European silver fir – an alternative for the dying Norway spruce in Białowieża Forest?

The material of this study, conducted in 2019 and 2020, was 30 silver fir trees (this is all that remains of an experience that began one century ago). Most of them were represented by old (mother) trees (18 pcs.) and other trees (12 pcs.) were the adult progeny of old firs (Fig. 1). These 12 trees were no longer located in the undergrowth and were also clearly not maternal trees. The present work focused on the study of the mother trees. In 1995, the 1.8-hectare site was fenced, where grew the old silver fir trees and their natural regeneration of different ages (Fig. 1), as well as the living and dead other tree species (Fig. 2).

Figure 1

Figure 2

The trees were located in Białowieża Forest District, subcompartment 498 Сi. The GPS coordinates of the centre of the site was E 23°48’53.2’’, N 52°40’35.3’’ (Fig. 3).

Figure 3

The type of forest was LMśw – fresh, mixed broad-leaved forest, and the soil was rusty brown, not degraded. According to EFTC (European Forest Type Classification 2006), based on the EUNIS (European Union Nature Information Scheme) forest classification, the type of forest can be classified as mesotrophic deciduous forests. The World Reference Base for Soil Resources (WRB) includes this soil in the Arenosols soil group (Bednarek and Prusinkiewicz 1997; IUSS Working Group 2014).

The following dates of planting (sowing) of the firs studied appear in different publications: 1900–1910 (Korczyk 2008), 1928–1930 (Korczyk 1999) and 1928–1932 (Korczyk 1995). Thus, the age of the trees, according to the above data, can vary from 87 to 119 years. According to a description of the stand, the age of the trees is said to be 96 years (http://www.bdl.lasy.gov.pl/portal). The unequal value of this important forestry indicator in different sources prompted us to conduct our own research.

No information was found on the initial characteristics of the forest cultural area – planting (sowing of seeds) under the canopy of an old forest stand, at the opening of the old stand without shelterwood, in a pioneer crop, under the canopy of naturally renewed pioneer breeds, or in cutting. Unfortunately, the documentation was most likely lost without a trace during World War II. Lack of this important information also prompted us to try to reconstruct the conditions under which these trees appeared.

It should be noted that at the time of the study, no forest stand existed in the forestry sense of the word on the fenced surface. Initially, it was a mixed (group placement) forest stand of different ages. In addition to silver fir, in its composition were also Norway spruce (main breed), Scotch pine, common oak and common birch (admixtures). However, by the time of the study, already at the age of maturity of the forest, it had almost completely disintegrated. The main reason was the massive drying of Norway spruce that is characteristic of the Forest as a whole. There was drying out of pine and oak (Fig. 2). Many dead trees lay on the ground while most of them were still standing.

The silver fir and hornbeam (Carpinus betulus L.) were represented most in the undergrowth, also found was Norway spruce (Picea abies (L.) Karst), Scots pine (Pinus sylvestris L.) and common oak (Quercus robur L.). The under-forest was formed by European mountain ash (Sorbus aucuparia L.), European hazel (Corylus avellana L.), red elder (Sambucus racemosa L.), raspberry (Rubus idaeus L.) and dewberry (R. caesius L.). The living ground cover was composed of small nettle (Urtica dioica L.), bracken fern (Pteridium aquilinum (L.) KUHN), male fern (Dryopteris filix-mas L.) Schott), common speedwell (Veronica officinalis L.), wild strawberry (Fragaria vesca L.), sweet woodruff (Asperula odorata L.), yellow archangel (Galeobdolon luteum L.) and common hepatica (Hepatica nobilis Mill.). Other species were also present, but only sporadically.

The age of old trees (as of 2020) was determined using several methods, including non-standard ones:

The study of the state of silver fir was carried out on the basis of a comparative analysis, by way of matching it with the wood species with which it grew. We studied the dynamics of the radial growth of silver fir and spruce over the past quarter-century. For this, the annual growth images scanned with a resolution of 600 dpi were analysed in the QGIS Software (Kotsan and Sevko 2019).

