Content uploaded by Maria Luiza Franceschi Nicodemo
Author content
All content in this area was uploaded by Maria Luiza Franceschi Nicodemo on May 16, 2016
Content may be subject to copyright.
279
Revista Árvore, Viçosa-MG, v.40, n.2, p.279-288, 2016
Allometric models for estimating aboveground...
ALLOMETRIC MODELS FOR ESTIMATING ABOVEGROUND BIOMASS AND
BIOMASS ALLOCATION OF CAPIXINGUI TREES (Croton floribundus Spreng.)
IN AN AGRISILVICULTURAL SYSTEM1
Maria Luiza Franceschi Nicodemo
2*
, Marcelo Dias Muller
3
, Antônio Aparecido Carpanezzi
4
and Vanderley
Porfírio-da-Silva
4
1
Received on 10.07.2014 accepted for publication on 16.12.2015.
2
Empresa Brasileira de Pesquisa Agropecuária, Centro de Pesquisa de Pecuária do Sudeste, São Carlos, SP - Brasil. E-mail:
<marialuiza.nicodemo@embrapa.br>.
3
Empresa Brasileira de Pesquisa Agropecuária, Centro Nacional de Pesquisa de Gado de Leite, Coronel Pacheco, MG - Brasil.
E-mail: <marcelo.muller@embrapa.br>.
4
Empresa Brasileira de Pesquisa Agropecuária - CNPF, Embrapa Florestas, Colombo, PR - Brasil. E-mail: <antonio.carpanezzi@embrapa.br>.
and <vanderley.porfirio@embrapa.br>
*
Corresponding author.
ht tp: //dx . doi .org/10. 1590 / 0100- 6762 2 0160 00200 010
ABSTRACT – The objective of this study was to select allometric models to estimate total and pooled aboveground
biomass of 4.5-year-old capixingui trees established in an agrisilvicultural system. Aboveground biomass distribution
of capixingui was also evaluated. Single- (diameter at breast height [DBH] or crown diameter or stem diameter
as the independent variable) and double-entry (DBH or crown diameter or stem diameter and total height
as independent variables) models were studied. The estimated total biomass was 17.3 t.ha-1, corresponding
to 86.6 kg per tree. All models showed a good fit to the data (R
2
ad
> 0.85) for bole, branches, and total biomass.
DBH-based models presented the best residual distribution. Model lnW = b0 + b1*lnDBH can be recommended
for aboveground biomass estimation. Lower coefficients were obtained for leaves (R
2
ad
> 82%). Biomass distribution
followed the order: bole>branches>leaves. Bole biomass percentage decreased with increasing DBH of the trees,
whereas branch biomass increased.
Keywords: Agroforestry systems; Integrated production systems; Modeling.
MODELOS ALOMÉTRICOS PARA ESTIMAÇÃO DE BIOMASSA AÉREA E
ALOCAÇÃO DE BIOMASSA DE CAPIXINGUI (Croton floribundus Spreng.) EM
UM SISTEMA SILVIAGRÍCOLA
RESUMO – O objetivo deste estudo foi selecionar equações alométricas para predição da biomassa aérea
total e nos compartimentos aéreos de capixingui com 4,5 anos estabelecidas em um sistema silviagrícola.
Foram tomadas medidas dendrométricas e utilizou-se o método destrutivo para a estimação da biomassa,
com a separação dos componentes em tronco, galhos e folhas. Foram ajustados e comparados quatro modelos
alométricos, sendo dois de simples entrada e dois de dupla entrada. Como variáveis preditoras foram considerados
o diâmetro à altura do peito (DAP), diâmetro de colo (D colo), diâmetro de copa (Dc) e altura total (H).
A biomassa total estimada foi de 17,3 t.ha-1, equivalente a 86,6 kg por árvore. Todos os modelos apresentaram
ajustes satisfatórios (R
2
aj
>0,85) para tronco, galhos e biomassa total. As equações para biomassa total que
tiveram como variável preditora o DAP apresentaram as melhores distribuições de resíduos. O modelo
lnW=b0+b1*lnDAP pode ser recomendado para estimativa de biomassa aérea. Para folhas, os coeficientes
foram menores (R
2
aj
>82%). A distribuição da biomassa nos componentes seguiu a ordem: tronco>galhos>folhas.
A percentagem de biomassa no tronco decresceu com o aumento do DAP das árvores enquanto que a biomassa
nos galhos aumentou.
Palavras-chave: Modelagem; Sistemas agroflorestais; Sistemas de integração da produção.
280
Revista Árvore, Viçosa-MG, v.40, n.2, p.279- 288, 2016
NICODEMO, M.L.F. et al.
1. INTRODUCTION
The selection of technologies for production systems
should consider ways to reduce the impacts on water
use, soil and biota and should be directed toward the
regeneration of ecosystem services which are associated
with increased biodiversity (TILMAN et al., 2011;
BALMFORD et al., 2012). Agroforestry systems based
on reconciling production and environmental conservation
have been encouraged by public policies such as the
Low-Carbon Agriculture Program (Agricultura de Baixa
Emissão de Carbono - ABC). Agrisilvicultural system
is among those, consisting of the combination of crops
and woody species (NAIR, 1985).
