Towards Silviculture Guidelines to Produce Large-Sized Silver Birch ( Betula pendula Roth) Logs in Western Europe

: Forest health problems arising from climate change, pests and pathogens are a threat to the main timber tree species. As a result, silver birch ( Betula pendula Roth) has become a precious asset for meeting oncoming forestry challenges in western Europe. However, silviculture guidelines to produce high-value birch logs in this region are lacking. Producing large-sized birch trunks requires crown release, i.e., removing crown competitors around selected target trees. These interventions are currently seldom carried out or else too late when the growth potential of the trees has already diminished. This study set out to ascertain the diameter at breast height (dbh) that could be reached by crown-released birch, determine dbh-associated crown diameters, and further characterize the gain obtained from early crown release on birch dbh growth. We measured 704 birch trees that had undergone crown release in 38 naturally regenerated pure birch stands in southern Belgium and in northeastern France. We then evaluated the variation in stem and crown diameter, and analyzed increments in response to the earliness of the interventions in three subsamples, also compared with control target birch. We found that trees with a dbh of 50 cm could be grown within 60 years. Based on crown diameter, to produce 40, 50 and 60 cm dbh trunk, the distance required between target birch trees at the end of the rotation was around 8, 10 and 12 m. With no intervention and in ordinary dense birch regenerations, the dbh increment was found to decline once the stand reached age 4–7 years. Starting crown release in stands aged 4–5 years can double the dbh increment of target trees and provide a continual gain that may last up to 20 years. When birch crowns are released after 9–12 years, it may already be too late for them to recover their best growth rate. Our contribution should help complete emerging guidelines in support of birch silviculture development.


Introduction
One of the key recommendations for adapting European forests to global change is to diversify tree species composition [1][2][3][4].Silver birch (Betula pendula Roth), which has mostly been neglected until now in western European forestry, is one candidate.According to FAO reports, the birch resource (of both Betula pendula Roth and Betula pubescens Ehrh.) in this region comprises 1-15% of all hardwood standing volume, depending on the country, and is sharply increasing [5].Moreover, socioeconomic and climatic context is favoring the expansion [6] of this colonizer, which regenerates profusely in forests after disturbances such as clearcuts, storm damage and disease [7][8][9][10].Birch thus offers an attractive option to help meet present and future challenges of forest management [6,9,10].
In western Europe, birch is set aside because it colonizes young forest plantations, acting as a strong competitor of the planted trees.Most exploited birch timber is also poorly sized and of low quality (wrong stem shape, rot and knots) [6].Without proper silvicultural operations, birch rarely achieves the qualities needed for the high-value timber market [6,9,10].The wood industry has therefore been unable to make gainful use of the 1.
Identify the general dbh growth response to crown release to predict the trunk dimensions that could be reached at age 60 years (maximum age to limit the risk of wood degradation), 2.
Quantify and compare the gain in dbh increment of different crown release scenarios starting very early (at age 1-5 years) or late (at age 9-19 years) with a scenario without silviculture (control), 3.
Fit an allometric relationship between dbh and crown diameter to design suitable crop birch silviculture scenarios (e.g., minimum distance between target trees or number of target trees per hectare according to the dbh objective).

Study Sites, Plot and Tree Selection
The study sites were located in southern Belgium and in northeastern France (Figure 1), in Europe's temperate oceanic bioclimatic zone [1].To take as many opportunities as possible, we gathered a heterogeneous dataset of target birch trees that had undergone crop tree silviculture and had originated from pure naturally regenerated birch stands.The target birch trees were selected based on their own vigor and dominance status, their stem quality (straightness, no forks, large branches or wounds) and their distance from other target birch trees.We analyzed the growth of 704 crown-released target birch trees (number of trees at the last measurement) and 110 control target birch trees aged 4-66 years located in 38 stands.Four different datasets were gathered for the needs of this study (Table 1): - The first dataset contained 87 target birch trees aged 8-41 years.They were sampled in five experimental plots.Crown release and measurements were carried out for 12 to 19 years (sites 1 to 5). - The second one contained 142 target birch trees in three experimental plots (sites 6 to 8).The first crown release was carried out at different ages ranging from 1 to 19 years.These were compared with 110 target birch trees in control plots that were identified but never crown-released.- The third one contained 429 target birch trees aged 7-42 years.They were sampled in 15 plots.Crown release was performed by the local forest manager (sites 9 to 23). - The fourth one contained 46 forest birch trees that had grown almost free of competition.Given their crown development and the height of the first living branch, they were considered as target birch even though they had not formally undergone crop tree silviculture.These trees were sampled in 15 sites and were aged 27-66 years (sites 24 to 38).
ios starting very early (at age 1-5 years) or late (at age 9-19 years) with a scenario without silviculture (control), 3. Fit an allometric relationship between dbh and crown diameter to design suitable crop birch silviculture scenarios (e.g., minimum distance between target trees or number of target trees per hectare according to the dbh objective).