For assessing tree vitality, relative crown length, growth trends, crown deformation and trunk quality, the modified IUFRO scale was used (Jaworski and Paluch 2007).

In this study, we used standard techniques usually practised in forestry. To determine the age, we calculated the number of tree rings on the core taken with the help of the incremental drill of Pressler from tree No. 18 (Fig. 1). But non-standard techniques were also applied.

The dbh (1.3 m) was measured with a calliper, as the height – using a SUUNTO PM-5/1520 height metre. The diameter data were distributed not by 4 cm thickness steps, as it is usually accepted, but by 5 cm ones, similar to the study by Korczyk (1999). This was done with the aim of a more objective comparative analysis of changes in this indicator, as well as the height and number of trees for more than 20 years.

The volume of trees of spruce, pine, oak, hornbeam and birch was determined approximately using the Denzin formula:

where:

V – tree volume (m³),

d – dbh (cm),

h – tree height (m) (Poschenrieder et al. 2016).

The dendrometric formula was used for a more accurate determination of the volume of silver fir trees:

where:

V – the volume of the tree (m³),

g_1.3 – the basal area at the height of 1.3 m (m²),

h – the height of the tree (m),

f_1.3 – form factor, taken from tables of Szymkiewicz (2001).

As part of the study, an attempt to extract biometric parameters of trees based on terrestrial laser scanning (TLS) data was made. During the field campaign, a selected group of trees consisting of five firs (group No. 3 in Fig. 1) was scanned using a terrestrial laser scanner – Trimble TX5. The TLS data were collected by using the multi-scan approach (Liang et al. 2016): 1 central position (in the centre of the selected group of trees) and 6 positions outside the group of trees (Fig. 4). Four artificial targets (spheres) were used in order to merge scans performed from multiple positions.

The process of scans co-registration and point could exporting was conducted in the FARO Scene software (FARO Technologies, Inc., USA). In the next stage, the co-registered point cloud was divided into individual trees in the CloudCompare software (Girardeau-Montaut 2020). The individual trees extraction was done manually in order to separate the intersecting crowns as accurately as possible, and thus, the points belonging to particular trunks and crowns of trees were selected.

Figure 4

Based on the preprocessed TLS data, the following dendrometric parameters of the trees were determined: tree height (TLS_H), dbh (TLS_DBH), basal area (TLS_BA), crown base height (TLS_CBH), crown projection area (TLS_CA), crown share (TLS_CS), crown volume (TLS_CVOL) and stem volume (TLS_VOL). The tree height was determined as the height of the highest registered point located within the crown of the particular tree. The height of the crown base was measured manually by the operator in CloudCompare software as the height measured to the first live branch of the crown of the tree. The dbh measurement was performed using an automatic RANSAC cylinder fitting algorithm (de Conto et al. 2017). Into a 10-centimetre slice of point cloud representing the tree trunk at a height of 1.25–1.35 m, a cylinder was automatically fitted, so as to it was possible to read the radius. The basal area has been calculated as the area of the cylinder base.

To calculate the stem volume, a method of building a three-dimensional trunk model was applied. Using the RANSAC cylinder fitting algorithm, a model consisting of a series of cylinders was created. Based on the generated model, it was possible to determine the total volume of cylinders, which was treated as the stem volume. It should be noted, however, that TLS technology has some limitations, and usually, it is not possible to rebuild the taper curve of the entire trunk, for example, due to the occlusion effect caused by mutual covering of branches in the upper parts of the tree crowns and by other neighbouring crowns (Wang et al. 2019). In order to minimize the underestimation of the stem volume caused by the inability to reproduce the entire taper curve, the authors used the diameter of the last visible section of the trunk and the height of the highest registered crown point to create a cone – a simplified model of the unregistered part of the trunk was generated (Fig. 5).

