The high tolerance to summer drought and a wide range of sites which European yew is able to occupy (Brzeziecki and Kienast, 1994; Thomas and Polwart, 2003; and Leuschner and Ellenberg, 2017) place it among tree species with high adaptation potential to climate change. Previously common tree species, present in Europe since the Early Pleistocene, disappeared from many areas (Uzquiano
Stress-tolerant life strategy (as proposed by Brzeziecki and Kienast, 1994)—high ability to withstand shade, slow growth, resistant hardwood and high longevity—makes this species also very suitable for dendrochronological purposes. Within its wide range of occurrence in Europe, different plant associations can be distinguished, which in most cases correspond to the
The radial growth response of understorey yew is a result of the interplay among several factors. Climatically, yew grows best in the humid mild oceanic climates, and on the contrary, its distribution is strictly restricted by continental climate (Thomas and Polwart, 2003). In understorey, the yew has stable ecological conditions and is protected from severe climatic fluctuations such as frosts and low humidity (Saniga, 2000). Yew has the ability to persist in unfavourable light conditions for a long time and to survive severe stem and crown damage. The high shade tolerance (Krol, 1978) and vegetative regeneration capacity (Suszka, 1978) promote its growth. It can be assumed that understorey yew would respond with a greater magnitude to canopy disturbances than over-storey trees (Nowacki and Abrams, 1997).
Among the factors, which negatively influence the growth performance of yew, the slow growth and susceptibility to browsing and bark stripping by ungulates have been identified as the most important (Dhar
Following the beginning of canopy cuttings, significant sex-related differences in growth variability of yew were recorded (Sedmáková
The main aim of the presented work is to explore the interactions between shading and crown vitality, stem damage and radial growth of understorey male/female yew trees growing in beech forest formations (mid-mountain conditions) to improve the effectiveness of forest management and restoration activities in European temperate forests, where yew represent an important biodiversity component. In our investigation, we hypothesise the following:
Growth differences should be expected between the sex-specific subpopulations of dioecious tree species possibly because of the higher cost of reproduction for female trees.
The magnitude of sex-related growth differences is increasing with increasing age (life stage).
The pronounced growth divergence will be manifested between categories of damaged and healthy-looking trees.
Within the damage categories, male trees will over-perform female ones.
The focus of our study is to extend knowledge and improve the conservation of yew in regularly managed forest stands providing mainly regulating and provisioning services with close-to-natural tree species composition, established by natural regeneration.
In Slovakia, until the 19th century, yew was commonly distributed, and its timber was exported to the entire Habsburg monarchy (Korpeľ, 1981). Since its occurrence gradually decreased in the half of the 20th century, the most valuable populations were conserved in established nature reserves. For the purpose of this study, we selected four localities in the Western Carpathians, Slovakia (Fig. 1, Table A1), with scattered occurrence of adult yew trees, composing up to 15% of the stand basal area. The localities are part of the limestone beech forests (
Description of the study sites
Study area | Lackov Grun | Pavelcovo | Strazov | Velka Fatra |
---|---|---|---|---|
Orographic unit | Stare Hory Mts. | Stare Hory Mts. | Strazov Mts | Great Fatra |
Name | LAG | PAV | STR | VFA |
Type of management | managed forest with the main protection function—soil erosion, water retention | managed forest with the main wood production function, 7% of plots are special purpose forests with main protection function | 52%—managed forest with the main protection function—soil erosion, water retention 48%—managed forest with the main wood production function | managed forest with the main protection function—soil erosion, water retention, 10%—managed forest with the main wood production function |
Altitude (m asl) | 603–688 | 640–720 | 710–930 | 940–1,030 |
Exposition | NE | S, NE | NW–N–NE | NW–N–NE |
Slope inclination | 20–40° | 10–30° | 25–45° | 20–40° |
Ground surface | sampling plots had no parent rock present on the surface ground | sampling plots had no parent rock present on the surface ground | 13% of sampling plots had parent rock present on the surface ground | 25% of sampling plots had parent rock present on the surface ground |
