Long-Term Cumulative Effect of Management Decisions on Forest Structure and Biodiversity in Hemiboreal Forests

Abstract

We evaluated the long-term impacts of various forest management practices on the structure and biodiversity of Estonian hemiboreal forests, a unique ecological transition zone between temperate and boreal forests, found primarily in regions with cold winters and moderately warm summers, such as the northern parts of Europe, Asia, and North America. The study examined 150 plots across stands of different ages (65–177 years), including commercial forests and Natura 2000 habitat 9010* “Western Taiga”. These plots varied in stand origin—multi-aged (trees of varying ages) versus even-aged (uniform tree ages), management history—historical (practices before the 1990s) and recent (post-1990s practices), and conservation status—protected forests (e.g., Natura 2000 areas) and commercial forests focused on timber production. Data on forest structure, including canopy tree diameters, deadwood volumes, and species richness, were collected alongside detailed field surveys of vascular plants and bryophytes. Management histories were assessed using historical maps and records. Statistical analyses, including General Linear Mixed Models (GLMMs), Multi-Response Permutation Procedures (MRPP), and Indicator Species Analysis (ISA), were used to evaluate the effects of origin, management history, and conservation status on forest structure and species composition. Results indicated that multi-aged origin forests had significantly higher canopy tree diameters and deadwood volumes compared to even-aged origin stands, highlighting the benefits of varied-age management for structural diversity. Historically managed forests showed increased tree species richness, but lower deadwood volumes, suggesting a biodiversity–structure trade-off. Recent management, however, negatively impacted both deadwood volume and understory diversity, reflecting short-term forestry consequences. Protected areas exhibited higher deadwood volumes and bryophyte richness compared to commercial forests, indicating a small yet persistent effect of conservation strategies in sustaining forest complexity and biodiversity. Indicator species analysis identified specific vascular plants and bryophytes as markers of long-term management impacts. These findings highlight the ecological significance of integrating historical legacies and conservation priorities into modern management to support forest resilience and biodiversity.

1. Introduction

The concept of forest ecological memory includes variety of modifications (called forest legacies) generated by both natural disturbances and anthropogenic management along with temporal aspects [1,2]. Historically there has been a significant shift in forest ecosystems from natural disturbances to more intense human-driven management approaches such as clear-cutting, selective harvesting, plantings, and understory maintenance [3]. Nowadays more forest land is protected [4] and clear cutting is transitioning into selective cutting or continuous cover forestry (CCF) [5]. CCF is not a new idea in forest management but there has been renewed interest in it for sustainability requirements [6,7]. It has also been found that sustainable forest management needs more region- and forest site type-specific targets [4,8].

Using historical data to quantify environmental impacts continues to be controversial [9], because of uncertain spatial accuracy, dates, and low image quality [10]. On the other hand, implications of historical management on forest structure and biodiversity are undeniable [11,12]. For example, ecosystem management has been shown to lead to retrogressive succession [13,14] and a simplified forest structure [15] but historical (ancient) semi-natural habitats, such as woodlands or grasslands, can support diverse communities and are key elements for biodiversity [16,17,18].

Already two decades ago, researchers recommended that anthropogenic disturbances in mature and old forest stands should receive more attention compared to the more extensively studied stand replacement cutting and natural disturbances [19,20,21]. Intensive structural and compositional changes also occur during the (re)establishment and growth of the stand [22], overshadowing the minor effects of internal stand modification and maintenance activities [23]. The manipulation of forest density, tree species composition, and other structural properties impact forest ecosystem complexity and various vegetation layers [24,25], and these effects are forest type-specific [26]. Forest management activities not only decrease habitat quality through the reduction in deadwood -a critical structural element for many species [27]—but also alter the species composition of field layer vegetation, bryophytes, lichens, and wood-inhabiting polypore fungi [26,28,29].

Estonia’s forest ecosystems have undergone significant management transitions since the late 19th century. The study seeks to clarify the accumulating effect of forest management practices over the life-cycle time of an average stand in Estonia, tracing a cascade of management decisions from the period of the Russian Empire (beginning of the 20th century) to the current framework under the Republic of Estonia (Figure 1, Appendix A).

