Structural Heterogeneity of European Beech (Fagus sylvatica L.) Stands at Its Northernmost Limits

Abstract

The expansion of European beech to northeastern regions due to climate change is anticipated, especially if assisted migration techniques are employed. Marginal populations of European beech are exposed to unfavorable growing conditions that are challenging for their survival and multifunctionality. Under such conditions, the structural complexity of stands is a critical factor that supports the sustainability of these populations. In this study, five stands of European beech in Latvia, which are currently the most northeastern stands in Europe, were investigated. In each of the stands, two sample plots (area 500 m²) were randomly established. The dimensions of trees, stem quality features, and spatial structure of the stands were assessed. The stands varied in density but were found to be productive as indicated by comparable tree dimensions to those in core populations. The studied beech stands displayed low species mingling and tended towards monospecies composition, with some structural diversification likely due to small-scale disturbances and varying stand densities, suggesting that spatial diversity was influenced by species composition and competition among trees. The analyzed European beech stands were in the maturing phase, but displayed diverse diameter and height structures, indicating that natural ecological processes were occurring, akin to those found in non-marginal regions. The stem quality of the trees was intermediate, with frequent occurrences of ramicorn, epicormic branches, and forking (41.8%, 53.5%, and 26.3%, respectively), while stem cracks were rare (4.6%). However, these features can provide crucial microhabitats for biodiversity. Therefore, European beech has the potential for diversification in forestry and ensuring sustainability at the edge of its range expansion. The main implications of this study highlight the diverse structural characteristics of the European beech stands, indicating the influence of species competition and small-scale disturbances, providing valuable insights for forest management and conservation strategies. Although, this study has a few potential limitations that should be considered, including the relatively small sample size and the absence of long-term data.

1. Introduction

European beech (Fagus sylvatica L.) is a widespread deciduous tree species [1,2] that dominates most of the forest stands in Central Europe, occupying up to 30% of the forest area [3]. It has high ecological, as well as economic importance [4,5]. Due to its low cost of carbon storage, European beech is often preferred for climate change mitigation [6]. By combining European beech trees with pure coniferous forests, not only productivity can be enhanced, but also the risks caused by disturbances and droughts can be significantly reduced [7]. This dual advantage underscores the ecological resilience of European beech. Furthermore, its significant role in conserving biodiversity further highlights the multifaceted benefits this species brings to forest ecosystems [8]. From an economic perspective, European beech holds significant importance due to its wide-ranging applications in industries such as furniture, veneer, paper production, and fuel [9,10].

The distribution of beech is mainly limited by winter temperature and summer precipitation [3,11,12]. Due to climatic change, both limiting factors are projected to decrease; therefore, northeastern expansion (and southern reduction) of the species distribution is projected [11,12,13]. Accordingly, beech might potentially expand its distribution into the eastern Baltics [10,14] and increase its forestry potential there, particularly if facilitated by assisted migration [15]. The approximate total area of beech stands in Latvia is 52.4 ha (25.9 ha pure stands; 26.5 ha mixed stands), primarily concentrated in Western Latvia, while their presence in Eastern regions is limited by the colder winter temperatures [16]. The forest management in these stands mainly consists of deadwood removal [17].

To promote forest resilience to natural disturbances in the potential distribution area, structural complexity of stands is relevant. It is an important driver for ecosystem stability, forest productivity [7], species richness [18], as well as regulation of microclimate [19].

The formation of multilayered canopies in forests enhances their ability to withstand structural disturbances [20] and provides diverse understory conditions for habitats necessary for different taxa [18,21]. Stand structure primarily is characterized by tree diameter, height, and age distribution within the stand, but also by deadwood amount and microhabitats [22,23]. Abiotic factors may promote damage to trees possibly resulting in ecologically important microhabitats, such as cavities, saproxylic fungi, dendrotelms, and cracks [24]. In addition to the main root, stem, and branch structure of a tree, microhabitats add to the variety of structures on trees. These physical indicators, known as microhabitats, host a substantial amount of biodiversity [22].

