	13.3	3.3	2.03	0.0026
	Weigela subsessilis	12.5	4.6	2.38	0.0132
	Abies holophylla	11.9	3.5	2.38	0.0162
	Corylus heterophylla	11.8	4.9	2.75	0.0308
	Callicarpa dichotoma	11.6	3	1.82	0.0074
	Ilex macropoda	9.2	2.9	1.84	0.0172
	Saussurea seoulensis	7.4	3.3	2.04	0.0290
	Crepidiastrum chelidoniifolium	7.1	2.9	1.92	0.0212
	Cryptotaenia japonica	7.1	2.6	1.75	0.0404
	Acer ukurunduense	7.1	2.6	1.77	0.0390
	Galium odoratum	7.1	2.6	1.79	0.0442
	Matteuccia struthiopteris	7.1	2.6	1.79	0.0442
9	Oplismenus undulatifolius	79.3	8	2.64	0.0002
	Castanea crenata	44.4	6.4	2.66	0.0002
	Smilax china	25	6.7	2.87	0.0006
	Zanthoxylum schinifolium	19.1	8.1	2.8	0.0052
	Cornus controversa	19	5.7	2.6	0.0024
	Lindera erythrocarpa	15.2	6	2.58	0.0110
	Phryma leptostachya var. oblongifolia	12.7	5.1	2.92	0.0244
	Pueraria lobata	11.7	4	2.4	0.0154
	Osmunda japonica	11.5	3.6	2.28	0.0154
	Morus bombycis	9	4.3	2.29	0.0462
	Akebia quinata	8.6	3.5	2.29	0.0352
	Carpinus laxiflora	8.4	3.8	2.25	0.0422
	Impatiens textori	8.4	3.5	2.15	0.0380

). Furthermore, consistent with the indicator species detected in this study, F. sieboldiana, L. maximowiczii, T. trichocarpum, A. keiskeana, and Morus bombycis were determined to be the characteristic species in the P. densiflora communities [44].

3.2. Site Conditions per Community Type

The correlations between community types and non-biological environmental factors were analyzed using the Coincidence method (Figure 4). Regarding topography, most communities were located on slopes and ridges, presumably because our study focused on the Baekdudaegan ridge areas. The topography varied from valleys to mountain tops in C7 and C8, whereas C9 was mainly located in the low- to mid-slope areas. Community types C1–C6 were approximately 600–1000 m above sea level, whereas C7 and C8 were mostly at 200–1000 m above sea level and C9 was at 400–800 m above sea level. Slopes were observed with varying distributions across all communities except C6, C8, and C9, which were mainly found below 25°. Meanwhile, bare rock areas accounted for 0%–10% in most communities, with C2 showing the highest (70%) percentage. C9 was located at the lowest latitude, occupying the most area in the Bokseongyijae–Seongsamhae part of the Jirisan region; in turn, C3 was found at an intermediate latitude, from 36°00′ to 36°50′. Concerning aspects, C3, C4, C6, and C7 were determined to be located on south-facing slopes, coinciding with the optimal distribution area of pine trees.

3.3. Layer Structures per Community Type

The IV values of the nine communities classified based on cluster analysis were estimated (

Table 2. Importance values of major species in community types. Abbreviations: T stands for tree layer; ST stands for subtree layer; S stands for shrub layer; and H stands for herb layers.

