Material Selection for Restoration of Genetic Diversity of Abies koreana on Mt. Jirisan in South Korea

Abstract A strategy is required for selecting appropriate materials for the restoration of Abies koreana on Mt. Jirisan, where the habitat of A. koreana is continuously shrinking. The current study aimed to analyze the genetic characteristics of A. koreana in three subpopulations (Banyabong, Byeoksoryeng, and Cheonwangbong) on Mt. Jirisan using 10 nuclear simple sequence repeat (nSSR) markers and calculate the sampling distance for each subpopulation for avoiding genetically similar samples. Based on the calculated sampling distance, we proposed the size of a sample containing more than 95% of the alleles at a frequency greater than 0.05. AMOVA showed that the difference in genetic variation across subpopulations of A. koreana on Mt. Jirisan was small, approximately 3% of the total. Spatial genetic structure analysis results suggested that it would be appropriate to collect samples of the Banyabong subpopulation at intervals of 10 m or more, when sampling A. koreana, whereas for the Byeoksoryeong and Cheonwangbong subpopulations, samples should be collected at intervals of 20 m or more. Results of random sampling of 5 to 30 individuals indicated that, by applying a 10 m distance within the Banyabong subpopulation, more than 95% of the total alleles with a frequency ≥ 0.05 were secured when more than 25 individuals were extracted. Therefore, as a restoration strategy for A. koreana on Mt. Jirisan, we proposed the collection of more than 25 samples, keeping 10 m distance within the Banyabong subpopulation, which has a relatively high genetic diversity.


Introduction
In the 20th century, global surface temperature increased by 0.7 °C (IPCC 2007), resulting in altered weather variables, such as precipitation (Bukhari & Bajwa 2011).The sub-alpine forest, located between timberline and treeline, is vulnerable to changed weather variables because it has poor environmental conditions, such as low temperatures, strong winds, and dry soil (Lim et al. 2006;Nagy & Grabherr 2009;Bukhari & Bajwa 2011;Kim et al. 2019).Mean temperature of the sub-alpine coniferous area, consisting of the dominant species of sub-alpine forest in Korea, increased by 0.8 °C over the past 30 years (1971-2000; 6.4-7.2 °C); this is higher than the average temperature increase rate in Korea (Kim & Lee 2013;Lim et al. 2019).The coniferous species in sub-alpine forest experienced physiological stresses, such as water stress due to increased evapotranspiration and imbalance in photosynthesis and respiration due to global warming, leading to a decrease in growth and an increase in dead trees (Park & Seo 1999;Koo et al. 2001;Kong 2002;Lim et al. 2006;Kim et al. 2019;Lim et al. 2019;Jiao et al. 2020;Kim et al. 2021).The sub-alpine coniferous area in Korea decreased by approximately 25% in 20 years   (Kim et al. 2019;Lim et al. 2019).The populations gradually decreased in size and increased in spatial isolation, resulting in a decrease in genetic diversity due to genetic drift or inbreeding (Young et al. 1996;Tang et al. 2008;Ahn et al. 2017).Low levels of genetic variation within a population limit their ability to adapt to environmental changes, and reduce seed production and seedling emergence, resulting in population decline and ultimately extinction (Booy et al. 2000;Hamrick 2004;Tang et al. 2008;Broadhurst & Boshier 2014).Therefore, preserving genetic diversity is important for the maintenance and conservation of sub-alpine coniferous forest in preparation about changing environmental conditions due to climate change (Kim et al. 2019).
Abies koreana, a native species of Korea, is a representative sub-alpine conifer in Korea that grows at altitudes higher than 1,000 m in Mt.Jirisan, Mt.Hallasan, Mt.Deogyusan, and Mt.Gayasan (Kim & Choo 2000;Chun et al. 2021).The Korean fir, A. koreana, had been designated as "endangered" by the International Union for Conservation of Nature (IUCN) Red List in expected heterozygosity; genetic diversity; population genetics; sampling strategy; simple sequence repeat; sub-alpine coniferous 2011 (Kim et al. 2011).Mt.Jirisan, the highest mountain in the southern part of Korean peninsula, covers the largest area, with 43% (5,198 ha) of the total area of sub-alpine coniferous in Korea, and A. koreana as the main species (Lim et al. 2019;Chun et al. 2021).Kim et al. (2021) reported that A. koreana of Mt.Jirisan has been declining in population, and dying every year since 2009.Research has shown that 49.9% of A. koreana in Mt.Jirisan has disappeared over the past 10 years (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018) (Chun et al. 2021), and seedling emergence has decreased by 22.4% in the last 2 years (Kim et al. 2018).In order to mitigate the loss of A. koreana habitat, appropriate restoration strategies, considering genetic diversity, would be required (McKay et al. 2005).
In order to use native plants as restoration materials, genetic characteristics of the populations need to be understood (McKay et al. 2005).The number of alleles in each population is a key indicator of optimal sampling (Marchall & Brown 1975).Although it is recommended to have as many alleles as possible to explain the population, an efficient sampling strategy would still be required, since manpower and costs are limited in the field.Our sampling strategy involved more than 95% of alleles composing in the target population with a frequency ≥ 0.05 (Marshall & Brown 1975).
In a previous study, applying the same sampling strategy to A. koreana populations on Mt.Hallasan, an appropriate strategy representing the genetic characteristics of A. koreana on Mt.Hallasan was proposed; more than 35 individuals were suggested to be collected at a distance of 5 m to 10 m each (Chae et al. 2022).Mt.Hallasan is located on an island and has one main peak, whereas Mt.Jirisan is located inland and has many peaks, and sub-alpine conifer species are continuously distributed along the ridge, forming the largest sub-alpine coniferous area in Korea.According to Hong et al. (2011), the two A. koreana populations differ in genetic characteristics.Therefore, it would be desirable to introduce different sampling strategies for both.A. koreana on Mt.Jirisan decreased by 14.6% over the past 20 years, requiring an appropriate restoration strategy (Lim et al. 2019).
This study aimed to analyze the genetic characteristics of A. koreana on Mt.Jirisan using nSSR markers and calculate the sampling distance for each subpopulation to reduce the number of genetically similar samples.Finally, based on the calculated sampling distance, we proposed the number of effective sample size, including appropriate genetic diversity of A. koreana on Mt.Jirisan population.