Study Sites, Plot and Tree Selection
The study sites were located in southern Belgium and in northeastern France (Figure 1), in Europe's temperate oceanic bioclimatic zone [1].To take as many opportunities as possible, we gathered a heterogeneous dataset of target birch trees that had undergone crop tree silviculture and had originated from pure naturally regenerated birch stands.The target birch trees were selected based on their own vigor and dominance status, their stem quality (straightness, no forks, large branches or wounds) and their distance from other target birch trees.We analyzed the growth of 704 crown-released target birch trees (number of trees at the last measurement) and 110 control target birch trees aged 4-66 years located in 38 stands.Four different datasets were gathered for the needs of this study (Table 1): - The first dataset contained 87 target birch trees aged 8-41 years.They were sampled in five experimental plots.Crown release and measurements were carried out for 12 to 19 years (sites 1 to 5).

-
The second one contained 142 target birch trees in three experimental plots (sites 6 to 8).The first crown release was carried out at different ages ranging from 1 to 19 years.These were compared with 110 target birch trees in control plots that were identified but never crown-released.

-
The third one contained 429 target birch trees aged 7-42 years.They were sampled in 15 plots.Crown release was performed by the local forest manager (sites 9 to 23).

-
The fourth one contained 46 forest birch trees that had grown almost free of competition.Given their crown development and the height of the first living branch, they were considered as target birch even though they had not formally undergone crop tree silviculture.These trees were sampled in 15 sites and were aged 27-66 years (sites 24 to 38).The study sites were mostly representative of fresh or slightly dry soils with poor or medium nutrient status (Figure 2), the most frequent soils in which birch trees grew well in the study area.The plots were located at 77-549 m (mean: 354 m) above sea level (Table 1).
Table 1.Description of the monitoring period, silvicultural treatment (age of first crown release) and ecological conditions of each plot.For each plot is also given the final number of monitored target trees together with the mean, minimum and maximum dbh at the last measurement.In cases of heart decay, the number of tree rings could not be counted and so tree age could not be estimated.

Total
The study sites were mostly representative of fresh or slightly dry soils with poor or medium nutrient status (Figure 2), the most frequent soils in which birch trees grew well in the study area.The plots were located at 77-549 m (mean: 354 m) above sea level (Table 1).Table 1.Description of the monitoring period, silvicultural treatment (age of first crown release) and ecological conditions of each plot.For each plot is also given the final number of monitored target trees together with the mean, minimum and maximum dbh at the last measurement.In cases of heart decay, the number of tree rings could not be counted and so tree age could not be estimated.The first crown release occurred in the sampled stand when the birch trees were 1-30 years old (mean: 10 years) (Table 2).Crown releases were repeated as often as necessary by the crown-touching method [32], i.e., all the competitive trees that touched the target birch crown were felled.Neighboring competitors that could touch the target birch with movements caused by the wind were also felled to prevent damage due to birch crown collisions [33].This subjective method was preferred to the cutting-distance method that systematically removes neighboring trees within a set distance from the target tree [32].The crown-touching method was found to be more efficient as it takes into account all particular cases such as bending trees [32].

Site
Table 2. Description of the studied crown released target birch trees at their last measurement (Only one measure per tree): Some target birch were harvested earlier than the others within their plot to prevent crown contact between target trees.They had been predesignated as extra target birch for a further designation or because of underestimation of crown development.For each target birch, we measured once or periodically: girth at breast height with a tape, tree height and the height of the base of the living crown with a Vertex IV hypsometer, and eight crown radii (i.e., the distance between trunk center and the vertical projection of the branch tip in the eight cardinal and intercardinal directions) with a compass, a Vertex IV hypsometer, and a Suunto clinometer to ensure vertically upward sighting.
Tree age was determined by counting the number of annual tree rings on wood discs taken at the stem base of minimum two dominant birch trees per stand.The rings were accurately counted with WinDENDRO software after sanding the wood discs to grain 320 and scanning at very high resolution (with an Epson Expression 10000 XL scanner).The age of some trees could not be estimated because the heart of the disc was rotten.

