Forest Ecology and Management 260 (2010) 1906–1913

Contents lists available at ScienceDirect

Forest Ecology and Management

journal homepage: www.elsevier.com/locate/foreco

Estimation of biomass and carbon stocks in plants, soil and forest floor in
different tropical forests
Juan Carlos Loaiza Usuga a,∗ , Jorge Andrés Rodríguez Toro b ,
Mailing Vanessa Ramírez Alzate b , Álvaro de Jesús Lema Tapias c
a
Laboratory of Hydrology and Soil Conservation, Forestry and Technology Centre of Catalonia, Solsona, Spain
b
Faculty of Forest Sciences, University of Concepción, Chile
c
Forestry Sciences Department, National University of Colombia, Medellin, Colombia

a r t i c l e i n f o a b s t r a c t

Article history: An accurate characterization of tree carbon (TC), forest floor carbon (FFC) and soil organic carbon (SOC)
Received 17 May 2010 in tropical forest plantations is important to estimate their contribution to global carbon stocks. This
Received in revised form 23 August 2010 information, however, is poor and fragmented. Carbon contents were assessed in patula pine (Pinus
Accepted 23 August 2010
patula) and teak (Tectona grandis) stands in tropical forest plantations of different development stages
in combination with inventory assessments and soil survey information. Growth models were used to
Keywords:
associate TOC to tree normal diameter (D) with average basal area and total tree height (HT ), with D and
Biomass
HT parameters that can be used in 6–26 years old patula pine and teak in commercial tropical forests as
Carbon
Forest floor
indicators of carbon stocks. The information was obtained from individual trees in different development
Tropical forests stages in 54 patula pine plots and 42 teak plots. The obtained TC was 99.6 Mg ha−1 in patula pine and
Soils 85.7 Mg ha−1 in teak forests. FFC was 2.3 and 1.2 Mg ha−1 , SOC in the surface layer (0–25 cm) was 92.6
and 35.8 Mg ha−1 , 76.1 and 19 Mg ha−1 in deep layers (25–50 cm) in patula pine and teak, respectively.
Carbon storage in trees was similar between patula pine and teak plantations, but patula pine had higher
levels of forest floor carbon and soil organic carbon. Carbon storage in trees represents 37 and 60%
of the total carbon content in patula pine and teak plantations, respectively. Even so, the remaining
percentage corresponds to SOC, whereas FFC content is less than 1%. In summary, differences in carbon
stocks between patula pine and teak trees were not significant, but the distribution of carbon differed
between the plantation types. The low FFC does not explain the SOC stocks; however, current variability
of SOC stocks could be related to variation in land use history.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction Preliminary results in forests show that traditional carbon esti-

The majority of carbon in the terrestrial pool is stored below
ground in soils (Janzen, 2004). Total global carbon in soils consti-
tutes between 1500 and 2000 Gtons; the majority of it stored in
forest biomes (Janzen, 2004; Smith, 2004). Greenhouse gas emis-
sions and the climatic change risk, make a reduction in atmospheric
CO2 concentrations necessary. However, most of the estimations
and assumptions realized are uncertain for many ecosystems
(Hendricks and Turkenburg, 1997; IPCC, 2000). In the last few years
in particular, there has been increasing interest in the quantifica-
tion of the biomass of forest ecosystems and its potential carbon
fixation (Chave et al., 2005; Fearnside, 1996; Schulp et al., 2008;
Sierra et al., 2007).
∗ Corresponding author at: Puig Salner 14, E-17486, Castello D’Empuries, Spain.
Tel.: +34 972250142; fax: +34 973480431.
E-mail address: jc.loaiza@ctfc.es (J.C.L. Usuga).
mation methodologies tend to under or overestimate real carbon
stocks in different forest types (Tan et al., 2007; Federici et al., 2008;
Schulp et al., 2008). Sometimes, these erroneous estimations are
due to inaccurate assumptions of layers soil properties and the dis-
tribution of soil type with dominant vegetation (Milne and Brown,
1997). The results also indicate the relevance and a high hetero-
geneity in soil organic carbon (SOC), probably associated with the
variability in topography, stoniness, parent material, soil depth and
microclimate, on a local scale. While knowledge is being gathered
in different regions of the world, few results are available from
tropical areas. According to Sierra et al. (2007), the estimations of
total carbon stocks in tropical forests obtained from aboveground
biomass tend to largely underestimate total carbon stocks. An accu-
rate estimation of the magnitude of carbon stocks in the different
components of forest ecosystems in tropical climates is therefore
essential. More specifically, the estimation of the carbon stored in
the aerial and subterranean part of the tree as well as in the fallen
leaves and litter, allows one to understand the carbon dynamic and

