The goal of this study is to assess the impact of different thinning approaches for coppice conversion into high forest of Turkey oak stands in Italy. The stand structure and the tree/shrub diversity were analyzed in 27 long-term monitoring plots from 7 experimental trials in the Colline Metallifere district (Tuscany, Central Italy) to verify the consistency of the original cultivation goals with the current stand structures. Three different approaches were applied from 1969 onwards: thinning from below, selective thinning, and no-management. Three indexes of specific diversity (Specific Richness, Shannon index and Importance Value) and two indexes of vertical diversity (Vertical Evenness and Coefficient of variation of tree height) were used to analyze and compare the outcome of management practices. The results showed a significantly higher dimensional variability and basal area, and a more complex vertical diversity in control plots and in the plots subject to selective thinning, as compared with plots subject to thinning from below. Tree species richness was high in all plots, independently of the thinning type applied. Based on our results, we suggest Turkey oak-dominated transitory stands to be initially managed by thinning from below, which is easy to be implemented and economically feasible. Selective thinning may be applied later with the aim of promoting sporadic but valuable tree species and increasing tree species diversity.
Turkey oak coppices (
The main goals of coppice conversion into high forest, which has been undertaken in most countries since the second half of the 1900s on a share of coppice area in the public domain, were to: (i) increase the value of semi-natural forests,
Within the wide cover of Turkey oak-dominated stands, both coppice and high forest systems may provide significant contributions to the current demand of renewable energy and contribute to biodiversity conservation, soil protection, carbon sequestration, landscape quality and recreational value (
The priority goal of the conversion of Turkey oak coppice into high forest is to re-establish the original physiognomical, structural and tree compositional diversity. This is why its implementation should be limited to well-served and fertile areas, and special attention should also be devoted to valuable tree species other than Turkey oak (
This paper analyses for the first time a large dataset from seven experimental trials located in the Colline Metallifere - a main forest district in Tuscany (Central Italy) characterized by hilly and mountainous marginal contexts - with the aims of: (i) describing the methods applied in the research trials established from 1969 onwards by the current CREA (Council for Agricultural Research and Economics); the main differences between trials were type, intensity and frequency of thinning and/or age of first thinning; (ii) analysing the impacts of the different silvicultural approaches on stand structure and tree/ shrub compositional diversity, to verify the consistency of current stand structures with the original cultivation goals,
The dataset originates from 27 permanent plots for a total monitoring area of 5.6 ha distributed in 7 long-term experimental trials in the Colline Metallifere, South-West of Tuscany (Central Italy). This area is the main and wider system facing the Tyrrhenian coast, ranging from the plains up to an elevation of 1060 m a.s.l. Bedrock is made by argillite and clay-calcareous flysch (http://sit.lamma.rete.toscana.it/websuoli/). Soils are deep to shallow with a reduced differentiation of the vertical profile (
The main site parameters and characteristics are reported in
The silvicultural approaches varied across sites as a function of the final conversion goal: (i) building up pure, one-layered Turkey oak stands; (ii) establish Turkey oak high forests with a complementary layer of other tree species; (iii) rearrange into mixed stands where other valuable broadleaves are pooled to Turkey oak. Two different thinning methods and the no intervention choice were here applied, as described below.
(a) Thinning from below (Be) applied in 10 plots. This is the most common practice for coppice conversion into high forest. All the trees living in the dominated layer and, partly, in the co-dominant layer if badly-shaped, damaged or competing with dominant trees, were removed.
(b) Selective thinning (Se) applied in 12 plots. A positive selection was undertaken here to improve the growth of best-shaped trees removing only their direct competitors. Selected trees included the dominant species Turkey oak, but also the other valuable tree species, where present. This thinning method was implemented with a different timing within the trials: a relatively high number of candidate trees (1100 and 1500 ha-1) were selected in a juvenile stand age (20 years) in CS; a limited number of target trees was chosen in the other cases (80-150 ha-1) at higher stand ages (42 to 56 yrs).