Statistical Analyses
Firstly, we used the whole dataset, except for the control target birch and the birch that could not be aged, to graphically analyze patterns of dbh evolution by individual, plot, site, or as a whole.This procedure enabled us to discuss our results considering the specific features of the different datasets and sites.
Secondly, we used the whole dataset, except for the control target birch trees, to model the allometry between tree crown diameter and dbh with a linear mixed-effects model (Equation ( 1)) taking into account that the observations were grouped by plot.Crown diameter was chosen as the response variable because measuring crown diameter is subject to greater error than measuring dbh.In addition, this choice was also justified by the fact that, generally, crown diameter is unknown and must be estimated while tree dbh is readily measured or available.The model was fitted with only one measure per tree.We chose the last measured crown diameter to use data from trees that had grown for as long as possible under crown release.Model fit was evaluated by computing the root mean square error (RMSE).
where C ij is the crown diameter of individual i (m) in plot j, β 0 and β 1 are fixed-effects parameters, dbh ij is the diameter at breast height of individual i (cm), α j is the random between-plot effect, and ε ij is the residual within-plot variation.
Thirdly, the effect of the earliness of crown release on dbh increment was analyzed with the data from sites 6, 7 and 8, including the control target birch trees.ANOVA and Tukey tests were carried out to test the significance of the effect of each treatment.Statistical assumptions were verified.As treatment varied across sites, and each had its own natural regeneration density, these analyses were repeated for each site.

Descriptive Statistics of Tree Increment
Taking all the measurements of each crown-released tree, the ages of the target birch ranged between 4 and 61 years and their dbh ranged between 1.2 and 58.6 cm (Figure 3) with an average of 18.6 cm considering their last measure.Most observations concerned birch younger than 35 years (Figure 3) as the mean age at the last measurement was 20 years (Table 2).The mean age at the first crown release was 10.4 years (Table 2), but the experiments with very young birch (e.g., sites 7 and 8) contained large numbers of individuals.
The dataset showed a wide dispersion of dbh growth trajectories.For example, at around age 40 years, target birch dbh ranged between 18 cm and 56 cm (Figure 3).
Our results showed that the best growing target birch could reach 40 cm in dbh at age 30-40 years or 50 cm in dbh at age 40-60 years.Nevertheless, some target birch trees failed to reach 10 cm in dbh at age 20 years.

Crown Development
A strong linear relationship was found between crown diameter and dbh (Figure 4 and Table 3).The root mean square error was 0.61.The data distribution is quite close to the regression line.For a given dbh, most of the crown diameters do not deviate by more than 1 m from the prediction, and the maximum range is about 2 m except for a few outliers (Figure 4).

Crown Development
A strong linear relationship was found between crown diameter and dbh (Figure 4 and Table 3).The root mean square error was 0.61.The data distribution is quite close to the regression line.For a given dbh, most of the crown diameters do not deviate by more than 1 m from the prediction, and the maximum range is about 2 m except for a few outliers (Figure 4).