Canopy cover estimation | 0.7–1.0 all plots placed inside the forest stands | 0.7–1.0 57% of plots placed inside the forest stand 43% of plots placed in cut forest where yew trees were present in group of trees of remaining mature stand | 0.4–1.0 83% of plots placed inside the forest stands 17% of plots placed in cut forest where yew trees grow solitary or in small groups of trees | 0.7–1.0 all plots placed inside the forest stands |
Sign of disturbance (based on stumps being present on the sampling plot) | all plots | all plots except two | all plots | 1/3 of plots were recorded as having no disturbance sign |
Relief | 93% of plots were established in the middle and upper part of the hill (slope) | except one, all plots were established in the middle and upper part of the hill (slope) | plots were established mainly (74%) in the middle part of the hill (slope) | plots were established in the middle and upper part of the hill (slope) |
Number of plots | 30 | 30 (13-fieldmapped) | 23 | 31 |
Size of the plot (m2) | 500 | 500 (one plot—2,500) | 500 (1x300; 2x200) | 500 |
Radius (m) | 12.62 | 12.62 | 12.62 (1x9.77; 2x7.98) | 12.62 |
Number of tree species | 13 | 10 | 9 | 9 |
Year of the sampling | 2017 | 2015 | 2016 | 2017 |
Disturbance | small scale signs of disturbance—selfthining or windthrow of trees | regeneration cut has begun in forest stands | small scale signs of disturbance—selfthinning 17% of plots—finished regeneration cut | small scale signs of disturbance—selfthining or windthrow of trees 6% of plots—regeneration cut |
We surveyed four localities in 2015–2017. Starting near the upper border of the stand, we walked along the contour line (isohypse) and marked a point every 40 m. If a mature yew with the diameter at breast height (DBH) ≥15 cm was present in a distance less than 3 m, a circular sampling plot of 500 m2 with the yew in the centre—hereafter called the ‘central yew’—was established. After reaching the stand border, we moved 40 m down the slope and walked again following the isohypse. Using this pattern, we covered the entire stand area and established 114 research plots in total. On the plots, the positions of all living trees with DBH above 8 cm, standing dead trees, snags and stumps were recorded by FieldMap (IFER—Monitoring and Mapping Solutions, Ltd., http://www.fieldmap.cz). For every standing tree, several basic characteristics were recorded: tree species, DBH, height and height of the crown base (for living trees only). In the case of yew trees, moreover, we recorded four perpendicular crown radii (x1–x4), sex, growth form as well as degree and type of stem and crown damage. The degree of crown damage (relative percentage of needle loss) was visually assessed and classified into four classes: 1—undamaged crown with transparency < 10%; 2—weakly damaged crown with transparency up to 25%; 3—damaged crown with transparency from 25% to 50%; and 4—heavily damaged crown with transparency > 50%. Similarly, damage to stems by herbivores was categorised into four classes. Class one corresponds healthy undamaged stem, class two with the weakly damaged stem up to 25% of stem circumference, class three with damaged stem from 25% to 50% of stem circumference and class four with the heavily damaged stem when more than 50% of stem circumference is bark stripped. For central yew, crown length and width were calculated. In the case of stumps and standing dead trees, we recorded tree species, diameter (at the top in case of stumps) and height.
For the dendrochronological analysis, we collected increment cores from all 114 central yews and 36 additional mature yews located on sampling plots at four localities, for a total of 150 cored yew stems. Two increment cores in perpendicular directions were extracted from each tree at 1.3 m above ground (294 utilisable cores), with the aim of avoiding reaction wood. Cores were mounted on wooden bars, dried and sanded. Samples were scanned using Epson Expression 10,000 XL scanner, and ring-widths were measured to the nearest 0.001 mm using WinDENDRO™ software (Regent Instruments Inc., Québec, Canada). Individual ring-width series were carefully cross-dated (Yamaguchi, 1991) and checked for missing rings and other cross-dating errors with the COFECHA software (Holmes, 1986). Following the cross-dating, pith age of yew trees at coring height was recorded from individual tree-ring series. In the case of partial cores (without pith) with inner arcs presented on the core, the age was estimated using the graphical method (Duncan, 1989). The age estimation was omitted for yew trees with partial cores without pith and inner arcs (5.4% of yew trees).
Stand basal area per hectare G (m2ha−1) of all tree species (including yew) was calculated for living trees Gtrees, standing dead trees and stumps Gnekro and for living beech trees Gbeech. Proportion of beech Pbeech determined as a ratio of Gbeech/Gtrees and expressed in percentage was also calculated at each sampling plot.
