Monitoring Intra-Annual Wood Formation of Pinus nigra J.F. Arnold (Black Pine) to Understand the Fire Seasonality in Western Anatolia

: Recent climate and societal changes have increased wildﬁre activity and prolonged the ﬁre season in many regions of the world. The precision of ﬁre seasonality analysis from tree-ring records can be improved by complementing the subjectively determined intra-ring position of ﬁre scars with more precise studies of wood formation. With this aim, we monitored the wood formation dynamics of Pinus nigra J.F. Arnold (black pine) trees along a climatic gradient in western Anatolia to better understand the wood formation for the interpretation of ﬁre seasonality. Wood microcores were collected from April to November 2021 from trees at four sites across (from north; the Black Sea climate in Bolu to the south; and the Mediterranean climate in Isparta) the areas where previous ﬁre history reconstructions were conducted. These previous studies showed that most ﬁres occurred during the latewood formation period. We found that matured latewood tracheids were observed between September (August) and November, thus suggesting that these ﬁres occurred during late summer and fall. Our results show the importance of temperature and water availability for the timing of earlywood and latewood formations. These ﬁndings can be used to better inform planning activities for ﬁre management and as a proxy to reconstruct past ﬁre seasonality.


Introduction
Fire seasonality is a crucial parameter of fire regimes and is still uncertain for most pyrogenic forest ecosystems [1,2].Understanding the fire regime at the regional level requires a deep knowledge of fire seasonality, indicating the part of the year when fires are more likely to occur.Dendrochronological techniques are widely used in terms of providing robust, annual, high-resolution, and century-long records on historical fire occurrence and seasonality at different spatiotemporal scales.However, the interpretation of the intra-annual position of fire scars is a coarse and uncertain way to determine the seasonality of fires.In dendrochronology-based fire history reconstruction studies, the seasonality of fire occurrence is commonly derived from the intra-annual position of fire scars within the annual ring.This often-used classical fire seasonality classification method is highly subjective [3,4].Wood formation is a complex growing process involving cell production, enlargement, and differentiation over several months during a vegetation period that reflects the seasonal cambium activity [5].Thus, additional studies based on Forests 2024, 15, 494 2 of 13 wood formation dynamics (i.e., xylogenesis) are needed to determine the fire seasonality of a region through the year in order to precisely interpret the occurrence of fire scars within a growing season.Since the timing of xylem formation varies according to species, location, and climate, xylogenesis studies can be used to estimate the timing of the formation of the different ring sections, such as earlywood, latewood, and the transitional part between them, to determine the seasonality of fire [2,6].
Pinus L. is one of the most ecologically and economically important genera of trees in the Mediterranean basin, where pine forests frequently experience surface fires.Despite many xylogenesis studies in pines already existing, most have focused on Pinus sylvestris L. [5,7,8] and only two studies have examined the interannual dynamics of the wood formation of the circum-Mediterranean pine species Pinus nigra J.F. Arnold (black pine) [9,10].The number of studies monitoring intra-annual growth and cambial activity is considerably limited in Türkiye, and this research only involves dendrometer measurements [11][12][13], xylogenesis research [9], and cambial phenology [14].However, site-specific studies in the natural distribution of black pine forests, analysing the mechanisms of the cellular structure of xylem growth dynamics of trees in Türkiye, have hitherto been lacking.
Previous studies of tree-ring-based fire history reconstruction in western Anatolia [15][16][17] highlighted a shift in fire frequency with both the influence of climate (the effect of long and severe drought on fuel properties and fuel moisture) and human-induced suppression activities on combustible material.These studies also showed that most of the fires occurred in the latewood formation period and suggested that the seasonality of fires remained unchanged over the last 600 years across this climate gradient.These findings on and future predictions of fire frequency increase the significance of the studies on the seasonality of fires to understanding the multi-dimensional and coupled nature dynamics of fires.Wood formation depends on climate conditions at the site and at annual level.Since this study is part of a broader project that complements previous findings related to dendrochronologybased historical fires in the region, here, we investigated the timing of wood formation along the transect of previously studied historical fire reconstruction sites.For this, the xylem formation of black pine trees was monitored during one vegetation period at four sites along a climate gradient in the western Anatolian mountains.This is the first study aiming to improve the knowledge of fire seasonality by studying the wood formation of black pine trees in Türkiye.Additionally, although environmental drivers or complex physiological processes were not comprehensively analysed in this study, the seasonal wood formation dynamics were studied preliminarily to provide direction for other xylogenesis studies for Türkiye.The specific objectives of this research were: (i) to monitor the phenology of the wood formation of black pine and estimate the earlywood and latewood formation periods, and to also interpret fire seasonality from tree-ring-based reconstructions; (ii) to assess differences in the growth patterns in black pine trees growing along a climatic gradient, including the three main climate types in western Anatolia; and (iii) to understand the effect of short elevation on cambium activity in one of the study sites.

