DOI:10.54614/forestist.2022.

21031
Forestist 2022: 72(2): 105-111 Original Article

Diametric Increment Patterns of Open-Grown and Under Competition

Pinus taeda L. Trees in Southern Brazil
Cristine Tagliapietra Schons1 , Emanuel Arnoni Costa2 , Lorena Oliveira Barbosa3 , César Augusto Guimarães Finger1,4
1
Graduate Program in Forest Engineering, Federal University of Santa Maria (UFSM), Rio Grande do Sul State, Brazil
2
Department of Forest Engineering, Federal University of Uberlândia (UFU), Minas Gerais State, Brazil
3
Graduate Program in Forest Engineering, Federal University of Lavras (UFLA), Minas Gerais State, Brazil
4
Graduate Program in Agroecosystems, Federal Technological University of Paraná (UFTPR), Paraná State, Brazil

Abstract
The main goal of this study was to evaluate the patterns of diametric increment of open-grown and under competition
trees of Pinus taeda L. as an expression of breast height diameter. The dendrometric variables and the diameter increment
of these trees were measured. The increments were obtained by measuring tree rings on cores extracted at breast height
using a Pressler borer. An allometric model was fitted to describe the percentage periodic annual increment in diameter
(PAId%) as a function of the initial diameter (di) for two tree groups. The analysis of covariance (ANCOVA) allowed com-
paring the effect of intercept (α0) and slope (α1) of the fitted equations between groups. The overall average of PAId5 was
0.83 cm year−1 for open-grown trees and 0.57 cm year−1 for under competition trees. The ANCOVA showed differences
in intercept (α0) − (Pr > F = 0.0066) and not in slope (α1) − (Pr > F = 0.8883) for the adjusted regression equations. This
finding enabled the development of a singular equation to express the increment, using dummy variables, described
 = exp é6.3725 + 0.7982 ´ D1 - 1.6212.ln d + 0.1960 ù , with R2 = .69 and Syx = 2.4%. The results of this study con-
as: PAId% ê ( )i
2 û
ú
ë
tribute to highlight the influences of the growth space in the increment of the species, as well as its maximum potential,
elements that are very important for the monitoring and decision-making process in silvicultural activities of individual
P. taeda trees.
Keywords: Annual tree rings, forest management, individual trees, modeling, potential growth

Introduction

Loblolly Pine (Pinus taeda L.), native to the southeastern region of the United States, is an important timber
species, and it has been widely planted in tropical and subtropical regions around the world, such as in South
Africa, New Zealand, Australia, and Brazil (Albaugh, 2018). In southern Brazil, Pinus taeda has been one of the
most planted species since the 1960s, due to its adaptability to climatic conditions, its diversity in use, and rapid
growth (Bognola et al., 2008). Forests of this genus currently cover an area of approximately 1.43 million hectares
in this region, concentrating around 87% of the total pine plantations in the country (Ibá, 2020).
Cite this article as:
Schons, C. T., Costa, E. A., Considering, therefore, the relevance of this forest production, it is essential to understand the growth patterns
Barbosa, L. O., & Finger, C. A. G. and factors that affect the trees’ increment for researchers and managers to assess and improve silvicultural
(2022). Diametric increment practices (Lhotka, 2017). In this sense, the estimation and projection of growth in a stand can be carried out by a
patterns of open-grown and variety of methods, depending on the required detail, and therefore applicable to the forest, size classes, or trees.
under competition Pinus taeda In many of the studies developed so far for the genus Pinus, and especially for the species P. taeda, the growth
L. trees in southern Brazil.
Forestist, 72(2), 105-111. modeling has been based on size classes models, characterized by the use of a probability density function,
as in the studies of Eisfield et al. (2005), Elesbão and Schneider (2011), and Gomes et al. (1997). More recently,
Corresponding Author: although still scarce, initiatives to consider the tree as the basic sampling unit for the species have emerged
Cristine Tagliapietra Schons
e-mail: cristschons@gmail.com
(Miranda, 2016; Zimmermann et al., 2016), then linked to the concept of tree-level modeling, with expected
gains in accuracy in relation to other approaches. For these tree-level models, the accepted hypothesis is that
Received: July 18, 2021 the total growth of the stand is given by the individual sum of each tree that composes it (Hasenauer, 2000;
Accepted: October 27, 2021
Kuehne et al., 2019).
Content of this journal is licensed
under a Creative Commons Attribution-
NonCommercial 4.0 International Licence.