Figure 1. The timescale (state, year) and data (data sources) used for classifying (origin, management, conservation) study areas in Estonia.

Furthermore, Estonia’s forest management has evolved from local collective farm and forest district management to a broader sector-based approach. For example, initially young forests (up to 20 years) undergo early tending and precommercial thinning [30], most systematically in planted and sowed conifer stands. This approach has started to evolve into more flexible cleaning and precommercial thinning of young stands, including those with naturally regenerated deciduous trees. Historical practices, such as seed tree harvest and selection thinning (developed by [31]), have been replaced by sanitation cutting across all ages and the more active use of clearcut and shelterwood silviculture in mature stands [32,33]. Recreational forest use has also gained prominence [34], with hiking and outdoor vacations complementing traditional activities such as berry/mushroom picking and herb gathering.

The Estonian Environmental Strategy 2030 indicates a complex turn towards sustainability, biodiversity conservation, and multifunctional forest use, marking a significant transformation need in Estonia’s forest stewardship. The implications of this transition are necessary, not just for the sustainability of forest resources but also for Estonia’s socio-economic landscape, setting a guideline for forest management in similar biogeographic contexts. Using a historical perspective, we examine the possibilities to shift from even-aged management to diverse, conservation-oriented strategies that align with modern ecological principles.

The first protected area in Estonia was established in 1910, the forest-oriented areas much later, and the establishment of new areas has continued through the 20th century [35]; however, in our study areas, the first conservation regulations began in 1981. Historically, forest protection zones were primarily managed at the district level, where forest land varied from 2000–4000 ha. Large nature preserves and landscape protection areas imposed various mild management restrictions. In addition to generic nature reserves and protection zones, specialized protected areas were established with the focus on water resources, natural maintenance, and key habitat protection. Since the 2000s, many areas have been reclassified as conservation zones or key habitat protection zones, bringing more specific regulations regarding cutting limitations and usage restrictions in sensitive areas such as road and water protection zones, recreational forests, and reserve coupes. The Nature Conservation Act (2004) categorizes protected areas into strict nature reserves, conservation zones, and limited management zones, specifying management restrictions in each.

In examining the long-term effects of forest management practices on the structure and biodiversity of Estonian hemiboreal forests, we categorize our study areas based on their stand origin, historical management, recent management, and conservation status (Figure 2).

Figure 2. Study outline. Data (2015–2023) from 150 forest plots (blue) across various site types (white) and supplemented by historical data (orange) were used to classify plots into categories of stand origin (grey), historical management (brown), recent management (yellow), and conservation status (green).

Our study addresses the critical knowledge gap in understanding how long-term management practices, including historic and recent interventions, influence forest biodiversity and structure in Estonian hemiboreal forests. We hypothesize that multi-aged forests, as opposed to even-aged ones, will display higher biodiversity and structural complexity due to increased ecological continuity and varied habitat conditions. Additionally, we hypothesize that recent forest management will reduce biodiversity, while conservation practices will support higher deadwood volumes and species richness. We focused on the following research questions:

• How does multi-aged forest management influence biodiversity and forest structure compared to even-aged forests?
• What is the impact of historic management practices on current biodiversity and structural elements?
• How do recent management interventions affect deadwood volumes and species composition?
• How does conservation status contribute to biodiversity preservation in Estonian forests?

These objectives are essential for guiding future forest management strategies that balance production, conservation, and biodiversity goals.