Study Sites
The study sites were located in western Anatolia, near Bolu (KIR), Kütahya (KUTH, and KUTL), and Isparta (HAV) (Figure 1; Table 1).These sites were chosen due to their proximity to ten sites where fire tree-ring-based reconstructions were developed to reveal the spatiotemporal patterns of the sites [15,16] in terms of both ecological and environmental conditions (Figure 1).The sites along this climate gradient represent the three main climate types in western Anatolia [18].The area spans from the Mediterranean in the south (HAV) to the Black Sea climate in the north (KIR).The KIR site near Bolu is located in the transition zone from the continental to the Black Sea climate, which receives precipitation throughout the year and is characterised by cool summers, warm winters on the coast, and cold, snowy weather in the mountains.The HAV site near Isparta is located in the Mediterranean climate.This has hot and dry summers as well as cool and wet winters throughout the year.The KUTH and KUTL sites are near Kütahya and fall under the Marmara transition climate conditions, which are cooler than the Mediterranean weather, warmer than the continental climate in winter, and drier than the Black Sea climate types.In order to understand the potential effects of elevation, even though it represents a short elevation gradient, on cambium activity in black pines, two different sites were sampled in Kütahya at different elevations dubbed KUTL (1160 m a.s.l.) and KUTH (1430 m a.s.l.).The soil type of all sites is classified as non-calcareous brown forest soil [19].The sites along this climate gradient represent the three main climate types in western Anatolia [18].The area spans from the Mediterranean in the south (HAV) to the Black Sea climate in the north (KIR).The KIR site near Bolu is located in the transition zone from the continental to the Black Sea climate, which receives precipitation throughout the year and is characterised by cool summers, warm winters on the coast, and cold, snowy weather in the mountains.The HAV site near Isparta is located in the Mediterranean climate.This has hot and dry summers as well as cool and wet winters throughout the year.The KUTH and KUTL sites are near Kütahya and fall under the Marmara transition climate conditions, which are cooler than the Mediterranean weather, warmer than the continental climate in winter, and drier than the Black Sea climate types.In order to understand the potential effects of elevation, even though it represents a short elevation gradient, on cambium activity in black pines, two different sites were sampled in Kütahya at different elevations dubbed KUTL (1160 m a.s.l.) and KUTH (1430 m a.s.l.).The soil type of all sites is classified as non-calcareous brown forest soil [19].[20].Climograph of study areas from (A) Kütahya, (B) Bolu, and (C) Isparta meteorological station .Bars indicate precipitation (mm), and red, green, and blue lines represent the maximum, mean, and minimum temperature ( • C), respectively.