105
 Schons et al. Increment Patterns of Pinus taeda L. Trees
Forestist 2022: 72(2): 105-111

In this context, one of the most important factors to simulate the mortality due to overstocking and the second consisting of solitary
growth of individual trees is the evaluation of the competition effects, trees expressing their empirical maximum tree dimensions, without
that is, the influence of population characteristics where the tree the influence of competitors throughout their development. Both are
grows. Competition among trees for resources affects the growth rate located on the same exposure and type of soil.
of woody species, due to limited access to water, light, and nutrients
(Johnson & Smith, 2009). This limitation is associated with the growth Data Collection
space, which refers to the availability of all the resources necessary for a Ten trees were sampled in each condition type (20 trees in total). The
tree to exist in a given site, and the extent of this space is influenced by diameter at breast height, considering 1.3 m from ground level (d, in
the surrounding trees (Costa et al., 2015). cm), was measured with a diametric tape and the height (h, in m) with
a Vertex IV hypsometer. Additionally, two increment cores per tree were
Conceiving that the degree of competition in a forest determines sampled with an increment borer of 0.4 cm in diameter. The samples
the individual growth dynamics of trees, then, open-grown trees can were air dried and sanded for later analyses of the tree rings and evalu-
be considered as a reference for the maximum potential of a species, ation of diametric increment. The characteristics of the sampled trees
in contrast to trees growing under competition (Hasenauer, 1997). are shown in Table 1.
Consequently, understanding and comparing the development of
open-grown and under competition trees supports the development Calculation of Increments
of management guidelines, helping in the decision-making process The evaluation of the average rates of periodic annual increment in diam-
associated with interventions in forest stands. eter (PAId) in 5, 10, 15, and 20 years was determined by the expression:
PAId = ( di + Dt - di ) / Dt
Furthermore, a prerequisite for studies within this perspective is obtain- (1)
ing reliable data. In this space, the analysis of tree rings is an important
where PAId is the periodic annual increment in diameter in cm.ano−1,
tool with a potential for rapid response, providing information on the
di+Δt is the current diameter in cm, di is the initial diameter in cm, and Δt
growth rates and age of trees. Studies that use this tool are called den-
is the annual variation.
drochronological, as they investigate changes over time, through the
responses expressed by trees (Andreacci et al., 2014). The percentage periodic annual increment in diameter (PAId%) was
determined to be used in the modeling as a function of the initial
Modeling of future increment based on a current diameter is prac- diameter (di). This type of relationship allows to predict the future
tical in comparison to the models that use other independent vari- increment based on a current diameter, which is calculated by
ables as crown diameter, tree height, age and stand density, indices of the expression:
competition, among other variables. Therefore, considering the great
relevance of the P. taeda species for the forestry sector, the present éd -d ù
PAId% = ê i + Dt i ú ´ 100
study aimed to evaluate the patterns of diametric increment of open- di
ë û (2)
grown and under competition trees as an expression of breast height
diameter (d). where PAId% is the percentage periodic annual increment in diameter,
di+Δt is the current diameter in cm, and di is the initial diameter in cm.
Methods
Model Used to Describe the PAId%
Study Area The allometric model (Eq. 3) was linearized (Eq. 4) and used to describe
The study area is located on the campus of the Federal University the relationship between PAId% as a function of di of open-grown
of Santa Maria (UFSM), in the Central Depression region of the Rio and under competition trees. The model adjustment in the linearized
Grande do Sul state, municipality of Santa Maria, Brazil, between form facilitates the estimation of the regression coefficients, tends to
coordinates 29° 43’ 10”–29° 43’ 21”S and 53° 42’ 42”–53° 42’ 59”W, minimize or eliminate heteroscedasticity, in addition to facilitating the
approximately 95 m above sea level. According to Köppen’s clas- analysis of covariance and, subsequently, the development of a single
sification, the region’s climate is humid subtropical (Cfa), character- regression model with Dummy variables. However, when a nonlinear
ized by hot summers and well-distributed precipitation throughout model is linearized using logarithm properties, it is necessary to cor-
the year, with an average annual precipitation {30 years} of 1720 mm rect the logarithmic discrepancy (Baskerville, 1972), estimating with the
(Alvares et al., 2013). The average annual temperature is 19.1°C, with proper correction in the original scale (Eq. 5):
average temperatures of 32.0°C and 9.0°C during the hot and cold PAId% = b 0 di * e i
months, respectively (Heldwein et al., 2009). The soil in the region is (3)
classified as an Ultisol, and it is present in the sampling area generally ln (PAId% ) = a 0 + a1.ln ( d i ) + e i
in a condition of low natural fertility with a high degree of compaction (4)
 = exp éaˆ + aˆ .ln d + sˆres. ù
(Brun et al., 2012). PAId% ê 0
êë
1 i
2 ûú
( )
ú
(5)
Two contrasting conditions were evaluated: (i) a dense 50-year-old P.
taeda plantation with a current density of approximately 550 trees.ha−1 where PAId% is the percentage periodic annual increment in diameter,
(growing in competition), with basal area of ≈68.0 m2.ha−1, average di is the initial diameter in cm; β0, β1 = are the regression coefficients
height of ≈32.5 m; and (ii) a nearby arboretum (<0.3 km) with scattered in nonlinear form, α0, α1 = are the regression coefficients in linear-
P. taeda trees, under a density of less than 30 trees.ha−1 and a grassy ized form, εi = is the random error, aˆ 0, aˆ1 = are the estimation of the
understory (open-grown). There were no thinning interventions in any equation in linearized form, sˆres. is the model residual variance in
of the stands, the first being characterized by the occurrence of high linearized form.

