Genetic Diversity of Five Broadleaved Tree Species and Its Spatial Distribution in Self-Regenerating Stands

Abstract

European forest ecosystems are currently subject to various disturbances and shifts in land-use legacies. To be able to forecast the trends and consequences of the changes in genetic diversity following these disturbances, it is of crucial importance to understand the genetic dynamics of natural tree populations. The present study aimed at determining the extent and spatial distribution of genetic diversity in five common broadleaved tree species in Lithuania in both mature (putatively maternal) stands and in natural regeneration (juveniles) of the respective species. The genetic diversity of Quercus robur, Betula pendula, Populus tremula, Alnus glutinosa, and Fraxinus excelsior was assessed using eight nuclear microsatellite loci for each species; 417 samples of regenerating juveniles and 141 samples of putatively maternal trees were analyzed in total. The investigated populations of self-regenerating Q. robur, B. pendula, A. glutinosa, and F. excelsior juveniles showed spatially random genetic structures, while P. tremula regenerated mostly via root suckers and formed clonal groups. The genetic diversity in regenerating juveniles of all species was as high as in putatively maternal stands. The detection of adequate (substantial) genetic diversity in the studied regenerating populations of these five broadleaved tree species suggests that in Lithuania these species have a good potential to adapt to changing environmental conditions.

1. Introduction

The stability and functionality of forest ecosystems greatly depend on the genetic diversity of the tree populations [1]. Natural forest ecosystems in Europe have been heavily affected by humans for a long period of time, which may have resulted in reduced population sizes and loss of genetic diversity in certain tree species. Widespread forest tree species are able to maintain high genetic diversity due to efficient gene flow, large effective population sizes, long generation time, and successful strategies of natural regeneration [2,3,4]. However, there is a serious concern as to whether commercial tree species will be able to sustain equal levels of genetic diversity under current intensive forest management practices [5,6], especially under rapidly changing environmental conditions [2]. Currently, the genetic diversity of European forest tree species is sufficient for their survival and adaptation to a changing climate [7,8]. However, the foreseen acceleration of climate change may disturb their evolutionary mechanisms, which currently help forest tree populations to adapt naturally to changes in environmental conditions.

Trees are among the most genetically diverse terrestrial organisms on Earth [3]. The genetic diversity and structure of a forest population are modified by demographic factors, including the clustering and density of relatives and their spatial distribution [9]. Severe natural or anthropogenic disturbances—such as windthrows, stand self-thinning, disease or pest outbreaks, forest fires, and fellings—may considerably reduce the genetic diversity in tree populations [1]. The size of plant populations, their fitness, and their genetic diversity often display positive correlations [10]. However, some studies have failed to demonstrate a clear correlation in this respect; for example, Buiteveld et al. [11] found no drastic genetic diversity loss after reduction in the population size of European beech (Fagus sylvatica L.) following different forest management practices. The observed phenomenon was explained by a sufficient number of reproducing trees being retained in the investigated populations.

Tree populations regenerating after severe natural or anthropogenic disturbances can undergo sharp reductions in genetic diversity due to a genetic bottleneck and/or a founder effect [5,12]. The reduction in genetic diversity may be greater at clearcut or unforested sites, where regenerating juvenile trees are affected by harsher microclimatic conditions (e.g., spring frosts, droughts, high summer temperatures, etc.) than seedlings that regenerate under stand shelter [13]. The genetic diversity of populations is collectively affected by natural selection, genetic drift, and gene flow that either promote or hamper local and range-wide adaptation [1]. The gene flow results in different combinations of rare alleles in maternal and progeny populations [14]. It is assumed that rare alleles—especially those with small effects—may possess attributes that are useful for adaptation to changing environmental conditions [15,16].

Information on the spatial and temporal distribution of genetic diversity is crucial for efforts to preserve genetic diversity in forest tree populations and, thus, for the survival and stability of entire forest ecosystems [17]. The spatial distribution of genetic diversity in forest stands has a certain impact on the genetic properties of new forest generations, as the majority of flowers of the maternal tree are fertilized by pollen from neighboring trees that grow within a 15–30 m radius around it [18,19]. Moreover, natural selection is reflected by the temporal distribution of genetic diversity in forest stands, which is normally reduced in juvenile trees with increasing stand age [20,21,22].