0378-1127/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.foreco.2010.08.040
 J.C.L. Usuga et al. / Forest Ecology and Management 260 (2010) 1906–1913
Fig. 1. Location of the study sites. All of the experiments were carried out in tropical
forest plantations.
the real carbon content on each ecosystem component in detail
(Cerri et al., 2003; Fearnside, 1996, 2004; Fearnside and Imbrozio
Barbosa, 1998; Sales et al., 2007; Sierra et al., 2007). This fragmented
knowledge makes it impossible to obtain more accurate balances
for each ecosystem type.
Several methodologies exist to quantify the carbon stocks
of the soil and vegetation. Different modelling techniques and
methodologies can be used reliably in tropical forest to predict
aboveground tree biomass across a broad range of ecosystems
(Brown, 2002; Chave et al., 2005). A large number of growth models
have already been published by Elfving and Kiviste (1997), Kangur
et al. (2007) and Kiviste et al. (2002). Most of these studies have
been associated with forestry inventory programs and dasometric
studies (Chave et al., 2005; Fearnside and Imbrozio Barbosa, 1998;
Montero and Kanninen, 2003; Schulp et al., 2008; Sierra et al., 2007).
These works have obtained a partial determination of the carbon
content in the ecosystems studied. These authors have only mea-
sured the carbon content in a few ecosystem components and the
missing information to obtain the total carbon balance was extrap-
olated or taken from literature. Additionally, in each of these studies
there were methodological adaptations and approximations, which
in most cases could lead to an erroneous estimation of the poten-
tial to capture carbon in forestry ecosystems. Furthermore, due to
these data generally being acquired through interpolation, remote
sensing or model simulation, sometimes erroneous assessments
of carbon stocks in the system exist, associated with non-linear
variable estimations, as well as problems with the application of
spatial and temporal scales (Kuikman et al., 2003; Paustian et al.,
1997). Due to the high variability among tropical plantations in age,
genetical factors, management of the commercial plantations and
environmental conditions, additional information to improve the
1907
total carbon stock estimation in these ecosystems is needed (Sierra
et al., 2007).
The aim of this paper is to quantify the effect of: site factors,
forest management, environmental conditions and tree species on
carbon stocks in mineral soil, forest floor, and tree. At the same time
we evaluate the application of forest inventories and soil survey
information for carbon stocks inventory improvement. The total
carbon stocks in tree, forest floor and mineral topsoil were analyzed
in different tropical plantations, i.e. in patula pine (Pinus patula) and
teak (Tectona grandis) stands. In situ observations at a number of
test sites in four different zones in Colombia were used for this
purpose.
2. Methods
2.1. Site description
For the study of carbon stocks in commercial plantations of
Colombia, four different zones were studied. The spatial distribu-
tion of the study sites appears in Fig. 1 with their characteristics
summarized in Table 1. The sampling zones cover the main Colom-
bian forest types and representative age rank of typical commercial
forests. The total area of 3712 ha studied, did not undergo a compre-
hensive plan for forest management. The study regions are in the
Caribbean plains, West and Central Andes. The topography consists
of hills and mountains with slopes often higher than 20%, and an
altitude ranging from 30 to 2160 m. Soils are deep, with generally
low fertility. A description of the complete soil survey can be found
in Cortés et al. (1982). The predominant land use is extensive live-
stock and agriculture (cassava, corn, banana and coffee). Other land
uses are natural forests and coal mining.
2.2. Experimental design
The network of plots established covers the main types of
planted forest and the age distribution of typical commercial forests
in the Andes and the Caribbean regions in Colombia. The monitoring
plots were distributed randomly in 54 patula pine plots and 42 teak
plots. The dasometric and morphologic characteristics conditions
of each individual tree were recorded on each plot of 500 m2 . The
carbon stocks in aerial biomass, forest floor, roots and soil profile
were measured during a campaign from January to December 2006.
To calculate the biomass sampled, an adaptation of Schlegel et al.
(2000) methodology was used. The selected trees represent mean
characteristics of the population in diameter, height and morphol-
ogy, all this information is complemented by the annual inventory
reports conducted from 1989. The total number of trees per site
was set establishing chronosequences with plots, distributed over
a wide rank of growth, density and ages.
2.3. Forest components quantification
For the aerial biomass sampled, moist weight in wood, foliage,
fruits, fine (≤2.54 cm diameter) and coarse (>2.54 cm diameter)
branches were measured in the field. For quantification of the total
biomass of the tree, the wood was cut into three pieces and each
piece was weighed; afterwards a proportionately distributed, 5 cm
thick sample was cut. The total weight of each part of the tree
(leaves, branches, trunk and roots) was estimated by measuring
the total fresh weight in the field and drying (at 60 ◦ C until con-
stant weight) samples in the laboratory to determine moisture
content. Later, a dry biomass subsample was taken for the total car-
bon content determination according to methodology described by
Montané et al. (2006). The total carbon per hectare was determined
according to the available information from the forestry inventory
of the different studied sites. The carbon content, in proportion
1908 J.C.L. Usuga et al. / Forest Ecology and Management 260 (2010) 1906–1913

Table 1
Description of the field study sites included in this study.

