INTRODUCTION

L. wilfordii is known as a specialist plant mainly distributed in high-latitude peatlands(Tamura et al., 2016). Peatlands are the result of dead plants being decomposed depending on topographical and climatic factors(Page et al., 2006;Tarnocai and Stolbovoy, 2006;Kim et al., 2015). Therefore, they are widely distributed in high-latitude regions, whereas they are rarely found in the mid-latitude regions with temperate climates(Lappalainen, 1996; Choung et al., 2009). In the mid-latitude regions, peatlands were formed mainly in high altitude regions, especially alpine regions, which were an shelter for a rear-edge population of northern plants retreating after the ice age(Hampe and Petit, 2005). These remaining peatlands provide an important site for the maintenance and conservation of plant diversity(Lienert et al., 2002;Joosten and Clarke, 2002;Bragg and Lindsay, 2003;Kim et al., 2015).

The specialist plants have limited distribution under specific environmental conditions(Dupré and Ehrlén, 2002;Işık, 2011). The mid-latitude peatlands provide highly isolated sites, like an islands floating on land(MacArthur and Wilson, 1967;Lienert, 2004). A highly isolated population cannot expect the immigrations of individuals from the neighboring population(MacArthur and Wilson, 1967). As a result, it has a high risk of extinction(MacArthur and Wilson, 1967;Dupré and Ehrlén, 2002). Although such species exhibit low genetic diversity and high genetic uniqueness, specialist plant distributed in highly isolated habitats have maintained their populations for long periods of time(Lammi et al., 1999;Schmidt and Jensen, 2000). There is a limit to assessing extinction risk based solely on the low genetic diversity of the plant population(Wiens, 1997;Oostermeijer et al., 2003). Therefore, for the conservation of specialist plant distributed in isolated sites, information on the ecological process of the population dynamics, which is represented by interactions of species attributes and environmental factors, including the lifespan and life history of the species with genetic characteristics, must always be included.

The extent of a sessile plant population is absolutely determined by the size and quality of the habitat(Dupré and Ehrlén, 2002). In particular, since specialist plant species are highly dependent on their habitat, the size and quality of the habitat are related to the extent of the population(Root, 1998). The habitat environment changes over time(Whittaker, 1953;Connell and Slatyer, 1977). If a site suitable for the growth of a specific plant is not maintained in the ecological process of succession, it will inevitably become extinct(Lerouxa et al., 2007). Consequently, the persistence of isolated plant populations can be described in terms of non-equilibrium co-existence mechanisms within the context of spatio-temporal series(Pickett, 1980). In other words, the natural delays or reversals of the successional process at a certain level can be a major factor(Denslow, 1980;Hobbs and Huenneke, 1992). Hence we decided to introduce a minimum dynamic area into the establishing of protected areas, including the minimal ecological processes required for the continuation of the plant population to be conserved(Pickett, 1980;Damschen et al., 2006). Therefore, it is essential to understand the internal dynamics of extinction and re-establishment on a regional scale involved in the sustainability of isolated plant populations(Pickett and Thompson, 1978).

In Korea, L. wilfordii has been reported in two distribution areas(Korea National Arboretum(KNA), 2012; National Institute of Biological Resources(NIBR), 2012). These distribution areas are about 80 km apart. Recently, genetic studies of L. wilfordii distributed in Korea have shown low genetic diversity(Kim et al., 2019). As previously proposed, since assessing the extinction risk and sustainability of certain plants that need conservation by using only by genetic characteristics has limits, in this study we attempted to discuss the following questions:

MATERIALS AND METHODS

1. Species

L. wilfordii is a perennial plant belonging to the Caryophyllaceae(Figure 1A;Tamura et al., 2016). Plants glabrous or pilose with intermixed sparse multicellular eglandular hairs(Flora of China(FOC), 2020). The leaves are opposite, have no petioles, ovate-lanceolate or lanceolate, and margin thickly ciliate. Stems simple or branched above, has nodes with fracturable. It usually grows straight and reaches 50 to 80 cm in height(FOC, 2020). Numerous filamentous roots are developed at the main and lateral roots. A well-grown individual may form a large clusters with a plurality of buds from a single individual(Figure 1A). The flowers are bright red, have inflorescences at the tip of the stalk, and bloom from mid-July to late August(Figure 1B). The fruits grow into a long oval shape and become brown after about forty to fifty days after the flowers fall(Figure 1C). The mature capsules split open, and seeds are scattered by wind or external stimuli(Figure 1C; Author's observation). The seeds are dark brown, and each capsule usually contains forty to sixty seeds. The size of the seeds is 1.5 mm × 2 mm, and has hook-shaped protrusions on the surface(Figure 1D;Ryu et al., 2017). It shares a reproduction by seed propagation and a vegetative propagation that produce multiple buds by extending a short stalk(Bae et al., 2014;Tamura et al., 2016). The scattered seeds usually germinate in the next spring of the following year. The seedlings differ depending on the locational conditions, and in an environment suitable for growth, they may bloom from early September to late September of the year when they germinate(Figure 1E~1G).