Layer	Species	Community Type
		1	2	3	4	5	6	7	8	9
T	Pinus densiflora	72.3	75.3	55.7	67.2	71.2	59.6	68.8	57.2	71.7
	Quercus mongolica	17.7	18.7	10.9	20.1	15.5	14.3	11.0	17.8	0.8
	Quercus variabilis	4.3	-	17.5	4.4	5.3	8.0	3.8	5.6	0.8
	Castanea crenata	-	-	-	-	-	-	-	-	11.9
	Others	5.7	6.0	15.9	8.3	8.0	18.1	16.4	19.4	14.8
	Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
ST	Quercus mongolica	43.4	41.4	35.6	39.7	37.5	26.2	17.1	11.8	13.3
	Acer pseudosieboldianum	4.2	3.1	-	11.0	8.4	10.4	-	38.7	-
	Pinus densiflora	10.8	11.2	14.0	3.3	10.8	7.1	7.3	-	10.3
	Fraxinus sieboldiana	15.3	22.5	-	13.6	8.1	-	7.4	-	-
	Fraxinus rhynchophylla	-	-	-	7.0	-	16.1	5.8	3.9	-
	Quercus variabilis	3.0	-	12.1	3.4	2.5	4.2	3.8	-	5.4
	Cornus controversa	-	-	-	-	-	3.2	4.9	-	13.4
	Castanea crenata	-	-	-	-	-	-	-	-	22.2
	Styrax obassia	-	-	-	-	-	-	8.0	5.2	-
	Others	23.3	21.8	38.3	22.0	32.7	32.8	45.7	40.4	35.4
	Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
S	Quercus mongolica	7.3	9.3	12.1	9.9	5.3	3.5	3.4	3.2	3.7
	Rhododendron mucronulatum	12.0	9.1	5.7	4.5	35.7	7.9	-	-	12.6
	Rhododendron schlippenbachii	45.3	20.3	7.7	16.8	8.4	-	-	4.8	-
	Sasa borealis	-	-	-	14.3	-	-	-	3.5	-
	Toxicodendron trichocarpum	-	2.7	-	-	-	-	6.8	3.7	-
	Zanthoxylum schinifolium	-	-	11.2	-	-	-	-	-	11.8
	Lindera obtusiloba	-	-	-	-	5.4	-	8.5	8.8	5.0
	Lespedeza maximowiczii	-	-	-	-	-	22.7	-	-	-
	Fraxinus sieboldiana	10.4	21.2		8.5	4.8		5.6
	Stephanandra incisa	-	-	-	-	-	-	-	5.7	-
	Others	25.0	37.4	63.3	46.0	40.4	65.9	75.7	70.3	66.9
	Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
H	Vaccinium hirtum var. koreanum	-	-	-	-	7.5	-	-	-	-
	Spodiopogon sibiricus	-	-	-	-	-	-	4.9	-	-
	Smilax china	-	-	-	-	-	-	-	-	4.5
	Sasa borealis	-	-	-	44.4	-	-	-	11.3	-
	Rubus crataegifolius	-	-	-	-	-	-	-	-	3.3
	Rhododendron schlippenbachii	8.4	3.2	1.2	1.4	2.3	-	-	1.1	-
	Rhododendron mucronulatum	1.9	1.1	1.2	0.5	8.0	-	2.7	-	0.6
	Polygonatum odoratum var. pluriflorum	-	-	-	6.5	-	-	-	-	-
	Oplismenus undulatifolius	-	-	-	-	-	-	4.7	-	32.6
	Carex siderosticta	5.3	-	-	-	-	5.4	-	7.2	-
	Carex pediformis	-	-	-	-	-	-	-	10.9	-
	Carex humilis var. nana	14.1	8.4	28.5	-	-	-	14.1	-	-
	Calamagrostis arundinacea	-	15.4	-	-	-	-	-	-	-
	Lespedeza maximowiczii	-	-	-	-	-	7.6	-	-	-
	Tripterygium regelii	-	-	-	2.2	-	4.4	-	-	-
	Artemisia keiskeana	-	-	3.7	-	-	-	-	-	-
	Others	70.3	71.9	65.4	47.2	82.2	82.6	73.6	69.5	59.0
	Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0

