Total aboveground biomass in central Amazonian rainforests:
a landscape-scale study
Henrique E.M. Nascimentoa, William F. Laurancea,b,*
aBiological Dynamics of Forest Fragments Project, National Institute for Amazonian Research (INPA),
C.P. 478, Manaus AM 69011-970, Brazil
bSmithsonian Tropical Research Institute, Apartado 2072, Balboa, Republic of Panama
Received 16 May 2001
Abstract
Amazonian forests play a key role in the global carbon cycle, but there is much uncertainty about the quantity and
distribution of carbon stored in these forests. We quantified total aboveground dry biomass (TAGB) in undisturbed central
Amazonian rainforests, based on detailed estimates of all live and dead plant material within 20 1 ha plots spanning an
extensive (ca. 1000 km2) study area. TAGB values in our study area were very high, averaging 397:7 
 30:0 Mg ha1. The
most important component of aboveground biomass was large (10 cm diameter-at-breast-height (DBH)) trees, which
comprised 81.9% of TAGB, followed by downed wood debris (7.0%), small trees, saplings, and seedlings (<10 cm DBH;
5.3%), lianas (2.1%), litter (1.9%), snags (1.5%), and stemless palms (0.3%). Among large trees, aboveground biomass
was greatest in intermediate-sized (20?50 cm DBH) stems (46.7% of TAGB), with very large (60 cm DBH) trees also
containing substantial biomass (13.4% of TAGB). There were no significant correlations between large tree biomass
and that of any other live or dead biomass component. An analysis based on the variability of our samples suggested
that just 3?4 randomly positioned 1 ha plots would be sufficient to provide a reasonable estimate of mean TAGB in a
landscape such as ours (with 95% confidence intervals being <10% of the mean). This suggests that efforts to quantify
Amazon forest biomass should be extensive rather than intensive; researchers should sample many geographically separate
areas with a few plots each, rather than sampling a small number of areas more intensively.
Published by Elsevier Science B.V.
Keywords: Amazon basin; Biomass; Carbon cycle; Carbon storage; Global warming; Terra-firme forest; Tropical rainforest
1. Introduction
The rapid conversion of tropical forests is a major
source of greenhouse gases such as carbon dioxide,
methane, and nitrous oxide, which are the principle
causes of global warming (Houghton, 1991; Fearnside,
2000). Nowhere is deforestation occurring more
rapidly than in Brazilian Amazonia, which contains
about 40% of the world?s remaining tropical rain-
forests (Laurance et al., 2001a). From 1995 to 1999,
deforestation rates in the Brazilian Amazon averaged
nearly 2 million hectares per year (INPE, 2000),
not including extensive forest areas degraded by
logging, ground fires, forest fragmentation, illegal
Forest Ecology and Management 168 (2002) 311?321
* Corresponding author. Present address: Biological Dynamics of
Forest Fragments Project, National Institute for Amazonian
Research (INPA), C.P. 478, Manaus AM 69011-970, Brazil.
Tel.: ?507-212-8252; fax: ?507-212-8148.
E-mail addresses: laurancew@tivoli.si.edu, wfl@inpa.gov.br
(W.F. Laurance).
0378-1127/02/$ ? see front matter. Published by Elsevier Science B.V.
PII: S 0 3 7 8 - 1 1 2 7 ( 0 1 ) 0 0 7 4 9 - 6
gold-mining, and overhunting (Skole and Tucker,
1993; Laurance, 1998; Cochrane et al., 1999;
Nepstad et al., 1999).