The species is distributed in the Maritime Province of Russia, Jilin Province of China, Korea, and Hokkaido and Honshu of Japan(Tamura et al., 2016;Kim et al., 2019; FOC, 2020). In Korea, it is distributed only in the YWPA in Inje-gun, Gangwon Province, and in the wetlands in Daegwallyeong-myeon(DWL), Pyeongchang-gun, Gangwon Province(KNA, 2012;NIBR, 2012;Kim et al., 2015). Using the evaluation criteria of the IUCN Red List, both the National Institute of Biological Resources(NIBR) of the Ministry of Environment(MOE) and the Korea National Arboretum(KNA) of the Korea Forest Service evaluated it as an endangered(EN)(KNA, 2012;NIBR, 2012). The MOE has designated it as an endangered wild plant of class II since 2012 and is protecting it(MOE, 2012).

2. Study sites

In Korea, a total of six populations were investigated(Figure 2; Table 1). In the YWPA in Inje-gun, Gangwon Province, its distribution is divided into three major populations: Aegi yongneup, Jakeun yongneup, and Keun yongneup(Figure 2A; Table 1). In DWL, Pyeongchang-gun, Gangwon Province, three populations were investigated(Figure 2B; Table 1). A total of 3,838 individuals were observed throughout the distribution. Among them, 3,040 individuals were distributed in the YWPA, and 798 individuals in the DWL(Table 1). Each population differs in size, from a minimum of 10 to a maximum of 1,500 individuals(Table 1). This study was conducted on eight plots of Aegi yongneup, Jakeun yongneup, and Keun yongneup in the YWPA, and seven plots in the DWL(Figure 2; Table 1).

3. Methods

1) Comparison of the Vegetation Environment in the Habitats Between Two Areas

To compare the vegetation environment between the distribution areas, we installed a plot with the size of 4 ㎡(2 m × 2 m) by selecting the highest density sites of L. wilfordii in the YWPA and the distribution area in the DWL. We recorded the distributed plants using a phytosociological survey method(Braun-Blanquet, 1964). However, the coverage of the observed plants was not converted into coverage class, and the percentage was recorded as it was. We calculated the importance values(IV) by the sum of the relative frequency(RF) and the relative coverage(RC) (Curtis and Mclntosh, 1951). We arranged the plants distributed in a total of 21 plots by layers and importance values, and created a species composition table by modifying the tabulation method(Ellenberg, 1956). The species that not only has it observed more than ten times and had an importance value of 4.0 or more but also appeared together in two areas were classified as a Constant species. On the other hand, species that observed in more than 40% of the plots and had an importance value of 2.0 or more were classified into a Differential species.

2) Distribution Characteristics in the YWPA

The YWPA, which has a stable population structure, well-maintained naturalness, and relatively little human disturbance, was selected as the sites for the distribution characteristics survey. Plots with the size of 1 ㎡(1 m x 1 m) in Aegi yongneup, Jakeun yongneup, and Keun yongneup where they showed the highest density of L. wilfordii, were selected as the target plots. To compare the vegetation characteristics of the distribution sites of L. wilfordii, we set up continuous plots of 1 ㎡(1 m x 1 m) in a Keun yongneup including high moor, intermediate moor, and low moor(Figure 3). The continuous plots were set up in the northeast(Figure 3A; 15m) and southwest areas(Figure 3C; 40m), and in the central regions(Figure 3B; 117m) where the original shape of the peatlands has been artificially damaged by the installation of a skating rink in 1977. Each plant that observed in continuous plots was recorded using a phytosociological vegetation survey method (Braun-Blanquet, 1964). At this time, the coverage of the observed species was not converted into coverage class and was recorded using a percentage. We used the results of the survey to calculate the dominance index(C′; Simpson 1949), evenness index (J′; Pielo, 1969), speciesdiversity index(H′; shannon-weaver diversity index), and maximum species-diversity index(H′max).