The ecological importance of beech often has been related to microhabitats due to high shading, scarce ground cover vegetation, and stem/bark properties [7]. The existence of stem defects in the form of microhabitats, while ecologically beneficial, may reduce the value of timber. Therefore, it is essential to find a balance between economic and ecological benefits to ensure the long-term sustainability of the introduction of beech [22].

In Latvia, which is currently the most northeastern location of the species in Europe, beech was introduced in the middle of the 19th century and became naturalized in the 20th century [10]. Beech stands are productive, implying an increasing forestry potential of the species in Latvia [25]. Although many studies have been published on European beech within its distribution range [26], the characteristics of beech stands near its northeastern distribution range, which are the indicators of global environmental change [27], have been scarcely studied. Species growing on the border of the distribution range are more subjected to changing environmental conditions and thus grow in a more challenging state [27]. This study aimed to describe the structural heterogeneity within second-generation beech stands in Latvia. Stem features linked to production (presence of cracks, ramicorn branches, forking, height of the lowest dry and living branches) were assessed also considering their potential for biodiversity.

2. Materials and Methods

2.1. Study Area

The study was conducted in forests of the Forest Research Station in northwestern Latvia (57°14′ N, 22°41′ E) located beyond the current distribution range of European beech (Figure 1). Data were collected in five plantations with age of the stands ranging from 44 to 69 years (average 56). The studied stands had flat topography (80–105 m a.s.l.) with fertile (mesotrophic) and silty, but well-drained soils (Oxalidosa forest type). The climate was temperate, the mean annual temperature during 1989–2018 was +6.99 ± 0.6 °C; the mean monthly temperature ranged from −2.2 ± 3.0 °C in February to +17.3 ± 1.6 °C in July [28]. The mean annual precipitation was 693 ± 86 mm; the highest monthly precipitation occurred during the summer months (June–August; 72 ± 36 mm) [28]. The size of the studied stands was in the range of 0.2–2.2 ha.

European beech was introduced in Latvia from northern Germany, yet the precise origin of seed material is unknown. The analyzed trees were second-generation trees, propagated from seeds of the first-generation trees. Three to five-year-old seedlings were used to establish the plantations from 1951 to 1982. The trees were planted within 2 × 1.25 m (14_3) and 2 × 2 m grids (19_2, 21_16, 21_30, 21_42) [29]. In stand 14_3, a selective thinning operation took place 2 years before the assessment. In the remaining stands, trees with low vitality or those that had died were continually removed over time, although the exact quantity of wood that was harvested remains uncertain.

2.2. Sampling and Measurements

In each of the stands, two sample plots (area 500 m²) were randomly established. Tree height, diameter at breast height (DBH, accuracy 1 mm), height of the lowest dry and living branches (accuracy 0.1 m) were measured and species were detected for all canopy trees within the plots. Dead trees were also recorded. Visual assessment of beech stem quality features—the presence of ramicorn branches, forking, cracks, and epicormic shoots [10,30]—was carried out. The presence of epicormic branches was recorded at the lowest three meters of the stem. Forking was considered as a splitting of stem into two or more equal axes, which reduces the length of a commercial log’s section [31]. The presence of acutely angled branches up to 50% of height of the stem was recorded as a ramicorn branch, which degrades the commercial value of a log [31]. In total, 524 trees were evaluated (500 beech trees and 24 trees of other species). To describe the spatial structure of the stands, canopy tree coordinates were mapped.

2.3. Data Analysis

Tree aggregation (W_i), species mingling (M_i), and diameter differentiation indices (TD_i) [32] were calculated to describe the spatial stand structure. Nearest neighbor statistics (NNS), a robust analytical methodology that encompasses an in-depth analysis of the spatial arrangement of individual trees, was used [33]. Our study plots were distinguished by the absence of a predefined buffer zone. To mitigate the potential influence of edge effects stemming from neighboring trees beyond the plot borders, a correction method known as NN1 edge cutting was applied [34]. Only trees whose proximity to neighboring trees was less than their distance to the plot perimeter were suitable for inclusion in the calculations. This criterion was implemented to uphold the precision and reliability of the derived indices. The spatial configuration of a stand was characterized through an examination of the relationships between the reference tree and its four closest neighboring trees. This approach, validated by prior research, emerged as the most suitable methodology, particularly when dealing with a limited number of trees under evaluation [34,35,36].