The Effect of Crown-Release Earliness on Dbh
As expected, dbh increment was found to depend on silvicultural treatment (Figures 5-7 and Table 4).Figure 7A shows the mean growth curves by site and treatment (thicker lines) superimposed on the growth curve of all the sampled individuals up to 35 years (from Figure 3).Figure 7B shows variation in mean dbh increment in time by site and treatment.
In site 8, three treatments were applied: control, 5-year and 9-year, according to the age of the first crown release.When the trees were 6 and 7 years old, no statistical difference was found across treatments and dbh increment ranged between 0.6 and 0.8 cm/year (Table 4).Statistical differences were found once the trees were 8 years old.From age 8-9 years, the mean dbh increment of the trees released at age 5 years was about 0.9  In site 6, four treatments were applied: control, 1-year, 12-year and 19-year treatment according to the age of the first crown release.We observed that the earlier the first crown release, the larger was the mean dbh at the end of the monitoring period once the trees were 32 years old (Figure 5).At age 32 years, the average dbh of the target trees released when they were 1 year old was 34.6 cm (±3.1 cm standard deviation).Two out of the 10 birch even reached 40 cm in dbh.By contrast, in the 19-year treatment, the mean dbh was only about 22.9 cm (±4.1 cm) at 32 years old.At age 32 years, the average dbh was 7%, 21% and 62% greater than that of the control treatment for the 19-, 12-and 1-year treatment respectively.The mean dbh increments of the birches released at age 1 year was 1.3 cm/year (±0.7 cm/year) from 12 to 21 years.In the other silvicultural treatments, it ranged between 0.5 and 0.8 cm/year during this period.Birch released at age 12 years seemed to have gained little from the crown release.We did not observe any significant, sudden increase in the dbh increment of these trees but they did show a slightly greater mean dbh increment than those of the 19-year or the control treatment (Figure 5).Birch released at age 19 years seemed to have gained very little from late crown release, their dbh increment being comparable to that in the control treatment.
The differences between the mean dbh increments across treatments seemed to level off over time (Figures 5-7).The mean dbh increment was statistically significant across treatments considering a first period when the trees were 12-18 years old.During a second period when the trees were 19-25 years old, the dbh increment was statistically different across all the treatments except between the control and the 19-year treatments.Considering a last period, when the trees were 26-32 years old, the treatments showed quite similar increments but the trees in the 1-year treatment showed a slightly higher and significant dbh increment than those in the control treatment.
In site 7, two treatments were applied: the control and the 4-year treatment according to the age of the first crown release.The dbh of the target trees increased for 2 years after crown release and then stabilized at a value of about 1.2 cm/year (±0.2 cm/year) (Figure 5 and Table 4).In the control treatment, the mean dbh increments decreased continuously during the 6 years of monitoring.At the end of the monitoring period, when the trees were 10 years old, their mean dbh increment was half (0.6 cm/year ± 0.3) that of the trees that were previously crown-released (Table 4).
In site 8, three treatments were applied: control, 5-year and 9-year, according to the age of the first crown release.When the trees were 6 and 7 years old, no statistical difference was found across treatments and dbh increment ranged between 0.6 and 0.8 cm/year (Table 4).Statistical differences were found once the trees were 8 years old.From age 8-9 years, the mean dbh increment of the trees released at age 5 years was about 0.9 cm/year (±0.2 cm/year) whereas it was about 0.5 cm/year (±0.2 cm/year) in the other two treatments where no crown release had been carried out at that time (Figure 5 and Table 4).When the trees were older than 10 years, the dbh increment was significantly different in all treatments.Following the first crown release, mean dbh increment increased in the 9-year treatment and reached 0.8 cm/year (±0.2 cm/year) at age 10 years.The highest dbh increment was found in the 5-year treatment (1.0 ± 0.2 cm/year).The lowest was found in the control treatment (0.6 ± 0.2 cm/year) (Figure 5 and Table 4).

Limits of the Sample
To address our research questions, we built an original data set gathering information from different in situ experiments.This approach had some disadvantages.In particular, the number of observations was not well-balanced across sites, tree ages and treatments (Table 1 and Figure 3).It can be assumed that the between-and within-stand density variability before the first crown release influenced individual dbh/age ratio, and thereby the impact of the competition at a given age.We could therefore not use "age of first crown release" to distinguish the between-plot treatments as would have been appropriate in systematic plantations.Consequently, some analyses were carried out site by site.Also, for the most recent monitored trees and particularly for the younger trees, past competition history was well-known and crown release had been carried out earlier and more intensely, whereas for some other trees, particularly for the older ones, the past competition was not always known.Most of the trees that had been monitored for a longer period of time had likely experienced very different growing conditions, such as age (Table 1) and intensity of the first crown release, from those experienced by the last monitored trees.At this time, even if crown release started earlier and was stronger than for all the other common tree species, there was no knowledge of the real needs of birch species [9,26].Our study shows that the impact of late thinning on dbh growth in birch pure natural regeneration is quite low (Figures 5 and 6 and Table 4).