). Concerning the C1 vegetation type, the tree layer comprised P. densiflora (IV = 72.3) and Q. mongolica (17.7); the understory layer comprised Q. mongolica (43.4), F. sieboldiana (15.3), and P. densiflora (10.8); the shrub layer comprised R. schlippenbachii (45.3), F. sieboldiana (10.4), and Q. mongolica (7.3); and the herbaceous layer comprised C. humilis var. nana (14.1), R. schlippenbachii (8.4), and Carex siderosticta (5.3). As for the C2 vegetation type, the tree layer comprised P. densiflora (75.3) and Q. mongolica (18.7); the understory layer comprised Q. mongolica (41.4), F. sieboldiana (22.5), and P. densiflora (11.2); the shrub layer comprised R. schlippenbachii (21.2), R. mucronulatum (20.3), and Q. mongolica (9.3); and the herbaceous layer comprised C. arundinacea (15.4), C. humilis var. nana (8.4), and R. schlippenbachii (5.5). As for the C3 vegetation type, the tree layer comprised P. densiflora (55.7), Q. variabilis (17.5), and Q. mongolica (10.9); the understory layer comprised Q. mongolica (35.6), P. densiflora (14.0), and Q. variabilis (12.1); the shrub layer comprised Q. mongolica (12.1), Zanthoxylum schinifolium (11.2), and R. schlippenbachii (7.7); and the herbaceous layer comprised C. humilis var. nana (28.5), P. odoratum var. pluriflorum (6.5), and A. keiskeana (3.8). Pine and oak trees competed as high-light-demanding trees [45]. Hence, it was hypothesized that C1 and C2 would succeed a Q. mongolica community and C3 a Q. variabilis community, which is consistent with the findings of Cho and Lee [46], who studied the Hanuiryeong–Daetjae section of Baekdudaegan, and Lee et al. [47], who studied the Daetjae–Baekbongryeong section of Baekdudaegan, and reported that P. densiflora communities succeeded to deciduous Quercus communities, especially Q. mongolica communities. Regarding the C4 vegetation type, the tree layer comprised P. densiflora (62.7) and Q. mongolica (20.1); the understory layer comprised Q. mongolica (39.7), F. sieboldiana (13.6), and A. pseudosieboldianum (11.0); the shrub layer comprised S. borealis (14.3), Q. mongolica (9.9), and F. sieboldiana (8.5); and the herbaceous layer comprised S. borealis (44.4), with the highest IV value. S. borealis is characterized by mass flowering and death [48], and once the species has settled in a low forest layer, it dominates the area due to its wide distribution and long-term residence [49,50]. Tree species with low shade tolerance were negatively correlated with S. borealis, which allegedly promotes the emergence of late-succession species rather than suppressing early-succession species. These observations suggest the need to monitor changes in vegetation [49]. As for the C5 vegetation type, the tree layer comprised P. densiflora (71.2) and Q. mongolica (15.5); the understory layer comprised Q. mongolica (37.5), P. densiflora (10.8), and A. pseudosieboldianum (8.4); the shrub layer comprised R. mucronulatum (35.7), R. schlippenbachii (8.4), and Lindera obtusiloba (5.4); and the herbaceous layer comprised R. mucronulatum (8.0), V. koreanum (7.5), and Q. mongolica (3.0). R. mucronulatum in the shrub layer tends to emerge more often in lower elevation areas compared with R. schlippenbachii [13], which could account for the low latitude of R. schlippenbachii in C5 when compared with C1 and C2, with higher IV values of R. schlippenbachii. Regarding C6 vegetation, the tree layer comprised P. densiflora (59.6) and Q. mongolica (14.3); the understory layer comprised Q. mongolica (26.2), Fraxinus rhynchophylla (16.1), and A. pseudosieboldianum (10.4); the shrub layer comprised Lespedeza maximowiczii (22.7) and Tripterygium regelii (11.1); and the herbaceous layer comprised L. maximowiczii (7.6), Carex siderosticta (5.4), and T. regelii (4.4). Concerning the C7 vegetation type, the tree layer comprised P. densiflora (68.8) and Q. mongolica (11.0); the understory layer comprised Q. mongolica (17.1), Styrax obassia (8.0), and F. sieboldiana (7.4); the shrub layer comprised L. obtusiloba (8.5), R. mucronulatum (7.9), and T. trichocarpum (6.8); and the herbaceous layer comprised C. humilis var. nana (14.1), Spodiopogon sibiricus (4.9), and O. undulatifolius (4.7). Regarding the C8 vegetation type, the tree layer comprised P. densiflora (57.2) and Q. mongolica (17.8); the understory layer comprised A. pseudosieboldianum (38.7), Q. mongolica (11.8), and S. obassia (5.2); the shrub layer comprised A. pseudosieboldianum (21.6), L. obtusiloba (8.8), and Stephanandra incisa (5.7); and the herbaceous layer comprised S. borealis (11.3), C. pediformis (10.9), and C. siderosticta (7.2). A. pseudosieboldianum is a tree species found in fertile and moist soils, such as foothills and valleys [51]. This study observed this species in low- to mid-slope areas. As for the C9 vegetation type, the tree layer comprised P. densiflora (71.7) and C. crenata (11.9); the understory layer comprised C. crenata (22.2) and Cornus controversa (13.4); the shrub layer comprised R. mucronulatum (12.6), Z. schinifolium (11.8), and L. maximowiczii (7.1); and the herbaceous layer comprised O. undulatifolius (32.6), Smilax china (4.5), and Rubus crataegifolius (3.3).

