Of the total roughly two million hectares of forests, mixed uneven-aged silver fir (Abies alba Mill)—European beech (Fagus sylvatica L.) (hereafter fir-beech) forests are the most important forest type in Dinarides.
Given the overall fir distribution and its frequency in the mixture (the share in the forest ecosystem) along with the Pyrenean area and Apennines, the Dinaric area represents a distinct southeastern edge of distribution. Within the southeastern border areal in the Croatian Dinaric region, besides sites favourable for fir, there are also localities at the micro borders (edges) of its appearance.
Historically, the use of forests in Europe until the beginning of the 19th century was unplanned and extensive. Moreover, it was characterised by individual logging, often aiming at high quality trees with a large diameter, similar to unregulated uneven-aged management. In the mid-19th century, uneven-aged forest management occurred in the Dinaric region.
Reasons for the introduction of uneven-aged management in primeval old-growth forests are still unknown since, at that time, the clear-cut prevailed in the rest of Europe. It can be assumed that the driver for the establishment of uneven-aged management was the specific structure of fir-beech forests due to their high variability in soil depth and rockiness, resulting in a seemingly uneven-aged stand structure.
Previous Forest Management
Over the last 100+ years, management in Dinaric fir-beech forest has gone through several changes, from single tree and small group selection to ‘freestyle forest management’.
Even and uneven-aged management has its advantages and disadvantages. However, the general trend of European forestry and worldwide forestry is to emphasise the advantages of forest management that is close to nature.
Thus, in the middle of the 20th century, uneven-aged management in the Dinaric region prevailed, clear-cutting was forbidden, grazing harvest was limited and reduced. At present, more than a third (33.4 percent) of Dinaric forests is managed by an uneven-aged approach. Furthermore, in Croatia, the use of uneven-aged management in all fir forests (about 30 percent) is obligatory by forest policy regulations.
Due to different circumstances, uneven-aged management was not consistently performed or did not yield satisfactory results. Some of the causes were fir dieback affected by air pollution and SO2 emissions, absence of natural regeneration, herbivorous influence on young fir trees, constant increase of beech abundance, the impact of different ownerships and consequent management approaches and various degrees of nature protection (absence of management). Consequently, fir-beech forests with dissimilar structures occurred.
This was mostly due to different interpretations of uneven-aged management and environmental impacts. Recent research of Dinaric forests in southeastern Europe has revealed the existence of pronounced dynamics in stand structure and composition, both in uneven-aged and old-growth forests, with the particular stand dynamics influenced by the habitat conditions.
Numerous studies have acknowledged that changes in fir stand dynamics and structure are reflected by the increased proportion of large mature trees, difficulties with regeneration and fir decline. This was recorded in Croatia, Bosnia and Herzegovina, Slovenia and other Central European countries.
Therefore, in the Dinaric region, similarly to Central Europe, uneven-aged fir-beech forests are characterised by the large and accumulated growing stock of approximately 350 cubic metres/hectare , this being favourable for fir as a shade-tolerant species.
Forests Getting More And More Attention
Changes in fir forest ecosystems and stand dynamics can be related to the concept of forest decline, which was introduced in the early 80s of the last century in Central Europe. Polluted air, especially SO2 emissions in interaction with climate change (rise in temperature and drought) and biotic factors are considered to be the major causes of fir decline (reduction of vitality, increase of susceptibility and mortality rate) and growth decrease.
Due to recent changes and growing needs for the maintenance of forest resources, interest in and acknowledgement of importance of long-term monitoring in Europe has increased. However, until now, there has been no systematic monitoring and comprehensive research within the entire region of fir forest in Croatia.
Previous studies were mostly oriented to individual and isolated research in the context of adverse environment and management influences. Hence, the first national forest inventory (hereafter NFI) in Croatia, conducted from 2006 to 2009, was the main basis for the establishment and conduction of the entire monitoring system for fir forests in the Croatian Dinaric region.
This comprehensive approach consists of a repeated inventory on 74 permanent sample plots, detailed monitoring of climatic and environmental factors and tree growth on the base plots carried out in 2019.
Thus, through long-term monitoring, the influence of climatic and environmental factors and management regimes can be obtained for the study of stability enhancement of stand structures, maintenance and how to increase their adaptive ability to environmental changes.
The importance of primary (old-growth) forests for biodiversity maintenance and climate change mitigation has been enhanced with many recent studies.
Networks of such forest reserves under varying protection regimes provide the main requirements for the fulfilment of habitat and nature protection (i.e., Natura 2000 sites) within the concept of sustainable and multifunctional forest ecosystem management.
The obtained stand structure dynamics in long-term studies of old-growth forests could be useful for the emulation of natural processes in sustainable forest management.