2. Material and Methods

2.1. Study Region and Sample Plots

The study region is situated in eastern and southern Estonia (Figure 3). Estonia belongs to the hemiboreal vegetation zone [36]. The average annual precipitation is 550–750 mm per year⁻¹, with average temperatures ranging from 17 °C in July to −5 °C in February [37]. These forests are characterized by a mix of deciduous and coniferous trees, often including species like spruce, pine, birch, and aspen. Hemiboreal forests support a diverse range of flora and fauna, offering habitats that blend species typical of both boreal (northern) and temperate zones. This transitional nature of hemiboreal forests makes them particularly sensitive to environmental changes, offering a valuable indicator of ecological shifts due to climate and land-use changes. A recent study (2015–2023) selected 150 forest sites (plots with 15–30 m radius) in multiple forest areas. They belong to the Estonian Network of Forest Research Plots [38]. Study plots were located within each forest compartment, representing specific forest site type and different combinations of management histories (Figure 2 and Figure 3). The recent study focused on three high-productivity forest site types: Oxalis (55 plots), Oxalis-Myrtillus (hereafter Ox-Myrt) (43 plots), and Oxalis-Rhodococcum (hereafter Ox-Rhod) (52 plots) [39,40]. Stands were limited to being at least 65 years old (average tree species age on plots varied from 65–177 years). The plots were then categorized by representation of differently managed forests and current conservation states (Figure 2).

Figure 3. Location of the studied sample plots in Estonia. The map shows the geographic distribution of the 150 sample plots. Legend codes represent various management and conservation factors. O indicates stand historic origin, where 0 (solid fill) represents mixed-aged stands and 1 (pattern fill) represents even-aged stands. H represents historic management, with 0 (circle) for not managed and 1 (triangle) for managed. R stands for recent management, with 0 (green fill) for not managed and 1 (red fill) for managed. C represents conservation status, where 0 (red edge) indicates commercial forest and 1 (green edge) represents protected area.

In each plot, forest stands were characterized using the methodology of the Estonian Network of Forest Research Plots (ENFRP) [38]. Field works were carried out from June to September. Trees (including standing dead trees and broken dead trees at h > 1.3 m, i.e., snags) with a diameter at breast height (DBH) over 4 cm were recorded with the species, DBH, and height. In addition, all downed dead trees (logs) with a diameter over 10 cm at stump end were measured.

The sub-plot was positioned at the center of the stand plot. Pin-points were taken circularly extending from the center towards the perimeter at 1-m intervals. Ground-dwelling species of bryophytes and vascular plants were recorded. Their taxonomy followed the national reference textbooks [41,42]. Unidentified species were analyzed in the laboratory of the Estonian University of Life Sciences. Later, data on tree seedlings and bush species were excluded from the herb (field) layer data because they were also recorded in the forest understory data. In total, 199 vascular field layer plant species and 103 bryophyte species were identified.

2.2. Management History Assessment

Historical and recent forest management practices were assessed for the period from 1884 until recent survey (2015–2023 depending on plot). Management activities were categorized into binary variables to facilitate robust analysis (see Figure 2, Appendix A, Supplementary Materials). This approach reflects a simplification, acknowledging the continuum of management intensities; however, potential details would fall within the limits of main management steps, such as initiating the stand, maintenance of the stand for tree growth, and conservation. This historical assessment utilized a variety of sources (Appendix B), including historical maps, aerial photographs, and forest planning documents from the Estonian State Forest Center [43] and National Archives of Estonia [44]. Insights were improved by interviews with local forestry specialists (retired and working).

Management activities identified from State Forest maps and interviews ranged from early tending to selective cutting (Appendix A). For analysis, these activities were divided into the categories detectable and undetectable from aerial photographs. Activities such as large-scale cuttings were readily apparent on maps, contrasting with refined human interventions in forest stands that emerged from interview data. In our analysis, we focused on management actions detectable in aerial photographs, excluding the undetectable ones (Appendix A).

The timeline between historical to recent management was set to the 1990s to reflect sharp changes in the governmental system and forest management in Estonia. There was a significant shift from usage of clear-cuttings and wider use of forestry machines instead of manual labor. Alongside these shifts, forest conservation policy was revised following Estonia’s restoration of independence and joining the EU Natura 2000 legislation area, leading to an increase in strictly protected areas (Figure 1).