(c) Control (Co), no silvicultural treatment applied (5 plots). Stand evolution followed strictly the natural post-cultivation dynamics.
Survey varied in time and was balanced according to the main goal of the experiment. Indeed, while periodical surveys of tree density, growth and productivity were implemented almost everywhere, more detailed surveys such as individual tree mapping and canopy parameters measurements, were carried out in a few specific plots.
The following parameters were recorded in each plot before and after each thinning: stand age, individual tree diameter at 1.30 m (DBH), tree height (measured on a representative sample to build up the tree height-DBH relationship), and the tree species assessment.
Data processing provided the number of trees per hectare (N), the basal area (Ba, m2 ha-1), the dominant diameter (Dd, the diameter of 100 largest trees, cm) and the mean diameter (DBH, cm). The above parameters were computed separately for each of three layers,
A set of indexes describing stand structure, complexity and tree biodiversity were calculated (
Three indexes of species diversity and two indexes of vertical diversity were selected to analyze and compare the effects of different silvicultural treatments on tree/shrub composition and stand structure. The Specific Richness index (
where
The third index was the Importance Value (
where
All the species diversity indexes (
Among the vertical diversity indexes, we used the Vertical Evenness (
where
The analysis of variance (ANOVA) was applied to test for differences in the above parameters and indexes among treatments (control, selective thinning, thinning from below). A preliminary analysis was firstly made to verify the assumptions of parametric models (
A multivariate linear model was fitted on plot level data using the above parameters as predictors of tree biodiversity (
where
The two thinning treatments applied in the Turkey oak coppices were significantly different in terms of Ba% and N% (TT = 0.65 in Be e TT = 1.16 in Se). The age of the first thinning was delayed in the case of thinning from below (90% of plots thinned the first time between 20 and 40 yrs), while it was earlier (within the age of 20 yrs) under selective thinning (Se). Thinning intensity was higher for thinning from below (55%) than for selective thinning (45% -
At the first monitoring inventory, the site index of the trials (
The basal area varied from 27.38 m2 in the control plots to 29.69 m2 in Be, without any significant differences among treatments because the time elapsed since the first thinning from below and the first stand inventory allowed the recovery of the removed basal area.
Significant differences (p<0.05) in mean DBH were observed between the plots thinned from below (DBH = 13.0 cm) and those selectively thinned (DBH = 9.4 cm) or unthinned (DBH = 9.5 cm).
Turkey oak presence was high in all plots, but the differences were not significant.
The evaluation of stand dynamics as a function of the applied silvicultural treatment was undertaken by comparing the values of three quantitative indexes: the current increment of basal area (CiBa), the tree mortality rate (M), and the difference between the dominant and mean diameter (Dd-DBH), calculated over the period between the occurrence of last thinning and the last inventory. CiBa reflects the stand growth performance, the mortality rate is dependent on the tree competition process, and the parameter Dd-DBH indicates the effect of thinning type on tree size structure.
The current basal area increment recorded over the last period (
The difference between the dominant and mean diameter (Dd-DBH) highlighted a reduced variability among both plots and treatments at the first inventory (
Stands had similar ages, ranging from 52 to 63 years at the last inventory (
A different distribution of trees and basal area becomes evident by splitting data into layers. The upper and middle layers had a similar number of trees in all the three treatments, while differences among the treatments were significant (p < 0.05) in the lower layer (171 trees ha-1 in Be, 810 and 1010 trees ha-1 in Se and Co, respectively). The three treatments showed significantly different (p < 0.05) basal area only in the upper layer (Ba = 35.56, 22.62, 29.33 m2 ha-1 in Co, Be and Se respectively).
The impact of the applied silvicultural treatment on tree specific diversity is shown in
The Shannon index (
The analysis of
The silvicultural treatment also influenced the relative importance of the other tree species in the different layers. While oak maintained its dominance in the upper layer, the other tree species became prevalent in the middle (M) and lower (L) layers in both Co (IVother
The index describing the stand structural variability within the vertical layers (
The results of multivariate model analysis are reported in
Thinning from below had the largest impact on stand structure when compared to the control plots and those under selective thinning, with a negative effect on dimensional variability (Dd-DBH), stand density (N and BA) and vertical structural diversity (CV). In contrast, thinning from below had a positive influence on productivity (CiBA) due to the heavy reduction of competition within all layers.