Despite the important role of Amazonian forests in
the global carbon cycle (Fearnside, 1997a), data on
biomass and carbon storage in the region are clearly
inadequate (e.g. Brown et al., 1992a,b, 1995; Laurance
et al., 1999). Biomass estimates for Amazonian
forests have been the subject of considerable debate
(e.g. Brown and Lugo, 1984, 1992; Fearnside, 1985,
1986, 1992) because of limited data, methodological
differences among investigators, and small or incom-
plete measurements of biomass in some studies
(cf. Houghton et al., in press). Current estimates of
total carbon storage in Brazilian Amazonia vary by
more than a factor of 2, from 39 to 93 Pg C, largely as a
result of uncertainty in the quantity and spatial
distribution of forest biomass (Houghton et al., in
press). Houghton et al. (2000) concluded that 60% of
the uncertainty in their estimates of annual carbon
flux from Brazilian Amazonia resulted from varying
estimates of forest biomass. Clearly, there is a need
for additional measurements of biomass across large
expanses of the Amazon basin.
In a recent study, Laurance et al. (1999) assessed the
relationship between soil features and aboveground
biomass of live trees for 65 1 ha plots arrayed across
an extensive (ca. 1000 km2) central Amazonian
landscape. Biomass estimates in this study were
generated by measuring diameters of all large
(10 cm diameter-at-breast-height (DBH)) trees and
then using a correction factor to approximate biomass
of small (<10 cm DBH) trees. Here we provide a much
more comprehensive estimate of aboveground bio-
mass for the same study area, based on 20 randomly
selected 1 ha plots in which biomass of all live
and dead material (large trees, small trees, seedlings,
palms, lianas, downed wood debris, snags, litter,
aboveground root mat) was quantified. We also use an
improved allometric model to estimate the biomass of
large trees.
Our study had three goals: (1) To provide rigorous
estimates of total aboveground dry biomass (TAGB)
for intact forests in our study area; (2) To develop
better correction factors for studies in which only large
trees are measured; (3) To predict the number of 1 ha
samples needed to reliably estimate aboveground
biomass in these forests.
2. Methods
2.1. Study area and plots
The study area is a partially fragmented landscape
in the central Amazon, 80 km N of Manaus, Brazil
(28300S, 608W) at 50?100 m elevation (Lovejoy et al.,
1986). Rainforests in the area are terra-firme (not
seasonally inundated). Rainfall ranges from 1900 to
3500 mm annually with a pronounced dry season from
June to October. The forest canopy is 30?37 m tall,
with emergents to 55 m. Species richness of trees is
very high and can exceed 280 species (10 cm DBH)
per hectare (de Oliveira and Mori, 1999).
The dominant soils in the study area are xanthic
ferralsols, which are heavily weathered, acidic, and
very poor in nutrients such as P, Ca and K (Chauvel
et al., 1987). Similar nutrient-poor soils are prevalent
throughout much of the Amazon basin (Richter and
Babbar, 1991). Cation concentrations tend to be higher
in more clayey soils, which are prevalent in flatter
areas and ridgetops; these areas generally support
greater tree biomass than do gullies and slopes, which
have higher sand contents and lower cation concen-
trations (Laurance et al., 1999).
To conduct inventories of (TAGB), we sampled 20
1 ha plots located in large (>106 ha) tracts of intact
forest ?n ? 18 plots? and in the interior of a 100 ha
forest fragment ?n ? 2 plots?. The 20 plots were
selected using stratified random sampling to ensure
that they spanned the length and breath of our
1000 km2 study area and that pseudoreplication (i.e.
having multiple plots in the same general area) was
minimised. All plots were located in forest interiors
(>300 m from the nearest forest-pasture edge) to
eliminate any influence of edge effects on forest
biomass and dynamics (cf. Laurance et al., 1997,
1998a,b, 2000).
2.2. Large trees
For each plot, complete inventories of all large
(10 cm DBH) trees were conducted from 1997 to
1999. Large (10 cm DBH) palms were included in
these samples but are rare in our study plots (<1% of
stems, most of which are a single species [Oenocarpus
bacaba]). Tree diameters were carefully measured (to
the nearest 1 mm) using a DBH tape at 1.3 m height or
312 H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321
above any buttresses. DBH measurements were
converted to estimates of aboveground dry biomass
with an allometric model derived from 315 trees from
local rainforests (Chambers et al., 2001b; Table 1).