Moreover, for each sampling plot, we calculated a quantitative index of canopy closure (shading) based on the three-dimensional stereogeometry as proposed by Lieberman
Two-way ANOVA was employed to test the differences among localities and sex categories of the trees according to Gtrees, Gnekro, Gt, Gbeech and Pbeech (dependent variables). Additionally, mean values, their standard deviations and Pearson’s correlations among stand characteristics were calculated in the exploratory phase to reveal potential dependencies that can affect the final interpretations. For the same reasons, the relationship between stem and crown damage of European yew trees was tested using Pearson’s
To confirm that the crown damage is related to shading, the relationship between crown damage (categorical dependent variable) and stand characteristics (Gtrees, Gnekro, Gt, Gbeech and Pbeech—continuous predictor variables) was described by the generalised multiple regression model with ordinal multinomial distribution of response variable and logit (f(z) = log(z/(1-z)) link function. The model assumptions were checked by scatterplots showing relationships between observed values, predicted values and residuals. The overall quality of the model was assessed by several standard goodness-of-fit statistics (sum of deviations, Pearson’s
The basic dendrochronological characteristics were computed for each TRW series. The established standardised site chronologies were compared based on the mean correlations between trees (RBT) and expressed population signal (EPS). To explore sex-specific growth performance of yew trees, raw TRW chronologies were built for male and female trees within each site separately. In addition, cumulative growth curves were calculated based on the alignment of TRW series according to the pith age. For each age, we calculated the cumulative growth as a sum of previous TRWs, and for each subgroup, mean yield curves were obtained. Resulting mean chronologies were graphically compared and assessed. Similarly, mean raw TRW chronologies according to the crown and stem damaged and undamaged groups of trees were constructed and graphically compared within each site.
To test the second and third hypotheses, we used the general linear model. The model served for statistical detection of whether tree-ring widths (TRWs, dependent variable) of yews are influenced by the health status of the crown and by the stem damage caused. To remove age-dimensional differences among individual tree-ring series, diameter (DBH) entered the model as an independent continuous variable. Sex (male/female), crown status (damaged/undamaged) and stem status (damaged/undamaged) entered the model as categorical predictor variables. The three model versions were established using the TRW data from three different time periods covering last available calendar years on time series: one short-term period (2010–2015) covering last 5 most recent TRWs and two mid-term periods (2010–2015, 2000–2015) composed by the 10 or 15 most recent TRWs. Three periods were explored, and modelling results were combined because the information about the exact year of crown or stem damage in category damaged trees was not known. Therefore, a robust modelling approach was adopted.
To quantify the competition between a central yew tree
where
Another different and useful measure of competition that was used for the determination of release cutting rules was angle sineθ (component of canopy closure Gt):
,
where
Plotting the A-values against angle sineθ for all taller neighbours at each locality allowed to select the threshold A and sineθ value, allowing to formulate release cutting rules for individual yew tree. Threshold values were determined to remove the competitors having exponentially larger pressure on central tree indicated by both competition measures simultaneously.
In this approach, a neighbouring tree
Threshold A-value (Johann, 1982) is defined for each locality and determines the degree of release; other variables in formula 3 are described above. Tdist refers to the threshold distance for a given A-value below which a neighbouring tree is removed.
Yew trees differ among localities by age and the basic dendrometric characteristics of sampled trees (Table 1). Within localities on average, female yew trees possess smaller crowns and have lower heights, DBH and basal areas compared to male trees. In the localities PAV and STR, yew trees have mostly undamaged or weakly damaged crowns, while in LAG and VFA, undamaged crowns are lacking (Figs. 1 and A1). The largest degree of stem damage was recorded in the locality VFA. The degree of stem damage was related to the mean height of yew trees (Fig. A2). Results of testing the differences among localities and sex of the yew trees in the characteristics of surrounding stand Gtrees, Gnekro, Gt, Gbeech and Pbeech are summarised in Table A2 and Fig. 2. The interaction between locality and sex was not significant; therefore, results of the main effects of two-way analysis of variance are presented. Significant differences were found between female and male trees in stand characteristics Gtrees, Gbeech and Gt. Stands surrounding male trees showed significantly lower values than those surrounding female trees. No differences between male and female trees were found in Pbeech and Gnekro. Among localities, significant differences were found for all variables, and localities STR and PAV showed lower mean values of Gtrees and Gt and higher values of Gnekro compared to LAG and VFA. Locality PAV showed the lowest Gbeech compared to other localities. Based on the mean values and Pearson’s correlations among stand characteristics (Table 2), we can state that the surveyed stands show quite high proportion of beech. Positive relationship between stand basal area of trees (Gtrees) and stand basal area of beech trees (Gbeech) means that beech mostly participates in denser stand parts. At the same time, as it was expected, index of canopy closure (Gt) is positively related to Gtrees and negatively to Gnekro.