The selective thinning treatment allowed the maintenance of a high and structured tree species diversity (
Selective thinning, although requiring more attention, was successfully applied because the pre-existing tree specific diversity was composed of valuable tree species. In this case, thinning was applied early enough so as to allow the selection of candidate trees before the occurrence of tree reaction to thinning and the reduction of the commercial value of timber due to inter-individual competition. In general, selective thinning was less intensive than that from below.
The no-management option (control plots) or thinning in the dominant layer (
Both the maintenance of the dominated layer and the incremental reaction of selected trees led to high basal area values that were comparable to those in the control plots. The low mortality rate within plots thinned from below was due to the early reduction in competition in the main crop layer and the removal of the dominated and overtopped layers. Since high stand productivity is an index of efficiency and effective functionality (
Tree size structure in terms of mean DBH, dominant diameter, and their difference, is also directly related to the applied thinning type. Thinning from below led to low differences among trees in the stands, while selective thinning maintained an increased stand structural diversity and therefore high differences among trees, as compared to the control treatment. Moreover, the presence of large trees allow the maintenance of a wide diversity of bryophytes and lichens (
The maintenance/improvement of tree specific diversity is one of the objectives of the conversion of coppices into high forests. This approach was first applied to mountain coppices, where species richness is usually low and Turkey oak has been favoured since long time because more resistant to repeated coppicing, as compared to other species. Later, coppice conversion was applied in piedmont and hilly areas, where the stand composition is more complex and tree specific diversity may improve the adaptability and resilience of stands (
The results of this study suggest that stand diversity can be deeply affected by site characteristics, such as site index and tree species composition. However, thinning type plays a key role in regulating the tree density and therefore the structural diversity of the stand. Indeed, the heavy reduction in stem number and the vertical simplification due to thinning from below contributed to increase the share of
Our results show that tree species diversity and the time passed since the last thinning significantly affect the vertical stand structure in Turkey oak coppices. Thinning from below clearly simplified the vertical stand structure, though the prompt resprouting of thinned stools established a subsidiary layer a few years later, thus increasing the structural diversity of the whole stand.
Regarding the selective thinning treatment, the maintenance of the subsidiary layer, as well as the application of localized thinnings around a few valuable selected trees, allowed the conservation of both specific and vertical diversity, where this was a pre-existing feature (
The parameters and indexes analysed in this study are effective descriptors of the different phases of coppice conversion into high forests. The transitory phase analysed here should aim to enhance the ecological functions of the stands and their ecological value, making use of easily applicable forestry techniques. In the analysed stands, the silvicultural practices applied over the last decades were consistent with the original purposes of either (i) favouring Turkey oak or (ii) setting up mixed, more structured stands.
Turkey oak dominance in the studied coppices may be usefully reduced where other tree species are already present in the transitory phase. The post-coppice cultivation dynamics largely depends on site characteristics and local tree species composition. Wherever present, the dissemination and regeneration of other species should be promoted by suitable silvicultural practices, from selective thinning up to tree-oriented silviculture. Turkey oak-dominated stands may be managed, at least in the first phase, by thinning from below which is easy to implement and economically feasible. Selective thinning may be pursued later when target trees may be confidently identified, with the aim of promoting valuable tree species and favouring their conservation in the next cycles. The release of an adequate canopy cover in the upper layer may control both the resprouting and vitality of cut stools in fire-prone environments.
Natural regeneration in transitory Turkey oak coppices and their timber production are critical issues that are still open and will be approached in the next future.
We wish to thank all the technicians of CREA/CRA/ISSEL who have cooperated in the establishment of the monitoring network and in data collection. The authors received no specific funding for this work. The authors declare that no competing interests exist.