3.4. Species Diversity

Table 3. The Shannon–Wiener diversity index according to community type.

Community Type	H′	J′	1 − J′	H′max
1	1.9156	0.6720	0.3280	2.8691
2	1.8630	0.6584	0.3416	2.8291
3	1.8637	0.6801	0.3199	2.7549
4	1.7703	0.6075	0.3925	2.9132
5	1.7089	0.6005	0.3995	2.8502
6	1.7033	0.6434	0.3566	2.6377
7	1.7656	0.6387	0.3613	2.7583
8	1.6895	0.6541	0.3459	2.5961
9	1.9021	0.6609	0.3391	2.8692
Average	1.8118	0.6465	0.3535	2.8035

Abbreviations: H stands for species diversity; J′ stands for evenness; 1 − J′ stands for dominance; H’max stands for maximum species diversity.

shows the Shannon diversity (H′), H′max, evenness (J′), and dominance (1 − J′) indices for the nine P. densiflora communities in Baekdudaegan. The highest species diversity was observed in C1 (H′ = 1.9156) and the lowest in C8 (H′ = 2.5961). Evenness was highest in C5 (J′ = 0.3995) and lowest in C3 (J′ = 0.3199). Species diversity for the P. densiflora communities at the key sites in South Korea (Boeun-gun of Chungcheongbuk-do, Inje-gun of Gangwon-do, Wanju-gun of Jeollabuk-do, Yeongcheon-si and Uljin-gun of Gyeongsangbuk-do, and Ahnmyeondo of Chungcheongnam-do) ranged between 0.9171 and 1.5016, with H′max values ranging from 1.6435 to 2.1523 [52]. The species diversity of the dominant P. densiflora community in the ridges of the Nakdong River ranged from 0.6693 to 0.9772, with a H’max index of 0.8078–1.0942 [13]. These findings indicated that the Baekdudaegan P. densiflora communities had high H′ and H′max values.

4. Conclusions

Pinus densiflora forests with the same physiognomy show varying community structures and ecological characteristics; therefore, different conservation methods and techniques for maintaining and managing these vegetation communities are required [15]. Notably, as Baekdudaegan extends over multiple areas across South Korea with regional differences in climate and environmental conditions, a study on the unique features of P. densiflora communities is required. Therefore, the purpose of this study was to classify the community types and analyze the stand structures of P. densiflora communities in Baekdudaegan (from Hyangrobong to Cheonwangbong) to identify appropriate management strategies that consider the characteristics of each P. densiflora community.

Regarding management plans that consider site conditions, layer structures, and species diversity in vegetation types, C1, C2, and C5 require close monitoring of the changes in vegetation in the ridges and are predicted to succeed into deciduous forests in the future because these communities were located on the ridges and slopes at 700 m above sea level and in dry areas with abundant light, with species such as F. sieboldiana, R. schlippenbachii, and R. mucronulatum. These communities were identified as typical P. densiflora community types, with the tree layer dominated by P. densiflora; the shrub and herbaceous layers of C4 were dominated by S. borealis and required monitoring vegetation changes. Meanwhile, the tree layers of C3 and C6 contained Larix kaempferi and had been managed, but they required an ecological restoration plan based on secondary forest status, considering stand regeneration. In turn, C7 and C8 community types were distributed at relatively low altitude areas (i.e., 670 m above sea level, on average). They were dominated by species such as F. rhynchophylla, Q. serrata, L. erythrocarpa, and A. pseudosieboldianum, which inhabit areas with moist soil. These communities required vegetation monitoring in the valleys and were in a state of succession to deciduous forests. Lastly, C9 was found in the Jirisan region (Yukshipryeong–Seongshimjae) and requires region-specific management plans.

Baekdudaegan has boreal and temperate plants growing together; thus, it is a key biodiversity site on the Korean Peninsula [19], with P. densiflora communities being the area’s representative forest and vegetation type. Therefore, specific management plans should be developed for each community in the area, and such plans should consider the ecological features of the habitats, species composition, and areas of disturbance and distribution to ensure effective ecological conservation of resident P. densiflora communities. Furthermore, the North Korean portion of the Baekdudaegan should be considered and studied shortly.