Evaluation of Compression Wood Incidence Under Different Thinning Regimes in Late Rotation of Pinus taeda

Abstract

Compression wood (CW) negatively affects the industrial quality of Pinus taeda by causing distortion in sawn boards and is visually characterized by a darker reddish colour. Thinning is considered a key factor influencing its formation, but the reported effects have been inconsistent. This study evaluated CW incidence at final rotation under five thinning regimes: 500-200, 500-325, 800-600-400, 700-450, and 1000-650 trees.ha⁻¹. The defect was assessed on log ends, basal discs, and sawn boards. Although overall CW severity was low, regimes differed significantly. The 500-325 trees.ha⁻¹ regime showed the highest stain and board defects, while the 500-200 trees.ha⁻¹ regime of similar intensity had lower values, indicating that intensity alone does not explain the occurrence of defects. After thinning, CW in growth rings increased and was positively associated with ring width and negatively with stand density index, indicating that reduced competition and accelerated radial growth are linked to higher formation levels. Visible CW staining on log ends was moderately correlated with board defects, indicating its potential as a practical, low-cost indicator of log quality. Thinning affects compression wood through its impact on growth and stand structure. In addition to intensity, timing and the effect of the wind must also be considered. Moderately intensive regimes help minimize defects, although their practical adoption may be limited by commercial priorities.

1. Introduction

Pinus taeda is widely cultivated as a commercial forest species in various regions around the world, valued for its fast growth and suitability for solid wood production. Originally native to the southeastern United States, it has been successfully established in countries such as Argentina, Brazil, Uruguay, Venezuela, China, South Africa, Australia, and New Zealand [1]. The formation of CW has been recorded in South Africa in Pinus taeda plantations in the Mpumalanga and Limpopo provinces [2]. It has also been reported in Brazil, where it is recognized as a factor that compromises wood quality and limits its industrial utilization [3]. Beyond Pinus taeda, the occurrence of compression wood has also been reported in other conifer species cultivated in plantations. For instance, its presence has been documented in Pinus radiata in Chile [4,5], South Africa [6], and New Zealand [7,8]; in Pinus sylvestris in the UK [9] and Spain [10]; and in Pinus nigra in the UK [11,12].

In Uruguay, Pinus taeda is one of the most important forest species in terms of planted area, representing a significant portion of the 128,000 hectares occupied by pine trees, which accounts for 14.3% of the country’s total forested area [13]. Its primary use is for industrial processing to produce solid wood products, mainly intended for export.

These plantations have been developed from genetic material introduced from the United States and South Africa [14], under silvicultural regimes focused on high-quality wood production. A pre-commercial thinning is usually conducted in the early stages, followed by one or two commercial thinnings during the rotation.

Among the various factors affecting the industrial quality of this wood, compression wood is identified as having the most significant negative impact. Based on internal assessments and observations within the Uruguayan forest industry, CW is estimated to cause around a 10% loss of the potential foreign exchange earnings from Uruguay’s pine mechanical transformation sector. Compression wood is a type of reaction wood found in conifer species that exhibits markedly different physical, chemical, and mechanical properties compared to normal wood [15]. One of its most distinctive features is a darker reddish-brown colour (stain), resulting from its highly lignified, thick-walled tracheids [16]. CW causes defects in sawn boards and veneer sheets, such as distortion, cracking, reduced dimensional stability, and lower mechanical strength [17,18,19]. Consequently, it disrupts production processes, causing misalignments and even equipment damage, reducing both yield and final product quality.

CW formation can result from several factors, including wind, growth rate, genetic material, and thinning plays a particularly relevant role. Schweitzer [3] reported that severe thinning has been associated with increased CW formation in Pinus taeda. Similarly, Bamber and Burley [20] suggested that thinning can promote CW development by altering the vertical equilibrium of tree stems. Post-thinning canopy opening encourages lateral crown growth and may induce trunk leaning, both of which are conducive to CW formation. Increased spacing between trees also enhances wind exposure, potentially intensifying this effect. In this context, Cown [21] found that, in Pinus radiata, CW formation was directly related to growth rate increases following thinning. The author proposed that this could be linked to hormonal changes, such as elevated auxin production from crown expansion, or to increased stem oscillation from wind, with both acting as stimuli for reaction wood development.