Study site and plant materials
For three subpopulations (Banyabong, Byeoksoryeng, and Cheonwangbong) on Mt.Jirisan, 99 A. koreana samples per subpopulation were selected for this study (Figure 1).The height and diameter at breast height (DBH) of each individual were measured (Table 1), location of each individual was recorded using GPS (GPS map60CSx; Garmin, Schaffhausen, Switzerland), and the location information was used for spatial structure analysis.

DNA extraction and nSSR marker analysis
Genomic DNA was extracted from the leaves of A. koreana using the Plasmid SV mini kit (GeneAll Biotechnology, Seoul, Korea).Its concentration was measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA) and was stored at -60 °C.

Summary statistics
Genetic diversity indices (A: number of alleles; A e : number of effective alleles; H o : observed heterozygosity; H e : expected heterozygosity; F: fixation index) were calculated for each subpopulation.Genetic variation across subpopulations was confirmed using analysis of molecular variance (AMOVA).The Nei's genetic distance and pairwise F ST value among the subpopulations were calculated.All analyses were performed using GenAlEx v.6.5.The significance level of F ST was determined using the FSTAT v2.9.4 software with 3000 permutations (Goudet 1995).

Spatial genetic structure analysis
The spatial genetic structure was analyzed by evaluating the genetic relevance of each plant individual according to the spatial distance (Gelmi-Candusso et al. 2017).Spatial autocorrelation analysis was performed using location of the individual and GenAlEx v.6.5 (Smouse & Peakall 2012).Distance evaluation was performed by distance classes from 10 m to 100 m, and 999 permutations were performed for each evaluation with a 95% confidence interval (Smouse & Peakall 1999).

Sample selection strategies
In order to establish an optimal sampling strategy, results of the spatial genetic structure analysis were used as the target distance.Within the representative subpopulation of Mt.Jirisan, 5, 10, 15, 20, 25, and 30 sample sizes were randomly extracted 10 times each using Python (Chae et al. 2022).We calculated the number of alleles in each sample size and compared them with the number of alleles on Mt.Jirisan with a frequency ≥ 0.05; the mean comparison was performed using Duncan's test at p ≤ 0.05 with R (De Mendiburu & Simon 2015).

Genetic diversity and distances within subpopulations
Analysis of genetic diversity of A. koreana on Mt.Jirisan indicated the average of the number of alleles (A), the number of effective alleles (A e ), observed heterozygosity (H o ), expected heterozygosity (H e ), and fixation index (F) to be 11.1, 5.2, 0.570, 0.657, and 0.106, respectively (Table 3).Comparing genetic diversity across the subpopulations on Mt.Jirisan, we found the Byeoksoryoung subpopulation to show the highest genetic diversity (H e = 0.663).The Banyabong subpopulation had the highest number of effective alleles among the three subpopulations and showed the lowest fixation index (F = 0.073).The Cheonwangbong subpopulation showed the lowest genetic diversity (H e = 0.625).
The AMOVA results showed that the genetic variance across subpopulations was 3% of the total, whereas that among individuals within each subpopulation was 97% of the total (Table 4).The Nei's genetic distance across subpopulations was 0.040-0.047,and the F ST value between each subpopulation was 0.013-0.011(Table 5).