In Lithuania, the intensive forest management practices and frequent occurrence of various climate-change-related disturbances (e.g., blowdown, disease or pest outbreaks) that have taken place during the last three decades may have negatively affected the genetic diversity in the populations of certain tree species—especially those that have recently suffered from severe disease (e.g., the currently ongoing dieback of European ash (Fraxinus excelsior L.) [23], decline of common oak (Quercus robur L.) (a series of epidemics in 2004–2006, caused by an unknown pathogen(s) [24]), or pest outbreaks (i.e., attacks by spruce bark beetle (Ips typographus L.) on Norway spruce (Picea abies Karst.) in the mid-1990s [25]). Our recent study [26] demonstrated that, in Lithuania, juveniles of Norway spruce and Scots pine (Pinus sylvestris L.) self-regenerating in areas subjected to either natural disturbances (e.g., windthrows and subsequent clearfelling of the damaged spruce stand) or to a changed land-use legacy (e.g., self-regeneration of pine on abandoned agricultural fields) show a spatially random and equally high genetic diversity as in the putative maternal populations. It has been concluded that this high genetic diversity in the regenerating conifer populations can provide a basis for the formation of ecologically and evolutionary sound stands that are able to adapt to a changing climate. However, the situation with disturbed populations of broadleaved tree species is largely unknown. The main aim of the present study was therefore to determine the extent and spatial distribution of genetic diversity in five common and economically important broadleaved tree species in Lithuania, both in mature (putatively maternal) stands and in natural regeneration of the respective species in areas subjected either to natural disturbances (e.g., disease outbreaks and subsequent sanitary clearfelling of the affected common oak and European ash stands), anthropogenic disturbances (e.g., clearfelling in European aspen (Populus tremula L.) and black alder (Alnus glutinosa (L.) Gaertn.) stands), or a changed land-use legacy (e.g., silver birch (Betula pendula Roth.) regeneration on abandoned agricultural land). The study aimed at answering the question of whether the current genetic diversity of the natural regeneration of five broadleaved tree species in disturbed areas is adequate compared to that of mature tree populations of the same species, i.e., if the sustainability of Lithuanian broadleaved tree populations (secured by a certain level of their genetic diversity) is not compromised.

2. Materials and Methods

2.1. Study Sites

Five stands each representing one of the five economically important broadleaved tree species in Lithuania—namely, common oak, silver birch, European aspen, black alder, and European ash—were selected for this study. The selected stands represented habitats that are typical for these tree species in Lithuania. The common oak site represented a pure 130-year-old natural common oak stand devastated by an unknown disease in around 2004–2006 and clearfelled in 2007. At the time of sampling, the age of the self-regenerating oak seedlings was 5–8 years. The silver birch site represented an approx. 2–4-year-old self-regeneration of silver birch and Scots pine growing on abandoned agricultural land (see Figure 1 in the work of Verbylaitė et al. [26]). A natural mixed pine–birch stand (growing northwest of the investigated area) was the most likely seed source for self-regenerating birch juveniles; however, the possibility of seed influx from further growing stands (approx. 1 km away) should not be excluded. The European aspen site represented a pure 60-year-old natural European aspen stand that was clearfelled in 2013. The sampled self-regenerating aspen juveniles were 2–4 years old. The black alder site represented a pure 80-year-old natural black alder stand that was clearfelled in 2013. The sampled self-regenerating alder juveniles were 2–4 years old. The European ash site represented a pure 60-year-old natural European ash stand devastated by ash dieback disease and selectively felled in 2013. The sampled self-regenerating ash juveniles were 2–10 years old. More information about the study sites is presented in

Table 1. Main characteristics of the study sites.