Site Puerto Libertador Toluviejo Angostura Manizales

Tree specie Tectona grandis Tectona grandis Pinus patula Pinus patula
Area (ha) 1270 80 2100 262
Age 6–26 12–17 6–26 8–25
Trunk diameter (cm) 8–34 16–21 11–27 18–33
Trunk height (m) 9–26 20–23 9–16 14–30
Trees (ha−1 ) 997 1540 1412 936
Parent material Sedimentary rocks Sedimentary rocks Metamorphic rocks Volcanic ashes
Relief Elongated hill Rounded hill Dissected slope Rounded slope
Soil Typic dystrudepts Typic dystrudepts Typic dystrudepts Andic dystrudepts
Soil thickness >150 cm >150 cm 100–150 cm >150 cm
Ecological zone Tropical dry forest Tropical dry forest Tropical wet forest Tropical wet forest
Mean rainfall (mm) 1249 1000 1500 1495
Mean temperature (◦ C) 28 27 14 17
Altitude (m) 150 30 2300 2160
Old land use Gold mining Gold mining Extensive pasture Extensive pasture

to the weight of each biomass component, determines the weight
values. This total carbon fraction was calculated according to the
methodology described by Garcia-Pausas et al. (2007).
The tree root density (coarse, medium and fine roots) was esti-
mated by pulling up roots. A root sample for every size was collected
to calculate the total root content per tree. Soil stuck to the roots
was removed by hand. The forest floor was sampled in each plot,
at 3 points, that were chosen randomly. For the fine pieces, the
material was selected according to the state of decomposition. The
layers distinguished in the forest floor were litter layer (L), frag-
mented layer (F) and humus layer (H), using methodology described
by Duchaufour (1994). Litter sampling was conducted with 0.8 m
height network 0.8 m × 0.8 m grids in patula pine and 1.2 m × 1.2 m
for teak, sampling 3 points per plot. These plots were sampled 3
months later, where total litter per plot was measured to calculate
the total carbon in the litter. For each forest component a kilogram
of subsample was collected.
The soil samples were dried for 48 h at 105 ◦ C (SSS, 1992). To
quantify the carbon, a dry sample of 25 g was taken, and passed
through a 2 mm grinder (Willey Mill). Later, five subsamples of 3 g
each were collected in a drying chamber. The carbon determination
was performed according to the Walker and Black (1934) modified
method. Through the quantification of oxide reduction percentage,
the carbon concentration was determined, by means of calorimetry,
using an isotope radio mass Lambda 3B UV/Vis spectrophotometer
(Peking & Elmer). The obtained results were correlated, accord-
ing to a determination coefficient of 0.985 with the experimental
results obtained by means of the elemental analyser CHN 600 (Leco
Truspec). The carbon content in soil was calculated from the carbon
content of the sample multiplied by bulk density, the thickness of
the mineral soil layer and the area it represents according to the
soil survey information.
2.5. Growth models and dasometric variables analysis

The data needed for the calibration of growth models was

2.4. Soil sampled and treatment
Although long-term accumulation of carbon takes place in
deeper soil layers, we focus on mineral topsoil because this is the
main area of interest in spatial inventories of SOC given Kyoto
report requirements. Representative soil profiles were selected
according to information on a soil survey map (Cortés et al., 1982).
In total, 96 soil profiles were characterized according SSS (1993)
methodology. Soil sample depth was established based on SOC
stocks results in tropical forests (Batjes, 2008; Cerri et al., 2003;
Moraes et al., 1995; Sierra et al., 2007). In each layer, a kilogram
soil composite sample for soil carbon content was taken from a
depth of 0–25 cm (surface layer) and 25–50 cm (subsurface layer),
respectively. A soil core sampler was used to determine the bulk
density (SSS, 1992) in each layer.
obtained by continuous observation of permanent growth plots
from January to December 2006. Sixteen models were obtained
with a complete set of data for patula pine and teak commer-
cial plantations, and ninety-six trees in total for both species. The
Hossfeld I and Hossfeld I modified, Smalian, Todorovic I, VIII and
IX, Strand, Levakovic II and III, Gemesi, Weibull II, Van der Vliet,
Kövessy, Bass, Sevastyanov I and allometric regression models com-
monly used to estimate tree growth and stand development were
evaluated. For an overview of these models, refer to Elfving and
Kiviste (1997) and Kiviste et al. (2002). The Smalian, Kövessy and
Todorovic I growth adapted models were selected based on their
mathematical simplicity, applied relevance and good adjustment,
expressing Mg of C for trees (Ct), as a function of the D (in m) and
HT (in m). The statistics for the calibrated growth non-linear mod-
els with the highest adjustment to the relation between estimated

Table 2
Estimated regression coefficients based on non-linear modelling.

a b c f DW RMSE MAE AIC

Pinus patula
Smalian Eq. (1) 4.4340*** 3.8180** −1.0041*** nd 1.9854 0.0290 0.0194 −376.2
Kövessy Eq. (2) 0.0908* 2.5166 ns −0.2537 ns nd 1.9561 0.0303 0.0206 −371.6

Tectona grandis
Smalian Eq. (1) 3.8694*** 3.2411** −0.8133* nd 2.0706 0.0255 0.0187 −302.2
Todorovic I Eq. (3) 0.5253 ns −3.0952 ns 9.5787*** −0.6557 ns 2.0284 0.0260 0.0193 −298.5