Microcore Sampling and Tree-Ring Measurements
To monitor the cambial growth dynamics, microcore samples were collected using the Trephor tool (2 mm in diameter) [21].Six healthy dominant or co-dominant trees with straight stems were selected for sampling from each site for a total of 24 trees.Trees with asymmetrical stems, partially dead crowns, reaction wood, or suppression or damage by insects were avoided.The pervasive impact of COVID-19 and ongoing restrictions across the country further complicated our ability to conduct synchronous sampling based on the standard xylogenesis sampling procedure every two weeks.For this reason, trees were sampled at monthly intervals across the four sites from the end of April to the end of November 2021 (the last weeks of each month).We started collected samples in April due to the persistent snow coverage in northern and high-elevation sites, which imposed limitations on our transportation.Microcore samples were collected following a spiral pattern from a stem height of 1.3-1.6 m.Adjacent sampling points were separated by 5-10 cm to avoid possible wound effects [21].In the last month of microcore sampling, we measured the diameter at breast height (DBH ~1.3 m from the ground) and collected full increment cores from each tree.The full increment cores from both sides of each tree were also collected using an increment borer at ~1.30 m above ground.These cores were glued onto channelled wood boards and sanded until growth rings were clearly visible.Tree-ring widths were measured with CDendro software (version 9.7) [22] with the aim of defining the age of the sampled trees.The COFECHA software (version 6.06) was used to verify the cross-dating of tree-ring series [23].

Laboratory Methods and Microscope Observations
Microcores were placed in microcentrifuge tubes filled with 70% ethanol solution and stored at 4 • C to prevent tissue deterioration until being processed in the laboratory.Microcore samples were dehydrated in successive ethanol concentrations (70%, 90%, 96%, and 100%) and later embedded in paraffin blocks.Sections of 9 µm thickness were cut using a manual rotary microtome (Leica Biosystems, Frankfurt, Germany) following standard procedures [21].Thin sections were stained with a mixture of safranin and astra blue to properly visualise the lignified (stained red) and non-lignified (stained blue) cells and then fixed in permanent slides mounted with Eukitt ® .The prepared slides were scanned at 100-400× magnification with a digital camera (Leica DMC5400, Leica Microsystems, Wetzlar, Germany) coupled with a microscope (Leica DM6 B, Leica Microsystems, Wetzlar, Germany) using transmitted and polarised light and then analysed to determine the stages of cell development.
In each thin section, four cell development phases were distinguished [24]: cambial zone (CZ), enlarging (EN), wall-thickening (WT), and mature tracheids cells (MC) (Figure 2).In each sample, we counted the number of cells in each phase along at least three radial rows.Cells in the cambial zone were distinguished due to their blue-stained thin cell walls and small radial diameters.Enlarging cells, also stained in blue, were distinguished by their radial dimension being at least twice as wide as that of cambium cells.Wall-thickening cells still have protoplasts, glow under the polarised light, and have cell walls only partially stained red.Finally, matured cells have lignified walls, which are stained in red and glow under polarised light [24].To determine the transition from earlywood to latewood, we utilised the widely used Mork's index (MI) [25].For this, we measured the radial width of the lumen and double cell wall thickness for MI calculation in addition to analysing cambial phenology.Three radial rows were measured by using the open access ImageJ software (version 1.53) [26] to calculate the MI. cells still have protoplasts, glow under the polarised light, and have cell walls only partially stained red.Finally, matured cells have lignified walls, which are stained in red and glow under polarised light [24].To determine the transition from earlywood to latewood, we utilised the widely used Mork's index (MI) [25].For this, we measured the radial width of the lumen and double cell wall thickness for MI calculation in addition to analysing cambial phenology.Three radial rows were measured by using the open access ImageJ software (version 1.53) [26] to calculate the MI.

Data Analysis for Dynamics of Xylem Formation
The value from the cell counts of the wood formation phases was used to calculate the number of standardised cambium, enlargement, wall-thickening, and mature tracheid cells for each site.Phases were assessed for each tree and were computed in days of the year (DOY).Because different parts along the stem circumference grow at different rates, we standardised the observed number of cells measured in each phase by using the total number of tracheid cells of the previous year in each sample [27].Xylem growth was characterised via cumulative counts of enlargement, wall-thickening, and mature tracheid cells.We modelled the total of standardised cells (enlarging, thickening, and matured) by fitting a Gompertz function [27].The equation for the Gompertz function used was: where A is the upper asymptote, β is the time axis placement parameter, and  is the parameter for the rate of change.We estimated biologically meaningful growth parameters for each site, including the time of the maximum growth rate (tx), the maximum rate (gx), the average cell production rate (gm), and the dates on which 5% and 95% of the total annual growth were completed (DOY) for each tree (t5 and t95, respectively) as follows: m = 0.9 log(log(0.05)/ log(0.95)) (2)