Study Site (Stand-Forming Species)	Type and Year of Disturbance	Regional Division of State Forest Enterprise	Forest Site Type a	Latitude N Longitude E
Quercus robur L.	Sanitary clearfelling in 2007	Jonava	Oxalido-nemorosa	55°9.180′ 24°6.946′
Betula pendula Roth.	Abandoned agricultural land since 2013	Tytuvėnai	Vaccinio-myrtillosa	55°31.934′ 23°10.858′
Populus tremula L.	Regular clearfelling in 2013	Anykščiai	Oxalido-nemorosa	55°32.354′ 24°51.835′
Alnus glutinosa (L.) Gaertn.	Regular clearfelling in 2013	Prienai	Oxalido-nemorosa	54°43.946′ 23°53.748′
Fraxinus excelsior L.	Sanitary clearfelling in 2013	Kėdainiai	Carico-mixtoherbosa	55°13.560′ 23°58.127′

^a According to Karazija [27], based on key herbaceous species.

.

The seed sources for the self-regeneration were either seed trees left in clearcut sites (i.e., European aspen, black alder, common oak, and European ash) or maternal trees from an adjacent mature stand (as with silver birch). However, long-distance seed flow should be taken into consideration.

2.2. Sampling of Plant Material

Leaf samples for the genetic studies of regenerating juveniles of each of the five investigated tree species were collected in seven linearly distributed circular plots, each with a diameter of 3 m. The distance between each pair of plots was 25 m. Six juveniles of the respective investigated species were randomly sampled in each plot. In addition, six juveniles of the same species were sampled from outside of the circular plots, in two perpendicular opposite locations (three seedlings in each) situated at a 20 m distance from each circular plot (Figure 1). In this way, a total of 84 juveniles were sampled at each site representing one investigated tree species, except for silver birch, where only 81 juveniles were sampled due to the absence of birch seedlings in one circular plot. Such a sampling scheme was chosen to reveal any possible genetic relatedness among juveniles of the respective tree species growing in compact areas (clumps) of equal size, and to check for the presence of spatial genetic structures within and between these clumps. Moreover, at each site, 30 mature, well-developed (i.e., seed-producing) trees that possibly contributed seeds for the regeneration (i.e., putatively maternal trees) were sampled in surrounding ‘undisturbed’ stands of the respective species. The only exception was European ash, for which it was not possible to find ‘undisturbed’ sites because of the ongoing ash dieback epidemic and subsequent sanitary fellings in virtually every European ash stand. Moreover, in the case of European ash, we sampled only 21 putatively maternal trees that were retained at the clearcut sites. Some retained European aspen and black alder seed trees were also sampled at regenerating clearcut sites of the respective species.

The plant material was sampled in spring through summer 2017. Fully developed asymptomatic leaves were collected from every sampled tree and transported to the laboratory on the same day. Prior to processing, all leaf samples were stored at −20 °C.

2.3. Microsatellite Analysis

DNA extraction from the collected plant material was performed as described by Verbylaite et al. [26]. Eight simple sequence repeat (SSR) loci were chosen for genetic investigation of each tree species, as this number of loci helps in obtaining reliable results in population studies [28]. It has been shown that an even smaller number of unlinked SSR markers allows the detection of population structure at recent divergence times [29].

For genetic analysis of common oak, primers described by Dow et al. [30] (MSQ4 and MSQ13), Steinkellner et al. [31] (ssrQpZAG 9, ssrQpZAG 36 and ssrQpZAG 110), and Kampfer et al. [32] (ssrQrZAG 7, ssrQrZAG 20 and ssrQrZAG 11) were employed. Amplification of oak DNA was performed in two multiplex reactions as described by Dzialuk et al. [33]. Amplification of black alder DNA was performed in a multiplex reaction using eight primers as described by Drašnarová et al. [34]. For silver birch, the primers (L1.10, L2.2, L2.3, L2.7, L3.1, L5.4, L5.5, and L13.1) and PCR conditions were adopted from the work of Kulju et al. [35]. The primers used for the genetic analysis of European ash were adopted from the works of Lefort et al. [36] (FEMSATL4, FEMSATL8, FEMSATL10, FEMSATL11, FEMSATL12, FEMSATL16, FEMSATL19) and Brachet et al. [37] (M2.30). The primers used for European aspen analysis were adopted from the works of van der Schoot et al. [38] (WPMS03 and WPMS09), Smulders et al. [39] (WPMS14 and WPMS16), and from web sources [40] (PMGC14, PMGC2607, GCPM1532, GCPM1608). The PCR reactions for European ash and European aspen were performed separately for each sample, using the protocols described by Tereba et al. [41] and Bruegmann and Fladung [42], respectively.