Best results obtained from the Smalian non-linear model. DW: Durbin–Watson statistics, RMSE: root of mean square error, MAE: average absolute error, AIC: Akaike
information criterion; nd: no data. ns: not significant (p ≥ 0.05).
*
Significant (p < 0.05).
**
Very significant (p < 0.01).
***
Extremely significant (p < 0.001).
 J.C.L. Usuga et al. / Forest Ecology and Management 260 (2010) 1906–1913 1909

wood production and carbon content for patula pine and teak are Table 3
Estimated regression coefficients based on additively of non-linear equations.
shown in Table 2. These models have been well adjusted to esti-
mate the accumulated carbon amount in patula pine and teak trees Pinus patula Tectona grandis
from the information obtained in the forestry inventories in the RMSE 0.03886 0.02454
study zone. The total carbon is then determined, depending on the a1 11.2598*** 5.7936***
carbon content of each component of P. patula (Eqs. (1)–(4)) and T. b1 nd 5.3980***
grandis (Eqs. (5)–(9)). c1 nd −1.3072***
(D2 · HT ) a2 nd 59.8695***
Pinus Wood (w) Cw = (1) a3 nd 51.6195**
a1
patula b3 1777.6916*** 94.6036**
c3 nd −29.5329**
(D2 · HT )
Branches (fb) Cfb = (2) a4 nd 7061.8610***
b3 · (D2 · HT )
b4 555.0909*** nd
(D2 · HT ) c4 −255.131*** −8804.3300***
Leaves (l) Cl = (3)
b4 · (D2 · HT ) + c4 · (D2 · HT )
2 a5 27.5369*** 17.97184***
b5 8.7930* 21.8624**
(D2 · HT ) c5 nd −5.6521*
Roots (r) Cr = (4)
a5 + b5 · (D2 · HT )
ns: not significant (p ≥ 0.05); RMSE: root of mean square error; nd: no data.
(D2 · HT ) *
Significant (p < 0.05).
Tectona Wood (w) Cw = 2
(5)
grandis a1 + b1 · (D2 · HT ) + c1 · (D2 · HT ) **
Very significant (p < 0.01).
***
Extremely significant (p < 0.001).
(D2 · HT )
Coarse branches Ccb = (6)
a2
(cb)
ing the first 10 years in the wood, roots, coarse and thin branches,
(D2 · HT ) leaves and fruits. The wood has higher carbon content than the
Fine branches (fb) Cfb = 2
(7)
a3 + b3 · (D2 · HT ) + c3 · (D2 · HT ) roots, branches and foliage, respectively. The growth in teak tends
(D2 · HT ) to be lower in the tropics, while patula pine tends to have a faster
Leaves (l) Cl = (8)
a4 + c4 · (D2 · HT )
2 growth and higher yield (Gillespie, 1992; Weaver, 1993). Never-
theless, the carbon volume in teak is 33–60% higher in roots (10–26
(D2 · HT )
Roots (r) Cr = 2
(9) years) and 4–24% higher in wood (6–26 years). The carbon content
a5 + b5 · (D2 · HT ) + c5 · (D2 · HT )
in branches, leaves and fruits does not show significant differences
The estimated regression coefficients, based on the addition
between species. The relation between the observed age ranks and
of non-linear equations where parameters were finally obtained,
the increase of the trunk biomass and the root production found in
are in most cases extremely significant (p < 0.001) and shown in
this study is similar to the correlation results obtained by Weaver
Table 3. Finally, only the significant parameters of each compo-
(1993). The total carbon content, although not statistically signifi-
nent (obtained from the first convergence of the parameters) were
cant, is slightly higher in teak than in patula pine. This response is
considered.
associated with a higher basic density in teak which allows more
fixation of carbon with similar volumes of biomass.
3. Results and discussion
3.2. Organic carbon in forest floor
3.1. Tree organic carbon
The carbon stocks of the forest floor are shown in Fig. 2, and the
The organic carbon (OC) values for different component systems carbon stocks average of different surface layers and the carbon
are shown in Fig. 2. In the natural system, patula pine contributes content in the soil profile are shown in Table 5. The FFC repre-
25–43% of total carbon stocks and teak 51–67%. The mean carbon sents 0.5–1.2% of the total carbon in the system for patula pine and
content and standard deviation of the different parts of the trees are teak, respectively. The maximum values were registered between
shown in Table 4. The carbon content decreases progressively dur- 6 and 10 years, and 10 and 17 years, and the minimum values

Fig. 2. Carbon stocks (Mg ha−1 ) in all the sites, comparison between Pinus patula and Tectona grandis (6–10, 10–17, 17–26) years old stands. Total tree, forest floor (litter
layer + fragmented layer + humus layer) and mineral soil carbon stocks.
1910 J.C.L. Usuga et al. / Forest Ecology and Management 260 (2010) 1906–1913

Table 4
Mean carbon content of the different tree components (kg tree−1 ), mean values and standard deviation.

Years n Mean S.D.

6–10 10–17 17–26

Pinus patula
Wood 34.4 (17.4) 79.1 (44.6) 98.9 (57.3) (13, 20, 21) 76.0 51.5
Coarse 8.2 (5.8) 19.6 (18.7) 19.9 (18.1) (13, 20, 21) 17.0 21.4
branches
Fine 2.9 (2.4) 5.2 (4.4) 4.2 (2.5) (13, 20, 21) 4.3 3.4
branches
Leaves 4.4 (2.4) 5.5 (3.1) 4.7 (2.7) (13, 20, 21) 4.9 2.7
Cone/fruit 1.4 (0.7) 0.9 (1.1) 2.1 (1.6) (13, 20, 21) 1.5 1.3
Root 12.6 (10.3) 16.1 (8.1) 26.1 (8.7) (13, 20, 21) 19.2 10.4
Tree 63.9 (35.3) 126.4 (66.5) 156.0 (87.6) (13, 20, 21) 122.9 77.8

