Silvicultural Measures for the Protection of Early-Stage Forest Regeneration from Deer Browsing: A European Perspective

Abstract

Forests worldwide are increasingly affected by climate-driven stressors and large-scale disturbances that impair tree physiology, disrupt water and carbon balance, and increase mortality risk. In this context, successful natural and artificial regeneration is essential for maintaining forest continuity, carbon storage, and biodiversity. However, regeneration outcomes depend not only on site conditions but also on biotic pressures, especially browsing by cervids in temperate and boreal forests. The aim of this review was to identify and synthesize evidence on how silvicultural methods can reduce browsing damage in forest regeneration and to assess how these methods influence the underlying drivers of cervid pressure through stand structure and forage availability. We examine mechanisms operating at two spatial scales: at the microscale, regeneration type, planting density, structural heterogeneity, planting stock, and how species mixture influences browsing probability and intensity; at the macroscale, how cutting systems and the spatial and temporal arrangement of harvests shape foraging landscapes by concentrating or dispersing browse resources and edge habitats. The reviewed evidence shows that dense, structurally diverse natural regeneration can dilute browsing pressure, whereas uniform artificial regeneration may increase repeated damage, and that species composition and mixture patterns can either protect or expose palatable species. We conclude that integrating microscale regeneration design with landscape-level harvest planning can strengthen stand resilience, reduce dependence on fencing, and support climate-adaptive forest development. To the best of our knowledge, no previous review has synthesized this evidence across both micro- and macroscale silvicultural contexts. Although most of the studies included in this review originate from Europe, we believe that the knowledge presented here is relevant to the majority of boreal and temperate forests worldwide.

1. Introduction

Contemporary forestry and forest ecosystems worldwide are under increasing pressure from accelerating climate change, which intensifies the frequency and magnitude of phenomena that significantly disrupt their functioning [1]. Among the most important of these are large-scale natural disturbances such as droughts, fires, and insect outbreaks, as well as heat waves, episodes of extreme frost following thaw periods, hurricanes, and hydrological disturbances (waterlogging and soil desiccation) [2,3,4]. These disturbances can severely impair the success of forest regeneration by increasing mortality and suppressing the growth of seedlings and saplings while also disrupting regeneration establishment, e.g., through heat and drought stress, fire damage, windthrow and increased pressure from insect herbivores. Moreover, because such disturbances often interact and cascade, they can amplify regeneration failure and increase the need for adaptive silvicultural measures. As a result, trees become more susceptible to fungal pathogens and physiological stress, which leads to impaired water transport and carbon deficit and, consequently, to increased mortality of forest stands [3,4,5]. In this context, forest regeneration—both natural and artificial—is becoming increasingly important as one of the key mechanisms enabling the continuity of forest ecosystems functioning and their capacity to act as climate stabilizers [6,7,8]. Young tree generations are responsible not only for rebuilding biomass or absorbing carbon dioxide but also for creating future carbon stocks in wood and soil, which makes effective regeneration one of the foundations of climate-change mitigation strategies [9,10]. Without an effective and lasting regeneration phase, even structurally well-preserved stands gradually lose their ability to sequester carbon efficiently [11,12,13].

The success of forest regeneration is determined by complex interactions between abiotic and biotic factors [14]. Abiotic factors (light availability, soil properties, moisture, or temperature) significantly influence seed germination, seedling and sapling growth, and survival [15]. In particular, light conditions, directly shaped by stand structure and silvicultural treatments, regulate competition among tree species and thus determine the course of succession and the future species composition of the stand [16,17]. The influence of individual abiotic factors on the course of forest regeneration is relatively well described in the literature [18,19]. The role of biotic factors in forest regeneration is more complex [20], as it involves multilevel interactions among plants, animals, fungi, and microorganisms and their synergistic impact on weakened trees [21,22]. Their trajectory depends on site conditions and on the temporal dynamics of the ecosystem. These dynamics involve changes in the composition and intensity of organismal effects at different successional stages—from early regeneration phases to older stand stages—which can alter both the pace and direction of the regeneration process [23]. In addition, these relationships are often nonlinear and indirect, which makes it difficult to determine unambiguously their effect on regeneration success [24]. Biotic factors may either promote or hinder forest regeneration by influencing seed production, seed dispersal, germination, seedling and sapling survival, and their early growth. In particular, herbivory, pathogen pressure, and competition from understory vegetation can substantially reduce regeneration success, whereas mutualistic interactions with mycorrhizal fungi and seed-dispersing animals may enhance establishment under suitable environmental conditions [25]. As a result, biotic factors are a major source of uncertainty in forecasting forest regeneration [26].

In the temperate and boreal zones, one of the key biotic factors affecting the condition and species composition of regeneration patches is browsing by cervids [27,28]. Large herbivorous mammals influence forest regeneration both directly, by reducing sapling survival, and indirectly, through selective browsing of attractive species, which may lead to lasting changes in stand structure and diversity [29]. A growing number of studies indicate that the intensity and selectivity of cervid browsing depend on the combined effects of population-related and habitat factors, as well as stand structure [30]. A particularly important, and often underestimated, element of this system is silvicultural measures, understood as the set of planned forest management interventions used to establish forest regeneration to achieve the desired structure and species composition of the future stand. Such measures can significantly modify the availability of food resources (and foraging efficiency) for large herbivorous mammals (especially cervids), thereby influencing the spatial distribution of pressure exerted by these animals on forest regeneration [30,31]. As a result, even moderate cervid pressure may negatively affect the condition of regeneration and its ability to recruit into higher stand layers [32,33]. In this context, managing cervid pressure on forest regeneration should be treated as an integral part of steering the long-term stand development strategy [34,35].