Data Analysis for Dynamics of Xylem Formation
The value from the cell counts of the wood formation phases was used to calculate the number of standardised cambium, enlargement, wall-thickening, and mature tracheid cells for each site.Phases were assessed for each tree and were computed in days of the year (DOY).Because different parts along the stem circumference grow at different rates, we standardised the observed number of cells measured in each phase by using the total number of tracheid cells of the previous year in each sample [27].Xylem growth was characterised via cumulative counts of enlargement, wall-thickening, and mature tracheid cells.We modelled the total of standardised cells (enlarging, thickening, and matured) by fitting a Gompertz function [27].The equation for the Gompertz function used was: where A is the upper asymptote, β is the time axis placement parameter, and κ is the parameter for the rate of change.We estimated biologically meaningful growth parameters for each site, including the time of the maximum growth rate (tx), the maximum rate (gx), the average cell production rate (gm), and the dates on which 5% and 95% of the total annual growth were completed (DOY) for each tree (t5 and t95, respectively) as follows: gm = 0.9κA log(log(0.05)/log(0.95)) (2) The non-linear Gompertz function was fitted using the 'nlme' package in the R environment [28,29].We classified cells into either earlywood or latewood cells as MI = 2 × CWT/LD, where CWT is the cell wall thickness and LD is the lumen diameter [25].We classified cells with MI ≤ 0.5 as earlywood, cells with 0.5 < MI < 1 as transitioning between earlywood to latewood, and cells with MI ≥ 1 as latewood [30].

Cambial Activity across Sites
The wood formation phases showed that the cambium was already active in the KUTL (the Marmara transition climate) and HAV (the Mediterranean climate) sites at the beginning of sampling at the end of April 2021, as indicated by the higher number of cells in the cambium than observed during the dormant period (Figure 3).The maximum number of cells in the cambium was observed in early May (ca.DOY130) in KUTL (lower elevation; the Marmara transition climate), at the end of May (ca.DOY150) in the KIR (the Black Sea to continental climate transition), and at the beginning of June (ca.DOY160) in the KUTH (upper elevation; the Marmara transition climate).The number of enlarging cells peaked earlier in KUTL (ca.DOY150) than in the rest of the sites (ca.DOY160, June).First, cells in the wall-thickening phase were observed in May (KUTL, KUTH, and HAV sites), and were seen later in June in the KIR site.The number of wall-thickening cells peaked between late July (DOY210) in the HAV site and late August (ca.DOY230) in the KUTH and KIR sites.The first mature cells were observed at all sites at the end of June (ca.DOY180).Table 2. Biologically important growth parameters for xylem growth dynamics derived from the Gompertz functions fitted at tree-level for each site (n = 6; ranges for each parameter in all cases).A, upper asymptote; ß, time axis placement parameter; , parameter for the rate of change; t5 and t95, the dates at which 5% and 95% of the total annual estimated growth is completed (DOY); tx, the time for maximum growth rate occurrence (DOY); gx, the maximum growth rate; gr, the mean rate of cell production (cells/day); R 2 , coefficient of determination for each tree.DOY as day of the year.We fitted the Gompertz function to the cumulative cell numbers at the tree level to show the cumulative radial growth rates (Figure 4).The fastest xylem growth rates were observed in the KUTH site, while the lowest were observed in the KUTL site (Figure 4).The growth parameters showed that the trees completed 5% of the growth earlier in KUTL and HAV sites (ca.DOY110-124, respectively) than in KIR and KUTH sites (ca.DOY124-125), while it did not differ markedly within sites (Figure 5A).In total, 95% of the growth was completed by DOY250 in the KUTH site, by DOY268 in KUTL and HAV sites, and in DOY282 in the KIR site (Figure 5).The HAV site was the first one to start and end growth, while the KIR site was the last one.For the two sites positioned along an elevational gradient in Kütahya, the lower site (KUTL) started and completed annual growth a few days earlier than the upper-elevation site (KUTH) (Table 2).the timing of earlywood/latewood formation.Because the samples were collected at monthly intervals, only the beginning of earlywood/latewood formation and the matured earlywood/latewood tracheid cells could be distinguished.The transition from earlywood to latewood occurred at the end of July in HAV, KUTH and KUTL sites, and at the end of August in the KIR site.In the HAV, KUTH and KUTL sites, the mature latewood tracheid cell formation was observed at the end of August (Figure 6A).In the KIR sites, mature latewood tracheid cell formations were observed at the end of September.In the HAV site, we observed that most of the earlywood cell formation was fully formed in May.At the end of July, earlywood cell formation was completed and the newly formed cells started displaying anatomical characteristics intermediate between earlywood and latewood (i.e., transition wood).In the KUTH site, earlywood cell formation was almost fully completed in June.All earlywood cells matured and the transition to latewood was observed in August.In the KIR site, all the earlywood cell formations were completed by the end of July.

