Structural Features of Old Growth Forest from South Eastern Carpathians, Romania

Romania’s primary forests are of globally significant value due to their natural characteristics, as similar forests in some other parts of the world have been lost forever [1, 2]. Quasi-virgin forests are those that have been managed in the past but that over time have been left to develop naturally, gaining specific features such as mixed tree ages, the presence of development phases, and strong relationships between dendrometric elements (i.e., diameter at breast height DBH, height, and volume) [3, 4]. The study of these forests is of high value, as they are considered an important source of scientific information that can be used in the management of unevenaged forests [5]. Quasi-virgin forests have numerous ecological, scientific, economic, social, and cultural characteristics, and they provide shelter for numerous species of flora and fauna [6] that have disappeared from managed forests. Despite their importance, many quasi-virgin forests are partially or completely unprotected, and their surface area is decreasing. Given their importance, there is an urgent concern to protect all of these types of forests, in Romania as well as on a global scale [2, 7]. Recently, interest in sustainable forest management concerning biodiversity and the protection of nature has increased enormously [8]. Numerous scientists have highlighted the importance of these types of ecosystems in different locations around the world [2]. In Europe, interest in natural forest stands began to appear with the publication of G. Gayer's Silviculture Treaty AbsTRACT


INTRODUCTION
Romania's primary forests are of globally significant value due to their natural characteristics, as similar forests in some other parts of the world have been lost forever [1,2]. Quasi-virgin forests are those that have been managed in the past but that over time have been left to develop naturally, gaining specific features such as mixed tree ages, the presence of development phases, and strong relationships between dendrometric elements (i.e., diameter at breast height -DBH, height, and volume) [3,4]. The study of these forests is of high value, as they are considered an important source of scientific information that can be used in the management of unevenaged forests [5]. Quasi-virgin forests have numerous ecological, scientific, economic, social, and cultural characteristics, and they provide shelter for numerous species of flora and fauna [6] that have disappeared from managed forests.
Despite their importance, many quasi-virgin forests are partially or completely unprotected, and their surface area is decreasing. Given their importance, there is an urgent concern to protect all of these types of forests, in Romania as well as on a global scale [2,7]. Recently, interest in sustainable forest management concerning biodiversity and the protection of nature has increased enormously [8].
Numerous scientists have highlighted the importance of these types of ecosystems in different locations around the world [2]. In Europe, interest in natural forest stands began to appear with the publication of G. Gayer's Silviculture Treaty (1878), in which he proposed respecting natural laws in forest development. After World War II, Leibundgut carried out multiple studies in natural forest stands [9][10][11], promoting the importance of these special ecosystems at the International Union of Forest Research Organizations (IUFRO). In 1971, IUFRO established a research group focused on studying natural forest stands and, taking into account the remarkable results and work done by Leibundgut, appointed him chairman of this scientific group [12]. In the period from 1995-1999, research was carried out within the secular forest stands situated in the boreal area of the Scandinavian Peninsula and in the northern European region of Russia, which exhibited the dangers of forest loss due to excessive logging [13]. Other studies of this kind were carried out in the Perućica Forest Reserve in the territory of Bosnia and Herzegovina; that work concluded that these natural forests are clearly superior in terms of biodiversity and structure compared to managed forests [14]. In 2000, a research area of 10 hectares was installed in the Uholka-Shyrokyi Luh Reserve (total area of the reserve is 15,974 ha, of which approx. 9,000 ha are considered virgin beech forests) located in the Transcarpathian region of Ukraine [15]. The naturalness, uniqueness, and the high level of biodiversity of these forests have attracted the attention of politicians and led to changes in the legislative framework in order to better protect them [12]. In Croatia, laws adopted for the protection and maintenance of natural forests date back to the 18th century [16]. In Romania, quasi-virgin forests have been protected by modifying the legal framework [17]. The aim of this paper is to emphasize the structural characteristics and very high structural diversity of quasi-virgin forests located in the Curvature Carpathians region (Romania) and to highlight the knowledge of their special structural features. The results of research on primary forests are extremely valuable for developing sustainable forest management practices.