However, the use of silvicultural methods designed to promote regeneration resistant to cervid browsing remains uncommon in forest management practice. Efforts to counter browsing are based mainly on the use of physical barriers (e.g., fencing and individual tree shelters) [35], chemical treatments [29,36], and reductions in cervid abundance [37,38]. Although these measures may partly reduce pressure on regeneration, they do not remove the causes of the problem, namely, the disturbed relationship between cervid presence, forest management practices, and ecosystem functioning. For example, in some countries, including Poland, fencing is used relatively frequently; however, despite its high local effectiveness, it may generate substantial ecological costs [35,39]. The consequences of its use may include habitat fragmentation, barriers to animal movement, and increased browsing pressure on unfenced areas [40,41]. In the German federal state of Bavaria, approximately 2.3 million running meters of fencing around regeneration were maintained in state forests in 2025 [42]. Another important consequence of this approach is its high financial cost [43,44]. The scale of these costs in Europe is considerable, although data across countries are not fully comparable because of different reporting methods and differences in forest area. For example, in Poland the costs of protection against cervid damage amounted to PLN 306.7 million in 2024, i.e., about EUR 72.1 million [45]. In the Czech Republic (≈four times smaller in area than Poland), the analogous figure for all ownership categories is estimated at approximately CZK 1,451 million annually, i.e., about EUR 59.5 million [46]. In Scotland (also ≈four times smaller than Poland), the Forestry Grant Scheme budget for 2024 related to all forms of forest ownership amounted to GBP 45.4 million, i.e., about EUR 52.5 million [47]. These data show that physical and chemical protection of regeneration constitutes a costly component of forest management in different countries.

An alternative to tree protection is to use silvicultural methods, which operate on a different principle: instead of limiting animal—plant contact, they modify forest structure and its attractiveness to cervids [48]. By shaping species composition and the spatial and vertical structure, it is possible to establish regeneration with intrinsic resistance to browsing while also increasing resilience to environmental stressors, including prolonged drought, extreme temperatures, fungal pathogens [49], and insect pests [50]. An additional argument in favour of silvicultural methods is their compatibility with the aims of adapting forestry to climate change and mitigating its effects [51]. Many tree species that are highly attractive to cervids, such as silver fir (Abies alba Mill.) and sycamore maple (Acer pseudoplatanus L.), are at the same time characterized by high tolerance to drought and heat, as well as by having a high ecological and production value [52,53]. Their elimination from regeneration through excessive browsing leads to stands dominated by less browsed species that are often more sensitive to climatic stress, such as Norway spruce (Picea abies (L.) H.Karst.) [30,54,55]. Thus, cervid browsing can redirect regeneration trajectories towards species assemblages with lower climatic resilience, weakening the long-term adaptive capacity of forest ecosystems [56]. Because, as noted above, effective and stable associational resistance is a prerequisite for rebuilding biomass and sustaining long-term carbon sequestration in stands [57], silviculture aimed at reducing browsing pressure may not only protect young trees but may also strengthen ecosystem services in a broader sense [58,59].

Silvicultural methods can be considered at two complementary scales: the microscale, encompassing the internal characteristics of regeneration site [60], and the macroscale, associated with cutting systems and the spatial organization of stands [61]. At the microscale, species composition and planting density are particularly important because they affect both the selectivity of cervid browsing and the attractiveness of entire patches as foraging areas [62]. At this scale, cervid food preferences may shift the species composition of associational susceptibility towards taxa that are less attractive or more resistant to damage [35,63]. In addition, species composition and mixing patterns can determine neighbourhood effects: the presence of well-defended species may limit pressure on neighbouring attractive species, whereas the presence of preferred species may intensify it [64,65]. The type of planting stock may also matter because physiological traits of the planting material, including container-grown stock, can affect its food attractiveness and its ability to compensate for browsing damage [66,67,68].

At the macroscale, key stand characteristics are those shaped by cutting systems, which influence the development of the herb layer, understory, and advance regeneration, as well as the distribution of food resources across the landscape [69]. As a result, different cutting systems create distinct foraging conditions, determining how intensively cervids use particular regeneration patches [70]. The response of regeneration to browsing may differ between clear-cut areas and under canopy regeneration, partly because of differences in the availability of alternative food sources [62]. At the landscape scale, cutting systems can either concentrate herbivore pressure in specific parts of a forest complex or distribute it more broadly through a mosaic of regeneration patches with varying attractiveness [71,72]. Therefore, an effective silvicultural strategy for reducing browsing damage must integrate stand-level measures with spatial planning at both stand and landscape scales.

The aim of this study is to assess the effect of selected silvicultural methods on the intensity and selectivity of cervid browsing, with particular emphasis on the role of species composition, neighbourhood effects and forms of regeneration mixture, as well as cutting systems. To the best of our knowledge, no previous review has synthesized this evidence across both micro- and macroscale silvicultural contexts.