The tree diameter differentiation index (TD_i) describes the heterogeneity of tree size within a group of neighboring trees [33]. This index provides insights into the heterogeneity of tree dimensions within the studied group. Specifically, TD_i values range from low to high, with lower values, nearing 0, referring to a situation with trees of relatively uniform sizes. Conversely, higher values, nearing 1, signify a situation characterized by heterogeneity in tree dimensions [37]. This quantitative assessment offers an understanding of the structural diversity present within the examined stand. The tree aggregation index (W_i) serves as an instrumental parameter for discerning the regularity or irregularity of the horizontal spatial arrangement of trees within the stand [33]. Similar to TD_i, W_i values encompass a spectrum from low to high. Lower values, approximating 0, are indicative of a more structured and regular placement of trees, whereas higher values, nearing 1, point to a pattern characterized by a less structured, more irregular grouping of trees [37].

The species mingling index (M_i) describes spatial relationships between species of neighboring trees (species diversity). Furthermore, the species mingling index (M_i) stands as a measure for revealing the complex spatial relationships between different tree species within the stand, ultimately providing a measure of species diversity and cohabitation [33]. Lower values, approaching 0, signify a scenario characterized by a relatively homogenous distribution of tree species. Conversely, higher values, approaching 1, denote a situation wherein diverse tree species coexist in spatial proximity [37].

To allow for comparison among different stands, the relative height of the lowest living and dry branches in relation to tree height was calculated. This metric provided a standardized assessment that allowed for cross-stand evaluations. The mean marginal probability of occurrence of cracks, forking, ramicorn, and epicormic branches was estimated based on a simple linear mixed-effects model for living beech trees only. This model operated under the assumption of a binomial distribution of residuals and was linked by means of the logit function. Stand was used as the fixed effect and sample plot as the random effect [38]. The G-test was employed to analyze the differences in class values across various stands. This statistical test helps to determine if there are significant disparities in the distribution of categorical variables among the different stands. Data analysis was carried out in R v.4.1.3 [39] using the package “lme4” [38].

3. Results and Discussion

The studied beech stands had a slight admixture of pedunculate oak (Quercus robur L.), Norway spruce (Pice abies (L.) Karst.), and goat willow (Salix caprea L.), which were 4.6% of the recorded trees. The degree of admixture differed among the stands and some were even lacking it; therefore, the species mingling index was low, in the range of 0.00–0.08 (Figure 2A). Considering the age of the stands, the amount of deadwood was still negligible, as only 2.1% of the recorded trees were dead, most of them were Norway spruce. Accordingly, this confirms that beech tends to form monospecies stands [15]. Although the beech stands were planted regularly, structural diversification has been occurring likely due to small-scale disturbances [24], as indicated by intermediate tree aggregation indices (Figure 2A), which exceeded the expected values for a plantation. This process has been similar among the studied stands irrespective of their age, as suggested by comparable values of the index.

The density of the studied stands ranged widely (570–1520 trees ha⁻¹) due to differences in planting density and thinning (

Table 1. Characteristics of the stands. DBH—the mean diameter at breast height (cm), H—the mean height (m).

Stand No.	Area, ha	Density, Trees ha−1	Stand Age	DBH, cm	H, m	Basal Area, m2 ha−1
14_3	2.2	630 ± 71	69	27.6 ± 2.2	27.9 ± 1.7	41.4 ± 7.0
19_2	0.2	1520 ± 255	59	24.2 ± 1.8	22.0 ± 1.0	84.6 ± 10.2
21_16	1.7	570 ± 156	44	23.4 ± 3.4	23.7 ± 0.9	31.4 ± 6.7
21_30	0.4	1480 ± 57	65	19.4 ± 1.2	25.4 ± 1.7	49.9 ± 2.8
21_42	4.1	930 ± 71	44	20.0 ± 1.5	22.6 ± 0.6	35.1 ± 6.3

). The mean stand density was 1026 ± 122 trees ha⁻¹, which can be considered somewhat higher than the considered optimum for beech in the mid-part of the distribution range [33]. The wide range in stand density, influenced by factors like planting density and thinning, suggests the potential for optimizing tree stem characteristics for wood production through pre-commercial thinning [31]. Compared to stand density, the mean DBH in all stands showed a low range (19.4 ± 1 cm to 27.6 ± 2 cm) implying that the studied stands were representing a range of growing conditions. The mean DBH was 22 ± 4 cm, which is similar to that in the central part of the distribution range [40] implying favorable growth conditions outside the current distribution range.