The trees used to derive this model ranged from 5 to
120 cm DBH and were destructively sampled at a site
ca. 20 km southwest of our study area, in very similar
lowland terra-firme forest. The allometric model of
Chambers et al. (2001b) is a refinement of an earlier
model by dos Santos (1996), which overestimates the
biomass of large (60 cm DBH) trees while under-
estimating biomass of smaller trees.
2.3. Lianas
Aboveground dry biomass of lianas (woody vines)
was estimated in the 20 plots from 1997 to 1999
(Laurance et al., 2001b). For each plot, diameters of all
liana stems (2 cm DBH) were carefully measured at
1.3 m height with a DBH tape (to the nearest 0.1 mm),
then converted to a biomass estimate using an
allometric formula developed in an Amazonian rain-
forest in Venezuela (Putz, 1983; Table 1).
2.4. Small trees, palms, and small lianas
Data on small trees (including saplings and seed-
lings), palms, and small lianas were collected from
1999 to 2000 within a series of subsamples in each plot.
All data were converted to aboveground dry biomass
estimates on a per-hectare basis.
The DBH of small trees and palms (DBH from 5.0
to 9.9 cm) were measured in 13 quadrats
?20 m 	 20 m? per plot. To accomplish this each plot
was divided into 25 quadrats, and every second
quadrat was sampled (in a checker-board pattern) to
ensure nearly uniform coverage of the plot. DBH data
were converted to aboveground dry biomass estimates
using the formulas of Chambers et al. (2001b) for
primary-forest trees, Nelson et al. (1999) for succes-
sional trees, and Hughes (1997) for palms (Table 1).
The models of Nelson et al. (1999) were developed
using trees from our study area, while that of Hughes
(1997) was for Mexican tropical rainforests.
DBH data for saplings (1.0?4.9 cm DBH), stemless
palms, and small lianas (1.0?2.0 cm DBH) were
recorded within 13 sub-quadrats ?5 m 	 5 m? per plot.
One sub-quadrat was nested within the southwest
corner of each 20 m 	 20 m quadrat. Aboveground dry
biomass estimates were generated using the formulas
of Nelson et al. (1999) for successional-tree species
and Putz (1983) for lianas. For primary-forest
saplings, 40 individuals were randomly selected at
six different sites in our study area, then cut at ground
level and weighed. Samples of the trunk, branches,
and leaves were removed, oven-dried, and reweighed,
in order to estimate aboveground dry biomass for
each individual. These data were used to derive an
Table 1
Allometric formulas used to estimate aboveground dry biomass (kg ha1) for rainforest plants in this study, based on DBH or basal-area (BA)
dataa
Plant group Formula
Large (5 cm DBH) treesb expf0:370 ? 0:333?ln?DBH?
?0:933?ln?DBH?2  0:122?ln?DBH?3g
Small (1?5 cm DBH) primary treesc exp?1:7689 ? 2:3770 ln?DBH?
Small (1?10 cm DBH) successional treesd exp?2:4128 ln?DBH?  1:9968
Small (1?10 cm DBH) Cecropia treesd exp?2:4257 ln?DBH?  2:5118
Small (1?10 cm DBH), stemmed palmse fexp?0:9285 ln?DBH2? ? 5:72361:05001g=103
Lianasf base10?0:12 ? 0:91 log10?BA?
a ??exp?? indicates that the natural logarithm ?e ? 2:71828? is raised to the power indicated in the formula (e.g. exp?0:35DBH? ? e0:35DBH),
whereas ??base10?? indicates that the base-10 logarithm is used (e.g. base10?0:35DBH? ? 100:35DBH?.
b Chambers et al. (2001b).
c This study.
d Nelson et al. (1999).
e Hughes (1997).
f Putz (1983).