The plant community representing the peatlands of yongneup is Carex thunbergii var. appendiculata Trautv.(Kim et al., 2015). In the adaptation strategy corresponding to the moisture condition of the habitats, C. thunbergii var. appendiculata grows sideways in the high moor, but in a space where periodic flooding occurs it grows upward to form a tussock (Nishikawa, 1990). In the preliminary survey of the YWPA for this study, L. wilfordii was mainly distributed on the tussocks formed by C. thunbergii var. appendiculata. Accordingly, we randomly selected 110 tussocks that were occupied by L. wilfordii in the largest Aegi yongneup population. The height and the width in the longest direction of the selected tussocks were measured and used to assess the characteristics of the tussocks in which L. wilfordii is distributed. In addition, to investigate the spatial distribution characteristics of adjacent tussocks, we measured the distances between the tussocks occupied by L. wilfordii and the surrounding tussocks were measured. At this time, the distance measurement was based on the distance between the centers of the tussocks. Additionally, we randomly selected 69 tussocks that were unoccupied by L. wilfordii in the Keun yongneup, and measured the height and width in the same way as above. Since 1977, when artificial damage was caused by the installation of the skating rink, we used a 15-meter line that has been installed at the spot where restoration of the tussocks is underway to measure the height of the 24 tussocks bordering the line to estimate their age(Kang and Kwak, 2000).

In the preliminary survey for this study, we observed disturbances caused by the feeding activities of wild boars periodically in the habitat of L. wilfordii. To account for the affect that such feeding activity caused to the tussocks occupied by L. wilfordii, we installed two plots of 4 ㎡ around the tussocks occupied by L. wilfordii in the population of the Jakeun yongneup where disturbances were observed over the largest area in 2017 in order to observe subsequent changes. Throughout 2018, 2019, and 2020, we observed and photographed the re-establishment, growth, and flowering of the seedlings of L. wilfordii. We used the observed results to evaluate the internal dynamics involved in population persistence of L. wilfordii.

3) Evaluation of Growth Characteristics and Self-compatibility

To investigate germination and growth characteristics, we carried out an experiment from January 2017 to September 2018 at the Natural Environment Research Park in Gangwon Province, an Ex-Situ Conservation Institution for Endangered Wildlife of the Ministry of Environment. The seeds used in the experiment were harvested from plants artificially grown in the Natural Environment Research Park in Gangwon Province in September 2016, and then refrigerated(2-8℃) that Ex-Situ Conservation Institution was divided into two conditions: one was when seeds were scattered from capsules in winter, the other when seeds were scattered from capsules in spring. One experimental plot was first sown on January 9, 2017, and then left in the open air(exposed to natural conditions) for about eighty days before the start of the germination; the other was sown on April 5. On that day, we moved the sowing containers left in the open air to the greenhouse and managed them together. In the greenhouse, only rainwater was blocked, but the entrances and sides remained open. The soil used in the sowing experiment was a mixture of garden soil(TKS-2, Floragard, Germany, N; 330mg/L, P; 220mg/L, K; 400mg/L) and coarse sand(Mountain soil; mainly contained granites; absence of nutrients) in the proportion of five to five. We used a commercially available square tray as the sowing container(50/540mm x 280mm, respectively, upper; 45mm x 45mm, lower; 15mm x 15mm, height 50mm). A total of eight repetitions were made to fill a square tray with fifty holes, and then seeds were sown to a depth of about 3 mm. Watering was done immediately after sowing, and then again once a day or once every two to three days depending on weather conditions. We measured the germination rates three times, on April 21, May 2, and May 15, 2017.

We used seedlings germinated in 2017 to observe the growth state depending on the rate of soil organic matter. There were four types of soil organic matter conditions: 100% for horticultural soil(TKS2), 50/50 for horticultural soil and coarse sand, 100% for peat moss(Sunshine, Canadian Sphagnum Peat Moss), and 100% for coarse sand. We prepared ten plastic experimental containers(13.5 cm in height, 13 cm in internal diameter, and 15 cm in external diameter) for each condition with the prepared soil filled to the height of the containers. On May 15, 2017, we selected forty individuals that were big enough from among the growing seedlings, and planted each in prepared plastic containers. Since the seedlings used for planting grew in germination containers, germinated soil of about 70 cm³ was included(TKS-2/Sand, 5:5 each). Planted seedlings were maintained under the same conditions and provided with 50% shading to avoid the effects of high temperatures in summer. In the flowering season of September 09, 2017 and August 14, 2018, we measured the height of the above-ground part and the number of fruits produced for each individual that grew under each condition. In May 2018, we counted the number of growing buds for each individual.