In contrast, Cameron [11] found that unthinned trees exhibited significantly higher CW incidence, up to three times more, than those that were thinned. His findings indicate that high stand density leads to intense competition for space, prompting trees to prioritize height growth over stem and root development. As a result, CW development in dense stands is believed to be strongly influenced by phototropic movements in response to light competition.

Given CW’s detrimental impact on product quality, a thorough understanding of the factors contributing to its formation is essential. Among silvicultural practices, thinning is one of the main tools used in Pinus taeda management. However, most available studies have focused on other species or planting conditions and have reported differing outcomes regarding the influence of thinning on CW development. This variability across existing studies suggests that certain thinning regimes may be more effective than others in reducing CW formation. Our study design include thinning schemes that considers the contrasting number of suckering trees per hectare and the stages of the crop cycle at which thinning is applied. This allows us to evaluate a wide range of management situations, ultimately providing a comprehensive overview of the effects of these management practices.

Our hypothesis is that different thinning schemes affect compression wood formation due to mechanisms related to growth and competition among individuals. The primary objective of this study was to characterize compression wood formation in Pinus taeda under different thinning regimes and to analyze its relationship with growth parameters under northern Uruguay conditions. The specific objectives were as follows: (i) quantify the CW presence for each of the thinning schemes evaluated; (ii) measure the relationship between CW presence and individual and population growth; and (iii) determine the usefulness of the visual log grading as a predictor of the deformations observed in the boards after sawing.

2. Materials and Methods

2.1. Study Area

The study was conducted in a Pinus taeda thinning trial, located in the Tacuarembó department in northern Uruguay (31°33′22″ S, 55°44′40″ W) (Figure 1). The soils at the site are deep, sandy-textured, highly acidic, and have low natural fertility. According to the USDA soil taxonomy, they are classified as Hapludult and Hapludalf [22]. The distribution of rainfall during the historical period (2002–2025) averaged 1386.8 mm per year. The average temperature was 17.6 °C, with maximum and minimum values of 32.6 °C and 7 °C, respectively. The test site is located in a region where the most frequent and fastest winds come from the northeast to southeast directions with an average speed of 1.9 × 10⁵ and an average maximum speed of 3.2 × 10⁵ m.s⁻¹, respectively [23].

The stand was planted in 1997 using seed from a first-generation seed orchard in Coastal Plain region, Georgia, USA, at an initial density of 1000 trees.ha⁻¹ with a spacing of 4 m between rows and 2.5 m between trees. The stand was pruned five times and ≤40% of the live crown was removed in each intervention.

2.2. Trial Design and Thinning’s Regimes

The thinning trial was installed in the year 2000, when the stand was three years old. A total of 58 square plots of 30 m × 30 m were established using a randomized complete block design with three blocks defined by the initial basal area as follows: Block 1 (>4.47 m².ha⁻¹), Block 2 (3.96–4.47 m².ha⁻¹), and Block 3 (<3.96 m².ha⁻¹) (Figure S1, Supplementary Materials). A total of 19 different thinning treatments were originally implemented in the trial, varying in the number of interventions, timing (ages at which thinning occurred), intensity (final residual stand density), and the interval between successive thinnings.

For this study, five thinning regimes were selected (

Table 1. Description of the selected treatment.

Thinning Regime	Initial Density (Trees.ha−1)	Target Density after Thinning (Trees.ha−1)				Last Density (Trees.ha−1)
		Age (years)				Age (years)
		3	6	9	12	22
500-200	1000	500		200		204
500-325	1000	500			325	304
800-600-400	1000	800	600	400		393
700-450	1000	700			450	437
1000-650	1000				650	559

) These were chosen because they represent the range of residual stand densities relevant to current silvicultural practices in the Uruguayan forest sector, while the remaining regimes were either operationally impractical or too similar to others in terms of final stand density and thinning regime The number of rows and trees in each of the plots evaluated in this study are described in Table S1, Supplementary Materials.

2.3. Tree Selection and Field Sampling

To ensure a representative sample of tree sizes, individuals were selected based on the diameter at breast height (DBH, cm) distribution within each thinning regime. DBH was measured at 25 years of age for all trees within each plot, and diameter classes were defined for each thinning. Based on the relative frequency of each class, seven trees were selected per plot, following the class proportions established at the thinning regime level. This resulted in a total of 105 trees (21 per thinning regime, 7 per plot) which were manually felled using a chainsaw. After felling, a basal disc (~5 cm thick) was collected from each tree for visual assessment of CW the growth rings, and a 3.5 m log was extracted from the base of the stem for sawing and subsequent evaluation of CW in the resulting boards.