Parameters
Using Mork's index criteria, we estimated that the formation of latewood tracheid cells started between the beginning of June and mid-July (between DOY155-193) in the HAV site, and between mid-June and the end of August (DOY165-242) in the KIR site

Discussion
In this study, we aimed to understand the timing of fire occurrence in our region by determining the timing of earlywood and latewood formation by analysing the intra-annual xylem growth.Although this study is based on one-year monitoring at monthly intervals and based on preliminary and limited data, our findings are relevant to defining the fire regime of our study region and to interpreting the corresponding fire seasonality within a wide climatic gradient in western Anatolia.The formation of matured latewood tracheids was observed between September (August in HAV, Isparta and KUTL, lower elevation of Kütahya) and November in KUTH (higher elevation of Kütahya and KIR, Bolu).This showed that relatively cooler sites (higher-elevation and northern sites) formed latewood tracheids one month later than warmer sites.Previous reconstructions of fire seasonality in the region through the analysis of fires scars showed that the majority (87.1%) of the fires occurred in the period of latewood formation [15,16].The results on intra-annual xylem growth suggest that fires in black pine forests in the study region occur between late summer and early fall.However, it is still necessary to remember that our dataset comprises only one year at monthly observations.Even though we are able to obtain approximate periods for wood formation, we need longer and more frequent monitoring to increase the accuracy of model predictions.The proportion and timing of early/latewood formation were assessed in various regions by other studies, albeit for different purposes, to understand the influence of climate on wood formation dynamics.We found that the period of earlywood formation was longest in the KIR site (ca.90 days, between ca.DOY140 and ca.DOY230) and shortest in the HAV site (ca. 10 days, between ca.DOY140 and DOY150).This difference in the length of the earlywood formation period indicates that climate (precipitation and temperature in spring) and latitudinal effects are important controlling factors [31].Studies in Spain on Pinus sylvestris also revealed that cool and wet spring conditions enhance earlywood cell production [32] and can thus affect the duration of earlywood.The KIR site (the Black Sea climate) receives significant amounts of precipitation throughout the year, even in the driest months, compared to the other study sites, and this supports the influence of precipitation on the duration of earlywood formation.The transition from earlywood to latewood may be related to water availability because low water content decreases the turgor pressure in the cells, directly affecting the enlarging phase.This causes smaller, denser, and mechanically stronger tracheid cells to form [33,34]. Latewood cell production starts earlier in the south than in the north, which is probably due to warmer summer temperatures and lower precipitation in southern sites [35].The sites, which is to say HAV under the Mediterranean climate and the KUTL low-altitude site under the Marmara transition climate, have relatively lower precipitation and higher temperature in summer, causing to start latewood formation earlier.Thus, the timing of latewood formation in the sites might be affected by the combination of low summer water availability and the influence of high summer temperatures (Figure 6C), highlighting the importance of both climate parameters.We also found relevant results from a previous xylogenesis study on the timing of latewood formation in black pine trees (in southeastern Spain, between August and October) [10].The results also agree with earlier xylogenesis research in Türkiye on two black pine trees in Istanbul in 1957 and 1958, which showed that earlywood was formed between mid-April and mid-June, while the latewood cells were formed between mid-June and late September [9].Overall, as mentioned earlier, we also still need longer and more frequent measurements to increase the certainty of annual changes in timing of earlywood and latewood formation.
The cumulative number of tracheids showed noticeable differences in xylem growth among the sites.The highest xylem growth was observed in KUTH at the upper elevation (1430 m a.s.l), and the lowest in KUTL was seen at the lowest elevation (1160 m a.s.l).This might be related to elevation differences, even though the elevation difference is short, as well as latitudinal differences between sites.On the other hand, the different growth patterns in the Gompertz model fitted at the tree level for KUTH01 and KIR05 samples may be due to those trees being relatively younger compared to other trees.KIR05 was an especially fast-growing tree.
The onset of cambium activity in the spring is related to warmer climatic conditions or stem heating [5,33,[36][37][38].Since spring temperatures affect the timing of snow melting and the beginning of the vegetation period, a clear relationship between temperature and the resumption of cambial activity after winter has been reported for several conifer species under different environmental conditions [33,39,40].These effects explain the earlier onset of cambial activity in the HAV site, with results suggesting that cambial onset is triggered by temperature as the HAV area has a warmer local climate than other sites.The response of cambium activity onset to temperature affects total tree-ring growth, as shown in dendrochronological studies in western Anatolia, where the radial growth of black pine trees [41] and other tree species [42] was partly driven by early spring temperatures (March-April).More research, including detailed wood anatomical traits and ecophysiology-based measurements (soil water content, evapotranspiration rates, etc.), should be conducted in order to reveal the effect of hydraulic adaptation, especially in regions with contrasting climatic conditions.The number of enlarging cells peaked between the end of May and early June (ca.DOY 150) at the time when the highest photosynthetic activity commonly occurs in response to longer, warm days and available soil moisture [43].
The cell enlargement phase is highly dependent on soil water availability [33,44,45] because low water availability directly restricts the enlargement phase of tracheids during xylem formation due low cell turgor [34].The highest number of enlarging cells is most likely closely related to May-June precipitation in the study region.As reported in several previous dendrochronological studies, May-June precipitation is the main driving factor in the growth of black pine forests in western Anatolia, which also supports our results as regards water availability [42].