2. Materials and Methods

The literature review drew on Google Scholar (Google LLC, Mountain View, CA, USA), ResearchGate (ResearchGate GmbH, Berlin, Germany), and ScienceDirect (Elsevier B.V., Amsterdam, The Netherlands) to identify studies on cervid-induced tree damage in forest regeneration, with particular emphasis on Europe. The search combined keywords referring to cervids, browsing pressure, tree damage, forest regeneration, and major forest-forming tree species. Terms included, among others: ungulates, deer, moose, roe deer, browsing, forest regeneration, plantation, tree damage, and the names of principal European species such as Scots pine (Pinus sylvestris L.), spruce, fir, European beech (Fagus sylvatica L.), pedunculate oak (Quercus robur L.), and sessile oak (Quercus petraea (Matt.) Liebl.). The query was then expanded to include the term Europe and individual European country names. In addition, the reviewed literature was used to compile

Table 1. Comparative characteristics of selected cervid species relevant to browsing pressure in forest regeneration.

Ungulate Species	Roe Deer Capreolus capreolus	Red Deer Cervus elaphus	Moose Alces alces
Feeding strategy	Concentrate selector; highly selective for high-quality, low-fibre forage [73,74].	Intermediate feeder; flexible along the grazer–browser continuum [73].	Concentrate selector; large browser, strongly associated with woody browse and early-successional forage [73].
Digestion physiology	Small rumen–reticulum, rapid passage, morphophysiology associated with selective feeding on easily digestible plant parts [73].	Intermediate morphophysiology among ruminants; capable of using more fibrous forage than roe deer while retaining substantial dietary flexibility [73,75].	Browser-type ruminant adapted to browse-rich diets; large body size allows extensive use of woody forage [73].
Home-range size	Most often, home ranges are several dozen hectares in continuous forests, but they may locally exceed 100 ha under less typical or poorer habitat conditions [76]. Across European forested landscapes, relatively small home ranges compared with larger cervids [74,76].	Most often, home ranges span from several hundred to several thousand hectares, and in forest–agricultural landscapes they are clearly larger than in forest-dominated landscapes [77].	Most often, home ranges cover several to several dozen km2, representing the largest spatial scale among these three species; they are often larger in summer and more concentrated in winter [78].
Habitat-use intensity	Generally high in structurally diverse forest patches that combine cover and foraging opportunities. In a managed temperate mixed forest, local forest structure (canopy openness, tree-species richness, vertical complexity) explained habitat use better than small-scale food abundance [79].	Typically high where forage and cover are closely juxtaposed. Habitat use in human-modified landscapes increases in productive open habitats but remains strongly tied to access to forest cover and edge structures [77].	Very high in managed boreal landscapes offering young stands, forest edges and clear-cuts [78].

, which summarizes the general comparative characteristics of selected cervid species relevant to forest regeneration, including feeding strategy, digestion physiology, typical home-range size and habitat-use intensity.

The review considered only peer-reviewed articles in English published from 1970 to December 2025, excluding non-peer-reviewed and popular science materials. Screening proceeded in stages. After removing duplicates, titles, abstracts, and keywords were screened; papers meeting the topic criteria underwent full-text assessment. At this stage, we included studies on browsing pressure, regeneration damage, and relationships between the intensity of ungulate impact and habitat, stand, or landscape characteristics. We also applied snowball sampling by examining the reference lists of papers selected for full-text assessment, which yielded additional sources not captured in the initial search.

In total, 315 publications were identified, of which 177 were eligible for full-text analysis. The earliest included a study dated from 1976 and the most recent from 2025. A clear majority (68%) of the analysed studies were published between 2004 and 2022. From each eligible publication, we extracted information on the cervid species or groups investigated, the type and intensity of damage, the type of natural regeneration or plantation, the tree species involved, the spatial scale, the methodology applied, and conclusions relevant to forestry practice. The material was then synthesized qualitatively to identify factors that increase or reduce ungulate pressure and to evaluate silvicultural measures that may lower the risk of damage in forest regeneration. The results of the review are presented in the sections below.

3. Selected Silvicultural Measures to Reduce Cervid Browsing Pressure in Forest Regeneration

3.1. Type of Regeneration and Planting Material

The consequences for browsing intensity in regeneration are already determined at the establishment stage. Among the factors influencing browsing pressure, the regeneration method is pivotal. Numerous studies report differences in browsing pressure between artificial and natural regeneration [64,80,81]. Factors shaping success or susceptibility to cervid pressure include tree density, age and height, the origin of planting material, and the presence of alternative vegetation.

Successful natural regeneration typically exhibits higher density, greater age and height class diversity, and strong spatial heterogeneity, which can produce a dilution effect that lowers the probability of damage to individual trees [82,83]. Lettenmaier et al. [84] showed that, under natural regeneration, high densities of young trees can mitigate browsing impacts—even under strong cervid pressure, the primary outcome was a shift in species composition rather than a decline in total tree numbers. Similarly, Borowski et al. [85] found that as natural regeneration density increases, the proportion of browsed individuals decreases, with the effect particularly marked in more susceptible species. This pattern reflects a disproportionate relationship between shoot availability and browsing intensity: greater sapling abundance distributes pressure across more stems, thereby reducing damage to any single tree [55,86,87]. These results indicate that where natural regeneration success is limited, browsing pressure concentrates on relatively few trees, whereas high sapling density spreads pressure more evenly and substantially lowers per tree damage. In addition, soil scarification often increases natural regeneration success, which may indirectly reduce cervid damage [88]. Another element of natural regeneration that reduces browsing is age and height-related heterogeneity [89,90,91]. Ungulates select specific height classes and development stages, commonly preferring trees of intermediate height—between ground level and the upper limit of their browsing reach [92]. In heterogeneous natural regeneration, where saplings of varying ages and heights co-occur, browsing becomes less efficient [68]. Consequently, vertical and spatial heterogeneity promotes the temporary or permanent escape of some individuals from browsing, increasing the probability that they recruit into higher stand layers [62,86]. Browsing intensity may also be negatively affected by competing vegetation, the presence of which on non-scarified sites may reduce browsing pressure [93].