Management and conservation information was classified into categorical variables with four levels (Figure 2). Plots can be characterized either as commercial forests (55 plots) or protected sites (95 plots), including areas within the Natura 2000 network in Estonia. All the plots (96) in Natura 2000 forest sites represented the 9010* “Western Taiga” habitat type. Conservation information was categorized to reflect both commercial and protected areas. Protected sites are unmanaged according to the Nature Conservation Act (2004), which includes strict nature reserves and wilderness conservation areas. Commercial sites also include protected areas that permit some forms of forest management activities (limited management zones). We would like to note that the Natura 2000 habitat plots were surveyed in 2015, and by 2018, some of them (10 plots) were also managed as commercial forests. Currently, these Natura 2000 sites in Estonia are designated as Sites of Community Importance (SCI) but are expected to be reclassified as Special Areas of Conservation (SAC) [45]. The current management and protection statuses of the plots were used at the time (2015–2023) of recent survey. Classifications and main characteristics with found management histories for each plot can be found in Supplementary Materials.

2.3. Data Analysis

To explain the ecological requirements of vascular plants and bryophytes, we applied Ellenberg [46] indicator values. These values were estimated as community-weighted means for light and moisture requirements for both groups and for soil fertility value for vascular plants. Pin-point counts were used as abundances.

The structural characteristic of the forest stands and estimates of species richness were analyzed using a general linear mixed model (GLMM) [47]. The GLMM estimation using the Type I test was applied to test the cascading effect of factors, starting from the forest site type (as environmental envelope), stand origin, historic management, recent management, and ending with the present conservation status. Forestry region was included as a random factor to address spatial clustering of study sites and management styles within historic and present forest districts. Post hoc comparison analysis of mean estimates within factor were conducted using Tukey’s multiple comparison test [48]. Analyses were preformed using the MIXED procedure implemented in SAS version 9.2 (SAS Institute Inc, Cary, North Carolina). To ensure comparability across variables, all continuous variables were standardized before analysis. The standardization allowed us to scale the predictors and the response variable appropriately, and while this typically constrains the effect sizes to a range between −1 and 1, certain variables exhibited strong associations, resulting in effect sizes and error bars exceeding this range. These larger effect sizes reflect the strong biological relationships between key environmental and management factors and the forest structure metrics under study. We assessed multicollinearity among the predictor variables using the variance inflation factor (VIF). All VIF values were below 5, indicating no significant multicollinearity. This ensures that the predictor variables are sufficiently independent, allowing for reliable coefficient estimation.

Non-metric multidimensional scaling (NMDS) was chosen as the ordination method to elucidate patterns in species composition. The species dataset included pinpoint counts representing the abundances of each species. Species compositional patterns in relation to stand origin, management, and conservation regimes were investigated using a multi-response permutation procedure (MRPP) [49]. In both analyses, the Bray–Curtis dissimilarity distance was applied on raw data, but the Euclidean distance was used on the species-plot semi-residual matrix, where the effect of site type and region was removed. Indicator species analysis (ISA) [50] was utilized to detect differences between same factors, with indicator values assessed for statistical significance through Monte Carlo permutation tests (1000 runs).

The NMDS, ISA, and MRPP analyses were executed using PC-ORD version 7.1 [51].

2.4. Manuscript Preparation

Generative AI technology was used in the preparation of this manuscript to assist with language editing, grammar correction, and structural refinement. Specifically, OpenAI’s ChatGPT was employed to enhance the clarity and readability of the text, ensuring grammatical accuracy and consistency in terminology. No AI-generated content replaced the author’s original scientific insights, data interpretations, or conclusions. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

3. Results

3.1. Environmental Envelope

The NMDS ordination (stress factor = 14.62, p = 0.004; Figure 4) resulted in a two-dimensional solution, capturing a significant portion (I axis 78%, II axis 10%) of the variation in field layer using raw logarithm data of vascular plants and bryophytes with species frequency > 3 on the plot (n = 204) and 25 environmental variables for species composition. The first axis is correlated with the plants’ requirements for soil fertility and light availability and the second axis is correlated with the plants’ requirements for soil moisture—these are conditions well related to the studied site types. It points out that the effect of forest site type on the analyzed species and structure is stronger and should be taken into account in the interpretation of the effects of the stand origin, management, or conservation (Figure 5 and Figure 6; Appendix F).