Basic description of sampled male and female trees of European yew at four localities in Slovakia
Locality | Sex | Trees | DBH (cm) | Height (m) | G(cm2) | Crown width (m) | Crown length (m) | Pith age (years) | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Nt | Mean | SD | Mean | SD | Mean | SD | Mean | SD | Mean | SD | Mean | SD | ||
♂ | 16 | 23.9 | 3.70 | 11.2 | 1.49 | 0.05 | 0.01 | 5.5 | 0.94 | 8.0 | 1.16 | 137 | 33.3 | |
♀ | 21 | 21.6 | 4.79 | 9.8 | 2.69 | 0.04 | 0.02 | 5.8 | 0.95 | 7.4 | 2.48 | 135 | 22.0 | |
♂ | 27 | 21.1 | 7.14 | 10.1 | 2.07 | 0.02 | 0.02 | 5.1 | 1.37 | 7.1 | 2.05 | 101 | 10.8 | |
♀ | 21 | 20.3 | 5.64 | 10.6 | 2.33 | 0.02 | 0.02 | 5.0 | 1.25 | 7.6 | 2.31 | 102 | 10.6 | |
♂ | 17 | 30.3 | 7.77 | 9.6 | 2.65 | 0.08 | 0.04 | 7.1 | 1.55 | 7.4 | 2.31 | 206 | 33.3 | |
♀ | 13 | 26.3 | 9.25 | 8.1 | 2.01 | 0.06 | 0.05 | 6.2 | 1.30 | 5.7 | 1.89 | 223 | 11.8 | |
♂ | 18 | 26.1 | 7.99 | 9.3 | 2.24 | 0.06 | 0.03 | 5.0 | 1.12 | 7.1 | 2.14 | 187 | 52.9 | |
♀ | 15 | 21.6 | 4.52 | 7.7 | 1.15 | 0.04 | 0.02 | 4.8 | 0.94 | 5.3 | 1.27 | 199 | 35.7 | |
○ | 2 | 27.0 | 4.17 | 8.8 | 0.78 | 0.06 | 0.02 | 4.7 | 0.40 | 6.2 | 1.06 | 235 | 5.0 | |
150 | 23.6 | 7.12 | 9.7 | 2.34 | 0.04 | 0.03 | 5.5 | 1.37 | 7.0 | 2.15 | 154 | 53.9 |
♂—male, ♀—female,
Results of ANOVA showing summary of all effects (locality and sex) of trees in stand characteristics Gtrees, Gnekro, Gt, Gbeech and Pbeech surrounding focal yew trees
Gnekro (m2ha-1) | Gtrees (m2ha-1) | Gt | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Df | SS | MS | F | SS | MS | F | SS | MS | F | ||||
1 | |||||||||||||
3 | |||||||||||||
1 | 0.21 | 0.21 | 0.0 | 0.96 | |||||||||
114 | 10557 | 92.6 | 5619 | 49.3 | 5900 | 51.7 | |||||||
118 | 15882 | 8567 | 10485 | ||||||||||
Df | SS | MS | F | SS | MS | F | |||||||
1 | |||||||||||||
3 | |||||||||||||
1 | 168.7 | 169 | 0.55 | 0.46 | |||||||||
114 | 4939 | 43.3 | 34859 | 305.8 | |||||||||
118 | 5697 | 40973 |
Gtrees—basal area of living trees, Gnekro—basal area of standing dead trees and stumps, Gbeech—basal area of living beech trees, Pbeech—proportion of beech, significant effects are marked bold, Df—degrees of freedom, SS—sum of squares, MS—mean square, F—test criterion,
Stand characteristics Gtrees, Gnekro, Gt, Gbeech and Pbeech surrounding central yew trees, their mean values and inter-correlations
Mean | SD | Gnekro(m2ha-1) | Gtrees(m2ha-1) | Gt | GBeech(m2ha-1) | PBeech (%) | |
---|---|---|---|---|---|---|---|
13.0 | 11.60 | 1.00 | 0.01 | ||||
31.1 | 8.50 | 1.00 | 0.04 | ||||
17.0 | 9.36 | 1.00 | 0.06 | ||||
18.6 | 6.98 | 1.00 | |||||
59.6 | 18.54 | 0.01 | 0.04 | 0.06 | 1.00 |
Gtrees—basal area of living trees, Gnekro—basal area of standing dead trees and stumps, Gbeech—basal area of living beech trees, Pbeech—proportion of beech, SD— standard deviation; significant correlations at
Crown damage was not related to stem damage based on the Pearson’s
Two-way table showing no relationship between stem and crown damage of European yew trees expressed in relative frequencies (Pearson’s chi-square = 15.69; Df = 9; p < 0.07)