2.4. Visual Analysis of Compression Wood on Log Ends

Both ends of each log were visually examined for CW presence and classified on a scale from 0 to 3 based on the proportion of the cross-sectional occupied by the stain (0: absent, 1: <1/3, 2: 1/3–2/3, 3: >2/3) (Figure 2). Each log was assigned a final score, obtained as the average of its two ends (basal and upper). All evaluations were performed by a single trained observer to avoid inter-rater variability, and the scoring procedure was calibrated in previous surveys of >300 trees from the same research project. Similar categorical approaches for visual CW severity have been used in conifer species [24,25]. The use of stain as a proxy for CW has been previously validated for Pinus taeda under similar site and management conditions, where stained areas consistently showed anatomical features of CW (shorter tracheids and thicker cell walls; Table S2, Supplementary Materials), in agreement with published characterizations of CW in conifers [26,27,28].

2.5. Preliminary Sampling for Juvenile–Mature Wood Transition

To determine the juvenile–mature wood transition, fibre length was analyzed in 10 basal discs representing different diameter classes from the study stand Measurements were conducted on earlywood tracheids from successive rings, samples were macerated following Kraus and Arduin protocol [29], and fibre length was measured. The transition ring was defined as the point where fibre-length values showed a clear stabilization trend, indicating the onset of mature wood formation. Based on this change in slope, the transition occurred around ring 11, consistent with previous reports for Pinus taeda [30,31]. This ring was used as the boundary to separate juvenile (inner) and mature (outer) boards during sawing (Figure S2, Supplementary Materials).

2.6. Sawing Methodology

Logs were sawn into 1-inch-thick boards using successive parallel cuts (Figure 3), corresponding to commercial dimensions. Boards were not edged, and therefore retained bark on their sides. From the 105 processed logs, between 9 and 15 boards were obtained per log, yielding a total of 1144 boards for evaluation. During this process, boards originated from the inner zone—expected to contain juvenile wood—were intentionally kept separate from those obtained from the outer zone, where mature wood was expected to be present, based on the estimated transition point determined through preliminary sampling (Section 2.6).

2.7. Visual Analysis of CW in Green Boards

Both faces of each board were visually examined for CW presence and classified on a scale from 0 to 3, based on the proportion of the surface area occupied by visible CW stain (0: absent, 1: <1/3, 2: 1/3–2/3, 3: >2/3). A final score was assigned to each board using the maximum and the average of the two faces.

2.8. Evaluation of Defects in Dried Boards

Boards were dried in a conventional kiln under controlled industrial drying conditions, following a standard softwood drying schedule. After drying, boards were evaluated for crook, bow, and twist, based on the UNIT 1461 standard [32]. Rather than using absolute deformation values, which do not necessarily reflect the board’s industrial usability, each board was classified into one of two practical categories: “with defects” or “without defects”. This classification was based on threshold values provided by a local industrial company, which defines acceptable limits for each type of deformation according to commercial quality requirements (Table S3, Supplementary Materials). These thresholds were developed in accordance with the UNE-EN 14081-1 standard [33] for structural timber grading.

2.9. Visual Analysis of CW in Discs

Basal discs from all selected trees were visually assessed to quantify the presence of CW. Each disc was divided into eight equal sectors by drawing equidistant radii, and CW was evaluated ring by ring. Each sector, defined as the portion of a ring between two consecutive radii, was scored based on the extent of visible stain (0: absent; 0.5: small but clearly identifiable stain; or 1: large and prominent stain).

Ring widths were measured along each radius using a ruler to estimate the area of each sector. These values were used to calculate the CW area per sector by weighting the stained proportion accordingly. Summing the stained sectors yielded the total CW area per ring, which was then expressed as a percentage of the total ring area. These results were used to calculate the average CW percentage per ring across trees within each thinning, and the total CW percentage per disc.