Figure 4. NDMS varimax ordination of 150 sample plots based on vascular plants and bryophytes. The first axis explains 78% of the variance (p = 0.004), while the second axis accounts for 10% of the variance (p = 0.004). The ordination was performed using raw logarithm data for species with a frequency >3 per plot (n = 204 species) and 25 environmental variables (see Appendix C for full list). Only environmental factors significantly related to the ordination axes (p < 0.05) are shown, with a cut-off of R² = 0.2 for vector inclusion. Plots are color-coded by site type: blue for Ox-Myrt, red for Ox-Rhod, and green for Oxalis. The pNDMS ordination without the effects of site type and region is available in Appendix F. The final stress value of the ordination is 14.62225, indicating the goodness of fit.

Figure 5. Heatmap showing GLMM analysis results (p-values) for each structural trait in different categories. The significant difference between sites was tested using the Type I model for structural traits. More detailed results in Appendix C.

Figure 6. General Linear Mixed Model (GLMM) Type I tests to evaluate the effects of interventions (stand origin, historic and recent management, conservation) on the traits (structural features and biodiversity components) of forest stands). Each point indicates the mean effect size with its confidence interval, showing influence across traits. Variance inflation factor (VIF) values for all predictors were below 5, indicating no multicollinearity among the independent variables.

3.2. Influence of Stand Origin on Forest Structure and Biodiversity

The results of General Linear Mixed Model (GLMM) analysis show that stand origin type predicts some forest structural and biodiversity features (Figure 5 and Figure 6). Specifically, mixed-aged forests had differences in average diameter of canopy trees (GLMM, p < 0.0001), a 16% smaller proportion of pine and 12% greater proportion of spruce in the stand, and a 17.8 m³/ha greater volume of lying deadwood (GLMM, p = 0.011) compared to even-aged stands (Appendix C, Figure 5). The basal area of canopy trees was 4.6 m² higher in mixed-aged stands. Also, bryophyte species richness was 4.9 species smaller in even-aged stands (GLMM, p = 0.0003).

The MRPP test (Figure 7) also showed differences in species composition between mixed-aged and even-aged forests’ origin (T = −7.1, p < 0.001). ISA (Appendix D; Appendix E; Table 1), conducted for each site type separately, identified species such as Dicranum majus, Rhizomnium punctatum, and Dicranum heteromalla with higher frequency in multi-aged forests (ISA, p < 0.05). On the other hand, species such as Melampyrum pratense thrived in even-aged stands (ISA, p < 0.001).

Figure 7. Multi-response permutation procedure (MRPP) results comparing species composition by different management regimes for stand origin, historic management, recent management, and conservation, with test statistic (T) and agreement (A) values.

Table 1. Summary of ISA (Indicator Species Analysis) results (p < 0.05) for different forest management regimes. The table summarizes species significantly associated with different management regimes for each habitat type. Habitats are indicated with superscripts: M for Ox-Myrt, O for Oxalis, and R for Ox-Rhod. For more detailed results of the ISA, please refer to Appendix D.