Eucalyptus is the most common tree species used
in agroforestry systems in Brazil (VENTURIN et al., 2010).
In an attempt to diversify the forest base by introducing
native species (RUSCHEL et al., 2003; CARPANEZZI
et al., 2010), other species have been tested in these
systems (NICODEMO et al., 2009; MELLOTO et al., 2009).
One such species is capixingui (
Croton floribundus
Spreng.), a pioneer tree species found mainly in seasonal
semideciduous forests in the southeastern region of
Brazil and in northern Paraná. This plant has some
interesting characteristic for agroforestry systems, such
as simple seedling production, high survival rates, rusticity,
and moderate to rapid growth. In addition, it is a melliferous
species which is of high value to the beekeeping industry
and its wood can be used for construction, cellulose
production, for medicinal purposes (CARVALHO, 2003;
GOUVEIA et al., 2007) and more recently for the production
of briquettes.
Agroforestry systems are dynamic and require
interventions such as thinning and pruning to regulate
the competition between its components (JOSE et al.,
2004). Therefore, estimating aboveground biomass from
simple measures is interesting for the producer since
it allows to predict the amount of timber forest products
to be obtained with the intervention. Forest biomass
can be evaluated by felled trees or by estimating biomass
using allometric models. Direct determination is labor
intensive, complex, time consuming, and expensive.
Allometric models are developed by fitting regression
equations (FEREZ, 2010). These models have been
developed for mixed forest stands (BROWN, 1997;
TIEPOLO et al., 2002; CHAVE et al., 2005; FEREZ, 2010)
and for individual species (KUMAR et al., 1998; KAONGA
and BAYLISS-SMITH, 2010), and are routinely used.
Technical projects involving agrisilvicultural system
include the estimation of the development of tree
components and expected revenue generation. Biomass
estimates are used to calculate the amount of carbon
captured by the system and to predict the amount of
potentially available forest product. Few allometric
equations have been developed for native Brazilian
tree species (SALIS et al., 2004; BARBEIRO et al., 2009;
PADILHA, 2011) and we found no model for native
tree species grown in agrisilvicultural system. The
objective of the present study was to select allometric
equations for estimating total aboveground biomass
of 4.5-year-old capixingui species established in an
agrisilvicultural system, as well as to evaluate the
aboveground biomass distribution in different
compartments.
2. MATERIAL AND METHODS
The study was conducted in the municipality
of São Carlos, SP (latitude 22°1’ South and longitude
47°53’ West). The climate of the region is classified
as Cwa-Awa (KÖPPEN), with the dry season from April
to September. The average annual temperature is 21.2
°C, average annual relative humidity is 75.6%, and
average annual rainfall is 1362 mm. The topography
is smooth, with slopes of 3% to 5% and a mean altitude
of 860 m. The area was formed by
Urochloa decumbens
on Red-Yellow Latosol. The agroforestry system was
implemented in January 2008, with five rows of trees
interspersed with annual crops. Each row was formed
by three lines of trees along the terrain level. The
trees were planted at a distance of 2.5 × 2.5 m, resulting
in 600 trees/ha. Annual crops were planted in the 17-
m inter-spaces between the bands of trees. The forest
species planted at random in the central line were:
angico (
Anadenanthera colubrina
), canafístula
(
Peltoph oru m dub ium
), ipê-felpudo (
Ze yheria
tuberculosa
), jequitibá-branco (
Cariniana estrellensis
),
and pau-jacaré (
Piptadenia gonoacantha
). To obtain
higher plant stems,
mutambo (
Guazuma ulmifolia
)
and capixingui were planted alternately in the marginal
lines, for a total of 200 trees/ha per species. The trees
were fertilized for the first 18 months. Silvicultural
practices included the combat of leaf-cutting ants
and herbaceous plant control. The crops planted in
the bands between rows were managed conventionally
and in these areas soil corrections were supported
by annual soil analysis.
281
Revista Árvore, Viçosa-MG, v.40, n.2, p.279-288, 2016
Allometric models for estimating aboveground...
Capixingui and mutambo were pruned in November
2010, 35 months after planting the trees. The pruning
criteria were the conservation of at least 50% of the
green crown, removing the branches below the point
where the stem measured 6 to 8 cm in diameter. The
crops planted after pruning showed low performance
and thinning was recommended to control competition.
Thinning was performed between June 13 and July
3, 2012. Because capixingui has a lower crown insertion
height and lower apical dominance, impairing mechanical
treatment of the plantation area, it was chosen for thinning.
The mean wood basic density of capixingui, which
was calculated by dividing dry weight by green volume,
was 0.36 g cm
-3
on the occasion of tree felling (Lima,
I.L., personal communication, May 7, 2013).