To quantify the post-thinning CW response over time, the area under the curve (AUC) of CW proportion was calculated from ring 9 onward using numerical integration. Therefore, AUC represents the cumulative CW proportion over time, combining both the magnitude and duration of CW formation after thinning.

2.10. Statistical Analysis

As several variables did not meet normality assumptions, non-parametric tests were applied. Kruskal–Wallis tests were used to evaluate the effect of thinning on the proportion of boards with defects, the CW stain score in boards, the CW stain score on log ends, and the CW proportion in growth rings. When significant differences were detected, pairwise comparisons were performed using Dunn’s test, applying a significance level of α = 0.05.

Spearman correlation coefficients (r_s) were calculated to explore the relationship between CW incidence at the growth ring level and several dendrometric and silvicultural variables, as well as between the proportion of boards with defects and stained area in rings, and between visual CW scores on log ends and boards with defects. The variables considered were Basal Area (G, m²·ha⁻¹), Stand Density Index (SDI) [34], Stand Age (Age, years), Diameter at Breast Height (DBH, cm), Total Height (Ht, m), Age at Thinning (AgeT, years), Number of Trees per Hectare (N), Ht/DBH ratio (m·cm⁻¹), and Ring Width (Grw, mm). All analyses were performed using InfoStat software version 2020 [34].

3. Results

3.1. Effect of Thinning on Growth and Competition Between Trees

In general, growth and competition results among individuals begin to diverge from age 6 onward, coinciding with the application of the first commercial thinning (Figure 4). Individual growth results (DBH and Ht) across thinning regimes were within expectations, being higher under wider spacings. The Ht/DBH ratio showed that trees under narrower spacings were more slender, since higher competition promoted greater height growth relative to DBH. The greatest change in this relationship is observed after the thinning applied at 9 years of age, and this is the moment from which an increase in the presence of CW begins to be detected. The G exhibited the opposite trend, as this parameter was directly related to stand density. Beyond the early years, the maximum Grw peak occurred after the thinning at age 12, when competition (SDI) was most reduced. Less intensive thinning regimes (1000-650 and 700-450 trees·ha⁻¹) spent most of the rotation under severe competition, while the more intensive regimes (500-200 and 500-325 trees·ha⁻¹) remained under increasing competition. Changes in tree populations across thinning regimes were not proportionally reflected in growth variables, either at the individual or per-hectare level.

3.2. CW Incidence by Thinning Regime

3.2.1. CW Stain on Log Ends

CW presence levels were generally very low across all thinning regimes; however, data analysis indicated significant differences among treatments (H = 14.5, p-value ≤ 0.0001) (Figure 5). The 500-325 trees·ha⁻¹ regime exhibited the highest mean CW stain proportion, significantly greater than the other thinning treatments. No significant differences were detected among the remaining regimes, which showed lower and more uniform values. Among the intensive thinning regimes, only 500-325 trees·ha⁻¹ showed elevated CW stain levels, while 500-200 trees·ha⁻¹ remained comparable to the lighter regimes. Moreover, except for the 500-325 trees·ha⁻¹ regime, both boards and log ends exhibited only a very reduced stained surface.

3.2.2. CW Distribution Across Annual Rings

Across all thinning treatments, an increase in CW proportion was observed in the years following thinning response (Figure 6). The magnitude and duration of this increase varied among regimes, but in all cases CW levels declined toward the final stage of the rotation.

The area under the curve (AUC) of CW proportion, calculated from ring 9 onward to represent the post-thinning period, did not differ significantly among thinning regimes (Figure 7). However, lower mean AUC values were associated with the moderate thinning regimes (800-600-400 and 700-450 trees.ha⁻¹), while both intensive (500-325 and 500-200 trees.ha⁻¹) regimes showed higher values. This was consistent with the log end stain results.

3.2.3. CW Stain in Boards

A significant effect of thinning regime was observed on the mean CW stain proportion in boards (Figure 8). Both the average stain across board faces (H = 121.89, p-value ≤ 0.0001) and the maximum (H = 115.27, p-value ≤ 0.0001) yielded consistent results. The 500-325 trees.ha⁻¹ regime showed the highest mean values, significantly greater than all other treatments. No significant differences were found among the remaining thinning regimes.