Figure 5

To determine the volume of the crown, points previously classified as the tree crown were used using the Convex Hull algorithm (Fig. 5). This process was performed in MeshLab software (Cignoni et al. 2008). The created shape presents a generalized three-dimensional crown envelope. The crown area and coverage was determined based on the projection of the crown contours using GIS tools. The crown share was calculated as a percentage of the crown area that is shared with neighbouring crowns. In addition, based on TLS data, a visual assessment of the crown apex shape was made to evaluate the development phase of a given tree and its potential for growth. In this case, the assessment scale provided by Jaworski and Paluch (2007) was used.

Finally, the results of TLS measurements were compared with the ground measurements performed by traditional methods.

In order to assess the condition of silver fir, a comparative analysis was carried out of some forestry parameters of tree species growing together with the silver fir.

The Norway spruce population was the most significant, but the individuals in the lower classes of the diameter to 15 cm (78%) (Fig. 6) prevailed. Most of the older spruce trees had dried out. There were only approximately 3% of trees with a diameter of 25.1–40 cm left.

Figure 6

The number of hornbeams was less than Norway spruce. The younger generation dominated in the classes of the diameter from 5.1 to 10 cm (88%). There were no individuals present with a diameter of more than 15 cm. At present, hornbeam is the most expansive species in the Forest (Вrzeziecki 2017). This is clearly visible in our case.

Single individuals of common birch were noted in steps from 40.1 to 50 cm. The admixture of this species was the smallest.

The predominant part of oak trees (50 pcs. or 75%) had a diameter not exceeding 35 cm. Only in single in dividuals (3 pcs. or 5%) was the diameter in the range from 85.1 to 100 cm.

Scotch pine was most significantly represented in the range of 35.1–55.0 cm (42 pcs. оr 79%). There were no individuals with a diameter of less than 20 cm. One tree survived with a diameter in the range of 90.1–95.0 cm.

An analysis of the diameter distribution (Fig. 6) showed that spruce, hornbeam and birch growing together with silver fir did not form a normal stand. The curves of the diameter distribution of pine and oak to a certain extent corresponded to this. But since the trees of these species dispersed throughout the site and there were few, they didn’t constitute a stand density even close to the minimum necessary degree. Therefore, as normal stands, pine and oak were also not represented.

A wide range of tree species on the site, as well as characteristics of under-forest and living ground cover, indicated rich soil conditions corresponding to the ecological and biological needs of silver fir.

Spruce dominated on the plot as in the case of living trees (55% of the total number or 261 pcs.) and in the case of dead trees (91% or 240 pcs.) (Fig. 2 and 7). Judging by the vitality of the still-living trees of this species, among which many were weakened and greatly weakened, they are likely to die in the future. As for the oak, the second largest species (14%), 12 individuals or 5% of the total number of all trees were killed. Pine, the third highest number species (11%), was characterized by a similar state to the previous species, with 3% of dead trees. The least represented was birch (4%), of which only one tree was dead.

Figure 7

As can be seen, the main forest-forming species for the Białowieża Forest represented on the plot were spruce, oak, pine and birch, although to varying degrees; these trees are vulnerable, however, to dying due to abiotic and biotic factors. No dead hornbeam trees, of which the number was 9% of the total, were found (Fig. 7). Judging by the abundance of the hornbeam undergrowth (Fig. 6), in the future, this species will increase significantly. The situation of the plot described above is a peculiar snapshot, which on the whole objectively reflects the current state of these species in Białowieża Forest.

Comparing the vitality of silver fir and other species, this does not raise concerns yet (Fig. 2 and 7): the number of healthy trees was 30% or 94%, only two trees were dead. Moreover, these two trees weren’t affected by a prolonged drought that weakened them, or by a subsequent attack of the bark beetle-typographer, as in the case of spruce. The reason, we believe, was as follows: a very tall silver fir tree (No. 6/4, 75 years old, and with a height of 31 m), and therefore unsustainable, was felled by the wind and, when it collapsed, broke the nearby tree (No. 6/3), which subsequently died.