Factor: Historic Origin
Level 0: Multi-Aged	Level 1: Even-Aged
Vascular plants: Goodyera repensR, Impatiens parvifloraO, Melampyrum nemorosumMR, Moehringia trinerviaO, Orthilia secundaO, Vaccinium vitis-idaeaO, Veronica chamaedrysO	Vascular plants: Calluna vulgarisR, Daphne mezereumO, Dryopteris expansaO, Galeobdolon luteumO, Galium odoratumO, Lathyrus vernusO, Milium effusumO, Pulmonaria obscuraO, Stellaria nemorumO, Viola rivinianaO
Bryophytes: Brachythecium oedipodiumOR, Brachythecium salebrosumO, Cirriphyllum piliferumO, Dicranum heteromallaM, Dicranum majusMR, Dicranum montanumOR, Lophocolea heterophyllaR, Nowellia curvifoliaOR, Plagiothecium curvifoliumR, Plagiothecium laetumR, Ptilidium ciliare, Ptilium crista-castrensisO, Ptilidium pulcherrimumO,R, Rhizomnium punctatumM, Sanionia uncinataR, Tetraphis pellucidaO,R	Bryophytes: -
Factor: Historic Management
Level 0: Not Managed	Level 1: Managed
Vascular plants: Molinia caeruleaM, Lathyrus vernusO	Vascular plants: Angelica sylvestrisM, Carex vaginataM, Convallaria majalisM, Lycopodium annotinumM, Melampyrum pratenseO, Moehringia trinerviaO, Mycelis muralisO, Orthilia secundaM, Rubus idaeusO
Bryophytes: Hypnum cupressiformeM	Bryophytes: Brachythecium oedipodiumM, Cirriphyllum piliferumM, Plagiochila asplenioidesO, Pleurozium schreberiO
Factor: Recent Management
Level 0: Not Managed	Level 1: Managed
Vascular plants: Calluna vulgarisR, Deschampsia flexuosaM, Festuca ovinaR, Fragaria vescaR, Goodyera repensR, Orthilia secundaO	Vascular plants: Aegopodium podagrariaM, Anemone nemorosaM, Angelica sylvestrisM, Athyrium filix-feminaM, Calluna vulgarisR, Carex digitataM, Carex vaginataM, Convallaria majalisM, Crepis paludosaM, Deschampsia cespitosaM, Deschampsia flexuosaR, Dryopteris filix-masO, Equisetum pratenseM, Equisetum sylvaticumM, Fragaria vescaM, Hepatica nobilisM, Orthilia secundaM, Rubus saxatilisM, Solidago virgaureaM
Bryophytes: Dicranum majusM, Dicranum montanumR, Lophocolea heterophyllaM, Nowellia curvifoliaR, Polytrichum communeM, Ptilidium pulcherrimumR, Sphagnum russowiiM	Bryophytes: Brachythecium oedipodiumR, Cirriphyllum piliferumM, Dicranum majusR, Dicranum montanumR, Dicranum scopariumR, Lophocolea heterophyllaR, Nowellia curvifoliaR, Plagiomnium affineM, Plagiomnium ellipticumMO, Plagiothecium curvifoliumR, Plagiothecium laetumR, Ptilidium ciliareR, Ptilidium pulcherrimumR, Rhodobryum roseumM, Sanionia uncinataR, Tetraphis pellucidaR
Factor: Conservation
Level 1: Protected	Level 0: Commercial
Vascular plants: Pteridium aquilinumO, Orthilia secundaO	Vascular plants: Dryopteris filix-masO, Impatiens parvifloraO, Stellaria nemorumO, Urtica dioicaO
Bryophytes: -	Bryophytes: Eurhynchium angustireteO, Plagiomnium ellipticumO

3.3. Impact of Historical Management

Forests with historical management showed higher tree species richness (GLMM, p < 0.01) and lower deadwood volumes (GLMM, p < 0.05) relative to historically unmanaged forests.

The ISA (Table 1) revealed significant associations between historically managed forests and certain species. Notable indicator species (Table 1, Appendix D) for historically managed forests was Angelica sylvestris (ISA, p < 0.05) and Calamagrostis arundinacea (ISA, p < 0.01). The bryophyte Cirriphyllum piliferum also showed high indicator values for historically managed forests (ISA, p < 0.01). Conversely, Brachythecium oedipodium and Hypnum cupressiforme (ISA, p < 0.05) implied their preference for more undisturbed conditions.

3.4. Effects of Recent Management

Understory tree species richness and basal area of spruce in sub-canopy trees showed a significant increase due to recent management activities (GLMM, p < 0.01). All variables connected to tree volume or basal area obviously showed lowering effects (Figure 6) under recent management (GLMM, p < 0.01). Also, all deadwood volumes were decreasing within plots where management after the 1990s was detected (GLMM, p < 0.01) (Appendix C).

ISA (Table 1, Appendix D) for recent management showed several species with strong relationships to recently managed forests. This suggests that management activities such as thinning, selective logging, and other interventions have an impact on analyzed species distributions. For instance, Anemone nemorosa showed a strong preference for recently managed areas, with a frequency of occurrence in these plots of 83% (ISA, p < 0.001). In the case of bryophytes, Dicranum scoparium was also found lot in recently managed forests, with a frequency of 94% (ISA, p < 0.001).