In contrast to natural regeneration, artificial regeneration is typically characterized by a uniform age–height structure and by saplings arranged in rows at fixed spacing, which makes browsing more predictable for cervids [94]. Such layouts increase browsing efficiency by reducing the energetic and time costs of foraging, favouring repeated use of the same patches and concentrating pressure on a limited number of saplings that simultaneously fall within the same “browsing window” [95]. Cervids prefer spatially predictable resources, and repeated browsing may keep plants within a vulnerable height class, raising the risk of subsequent damage [96]. This mechanism corresponds to the browse-trap phenomenon, in which chronic pressure prevents plants from attaining heights at which they escape herbivore pressure [97,98]. Moreover, in artificial regeneration, high browsing intensity often translates into qualitative losses (leader shoot deformation, multiple stemming) and quantitative losses (sapling mortality) [64,80,94]. Consequently, the classic silvicultural approach based on evenly aged, evenly spaced saplings—although it facilitates tending operations—substantially increases the risk of strong browsing pressure [27].

The type of planting material also affects how ungulates use artificial regeneration areas. Recent studies suggest that container-grown saplings may be browsed more readily—particularly in the first years after planting—than naturally regenerated saplings or bare-root stock [68]. As noted earlier, tree height is a key determinant of browsing frequency [99,100]. Shortly after planting, container-grown saplings often grow faster than bare-root saplings [66] and thus relatively quickly reach heights most favourable for cervid browsing. They also tend to show greater above-ground biomass increment, which may increase food attractiveness and favour selective browsing.

Thus, the type of regeneration is a key determinant of resistance to browsing. Where feasible, natural regeneration should be prioritized; when successful and dense, it reduces browsing pressure by dispersing it across numerous individuals [39]. High densities can be achieved with appropriate site-scarification treatments that favor germination and seedling survival [88]. However, for artificial regeneration on sites with strong competition from the herb layer, container-grown saplings are advisable, owing to better establishment and faster initial growth, although their rapid early height gain and greater above-ground biomass often make them more readily browsed in the first years after planting [68,101]. At the same time, the retention of some non-woody vegetation may reduce cervid use of saplings, whereas planting saplings that vary in height and age may better approximate a natural stand structure, thereby enhancing resistance to cervid damage [39,102].

3.2. Species Composition

The species composition of natural and artificial regeneration is another important factor determining the scale and spatial distribution of cervid browsing pressure. Although preferences of cervids for different tree species are well described in the literature [35,79,86,93], they have clear and direct implications for managing key elements of silviculture, including neighbourhood effects and mixture forms (see the review below). Therefore, we provide a short summary of the most important facts. As already mentioned, cervids affect young stands selectively, preferring some species and avoiding others, which over time alters stand composition, vertical structure, and successional trajectories [55,103]. As a result, the future stand structure may increasingly reflect current food preferences rather than forest site potential [104].

In Central and Northern Europe, species particularly exposed to intense and repeated browsing include deciduous trees such as sycamore maple, silver fir, Norway maple (Acer platanoides L.), field maple (Acer campestre L.), rowan (Sorbus aucuparia L. em. Hedl.), European ash (Fraxinus excelsior L.), wych elm (Ulmus glabra Huds.), small-leaved lime (Tilia cordata Mill.), and pedunculate and sessile oaks [30,104,105,106]. Their high attractiveness is generally related to high digestibility, high tissue nitrogen, and low levels of secondary compounds [107]. Browsing of these species often reduces height growth and biomass increment, disturbs crown architecture, and delays recruitment into higher stand layers [64]. For example, repeated browsing of the main shoot in silver fir may permanently suppress height growth and induce strong bushiness, reducing competitive ability and often preventing recruitment into the sapling layer [108]. In oaks, intensive browsing may keep individuals in a low shrub form for many years, increasing their susceptibility to competition from beech and spruce and likewise limiting recruitment [107,109]. In extreme cases, browsing may also cause substantial mortality, particularly in highly preferred species such as maples and ash [38,110]. Rowan, despite its sprouting capacity, may remain trapped below browsing escape height for many years [106]. Elm, lime, and hornbeam (Carpinus betulus L.) are also strongly browsed; in hornbeam and lime, biomass allocation shifts toward lateral shoots and sprouts, while axial growth declines, reducing their role in stand recruitment [30,105]. In silver birch (Betula pendula Roth.) and downy birch (Betula pubescens Ehrh.), browsing is usually less frequent, but it may still reduce growth and increase susceptibility to drought- and frost-related stress, although these species often recover better than oak or fir [38].

In contrast, most coniferous species appear less affected by browsing, either because they are less palatable to herbivores or because they exhibit greater tolerance to mechanical damage. This pattern is observed in spruce, pine, European larch (Larix decidua Mill.), and, to a lesser extent, Douglas fir (Pseudotsuga menziesii (Mirb.) Franco.) [28,111,112]. Their lower attractiveness reflects a combination of chemical and morphological traits, including higher lignin and resin content, phenolics and terpenoids, needle stiffness, and lower tissue nitrogen [113,114]. For example, Norway spruce may be browsed locally at a young age but usually soon exceeds browsing escape height (about 1.5–2 m), favoring its dominance in natural regeneration [104]. Scots pine, in turn, often avoids severe biomass losses owing to rapid early growth and low shoot digestibility, especially on coniferous sites [104,115]. Under strong and prolonged browsing pressure, European beech may compensate for losses through intensive lateral growth and shoot association, thereby gaining a regenerative advantage and increasing its share in the regeneration layer [106,116].