H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321 313
allometric formula to predict biomass based on
sapling DBH (R2 ? 96:2%; Table 1). For abundant
stemless palms (principally Attalea spp. and Astro-
caryum spp.), leaves of individual plants were divided
into three size-classes (0?2, 2?3 and >3 m height). The
number of leaves in each size-class was counted and
multiplied by an average leaf weight, derived by
randomly sampling 10 leaves in each size-class that
were oven-dried and weighed.
Seedling aboveground dry biomass was estimated
by counting seedlings (>50 cm height but <1 cm
DBH) within 13 samples ?2 m 	 2 m? per plot, with
one sample located in the southwest corner of each
sub-quadrat. Seedling counts were multiplied by
average seedling weight, derived by randomly sam-
pling 60 seedlings that were oven-dried and weighed.
2.5. Wood debris
Downed wood debris was non-destructively quan-
tified in 1999?2000 using the planar-intersect method
(Van Wagner, 1968; Brown, 1974). Fine (2.5?9.9 cm
diameter) and coarse (10 cm diameter) wood debris
fractions were estimated separately. Within each
20 m 	 20 m quadrat, two 15 m long perpendicular
transects were established (yielding 26 transects per
plot). For each transect, fine debris was sampled along
the first 5 m of the sample plane while coarse debris
was sampled along the entire plane. All wood particles
that intersected the planes were measured. Coarse
debris was separated into sound and rotten classes,
using criteria such as the condition of sapwood and
heartwood (Delaney et al., 1998). Dry necromass
was estimated with formulas in Brown (1974),
using wood densities of 0.69 g cm3 for sound coarse
debris (mean wood density for Amazonian forests;
Fearnside, 1997b), 0.34 g cm3 for rotten coarse
debris (Cummings, 1998) and 0.41 g cm3 for fine
debris (Cummings, 1998).
2.6. Snags
All plots were systematically surveyed on foot in
1999?2000 to identify trees that died but remained
standing. Dry necromass was estimated using the
formula of Chambers et al. (2001b), with values
reduced by 10% to compensate for the loss of leaves,
twigs and small branches (cf. Delaney et al., 1998). A
small proportion of snags (13%) had broken boles,
and for these a taper function was used to estimate
tree volume (volume ? basal area 	 estimated bole
height 	 0:78; Grac?a et al., 1999), which was then
multiplied by the mean density of sound wood
(0.69 g cm3) to yield a necromass estimate.
2.7. Litter standing crop
In each 20 m 	 20 m quadrat, two 50 m 	 50 cm
samples (26 samples/plot) were removed to determine
the biomass of litter (leaves, twigs, fruits) on the forest
floor (this included fine live roots above the soil
surface). These samples were located at the midpoint
of the southern and western margins of each quadrat.
Samples were collected during the 2001 wet season
(January?March) and weighed in the field. A portion
of each sample was then oven-dried, and the ratio of
dry to wet litter weight was used to estimate litter dry
mass.
2.8. Required sample size
We estimated the number of 1 ha samples needed to
adequately assess TAGB in our landscape as a function
of confidence-interval size, using the formula of Ott
and Mendenhall (1990, p. 504): N ? ?z2 s2?=E2,
where z depends on the desired confidence interval
(i.e. z ? 1:645 for 90% confidence interval, 1.96 for
95% confidence interval, etc.), s2 is the sample
variance and E the tolerable error in Mg ha1.
Required values for N were then plotted as a function
of tolerable error.
3. Results
3.1. Total biomass
TAGB in the 20 plots ranged from 305 to
432 Mg ha1, with a mean of 397:7 
 30:0 Mg ha1
(Table 2). When a data distribution was generated,
TAGB values were skewed to the left (Fig. 1); most
(85%) plots had values ranging from about 380?
430 Mg ha1, but three plots had relatively low values
(305?364 Mg ha1).