Plots containing 79 of the 802 existing capixingui
trees (10%) were delimited for measurements. The diameter
at breast height (DBH) and total height were determined
with a measuring tape (DBH = circumference/ð) and
a telescopic pole in May 2012. The DBH of trees with
multiple, bifurcated boles or crown formation below
a height of 1.30 m was calculated as follows: the basal
area for each bole was calculated and the sum of basal
areas was related to a cumulative value of DBH.
The capixingui trees were divided into four diameter
size classes to account for variations in the stand.
Next, three trees were selected per diameter size class
for biomass estimation. The following measurements
were obtained from each selected tree: stem diameter
(10-15 cm from the ground), DBH, crown diameter
obtained as the mean of orthogonal measurements taken
in the N-S and W-E directions, and total height.
The trees were cut with a chainsaw at 10 cm from
the ground onto a plastic canvas. First, leaves and
branches were removed. The branches were divided
into three diameter size classes: thin (< 3 cm), medium
(3 to 15 cm), and thick (> 15 cm). The leaves and branches
were weighed on a digital scale in the field immediately
after felling. Three subsamples of leaves and branches
were obtained from each size class to calculate the
dry matter. The bole was divided into sections of 1.5
m, numbered on the lower part of the tree. If there were
bifurcations, shorter segments were taken. The weight
of the segments was measured in the field with a digital
scale. Two subsamples were obtained from each segment
to calculate the dry weight. Thus, the subsamples were
oven-dried at 60°C until a constant weight was obtained.
Total biomass of the tree compartments was calculated
as the percentage of dry weight of the samples using
the following formula: dry weight = (dry weight/green
weight) * 100.
In addition, the data of the 79 trees were used
to develop a hypsometric equation that related height
to DBH. This equation reduces the time necessary to
characterize the capixingui production systems, provided
the conditions for which the equation was generated
are respected. The following model was used: lnHt
=
0 +
1*(1/DBH), where
Ht
= estimated total height,
0
and
1
= coefficients of the equation, and DBH
= diameter at breast height.
To estimate total biomass, four allometric models
were fit to the diameter data (DBH, stem diameter, crown
diameter), total height and biomass, including two double-
entry (two independent variables) and two single-entry
(one independent variable) models, according to: (1)
lnP = b0+b1*lnD+b2*lnH (Schumacher and Hall); (2)
lnP = b0+b1*lnD+b2*D (Brenac); (3) lnP = b0+b1*ln(D²Ht)
(Spurr); and (4) lnP = b0+b1*lnD (Koperzky and
Gehrhardt), where: P = biomass (t.ha
-1
); D: diameter
(diameter at breast height, cm; stem diameter, cm; crown
diameter, m); H: total height (m). Source: Soares et
al. (2007). In the case of the allometric equations used
to estimate the biomass of tree components, only DBH
was computed for the diameter data, using the same
models.
The models were selected based on the adjusted
coefficient of determination (R²
ad
), standard error of
the estimate (S
yx
), and graphic distribution of residuals.
The coefficient of determination expresses the total
variation in the data explained by the model, the standard
error indicates sampling-induced errors, and graphical
analysis of the residuals shows trends of biomass
underestimation or overestimation by the model.
3. RESULTS
The model of hypsometric relationship, lnHt =
0
+
1*(1/DBH), used to estimate total tree height, provided
an R
2
ad
= 0.72 and S
xy
= 0.094, with coefficients
0 =
2.4732 and
1 = -5.0503. Figure 1 illustrates the relationship
between height and diameter.
In general, the models tested for estimating total
aboveground biomass of the trees showed a good fit
to the data (R
2
ad
> 0.92; Table 1), which suggests they
are effective predictors of aboveground biomass.
282
Revista Árvore, Viçosa-MG, v.40, n.2, p.279- 288, 2016
NICODEMO, M.L.F. et al.
Figure 2 shows the residuals of all equations adjusted
as a function of the different variables studied. The
graphs confirm the better performance of the equations
adjusted as a function of DBH.
The models used for estimating biomass of tree
components showed a good fit to the data (R
2
ad
> 0.82;
Table 2). Model fit was better for total biomass, bole
and branches biomass than foliage biomass.
Biomass distribution in the different tr ee
components followed the order: bole>branches>leaves
(Figure 3).
4. DISCUSSION
4.1. Allometric equations for total biomass
The model of hypsometric relationship, lnHt =
0
+
1*(1/DBH), used to estimate total tree height, showed
good fit (Figure 1). In the study of Soares et al. (2011),
this model also showed the best fit for nine of 11 native
species from the city of Viçosa, Minas Gerais, with
R
2
ad
values ranging from 0.70 to 0.92 between species.
The possibility of using DBH to estimate tree height
is a time-saving approach for the characterization of
capixingui since height measurements are more labor
intensive and time-consuming than the measurement
of DBH.