The distribution of DBH differed between the stands. One of the stands with the youngest age (21_16) represented the most diverse diameter classes, none of which dominated, indicating patterns similar to natural stands [41]. This was also highlighted by the highest tree diameter differentiation index (0.46) which can be considered intermediate (Figure 2A). Yet the other stands, which were the same age or older, had lower diameter diversity (Figure 3A). These stands were dominated by trees with intermediate DBH (11–20 cm and 21–30 cm) likely due to higher density, and thus competition [7]. The lowest tree diameter differentiation index was estimated for the thinned stand 14_3 (0.27) (Figure 2A). The basal area correlated with the stand density, yet the mean basal area was 48.5 m² ha⁻¹, which can be considered as a high basal area compared to results from northeastern France [40]. For comparison, the basal area for 60-year-old mixed (Abies alba Mill.) beech stands growing in optimal ecological conditions ranged from 37.9 m² ha⁻¹ to 41.8 m² ha⁻¹ [37]. Stands with smaller tree dimensions in Sweden had a basal area ranging from 5.17 to 22.68 m² ha⁻¹ [38].

Tree height, which is a proxy for stand productivity [42], was similar within the stands ranging from 22.0 ± 1 to 27.6 ± 2 m, implying comparable growth potential irrespectively of differences in DBH. The mean overall height was 24 ± 4 m, exceeding that under optimal conditions [37], thus highlighting the forestry potential for beech in Latvia. Although the analyzed stands were planted, the distribution of tree height indicated that stratification of canopy layers has occurred, yet its degree differed among the stands. While some stands had a single layer canopy (e.g., 21_42), in other stands two layer canopy was already shown by two dominant height classes (Figure 3B, e.g., 19_2 and 21_30). From an economic perspective, this is undesirable as part of the trees growing in a multilayer canopy have reduced growth. Yet it is ecologically important, as it promotes different structures and microhabitats in the stand. The multilayer canopy formation and lower growth performance in the lowest layers indicate the necessity for thinning. Generally, such diameter, height distribution, and basal area indicate an economic potential for a stand; as in our case, the stands by the age of 69 years were dominated by valuable assortments [10]. Furthermore, for similar stand densities, larger diameters have been observed in stands in Latvia compared to Sweden [43].

The stem quality of beech in the studied stands was intermediate, as indicated by the height of living and dry branches as well as by the probability of stem features. The height of the lowest living and dry branches, which is an important characteristic of timber value and quality [44], ranged between 7.5–12.8 m and 6.1–13.6 m, comprising 41.1% and 15.3% of the tree height, respectively (Figure 3C). These values exceed those of somewhat younger trees in stands in the mid-part of species distribution (5.8 m) [31], suggesting a higher proportion of higher value assortments acquirable from a tree [10]. Although the initial stand density can affect the height of the lowest dry and living branches [44], such effects were not visible in the studied stands (Figure 3). The height of the lowest dry branch was not influenced by either the stand, the diameter, or the combination of these two factors (p > 0.05). On the other hand, the height of the lowest living branch is influenced by the diameter (p < 0.05), but not by the stand or the interaction between the stand and diameter (p > 0.05). A larger diameter suggests a stronger, more robust tree with a greater capacity to support lower branches. This is often because larger trees tend to have more resources available for growth, which can potentially allow for the development of lower living branches.

Frost cracks that are considered a particular issue under cool climates, as they permanently damage beech timber [45] and promote rot infections [10], were recorded on 4.6% of trees and were affected by stand conditions as they were in the range of 1.9–4.7% among the stands. This frequency can be considered low compared to the results in Romania, yet these stands were older [44]. On the other hand, frost cracks are important microhabitats for spiders, birds, and bats [46]. The probability of trees having cracks previously has been positively linked to forking [12], yet such a relationship was not evident (Figure 2B). The likelihood of a tree developing a crack appears to be affected by the stand (with a p-value of 0.04). However, when stands were compared, no significant differences were found.