Amplicon purification and subsequent PCR product analysis were performed as described by Verbylaite et al. [26]. Geneious prime computer software (http://www.geneious.com [43]) was used for fragment length analysis.

2.4. Statistical Analysis

The main genetic diversity parameters (i.e., number of alleles (Na), effective number of alleles (Ne), observed heterozygosity (Ho), expected heterozygosity (He), fixation index (F), and number of migrants per generation (Nm)) were calculated using GenAlEx version 6.51b2 computer software [44,45,46]. The number of multilocus genotypes (MLGs) per investigated species was determined using the GenAlEx software package. The same software package was also used to perform principal coordinate analysis (PCoA) using Nei’s genetic distance matrix, the Mantel test, and analysis of molecular variance (AMOVA), as well as to check for spatial genetic structure and deviation of the loci from Hardy–Weinberg equilibrium (HWE). The allelic richness of the mature and juvenile cohorts was calculated using FSTAT computer software [47]. The calculation of Ar was based on 30 diploid samples for all tree species except F. excelsior, where it was based on only 21 samples (due to the lower number of sampled mature F. excelsior individuals). The linkage disequilibrium method [48] was used to calculate the effective population sizes with NeEstimator v 2.1. computer software [49]. Latent genetic potential (LGP) was calculated as the difference between the total and effective number of alleles summed over all loci [50]. Statistical significance of genetic diversity parameters between mature and juvenile cohorts was tested using two-tailed Student’s t-tests. A Bayesian clustering approach was implemented using STRUCTURE version 2.3.4 [51,52,53,54] to infer the most probable number of clusters for each investigated species. An admixture model with correlated allele frequencies was assumed, and the maximum number of clusters (K) tested was 16, as the most likely value of K was expected to fall in this range, with 10 replications for each K. The length of the burn-in period and the Monte Carlo Markov chain (MCMC) was 100,000 iterations. STRUCTURE Harvester computer software [55] was used to identify the most likely number of clusters (using the delta K (ΔK) criterion).

3. Results

All sampled trees of common oak (n = 114), silver birch (n = 111), and European ash (n = 105)—including both putatively maternal trees and regenerating juveniles—represented unique MLGs. Black alder trees (n = 114) represented 113 MLGs (two adjacent trees—one putatively maternal and one juvenile—belonged to the same MLG), while European aspen trees (n = 114) represented 96 MLGs; this tree species formed clonal groups. The largest aspen clonal group comprised five juveniles sampled in plots 6a (three ramets) and 6c (two ramets), along with a maternal tree growing between these two sample plots (Figure 2; the ramets of this clone are indicated by the letter ‘M’). All investigated putatively maternal aspen trees represented unique MLGs, except for two that grew next to one another in the surrounding stand (Figure 2; these trees are indicated by the letter ‘K’).

All microsatellite loci chosen for this study were polymorphic and revealed 5 to 30 different alleles each. On average, 16.5 alleles were found for common oak, 17.9 alleles for silver birch, 8.5 alleles for European aspen, 12.0 alleles for black alder, and 22.6 alleles for European ash. The highest numbers of alleles were found in locus ZAG11 of common oak (20 alleles), locus L2.7 of silver birch (27 alleles), locus WPMS14 of European aspen (13 alleles), locus A37 of black alder (19 alleles), and locus Femsatl4 of European ash (30 alleles).

The results of principal coordinate analysis (PCoA) revealed small genetic distances between the regenerating juveniles and the putatively maternal trees in each investigated species (Figure 3). The percentages of genetic variation explained by the first three coordinates of the PCoA are given in

Table 2. The percentages of genetic variation in five broadleaved tree species explained by the first three coordinates of the principal coordinate analysis (PCoA; diagrams presented in Figure 3).