The mean height and DBH were 8.06 m and 17.55
cm, respectively. The frequency distribution obtained
for the capixingui stand was unimodal, with 85% of
the trees in the DBH class of 14 to 27.9 cm; 15% of
the individuals did not exceed a DBH of 13.9 cm. This
distribution is typical of even-aged planted forests
and reflects processes of competition and genetic
Table 1 –
Statistical parameters of the allometric models used for estimating biomass of 4.5-year-old capixingui established
in an agrisilvicultural system as a function of diameter at breast height (DBH, cm), stem diameter (Ds, cm),
crown diameter (Dc, cm), and total height (H, m).
Tabela 1 –
Parâmetros estatísticos da modelagem de equações alométricas para estimação de biomassa de capixingui
estabelecidas em um sistema silviagrícola, aos 4,5 anos de idade, em função do DAP – diâmetro à altura do
peito (cm), D colo - diâm etro do colo (cm ), Dc - diâmetro de copa (cm) e H – altura total (m).
Equation Coefficients R
2
ad
S
xy
0
1
2
Predictor variables: DBH and H
1 -2.721 8 2.617 2 -0.2518 0.9850 0.2098
2 -2.765 2 2.395 3 0.0076 0.9850 0.2099
3 -3.356 1 0.966 6 0.9801 0.2187
4 -2.8564 2.4771 0.9834 0.1999
Predictor var iables: Ds and H
1 -4.862 7 1.241 9 2.5268 0.9587 0.3484
2 -6.773 9 4.483 0 -0.1142 0.9507 0.3805
3 -4.643 7 1.093 3 0.9423 0.3724
4 -4.346 0 2.861 7 0.9168 0.4473
Predictor variables: Dc and H
1 -2.666 2 1.619 9 2.0887 0.9800 0.2427
2 -0.364 8 4.554 7 -0.4781 0.9710 0.2921
3 -1.599 5 1.129 8 0.9706 0.2658
4 -0.269 7 2.960 6 0.9436 0.3685
Where: R
2
ad
= adjusted coefficient of determination; S
xy
= standard error of the estimate.
Figure 1 –
Relationship between height and diameter at breast
height (DBH) of 4.5-year-old capixingui established
in an agrisilvicultural system.
Figura 1 –
Relação entre altura e diâmetro plantas de capixingui
em um sistema silviagrícola de 4,5 anos de idade.
283
Revista Árvore, Viçosa-MG, v.40, n.2, p.279-288, 2016
Allometric models for estimating aboveground...
variability of the species (RÉ, 2011). Carvalho (2003)
reported DBH values for two homogenous capixingui
stands. In Telêmaco Borba (PR, Brazil), the DBH was
7.8 cm at 8 years. The narrow row spacing (4,444 trees/
ha) may have compromised the development of these
trees. In contrast, in Ilha Solteira (SP, Brazil), the mean
DBH was 6.9 cm at one year after planted, at a density
of 1,111 trees/ha. It is possible that in the present
study the trees benefitted from the wider spacing and
the presence of fertilizer residue of the crops between
rows.
The estimated total biomass was 17.3 t.ha
-1
,
corresponding to 86.64 kg/tree. Approximately 96%
(16.7 t.ha
-1
) of this biomass was found in the upper
diameter size classes (14-20.9 and 21-27.9), where there
is a higher frequency of individuals per hectare. The
production of aboveground biomass was higher than
that obtained for 6-year-old
Croton urucurana
in a
reforestation area of naturally low fertility, which produced
on average 32.6 kg/tree at a density of 1,667 plants/
ha (FEREZ, 2010). In that study, the mean individual
total (aboveground + belowground) biomass production
Figure 2 –
Residual plot of the four models as a function of different predictor variables (diameter at breast height, DBH;
stem diameter, Ds; crown diameter, Dc).
Figura 2 –
Distribuição gráfica de resíduos para os quatro modelos testados em função de diferentes variáveis preditoras
(em função do DAP – diâmetro à altura do peito, D colo - diâmetro do colo, Dc - diâmetro de copa).
284
Revista Árvore, Viçosa-MG, v.40, n.2, p.279- 288, 2016
NICODEMO, M.L.F. et al.
of 20 species of a seasonal semideciduous forest was
a mean total biomass of 35 kg/tree (FEREZ, 2010). Soil
fertility and provenance, among other reasons, might
increase tree production.
For the four models studied (Tabela 1), the best
fit was obtained when DBH was included in the model,
with R
2
ad
values ranging from 0.985 for models 1 and
2 to 0.9801 for model 3. The use of crown diameter
also provided satisfactory fits, with R
2
ad
ranging from
0.98 for model 1 to 0.9436 for model 4. Stem diameter
was the least efficient variable to fit the models tested,
despite the observation of satisfactory R
2
ad
values which
ranged from 0.9587 for model 1 to 0.9168 for model
4. The lowest S
xy
values were obtained for the models
adjusted as a function of DBH (0.1999 for model 4 and
0.2187 for model 3). When the models were adjusted
as a function of crown diameter, S
xy
ranged from 0.2427
(model 1) to 0.3685 (model 4). The highest S
xy
values
were obtained when stem diameter was used (0.3484
for model 1 and 0.4473 for model 4).