We selected fifteen well-grown individuals with three or more stalks from among the individuals sown in 2017 and grown in 50% horticultural soil and 50% coarse sand. On July 4, 2018, we selected a stalk from each individual before the start of the flowering. We installed insect nets(0.2 mm) on the inflorescences formed on the selected stalks to block the visit of pollinators. The remaining stalks were in the open to allow pollinators to visit. On August 14, 2018, when the fruits were matured, we measured and compared the number of seeds per each fruit produced in inflorescences where the visit of pollinators was blocked and inflorescences where the visit of pollinators was allowed, and then measured the weight of seeds according to their respective conditions(100 seeds: AND balance HR-200, Japan).

4) Statistical Analysis

We did Detrended correspondence analysis(DCA) ordination using vegetation data collected in the habitats(Canoco 4.53, Microcomputer Power, USA; Lepš and Šmilauer, 2007). The data collected from the tussocks in the YWPA and the data collected for the self-compatibility experiment were compared by the two-sample t-test. In the comparison of seed germination rates(n = 8) by sowing period, the samples did not satisfy normality; so we used the Mann-Whitney U Test. To compare the height of the individuals(n = 40), fruit production(n = 40), and the number of buds(n = 40) according to the organic material in the collected soil in the cultivation experiment conducted in 2017, we examined its normality by the Kolmogorov-Smirnov and Shapiro-Wilk tests, and later did a one-way analysis of variance when normality was satisfied. The differences between the samples were compared in terms of the Tukey HSD test as a post hoc test. However, in a two-year cultivation experiment, individuals that died because of poor drainage(n = 6) were observed in 2018 and were excluded from statistical analysis. Therefore, we used the Kruskal-Wallis H Test, because the normality of both the Kolmogorov-Smirnov and Shapiro-Wilk tests was not satisfied for the fruit production(n = 34) data collected during the 2018 cultivation experiment. In addition, we used the Mann-Whitney U test for comparison of the samples. IBM SPSS Statistics(v25, Hearne Software, USA) was used for all statistical analysis.

RESULTS

1. Comparison of the Vegetation in the Habitats between Two Areas

There was an obvious difference in vegetation between the YWPA in Inje-gun, Gangwon Province and the distribution area in the DWL, Pyeongchang-gun, Gangwon Province(Table 2). The distribution areas of both regions showed the highest importance values in the order of L. wilfordii(18.1), Persicaria thunbergii(13.0), Caltha palustris (11.1), Ostericum koreanum(11.0), Astilbe chinensis(8.5), Ostericum maximowiczii(6.4), and Persicaria siboldii (4.7)(Table 2). Using the tabulation methods to classify the species that observed in the two distribution areas, we found that the YWPA showed the highest importance values in the order of C. thunbergii var. appendiculata (14.8), Sanguisorba tenuifolia(13.1), Calmagrostis purpurea (6.2), Sium sisarum(4.7), Viola verecunda(3.1), Carex jaluensis (2.8), and Equisetum arvense(2.3). In the distribution area of DWL, Carex forficula(5.8), Angelica decursiva(3.0), and Pilea mongolica(2.2) were the Differential species that distinguished the two distribution areas(Table 2).

2. Distribution Characteristics in the YWPA

By conducting DCA ordination using the material collected from the continuous plots installed linearly in the distribution area and in the Keun yongneup, we distinguished the points in which L. wilfordii is distributed(Figure 4). According to the results of the vegetation survey conducted linearly at the three spots, first, region A(15 plots) showed the highest importance values in the order of C. thunbergii var. appendiculata (38.3), S. tenuifolia(26.0), Molinia japonica(14.6), P. thunbergii (13.5), O. koreanum(12.0), C. palustris(10.6), and L. wilfordii (8.2). Region B(117 plots) showed the highest importance values in the order of C. thunbergii var. appendiculata (39.7), S. tenuifolia(36.3), M. japonica(33.9), O. maximowiczii (11.8), Heloniopsis koreana(6.4), Parnassia palustris(5.8), Trientalis europaea var. arctica(4.8), and S. sisarum(4.6); the importance value of L. wilfordii was 2.2. Region C(40 plots) showed the highest importance values in the order of S. tenuifolia(36.3), C. thunbergii var. appendiculata(29.9), Persicaria sagittata(18.7), J. effusus (12.6), and C. langsdorfii(9.4); the importance value of L. wilfordii was 4.6. That is, L. wilfordii was not distributed throughout the linear plots, and only a few of them showed high coverage. L. wilfordii observed in only 22 plots of the total(172 plots), and these plots belonged to the low moor around the edge of the peatlands(Figure 4). The individuals of L. wilfordii did not appear in the high moor, which developed from peat accumulation in the YWPA(Figure 4). In other words, the space in which L. wilfordii was distributed was the low moor located at the edge of the high moor(Figure 4).