3.2.4. Board Defects

No significant differences were found in the proportion of boards, with defects between the innerwood (up to ring 11) and outerwood (beyond ring 11) regions (H = 0.79, p-value = 0.37) (Figure 9). The proportion of boards with defects was similar in both radial positions, with relatively high averages and considerable variability in each region.

Significant differences were observed in the proportion of boards with defects among the different thinning regimes (H = 12.37, p-value = 0.0148) (Figure 10). The 500-325 trees.ha⁻¹ thinning regime showed the highest mean value (61.9%), significantly higher than all other regimes. It was followed by the 500-200 trees.ha⁻¹ regime, which, although not significantly different from 500 to 325 trees.ha⁻¹, had a much lower proportion (17.6%). The 800-600-400 (13.2%), 1000-650 (10.3%), and 700-450 trees.ha⁻¹ (5.9%) regimes showed lower proportions of defects, with no significant differences among them or compared to the 500-200 regime. The 700-450 trees.ha⁻¹ regime had the lowest proportion of defects and was significantly different from the 500-325 trees.ha⁻¹ regime.

A significant effect of thinning regime was also observed when considering the proportion of trees yielding boards with severe defects, defined as those in which more than 50% of their boards were classified as boards with defects (H = 7.36, p-value = 0.0334) (Figure 11). The mean proportion of such trees also varied across treatments. The 500-325 trees.ha⁻¹ regime again showed the highest incidence, with over 80% of trees affected. In contrast, the remaining thinning treatments exhibited much lower and more homogeneous proportions, particularly under the moderate thinning regimes.

3.3. Relationship of Dasometric and Silvicultural Variables with CW at the Growth Ring Level

To explore these relationships, the dynamics of stand and dendrometric parameters were analyzed throughout the rotation. Correlation analysis identified notable associations between silvicultural and dendrometric variables and the proportion of CW in basal stem discs (Figure 12). The most pronounced negative associations were found for stand-level variables, such as G, SDI, and Age (Figures S3 and S4, Supplementary Material). In contrast, Grw exhibited the strongest positive correlation (r_s = 0.52), indicating a tendency for greater radial growth (reflecting lower inter-tree competition) to be associated with increased CW formation. Other variables, including Age and AgeT, showed moderate negative correlations, while N displayed a weaker but still significant positive relationship. Individual growth variables such as DBH and Ht also showed moderate negative correlations with CW incidence.

3.4. Predictive Value of Visual CW Assessment on Log Ends

Significant differences were also observed when grouping logs by average CW stain category (0, 0.5, 1) (H = 12.6, p-value < 0.0001) (Figure 13a). Logs classified in categories 0.5 and 1 had higher proportions of boards with defects, while logs in category 0 showed the lowest values. A moderate positive correlation (r_s = 0.44, p-value ≤ 0.0001) was found between the CW stain score on log ends and the proportion of boards with defects (Figure 13b). No significant relationship (r_s = 0.12, p-value = 0.23) was detected between CW proportion in basal discs and the proportion of boards with defects (Figure 14).

4. Discussion

The results of this study provide new insights into CW development in Pinus taeda plantations under different thinning regimes in northern Uruguay. The findings support the hypothesis that thinning influences CW formation, although the relationship observed between both parameters is not particularly strong. The detection of a low proportion of CW suggests that its association with some silvicultural parameters is of limited magnitude. Thinning is a common management practice in these plantations, which highlights the need to understand its effect on CW formation. Previous studies conducted in the planting areas of Uruguay have shown that the presence of CW generally does not exceed 10% of the total wood volume produced. According to Bendtsen and Senft [35], Pinus taeda is highly susceptible to CW formation, particularly during the early years of growth. At these stages, CW formation occurs simultaneously with juvenile wood [30], although the two wood types differ in their chemical composition and anatomical features and arise from different causes [35]. Therefore, studying CW formation at later stages, coinciding with responses to management interventions such as pruning and thinning, is of particular interest. In general, thinning may affect growth-related variables (which are the primary target of this practice) but also crown structure, branch size, and stem inclination [9]. These factors have been identified as sources of asymmetry in light availability, ultimately causing mechanical imbalances within the stem [36]. In addition, wind exposure plays a role, particularly in stands with wider spacing, where trees are more exposed [37]. For these reasons, it is necessary to examine the relationship between silvicultural parameters and the intensity and timing of CW formation. Identifying management variables that reduce the occurrence of this abnormal wood could have a positive effect on wood quality under these production conditions. Nevertheless, results reported in the literature indicate difficulties in isolating the effect of single variables, as growth conditions strongly interact with site characteristics [38].