Spatial genetic structure
Patterns of spatial genetic structure of A. koreana subpopulations on Mt.Jirisan are shown in Figure 2. If the observed mean pairwise genetic distance in each distance class was greater than or less than the 95% confidence interval (generated by 1000 permutations), the individual was considered to be randomly distributed.Individuals within the Byeoksoryeng subpopulation were genetically similar when distributed within a distance of 20 m, whereas those within a distance of 20 m-60 m were randomly distributed (Figure 2a).Some genetically similar structures were found in individuals with distances in the range of 60 m-70 m, whereas those with distances over 80 m were randomly distributed.Within the Banyabong subpopulation, individuals over a distance of 10 m had the same genetic structure, and those with a distance of 10 m-30 m were randomly distributed (Figure 2b).Although some genetically similar structures were found in individuals with distances between 30 m and 40 m, those with distances > 40 m were randomly distributed.Within the Cheonwangbong subpopulation, individuals distributed over 20 m distance had the same genetic structure, and those with a distance between 20 m and 30 m were randomly distributed (Figure 2c).Although some genetically similar structures were found in individuals with distances between 30 m and 40 m, those with distances greater than 40 m were randomly distributed.

Sample size selection
For sample size selection of restoration materials containing alleles with a frequency ≥ 0.05, the sampling distance presented above was applied and the optimal sample size was determined.As seen from the previous genetic distance analysis across subpopulations, the difference between each subpopulation was small.Therefore, the appropriate number of samples was analyzed for the Banyabong subpopulation, which had the highest number of alleles among the three subpopulations.Random sampling of 5 to 30 individuals, by applying a 10 m distance, in the Banyabong subpopulation showed that more than 95% of total alleles with a frequency ≥ 0.05 were secured when more than 25 individuals were extracted (Figure 3).