Tectona grandis
Wood 40.7 (34.5) 81.1 (36.2) 105.2 (56.3) (9, 24, 9) 77.6 45.4
Coarse 2.7 (2.3) 11.2 (7.8) 19.6 (13.8) (9, 24, 9) 11.2 10.2
branches
Fine 2.7 (1.5) 6.8 (3.6) 10.6 (7.6) (9, 24, 9) 6.8 5.1
branches
Leaves 1.5 (1.4) 3.8 (4.2) 3.2 (1.9) (9, 24, 9) 3.2 3.4
Cone/fruit nd 0.07 (0.12) 0.03 (0.10) (9, 24, 9) 0.05 0.10
Root 9.1 (3.1) 22.3 (10.0) 37.7 (15.7) (9, 24, 9) 22.7 14.0
Tree 56.7 (38.8) 125.3 (52.8) 176.3 (93.1) (9, 24, 9) 121.5 71.7

n: number of sampled units, S.D.: standard deviation in brackets; mean and S.D. all samples; nd: no data.

between 17 and 26 years, and 6 and 10 years in patula pine and
teak, respectively. The results indicate high carbon stocks in the
litter layer with respect to the humus layer and in the materials
with a medium percentage of decomposition (fragmented layer).
The FFC in the study zones showed lower values than the results
reported in planted forestry areas in temperate zones (Schulp et al.,
2008; Vesterdal et al., 2008). The results suggest a lower organic
matter mineralization in P. patula than in teak plantations. For pine
forest Rigobelo and Nahas (2004) found low organic carbon con-
tributions from litter fall, associated with low microbial activity,
moisture conditions and organic matter quality. Moreover, studies
in teak plantations showed an increase in the annual litter fall as the
plantation grew older; it was completely mineralized in less than
6 months (Egunjobi, 1974). To summarize, the FFC represented a
low percentage of carbon in the system, 0.92 and 0.84% of total
carbon stocks in P. patula and teak plantations, respectively. In a
26-year period, the content tends to be constant for FFC stocks and
less significant with respect to the total content of carbon in the
whole system. It is therefore necessary to reconsider the impor-
tance of FFC content in the total carbon balances in these kinds of
ecosystems.
3.3. Soil organic carbon
The edaphic carbon stocks and its interrelation with the other
components (forest floor and trees) can be observed in Fig. 2.
The mean carbon stocks and standard deviation at 0–25 cm and
25–50 cm depths can be seen in Table 5. There, 57–73% of total car-
bon under patula pine is accumulated, and between 31 and 48%
is stored under teak plantations. The soil stored the highest car-
bon volume in the forest (Fearnside and Imbrozio Barbosa, 1998;
Federici et al., 2008; Schulp et al., 2008; Sierra et al., 2007). Soils
under patula pine carbon enrichment can be associated with con-
stant accumulation of vegetation material during the year. As teak
is a deciduous species, this reserve occurs seasonally and is smaller.
The carbon stock in the first 50 cm of the soil is 1.5–3.6 times
larger in soils under patula pine, with respect to teak. The low
SOC stocks in teak related to high temperatures and precipitation

Table 5
Mean carbon stocks (Mg ha−1 ) in the forest floor and mineral soil and over L–F–H and 0–25 and 25–50 cm layers and bulk density (g cm−3 ) in brackets, mean and standard
deviation.

Years n Mean S.D.

6–10 10–17 17–26

Pinus patula
Litter layer 0.34 (0.27) 0.05 (0.07) 0.26 (0.20) (13, 20, 21) 0.20 0.22
Fragmented layer 0.07 (0.18) 0.39 (0.77) 0.27 (0.53) (13, 20, 21) 0.27 0.58
Humus layer 2.71 (2.09) 1.53 (2.00) 1.07 (0.49) (13, 20, 21) 1.64 1.71
Mineral soil (0–25 cm) 90.8 (22.48) 90.2 (41.84) 82.2 (37.0) (13, 20, 21) 87.2 35.7
Mineral soil (25–50 cm) 52.2 (29.3) 57.0 (25.76) 60.0 (32.5) (13, 20, 21) 57.0 25.5
Bulk density (0–25 cm) 0.99 (0.25) 0.93 (0.18) 1.00 (0.21) (13, 20, 21) 0.97 0.21
Bulk density (25–50 cm) 1.00 (0.27) 0.96 (0.21) 1.03 (0.19) (13, 20, 21) 1.0 0.22

Tectona grandis
Litter layer nd 0.01 (0.02) 0.12 (0.32) (9, 24, 9) 0.03 0.15
Fragmented layer nd 0.56 (0.74) 0.22 (0.41) (9, 24, 9) 0.37 0.63
Humus layer 0.60 (0.43) 1.35 (0.81) 1.21 (1.25) (9, 24, 9) 1.16 0.90
Mineral soil (0–25 cm) 21.7 (12.23) 35.5 (15.83) 39.4 (15.07) (9, 24, 9) 33.4 15.96
Mineral soil (25–50 cm) 19.0 (15.55) 11.9 (11.87) 18.1 (13.43) (9, 24, 9) 14.7 13.14
Bulk density (0–25 cm) 1.45 (0.18) 1.43 (0.31) 1.46 (0.09) (9, 24, 9) 1.44 0.25
Bulk density (25–50 cm) 1.35 (0.15) 1.55 (0.21) 1.46 (0.13) (9, 24, 9) 1.49 0.20

n: number of sampled units, S.D.: standard deviation in brackets; mean and S.D. all samples; nd: no data.
 J.C.L. Usuga et al. / Forest Ecology and Management 260 (2010) 1906–1913 1911

Fig. 3. Relation between bulk density and soil carbon content in forest soils. 0–25 cm and 25–50 cm layers for Pinus patula and Tectona grandis in tropical forests plantations.