106
 Schons et al. Increment Patterns of Pinus taeda L. Trees
Forestist 2022: 72(2): 105-111

Table 1.
Dendrometric and PAId Characteristics of Open-Grown and Under Competition Trees of Pinus taeda L.

Tree Condition n d h PAId5 PAId10 PAId15 PAId20

1 Open-grown trees 25 40.1 20.4 0.82 ABC
0.82 AB
0.91 BC
1.04 BCD

2 24 42.0 17.5 0.77 ABC

0.85 AB
1.13 ABC
1,29 BC

3 20 49.7 15.7 1.17 A

1.10 A
1.40 A
1.73 A

4 25 62.0 18.0 1.00 AB

1.11 A
1.30 AB
1.47 AB

5 38 63.3 20.0 0.48 C

0.39 C
0.45 D
0.51 E

6 36 66.0 21.4 0.61 BC

0.63 BC
0.70 CD
0.74 DE

7 36 69.6 20.0 0.88 ABC

0.81 AB
0.91 BC
0.93 CDE

8 34 73.6 18.6 0.72 ABC

0.73 BC
0.92 BC
0.93 CDE

9 36 78.0 20.8 0.97 AB

0.78 AB
0.97 ABC
1.09 BCD

10 35 98.0 35.0 0.90 ABC

0.95 AB
1.10 ABC
1.24 BC

Overall average 31 64.2 20.7 0.83b 0.82b 0.98a 1.10a

1 Under competition trees 46 41.1 – 0.51 BCD
0.47 BC
0.56 BC
0.55 BC

2 41 41.5 30.2 0.61 BC

0.56 B
0.56 BC
0.57 BC

3 48 42.9 – 0.54 BCD

0.52 BC
0.45 C
0.45 C

4 46 45.0 32.0 0.69 B

0.57 B
0.52 BC
0.52 BC

5 46 46.8 – 0.31 D
0.33 C
0.38 C
0.47 C

6 46 47.1 – 0.46 BCD

0.48 BC
0.55 BC
0.54 BC

7 47 48.6 – 0.37 CD
0.35 C
0.47 BC
0.50 BC

8 44 52.5 – 0.56 BC
0.51 BC
0.57 BC
0.57 BC

9 47 55.0 33.4 0.58 BC

0.62 B
0.66 B
0.66 B

10 45 69.4 34.5 1.08 A

0.93 A
0.92 A
0.91 A

Overall average 46 49.0 32.5 0.57a 0.53a 0.56a 0.58a

Note: n = number of tree rings measured at 1.3 m above ground; d = diameter at breast height in cm; h = total height in m; PAId5,..,20 = average periodic annual incre-
ment in diameter considering 5, 10, 15, and 20 years in cm; – = trees whose heights have not been determined. Means followed by the same letter in the column
(uppercase) and row (lowercase) for each group do not differ from each other (α = 5%) by the Tukey’s test.

Analysis of Covariance is the regression coefficient associated with the level for the group of
Analysis of covariance (ANCOVA) was used to verify the exis- open-grown trees, gˆ1éunder competition trees and open-grown treesù is the
ë û
tence or not of difference between intercept (α0) and slope (α1) of a regression coefficient associated with slope for the group of under
group of fitted equations (Milliken & Johnson, 2002). In the pres- competition trees and open-grown trees, sˆres. is the model residual
ent study, this analysis compared the adjusted regression equa- variance in linearized form, definition of Dummy variables: if D1 = 0, the
tions (Eq. 4) for the group of open-grown trees with those of under equation is for under competition trees, and if D1 = 1, the equation is for
competition trees. open-grown trees.