4.1. Influence of Thinning on Growth and Competition Among Trees

The effects of thinning were evident across all variables analyzed, which is consistent with evaluations conducted on Pinus taeda, including those from this trial [39,40,41,42]. A fairly strong indirect relationship was observed between the rate of individual tree growth and thinning intensity. The most pronounced changes occurred between years 12 and 15, a period characterized by the lowest levels of competition across all treatments. Up to this stage of the rotation, the more intensive thinning regimes (500-325, 500-200, and 800-600-400 trees.ha⁻¹) achieved G and N values that allowed the growth potential of this species to be maximized. In contrast, the other thinning regimes entered stages of severe competition or “overstocking” from the early years [43]. These differences in site occupancy may reflect varying levels of wind exposure and their possible association with CW formation, which will be analyzed in the following sections.

4.2. Influence of Thinning Intensity and Timing on CW Formation

The low proportion of CW detected across all thinning regimes may be explained by the seed source evaluated in this case (Coastal Plain, USA), since the incidence of this parameter has been reported by several authors [44,45]. The high proportion of CW observed under one of the most intensive thinning regimes (500-325 trees·ha⁻¹) is consistent with previous studies showing that abrupt changes in spacing and canopy structure after thinning. This can destabilize stem posture through crown imbalance or stem inclination, and thereby promote reaction wood formation, particularly under increased wind exposure [5,38]. This thinning regime combines advanced age and a low number of remaining trees, which alters the conditions of competition between individuals, as will be analyzed later.

However, another highly intensive regime (500-200 trees.ha⁻¹) showed a notably lower incidence of CW, suggesting that thinning intensity alone does not fully explain the patterns observed. This could be because in situations of greater competition from neighbouring trees, stems tend to deviate toward areas with increased light availability

The higher levels of CW observed during the early years of growth were not related to the management variables evaluated in this study and, according to Larson et al. [30], are associated with the action of growth-promoting hormones. The temporal distribution of CW in growth rings indicated an increase in CW formation during the years following thinning across all regimes. This pattern suggests that the post-thinning period may represent a critical window for CW development. These results are consistent with numerous studies reporting that CW formation tends to increase in the years following thinning [2,46]. According to some authors, this effect is attributed to elevated levels of indole-3-acetic acid in the cambial region of several conifer species, including pines [47]. These hormonal changes are also associated with wind action, which may be more evident in more intensively thinned stands. Stem curvature due to crown weight in sites frequently exposed to wind has also been reported as a contributing factor [37]. Additionally, some authors have linked stem taper with CW formation, although through mechanisms other than tree population density [48]. Results obtained from a trial of different pruning schemes located near our thinning trial show that CW presence occurs to a greater extent on the side opposite the predominant wind direction in that area (See Figure S5, Supplementary Materials). Records from 2009 onwards of wind direction and speed in the test area show that the most intense and most frequent winds occur from the south–southeast areas [23].

CW levels in the final stages of growth decreased markedly, indicating the disappearance of predisposing factors for its formation regardless of the thinning regime applied. The highest incidence of CW following thinning was detected around ages 18–20, but under the more intensive thinning regimes this effect persisted over longer periods. The reduction in CW presence after the thinning phase is consistent with the findings of Cown [49] and Voorhies [50], who evaluated P. radiata and P. ponderosa, respectively.

4.3. Implications of CW on Product Quality

The results obtained from board defects followed the same trend as the visual observations of CW presence on the log faces. The defects observed in boards containing CW are primarily attributed to greater longitudinal shrinkage compared to normal wood [2,51]. These shrinkage effects, which are more evident after drying, are mainly due to three characteristic features of CW relative to normal wood: (i) a larger microfibril angle in the S2 layer of tracheid walls [52,53], (ii) higher lignin and polysaccharide content [54,55,56], and (iii) greater tracheid wall thickness [57]. Several studies on Pinus radiata have also shown that the microfibril angle can be influenced by spacing, with more intensive thinning leading to an increase in this variable [58,59,60,61]. The increase in CW content can also affect wood properties that have implications for its end use. Results obtained by Diaz-Vaz et al. [4] show that CW has higher basic density values than normal wood, although these differences depend largely on the level of this abnormal wood. This is mainly explained by an increase in the thickness of the tracheid walls, which in turn leads to higher compressive strength values parallel to the grain [62] while simultaneously decreasing the modulus of elasticity due to alterations in the microfibril angle [63], as mentioned previously. Another negative effect of the presence of CW is a reduction in the drying rate of boards in the order of 40 to 100% [64].