There were also differences in the species-diversity index between the space corresponding to the high moor and the space corresponding to the low moor(Table 3). The plots showing the characteristics of the high moor had a species density of 8.51, a dominance index of 0.66, an evenness index of 0.67, a species-diversity index of 1.41, and a maximum species-diversity index of 2.11(Table 3). On the other hand, the plots showing the characteristics of the low moor had a species density of 10.20, a dominance index of 0.25, an evenness index of 0.73, a species-diversity index of 1.68, and a maximum species-diversity index of 2.31(Table 3). In summary, the areas corresponding to low moor had a lower dominance index and a higher species-diversity index than high moor had(Table 3).

We measured the average height and width of the tussocks developed in the YWPA to investigate the role of the tussocks formed by C. thunbergii var. appendiculata in the distribution of L. wilfordii. There were differences between the average height of the tussocks in which L. wilfordii is distributed(29.3±11.9 cm) and the average height of the tussocks in which L. wilfordii is not distributed(23.0±7.2 cm)(two sample t-Test, t=-4.413, p<0.001; Table 4). There were also differences between the average width of the tussocks in which L. wilfordii is distributed(20.3±8.8 cm) and the average width of the tussocks in which L. wilfordii is not distributed(26.2±12.9 cm)(two sample t-Test, t=2.608, p<0.05; Table 4).

3. Evaluation of Growth Characteristics and Selfcompatibility

1) Growth Characteristics

The cumulative germination rate under the sowing conditions on January 9, 2017, was 50.0 ±5.4%. The cumulative germination rate under the sowing conditions on April 5, 2017, was 53.5 ±6.8%, and there was no statistical difference between the two sowing conditions (Mann-Whitney U Test=9.500, p=0.686; Figure 5A).

Evaluating the growth characteristics according to the organic matter contained in the soil, we found that the average height of the above-ground part of the individuals did not differ in 2017(one-way ANOVA, F = 0.584, p = 0.683; Figure 5C). Organic matter remaining in the soil used at the time of seeding did not make differences in the growth of any individuals. On the other hand, in 2018, there were differences in the average height of the above-ground part of individuals depending on the amount of soil organic matter(one way ANOVA, F=39.348, P<0.001; Figure 5E). There were differences between the TKS2 100% condition(71.4±15.3) and the TKS2/Sand 5:5 condition(53.1±9.2)(Tukey’ HSD test, p<0.05). The TKS 100% condition was also different from the Peat 100% condition(Tukey’ HSD test, p<0.001) and Sand 100% condition(Tukey’ HSD test, p<0.001). There were also differences between TKS2/S and 5:5 conditions, and the Peat 100%(Tukey’ HSD test, p<0.001) and Sand 100%(Tukey’ HSD test, p<0.001) conditions. However, there were no differences between the Peat 100% condition and the Sand 100% condition(Tukey’ HSD test, p=0.633).

The fruit production also depended on the amount of organic matter in the soil(Figure 5D, 5F). There were differences between the total number of fruits produced under each condition in 2017(one way ANOVA, F=7.511, P<0.05) and in 2018(Kruskal-Wallis H Test, x²=26.669, P<0.001)(Figure 5D, 5F). There were no differences between the total number of fruit produced under conditions of TKS2 100%(11.5±5.2) and TKS2/Sand 5:5(11.8±7.5) (Tukey’ HSD test, p=0.999; Figure 5D). There were also no differences between the Peat 100% condition(4.8±1.4) and the Sand 100% condition(3.9±3.1)(Tukey’ HSD test, p=0.976). On the other hand, the TKS2 100% condition was different from the Peat 100% condition (Tukey’ HSD test, p<0.05) and the Sand 100% condition(Tukey’ HSD test, p<0.01). In addition, the TKS2/Sand 5:5 condition was different between the Peat 100% condition(Tukey’ HSD test, p<0.05) and the Sand 100% condition(Tukey’ HSD test, p<0.01).