Crown Development of Target Trees
Crown diameter is one of the determining factors of dbh growth, linked to the photosynthetic capacity of a tree [39].We found a strong linear relationship between dbh and crown diameter (Figure 4) as already reported for some shade-intolerant pioneer species [23,38].Our model forecasts a larger crown diameter for a given dbh with a slightly higher slope than the model of [23,38] (Figure 4), established on naturally growing birch in forests in England, Scotland and Poland.This is probably because our sample focuses on crown-released target birch, consistent with the observations of Hein et al. [9].Our model of crown development is therefore designed for birch under crop tree silviculture and can be used for managing the crown space needed for optimal crown development in western European conditions.For example, if the goal is to produce 40, 50 and 60 cm dbh trunks, crown diameter at the end of the rotation, and thus the distance required between target birch, is around 8, 10 and 12 m, respectively (Figure 4).

Diameter at Breast Height Growth According to Age of First Crown Release
Tree competition can rapidly alter the dbh growth of naturally regenerated birch.Our results confirmed the findings of Prévosto et al. [21,22] and Lemaire [17].Without silvicultural intervention, the dbh growth can be reduced as early as age 4-7 years.We particularly evidenced this effect on birch aged 4-5 years in sites 7 and 8 (Figure 5 and Table 4).In these sites, between ages 4 and 10 years, mean dbh of the earliest crown released target trees grew on average 40% faster than those in the controls.At age 10 years, average dbh of the crown-released trees was about 1 cm greater than that of the controls.When the first crown release occurred a few years later, at age 9 years instead of 5 in site 8, target trees did not completely recover their dbh growth level in the following two years (Figure 5 and Table 4) and may keep a lower growth level thereafter.
Besides spurring the very early growth of target trees, early crown thinning may also have long-lasting effects [23].Between ages 12 and 24 years, the highest dbh increments were always observed for the trees that were released the earliest in site 6.At age 26 years, the trees that were released very early (from 1 years of age) were on average 9 cm larger than those released later, from age 12 years.The difference from the 19-year and the control treatments were 11 cm and 13 cm, respectively.There was thus no great gain between these last three treatments.Early crown release is therefore a way to reduce forest rotation [19,20] and thereby ensure that large logs are produced before wood depreciation occurs.

Timber Production Objective
In the ecological conditions studied, our results highlighted that trunks around 50 cm in dbh could be produced in less than 60 years, which is considered as the maximum harvest age, as older trees are too prone to wood rot and discoloration [8,9,27].This statement is a graphical interpretation (Figure 3) mainly based on the nine trees meeting these two criteria simultaneously in our dataset.Very few large-sized birches and birches older than 40 years were measured.Moreover, we can assume that most of the target trees older than 20 years probably did not receive a sufficiently dynamic treatment, in line with the negative consequences of competition on diameter growth, as discussed above.It can thus be considered that this objective is not overambitious.With appropriate crown releases, the objective of 60 cm dbh logs proposed by Hein et al. [9], Wilhelm and Rieger [26], Vanhellemont et al. [13] or in HOMBURG1 [40] seems realistic.Following market studies [9,10,26] and high-value birch log sales in Saarland [40], this size seems to correspond to the maximum target dbh for high-value silver birch logs.Alternatively, harvestable commercial birches of 40 cm dbh could be produced in less than 40 years.

Silvicultural Implications
Our study confirms that it is possible to produce high-value birch trunks with a dbh of 50 cm within a forest rotation of about 50-60 years.Furthermore, the size of trees, even for birch, and even for a short term in managed forests, plays an important role in the diversity of species that depend on it [41].Having such large logs rapidly in comparison with more shade-tolerant species is beneficial not only from an economic, but also from an ecological point of view.After this age, the risk of wood coloration and rot becomes high.Applying a crop tree silviculture with early crown release appears to be a good way to achieve this goal.Given the crown diameter of 50-year-old birch (Equation (1), Figure 4), forest managers can select about 100 target trees/ha.From a diameter growth point of view, crown release should start very soon after the canopy closes, i.e., as early as 4-5 years in ordinary dense pure birch natural regeneration, to avoid any likely persistent reduction in dbh growth of the target trees.This is most important for birch silviculture unlike some other species such as oak (Quercus sp.) or beech (Fagus sylvatica L.) e.g., [26] and even for other fast-growing species such as maple (Acer pseudoplatanus) or black alder (Alnus glutinosa) [42].Moreover, determining the optimal age of the first crown release depends on several factors including initial stand density, growth rate in relation with site conditions, desired log size and quality [6,28], and the cost of the early crown releases and associated pruning.