Model With Dummy Variables Used to Describe the PAId% Goodness-of-Fit Statistics
The estimation of the PAId% model using Dummy variables, with cor- The goodness-of-fit of the generated equations was evaluated using
rection for the logarithmic discrepancy, used to distinguish the group the coefficient of determination (R2 – Eq. 7), standard error of the esti-
of open-grown trees and under competition trees is defined by the fol- mate (Syx – Eq. 8), and graphical residual analysis (Eq. 9), according to
lowing expression (Eq. 6): the following expressions:
 = exp égˆ é
å ( yi - y˘ i )2 ùú
n
PAId% ê 0 éunder competition treesù + gˆ0 éopen-grown treesù ´ D1 ê
ë ë û ë û 2
R = ê1- i =1
ú
å
n
sˆ ù ê ( yi - y )2 úû
+gˆ é
1 under competition trees and open-grown trees ù
ë û
( )
.lln di + res ú
2 úû
ë i =1 (7)

å
n
(6)
i =1
( yi - y˘ i )2
 Syx =
where PAId% is the estimation of the percentage periodic n- p
(8)
annual increment in diameter, di is the initial diameter in cm;
gˆ é Residuals = yi - y̆i
0 under competition trees ù is the regression coefficient associated with the (9)
ë û
intercept for the group of under competition trees, g 0 éopen-grown treesù
ˆ
ë û

107
 Schons et al. Increment Patterns of Pinus taeda L. Trees
Forestist 2022: 72(2): 105-111

Table 2.
Adjustment and Precision Statistics of the Regression Models to Describe the PAId ̂% of Open-Growth and Under Competition Pinus taeda L. Trees

ă 0 ă1
Condition F value Pr > F R2 Syx (%)
Open-grown trees 7.1507 −1.6159 750.5 <0.0001 0.70 3.2
(<0.0001) (<0.0001)
Under competition trees 6.3913 −1.6267 1026.7 <0.0001 0.72 1.9
(<0.0001) (<0.0001)
Note: PAId% = percentage periodic annual increment in diameter.

where y1 is the observed value, y̆i is the predicted value, y is the aver- between the groups, with values of 0.39–1.47 cm.year−1 in open-grown
age of observed value, n is the number of observations, and p is the trees and 0.31–1.08 cm.year−1 and in under competition trees.
number of coefficients in the model.
Analyzing the overall averages of the PAId of open-grown trees, no
Statistical Analysis statistical differences were found between the periods of 5 (0.83b
Analyses were performed using the Statistical Analysis System (SAS cm.year−1) and 10 years (0.82b cm.year−1). Meanwhile, the increments
Institute, 2004). The ANOVA procedure was used to compare the PAId for the 15 (0.98a cm.ano−1) and 20 years (1.10a cm.ano−1) have statis-
considering 5, 10, 15, and 20 years in the open-growth trees and under tical difference from the younger periods (Table 1). With increasing
competition trees. After applying the ANOVA, when significance was age, the diameter growth rate of open-grown trees tends to decrease
found using the F test, the means were compared using the Tukey’s (≈0.27 cm.year−1—at age 15 years).
test (α = 5%).
For under competition trees, the overall average PAId did not differ sta-
The GLM procedure (PROC GLM) was used to perform ANCOVA. This tistically among the periods of 5 (0.57a cm.year−1), 10 (0.53a cm.year−1),
procedure uses the method of least squares to fit general linear models. 15 (0.56a cm.year−1), and 20 (0.58a cm.year−1) years, with a trend of stag-
Meanwhile, the REG procedure (PROC REG), which is a general-purpose nation in the growth rate in diameter (Table 1).
procedure for regression, allowed the adjustment of the individual
models of PAId% in linearized form and with Dummy variables. Adjusted Equations to Describe the PAId%
The equations adjusted for the group of open-grown trees and under
Finally, the model with Dummy variables was investigated for the collinear- competition trees showed coefficients of intercept ( ă 0 ) and slope
ity of the coefficients, considering the variance inflation value (VIF) lower ( ă1 ) significant (Pr > |t| = <0.0001) (Table 2). Competition trees have
than 5 as an indication of the absence of a severe collinearity effect. increments with less variation (Figure 1a) and equation with greater fit
(R2 = 0.72) and precision (Syx = 1.9%) when compared to open-grown
Results trees statistics (R2 = 0.70; Syx = 3.2%). The residual variance (s˘ res. ) used
to the correction of the logarithmic discrepancy was 0.2355 for open-
Characterization of Sampled Trees grown trees and 0.1698 for under competition trees.
On average, 31 and 46 tree rings of open-grown and under competi-
tion trees were measured, respectively (Table 1). The averages of the The equations fitted for each group with the correction of the logarith-
PAId, per tree, at 5, 10, 15, and 20 years, presented different amplitudes mic discrepancy are expressed by:

Figure 1.
Pattern of PAId% of Open-Grown and Under Competition Trees of Pinus taeda L: (a) Observed Data and (b) Trend of Predicted Equations.