In 2018, the total number of fruits produced under each condition was more clearly different from 2017 (Figure 5F). The TKS2 100% condition(67.0±14.5) was different between the conditions of TKS2/Sand 5:5(26.0±6.9; Mann-Whitney U Test, p<0.001), Peat 100% (3.9±4.4; Mann-Whitney U Test, p<0.001) and Sand 100%(1.9±4.4; Mann-Whitney U Test, p<0.001)(Figure 5F). Also, the TKS2/Sand 5:5 condition was different between the conditions of Peat 100%(Mann-Whitney U Test, p<0.001) and Sand 100%(Mann-Whitney U Test, p<0.001). On the other hand, there were no difference between the Peat 100% condition and the Sand 100% condition(Mann-Whitney U Test=29.500, p=0.123).

There were differences in the number of buds for each individual produced under each condition in 2018(one way ANOVA, F=13.270, P<0.001; Figure 5B). There were no differences between the conditions of TKS2 100%(13.9±7.3) and TKS2/Sand 5:5(9.4±3.2)(Tukey’ HSD test, p=0.105), and there were no differences between the conditions of Peat 100%(3.6±1.5) and Sand 100%(4.5±1.4)(Tukey’ HSD test, p=0.988; Figure 5B). Whereas, the TKS2 100% condition had clear differences between the conditions of Peat 100%(Tukey’ HSD test, p<0.001) and Sand 100%(Tukey’ HSD test, p<0.001). Likewise, the TKS2/Sand 5:5 condition was different from the Peat 100% condition(Tukey’ HSD test, p<0.05); however, there were no differences in the Sand 100% condition(Tukey’ HSD test, p=0.06).

In cultivation experiments in terms of organic matter in soil, the height of the stalk of L. wilfordii remained constant, even if the amount of organic matter in the soil was high(Figure 5C, 5E). While, the production of buds increased in terms of vegetative reproduction(Figure 5B). The amount of fruit also increased depending on the amount of organic matter in soil(Figure 5F). In this cultivation experiment, although some organic matter was contained in the process of preparing the initial seedling, the plant also grew and produced a few fruits even for Peat 100% and Sand 100% without any additional organic matter(Figure 5F). In other words, L. wilfordii could survive and reproduce even in spaces with less organic matter that plants can use.

2) Evaluation of Self-compatibility

L. wilfordii produced seeds and had self-compatibility even when pollinators were blocked(Table 5). However, there were clear differences between the number of seeds when we blocked the visit of pollinators(13.3±16.1) and the seeds produced the visit of pollinators(52.0±26.3) was allowed(two sample t-Test, t=-6.872 P < 0.001; Table 5). Although the plant relies primarily on seed production in terms of cross-pollination, minimum seed production is possible even without the visit of pollinators(Table 5).

There were also differences between the weights of the seeds produced in the inflorescences where the visit of pollinators was blocked(0.0361±0.0015) and those produced where the visit of pollinators was allowed(0.0342±0.0009)(two sample t-Test, t=3.383 P < 0.01; Table 5). The weight of the seeds produced in the inflorescences where the visit of pollinators was blocked was greater. On the other hand, the weight of the seeds produced in the inflorescences where the visit of pollen vectors was allowed was light and differed(Table 5).

DISCUSSION AND CONCLUSION

1. Discussion

The differences in vegetation between the two distribution areas were influenced by the origin of the wetlands and the relatively recent changes in the vegetation. The YWPA has developed peatlands and has experienced no significant changes in the vegetation for a long time, whereas the distribution area in the DWL, experienced rapid environmental changes because of the influx of xylophytes, reforestation programs, and development of pastureland according to the progress of the vegetation succession. As a result, in the YWPA, C. thunbergii var. appendiculata dominated, and both species distributed in peatlands and species distributed in low wetlands, which is formed in mountainous areas, were distributed(Table 2). While, in the distribution area in the DWL, trees, such as Salix koreensis, Fraxinus rhynchophylla and Pyrus pyrifolia, and a shrub, such as Salix koriyanagi had invaded and were growing(Table 2). In addition, the importance value of the C. forficula was high, and the coverage of P. thunbergii, which occupied the disturbed space, was also high(Author's observation; Table 2). L. wilfordii in that area was distributed in the protruding space generated by the broadleaf herbaceous plant distributed around the tussocks rather than in the tussocks generated by the C. forficula. Therefore, it was influenced by competition and showed large differences in population size.