As shown above (Fig. 2, 6 and 7), spruce, despite massive drying, still quantitatively continued to be the dominant species on the site. As well as silver fir, it is a dark coniferous species. It has similar requirements for soil and lighting conditions, and also a close rate of increase. The most significant difference is in the type of root systems: silver fir has a tap root system, with 4–5 powerful lateral roots directed down vertically, while Norway spruce has a pronounced superficial root system (Jaworski 2011).

Crucial for the understanding of the current state of these two species and future predictions is a retrospective assessment of the features of their growth. It was obtained based on an analysis of the annual radial increments. To minimize potential damage, cores were taken from only one old silver fir tree of a medium size (No. 18, see Fig. 1) and one Norway spruce (they both had approximately the same diameter and age). And only this circumstance explains the uncertainty of the data necessary for a full-fledged statistical analysis. Nevertheless, the available material made it possible to conduct a preliminary comparative analysis of tree-ring chronologies shown in the form of absolute radial increments (Fig. 8).

Figure 8

Of particular interest is the period preceding the beginning of the last mass drying of Norway spruce in Białowieża Forest in 2012 (Brzeziecki et al. 2018a), and further, up to the present. In 1994 was recorded a fairly significant synchronous decrease in the analysed parameter in both species. However, over the past quarter-century with its steadily warming trend, the dynamics of changes in the radial increment in these two species, as can be clearly seen from Fig. 8), were diametrically different. So, for example, the amplitude of the radial increment of silver fir varied from 0.14 to 0.32 cm. While in Norway spruce, it was an order of magnitude less – only from 0.04 to 0.16 cm. These facts confirmed our assumption stated above that the remaining Norway spruce trees, most of which are currently weakened and severely weakened, are likely to die in the future.

The state of mother trees with a division into separate positions, according to quality indicators, is presented in Table 1.

Analysing the obtained data, we can conclude that the qualitative characteristics of the mother trees correspond to their age, and in most cases, the assessment of indicators is closer to the positive maximum.

The relatively good condition of the trees led to the stability of their number, supported by documentation. As follows from a comparison of the data of Korczyk (1999) and our own comparison, in more than 20 years, only two trees have died due to the mechanical damage incurred as shown above. Unfortunately, we did not have data on the initial number of trees on the site and on the subsequent change in their number up to the time of the study.

The expressed positive temporal dynamics of the height and diameter of silver fir trees should also be clearly noted. If, more than 20 years ago, the height of the tallest tree was 25 m, then in 2019 it was already 33 m. The diameter of the thickest tree was in the class of diameter of 35.1–40.0 cm, and it had moved to the class of 60.1–65.0 cm at the time of the study (Fig. 9).

Figure 9

In carrying out the total enumeration of trees, their main parameters and average statistical indices were determined (Tab. 2 and 3). As can be seen, by the size of the diameter, old trees were distributed over a wide enough class of diameter – from 25.1 to 65.0 cm. More than half of the trees (55%) were concentrated in the class from 30.1 to 45.0 cm. The lower class 25.1–30.0 cm was represented by two trees, and in the higher class from 50.1 to 65.0, four individuals were recorded (No. 1, 6, 11 and 19). They had the most significant height (from 29 to 33 m) and the volume of the trunk (from 3.86 to 5.46 m³).

The variability of the diameter and height was characterized by variation coefficients of 25.2% and 9.9%. The coefficient of variation of the volume of the trunks was 56.8%.

Individual No. 1 (Fig. 1), in our opinion, can be considered a candidate for advantage trees. Possessing the highest taxation indices (Tab. 2, Fig. 5B), at an age of about 100 years, it continued to abundantly bear fruit.