Mean, standard error and standard deviation of current basal area increment (above) and mortality rate (below) recorded over the last period at each silvicultural approach. Significant differences (p<0.05) are marked with different letters.
Differences between dominant diameter (Dd) and mean diameter (DBH) according to the applied treatment at the beginning (above) and end (below) of monitoring time. Significant differences (p<0.05) are marked with different letters.
Box-plots of specific richness (
Box-plot of the Shannon index in the different vertical layers per silvicultural treatments. Significant differences (p<0.05) are marked with different letters.
Box-plots of the coefficient of variation (
Main climatic data. (Ta): mean annual temperature; (Ts): mean summer temperature; (Ra): mean annual rainfall; (Rs): mean summer rainfall; (Ia): Ra/(Ta+10) mean annual dryness index (de Martonne); (Sd): summer dryness and soil type at the research sites.
Researchsites | Meteorological site (elev.) | Ta (°C) | Ts (°C) | Ra (mm) | Rs (mm) | Ia | Sd (days) | Soil Type |
---|---|---|---|---|---|---|---|---|
FON FOS FOT POG TRO | Monterotondo(670 m a.s.l.) | 12.1 | 20.5 | 1112 | 113 | 50 | 60 | Typic Ustorthents, loamy-skeletal, mixed, calcareous, mesic, shallow |
MR | Chiusdino(450 m a.s.l.) | 13.4 | 21.4 | 827 | 114 | 35 | 62 | Typic Dystrusteps, loamy-skeletal, siliceous, mesic |
CS | Monteverdi(300 m a.s.l.) | 13.3 | 21.3 | 896 | 92 | 38 | 81 | Typic Ustorthents, fine-loamy, mixed, calcareous, mesic |
Main site characteristics, design of trial, applied silvicultural approach and monitoring time at the 7 experimental sites. (Co): no-management (control); (Be): thinning from below; (Se): selective thinning.
Characteristics | Treatment | TRO | FON | FOT | FOS | POG | MR | CS |
---|---|---|---|---|---|---|---|---|
Elevation (m a.s.l.) | 720 | 610 | 585 | 580 | 570 | 325 | 305 | |
Aspect | N | NO | NE | E-NE | E-NE | O | N-NO | |
Slope (°) | 8 | 10 | 8 | 10 | 10 | 12 | 15 | |
Plots (n) | 2 | 2 | 2 | 1 | 2 | 6 | 12 | |
Plot size (m2) | 2500 | 2500 | 3200 | 5000 | 4700 | 2400 | 900 | |
Silvicultural approach | Co | - | - | - | - | 1 | - | 4 |
Be | 1 | 2 | 1 | 1 | 1 | 4 | - | |
Se | 1 | - | 1 | - | - | 2 | 8 | |
Thinnings (n) | Co | - | - | - | - | - | - | - |
Be | 1 | 1 | 2 | 2 | 1 | 1 | - | |
Se | 2 | - | 2 | - | - | 2 | 2 | |
Stand age at the first inventory (yrs) | 42 | 34 | 47 | 38 | 35 | 55 | 20 | |
Stand age at the last inventory (yrs) | 50 | 47 | 54 | 45 | 55 | 57 | 65 |
Thinning trials (type, frequency and intensity) at the experimental plots. Significant (p<0.05) differences in rows are marked with different letters.