Degree | Stem damage | ||||
---|---|---|---|---|---|
3.33 | 2.00 | 30.67 | |||
4.67 | 40.67 | ||||
2.00 | 6.67 | 4.00 | 23.33 | ||
0.67 | 2.00 | 2.67 | 0.00 | 5.33 | |
18.00 | 43.33 | 28.00 | 10.67 | 100.00 |
Observed frequencies > 7% are marked bold in a two-way table, relative frequencies were calculated from the total number of observations (Nt=150), Df—degrees of freedom,
Parameter estimates and goodness-of-fit statistics of the generalised linear regression model applied to test the relationship of stand canopy closure (shading) and crown damage (transparency) of yew adult trees.
Parameter estimates | |||||||
---|---|---|---|---|---|---|---|
Effect | Level of effect | Column | Estimate | SE | |||
Intercept n1 | 1 | 0.11 | 0.36 | 0.77 | 0.77 | ||
Intercept n2 | 2 | 2.48 | 0.44 | 0.00 | 0.00 | ||
Intercept n3 | 3 | 4.94 | 0.65 | 0.00 | 0.00 | ||
Gt | 4 | −30.09 | 0.02 | 0.00 | 0.00 | ||
Scale | 1.00 | 0.00 | |||||
Deviance | Scaled deviance | Pearson’s chi2 | Scaled P. chi2 | AIC | BIC | Loglikelyhood | |
Df | 359 | 359 | 359 | 359 | |||
Statistic | 263.04 | 263.04 | 383.01 | 383.01 | 271.04 | 282.23 | −131.52 |
Stat/Df | 0.73 | 0.73 | 1.07 | 1.07 |
SE—standard error,
In general, established chronologies expressed very dissimilar recent growth trends among localities and differed in basic dendrochronological statistics. The mean chronology at the locality LAG showed the highest RBT and the strongest population signal. On the contrary, the lowest RBT and the highest variation in TRW values were within the locality STR (Table 4). Based on the graphical comparison of male and female mean growth curves within localities, significant differences in growth occurred in certain periods, during which female trees showed contrasting growth trends compared to those of male trees (Fig. 3). Such periods lasted approximately 10 years, during which time sequences of sex-specific mean chronologies and corresponding bands of standard error (SE curves) did not overlap under sufficient sample depth. The exception is locality LAG, where mean sex subchronologies show similar time course of growth.
Basic statistics of European yew tree-ring series and standardised chronologies for the common periods in four localities of Slovakia
Locality | Raw series | Standardised chronology | |||||||
---|---|---|---|---|---|---|---|---|---|
Time span | Nt | MSL | TRW | SD | MSs | AC1 | RBT | EPS | |
1766-2015 | 37 | 127 | 0.713 | 0.44 | 0.251 | 0.81 | 0.526 | 0.976 | |
1897-2015 | 48 | 94 | 0.885 | 0.52 | 0.257 | 0.76 | 0.303 | 0.952 | |
1782-2015 | 30 | 204 | 0.551 | 0.32 | 0.236 | 0.83 | 0.187 | 0.869 | |
1755-2015 | 35 | 185 | 0.495 | 0.27 | 0.270 | 0.88 | 0.224 | 0.967 |
Nt—number of sampled trees, MSL—mean series length, TRW—mean tree-ring width (mm), SD—standard deviation, MSs—mean sensitivity of individual series, AC1— first-order autocorrelation coefficient, RBT—mean correlations between trees, EPS—expressed population signal calculated for the common interval, PAV—Pavelcovo; LAG—Lackov Grun; STR—Strazov; and VFA—Velka Fatra.