L. wilfordii in the YWPA was distributed in the low moor formed at the edge of Keun yongneup, the low moor remaining in the Jakeun yongneup, and the low moor developed in the Aegi yongneup(Table 2; Figure 6). As a result, the size of the space in which L. wilfordii can be distributed is closely related to the size of the population(Table 1). The low moor had a lower dominance and a higher species-diversity index than the high moor had, because the tussocks had increased the spaces where various plants can be distributed(Peach and Zedler, 2006;Yu et al., 2006). In addition, we found that the periodic flooding and repeated drying periods facilitate more vigorous decomposition of organic matter than that of the high moor, and that plants with high resource demand can be distributed. At the same time, the high species diversity in the low moor resulted from the development of the tussocks at certain distances(van de Koppel and Crain, 2006). The average distance between the tussocks in which L. wilfordii is distributed was 71.1±9.4 cm (Skewness: 0.007; Kurtosis: -1.210; Table 4). We observed regular spatial patterns relatively frequently in the peatlands, and it has been reported that the tussocks of Carex stricta Lam. showed regular spatial patterns at distances of 60 cm on average(van de Koppel and Crain, 2006). The tussocks of C. thunbergii var. appendiculata also showed similar spatial patterns, presumably because the distances between the tussocks are related to the length of the leaves of C. thunbergii var. appendiculata(Author's observation). Therefore, the spaces between the tussocks were evaluated as being not strongly blocked by light; hence they could function as spaces where plants could immigrate and establish.

The tussocks in which L. wilfordii is distributed are taller, and the width decreases relatively as the height increases; that was related to the water level during the inundation and rainfall periods involved in the growth and development of the tussocks. These differences are not directly applicable to the evaluation of the distribution position of L. wilfordii. In other words, the elapsed period and its relationships are more important factors after the influx and settlement of L. wilfordii and the formation of the tussocks. Nevertheless, we expect that the space in the tussocks in which L. wilfordii is not currently distributed could be transformed into a potential distribution site in the future.

It has been reported that seeds generally tend to increase their germination rates when exposed to low temperatures long enough while exposed to moisture in mid-latitude regions where winter exists(Ryu et al., 2017). However, L. wilfordii did not differ in germination rate even when exposed to low temperatures. On the other hand, not all the seeds produced would germinate all at once, and about 50 ±5% of them would produce latent-soil seed banks. At that time, the seeds of L. wilfordii were expected to physiologically have a dormancy mechanism(Bae et al., 2014;Ryu et al., 2017). The characteristics of these seeds were evaluated as a strategy for sustaining populations in highly uncertain environmental conditions.

In the distribution areas of L. wilfordii, irregular environmental conditions, such as rainfall, fog, and wind, can constrain the activity of pollinators. Therefore, we presumed that the plants have a growth strategy that has self-compatibility and therefore guarantees a minimum amount of seeds. These results were consistent with the recent study of L. wilfordii distributed in Japan(Tamura et al., 2016). Therefore, in the distribution area, all seeds were produced in inflorescences regardless of whether pollinators visited or not. Because of differences in the size of seeds produced depending on whether pollinators visit or not, we expected that seeds of different sizes would scatter in the distribution area. Although we did not do germination experiments to distinguish seeds produced under different conditions, the results showed an average germination rate of 50.1±5.6%, presumably because of having different germination characteristics that depend on the condition or size of the seeds produced. We considered the germination characteristics and dormancy of seeds to be important information for evaluating the sustainability of plant populations regenerated by seeds.

2. Conclusion

In Korea, as already mentioned, L. wilfordii is distributed only in two areas: the YWPA in Inje-gun, Gangwon Province, and two areas in the DWL(Figure 7). The two distribution areas are about 80 km apart and are evaluated as an isolated population with no pollen or seed dispersals between them(Kim et al., 2019).

Highly isolated populations are at a high risk of extinction(Lienert, 2004;Aguilar et al., 2008). In order for an isolated population to last for a long time, the minimum population size must be maintained(Pickett and Thompson, 1978;Pickett, 1980). Also, the dynamic process of extinction and resettlement within the population must be well done(Wiens, 1997;Hanski, 1998). Among the two populations of L. wilfordii, we evaluated the YWPA in Inje-gun, Gangwon Province as having a relatively stable population in size(Figure 7A). On the other hand, the population in the DWL, Pyeongchang-gun, Gangwon Province was small, and each subpopulation was also small(Figure 7B). The YWPA is a protected area where the quality of the habitat is well maintained, and each subpopulation is responsible for complementary functions and can be used as a source of seeds. Therefore, we expected that the population would be sustainable in the YWPA. On the other hand, the quality of the habitat in the DWL, has worsened because of the vegetation succession and the implementation of reforestation programs. In recent years, the damage pressure has been increased by the stamping caused by visits of native-plant and photographic clubs(Author's observation). In particular, the quality of habitats continues to deteriorate because of the closure of tree crowns by the increase in large herbaceous plants that are estimated to have a competitive advantage and the growth of vigorous trees, and the size of the population is expected to gradually decrease as the space required for the establishment of seedlings decreases. The decrease in available space for resettlement, the decrease in population numbers, and the high isolation between subpopulations consequently cause low genetic diversity and increase the chances of regional extinction(Kim et al., 2019).