(volume and intensity), and old soil uses (mining), which have a
negative effect on organic matter accumulation (Jaramillo et al.,
1994). Other parameters like soil hydraulic properties, thickness,
fertility (Gillespie, 1992; Weaver, 1993), and runoff limit the effi-
ciency of this species (Cortés et al., 1982). Under teak plantations,
runoff is 11 times more than in natural forests (Weaver, 1993). To
summarize, the soil under patula pine (old extensive pasture) has
higher carbon stocks than the soil under teak (mining areas). SOC
is 48–73% of the total carbon in the system, and it is regulated by
the soil fertility and past soil uses (Table 1).
The soils contain high organic matter content associated with
large carbon stocks in the soil profile, which is where the highest
carbon volume is accumulated in the system. When both land uses
were compared, changes in mean bulk density in the studied depth
intervals were observed (Fig. 3). The bulk density in soils under pat-
ula pine and teak in the studied period vary slightly. Patula pine had
lower bulk density values and higher carbon content than teak; the
differences between land uses were not significant. The soil carbon
content in patula pine plantations was approximately 80 Mg ha−1
higher than in teak plantations. This response could be associ-
ated with the previous soil uses. Other studies in the Netherlands
showed that historical land use has a stronger association with SOC
variability than present-day land use (Schulp and Verburg, 2009).
Patula pine plantations were established in soils which had tra-
ditionally been under pasture, so bulk density changes in these
sites could be associated with an increase in organic matter and
the presence of volcanic ashes. Whereas under teak plantations,
bulk density changes are associated with losses in soil quality, due
to previous land use (coal mining). This exploitation process tends
to destroy soil aggregates, resulting in a loss of the superficial lay-
ers, creating compacting and erosion process problems, increasing
physical and chemical soil quality loss (Alfsen et al., 2001; Jaramillo
et al., 2004). The results indicate that bulk density could be an useful
indicator of previous soil uses and soil quality changes.