In the Croatian Dinaric region, there are no fir and beech forests that have been out of forest management activities for a long past period (more than 80 years), except for one or two small isolated primary forest patches.
Furthermore, the larger part of the forest region, including nature parks, has been under the implementation of Natura 2000 requirements and forest management restrictions only for the last decade.
However, the area of about 30,000 ha of forest within two national parks has been out of any management activities during the last three decades. Hence, we believe that these forests could be appropriate for studying initial natural processes and stand structures that are closer to the old-growth forest structure, and that they would differ from managed forests.
Besides unmanaged forests within the national parks, three types of managed forests in the region can be distinguished by site characteristics, management systems and ownership category.
The aim of this study was to compare current stand structure characteristics, past dynamics and stand stability of unmanaged forests within national parks (UM-NPF) and managed forests (state forests on carbonate bedrock (M-S1F), state forests on non-carbonate bedrock (M-S2F) and private forests (M-PF)).
We hypothesised that (1) stand structure characterised by high basal area, a large share of beech, a large number of large live trees and snags and large mean diameter at breast height (DBH) in unmanaged forests would be recognised as old-growth attributes; (2) current stand structure characteristics and recent past dynamics of stand structures would differ between unmanaged and managed forests and also between different types of managed forest.
Study Area & Analysis Methods
The study was carried out in the entire range of fir-beech forests within the Croatian part of Dinaric Alps. These forests encompass 304,000 ha of forest area including the dominant forest community classified as Abieti-Fagetum and Blechno-Abietetum with a share of 96.4 percent and 3.6 percent respectively where beech, fir and Norway spruce (Picea abies (L.) Karst) are dominant species.
Altitudinal vegetation belt ranges between 600 and 1,100 m above sea level. Limestone–dolomite substrates with basic soils (cambisols and leptosols) prevail, while silicate bedrock with dystric cambisols and podzols are less represented.
Region climate is characterised by relatively low average annual temperature (6–10 °C), high humidity (>80%) and high annual precipitation (1500–2500 mm). The amount of precipitation gradually decreases from the north west to the south east part of the region, while the average annual temperature increases.
In the region, small-scale private forests account for 4.6 percent of the total forest area, while 95.4 percent are state forests including managed forests and different forest categories of nature protection under restricted management (nature parks) or without any forest management activities (strict reserves, national parks).
The two national parks (Plitvce Lakes and Risnjak) encompass nine percent of state forests that have been out of management for the last three decades. Due to variety of site and stand characteristics and past management approaches, the area is characterised by single-group selection management system (60%), and different even-aged and irregular (transitional) forest stand structures (40%).
Selection management in fir-beech stands is characterised by single stem (heavy sites, high rate of rockiness and large slopes) and small group felling (diameter of canopy gaps between one and two canopy heights) in more favourable site conditions and silicate bedrock and with 10-years cycles of an average cutting intensity of 21 percent or 76.5 cubic metres of cut volume per hectare.
Sample plots from the NFI, established and inventoried in 2006 and 2007 (725 plots or 11.6 percent of overall NFI plots within the entire Dinaric forests), were used for the sampling design.
Within each of the previously defined strata (type of management regime), sample plots were randomly selected from the set of NFI plots. A minimum of 15 plots by defined forest strata and share of fir growing volume at least 30 percent were main requirements of the plot selection. Still, due to the insufficient number of sample plots where fir is represented at all, only five plots were selected in private forests.
On 74 selected sample plots, re-measurement (second inventory) was carried out in 2019 using approaches similar to the first inventory. The four concentric plots were used to assess tree-specific variables.
We recorded all standing trees (living and dead) that exceeded a pre-established diameter at a breast height (DBH) threshold in each circle: 5 cm for the 3.5 m radius circle, 10 cm for the 7 m radius circle, 30 cm for 13 m radius circle and 50 cm for the 20 m radius circle.
The following tree-specific variables were measured and assessed: tree species, tree position (azimuth, horizontal distance and inclination), DBH, tree height (only in the first inventory), height up to the crown base and two crown diameters of firs (only in the second inventory), tree canopy layer, stem quality, bole damage, crown damage, crown defoliation and the ‘stump age’ of a cut tree.
Furthermore, to estimate the consistent growing volumes between the two inventories (mortality, increment, recruitment and cut), the following sampling categories were assessed for each tree in the second inventory: survivor tree, on-growth tree, ingrowth tree, died tree, cut tree and wrong tree sampling (included, excluded) in the first inventory.