Conclusions
In this study, the primary objective was to enhance our understanding of the findings obtained from previous studies in these sites, thereby providing valuable insights into latewood formation via an analysis of extensive data.It is essential to underline that the data acquired in this study, although preliminary and limited, hold significant importance due to their expansive coverage across western Anatolia, presenting a unique and substantial contribution to the existing body of initial knowledge.We were able to determine fire seasonality indirectly by estimating the timing of earlywood and latewood formation periods.This is the first xylogenesis study to expand the climate gradient of black pine forests in Türkiye with the purpose of estimating fire seasonality.Our previous results from retrospective fire history reconstruction underline that most of the fires in western Anatolia occurred in the latewood formation period and that this seasonality remained unchanged over the 600 years.This study on cambial development allowed us to precisely establish that late summer and fall (period August-November) should be considered the period of a high fire risk.Our results are able to partly define the seasonality of fires for the region better than the intra-ring position of fire scars since the although the dataset for the latter is coarse and sparse.The knowledge gathered from this study about fire risk months will help to make an important contribution to defining the regional fire regime in the literature.However, we still need longer-term and more frequently monitored data to improve our knowledge on fire dynamics, increase the certainty of model predictions, and carry out planning activities more effectively for fire managers in black pine forests at risk of fire in Türkiye.The data and method from this study can be applied to the reconstruction of new proxies to advance the long-term knowledge of the seasonality of past fires in order to improve our knowledge of future fire risks and develop more pragmatic fire management plans for different regions.