From a silvicultural standpoint, species composition should be managed so as to maintain regeneration continuity despite selective browsing. In practice, this means avoiding monocultures of highly attractive forage species and promoting mixed regeneration [117], in which heavily browsed species are introduced in smaller proportions and combined with less attractive or more tolerant species, such as spruce, pine, larch, and Douglas fir, as well as species with high compensatory capacity, such as beech. This reduces the risk of eliminating browse-sensitive species. Importantly, species particularly vulnerable to browsing often also belong to those better adapted to environmental stresses associated with climate change [118,119,120]. Therefore, their effective protection against ungulate pressure is important not only from the production perspective but also from the adaptive one, as it supports the development of stands that are more resistant to future climatic conditions and increases the stability and resilience of forest ecosystems [121,122]. Detailed mechanisms and silvicultural recommendations related to associational resistance/susceptibility and forms of species mixture are discussed in the following chapters.

3.3. Neighbourhood Effects

Another element of regeneration characteristics that can be used in silviculture to strengthen regeneration resistance to browsing is the neighbourhood of species with differing attractiveness. Importantly, neighbourhood effects may both reduce browsing (attractive species for cervids growing next to unattractive ones) and intensify it—when unattractive species occur in the vicinity of attractive ones. Effects related to the species neighbourhood are usually context-dependent and may change (from resistance to susceptibility) with spatial scale [123], species proportions, and cervid pressure [124,125,126]. For example, cervids may prefer beech if it grows in combination with spruce or pine but avoid it when it occurs next to other, more attractive broadleaved species, such as sycamore maple or ash [127], as cited in [128].

One component of neighbourhood effects is associational susceptibility, which in the context of cervid pressure means a situation in which the presence of some tree species in the neighbourhood increases the probability of browsing of the focal species (the species affected by neighbouring species) [90]. This susceptibility is most often linked to the attraction effect when the admixture of species preferred by cervids increases overall foraging activity in a given part of the regeneration [90,129]. An example of associational susceptibility is increased browsing of conifer species in regeneration containing attractive broadleaves. In boreal and Scandinavian forests, it has been found that the distribution of damage in pine or spruce regeneration may depend on the distribution of preferred tree species, such as willows (Salix spp.) or rowans [130,131]. Broadleaved species attract cervids, and intensive use of the area results in the browsing of conifers [129].

Another key mechanism of neighbourhood effects is the associational resistance of trees [129,132,133,134]. In theoretical terms, this mechanism is sometimes broken down into three effects: (i) masking/reduced visibility of focal species; (ii) resource dilution, numeric dilution; and (iii) the influence of neighbours of low attractiveness (mechanical/chemical barrier, “nurse plants”). These effects are often described jointly. They may be particularly important in the first years after regeneration establishment, when saplings are young/low and the selective search for food is relatively efficient [106]. The proximity of spiny/thorny plants may act as a mechanical barrier and/or a masking element and can significantly reduce damage to plants growing next to them [135,136]. Under European conditions, species such as blackthorn (Prunus spinosa L.), hawthorns (Crataegus spp.), and wild roses (Rosa spp.) are often indicated as protective plants of this kind (masking-barrier function) [136]. It should be noted, however, that these species may simultaneously create competition for light and water and may subsequently require selective removal [137,138]. Another species indicated as protective at the European scale is sweet briar (Rosa rubiginosa L.). The shelter provided by this shrub may reduce browsing and increase the survival of saplings, for example, of fir or sycamore maple [139]. A sheltering function may be performed not only by individual species but by entire assemblages of accompanying vegetation, through reducing young tree accessibility or detectability. In this context, strongly preferred species, such as rowan, are so attractive that the presence of shelter does not significantly reduce their browsing, while less frequently browsed species, such as spruce, are subject to little pressure regardless of surrounding conditions. In species of intermediate preference, such as beech, however, the traits of neighbouring vegetation may determine the profitability of choosing them as food [79]. An equally good example of the practical interaction between masking and dilution is a field experiment from southern Sweden, in which the presence of raspberry (Rubus idaeus L.) and blackberry (Rubus fruticosus L.) significantly reduced browsing pressure on oak saplings. The mechanism can be interpreted, on the one hand, as masking because the shrub layer limits sapling detectability and hindered access to them, and, on the other, as an expression of numeric dilution when the increased complexity and density of plants around the browsed species reduces the probability that a given sapling will become the object of browsing [140]. Similar results were obtained in other studies involving blackberries, where their development over time was associated with a decline in deer pressure on birch and willow saplings [141]. Studies from Spain additionally showed that the effectiveness of “protective shrubs” depends on the type of defence and the surrounding conditions: shrubs physically protected (by thorns) may be effective in wet years when food availability is relatively high. Under stress and food shortage, however, chemically defended shrubs (e.g., of low digestibility) may be more effective because cervids more readily overcome the physical barrier when access to higher-quality food is limited [142,143].