Discussion
Genetic diversity is influenced by ecological factors, such as the geographical distribution and breeding methods of species (Ueno et al. 2000;Shin et al. 2014).The A. koreana habitat on Mt.Jirisan is fragmented, and young trees are rapidly declining in number; therefore, genetic diversity is likely to decrease (Kim et al. 2018).Hence, it is necessary to establish an appropriate restoration strategy to maintain and conserve the genetic diversity of A. koreana on Mt.Jirisan.
Large populations of A. koreana in Korea include Mt.Jirisan, Mt.Hallasan, and Mt.Deogyusan, and small populations with small distribution areas include Mt.Geumwonsan, Mt.Wolbongsan, and Mt.Baegunsan  (Lim et al. 2019;Lim et al., 2020).According to Kwak et al. (2017), A. koreana populations located inland, including Mt. Jirisan, Mt.Deogyusan, and Mt.Geumwonsan, and A. koreana populations located on the island, including Mt. Hallasan, have genetically different patterns.In addition, the large population of A. koreana in Mt.Jirisan was reported to have relatively stable genetic diversity (H e = 0.656), while the small population of A. koreana in Mt.Geumwonsan was evaluated to have relatively low genetic diversity with H e = 0.612.In particular, Lim et al. (2020) reported that the genetic variation of A. koreana seedlings in Mt.Geumwon was H e = 0.592, F = 0.318, indicating relatively low genetic diversity and high risk of inbreeding.
In the current study, the H e value of A. koreana on Mt.Jirisan was 0.657.The genetic diversity of A. koreana on Mt.Jirisan we observed (H e = 0.672) was similar to that reported by Ahn et al. (2017) but was lower than that reported by Kwak et al. (2017) (H e = 0.778).The H e , an index representing the genetic diversity within a population, is estimated via the allele frequencies calculated by the marker (Hartl & Clark 1997).Therefore, differences in the results from various studies can be considered to be based on the type and number of markers used in the analysis.
The AMOVA results showed the difference in genetic variation across subpopulations of A. koreana on Mt.Jirisan to be 3% of the total, indicating the genetic difference to be small across the subpopulations (Table 4).According to a previous study (Chae et al. 2022), the subpopulations of A. koreana on Mt.Hallasan showed a difference of 4% in genetic variation, similar to the subpopulations of A. koreana on Mt.Jirisan.In addition, the genetic distance (F ST ) across the subpopulations of A. koreana on Mt.Jirisan was 0.013-0.011(Table 5).The results indicated little genetic difference across the subpopulations of A. koreana on Mt.Jirisan.According to a previous study, the F ST across the three subpopulations on Mt.Jirisan was 0.014, and the optimal population on Mt.Jirisan, estimated by the Pritchard method, was reported to be 1 appropriate (Ahn et al. 2017).Therefore, it should not be difficult to secure most of the alleles, without considering the subpopulation, when selecting the restoration material of A. koreana on Mt.Jirisan.
An understanding of the spatial genetic structure of populations is required to maximize the genetic diversity or minimize consanguinity; knowledge of the spatial genetic structure can be used to develop breeding programs and in-situ conservation of natural populations (Ueno et al. 2000).In this paper, while selecting materials for the restoration of A. koreana on Mt.Jirisan, spatial autocorrelation analysis was performed on the genetic variation of each subpopulation on Mt.Jirisan to avoid genetically similar samples.Analysis results showed that it would be appropriate to take samples of the Banyabong subpopulation at intervals of 10 m or more when sampling A. koreana, whereas for the Byeoksoryeong and Cheonwangbong subpopulations, it would be appropriate to collect samples at intervals of 20 m or more (Figure 2).Chae et al. (2022) had proposed a strategy to sample A. koreana on Mt.Hallasan, at a distance of 5 m in the Bangaeoreum subpopulation and at 10 m in the Yeongsil and Jindalleabat subpopulations.The A. koreana subpopulations on Mt.Jirisan are characterized by lower density (number of A. koreana per 1 ha on Mt.Jirisan, 320.5; number of A. koreana per 1 ha on Mt.Hallasan, 647.9; (Lim et al. 2019) and higher height than A. koreana on Mt.Hallasan (average height of A. koreana of Mt.Jirisan, 15.85 m; average height of A. koreana on Mt.Hallasan, 4.17 m) (Chae et al. 2022).According to Epperson (1992), lower the density of the forest, farther can the pollen or seeds be dispersed, which might explain the results of the spatial autocorrelation analysis of population.In addition, higher plant heights can cause seed dispersion to occur farther away, which can change the results of spatial autocorrelation analysis (Thomson et al. 2011).Therefore, it is considered that the geographical distance of genetically similar individuals of A. koreana of Mt.Jirisan is farther than that of genetically similar individuals of A. koreana of Mt.Hallasan.
For the purpose of sampling 95% alleles of the total genetic information of A. koreana on Mt.Jirisan with a frequency ≥ 0.05, we calculated the appropriate number of restoration materials.Results showed that if more than 25 samples are collected at a distance of 10 m for the Banyabong subpopulation, with relatively high genetic diversity, the latter of A. koreana on Mt.Jirisan would be stably secured.Lim et al. (2019) reported that restoration efforts are urgently required for the endangered sub-alpine coniferous species A. nephrolepis, Picea jezoensis, Thuja koraiensis, Juniperus chinensis, Picea pumila, and Taxus cuspidate, in addition to A. koreana.Our present approach can be applied to develop conservation and restoration strategies for coniferous species other than A. koreana that are vulnerable given the changing and hostile environmental conditions.

Conclusions
A. koreana populations on Mt.Jirisan need appropriate maintenance and conservation strategy to survive the changing environmental conditions resulting from climate change.Restoration using native species with genetic diversity adapted to the environmental conditions of their native habitat would enable long-term restoration success.We established restoration strategies via the analysis of genetic diversity and spatial genetic structure of A. koreana on Mt.Jirisan.In this study, genetic diversity of the three subpopulations on Mt.Jirisan was found to be relatively high, and genetic variation across subpopulations was rarely seen.Therefore, the restoration strategy for A. koreana on Mt.Jirisan would involve the collection of more than 25 samples at a distance of 10 m each within the Banyabong subpopulation.

Figure 2 .
Figure 2. results of spatial structure analysis for Abies koreana using 10 nuclear simple sequence repeat (nssr) markers.the r value is an autocorrelation coefficient.the solid line represents the correlation between genetic distance and geographic distance.the dotted lines represent 95% confidence intervals.(a) Byeoksoryeng; (b) Banyabong; (c) cheonwangbong.

Figure 3 .
Figure 3. comparison of allele numbers with a frequency ≥ 0.05 in Abies koreana on Mt.Jirisan according to the sample size extracted based on the sampling strategy.Values with different letters are significantly different according to duncan's test at p ≤ 0.05.

Table 2 .
characteristics of 10 nuclear simple sequence repeat (nssr) markers used in the study
A: number of alleles; A e : number of effective alleles; H o : observed heterozygosity; H e : expected heterozygosity; F: fixation index.

Table 4 .
assessment of genetic differentiation of subpopulations by analysis of molecular variance (aMoVa).

Table 5 .
nei's genetic distance below the diagonal and pairwise F st values above the diagonal among the three subpopulations of Abies koreana on Mt.Jirisan.