The circle plot radius of 13 m was used to assess lying dead wood according to the number of stems, diameter classes, degree of decomposition and tree species (conifers, broadleaved). The two small concentric circle plots (radii of 1 and 2 m) within the 13 m radius circle plot were used to assess seedling and sapling density. The large circular plot radius of 25 m was used for the assessment of management-specific, site-specific and stand-specific variables.
Calculation of quantitative stand structure characteristics on the sample plot level (tree, basal area and growing volume distribution by tree species and diameter classes) were based on per hectare values for the stem number, represented by each sampled tree. Each sampled tree was defined with its tree species, DBH, tree sampling category (survival, on-growth, ingrowth, died, cut) and corresponding circle plot and on summing per hectare values of all sampled trees.
The obtained data from the 2019 inventory year were compiled and averaged from sample plot level and represented in table and figure forms to visualize and compare current structural characteristics of unmanaged and managed forests.
Statistical differences between the studied forests in terms of current characteristics (variables) and stand structure dynamics were assessed with Mann-Whitney U tests. Diameter distributions for total and fir tree numbers were obtained as average values of five cm DBH class midpoints and presented with logarithmic scale (base 10).
The shapes of the DBH distributions were analysed and described following the methodology presented in Leak. Stand structure was evaluated with the estimation of closeness of the diameter distributions for total tree number to the inverse J-shaped distribution. The LikeJ index was estimated. Stand density index (SDI) was calculated by the summation method for the uneven-aged stands.
Components of stand dynamics between the two inventories at the sample plot level (growing volume, standing dead volume, mortality, increment, recruitment, cut and net changes) were estimated using procedures.
In total, 34 site-specific and stand-specific attributes were computed in each sample plot (2019 inventory).
The attributes for management regime types were statistically compared using nonparametric Mann-Whitney U tests. The analysis was performed in Statistica 13.5 software.
In order to identify possible key gradients of structural variation between unmanaged forests in national parks and managed forests, a principal component analysis (PCA) for site, stand and management specific attributes was carried out. PCA was performed using the function PRCOMP in R software.
Characteristics Of Current Stand Structures
Unmanaged forests (UM-NPF) showed significantly higher overall growing volume (666 m3 ha−1), overall stand basal area (48.9 m2 ha−1), density of large (84.8 trees ha−1) and very large trees (23.0 ha−1), overall mean DBH (40.6 cm) and lower LikeJ index (3.7) compared to managed M-S1F forests.
No significant differences between UM-NPF and state M-S2F forests in growing volume, overall tree density, density of large and very large trees, overall mean DBH and LikeJ index were observed.
Managed private forests (M-PF) exhibited significantly higher tree density (884 trees ha−1), stand density index (892), small diameter trees (688.6 trees ha−1) and LikeJ index (7.5) and the lowest overall mean DBH. Dead wood and density of stand regeneration differed between all types of management regimes, but not significantly.
The best stand regeneration regarding overall seedling (h ≤ 130 cm, ha−1) density was observed in M-S2F and M-S1F forests, with more than 30,000 and 13,000 seedlings, respectively, while considerably lower seedling density (two times or more) was obtained in unmanaged forests and managed private forests.
Unmanaged and managed private forests showed similarity in the relative tree density and basal area of beech, which was higher than in managed state forests where the relative basal area of fir reached 50 and 75 percent, respectively.
A large share, in relative tree density, of beech and other species was obtained within the first size class (10 ≤ DBH < 30 cm) in all forest types with the achieved share of fir of about 30 percent in managed forests and only 13 percent in unmanaged forests. The share of fir relative density showed a systematic increase along with successive tree size classes. A similar pattern was obtained for both unmanaged forests and three types of managed forests.
Considering diameter distribution for all tree species, unmanaged and M-S2 forests showed similar increasing Q shape (IQ) with a moderate initial reduction in the number of trees per DBH classes, followed by the increased decline of tree number in DBH classes > 60 cm. A more or less constant (negative exponential) and variable (rotated sigmoid) reduction rate of overall tree number per diameter classes was observed in M-S1F and M-PF forests.
Unmanaged and managed forests showed slight differences by PCA. Along a direction from managed to unmanaged forests, there was a decreasing of overall fir and beech and other broadleaves (OB) harvest volume, decreasing of recruited trees per ha annually and increasing of total basal area, average DBH, crown defoliation of firs and basal area of died beech trees between two inventories and the number of very large trees.
The line of the two types of management gradient was relatively orthogonal to overall dead wood, standing dead wood, volume and number of large and very large snags, basal area of died firs between two inventories, number and volume of very large live trees.
Dynamics Of Stand Structure
With the larger and similar initial growing volume (GV) of approximately 610 m3 ha−1 obtained in UM-NPF and M-S2F forests, during the period between two inventories, the growing volume increased in the unmanaged forests while decreased in the managed forests.