Fig. 4. Relation between tree biomass (kg) and tree carbon content (kg) in tropical forests plantations. Each dot corresponds to an individual weighed tree.
1912 J.C.L. Usuga et al. / Forest Ecology and Management 260 (2010) 1906–1913
3.4. Organic carbon in the system
In patula pine and teak trees most of the carbon is in the trunk,
followed by the roots and branches. In both species, the compo-
nent that has the minor carbon fraction is the foliage (Table 4).
The correlation between biomass and carbon content in the stud-
ied species, has a lineal trend, with coefficient of determination (R2 )
values between 0.98 and 0.99. Comparing both variables it can be
observed that having the same biomass amount (kg), the carbon
content is 2% higher in teak than in pine (Fig. 4). The underground
biomass has similar values of carbon content in both species. The
forest floor (L, F, H) has the same tendency (Table 5). L values
are really close to the values found in F, increasing moderately
in H; this is evidence of a late carbon liberation process, when
this has been fixed into the tree structures. Nevertheless, when
a certain decomposition level is reached, carbon fraction dimin-
ishes; due to the material liberating the carbon stored in the xylem
and in the structure. In patula pine a small variation exists in the
process, probably because the material goes to an intermediate
decomposition state, diminishing carbon content due slightly to
the presence of resins, which makes the initial liberation of other
elements difficult (Duchaufour, 1994). On arriving at phase H, the
material reaches a higher decomposition state, lightly increasing
the carbon fraction. The total balance of carbon in the soil tends to
concentrate in the first 25 cm. Patula pine constantly accumulates
needles on the surface; while teak, a deciduous species, loses more
than 70% of the foliage during 3 months each year. However teak
reaches 1.59 Mg ha−1 day−1 of C, compared to 0.36 Mg ha−1 day−1
registered in patula pine for the studied period. The use of growth
models in patula pine and teak plantations, allows a calculation of
between 25.7 and 67.5% of the total carbon in the system, which
remain in the trees. The soil provides a considerable percentage of
all the carbon in the system (31.3–73.1%). The remaining percent-
age 0.52–1.25% corresponds to FFC. In spite of the average values
of the different components tending to be similar, the total volume
of biomass of each component is directly related to its importance
in the overall carbon balance within the tropical forestry ecosys-
tems. Pastures tend to accumulate high carbon stocks (Fearnside,
1996), when the trees were planted the SOC stocks in the pat-
ula pine areas were much higher than in the teak stands. What is
more, the stands are relatively young while SOC stocks only change
slowly over time (Fearnside, 1996; Schulp and Verburg, 2009). The
results obtained show that the carbon fraction in the tree has a
dynamic behaviour, whereas in the roots, an average value can be
considered because of its stable carbon content. For inventories of
underground biomass an average value per species and age class
can be used safely, for a correct calculation of tree biomass, growth
models are needed. The carbon values found in this study differ
to the results obtained by other authors on soils and forest floor
(Fearnside, 1996; Nabuurs et al., 2008; Schulp et al., 2008; Vesterdal
et al., 2008) in forestry ecosystems. The carbon content in the study
sites have a high variability and are conditioned by the past soil
uses, before the establishment of the plantation, and are also sub-
ordinate to the actions of the former soil factors and climatological
conditions.
4. Conclusions
The methodology developed incorporates factors such as
intraspecific variability, type, age and site conditions (soil and cli-
matic variability), that are reflected in the variables measured in the
field. The highest organic carbon content values were obtained in
mineral soil followed by tree and forest floor. The low FFC in trop-
ical plantations make it necessary to re-evaluate the importance
of this parameter in forest carbon stock balances. The TC is slightly
higher in teak than in patula pine, however, it is not statistically sig-
nificant. The results show the importance of calculating different
organic carbon fractions in trees due to their dynamic behaviour.
The higher SOC stocks under the P. patula stands compared to the
T. grandis stands could be attributed to the fact that the P. patula
stands have been under pasture before planting the trees. The SOC
stock has a high variability between species, and it is conditioned by
the historical land uses, before the establishment of the plantation.
The use of forest and soil survey information, during a relatively
short period for the estimation of TC in tropical ecosystems, can lead
to a good estimation of tree carbon balance components. In future
sampling or inventory campaigns, the spatial variation of land use
and management history could be incorporated to improve SOC
estimations. These carbon estimations should provide important
information for forest managers and policy makers on a regional
and national scale.
Acknowledgements
The research was funded by ARGOS S.A, El Guásimo S.A. y Caribe
S.A reforest enterprises. We would also like to thank to the technical
staff of the National Centre of Coffee Research (CENICAFE) and the
National University of Colombia (Medellín) for making this research
possible. Our gratitude to Maria Casamitjana and Leigh Sanders for
their improvements to the English used and Nynke Schulps for the
final remarks.
References
Alfsen, K., Bye, T., Glomsroed, S., Wiig, H., 2001. Soil degradation and economic
development in Ghana. Environment and Development Economics 2, 119–143.
Batjes, N., 2008. Mapping soil carbon stocks of Central Africa using SOTER. Geoderma
146, 58–65.
Brown, S., 2002. Measuring carbon in forests: current status and future challenges.
Environmental Pollution 116, 363–372.
Cerri, C.E.P., Coleman, K., Jenkinson, D., Bernoux, M., Victoria, R., Cerri, C.C., 2003.
Modeling soil carbon from forest and pasture ecosystems of Amazon, Brazil.
Soil Science Society of America Journal 67, 1879–1887.
Chave, J., Andalo, C., Brown, S., Cairns, M., Chambers, J., Eamus, D., Fölster, H., Fro-
mard, F., Higuchi, N., Kira, T., 2005. Tree allometry and improved estimation of
carbon stocks and balance in tropical forests. Oecologia 145, 87–99.