MATeRIAls AND MeTHODs
Four permanent research plots (Plots A, B, C and D) were established in the Penteleu Mountains in the Curvature Carpathians Region, Romania ( Figure 1).
The forest stands where the permanent research plots were located have not been influenced by human activity for a very long time, according to the criteria for identifying quasi-virgin forests [17,18]. The permanent research plots were established during the period 2015-2018 and have dimensions of 100×100 m (1 ha). The perimeter of each plot was delineated using a Global Positioning System (GPS). All trees in each permanent research plot with a DBH>8 cm were inventoried by measuring the DBH and height (H) and establishing the tree's social position in the canopy. Tree height was measured using an ultrasonic hypsometer (Vertex IV) and DBH using a measuring tape. Based on the field inventory of the forest stands, their structural characteristics were processed and analysed. First the DBH distribution was analysed, then the structural biodiversity of the studied forest stands. It has been established that in the absence of human intervention, natural dynamics lead to very diverse forest stand structures [19,20]. One of the important aspects of diversity in forests is tree size variability. Gini [21] and Camino [22] are two relevant indices of forest structure based on dispersion estimates of tree size [20]. The structural homogeneity of the studied forest plots was tested using the Camino and Gini indexes, and a graphic representation was made using the Lorenz curve [23]. The Gini coefficient has proven to perform better as an indicator of forest structure than Shannon's diversity index, Simpson or other indices [24]. Therefore, the Gini coefficient is recognised as the best estimator of stand structure based on DBH [25,26].
The volume of each tree was determined using the following formula: where b 0 , b 1 , b 2 , b 3 and b 4 are nationally (Romanian) specific coefficients for each species. The experimental DBH distribution was fitted using different theoretical distribution functions (beta, gamma, gamma 3P, gamma 3P mixt, loglogistic 3p, lognormal 3P and Weibull 3p). To estimate the goodness of fit of the theoretical distributions to the measured DBH values, the χ2 criterion, Kolmogorov Smirnov (KS) [28] and Anderson Darling (AD) [29] statistical tests were used. Analyses were performed using Microsoft Excel Software, Mathwave -EasyFit Distributions, IBM SPSS Statistics, and packages fitdistrplus [30] and mixdist [31] of the software R.

ResUlTs AND DIsCUssION
All the permanent research plots exhibited uneven-aged structure, a large number of trees and a high volume per hectare ( Table 1).

Fitting of Experimental DBH Distribution Related to Number of Trees
To identify the best theoretical function to fit the distribution of the collected field data, beta, gamma, gamma   Figure 2). In the case of research plot C, a bimodal experimental distribution was observed. For this particular situation where the forest stand could not be fitted with the studied theoretical functions, a combination of two gamma functions was used. The bimodal distribution is more accentuated in research plot C than in the other plots and is related to the plot's history [32,33] and events like fire [34], wind disturbance or other biotic and abiotic factors. The KS test and χ 2 criterion for goodness of fit for plot A showed no differences between the experimental and theoretical distributions for the lognormal 3P function. The AD test showed no differences between the experimental and theoretical distributions for the Weibull 3P, gamma 3P and lognormal 3P functions. In the case of plot B, all of the goodness of fit tests indicated that the experimental DBH distribution followed the theoretical distributions except for the gamma distribution, where the statistical tests showed significant differences between the experimental and theoretical distributions. In the case of plot D, the relationship between DBH distribution and number of trees was analysed using the lognormal 3P, gamma 3P and beta functions. None of these theoretical functions adjusted the experimental DBH distribution (Table 2). Using the χ 2 criterion, in the case of plot C, the theoretical frequencies resulting from the mixed gamma 3P function were significantly different from the experimental distribution (p>0.05) ( Table 3). The experimental DBH distribution has a descending form, with the highest numbers of trees in small DBH categories and a shape similar to a reverse "J" [35], which is specific to the structure of quasi-virgin forests. All research plots exhibited high variation in DBH, over 90 cm, an aspect specific to uneven-aged stands [36,37].

structural biodiversity Analysis of studied Forest stands
To test the biodiversity of the studied stands, the Gini (G) and Camino (H) indexes were calculated, and for graphical analysis, a Lorenz curve was generated (Figure 3). The Lorenz  curve and the Gini indexes indicate that the studied stands have high structural biodiversity, which is specific to this type of forest ecosystem. The Gini index of the plots ranged from 0.69-0.71 which is very close to the index's maximum value of 1 [24].       forest stands in the present research are close to the values obtained for other uneven-aged forests, which emphasizes that the studied stands are characterized as uneven-aged stands [38][39][40].

CONClUsIONs
Based on the results presented in this research, it can be concluded that due to their structural complexity, quasivirgin forests represent a very good scientific base for studying the natural structure and dynamics of a forest, and can be considered as "real laboratories in situ". The presence of large-DBH trees and fast development phase alternations are signs of strong dynamics as well as a great capacity to regenerate after natural competition processes.
The protection of quasi-virgin forests should be a precondition for successful scientific research in natural science. To develop improved forest management practices, it is very important to understand the structural principles and development of natural forests. The legislative system in Romania for protection of natural areas where we still encounter natural ecosystems that are not influenced by humans must generally be improved, and very valuable quasi-virgin forest areas must be included among other natural conservation areas.