In UM-NP forests, the increase of overall growing volume was higher compared to fir growing volume, while in M-S2 forests higher decrease was for overall GV than fir GV. The initial growing volume in M-S1F and M-PF forests was also similar but much lower than in the other two forest types, which slightly decreased and increased in M-S1F and M-PF forests, respectively.
The gross volume increment (GVI) in M-S1F forests was significantly lower than in UM-NPF and M-S2F forests in which GVI amounted to approximately 10 m3 per ha annually. This resulted in similar net volume increment (around 5 m3 ha−1 annually) and significantly smaller natural loose in M-S1F forests. M-PF forests were characterised by the highest net volume increment and lowest natural loose.
The highest relative natural loose of fir (>70%) was in both managed state forests, while in M-PF forests beech represented overall natural loose. Furthermore, DBH structure of natural loose showed that in both managed state forests the highest rate of natural loose was in medium diameter trees (DBH 30–50 cm), while in M-PF and UM-NPF forests was in very large diameter trees (DBH ≥ 70 cm).
The managed state forests, M-S2F and M-S1F, were characterised by cut volume larger than net volume increment, amounting to 178 and 113 percent of net volume increment, respectively. In M-PF and UM-NPF, the cut volume amounted to 48 and 18 percent (trees harvested on plots close to walkways due to safety of visitors) of net volume increment, respectively.
The share of fir in total cut volume was the highest in all forest management regime types, while DBH structure of cut trees differed among studied forests. The net changes corresponded to differences between net volume increment and cut volume.
An increase of an average stand density in UM-NPF and M-PF between two inventories was indicated with increased SDI of 4.42 and 5.51 percent, respectively. As expected, both types of managed state forests showed a decrease in SDI, which was almost three times larger in M-S2F.
In all forest management types, the past dynamics of stand structure represented by LikeJ index change showed negative trends. The smallest change was in M-PF, where DBH structure was the closest to inverse J-shaped distribution, while the largest change was observed in M-S2F and UM-NPF with the lowest LikeJ index.
The best stand regeneration, indicated by recruitment, was in M-PF, while M-S1F showed poorest stand regeneration, especially for fir with on average only 0.2 firs recruited per hectare yearly. Similar stand regeneration was observed in UM-NPF and M-S2F with 3.3 and 3.9 recruited trees per hectare, respectively.
Moreover, in all forest management types, annual changes in the number of small diameter trees were negative with approximately 6.5 (UM-NPF and M-S2F) and 10.5 (M-S1F and M-PF) trees decreased annually.
Mortality represented by percent of died trees per ha annually showed that the highest rate for overall, fir and beech trees was in M-S2F and significantly different than in other three forest management types, while in M-PF the mortality rate (except beech) was lowest. Unmanaged forests and M-S1F showed a similar rate for fir and overall mortality of approximately 1.9 percent.
Further Study On Stand Required
According to our study, we provided the first comprehensive landscape-level results of stand structure and recent stand dynamics within fir-beech stands in the Croatian Dinaric region that support the research hypotheses stated above.
We concluded that the current stand structure characteristics differ between forests withdrawn from management and most representative state managed forests. It would be easier to distinguish these forests if there were longer periods of forest use abandonment in the recently unmanaged forests and if more intensive harvests were applied in the managed forests.
As expected, some quantitative and qualitative differences caused by different site conditions, management objectives and approaches and natural disturbances were revealed in the studied types of managed forests.
The recent stand structure dynamics obtained with this fir-beech forests monitoring system can be understood as a result of long-term forest dynamics over past periods and as a ‘potential state’ for future forest dynamics that can be driven by spontaneous (natural) processes and/or different management approaches.
The gradual accumulation of standing volume, an increase of large diameter live trees and large snags as main characteristics of long-term past dynamics have supported unmanaged forests recently, unlike managed forests where these processes were slowed down or stopped, indicating a recent difference in structures of stand dynamics between unmanaged and managed forests.
The achievement of high accumulated basal area and stand density, a large share of beech, a large number of large live trees and snags and large mean DBH that can be recognised as old-growth attributes in forests withdrawn from forest management only during the last three decades indicates the large potential of fir-beech forests to sustain multifunctional forest management.
The present results alongside continuous fir-beech forest monitoring may be used to support silviculture for the conversion of higher-stocked stands into stands with a balanced structure, designation and maintenance to achieve an old-growth structure in the Dinaric region.
Moreover, further research on long-term projections of stand structures and stand dynamics based on various management objectives simulations, selection management systems and environmental factors (climate change, emissions and natural disturbances) should also be considered in order to support planning and management in fir-beech forests under demanding and changeable ecological and socioeconomic circumstances.