For the models adjusted as a function of DBH,
the statistical parameters were closely similar (0.9801
< R
2
ad
< 0.9850 and 0.1999 < S
xy
< 0.2099). Among the
double-entry models (1 and 3), model 1 showed the
best fit, whereas among the single-entry models
(2 and 4), model 4 provided the best R
2
ad
and S
xy
values.
For the models adjusted as a function of stem diameter
and crown diameter, model 1 (double entry) showed
the best fit.
For the equations adjusted as a function of DBH
(Table 1 and Figure 2), model 4 (single entry) was similar
to the other models, a finding that supports its indication
for biomass estimation since it is simpler while presenting
the same efficiency. Barbeiro et al. (2009) also
recommended the use of a single-entry equation using
DBH to estimate total biomass of
Nectandra grandiflora
,
with an R
2
ad
of 0.9458 and S
xy
of 0.1969. However, total
biomass included belowground biomass in that study.
Kumar et al. (1998) tested similar models for estimating
Table 2 –
Statistical parameters of the allometric models used for estimating biomass of tree components of 4.5-year-old
capixingui established in an agrisilvicultural system.
Tabela 2 –
Parâmetros estatísticos da modelagem de equações alométricas para estimação de biomassa de componentes
de capixingui estabelecidos em um sistema silviagrícola, aos 4,5 anos de idade.
Equation Coefficients R
2
ad Sxy
0
1
2
Bole
1 -15.0651 0.6805 2.4295 98.79 0.1732
2 - 2.7771 2.2878 -0.0394 95.96 0.316 2
3 - 5.5287 0.6790 95.85 0.3041
4 - 1.7473 1.5941 93.16 0.3906
Branches
1 -20.1640 0.7204 3.1347 92.48 0.5340
2 - 4.4605 2.8962 -0.0566 90.15 0.611 5
3 - 7.4997 0.8100 88.01 0.6101
4 - 2.9805 1.8992 85.00 0.6825
Leaves
1 -16.9614 0.7911 2.3654 85.87 0.6664
2 - 4.8686 2.2690 -0.0334 83.25 0.725 5
3 - 7.9701 0.7149 82.19 0.6767
4 - 3.9952 1.6806 81.66 0.7202
Figure 3 –
Distribution of aboveground biomass according
to diameter size class.
Figura 3 –
Distribuição da biomassa aérea por classe diamétrica.
285
Revista Árvore, Viçosa-MG, v.40, n.2, p.279-288, 2016
Allometric models for estimating aboveground...
biomass in silvopastoral and woodlot experiments
involving different tree species. The results obtained
for the single-entry model (biomass as a function of
DBH) and double-entry model (biomass as a function
of DBH and height) were similar and satisfactory.
According to Kaonga and Bayliss-Smith (2010), biomass
depends mainly on tree diameter since the bole
accumulates most of the biomass, although the inclusion
of height in diameter-based prediction equations may
describe more accurately the variability in aboveground
biomass. The authors suggested that tree biomass of
other species with similar diameter and height ranges
and crown architecture could be estimated using
regression equations developed for different species
grown in the same environment provided they are
regularly calibrated. However, wood density is another
important factor to be considered when selecting the
equations (CAMPBELL et al., 1985).
4.2. Allometric equations for biomass of tree components
The best fits for estimating biomass of tree
components (Table 2) were obtained with double-entry
equation (1) (lnW = b0 + b1*lnDBH + b2*lnH), which
showed a high capacity to explain the variability in
the trees included in the sample. Although the double-
entry models (as a function of DBH and H) provided
the highest R
2
ad
and lowest S
xy
, the single-entry models
(including only DBH as the independent variable) also
showed satisfactory fits for bole biomass, which
correspond to the largest part of the tree (Figure 3).
Barrichello et al. (2005) obtained satisfactory fits (0.83
< R
2
ad
< 0.99 and 0.125 < S
xy
< 0.496) for the single-
entry equation lnW = b0 + b1*lnDBH (Koperzky and
Gehrhardt) when estimating tree component biomass
of acácia-negra (
Acacia mearnsii
), including leaves
(R
2
ad
= 0.94; S
xy
= 0.308). Barbeiro et al. (2009) attributed
the high S
xy
of the equations developed for the leaf
component of
Nectandra grandiflora
to the large
heterogeneity in foliage biomass of similar size trees,
which presented irregular crown architecture, a fact
that may also apply to capixingui. The authors suggested
that the models tested were not adequate for predicting
leaf biomass.
4.3. Biomass partition
The proportion of biomass accumulated in the
bole (Figure 3) decreased with increasing diameter size
class (from 65.4% in class 0-6.9 cm to 47.88% in class
21-27.9 cm) and biomass accumulated in the branches
increased (from 27.95% in class 0-6.9 cm to 44.54%
in class 21-27.9 cm).