108
 Schons et al. Increment Patterns of Pinus taeda L. Trees
Forestist 2022: 72(2): 105-111

The trend of observed data and estimated equations of PAId% for

both groups presented in Figure 1a and b, showed a difference of
≈20% between the estimated values of PAId% at a di equal to 10 cm
and still have 0.85% difference at a di of 70 cm. Additionally, the resid-
ual graphical analysis (Figure 2) shows a good residual dispersion,
with low error amplitude and maintaining a proportionality between
the under and overestimates, especially for the estimates under com-
petition trees.

Analysis of Covariance
Significant differences were found in the intercept of the equations (α0)
− (Pr > F = 0.0066) between the groups of open-grown and under com-
petition trees (Table 3). The same was not true for the slope (α1) − (Pr >
F = 0.8883). Therefore, a single regression equation, with the same slope,
can be used for both groups.

Equation With Dummy Variable Fitted to Describe the PAId%

The equation fitted with the Dummy variable showed all the estimated
coefficients to be significant (Pr > |t| = <0.0001) (Table 4). The statistics
Figure 2. calculated for the original scale of the equation showed adjustment
Residual Graphical Analysis for the Model Evaluated for the PAId% (R2 = 0.69) and precision (Syx = 2.4%) with no severe collinearity of its
Estimation. coefficients (VIF < 5).


The fitted equation with Dummy variables allows estimating the PAId%
 = exp é7.1507 - 1.6159.ln d + 0.2355 ù
Open-grown trees : PAId% ê
ë
i
2 û
ú ( ) for both tree groups simultaneously, and with the necessary correction
of the logarithmic discrepancy, expressed by:

 = exp é6.3913 - 1.6267.ln d + 0.1698 ù .  = exp é6.3725 + 0.7982 ´ D1 - 1.6212.ln d + 0.1960 ù .

Under competition trees : PAId% ê
ë
i
2 û
ú( ) PAId% ê i ( )
2 û
ú
ë

Table 3.
ANCOVA to Evaluate the Effect on Intercept (α0) and Slope (α1) of the Fitted Equations to Describe the PAId% of Open-grown and Under Competition Trees of Pinus
taeda L

Variable Condition Intercept (α0) − Pr > F Slope (α1) − Pr > F

ln(PAId%) Open-grown trees versus under competition trees 0.0066 0.8883ns
Note: ANCOVA = analysis of covariance; PAId% = percentage periodic annual increment in diameter; ns = not significant.

Table 4.
 of Open-Grown and Under Competition Trees of
ANOVA and Adjustment and Precision Statistics of the Equation With Dummy Variables to Describe the PAId%
Pinus taeda L

SV DF SS MS F value Pr > F
Model 2 366.96 183.48 935.9 <0.0001
Residual 762 149.39 0.1960
Total 764 516.35
Coefficients Estimated Pr > |t|
6.3725 <0.0001
gˆ
0 éunder competition trees ù
ë û

0.7982 <0.0001
gˆ ´ D1
0 éopen-grown trees ù
ë û

−1.6212 <0.0001
gˆ é
1 under competition trees and open-grown trees ù
ë û
( )
.ln dt

R2 = 0.69
Syx = 2.4%
Note: ANOVA = analysis of variance; PAId% = percentage periodic annual increment in diameter. SV = sources of variance; DF = degrees of freedom; SS = sum of
squares; MS = mean square.