Species	Percentage of Variation	Coordinate 1	Coordinate 2	Coordinate 3
Quercus robur	%	22.68%	19.21%	16.52%
	Cumulative %	22.68%	41.89%	58.41%
Betula pendula	%	27.50%	16.81%	15.54%
	Cumulative %	27.50%	44.32%	59.86%
Populus tremula	%	28.84%	26.28%	15.17%
	Cumulative %	28.84%	55.12%	70.29%
Alnus glutinosa	%	21.35%	19.83%	17.01%
	Cumulative %	21.35%	41.18%	58.19%
Fraxinus excelsior	%	23.98%	22.07%	15.16%
	Cumulative %	23.98%	46.05%	61.21%

.

The distribution of points on the scatterplots (Figure 3) shows Nei’s genetic distances among the investigated individuals. The distribution of MLGs of common oak, silver birch, and European aspen regenerating juveniles was slightly more scattered than the MLGs of putatively maternal trees of the same species (Figure 3A–C, respectively). The MLGs of oak juveniles (even those sampled from compact circular plots) and putatively maternal oak trees revealed no clear grouping pattern and overlapped to great extent (Figure 3A). A similar scattered pattern was found for silver birch (Figure 3B), black alder (Figure 3D), and European ash (Figure 3E), while the MLGs of European aspen juveniles showed a grouping pattern in most circular plots (Figure 3C). The MLGs of putatively maternal European aspen trees showed a scattered pattern of distribution and overlapped to a great extent with the MLGs of the regenerating juveniles.

The Mantel test showed weak and non-significant correlation between the genetic and spatial distances of the regenerating juveniles of all studied species except for European aspen, where a low yet significant correlation was found (r = 0.171, p < 0.001) (Supplementary Materials S1). The spatial correlograms of juveniles of all five tree species revealed random spatial distribution, and the correlograms were not significant.

The main genetic diversity parameters (Na, Ne, Ho, He, F, Ar) of the regenerating juveniles and putatively maternal trees (

Table 3. Main parameters of genetic diversity, calculated for putatively maternal (mature) trees and self-regenerating juveniles (young) of common oak (Quercus robur L.), silver birch (Betula pendula Roth.), European aspen (Populus tremula L.), black alder (Alnus glutinosa (L.) Gaertn.), and European ash (Fraxinus excelsior L.). N, sample size (number of sampled trees); Na, mean number of alleles; Ne, effective number of alleles; Ho, observed heterozygosity; He, expected heterozygosity; F, mean fixation index; Ar, allelic richness; CI, confidence interval; LGP, latent genetic potential; ±SE, standard error.

Sample Group	N	Na ± SE	Ne ± SE	Ho ± SE	He ± SE	F ± SE	Ar ± SE	Effective Population Size (95% CI)	LGP ± SE
Q. robur mature	30	12.125 ± 0.972	7.241 ± 0.995	0.762 ± 0.051	0.839 ± 0.026	0.093 ± 0.051	12.13 ± 0.909	129.8 (73.7–421.1)	4.884 ± 0.412
Q. robur young	84	14.750 ± 0.675	6.666 ± 0.969	0.769 ± 0.038	0.817 ± 0.036	0.057 ± 0.028	11.44 ± 0.689	102.9 (86.3–125.7)	8.084 ± 0.791
B. pendula mature	30	11.500 ± 1.210	4.281 ± 0.583	0.754 ± 0.045	0.731 ± 0.039	−0.042 ± 0.062	11.50 ± 1.132	109.6 (62–361.4)	7.129 ± 0.893
B. pendula young	81	16.000 ± 2.000	5.459 ± 0.799	0.853 ± 0.021	0.789 ± 0.029	−0.094 ± 0.052	11.75 ± 1.153	208.5 (156.3–304.3)	10.541 ± 1.572
P. tremula mature	30	7.000 ± 0.845	3.295 ± 0.450	0.829 ± 0.080	0.655 ± 0.047	−0.251 ± 0.072	7.00 ± 0.791	26.9 (18.4–43.5)	3.705 ± 0.531
P. tremula young	84	7.375 ± 0.596	3.022 ± 0.555	0.759 ± 0.109	0.598 ± 0.065	−0.214 ± 0.123	6.01 ± 0.408	13.5 (11.4–15.8)	4.353 ± 0.521
A. glutinosa mature	30	9.750 ± 1.098	4.562 ± 0.575	0.833 ± 0.063	0.740 ± 0.051	−0.128 ± 0.048	9.75 ± 1.023	66.6 (41.5–143.5)	5.188 ± 0.557
A. glutinosa young	84	10.500 ± 1.648	4.229 ± 0.642	0.826 ± 0.067	0.704 ± 0.061	−0.194 ± 0.079	8.32 ± 1.114	793.9 (284–∞)	6.271 ± 0.984
F. excelsior mature	21	12.375 ± 1.580	7.027 ± 1.440	0.625 ± 0.064	0.795 ± 0.051	0.223 ± 0.053	12.38 ± 1.478	∞ (138.3–∞)	5.348 ± 0.944
F. excelsior young	84	20.875 ± 2.341	8.252 ± 1.666	0.665 ± 0.062	0.825 ± 0.050	0.196 ± 0.054	13.31 ± 1.413	1054.4 (504–∞)	12.623 ± 1.560