Foster and Melo (2007) collected data from 120
trees belonging to 44 species. DBH ranged from 4.5
to 57.5 cm in nine heterogeneous reforestation areas
and tree age ranged from 5 to 36 years. Three capixingui
specimens, considered a fast-growing species, were
sampled. Aboveground biomass accounted for 75.5%
of total biomass. Trees with this growth pattern presented
on average 6.5 ± 4.8% of their biomass in leaves (8.2%
of aboveground biomass), 39.9 ± 16.9% in branches
(49.8% of aboveground biomass), 33.8 ± 14.7% in the
bole (42.1% of aboveground biomass), and 19.8 ± 6.7%
in roots. No detailed data according to species or age
were available. These results are compatible with the
present findings, reflecting a greater maturity of the
individuals collected in that study. Results similar to
those obtained here have been reported by Kumar et
al. (1998) for different species grown in silvopastoral
systems in India. Furthermore, Drumond et al. (2007)
and Silva and Sampaio (2008) observed the same pattern
for woody caatinga species. Barrichello et al. (2005)
and Schneider et al. (2005) reported the same pattern
for acácia-negra in southern Brazil.
Foliage biomass increased by 25% with increasing
diameter size class (from 5.7% in class 0-6.9 cm to 7.6%
in class 21-27.9), although the contribution of this
component to total biomass was still low. Kaonga and
Bayliss-Smith (2010) found that less than 3% of
aboveground biomass is allocated to leaves in tropical
forest species, which is probably a consequence of
tree architecture, phenology and age. Barichello et al.
(2005) showed that leaves accounted for 3.39% of
aboveground biomass in an 8-year-old acácia-negra
stand. According to the authors, carbohydrates are
allocated to the crown during the early stages of forest
development and their proportions in wood and bark
increase gradually with age. This observation is supported
by the findings of Caldeira et al. (2001) who detected
20% dry matter of total aboveground biomass in the
leaves of a 2.4-year-old acácia-negra stand.
The spacing used in agroforestry systems may
alter biomass distribution when compared to denser
systems. Gutmanis (2004) studied a silvopastoral system
with two densities of 30-year-old
Pinus elliottii
. The
author observed 4.65% of aboveground biomass in
the needles and cited data from Soave (1990), who found
286
Revista Árvore, Viçosa-MG, v.40, n.2, p.279- 288, 2016
NICODEMO, M.L.F. et al.
7.4% of aboveground biomass in the needles of a 31-
year-old pure
P. elliotti
stand. The bole accounted
for 77-78% of aboveground biomass in the silvopastoral
system and for 79% in the pure forest plantation. The
contribution of leaves was higher in wider spaced
plantations, which is a characteristic of integrated systems.
Although Gutmanis (2004) concluded that the
greater allocation of energy to branch production is
unfavorable in silvopastoral systems focusing on wood
products, the fact that that capixingui exhibits sympodial
growth and is a heliophilic and pioneer species should
be taken into consideration. After a certain amount
of biomass is accumulated in the stem/bole, energy
is allocated to the formation of the canopy (branch
production), which is expected for a sympodial species.
This can be seen in Figure 3, in which the proportion
of branch biomass in relation to bole biomass increases
in larger diameter size classes.
The partitioning of woody biomass between stem
and branches decreased with increasing DBH of the
trees, increasing the amount of woody biomass in
branches and limiting the bole height. As a consequence,
the use of this species in agrisilvicultural system would
be restricted, especially for carbon sequestration and
immobilization in wood products, since pruning is a
necessary silvicultural practice to regulate competition
for light exerted by the tree canopy on agricultural
or forage components of these systems. The practice
of pruning 50% of the green crown in trees with a DBH
> 14 cm could lead to the removal of more than 25%
of woody biomass, corresponding exactly to the fraction
where the highest carbon concentration is found. Rather
than using pruning to control competition in systems
composed of capixingui, the thinning of this species
could be predicted and the material collected could
be used for the production of briquettes.
5. CONCLUSIONS
All models provided a satisfactory fit for predicting
biomass. Therefore, we recommend the use of model
4 (lnW = b0 + b1*lnDBH) for estimating tree biomass
in capixingui grown in an agrisilvicultural system because
of its sim plicity and easy data collection and
transformation. The variable DBH should be used since
it provided the best fit of the models tested.
The estimated total biomass of capixingui was 17.3
t.ha-1, corresponding to 86.6 kg per tree. Aboveground
biomass production follows the order bole>branches>leaves
in 4.5-year-old capixingui. The contribution of branches
increases with diameter size class.
6. REFERENCES
BALMFORD, A.; GREEN, R.; PHALAN, B. What
conservationists need to know about farming.
Proceedings of the Royal Society B
,
v.279, p.2714-2724, 2012.
BARBEIRO, L.S.S.; VIEIRA, G.; SANQUETTA,
C.R. Equações para estimativa da biomassa
individual de Nectandra grandiflora Ness (canela-
amarela).
Floresta
, v.39, n.4, p. 833-843, 2009.