109
 Schons et al. Increment Patterns of Pinus taeda L. Trees
Forestist 2022: 72(2): 105-111

Discussion trees tend to reduce growth in the same way as under competition
trees, changing only the level of the equations. In this sense, manage-
Competition between trees is one of the strongest determinants of ment planning and interventions in forests are important to control
tree size and stem increment. In this sense, open-grown trees repre- competition and accelerate growth, in the search for avoiding losses,
sent the empirical maximum tree dimensions, providing indications on which can gain even greater relevance in a context of the increasing
the space “expected” for a given species to grow without interference demand for pine wood for solid end-uses, requiring large stems from
from other trees, that is, without harming its growth. Applications of longer rotations (Topanotti et al., 2021). Competition, from this angle,
studying trees in this type of condition extend from the development can be anticipated, postponed, or eliminated in intensely managed
of management guidelines (such as thinning planning) to competition rotations (Ferreira et al., 2020). In addition, this identification of com-
modeling and crown closure (Cavalli & Finger, 2017; Costa et al., 2018; mon slope between curves and differences only in level also makes it
Grote et al., 2020; Hasenauer, 1997; Li et al., 2008; Salih et al., 2019). easy to change the intercept of the equations to different other levels of
Furthermore, in the context of growth modeling, open-grown trees are competition, for example, intermediate to those studied in the present
used to describe potential growth, allowing to examine the behavior of study, thus contributing to species growth estimates and aiding forest
trees without competition and then compare changes in vigor due to management activities.
competition (Weiskittel et al., 2011).
The use of Dummy variables allowed dividing the PAId% estimates into
P. taeda is a fast-growing conifer species that presents different charac- the two groups of trees (open-grown and under competition), with a
teristics in response to competition. Differences of up to ≈20% were tendency to decrease the increment rates as the di increases. This char-
observed in the present study for the species when comparing the acteristic reveals a mature phase of the trees, where the juvenile phase
increments of open-grown and under competition trees. This growth of accelerated growth was not observed in this study, as it probably
rate response is clearly associated with differences in the external mor- occurred before 10 cm in diameter (Weiskittel et al., 2011).
phology of the trees, reflecting on the conformations of the growth
rings (Tomazello Filho et al., 2017). In general, open-grown trees have The relationship modeled in this study provided predicting the future
larger diameters and less investment in height, while under competi- increment of the species based on the current diameter, with a view
tion trees generally express smaller diametric dimensions and greater to practical application, and clearly showed the influence of the
heights, in order to overcome the competition imposed by surrounding growth space, without considering, however, other variables, such
trees and achieve greater light availability, as characterized in Table 1. In as crown diameter, tree height, age, and site (Schröder et al., 2002;
addition, the larger crown dimensions expected for open-grown trees Zimmermann et al., 2016). These additional variables should be the
also have a substantial influence on these reported growth differences, focus of further studies in order to reinforce these findings and contrib-
reflecting the greater availability of resources. ute to refining the assessments and predictions of the species, associ-
ated with the planning of forest management actions.
Investigating the growth of trees by using the allometric approach
avoids time-consuming experiments, and it also delivers valuable infor-
mation when long time-series on individual tree growth is not avail- Conclusion
able (Hein & Spiecker, 2008). In the present study, we found a strongly
significant relationship (R2 ≥ 0.70) for PAI% model as a function of the Growth differences of up to ≈20% were observed in the present study
initial diameter (di) in both conditions studied, highlighting the possi- for the species when comparing open-grown and competing trees.
bility of an accurate prediction of the increment based on the initial
diametric dimension of the tree. Slightly higher estimates for the PAId% The groups of open-grown and under competition trees differ from
equations were observed for the condition of under competition trees each other only in intercept for the fitted equations, and a single regres-
(Table 2 and Figure 1), probably as a consequence of the greater homo- sion model with a Dummy variable is proposed to describe the variation
geneity observed in this condition (less variation in data), characteristic of the percentage periodic annual increment (PAId%) as a function of
of planting trees with the same age, uniform canopy and little varia- the diameter.
tion in diameter and height. However, the high goodness-of-fit also
observed for open-grown trees in conditions of greater heterogeneity The results of this study were important for the clear characterization
reinforces the ability of the generated models. of the influence of the growth space on the species increment, as well
as for the investigation of its maximum potential. In addition, further
The significant difference in intercept found in the ANCOVA between studies will help to reinforce these findings, especially through the
open-grown and under competition trees groups especially encompasses inclusion of additional variables, which may contribute to refining the
differences in PAId% at di close to 10 cm. This represents that the density predictions, therefore contributing to the monitoring and execution of
and greater space for growth interfere with the initial growth of trees. In silvicultural activities of individual P. taeda trees.
this sense, planting spacing is a crucial decision to avoid early competi-
tion (Akers et al., 2013; Aspinwall et al., 2011). This also corroborates with Peer-review: Externally peer- reviewed.
the findings of Dobner et al. (2019), who in a study comparing different
levels of crown thinning in P. taeda stands, observed significant differences Author Contributions: Concept – C.T.S., E.A.C.; Design – C.T.S., E.A.C.; Supervision –
in tree rings width from 6 years of age onward between trees with and C.A.G.F.; Materials – C.T.S.; Data Collection and/or Processing – C.T.S., E.A.C.;
Analysis and/or Interpretation – C.T.S., E.A.C., L.O.B.; Literature Search – L.O.B.;
without thinning, in a clear association with the available growing space.
Writing Manuscript – C.T.S., E.A.C., L.O.B.; Critical Review – C.A.G.F.