Conclusions
In the western European forestry context, silver birch, which naturally regenerates profusely, is a potentially useful resource to face ongoing changes.Our analysis of the dbh growth of 814 target silver birch trees in 38 naturally regenerated stands confirms that large-sized, high-value logs can be produced.
However, achieving this goal requires a specific silvicultural treatment adapted to the growth pattern and the sensitivity to crown competition of birch.Logs must be produced rapidly to limit the risk of wood coloration and rot, which appears as early as age 60 years.Birch needs dynamic silviculture to gain from its high but unsustainable early growth potential and to limit the rotation period.
In the context of dense natural regenerations (e.g., after clearcut), our findings are that crop tree silviculture with very early crown release, from age 4-7 years, is the key to producing large marketable trunks.Based on crown development of target trees, we also provide first guidelines for the spacing between target trees.

Figure 1 .
Figure 1.Site location and climatic gradient (mean annual temperature an annual rainfall) in the study area.Figure 1. Site location and climatic gradient (mean annual temperature an annual rainfall) in the study area.

Figure 1 .
Figure 1.Site location and climatic gradient (mean annual temperature an annual rainfall) in the study area.Figure 1. Site location and climatic gradient (mean annual temperature an annual rainfall) in the study area.

Figure 2 .
Figure 2. Soil nutrient and moisture levels of the sampled plots (site number) and birch production aptitude (cell color) based on the Afforestation Guide of Wallonia [31].

Figure 2 .
Figure 2. Soil nutrient and moisture levels of the sampled plots (site number) and birch production aptitude (cell color) based on the Afforestation Guide of Wallonia [31].

ForestsForests
2021, 12, x FOR PEER REVIEW 7 of 17 Our results showed that the best growing target birch could reach 40 cm in dbh at age 30-40 years or 50 cm in dbh at age 40-60 years.Nevertheless, some target birch trees failed to reach 10 cm in dbh at age 20 years.

Figure 3 .
Figure 3. Temporal variation of target tree dbh (trees from the control plots not included).The gray lines connect the successive measurements of a tree.

Figure 3 .
Figure 3. Temporal variation of target tree dbh (trees from the control plots not included).The gray lines connect the successive measurements of a tree.

Figure 4 .
Figure 4. Scatterplot between tree dbh and crown diameter.The modeled relationship is shown by the solid gray line (Equation (1)).The relationships fitted by Hemery et al.[38] is shown by the dashed gray line.

Figure 4 .
Figure 4. Scatterplot between tree dbh and crown diameter.The modeled relationship is shown by the solid gray line (Equation (1)).The relationships fitted by Hemery et al.[38] is shown by the dashed gray line.

Figure 5 .
Figure 5. Evolution of dbh (cm) and dbh increment (cm/year) with tree age for each silvicultural treatment of sites 6, 7 and 8.Figure 5. Evolution of dbh (cm) and dbh increment (cm/year) with tree age for each silvicultural treatment of sites 6, 7 and 8.

Figure 5 .
Figure 5. Evolution of dbh (cm) and dbh increment (cm/year) with tree age for each silvicultural treatment of sites 6, 7 and 8.Figure 5. Evolution of dbh (cm) and dbh increment (cm/year) with tree age for each silvicultural treatment of sites 6, 7 and 8.

Figure 6 .
Figure 6.Boxplot of the mean annual dbh increment during age intervals 12-18, 19-25, and 26-32 years in site 6.Significant differences were tested with a Tukey test.Significantly different groups are denoted with different letters.

Figure 6 . 17 Figure 7 .
Figure 6.Boxplot of the mean annual dbh increment during age intervals 12-18, 19-25, and 26-32 years in site 6.Significant differences were tested with a Tukey test.Significantly different groups are denoted with different letters.Forests 2021, 12, x FOR PEER REVIEW 12 of 17

Figure 7 .
Figure 7. Variation in dbh and dbh growth in sites 6, 7 and 8 across treatments.(A) shows the mean growth curves by site and treatment (thicker lines) superimposed on the growth curve of all the sampled individuals up to 35 years (from Figure 3).(B) shows variation in mean dbh increment in time by site and treatment.

Table 3 .
Parameters of the mixed-effects crown model.

Table 4 .
Variation of the mean dbh increment across sites, treatments and monitoring years.p-value levels of the ANOVA and Tukey tests are indicated next to the mean values (***: p-value < 0.001, *: 0.01 < p-value < 0.05, n.s.(nonsignificant): p-value > 0.05).Significantly different groups are noted with letters.