) were calculated in order to compare these parameters between two tree generations. All investigated species had a larger mean number of alleles in juveniles than in putatively maternal trees, but statistically significant differences between the mature (maternal) and regenerating (juvenile) cohorts were found only for Na in Q. robur and F. excelsior (p = 0.045 and p = 0.010, respectively). Higher effective number of alleles and allelic richness were found in mature cohorts, except for B. pendula and F. excelsior, for which these parameters were higher in juveniles; however, for all investigated species, the differences were statistically non-significant. The observed and expected heterozygosity in regenerating juveniles of both B. pendula and F. excelsior were slightly higher than these parameters in putatively maternal trees, while for the other tree species (i.e., Q. robur, P. tremula, and A. glutinosa) the situation was reversed. However, the differences were small and non-significant. The inbreeding coefficient (F) for Q. robur and B. pendula was close to zero. P. tremula and A. glutinosa had negative inbreeding coefficient values, indicating the excess of heterozygotes, while for F. excelsior the inbreeding coefficient value was positive, indicating an increase in homozygotes (Table 3). The regenerating juveniles of all investigated tree species had more private alleles than the putatively maternal trees: in Q. robur, 35 vs. 14; in B. pendula, 51 vs. 15; in P. tremula, 12 vs. 9; in A. glutinosa, 18 vs. 12; and in F. excelsior, 82 vs. 14 private alleles, respectively. The latent genetic potential (LGP) for all investigated tree species was found to be higher in the young generation than in the maternal one. LPG is an important genetic parameter that reflects the aptitude of a population to preserve adaptability under changing environmental conditions [56,57].

In all investigated tree species, most of the genetic differentiation (79%–99%) was found within individuals (F_IS; within species juveniles and putatively maternal trees pooled together), and only a small (i.e., 1% or less) differentiation was found between the regenerating juveniles and putatively maternal trees of each species (F_ST) (

Table 4. Coefficients of genetic differentiation and percentage of AMOVA (%), given in brackets, calculated for eight microsatellite loci in five broadleaved tree species: common oak (Quercus robur L.), silver birch (Betula pendula Roth.), European aspen (Populus tremula L.), black alder (Alnus glutinosa (L.) Gaertn.), and European ash (Fraxinus excelsior L.). F_IS, genetic differentiation within sampled trees of the respective species; F_IT, genetic differentiation among sampled trees of the respective species; F_ST, genetic differentiation (genetic distance) between regenerating juveniles and putatively maternal trees; coefficients of genetic differentiation are given with standard errors (± SE).