BARICHELLO, L.R.; SCHUMACHER, M.V.;
VOGEL, H.L.M. Quantificação da biomassa de um
povoamento de Acacia mearnsii De Wild. na
região sul do Brasil.
Ciência Florestal
, v.15,
n.2, p.129-135, 2005.
BROWN, S.
Estimating biomass and
biomass change in tropi cal forests
: a
primer. Rome: FAO, 1997. 55p. (Forestry paper, 134).
CALDEIRA, M.V.W.; SCHUMACHER, M. V. ;
RONDON NETO, R. M. Quantificação da biomassa
acima do solo de Acacia mearnsii De Wild.,
procedencia Batemans Bay – Austrália.
Ciência
Florestal,
v.11, n.2, p.79-91, 2001.
CAMPBELL, J.S.; LIEFFERS, V.J.; PIELOU, E.C.
regression equations for estimating single tree
biomass of trembling aspen: assessing their
applicability to more than one population.
Fo re st Ecology and Manage men t
, v.11,
p.283-295, 1985.
CARVALHO, P.E.R.
Espécies arbóreas
brasileiras
. Brasília: Embrapa Informação
Tecnológica; Colombo, PR: Embrapa Florestas,
2003. 1039p.
CARPANEZZI, A. A.; NEVES, E. J. M.; AGUIAR, A.
V.; SOUSA, V. A. Espécies lenhosas alternativas para
fins econômicos no Paraná. In: SEMINÁRIO DE
ATUALIZAÇÃO FLORESTAL, 2.; SEMANA DE
ESTUDOS FLORESTAIS, 11., 2010, Irati.
Anais
...
Irati: UNICENTRO, 2010. 9p. Disponível em:<http://
anais.unicentro.br/sef2010/pdf/palestras/
Carpanezzi.pdf>. Acesso em: 03 nov. 2010.
287
Revista Árvore, Viçosa-MG, v.40, n.2, p.279-288, 2016
Allometric models for estimating aboveground...
CHAVE, J.; ANDALO, C.; BROWN, S.; CAIRNS, M.
A.; CHAMBERS, J. Q.; EAMUS, D.; FOLSTER, H.;
FROMARD, F.; HIGUCHI, N.; KIRA, T.; LESCURE, J.
P.; NELSON, B. W.; OGAWA, H.; PUIG, H.; RIERA,
B.; YAMAKURA, T. Tree allometry and improved
estimation of carbon stocks and balance in tropical
forests.
Oecologia
, v.145, n.1, p.87-99, 2005.
DRUMOND, M. A.; SALVIANO, L. M. C.;
CAVALCANTI, N. DE B.; PEREIRA, L. G. R.
Produção, distribuição da biomassa e composição
bromatológica da parte aérea da faveleira.
Revis ta Brasilei ra de Ciê nci as
Agrárias
, v.2, n.4, p.308-310, 2007.
JOSE, S.; GILLESPIE, A.R.; PALLARDY, S.G.
Interspecific interactions in temperate
agroforest ry.
Agroforestr y Sys tems
, v.6 1-62,
n.1, p.237-255, 2004.
FEREZ, A.P.C.
Efeito de práticas
silvicultur ais s obre as tax as iniciais
de s equestro de ca rbon o em planti os
de res tauração d a M ata Atlâ ntica
.
Piracicaba: 2010. 104f. Dissertação (Mestrado em
Ciências Florestais) - Escola Superior de
Agricultura Luiz de Queiroz, Piracicaba, 2010
FORSTER, H.W.; MELO, A.C.G. Biomassa aérea e
de raízes de reflorestamentos heterogêneos no
vale do Paranapanema, SP.
IF. Série
Registros
, n.31, p.153-157, 2007.
GOUVEIA, T. C.; FLORSHEIM, S. M. B.;
PASTORE, J. A.; AGUIAR, O. T.; LIMA, I. L.
Morfologia, anatomia do lenho e densidade
básica de Croton floribundus Spreng e Croton
macrobothrys Baillon.
IF. Série Regi stros
,
n.31, p.45-49, 2007. (Edição dos anais do 1º
Seminário de Iniciação Científica do Instituto
Florestal, São Paulo, 2007)
GUTMANIS, D.
Es toqu e de car bono e
di nâmi ca ecof isiol ógic a em sist emas
silvipastoris
. 2004. 142f. Tese (Doutorado em
Ciências Biológicas) – Universidade Estadual
Paulista, Rio Claro, 2004.
KAONGA, M.L.; BAYLISS-SMITH, T.P.
Allometric models for estimation of aboveground
carbon stocks in improved fallows in eastern
Zambia.
Agroforestry Systems
, v.78, p.217-
232, 2010.
KUMAR, B.M.; GEORGE, S.J.; JAMALUDHEEN,
V.; SURESH, T.K. Comparison of biomass
production, tree allometry and nutrient use
efficiency of multipurpose trees grown in
woodlot and silvopastoral experiments in Kerala,
In dia.