In recent mid-term period (2000–2015), the difference in growth between male and female trees became very apparent after regeneration cut on localities PAV and STR. The sex-related growth divergence becomes even more evident, when plotting subchronologies according to their cambial age (Fig. 4). Cumulative mean growth curves of STR and VFA express higher divergence compared to those of LAG and PAV. In localities STR and VFA, sampled yew
trees are on average twice as old as those in LAG and PAV (Table 1). Additionally, STR and VFA show higher within-site difference between male and female diameters (Table 1). Within-site significant differences in growth are apparent also for stem and crown damage-specific subchronologies (Figs. 5 and 6). However, growth divergence is not so pronounced. The subchronologies do not express opposite growth trends.
Using the general linear model, we revealed significant negative effects of crown defoliation and stem damage caused by bark stripping on growth performance of European yew (Fig. 7). Defoliation of tree crowns has led to recent short-term (2010–2015) and mid-term (2005–2015, 2000–2015) growth decline, regardless of the stem damage and sex. Unlike male trees, female trees with healthy crowns responded to stem damage with pronounced growth decrease. Surprisingly, within the category of damaged crowns and stems, there was no difference in radial growth between female and male trees.
Increased stand canopy closure (Gt or shading) relates to reduced dimensions of yew trees (DBH, height, crown width, crown length). Negative relationship between Gt and tree height and crown width was not significant in the locality STR, most likely because of narrow range of mainly low Gt values. Increased competition (A-value) reduces DBH of yew trees significantly. The correspondence between competition and canopy closure is shown in two localities with contrasting values of stand characteristics. Unlike the locality STR, locality LAG has among all studied localities the highest average value of stand basal area, canopy closure index and the lowest stand basal area of death trees (Fig. 2).
Based on the range of A-values, the competition pressure on yew is higher in locality LAG compared to that in locality STR (Fig. 8). The majority of trees, surrounding the central yew in the centre of sampling plot, show A-value of competition index in the range of 0–2 in both localities. For A-value above 2.0, the sine value of the respective angle reaches Gt mainly of more than 0.8. Thus, selection of the trees with the threshold value (Tdist) below 0 (at the A-value equal to 2.5) and Gt value equal to or above 0.8 would reduce the competition pressure and release crowns of yew from shad ing. Rules for determination of threshold values applied in cutting scheme on a particular locality are illustrated in Fig. 9 (illustrated for three selected sampling plots).
In the locality LAG, such selection represents an average reduction in Gt by 2.75 (maximum 5.81), intensity of cut calculated from basal area would be on average 11.3% (maximum 30.9%), and intensity of cut calculated from volume would be on average 14.9% (maximum 57%). In the locality STR, the respective threshold values for Gt represent 1.0 (max. 3.6), cutting intensity calculated for basal area 3.5% (max. 12.4%) and from stand volume 16.8% (max. 53%). While in the locality STR in 50% of sampling plots, the cut is not indicated; in LAG, only one sample plot is left without the cutting prescription.
This study demonstrates divergent growth between male and female trees (expressed in absolute terms) and the negative effects of crown defoliation and stem damage on growth performance of European yew.