Within the management, the disparity was similar between historically (MRPP, T= −3.2, p < 0.01) and recently (MRPP, T= −3.4, p < 0.01) managed vs. unmanaged forests (Figure 7).

3.5. Protected Areas Outcomes

Protected areas exhibited higher average diameter of canopy trees compared to commercial forests, indicative of the positive impact of these practices on preserving larger trees (GLMM, p < 0.01). The volume of lying deadwood was significantly higher (GLMM, p < 0.0001) in protected areas, averaging 59.1 m³ ha⁻¹ (Figure 5, Appendix C) compared to much lower average 22.3 m³ ha⁻¹ in commercial forests. Similarly (GLMM, p < 0.01), the total average volume of deadwood (77.6 m³ ha⁻¹) in protected areas was 88% higher than in commercial forests (41.2 m³ ha⁻¹). Bryophyte species richness was slightly higher in protected areas (GLMM, p < 0.01), contrary to the decrease in vascular plant species richness (GLMM, p < 0.05).

Indicator Species Analysis (ISA) also provided insights into how species that are indicative of conservation areas reflect the protective management regime’s impact on maintaining or increasing biodiversity within these forest ecosystems. Melampyrum pratense and Hylocomium splendens were significantly associated with conservation areas, showing a high frequency of 100% (ISA, p < 0.05). These species are quite usual in Estonian forest ecosystems and despite finding protected species on some protected sites they did not occur in our ISA results. Notably, the species composition of commercial forests exhibited significant divergence from protected forests (MRPP, T = −9.1, p < 0.001).

4. Discussion

Our study of Estonian hemiboreal forests shows how long-term management practices influence forest structure and biodiversity, which is crucial for designing effective forest management policies. Our findings correlate with previous research on living and dead tree densities [52,53]. Contrary to study [27], our protected areas exhibited significant differences in deadwood volumes compared to commercial forest areas, indicating positive effects associated with protection. Our hypothesis that mixed stands would exhibit greater structural diversity, and that managed stands would have lower levels of forest structures, particularly deadwood, was confirmed. Contrary to our initial assumptions, ground vegetation species richness was not significantly affected by management. Bryophyte species richness was higher in protected areas, though the richness of herb layer species decreased.

Previous studies in boreal forests [26] have detected a nonlinear change in species composition response, indicative of a significant resilience in medium productivity site types. Also, Ref. [54] found a likely indirect pathway of edge effects through overstorey loss which led to shrub cover loss in the long term. This resilience may be attributed to the dominance of shrub and moss species like Vaccinium myrtillus, Vaccinium vitis-idaea, and Hylocomium splendens. Because of their broad ecological niches, these species are more tolerant of disturbances, thus significantly contributing to the overall resilience of the forest stands [55,56]. These findings offer insights into the ecosystems’ natural resilience to anthropogenic impacts. Our results also support this viewpoint showing broad ecological niche species in different management regimes. For example, the presence of species like Orthilia secunda across various habitat conditions highlights their ecological adaptability, reflecting the intricate interplay between species and their environments. These dynamics are possibly influenced by the unique root systems of these plants, which engage in symbiotic relationships with mycorrhizal fungi, enhancing nutrient and water uptake. In turn, the fungi benefit from the carbohydrates produced by the plants through photosynthesis. The versatility of these species offers a valuable means to monitor diverse ecological states and assess the efficacy of various management approaches.

The transition in forestry processes has significantly impacted post-Soviet countries for decades [57,58,59,60]. Our research shows that forestry management actions such as sanitation, selection cutting, and thinning did not modify vascular and bryophyte species richness significantly. These findings do not suggest that recent management actions may become significant later, potentially being more substrate-based [61], or promoting vegetation growth [62]. Historical management actions conducted over 30 years ago have been shown to facilitate the restoration of vegetation composition in these stands, illustrating the resilience of historically and moderately managed forest stands [22]. This resilience is similar to our study observations in protected areas, where an increase in bryophyte richness and a decrease in vegetation richness indicate changes in substrate and light conditions.