Forest Ec olog y a nd Mana geme nt
,
v.112, n.1, p.145-163, 1998.
MELOTTO, A.; NICODEMO, M.L.F.;
BOCCHESE, R.A.; LAURA, V.A.; GONTIJO
NETO, M.M.; SCHLEDER, D.D.; POTT, A.;
PORFÍRIO-DA-SILVA, V. Sobrevivência e
crescimento inicial em campo de espécies
florestais nativas do Brasil Central indicadas
pa ra sistemas silvipastoris.
Revista Árvore
,
v.33, p.425-432, 2009.
NAIR, P.K.R. Classification of agroforestry
sy ste ms.
Agrofores try Systems
, v.5, p .97-
128, 1985.
NICODEMO, M.L.F.; PORFÍRIO-DA-SILVA, V.;
SANTOS, P.M.; VINHOLIS, M.M.B.; FREITAS,
A.R.; CAPUTTI, G. Desenvolvimento inicial de
espécies florestais em sistema silvipastoril na
Regiã o Sudeste.
Pesquisa Florestal
Brasileira
, v.60, p.89-92, 2009.
PADILHA, D.R.C.
Fitossociologia e
estim ativas da biomassa aérea e de
carbono em Chaco Florestado no
Brasil
. 2011. 86f. Dissertação (Mestrado em
Biologia Vegetal) - Universidade Federal de Mato
Grosso do Sul, Campo Grande, 2011.
RÉ, D.S .
Equaç ões v olumétricas e
produção de made ira em pl ant ios
mi sto s visando à restauração da
fl oresta estacional semidecidual
. 2011.
65f. Dissertação (Mestrado em Ciência Florestal) -
Universidade Estadual Paulista, Botucatu, 2011.
RUSCHEL, A. R.; NODARI, E.S.; GUERRA, M.P.;
NODARI, R.O. Evolução do uso e valorização das
espécies madeiráveis da Floresta estacional
decidual do Alto-Uruguai, SC.
Ciência
Florestal
, v.13, n.1, p.153-166, 2003.
SALIS, S. M.; ASSIS, M. A.; MATTOS, P. P.;
PIÃO, A. C. S.
Equações para estimar
biomassa e v olume de madeira em
cerradões do Pantan al da
Nhecolândia, MS.
Corumbá: Embra pa
Pantanal, 2004. 4p. (Comunicado Técnico, 50).
288
Revista Árvore, Viçosa-MG, v.40, n.2, p.279- 288, 2016
NICODEMO, M.L.F. et al.
Disponível em: <http://www.cpap. embrapa.br/
publicacoes/online/CT50.pdf>. Acesso em: 18
out. 2012.
SCHNEIDER, P.R.; FINGER, C.A.G.;
GIACOMELLI SOBRINHO, V.; SCHNEIDER,
P.S.P. Determinação indireta do estoque de
biomassa e carbono de povoamentos de acácia-
negra (
Acacia mearn sii
De Wild).
Ciência
Florestal,
v.15, n.4, p.391-402, 2005.
SILVA, G.C., SAMPAIO, E.V.S. Biomassas de
partes aéreas e m plantas da caa tinga.
Revista
Árvore
, v.32, n.3, p.567-575, 2008.
SOARES, C.P.B.; PAULA NETO, F.; SOUZA,
A.L. Equações hipsométricas, volum étricas e de
taper para onze esp écies nativas.
Revista
Árvore
, v.35, p.1039-1053, 2011.
SOARES, C.P.B.; PAULA NETO, F.; SOUZA,
A. L.
Den d rometr ia e in v entári o
florestal
. Viçosa, MG: UFV, 2007. v.1. 276p.
SOAVE, R.C.F. Influência de diferentes tiposde
extração de resina na ciclagem de nutrients e no
crescimento de indivìduos de Pinus elliottii
Engelm. var. elliottii (Mogi Guaçú – SP). Rio
Claro: 1990. Dissertação (Mestrado em Ciências
Biológicas) – Instituto de Biociências,
Universidade Estadual Paulista. Rio Claro, 1990.
TIEPOLO, G.; CALMON, M.; FERETTI, A.R.
Measuring and monitoring carbon stocks at the
Guaraqueçaba climate action project, Paraná,
Brazil. In: International Symposium on Forest
Carbon Sequestration and Monitoring. Taipei,
Taiwan.
Anais
... Tai pei: Taiwan Forestry
Research Institute, 2002. p.98-115.
TILMAN, D.; BALZER, C.; HILL, J.; BEFORT,
B.L. Global food demand and sustainable
intensification of agriculture.
PNAS
, v.108,
n.50, p. 20260-20264, 2011.
VENTURIN, R.P.; GUERRA, A.R.; MACEDO,
R.L.G.; VENTURIN, N.; MESQUITA, H.A.
Sistemas agrossilvipastoris: origem,
modalidades e modelos de implantação.
InformeAgropecuário
, v.31, p.16-2 4, 201 0.