Our results suggest that the differentiation in diameter growth between male and female trees likely began when the yew trees reached the life stage of sexual maturity, probably because of the assumed greater reproductive effort of female than male trees. This is in accordance with the findings of Cedro and Iszkuło (2011). In our study, the within-locality-detected differences between male and female trees in mean diameter are greater in older populations (STR, VFA) compared to those in younger ones (PAV, LAG, Table 1). On mean TRW chronologies, older populations showed more often the time sequences with contrasting growth divergence and lower between-tree correlations. Therefore, there are differences in the cumulative diameter growth as well (Fig. 4). This is in agreement with our first and second hypotheses. For reproducing trees, the long-term diameter growth trend related to age and size of trees is sex specific. Dioecy—an evolutionary reproductive strategy—likely aimed at attracting many seed dispersers, though the production of extraordinarily heavy crops of fleshy seeds, arils or cones is costly (Givinish, 1980). Thus, with progressing age of trees within the population and to this related time length of production of reproductive organs, the magnitude of differences in growth performance between female and male trees would increase. Scientific findings are relatively consistent in that male trees outperform female trees, have higher growth rates and are larger (Obeso, 2002; Mitchell
In our case, rapid increases in radial growth are likely responses of yew trees to canopy disturbances (either artificial cutting operations or natural disturbances) based on the known releases identified in later life stages (Fig. 3). In the locality PAV, regeneration cut (in 2000) triggered production of reproductive organs and hence growth divergence. Before the beginning of the regeneration cuttings, female and male yew individuals displayed similar growth rates (Sedmáková,
In the temperate forests of northern Europe, the yew is under threat mostly because of low light availability and high herbivory pressure (Dhar
Concerning the stem damage, interesting results on yew reports study from the Himalayas in India. Depending on the depth of the bark removed, it shows that bark removal up to 25% and even up to 50% around the stem circumference does not affect the survival of yew trees largely. The high mortality rates (over 50% decline) were recorded when bark was removed at the depth greater than 4 mm or around more than 75% of the stem circumference, which could be considered as critical limit for yew tree survival (Purohit
The size dependence would apply also to radial growth. We expect that the same degree of stem damage and the same degree of crown transparency would cause in absolute terms higher decrease in radial growth for larger- and faster-growing trees. Accordingly, differences in radial growth detected on raw mean chronologies in our work could simply be assigned to differences in sizes of trees between the categories of damaged and undamaged trees. To evaluate the effect of crown and stem damage on radial growth separately for female and male trees, differences in growth related to the size of trees need to be removed.
Based on the applied general linear models (Fig. 7), we register considerable recent growth decline in damaged European yew trees compared to undamaged ones in the limestone beech forests of the selected geographical region (in relative terms). Our results stress the importance of light being the most limiting factor as the highest growth divergence was found between trees with healthy-looking crowns (undamaged crowns) and trees showing certain degree of transparency (damaged crowns). The high shade tolerance did not compensate for recent growth decline, and there is no difference in growth performance between the sexes, although in the case of healthy looking trees, female trees even slightly overperform male ones. However, this difference is not statistically significant. Crown transparency is a more important factor causing the growth decline than bark stripping. On the other hand, vegetative regeneration capacity of yew showed to be sex related. Male trees compensated for recent growth decline caused by herbivory (combination of damaged stem and undamaged crown), and their growth was similar to the growth of healthy male trees (combination of undamaged stem and undamaged crown). Unlike male trees, the growth of female trees with damaged stems and undamaged crowns decreased substantially in comparison with healthy-looking female trees. Detected growth divergence may relate to higher cost of reproductive effort of female trees. At the time of sampling, all female trees showing no signs of crown transparency (with or without damage of stems caused by herbivory) were producing ‘fruits’. Overall, these findings are in agreement with the third hypothesis and disagreement with the fourth hypothesis. However, in one combination, in case of male trees, there was no pronounced difference between damaged and healthy-looking trees, and in this combination, male trees significantly overperformed female ones.