Covariance analysis allows us to assess the need or not to use differ-

Declaration of Interests: The authors have no conflicts of interest to declare.
ent equations for sites, areas, diameter classes, ages, among others.
The identification of non-significant differences between the slopes Funding: The authors declared that this study has received no financial
of the adjusted equations demonstrates that, over time, open-grown support.

110
 Schons et al. Increment Patterns of Pinus taeda L. Trees
Forestist 2022: 72(2): 105-111

References • Hasenauer, H. (1997). Dimensional relationships of open-grown trees in

Austria. Forest Ecology and Management, 96(3), 197–206. [CrossRef]
• Akers, M. K., Kane, M., Zhao, D., Teskey, R. O., & Daniels, R. F. (2013). Effects • Hasenauer, H. (2000). Princípios para a modelagem de ecossistemas flo-
of planting density and cultural intensity on stand and crown attributes of restais. Ciência and Ambiente, 20, 53–69.
mid-rotation loblolly pine plantations. Forest Ecology and Management, 310, • Hein, S., & Spiecker, H. (2008). Crown and tree allometry of open-grown ash
468–475. [CrossRef] (Fraxinus excelsior L.) and sycamore (Acer pseudoplatanus L.). Agroforestry
• Albaugh, T. J., Fox, T. R., Maier, C. A., Campoe, O. C., Rubilar, R. A., Cook, R. L., Systems, 73(3), 205–218. [CrossRef]
Raymond, J. E., Alvares, C. A., & Stape, J. L. (2018). A common garden experi- • Heldwein, A. B., Buriol, G. A., & Streck, N. A. (2009). O clima de Santa Maria.
ment examining light use efficiency and heat sum to explain growth dif- Ciência and Ambiente, 38, 43–58.
ferences in native and exotic Pinus taeda. Forest Ecology and Management, • Ibá – Indústria Brasileira de Árvores. (2020). Relatório anual. Annual Report.
425, 35–44. [CrossRef] 66 p.
• Alvares, C. A., Stape, J. L., Sentelhas, P. C., de Moraes Gonçalves, J. L. M., & • Johnson, J. E., & Smith, D. W. (2009). Principles of regeneration silviculture in
Sparovek, G. (2013). Köppen’s climate classification map for Brazil. Meteor- Virginia. Virginia Cooperative Extension, 20 p.
ologische Zeitschrift, 22(6), 711–728. [CrossRef] • José Brun, E., Ferreira da Rosa, S., Roppa, C., Valdir Schumacher, M., & Gizele
• Andreacci, F., Botosso, P. C., & Galvão, F. (2014). Sinais climáticos em anéis König Brun, F. (2012). Avaliação nutricional de espécies nativas utilizadas na
de crescimento de Cedrela fissilis em diferentes tipologias de Florestas arborização do campus da Universidade Federal de Santa Maria-RS. Revista
Ombrófilas do Sul do Brasil. FLORESTA, 44(2), 323–332. [CrossRef] da Sociedade Brasileira de Arborização Urbana, 7(1), 89–111. [CrossRef]
• Aspinwall, M. J., King, J. S., Mckeand, S. E., & Bullock, B. P. (2011). Genetic • Kuehne, C., Weiskittel, A. R., & Waskiewicz, J. (2019). Comparing perfor-
effects on stand-level uniformity and above- and belowground dry mass mance of contrasting distance-independent and distance-dependent
production in juvenile loblolly pine. Forest Ecology and Management, 262(4), competition metrics in predicting individual tree diameter increment and
609–619. [CrossRef] survival within structurally-heterogeneous, mixed-species forests of North-
• Baskerville, G. L. (1972). Use of logarithmic regression in the estimation of eastern United States. Forest Ecology and Management, 433, 205–216.
plant biomass. Canadian Journal of Forest Research, 2(1), 49–53. [CrossRef] [CrossRef]
• Bognola, I. A., Ribeiro Junior, P. J., Silva, E. A. Ad, Lingnau, C., & Higa, A. R. • Lhotka, J. M. (2017). Examining growth relationships in Quercus stands: An
(2008). Modelagem uni e bivariada da variabilidade espacial de rendimento application of individual-tree models developed from long-term thinning