On average, large (10 cm DBH) trees comprised
81.9% of TAGB, followed by coarse downed wood
314 H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321
Table 2
Components of aboveground dry biomass for 20 1 ha plots in central Amazonian rainforesta
Plot Large trees Small trees Stemless palms Lianas Coarse debris Fine debris Snags Litter Total
1101 346.23 18.50 1.90 9.33 23.29 3.35 9.27 7.90 419.77
1102 255.28 24.02 1.48 6.10 28.74 2.74 12.57 7.72 338.64
1103 299.90 25.25 1.11 9.43 13.67 2.30 4.30 7.63 363.58
1113 359.60 21.85 2.23 4.63 30.14 2.91 1.60 6.73 429.69
1201.1 307.72 20.57 1.14 7.72 29.18 1.59 4.22 10.22 382.35
1201.2 335.72 25.28 1.02 9.41 18.79 2.72 6.80 6.06 405.80
1201.3 238.11 18.52 1.20 10.90 23.50 1.53 3.16 8.05 304.98
1301.1 327.20 19.67 1.37 6.69 33.06 4.47 1.95 7.20 401.60
1301.5 340.74 18.44 2.25 6.41 35.81 3.05 2.19 5.82 414.71
1301.7 329.58 25.47 1.39 8.05 36.15 3.47 6.09 7.16 417.34
1301.9 323.29 19.95 1.24 6.48 34.12 3.43 4.55 6.65 399.69
1501.1 348.29 18.78 1.81 7.94 14.13 3.19 8.58 7.07 409.78
1501.2 349.92 18.08 1.00 9.28 21.05 4.10 8.31 9.87 421.61
1501.3 334.37 19.45 0.85 9.04 25.53 3.26 8.18 8.73 409.41
2303.5 352.61 22.80 1.49 5.87 25.65 2.72 13.40 7.18 431.72
2303.6 352.36 22.32 1.12 5.75 11.91 3.00 7.33 8.93 412.74
3402.1 338.88 19.42 1.46 9.58 18.76 4.00 2.26 6.69 401.03
3402.3 317.11 21.42 0.32 9.07 31.30 3.05 10.81 5.93 399.01
3402.5 329.42 19.05 0.80 10.68 21.99 3.61 2.03 7.29 394.86
3402.7 323.89 23.28 0.43 13.67 19.44 2.61 5.29 6.92 395.25
Mean 325.51 21.11 1.28 8.30 24.81 3.05 6.14 7.49 397.68
a Biomass estimates (Mg ha1) are shown for large (10 cm DBH) trees, small trees (including saplings and seedlings; <10 cm DBH),
stemless palms, lianas (1 cm DBH), coarse downed wood debris (10 cm diameter), fine downed wood debris (2.5?9.9 cm diameter), snags,
and litter (leaves, fruits, flowers, fine aboveground roots).
Fig. 1. Frequency distribution for TAGB in 20 1 ha plots in intact rainforests of the central Amazon.
H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321 315
debris (6.2%), small trees, saplings, and seedlings
(5.3%), lianas (2.1%), litter (1.9%), snags (1.5%), fine
downed wood debris (0.8%), and stemless palms
(0.3%). Nearly nine-tenths of TAGB (89.6%) was live
material, with the remainder comprised by wood
debris, snags and litter (Table 2).
Among large trees, most of the aboveground
biomass was concentrated in abundant, intermedi-
ate-sized (20?50 cm DBH) stems, which contained
46.7% of TAGB (Table 3). Although uncommon
(averaging <10 stems ha1), very large canopy and
emergent trees (60 cm DBH) contained 13.4% of
TAGB, with the remainder (12.2%) in the smallest
(10?20 cm DBH) size-class.
3.2. Correlations among biomass components
Surprisingly, there was no significant correlation
between the biomass of large (10 cm DBH) trees and
that of any other biomass component (small trees/
seedlings, lianas, stemless palms, downed wood
debris, snags, litter) (r < 0:25, P > 0:30 in all cases).
There was also no significant relationship between
large-tree biomass and all other live-plant biomass
?r ? 0:24?, total necromass ?r ? 0:06?, or all
other live-plant biomass and necromass combined
?r ? 0:16? (P > 0:30 in all cases; Pearson correla-
tions).