4.4. Relationship Between CW and Growth Dynamics

The correlation analysis suggests that CW development may be influenced by a combination of factors related to both tree growth dynamics and stand structure. Although several variables were statistically associated with CW incidence, such as G, SDI, and Grw, most relationships were of moderate strength. The apparent contradiction in the relationship obtained between individual growth rates (obtained with more intensive thinning) and the presence of CW could indicate that the correlation values obtained (−0.45 and −0.37) are not due to a biological cause. Therefore, individual growth rates do not appear to be a relevant factor in the formation of this type of wood. At the time of the second intervention, the 500-200 trees.ha⁻¹ regime (thinned at age 9) was still in a phase of moderate-to-low competition. On the other hand, the 500-325 trees.ha⁻¹ regime (thinned at age 12) had already reached higher competition pressure by the time of intervention. Among all the regimes evaluated, this was the only case where competition level changed from severe to increasing at the time of the second thinning. This was reflected in the change in basal area growth rate, which may help explain the higher presence of CW [65]. According to de Villiers [66], thinning at advanced stages of the rotation tends to result in a high proportion of suppressed trees (due to elevated competition), which are relatively flexible and less stable.

In addition to these variables, other factors, such as wind action—which tends to be more evident under wider spacing—have also been reported to influence CW formation. From the stand managementperspective, this outcome is particularly relevant, as it suggests that moderate or light thinning may help minimize CW formation and its negative impact on product quality.

Among the treatments evaluated in this study, the 500-325 trees.ha⁻¹ regime most closely resembles those applied in commercial practice in terms of rotation timing and residual stand density. From the perspective of yield and wood quality, this regime recorded the highest levels of individual growth, as well as one of the highest levels of CW and board defects. In this regard, the implementation of some intermediate thinning should be considered, for example, around the age of 6, with a remaining population of 650 trees.ha⁻¹. This would prevent a more abrupt change in growth rates and exposure to wind. With a thinning scheme of around 650-450-325 trees.ha⁻¹, it would be possible to achieve high DBH values, which is one of the desired objectives, in addition to a low presence of CW. Implementing additional thinning would increase operating costs, so it would be necessary to analyze the cost–benefit ratio of this type of thinning scheme.

4.5. Value of Visual Assessment on Log Ends

Given the similarity observed in CW stain values at the ends of logs, visual grading does not appear to be sufficiently precise to quantify CW presence in all cases [67]. This limitation is particularly relevant when CW occurs at low to moderate levels, as was the case in this study.

Although the correlation was moderate (r_s = 0.44, p value < 0.0001), these findings confirm that external visual cues can serve as effective proxies in operational settings, particularly when more advanced assessments are not feasible. It should be noted, however, that not all trees with severe CW exhibit the characteristic darker staining bands [2]. Based on the results obtained in Uruguay under conditions similar to those of this study, identifying severe levels of CW presence improves the prediction of defects observed after sawing. However, the visual grading of logs is not very accurate as an indicator of the presence of WC in cases of low levels of this type of abnormal wood.

5. Conclusions

CW occurrence in Pinus taeda was influenced by thinning regime and subsequent stand development. Moderate thinning regimes resulted in lower CW-related defects, indicating that thinning can be used as a management tool to improve wood quality. Therefore, operational thinning strategies should be evaluated for their ability to minimize CW occurrence while maintaining diameter growth objectives. Competition conditions, thinning time, and wind exposure are the factors that together have the greatest relationship with the formation of CW. Visible CW stain on log ends can assist in identifying higher-risk logs, but this should be used as a supportive, not exclusive, screening tool. The overall moderate relationships observed confirm that CW is multifactorial, highlighting the need for future research integrating silvicultural context with tree structural (e.g., stem lean and crown asymmetry) and environmental (e.g., wind exposure) factors to better explain CW formation and improve wood quality.