Cortés, L.A., Guevara, C.J., Cortés, B.M.A., Palacino, A.M., 1982. Mapa de suelos de
Colombia. Instituto Geográfico Agustín Codazzi (IGAC), Bogotá.
Duchaufour, P., 1994. Pédologie: sol, végétation, environnement, fourth ed. Masson,
París.
Egunjobi, J., 1974. Litter fall and mineralization in a teak Tectona grandis stand. Oikos
25, 222–226.
Elfving, B., Kiviste, A., 1997. Construction of site index equations for Pinus sylvestris
L. using permanent plot data in Sweden. Forest Ecology and Management 98,
125–134.
Fearnside, P., 1996. Amazonian deforestation and global warming: carbon stocks in
vegetation replacing Brazil’s Amazon forest. Forest Ecology and Management
80, 21–34.
Fearnside, P., 2004. Are climate change impacts already affecting tropical forest
biomass? Global Environmental Change A 14, 299–302.
Fearnside, P., Imbrozio Barbosa, R., 1998. Soil carbon changes from conversion of
forest to pasture in Brazilian Amazonia. Forest Ecology and Management 108,
147–166.
Federici, S., Vitullo, M., Tulipano, S., Lauretis, R., de Seufert, G., 2008. An approach
to estimate carbon stocks change in forest carbon pools under the UNFCCC: the
Italian case. iForest: Biogeosciences & Forestry 1, 86–95.
Garcia-Pausas, J., Casals, P., Camarero, L., Huguet, C., Sebastia, M., Thompson, R.,
Romanya, J., 2007. Soil organic carbon storage in mountain grasslands of the
Pyrenees: effects of climate and topography. Biogeochemistry 82, 279–289.
Gillespie, A.J.R., 1992. Pinus patula Schiede and Deppe. Patula pine. Res. Note SO-ITF-
SM-54. USDA Forest Service, Southern Forest Experiment Station, New Orleans.
Hendricks, C., Turkenburg, W., 1997. Towards meeting CO2 emission targets: the
role of carbon dioxide removal. IPTS Report 16, 13–21.
IPCC, 2000. In: Watson, R.T., Noble, I.R., Bolin, B., Ravindranath, N.H., Verardo, D.J.,
Dokken, D.J. (Eds.), Land Use, Land-Use Change, and Forestry. Cambridge Uni-
versity Press, UK.
Janzen, H., 2004. Carbon cycling in earth systems: a soil science perspective. Agri-
culture, Ecosystems & Environment 104, 399–417.
Jaramillo, D.F., Parra, L.N., González, L.H., 1994. El recurso suelo en Colombia:
distribución y evaluación. Instituto de Ciencias Naturales y Ecología (ICNE),
Universidad Nacional de Colombia, Medellín.
 J.C.L. Usuga et al. / Forest Ecology and Management 260 (2010) 1906–1913
Jaramillo, S., Silva, J., Osorio, N.W., 2004. Potencial simbiotico y efectividad de hon-
gos micorrizo arbusculares de tres suelos sometidos a diferentes usos. Revista
Facultad Nacional de Agronomía, Medellín 57, 2205–2217.
Kangur, A., Sims, A., Jõgiste, K., Kiviste, A., Korjus, H., von Gadow, K., 2007. Compara-
tive modeling of stand development in Scots pine dominated forests in Estonia.
Forest Ecology and Management 250, 109–118.
Kiviste, A., Álvarez, J., Rojo, A., Ruíz, A., 2002. Funciones de crecimiento de aplicación
en el ámbito forestal. Instituto Nacional de Investigación Tecnología Agraria
Alimentaria (INIA), Madrid.
Kuikman, P.J., de Groot, W.J.M., Hendriks, R.F.A., Verhagen, J., de Vries, F., 2003. Stocks
of C in Soils and Emissions of CO2 from Agricultural Soils in The Netherlands.
Alterra-rapport 561, Wageningen.
Milne, R., Brown, T., 1997. Carbon in the vegetation and soils of Great Britain. Journal
of Environmental Management 49, 413–433.
Montané, F., Loaiza, J.C., Casals, P., 2006. Shrub encroachment does not decrease
sequestered soil carbon in mountain grassland in Spain. In: Price, M.F. (Ed.),
Global Change in Mountain Regions. Sapiens Publishing, UK, pp. 200–201.
Montero, M., Kanninen, M., 2003. Biomasa y carbono en plantaciones de Terminalia
amazonia en la zona Sur de Costa Rica. Revista Forestal Centroamericana 39,
50–56.
Moraes, J., Cerri, C., Melillo, J., Kicklighter, D., Neill, C., Steudler, P., Skole, D., 1995.
Soil carbon stocks of the Brazilian Amazon basin. Soil Science Society of America
Journal 59, 244–247.
Nabuurs, G, van Putten, B., Knippers, T., Mohren, G., 2008. Comparison of uncer-
tainties in carbon sequestration estimates for a tropical and a temperate forest.
Forest Ecology and Management 256, 237–245.
Paustian, K., Levine, E., Post, W.M., Ryzhova, I.M., 1997. The use of models to integrate
information and understanding of soil C at the regional scale. Geoderma 79,
227–260.
Rigobelo, E., Nahas, E., 2004. Seasonal fluctuations of bacterial population and micro-
bial activity in soils cultivated with eucalyptus and pinus. Scientia Agricola 61,
88–93.
1913
Sales, M., Souza Jr., C., Kyriakidis, P., Roberts, D., Vidal, E., 2007. Improving spatial
distribution estimation of forest biomass with geostatistics: a case study for
Rondonia, Brazil. Ecological Modelling 205, 221–230.
Schlegel, B., Gayoso, J., Guerra, J., 2000. Manual de procedimientos para muestreos
de biomasa forestal. Universidad Austral de Chile, Valdivia.
Schulp, C., Nabuurs, G., Verburg, P., de Waal, R., 2008. Effect of tree species on
carbon stocks in forest floor and mineral soil and implications for soil carbon
inventories. Forest Ecology and Management 256, 482–490.
Schulp, C., Verburg, P., 2009. Effect of land use history and site factors on spa-
tial variation of soil organic carbon across a physiographic region. Agriculture,
Ecosystems & Environment 133, 86–97.
Sierra, C., del Valle, J., Orrego, S., Moreno, F., Harmon, M., Zapata, M., Colorado, G.,
Herrera, M., Lara, W., Restrepo, D., 2007. Total carbon stocks in a tropical forest
landscape of the Porce region, Colombia. Forest Ecology and Management 243,
299–309.
Smith, P., 2004. Soils as carbon sinks: the global context. Soil Use and Management
20, 212–218.
SSS, 1992. Soil Survey Laboratory Methods Manual. Soil Conservation Service U.S.,
Department of Agriculture, Rep 42, Washington, DC.
SSS, 1993. Soil Survey Manual, Soil Survey Division Staff. Soil Conservation Service
U.S., Department of Agriculture, Handbook 18.
Tan, K., Piao, S., Peng, C., Fang, J., 2007. Satellite-based estimation of biomass carbon
stocks for northeast China’s forests between 1982 and 1999. Forest Ecology and
Management 240, 114–121.
Vesterdal, L., Schmidt, I., Callesen, I., Nilsson, L., Gundersen, P., 2008. Carbon and
nitrogen in forest floor and mineral soil under six common European tree
species. Forest Ecology and Management 255, 35–48.
Walker, A., Black, J.A., 1934. Estimation of organic carbon by the chromic acid titra-
tion method. Soil Sciences 37, 29–31.
Weaver, P.L., 1993. Tectona grandis L.f., teak, Res. Note SO-ITF-SM-64. USDA Forest
Service, Southern Forest Experiment Station, New Orleans.