The analysis did not show a significant correlation between tree diameter and the likelihood of having cracks, although it was indicated by wood density (where smaller annual tree growth corresponds to higher density) that these trees should generally have a lower probability of developing cracks, except in certain exceptional cases [47]. While there is an overall effect of stand on crack development, the individual stands in this study may be relatively similar in factors influencing crack formation, potentially making it challenging to distinguish significant differences. Additionally, the small sample size or high data variability could hinder the detection of distinct stand differences.

When the main goal is timber production (especially high value assortments), forking and ramicorn branching are subjects of negative selection as they reduce the amount of saw logs and increase knottiness [10,48]. Although, from an ecological point of view, forking is a valuable trait as it promotes dendrotelm formation between forks, serving as a habitat for a complex of related species [47]. In the studied stands, forking was quite frequent—on 26.3% of trees on average. Nevertheless, the frequency of forking differed by stand conditions as indicated by the frequency in the range of 17.5–35.7% among the studied stands. The likelihood of a tree having forking is influenced by the stand (p < 0.05). The stands 14_3-21_30 and 21_30-21_42 exhibited distinctions from one another in this regard. The reasons for these findings could be influenced by a combination of factors, including varying environmental conditions, competitive interactions, and potentially unmeasured variables that collectively contribute to the observed differences in forking frequency across the studied stands [48].

Ramicorn and epicormic branches were the most frequent stem features in the studied beech stands being present on 41.8% and 53.5% of trees on average, respectively, which is similar to other studies [48]. From an ecological point of view, ramicorn and epicormic branches are considered microhabitats important for the nesting of small birds [48]. The occurrence of epicormic branches was more affected by stand conditions (p < 0.05) compared to ramicorn branches, as indicated by the high range among the stands (26.9–66.8% and 31.7–52.0%, respectively). The occurrence of ramicorn branches has been negatively related to the stand density due to self thinning [49]; however, stands with the highest density (19_2; 21_30) did not have the lowest probability of trees having ramicorn branches (Figure 2B). Although the formation of epicormic branches has been linked to thinned stands [31], the thinned stand (14_3) had the lowest probability of epicormic branches (Figure 2B). This might be explained by the short period that had passed since the thinning or relatively low intensity of management [49]. The probability of a tree being free from defects is affected by the stand, with statistical significance (p < 0.05). Notably, stands 14_3 and 19_2 exhibit a difference, which is likely attributed to the thinning operations carried out in stand 14_3, where trees with defects were likely harvested.

This study emphasizes the varied conditions within these stands. The comprehensive examination of tree diversity, structure, and productivity relationships across long-term experimental plots and various spatial scales presents a promising avenue for advancing forest management strategies [50]. To draw specific conclusions, it is necessary to include an additional dataset with various influencing parameters like stand density, tree age and diameter, age, etc. being represented. However, the current study offers valuable initial insights.

4. Conclusions

The studied stands, situated beyond the established range of European beech, displayed comparable productivity and structural diversity to core populations, indicating a northeastward shift in the species distribution. This suggests a potential for diversification in forestry practices, bolstering sustainability at the expansions edge. The noted intermediate quality of the beech stems in the examined stands shows potential for acquiring more valuable assortments. The relatively low incidence of frost cracks, influenced by stand conditions, showcases the stands’ resilience to timber damage, a crucial aspect of sustainable forestry practices. The prevalence of forking underscores its ecological importance, fostering dendrotelm formation and creating vital habitats for associated species. This ecological role underscores the broader significance of structural features in supporting biodiversity within these forest stands. For future investigations, emphasis should be placed on long-term studies to monitor changes in structural diversity and biodiversity over time. Expanding research to diverse regions will enable assessments of the influence of climate, soil, and management practices. Additionally, in-depth studies on the ecological roles of microhabitats will contribute to a deeper understanding of overall biodiversity and ecosystem functioning. There are a few potential limitations to consider in this study, such as the relatively small sample size and the lack of long-term data. While these limitations may affect the generalization of the study findings to other regions, they highlight the need for additional research in this field.