Species	FIS (%)	FIT (%)	FST (%)
Q. robur	0.075 ± 0.037 (91.6%)	0.085 ± 0.036 (7.5%)	0.010 ± 0.001 (0.9%)
B. pendula	−0.07 ± 0.053 (98.9%)	−0.058 ± 0.054 (0.0%)	0.011 ± 0.002 (1.1%)
P. tremula	−0.239 ± 0.089 (99.2%)	−0.229 ± 0.090 (0.0%)	0.009 ± 0.002 (0.8%)
A. glutinosa	−0.158 ± 0.059 (99.7%)	−0.151 ± 0.058 (0.0%)	0.006 ± 0.001 (0.3%)
F. excelsior	0.208 ± 0.048 (79.0%)	0.217 ± 0.047 (20.6%)	0.011 ± 0.001 (0.4%)

). Of all investigated species, the genetic difference among the sampled trees (F_IT) was found only in common oak (7.5%) and European ash (20.6%)—the species that regenerated after sanitary clearfelling (Table 4). The low genetic distances (expressed as F_ST) revealed between the sampled regenerating and putatively maternal trees confirm the PCoA results, where the MLGs of regenerating juveniles and of putatively maternal trees overlapped to great extent (Figure 3). All of the investigated species showed a very high gene flow between putatively maternal trees and the young generation (Nm = 26.5, 26.3, 70.7, 47.4, and 26.0 for common oak, silver birch, European aspen, black alder, and European ash, respectively).

Most of the studied microsatellite loci in regenerating trees significantly deviated from Hardy–Weinberg equilibrium (HWE), although the situation varied among the investigated tree species. For example, all loci in European aspen and all but one (Femsatl4) in European ash significantly deviated from HWE, while in black alder most of the loci (A26, A37, A2, A22, and A35) were in HWE.

Bayesian clustering analysis was employed to infer the true number of genetic clusters in each investigated species. According to the ΔK values, the best grouping for all tree species except European aspen was achieved when K = 2 (the aspen MLGs were best grouped when K = 3). STRUCTURE analysis revealed spatially intermixed structures (for both putatively maternal and young regenerating trees) of the genetic clusters for all investigated species, regardless of the most likely number of clusters. To reveal the spatial genetic structures of the regenerating juveniles of each investigated species, spatial distribution maps were constructed (Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8). The spatial distribution maps were created using data on individual tree assignment (calculated by STRUCTURE) to the most likely number of clusters, as revealed by the ΔK criterion. Spatial distribution maps of the regenerating juveniles (Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8) showed no clear spatial structural pattern, where regenerating juveniles from the same circular plot (sampling group) were assigned with different proportions to one of two (or three in the case of European aspen) genetic clusters.

4. Discussion

The results of the present study showed high genetic diversity in the investigated common oak, silver birch, European aspen, black alder, and European ash populations. High genetic diversity in populations of the same tree species has also been reported in Europe by other authors (e.g., [58,59,60,61,62]). The present study showed that despite significant forest ecosystem disturbances in maternal stands, the new generations of all five investigated broadleaved tree species retained high (or exhibited even higher compared to the maternal generation) genetic diversity parameters. Possible loss of genetic diversity in the maternal generations (i.e., in putatively maternal stands) due to natural tree selection during ontogenesis was compensated by recombination and gene flow from more distant (non-sampled) sources. Our findings are consistent with the results obtained by other authors in surveys of the natural regeneration of forest tree species [14,26,63,64]. On the other hand, higher genetic diversity in regeneration tends to decrease throughout stand ontogenesis because of natural selection [20,21,22].

The observed deviation from HWE in the regenerating juveniles indicates an ongoing change in gene or genotype frequencies. The most stable species in this respect was found to be black alder, where only three loci deviated from HWE. The opposite situation was found with European aspen and European ash, where most of the loci significantly deviated from HWE. The HWE can be disturbed by various factors, including mutations, natural selection, non-random mating, genetic drift, and gene flow [13]. The most likely cause for the deviation in aspen and ash could be natural selection—European ash is experiencing high selection pressure due to ash dieback disease (e.g., [65,66]), while European aspen is a pioneer species that readily regenerates vegetatively and largely depends on retaining high levels of heterozygosity in seedlings [67,68]. The positive fixation index found for European ash may be an outcome of inbreeding and selfing due to decreased population size (bottleneck effect). The low fixation indices found for common oak and silver birch indicate random mating.

The results of the present study revealed very small genetic differentiation between regenerating juveniles and putatively maternal trees. All investigated tree species showed a very high gene flow between putatively maternal trees and the young (regenerating) forest generation. This result allows us to expect complete rare allele retention in naturally regenerating populations, along with a subsequent possibility for populations to adapt to rapidly changing climatic conditions (for further discussion, see [69]).