The number of adult progeny of old firs was extremely small – only 12 pcs. (6% of the total species stock in m³) per 18 mother trees, which strongly discords with abundantly represented young natural regeneration this species up to 0.5 m high on the plot now. We believe that this situation indicates the digression of this category of offspring in the present and the even possible elimination of A. alba in the future. At the same time, we note that none of the trees of the adult progeny of old firs had signs of weakening under the influence of biotic or abiotic factors. Exactly half of them were represented in the 10–15 cm class (Tab. 2). The diameter floor was in the class of a diameter of 5–10 cm, and the upper one was in 25–30 cm. The coefficient of variation of diameters (36.5%) was much larger than that of the mother trees, but this, in principle, corresponds to the characteristics of young trees. The coefficient of variation in height was even more significant (48.6%). The amplitude of its variation was from 4 to 22 m (Tab. 3). This fact, in our opinion, indicates the diversity of the age of the preserved adult seed offspring.

An analysis of the biometric tree parameters estimated using the TLS data showed that the difference in the results of determining the most important tree traits (H, DBH, V) between the traditional method (field measurements) and TLS data is not significant (Tab. 2 and 4). The TLS-estimated tree height and stem volume are underestimated compared to field measurements. On the other hand, the DBH values obtained from the TLS data are higher than the reference values collected in the field.

The application of the TLS technology allowed to objective and precise estimation of tree crown metrics. The field measurements of these parameters were not performed due to the high time consumption and difficulties in the acquisition process, and in this connection, at the initial stage of the study, we limited ourselves only to a visual assessment (Tab. 1). The length of crowns for all five trees ranges from 58% to 81.5% of their relative height, which corresponds to grades 1 and 2 (Jaworski and Paluch 2007). Moreover, the maximum crown length ratio (81.5%) was observed for the most powerful tree (No. 1), which is even slightly more than ¾ of its height. The average distance between the trees in the group No. 3 is ca. 5 m, and the total horizontal projection of the crowns is about 190 m². Moreover, approximately 35% of this area is overlapped by projections of crowns of adjacent trees.

It has to be highlighted that significant percentage of overlap in the horizontal plane is created by well-developed crowns also in the vertical direction (Tab. 4). This, in particular, is indirectly evidenced by the fact of the complexity of scanning the upper part of the trunks mentioned above (Fig. 5A). This is due to the frequent arrangement of branches and their dense needles. The shape of the crowns of the analysed trees resembles quite high (from 13 to 27 m) irregular, narrow pyramids pointed at the top or even cylinders (Fig. 4 and 5). At different vertical levels, they touch, overlap and thus fill with their volume quite a significant total space above the ground. It is located starting from about 6 m above its surface and to a height of 34 m. Moreover, it was observed that the contribution of individual trees to the formation of the total crown volume (sum of TLS_CVOL) is significantly different. The share of tree No. 1 accounts for almost half (42.8%) or 1546 m³, while the contribution of the other trees is much lower: from 7.6% to 23%. It is visible that the leader tree, which has most successfully realized itself in a life race lasting 100 years, due to its crown, currently has the most significant influence in this biogroup. This case confirms one of the main paradigms of forest growing, which says that the main forester activities should be aimed at creating comfortable conditions for leader trees. An analysis of the results obtained using TLS allowed us to reveal a fairly clear pattern that we have not previously encountered in the literature: a change in the stem volume positively correlates with a change in the volume of crowns. Or in our case: when the stem volume decreases by 2–6.7 times compared to the leader tree, the volume of their crowns decreases by 1.9–5.6 times (Tab. 4). For monitoring purposes, it is planned to repeat the scanning of the area in the next 5 years, so it will be possible to estimate the growth of trees and assess changes in this stand.

The described situation would be more revealing if the analysed isolated group of trees was a fragment of a stand that has a normal spatial structure over a fairly large area. That is, if it was surrounded by other silver fir trees or trees of other species. Nevertheless, even with this objectively far from fully representative example, it can be seen that in the conditions of the Białowieża Forest, A. alba can form a typical multi-level vertical structure of the mother canopy typical of its stands, which is certainly necessary for the normal development of shade-tolerant undergrowth. At present, it is completely open to lateral sunlight.