We expect the most pronounced sex-related differences in growth for strongly limited environmental conditions (low light availability in understorey). Under distribution limiting conditions, the percentage of female trees are found to positively relate with precipitation regardless of population age, which suggest their higher mortality rate on drier sites (Iszkuło
In our study, canopy closure is positively associated with the degree of crown transparency and the growth decline in yew. The number of healthy-looking yew trees in Slovakia is low, and the overall health status of yew is unfavourable. In forests of East–Central Europe, decreased growth and high rates of adults’ auto-reduction and extensive absence of regeneration threaten maintenance of yew populations (Linares, 2013). On the other hand, canopy openness positively relates to growth of European yew, and the size and growth rate of mother trees positively affect the amount of newly regenerated seedlings (Sedmáková
Based on the above-described reasons, although the hypothesis about worsened radial growth of damaged female trees is not confirmed, to enhance regeneration and improve growth of individual trees, importance of making silvicultural treatments for female trees as a priority is advisable. Releasing trees from competition after cutting will except for improved individual tree growth (Saniga, 2000) also reduce their mortality and ameliorate possible drought stress, what are facts important namely for female trees (Calev
To direct silvicultural prescription to be more specific with focus on individual yew trees, we propose a release cutting based on the combination of two competition criteria (Johan’s A-value and canopy closure index, Gt). Our aim is to provide a human-independent, quantitatively and spatially explicit, release cutting prescription (Pretzsch, 2009). The proposed selective cutting (thinning) gives the opportunity to promote individual yew trees and at the same time remains the rest of the stand without intervention or to incorporate proposed cutting within already defined forest management plan.
Intensity of cutting is defined by the selection of threshold values (Fig. 8). In case of the high canopy closure based on the selection of threshold values and higher intensity of cutting indicated (up to 20–25%, Fig. 9a), we suggest to separate the intervention into two consecutive cuttings with minimum time interval of 5–7 years. Sex-differing cutting intensity, interval and frequency can differentiate between higher susceptibility of female trees to stress by photoinhibition (Robakowski
The proposed cutting prescription in our study, however, neglects sex-related dimensional differences in yew trees, especially that female yew trees naturally have wider and shorter crowns than male trees. What needs to be mentioned is that selecting criteria are not species specific. The same cutting threshold value may indeed mean different levels of competition in case of different tree species. Species composition might be important to take into account because of different light transmissions. Canopies vary in composition and location above the forest floor, and in many cases, species-specific canopy transmission coefficients are not known. A more detailed estimation of competition may include additionally light transmission of neighbouring trees (Pretzsch, 2001). On the other hand, the study of Canham
Study of interactions between the stem and crown status, sex and growth performance of rare European yew growing in limestone beech formations in Central–East Europe provides several new findings:
Divergent absolute growth trends between male and female trees and the pronounced negative effects of crown defoliation and stem damage on the short- and mid-term radial growth performance of European yew.
The long-term radial growth trend related to age and size of trees is sex specific. After yew trees reach the physiological maturity and environmental conditions that are optimal to regeneration, significant sex-related differences in growth variability appear.
The considerable recent growth decline in damaged European yew trees compared to undamaged ones is revealed also in relative terms, after the exclusion of size-related effects from analysis. Crown transparency status is a more important factor causing the relative growth decline than stem damage.
Basically, no sex-related differences in growth of trees showing crown transparency were found, possibly due to exhaustion of resources
Regarding the need of specific management of female trees, female trees in most cases grow slower and are of smaller size than male ones in absolute terms. Under less optimal environmental conditions, female trees are expected to be more susceptible to stress. Female trees revealed worsened radial growth in comparison with male ones in case of stem damage caused by herbivory. Thus, to enhance regeneration and growth of yew populations, making silvicultural treatments for female trees a priority is recommendable. The release cutting rules to release the focused trees from canopy shading are formulated by innovative way.
Our findings can improve the effectiveness of forest management and restoration activities in European temperate forests, where yew adults are threatened by higher degree of shading and herbivory pressure. Many local population of yew are fragmented and isolated by distance, which further stress the need to improve their ecological conditions. Because of fragmentation and declining numbers of healthy individuals that reached maturity, the capability to safeguard population genetic variability may weaken. Data on growth rates along with survivor-ship of different-sized individuals in relation to canopy closure would likely provide further insights regarding the light dependency of adult yew trees. Damage to yew adults and regeneration caused by wild ungulates represents an important ecological and silvicultural issue along its entire natural range. The slow growth imposes yew trees to bark stripping for very long time, but this period can be shortened by cutting treatments aimed to improve growth and size of yew trees (as bark stripping damage seems to be tree size related).
In this study, we provide clear conservation message on the possibility to improve status of yew adults in managed forests with scattered occurrence of yew and so to maintain high species diversity of species-rich rare habitat—limestone beech forests. In the future, research could be directed to study in more detail intra- and inter-species competition of yew and how the competition may affect the vitality of individual yew trees in response to current climate change.