Regeneration resistance can be shaped not only by using shrubs but also through the appropriate choice of neighbouring tree species. For example, fir may exhibit significantly lower browsing intensity when it grows in the vicinity of less attractive species, such as beech or spruce [32]. It has also been shown that the neighbourhood of spruce may reduce browsing on sycamore maple [82] or oak [121,129]. It should be borne in mind, however, that the eventual neighbourhood effect is context-dependent. For instance, in situations where beech is surrounded by species of intermediate preference, such as Scots pine, and unattractive species, such as spruce, it will be readily used by cervids [127], as cited in [128]. Borkowski et al. [99] showed that young pine stands may function as a kind of browsing refuge for an admixture of oak and that this effect intensified with the age and growth of the pine. On this basis, the authors concluded that introducing oak—as a species more attractive as forage—should take place several years after pine regeneration when the pine stand has reached a height ensuring a protective effect [99].

In conclusion, from the perspective of silvicultural methods, associational resistance works best as a set of design measures consisting of: (i) maintaining or introducing temporary shrub and structural shelter, and (ii) using protective species—especially thorny or poorly digestible ones—in the immediate neighbourhood of target species while at the same time planning later reduction of competitive interactions. In practice, this means that associational resistance at the regeneration stage provides silviculture with tools to increase stand-level resistance. Instead of simplified patches that cervids can easily penetrate, shelter vegetation and species admixtures should be retained (at least temporarily) to reduce access to target species and increase foraging costs, thereby reducing both the probability and intensity of browsing.

3.4. Forms of Mixture

The susceptibility of regeneration to browsing is determined not only by species selection or the conditions of the direct neighbourhood between different species but also by the way they are arranged within the regeneration patch as a whole [144]. These relationships are much more complex than the framework of associational resistance/susceptibility [134]. The same set of species may be browsed to different degrees depending on whether it was introduced in a row-wise or group mixture.

In a row-wise mixture, species are arranged in a regular, repetitive pattern every few/several trees of the main species, which, from the perspective of cervids, creates an easy spatial pattern for foraging [30,145]. Such a way of establishing regeneration may maintain high predictability and accessibility of species attractive as forage and may also facilitate sequential movement of cervids along rows and repeated damage to successive individuals [146]. The results of experiments conducted in north-eastern Poland showed that young pine trees may partly protect oak against browsing. However, this effect was weaker in the row arrangement than in the arrangement of groups, and oaks growing in groups attained greater heights than individuals planted singly within pine rows [99,100]. Modifying the row arrangement by using double rows also does not necessarily suffice to effectively reduce damage if the saplings still remain arranged in a regular, easily recognizable pattern [147]. As a result, a row-wise mixture, although designed to increase diversity, may in the longer term undergo secondary homogenization as a consequence of the selective removal of species more readily browsed by cervids [148,149]. A similar problem may also occur in natural regeneration: in beech, a regular row-wise seedbeds soil preparation system did not eliminate browsing, leading to a significant reduction in associational of this species [150].

A separate form is the group mixture of species (group mixture; patch-scale mixture). In discussing this form of mixture, it is more useful to consider its size, rather than its name, since various terms are used in the literature [30,151]. The smallest groups (up to 0.01 ha) usually comprise several to several dozen saplings. They often function similarly to a row-wise mixture and despite the structural clustering of saplings, do not yet provide the protective effect resulting from growth in larger aggregations [148,152]. Because cervid browsing usually concentrates in selected fragments of a regeneration patch rather than being distributed evenly across the entire area, small clusters of attractive species may be subject to above-average browsing pressure [153]. Only growth in larger groups may perform a protective function. Group mixture most often corresponds to clusters with an area of 0.03–0.1 (less frequently up to 0.3) ha [154,155,156]. It is only at these spatial scales that browsing conditions may be modified more substantially at several levels simultaneously. First, a larger number of saplings within a group or patch may lead to a partial dilution of browsing pressure. For example, studies from south-eastern Poland showed that with increasing sapling density in a group, the proportion of browsed individuals decreased, which concerned primarily beech, fir, sycamore maple, hornbeam, and oak regeneration patches [85]. Second, larger and more diverse groups may reduce browsing risk [157]. In the broadleaved forests of Hainich National Park in Germany, the proportion of browsed saplings of oak, ash, and sycamore maple was lower in larger and more diverse groups [151]. Third, in larger groups there is a greater probability that at least some individuals will gain a height advantage beyond the zone of most intensive browsing. For example, in the Paneveggio reserve in Italy, cervids browsed mainly on regeneration below 150 cm, which indicates that individuals growing more quickly above this threshold had a greater chance of permanently entering higher height classes [149,158].

In silvicultural practice, especially in regions exposed to high cervid pressure, the inclusion of attractive forage species in row-wise mixtures should be avoided. It is more advantageous to use a group mixture while avoiding very small groups comprising only several to several dozen saplings because they do not always provide a protective effect and may still be exposed to strong cervid pressure. Larger groups, with a greater number of trees, favour a reduction in pressure and increase the chances that trees will grow beyond the browsing reach of cervids. Effective admixture planning should therefore combine species diversity with a spatial organization of the regeneration patch that makes it more difficult for cervids to locate the most attractive saplings, thereby reducing foraging efficiency and the long-term persistence of pressure on the same regeneration fragments.