Characteristics | Class/Stat | Thinningfrom below | Selectivethinning |
---|---|---|---|
Number of applied thinnings | - | 1-2 | 2 |
Age at the first thinning (% of plots) | ≤ 20 yrs | 10 | 67 |
20-40 yrs | 90 | 33 | |
Elapsed time (yrs) between thinnings | Mean ± SD | 20.00 ± 5.83 a | 20.00 ± 4.82 a |
Thinning type (TT) | Mean ± SD | 0.65 ± 0.15 a | 1.16 ± 0.35 b |
Intensity of thinning (% of plots, according to basal area removed) | ≤ 20 % | 25 | 32 |
20-35 % | 25 | 63 | |
> 35 % | 50 | 5 | |
Increment and Felling (% of felling to basal area increment) | Mean ± SD | 55.0 ± 11.8 a | 45.0 ± 15.4 a |
Main stand parameters (mean values ± standard error) for stand age, tree dominant height (DH), number of stems (N), basal area (Ba), mean tree diameter (DBH), Importance Value index for
Parameter | Control | Thinningfrom below | Selectivethinning |
---|---|---|---|
N of plots | 5 | 10 | 12 |
Age (yrs) | 23 ± 2.7 | 38 ± 2.1 | 25 ± 3.3 |
DH (m) | 18.0 ± 0.2 a | 18.6 ± 0.8 a | 17.6 ± 0.5 a |
N total (n ha-1) | 4347 ± 180 a | 2681 ± 400 b | 4448 ± 442 a |
Ba (m² ha-1) | 27.38 ± 1.24 a | 29.69 ± 1.80 a | 27.94 ± 1.47 a |
DBH (cm) | 9.0 ± 0.2 a | 13.0 ± 1.2 b | 9.4 ± 0.8 a |
IVQc | 0.70 ± 0.05 a | 0.71 ± 0.05 a | 0.82 ± 0.08 a |
Main stand parameters (mean values ± standard error) for tree age, dominant tree height (DH), number of stems (N), basal area (Ba), mean tree diameter (DBH) recorded at the last inventory (see
Variable | Layer | Control | Thinningfrom below | Selectivethinning |
---|---|---|---|---|
N of plots | 5 | 10 | 12 | |
Age (yrs) | 63 ± 4.5 | 52 ± 4.5 | 61 ± 5.6 | |
DH (m) | 25.7 ± 3.5 a | 20.7 ± 2.5 b | 24.7 ± 2.8 a | |
N total (n ha-1) | All layers | 2041 ± 533 a | 1082 ± 604 b | 1697 ± 785 a |
Upper layer | 666 ± 257 a | 620 ± 248 a | 550 ± 178 a | |
Medium layer | 365 ± 327 a | 291 ± 286 a | 338 ± 209 a | |
Lower layer | 1010 ± 325 a | 171 ± 232 b | 810 ± 734 a | |
Ba (m2 ha-1) | All layers | 39.99 ± 4.77 a | 24.99 ± 6.56 b | 33.49 ± 7.96 a |
Upper layer | 35.56 ± 5.51 a | 22.62 ± 4.84 b | 29.33 ± 6.44 c | |
Medium layer | 2.48 ± 1.51 a | 1.61 ± 1.93 a | 2.49 ± 1.56 a | |
Lower layer | 2.00 ± 0.79 a | 0.23 ± 0.30 a | 1.67 ± 1.47 a | |
DBH (cm) | 16.1 ± 2.7 a | 18.3 ± 3.6 a | 16.8 ± 3.3 a |
Multivariate model analysis of the selected diversity indices. (df): degrees of freedom; (S) specific richness; (N): tree density (number of tree ha-1); (DH): dominant height; (VE): vertical evenness; (SH): Shannon index.
ResponseVariable | Predictor | Sum ofSquare | ExplainedVariance (%) | Effect | df | F-value | Prob(>F) |
---|---|---|---|---|---|---|---|
Shannon Index | S | 0.889 | 35.52 | + | 1 | 56.705 | 8.24E-07*** |
N | 0.153 | 6.11 | - | 1 | 9.758 | 0.0006** | |
DH | 0.350 | 14.01 | + | 1 | 22.366 | 0.0001*** | |
VE | 0.481 | 19.24 | + | 1 | 30.717 | 3.58E-05*** | |
Residuals | 0.266 | 10.65 | - | 7 | - | - | |
Vertical Evenness | Time from last thinning | 0.097 | 7.98 | + | 1 | 8.602 | 0.0085** |
S | 0.127 | 10.52 | - | 1 | 11.336 | 0.0032** | |
SH | 0.606 | 49.85 | + | 1 | 53.738 | 5.98E-07*** | |
Residuals | 0.214 | 17.62 | - | 19 | - | - |