Understanding how the population continues in distribution areas where populations remain stable despite being distributed in high isolation is critical for the effective conservation of populations(Crawley and Ross, 1990;Kim et al., 2016). L. wilfordii is distributed in the low moor in which tussocks are developed in the YWPA, and the size of the spaces in which L. wilfordii can be distributed is associated with the size of each population(Figure 4, Figure 6, Fiugre 7). Tussocks function as distribution areas for L. Wilfordii and form micro-habitats. We observed that especially L. wilfordii was mainly distributed in the upper part of the tussocks and settled in the space that was at least more topographically protruding than was the surrounding area(Tamura et al., 2016; Figure 8). Based on these points, the average height of the tussocks in which the L. wilfordii is distributed was 29.3±11.9 cm. In addition, in 2018 and 2019, we observed that the period in which the space where L. wilfordii established was submerged in water was shortened in the space where L. wilfordii is distributed because there was little rainfall(Author's observation). According to several literatures, L. wilfordii is known to prefer watery conditions in wetlands(Tamura et al., 2016;Kim et al., 2019). However, the space in which L. wilfordii is mainly distributed in the tussocks is relatively prone to be dry. On the other hand, in the space where inundation occurs for a long time, it micro-topographically would go into extinction. Therefore, we found that L. wilfordii is a wetland plant that selectively occupies a specific space in the wetland, rather than a plant that Facultative-Wetland Specialist.

In 2017, therophytes such as P. thunbergii and P. siboldii mainly settled in the plot in the YWPA where strong disturbances were caused by the feeding activities of wild boars, and we observed that the seeds scattered from L. wilfordii distributed in the tussocks germinated and resettled on the slope and the protruding space of the tussocks. Only some of the seedlings that germinated in June 2018 survived, and two of the surviving seedlings bloomed in September 2018. Seedlings that survived in 2018 were observed as young growing individuals in June 2019. In 2019, 24 seedlings were observed within the seeds rain, centering on the tussocks in which L. wilfordii is distributed. Therefore, we considered that L. wilfordii distributed in the tussocks will continuously scatter seeds to the surroundings every year and occupy the distributable space generated within its reach(Figure 8). However, we expected that the success of the resettlement by scattering seeds would progress slowly over a long time. The disturbance caused by wild boars in the YWPA is presumed to result from their feeding mostly on the storage roots of S. tenuifolia and the rhizome of O. koreanum(Author's observation). Therefore, we expect that the disturbances in the habitats will occur periodically, and the unaffected tussocks will function as a source of seeds required for sustaining the population of L. wilfordii. Hence we think that the ecological process of the feeding activities of wild boars and subsequent re-establishment was an important process for the continuation of the population in the YWPA.

Various studies have reported on the function and role of tussocks in peatlands and the cyclic microsuccession of tussocks(Fetcher and Shaver, 1982;Verhoeven, 1992;Crain and Bertness, 2005;Peach and Zedler, 2006;Yu et al., 2006). Repeated formation, growth, and extinction of the tussocks produced by C. thunbergii var. appendiculata in the YWPA have been observed(Figure 8). Since 2010, seedlings of L. wilfordii have been observed to resettle and grow not on the upper part of the tussocks, but mainly in the protruding spaces or in the lower tussocks that are in the early stages of development. The tallest tussock(48.0 cm) in which L. wilfordii are distributed is estimated to have developed over 100 to 150 years. In other words, there is a space in the YWPA in which C. thunbergii var. appendiculata was removed by the artificial or natural disturbances in the past, but later resettled and the tussocks developed again. The average height of the tussocks in this space was 25.4 ±7.4cm, which probably took about fifty years. Thus, the tussocks show the cycle of formation, growth, and extinction, and the space produced at this time can be occupied by seedlings settling from the seeds of L. wilfordii. The YWPA had a well-organized ecological process. Therefore, the population can be sustained for a very long time if the present natural process is well maintained. However, if too much organic matter flows into the peatlands because of artificial factors, that will lead to gradual extinction, or rapid extinction in some cases, as in the current distribution area of L. wilfordii in the DWL. So, we suggest that management is necessary to maintain the current ecological processes well and to maintain the health of ecosystems, including animals and plants.