Our findings about higher tree species richness in historical management sites suggest that these actions were more nature-based, creating more diverse species compositions. The scarcity of bryophytes and the presence of a larger scale of generalist species related to forest origin compared to historic management further endorse this idea [63]. Similarly to the findings of [64], we agree that in landscapes with long-term structures, forest species are less limited by dispersal and more by habitat characteristics. The species composition is influenced by the persistent presence of light, moisture, and fertility in the stand, determined by forest habitat type.

Our results show that in mature forests, the effect of forest age is more related to individual trees than to the entire stand. Individual trees are especially important for vascular plant species, which depend heavily on light conditions influenced by selected trees in past management actions. Selective cutting of canopy trees improves light availability, favoring regeneration and leading to a denser understory with an altered composition [65,66].

The significance of bryophyte richness from mixed-species origin and site protection has likely resulted from lower light access and different substrate base. This highlights the importance of considering the abundance, size, and decay stage distribution of coarse woody debris, which are key characteristics of natural forests [27,53] and support biodiversity [67,68]. The volume of deadwood increased under protection which suggests an enhancement of habitat complexity. Similar results, that forest protection increases deadwood volume and bryophyte species diversity, were also found by [69]. Like [70] we saw that management had the strongest negative effects on deadwood structures that occurred predominantly in the most productive forests like our study sites.

However, it is essential to acknowledge several limitations that merit consideration. The distinction between recent and historical management practices, influenced by the evolution of forestry machinery, introduces a variable that could influence the comparability of data across time. Although our study assumes ecological consistency across the research areas, aside from the effects of different habitat types and passive conservation measures, this simplification may not fully capture the complex interplay of ecological processes influencing forest dynamics. Moreover, our temporal overview, while comprehensive, may not capture the entirety of long-term ecological changes or the delayed effects of past management practices on the current composition and structure of forests. Future research should aim to incorporate more specific methodological approaches that can differentiate among various management practices over time and assess their individual impacts. Additionally, expanding the geographical and ecological scope of the study could enhance the applicability of future findings.

5. Conclusions

This study has systematically examined the long-term effects of different forest management practices on the biodiversity and structural complexity of Estonian hemiboreal forests. The results clearly demonstrate that multi-aged origin forests exhibit greater biodiversity and structural complexity compared to even-aged stands. This is driven by factors such as the higher average diameter of canopy trees, greater volumes of lying deadwood, and extended ecological continuity. These elements collectively support diverse plant communities, including a higher richness of bryophyte species and greater understory diversity, highlighting the critical role of habitat heterogeneity in promoting biodiversity.

Historically managed forests were found to have higher tree species richness but lower volumes of deadwood, suggesting a trade-off between species richness and structural complexity due to past disturbances. These findings show the importance of considering the long-lasting impacts of historical management when developing current forest management strategies.

Forests that have not undergone recent management interventions exhibited significantly higher levels of deadwood and understory diversity, confirming that recent management activities—particularly those implemented after the 1990s—tend to reduce tree volume and deadwood, impacting forest structure and biodiversity. In contrast, protected areas showed higher average diameters of canopy trees and greater volumes of both lying and standing deadwood. The bryophyte species richness was also higher in protected areas, although the richness of herb layer species decreased. These results bring out the importance of conservation-oriented management in maintaining habitat complexity and supporting ecological functions.

Our research offers several novel insights into the resilience and adaptability of forest ecosystems. For instance, species with broad ecological niches, such as Orthilia secunda, were prevalent across diverse management regimes, indicating their ability to thrive in various habitat conditions. These findings emphasize the need to manage multi-aged and protected forest stands with a focus on maintaining structural complexity and biodiversity.

In conclusion, successful forest management requires the integration of ecological insights and conservation priorities. By fostering landscapes that are productive, sustainable, and rich in biodiversity, forest management practices can better support ecosystem resilience and contribute to the long-term preservation of biodiversity in hemiboreal forests.