de Pinus taeda L. Floresta, 38(2), 373–385. [CrossRef] experiments. Forest Ecology and Management, 385(1), 65–77. [CrossRef]
• Cavalli, J. P., & Finger, C. A. G. (2017). Modelling of upper crown exposed to • Li, F., Chen, D., & Lu, J. (2008). Individual-based competition analysis for
light of Cedrela fissilis (Vell.) open-grown trees by a non-destructive method. secondary forest in northeast China. Journal of Korean Forestry Society, 97(5),
Forestry, 90(2), 312–317. [CrossRef] 501–507.
• Costa, E. A., Finger, C. A. G., & Fleig, F. D. (2018). Modelagem do espaço de • Milliken, G. A., & Johnson, F. E. (2002). Analysis of messy data volume III: Analy-
crescimento para araucária. Ciência Florestal, 28(2), 725–734. [CrossRef] sis of covariance. Boca Raton: Chapman & Hall/CRC.
• Costa, E. A., Finger, C. A. G., & Hess, A. F. (2015). Modelo de incremento em • Miranda, R. O. V. (2016). Modelagem de árvores individuais para povoamentos
área basal para árvores de araucária de uma floresta inequiânea. Pesquisa não desbastados de Pinus taeda L. (Thesis, Universidade Federal do Paraná).
Florestal Brasileira, 35(83), 239–245. [CrossRef] Retrieved from https​://ac​ervod​igita​l.ufp​r.br/​handl​e/188​4/439​57.
• Dobner Jr., M., Nicoletti, M. F., & Arce, J. E. (2019). Influence of crown thinning • Salih, T. K., Saied, M. Y., & Wali, S. T. (2019). Determination of competition
on radial growth pattern of Pinus taeda in southern Brazil. New Forests, 50(3), factor for tree growth within a stand. Journal of the University of Duhok, 22(2),
437–454. [CrossRef] 177–188. [CrossRef]
• Eisfeld, R. L., Sanquetta, C. R., Arce, J. E., & Mello, A. A. (2005). Modelagem • Sas Institute Incorp. (2004). Statistical analysis system user’s guide. Version 9.1.
do crescimento e da produção de Pinus taeda L. por meio do processo de Cary: Statistical Analysis System Institute.
difusão. CERNE, 11(2), 167–177. • Tomazello Filho, M., Latorraca, J. VdF., Fischer, F. M., Muñiz, G. I. Bd, Melandri,
• Elesbão, L. E. G., & Schneider, P. R. (2011). Produção de Pinus taeda L. em J. L., Stasiak, P. M., Torres, M. A., Piccion, W. J., Hoffmann, H. A., & Silva, L. D.
povoamento desbastado na região dos Campos de Cima da Serra, Rio (2017). Avaliação da dispersão de sementes de Pinus taeda L. pela análise
Grande do Sul. Ciência Florestal, 21(1), 119–124. [CrossRef] dos anéis de crescimento de árvores de regeneração natural. Floresta e
• Ferreira, G. W. D., Rau, B. M., & Aubrey, D. P. (2020). Herbicide, fertilization, Ambiente, 24, 1–11, 24. [CrossRef]
and planting density effects on intensively managed loblolly pine early • Topanotti, L. R., Vaz, D. R., Carvalho, SdP. Ce, Rios, P. D., Tomazello-Filho, M.,
stand development. Forest Ecology and Management, 472, 118206. Dobner Jr., M., & Nicoletti, M. F. (2021). Growth and wood density of
[CrossRef] Pinus taeda L. as affected by shelterwood harvest in a two-aged stand in
• Gomes, F. S., Maestri, R., & Sanquetta, C. R. (1997). Avaliação da produção Southern Brazil. European Journal of Forest Research, 140(4), 869–881.
em volume total e sortimento em povoamentos de Pinus taeda L. submeti- [CrossRef]
dos a diferentes condições de espaçamento inicial e sítio. Ciência Florestal, • Weiskittel, A. R., Hann, D. W., Kershaw Jr., J. A., & Vanclay, J. K. (2011). Forest
7(1), 101–126. [CrossRef] growth and yield modeling. Oxford: John Wiley & Sons.
• Grote, R., Kraus, D., Weis, W., Ettl, R., & Göttlein, A. (2020). Dynamic coupling • Zimmermann, A. P. L., Costa, E. A., Schröder, T., & Fleig, F. D. (2016). Mode-
of allometric ratios to a process-based forestgrowth model for estimating lagem do incremento diamétrico de Pinus taeda em função de variáveis da
the impacts of stand density changes. Forestry, 93(5), 601–615. [CrossRef] copa e índices de competição. Floresta, 46(1), 115–122. [CrossRef]

111