Figure 1 .
Figure 1.Study area and average climate characteristics.Locations of microcore sampling for this study (yellow dots), and site locations of previous fire history reconstruction studies in western Anatolia (white dots).Stars show the location of meteorological stations.Blue-coloured areas present the distribution area of Pinus nigra J.F. Arnold (black pine) forests in Türkiye [20].Climograph of study areas from (A) Kütahya, (B) Bolu, and (C) Isparta meteorological station (1950-

Figure 1 .
Figure 1.Study area and average climate characteristics.Locations of microcore sampling for this study (yellow dots), and site locations of previous fire history reconstruction studies in western Anatolia (white dots).Stars show the location of meteorological stations.Blue-coloured areas present the distribution

Forests 2024, 15 , 494 7 of 14 Figure 3 .
Figure 3.The number of (standardised) cambium, enlargement, cell-wall-thickening, and mature cells was indicated with dots for the four sites during 2021.Fourth-order polynomial splines were used as smoothing methods for the monthly standardised value of the cell counts for all trees in each site and date (n = 6 trees at each sampling date).The shades indicate standard error for each site.DOY as day of the year.

Figure 3 .
Figure 3.The number of (standardised) cambium, enlargement, cell-wall-thickening, and mature cells was indicated with dots for the four sites during 2021.Fourth-order polynomial splines were used as smoothing methods for the monthly standardised value of the cell counts for all trees in each site and date (n = 6 trees at each sampling date).The shades indicate standard error for each site.DOY as day of the year.

Forests 2024, 15 , 494 8 of 14 Figure 4 .
Figure 4.The tree-level Gompertz models fitted to the cumulative number of cells during the 2021 growing season.The black dots show the corresponding days of the year for the beginning of latewood formation based on the Mork index criteria for the six different trees within each of the four sites.Different colours represent the different trees within the site.DOY as day of the year.

Figure 5 .
Figure 5. Dates of completion of (A) 5% and (B) 95% of the total annual growth for each site.Sites are ordered from north to south (left to right).Dots in different colours represent the outliers.Black dots represent the mean values.Different letters show the significant differences among means at p < 0.05.DOY as day of the year.

Figure 4 .
Figure 4.The tree-level Gompertz models fitted to the cumulative number of cells during the 2021 growing season.The black dots show the corresponding days of the year for the beginning of latewood formation based on the Mork index criteria for the six different trees within each of the four sites.Different colours represent the different trees within the site.DOY as day of the year.

Forests 2024, 15 , 494 8 of 14 Figure 4 .
Figure 4.The tree-level Gompertz models fitted to the cumulative number of cells during the 2021 growing season.The black dots show the corresponding days of the year for the beginning of latewood formation based on the Mork index criteria for the six different trees within each of the four sites.Different colours represent the different trees within the site.DOY as day of the year.

Figure 5 .
Figure 5. Dates of completion of (A) 5% and (B) 95% of the total annual growth for each site.Sites are ordered from north to south (left to right).Dots in different colours represent the outliers.Black dots represent the mean values.Different letters show the significant differences among means at p < 0.05.DOY as day of the year.

Figure 5 .
Figure 5. Dates of completion of (A) 5% and (B) 95% of the total annual growth for each site.Sites are ordered from north to south (left to right).Dots in different colours represent the outliers.Black dots represent the mean values.Different letters show the significant differences among means at p < 0.05.DOY as day of the year.

Figure 6 .
Figure 6.Dates at which the formation of (A) transition between earlywood and latewood and (B) latewood for each site, where black dots represent the mean values.(C) shows the meteorological station data of mean temperature and precipitation for summer (from June to August).Sites are ordered from north to south (left to right).Dots in different colours represent the outliers.Different letters show the significant differences among means at p < 0.05.DOY as day of the year.

Figure 6 .
Figure 6.Dates at which the formation of (A) transition between earlywood and latewood and (B) latewood for each site, where black dots represent the mean values.(C) shows the meteorological station data of mean temperature and precipitation for summer (from June to August).Sites are ordered from north to south (left to right).Dots in different colours represent the outliers.Different letters show the significant differences among means at p < 0.05.DOY as day of the year.

Table 1 .
Study sites and characteristics of Pinus nigra J.F. Arnold (black pine) trees.