Common oak and European ash—tree species affected by severe natural disturbances and subsequent clearfelling—showed substantial genetic differentiation among the sampled trees (Fit), while in silver birch, European aspen, and black alder no genetic differentiation was detected in this respect. The observed genetic differentiation between mature (putatively maternal) trees and the young generation of common oak and European ash may be a consequence of natural selection for resistance in these tree species, as regenerating juveniles might have originated from seeds of both survived (‘resistant’) and already-cut diseased maternal trees, while the older generation was represented by ‘resistant’ survivors only. Strong natural selection in diseased European ash stands might have reduced the effective population size; however, the results of the present study suggest that even after severe decline the effective population size remains sufficient.

Slightly wider (more scattered) distribution of common oak, silver birch, and European aspen regenerating trees (i.e., their MLGs) compared to the distribution of the putative maternal trees obtained in the PCoA (Figure 3A–C) most likely resulted from gene flow from more distant trees or stands. The MLGs of common oak, silver birch, black alder, and European ash regeneration showed no clear patterns of grouping, even when growing in compact circular plots, and their distribution overlapped to a great extent with the MLGs of the putatively maternal trees, indicating relatedness of regeneration to maternal trees, as well as the absence of pronounced spatial half- and full-sib family structures. Meanwhile, the MLGs of European aspen regeneration showed a clear grouping pattern in most circular plots (Figure 3C and Figure 6). This grouping pattern was expected, as in favorable conditions aspen readily regenerates from root suckers and forms clonal groups.

In all five investigated tree species, random spatial distribution of juvenile MLGs in the circular plots, along with juveniles growing next to one another belonging to different genetic clusters (Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8), indicated an accidental spatial genetic structure of the regenerating/emerging stands. This provides a good potential for the ecological stability of these stands in the future. All five investigated tree species, except for common oak (the seeds of which are dispersed by animals), are wind-pollinated and wind-dispersed; therefore, the spatially random genetic structure of the studied populations was not surprising, having been formed by an intermixed gene flow.

The higher latent genetic potential observed in the young generations reflects good adaptation potential retained in the investigated tree populations under the changing environmental conditions [56,57]. The relatively high LGP shows a high proportion of ‘latent’ genetic potential, which is present—albeit not operational—under actual climatic conditions, and which can play a significant role in future adaptation under changing environmental conditions [5,70]. The assessed effective population size varied for different tree species. For all tree species except P. tremula, the effective population size was high enough to ensure sufficient genetic diversity to be transmitted to future generations [71]. The low effective population sizes detected for P. tremula could be associated with the uneven sex ratio, where male individuals prevailed [72], along with the species’ high potential to regenerate vegetatively and, thus, lower reliance on generative reproduction [73].

Significant positive correlations between plant population size, fitness, and within-population genetic variation have been demonstrated by numerous authors (for references, see [10]). The substantial genetic diversity observed in the young generations of common oak, silver birch, European aspen, and black alder allows us to count on the formation of evolutionarily and ecologically sound stands that will sustain the species’ adaptability and plasticity, population stability, and overall ecosystem functionality. As for European ash, the future of this species will be highly dependent on its ability to develop resistance against the ash dieback disease [23]. The results of the present study showed an excess in homozygotes (as indicated by F in Table 3), suggesting a possible genetic bottleneck in the Lithuanian population of European ash. However, during the last decade of the 20th century, some young, naturally regenerated F. excelsior stands emerged, whose sanitary conditions over the last five years have been more or less satisfactory [74]. This brings some hope for the survival of Lithuanian F. excelsior populations despite the ongoing epidemic.

5. Conclusions

The adequate genetic diversity maintained in the young generations of common oak, silver birch, European aspen, black alder, and even European ash suggests that the regenerating populations of the studied tree species have good chances to adapt to the changing climatic conditions in Lithuania, although the future of the Lithuanian population of European ash is unclear because of the possible genetic bottleneck effect occurring due to the ongoing ash dieback epidemic across Europe.