3.5. Cutting Systems and Their Landscape Context

The last, but equally important, element of shaping regeneration resistant to cervid browsing is the spatial arrangement of cutting type and harvest sequence. Cross-scale analyses indicate that cervid pressure on regeneration is shaped simultaneously by factors operating at two levels of spatial organization. The first is the size of an individual regenerated area, while the second is connected with the spatial arrangement of regenerated patches in the forest landscape [27,28,159,160]. In addition, some studies point to the significance of the surroundings of regeneration, particularly mature stands. Such a neighbourhood arrangement provides proximity both to shelter and to forage, which favours the presence of cervids and, consequently, their pressure on regeneration [34,161].

Among cutting systems, clear-cutting is most often identified as the system generating high browsing pressure through the accumulation of young shoot biomass that provides high food availability [27,162]. This approach means that the boundary of easily accessible food directly adjoins security cover, which further increases the attractiveness of this type of regeneration for cervids [82,146,163]. In Pennsylvania forests, for example, it was shown that the level of damage to saplings such as maple or ash under clear-cut conditions may be significantly higher than in stands with a lower degree of opening, for example, after selection cuttings [80,164]. At the landscape scale, where possible, a dispersed arrangement of clear-cuts is recommended, thereby preventing high local concentrations of cervids that could lead to intensified pressure on regeneration [165].

More complex cutting systems, like patch-cutting and shelterwood systems may be included among regeneration systems based on partial overstory retention because in both cases regeneration depends on the presence of shelter from the parent stand. In patch cutting, regeneration arises in areas established under the shelter of, or in the immediate vicinity of, the parent stand, whereas in shelterwood systems it develops through gradual loosening of the canopy over a larger area [34,89].

In patch-cutting systems, an important factor affecting the level of browsing is the size of the patches. In the case of small patches (from several dozen to about 200 m²), the proportion of the edge zone is relatively high in relation to the whole area. Under such conditions, although the influence of the surrounding stand causes browsing to concentrate near the edges of regeneration, the pressure affects a relatively large part of the regeneration area [89,166,167]. As the patch area increases, the share of the edge zone decreases [168,169], and as a result the intensity of browsing relative to the regeneration area declines, often leading to increased survival of attractive broadleaved species [161]. It can be assumed that the influence of the surrounding stand weakens when its height is smaller than the diameter of the patch [169]. Under such conditions, cervid pressure may be reduced by the increasing share of the interior part of the patch [34,161]. Shelterwood systems based on gradual canopy opening, in turn, favour more stable regeneration conditions for late-successional species and reduce competition from the herb layer [168,170]. Slower, staged opening of the stand (about 20%–30% in the first entry) may reduce the food attractiveness of the area by maintaining a relatively high degree of shading and limiting development of the herb layer. A more intensive treatment (canopy reduction on the order of 50%–70% in a short period), by contrast, often leads to conditions similar to those after a clear-cut, resulting in higher cervid pressure and rapid development of competing ground vegetation [34,161].

At the landscape scale, for both types of regeneration system it is important to avoid excessive clustering of regeneration areas in time and space. Numerous small patches situated close to one another, just like neighbouring zones of strongly loosened canopy in shelterwood systems, may increase the probability of browsing in many regeneration areas simultaneously [79,159]. Therefore, the arrangement of patch-cuttings and shelterwoods should be as dispersed as possible, especially if they are located next to stands with a developed understory that provides security cover and thus increases animal presence [79,83]. Such an organization of cuttings makes it possible to preserve the advantage of partial canopy loosening over the effect of a large, one-time opening of the stand resembling clear-cut conditions.

From a practical standpoint, the most important recommendation seems to be limiting the use of clear-cuts, especially in regions with high cervid pressure. This appears all the more justified because the importance of early-successional species regenerated by this method is likely to decline under climate change [34,171]. Where limiting this form of regeneration is difficult, reducing its area is indicated as is designing clear-cuts in a dispersed arrangement so that an extensive post-harvest landscape favouring local increases in cervid concentration does not arise [27,28,164]. In patch-cutting systems, larger patches are more favourable because the share of the edge zone, where browsing is most concentrated, is smaller [34,168]. In shelterwood systems, a practical recommendation is to avoid rapid, large-scale canopy reduction, leading to an effect similar to clear-cutting, and instead to favour slower, staged thinning of the canopy [89,161]. In both systems, the simultaneous establishment of many neighbouring regeneration areas should be avoided, particularly near stands with a developed understory that provides cervids with security cover. As a result, the greatest possible dispersal of regeneration in time and space is recommended.

4. Conclusions

It seems that silvicultural methods may be an effective (although often underestimated) component of forest regeneration [172]. In practice, the protection of regeneration using these methods may be based on introducing regeneration and planning harvest operations in such a way as to reduce foraging attractiveness and ungulate concentration without the constant need to resort to mechanical or chemical protective measures. This means, above all, favouring solutions that create internal associational resistance through the shaping of species composition and structure (spatial and vertical) rather than merely limiting animal–plant contact and reducing population abundance. Ungulates are an important element of the ecosystem, shaping its structure (even if this does not align with timber production), and they are also indispensable for the conservation of large predators or non-consumptive wildlife use by humans (watching, photography). The above-mentioned internal stand-scale resistance can be achieved to a large extent through the use of natural regeneration (high density and an uneven-aged tree structure), including soil preparation treatments. A similar effect can also be achieved through direct seeding, which promotes high regeneration density and may better reproduce the mechanisms of natural regeneration than planting [173,174]. In the case of artificial regeneration, building heterogeneity (e.g., saplings of varying height/age) is extremely important because it reduces susceptibility to selective browsing. Appropriate design of species composition is also crucial, namely, using admixtures and neighbourhoods of less attractive species as shelter (masking, impeded access, associational resistance). Such admixtures in the form of spiny and thorny shrubs should be used moderately so that competition does not undermine their protective effect. In addition to the internal characteristics of regeneration, the landscape scale is also important [175], within which management measures should aim to limit the use of clear-cuts, especially in areas with high ungulate pressure. In situations where the use of clear-cutting is unavoidable, efforts should be made to reduce its area, avoid accumulating extensive openings in a single season, and design them in a dispersed pattern while maintaining spatial variation in stand cover and shelter. These solutions reduce conditions favourable to deer concentration and intensified browsing pressure. More favourable may be smaller openings, for example, patch cutting or shelterwood, and where possible, shaping the forest regeneration edge so that it does not create long edges that increase cervid pressure by combining food availability with another key resource for cervids, namely, security cover [176,177]. An overview of the main silvicultural measures for mitigating cervid browsing pressure in forest regeneration is provided in

Table 2. Silvicultural measures affecting cervid browsing pressure in forest regeneration.

Forest Regeneration Context	Management Method	Tree Species in Browsing Preference Order	References
Natural regeneration dense and structurally heterogeneous	Prioritize dense natural regeneration; maintain spatial and vertical heterogeneity; use scarification where needed to secure seedling establishment.	High: maple, fir, Norway maple, rowan, ash, elm, lime, oaks Intermediate: beech, birch Low: spruce, pine, larch, Douglas fir	[39,82,83,84,85,86,88,89,90,91,92]
Artificial regeneration uniform, even-aged planting	Avoid regular spacing and strongly uniform age/height structure; diversify planting layout and growth structure to reduce browsing predictability.		[27,80,94,95,96,97,98,102]
Artificial regeneration sites with strong herb-layer competition	Use container-grown stock only where establishment advantages are necessary; retain some non-woody vegetation when compatible with regeneration goals.	Most exposed among planted stock are palatable species, especially fir, maples, oaks, ash, and rowan; less attractive conifers are generally browsed less.	[66,68,99,100,101]
Species composition of regeneration	Avoid monocultures of highly palatable species; promote mixed regeneration with a lower proportion of browse-sensitive taxa and admixture of less attractive or more tolerant species.	High: maple, fir, Norway maple, rowan, ash, elm, lime, oaks Intermediate: beech, birch Low: spruce, pine, larch, Douglas fir	[30,35,38,55,64,93,104,105,106,107,108,109,110,111,112,113,114,115,116,117]
Mixed regeneration neighbourhood design	Use associational resistance by combining palatable species with less attractive neighbouring trees; avoid combinations that attract cervids into the regeneration patch.		[32,79,82,90,99,121,123,124,125,126,127,128,129]
Mixed regeneration temporary shrub shelter	Retain or introduce thorny or poorly digestible shrubs as temporary nurse vegetation; control later competition if necessary.	Strongly preferred species such as oak, fir, and maple may benefit most.	[79,135,136,137,138,139,140,141,142,143]
Mixed-species regeneration row-wise mixture	Avoid row-wise introduction of highly palatable species because regular spatial patterns increase detectability and repeated browsing.	Particularly risky for oak, maples, fir, ash, and rowan.	[30,99,100,145,146,147,148,149,150]
Mixed-species regeneration group mixture	Prefer group mixture; avoid very small groups; use larger and more diverse patches to dilute browsing pressure and improve escape above browsing height.	Grouping is most beneficial for beech, fir, maple, hornbeam, oaks, and ash; it is less important for already less preferred conifers.	[30,85,148,149,151,152,153,154,155,156,157,158]
Regeneration after clear-cutting	Limit clear-cutting where cervid pressure is high; where unavoidable, reduce cut size and disperse cuts spatially and temporally.	Clear-cuts particularly expose palatable broadleaves such as maple and ash; in general they increase access to preferred browse	[27,28,80,82,146,159,160,161,162,163,164,165]
Regeneration under partial canopy retention	Prefer larger patch cuts and shelterwood with gradual canopy opening; avoid rapid, heavy canopy reduction that creates clear-cut-like conditions.	Most beneficial for regeneration of palatable late-successional broadleaves and fir while reducing local browsing concentration.	[34,89,161,166,167,168,169,170]

.

There is an urgent need for further research on how silvicultural methods can reduce browsing pressure. One particular challenge in this context is the expected increase in the importance of broadleaved species, which are generally more resilient to climate change but, at the same time, experience substantially higher browsing pressure from cervids than their coniferous counterparts. Greater attention is also needed to understand how browsing resistance can be shaped across interacting spatial scales, from stand-level regeneration design to broader harvest and landscape patterns. Available evidence also suggests that no single plantation attribute can be considered universally neutral with respect to cervid foraging, as the effects of light conditions, local stand and landscape characteristics, and other silvicultural variables are strongly context-dependent. Moreover, important knowledge gaps also remain in context-dependent neighbourhood effects, including when species mixtures, protective shrubs, and shelter vegetation reduce browsing and when they instead increase the exposure of palatable trees or limit their development through competition. Further study is also required on the optimal spatial arrangement of mixtures, especially the role of patch structure, grouping, and structural heterogeneity in limiting repeated browsing. In addition, more evidence is needed on how planting material influences browse attractiveness and post-damage recovery, particularly in the case of container-grown versus bare-root saplings.