3.3. Estimating TAGB from large-tree biomass
Our results suggest that TAGB can be reliably
estimated using measurements of large trees alone.
Large trees comprised over 80% of TAGB and, not
surprisingly, there was a positive, linear relationship
between large-tree biomass and TAGB (TAGB ?
67:7 ? 0:981 large-tree biomass; R2 ? 94:3%,
P < 0:00001; linear regression analysis). According
to our data (Table 2), TAGB estimates can be derived by
increasing large-tree biomass by a correction factor of
22.2% (6.5% for small trees, saplings, and seedlings;
3.0% for lianas and stemless palms; 12.7% for dead
material). When we applied this correction, there was
very good agreement between measured and estimated
TAGB (Fig. 2). The average difference for our 20 plots
was just 2.1% (range 0.1?7.9%).
3.4. Predicting adequate sample size
The analysis of tolerable error (Fig. 3) suggests that,
for a landscape such as ours, three randomly
positioned 1 ha plots would provide a reasonable
estimate of mean TAGB if all components of above-
ground biomass are directly measured. Ninety-five
percent confidence intervals based on three plots
would be 
35 Mg, which represent errors of <9% in
magnitude (i.e. because TAGB averaged nearly
400 Mg ha1), whereas errors for four plots would
Table 3
Aboveground biomass, number of stems, and percentage of TAGB
for seven diameter-classes of large (10 cm DBH) trees in the
central Amazon. Data shown are mean values (
S.D.) for 20 1 ha
plots
Size-class
(cm)
Biomass
(Mg ha1)
Stems
(ha1)
Percentage
of TAGB
10?20 48:61 
 3:79 394.40 
 27.46 12.22
20.1?30 64:49 
 6:60 126.45 
 13.64 16.22
30.1?40 64:35 
 11:49 53.75 
 8.92 16.18
40.1?50 57:02 
 12:60 27.00 
 5.86 14.34
50.1?60 37:91 
 10:19 11.95 
 3.09 9.53
60.1?70 22:07 
 7:93 5.05 
 1.66 5.55
>70 31:05 
 12:76 4.70 
 2.00 7.81
Fig. 2. Relationship between measured TAGB and estimated TAGB
values (based on large-tree (10 cm DBH) biomass plus a 22.2%
correction factor to account for the other biomass components). The
line shows Y ? X.
316 H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321
be 
30 Mg (7.5% in magnitude). If TAGB is esti-
mated by measuring large trees only, then four plots
would be advisable to ensure that total sampling errors
(7.5% tolerable error plus 2.1% average error from the
correction factor) are <10%.
The necessary sample sizes rise rapidly if higher
levels of sample accuracy are needed (Fig. 3). For
example, to achieve 95% confidence intervals of

20 Mg (ca. 5% error), nine plots would be needed.
Likewise, 36 plots would be needed to achieve
confidence intervals of 
10 Mg, and 144 plots for
intervals of 
5 Mg.
4. Discussion
4.1. Total biomass
Our study provides the most comprehensive
estimate presently available for TAGB in central
Amazonian rainforests. A key conclusion is that
biomass in these forests is very high, averaging nearly
400 Mg ha1. It is notable that this estimate was
produced using an improved but relatively conserva-
tive allometric model (Chambers et al., 2001b) to
estimate the biomass of large (10 cm DBH) trees.
Our value is higher than most comparable estimates in
other parts of the Amazon (e.g. Klinge and Rodriguez,
1973; Uhl and Jordan, 1984; Uhl et al., 1988; Brown
and Lugo, 1990, 1992; Brown et al., 1992a,b, 1995;
Saloma?o et al., 1998), but clearly supports recent
estimates of high biomass for central Amazonian
forests (Fearnside, 1997a; Laurance et al., 1999).
Belowground biomass was not measured in this
study but can be estimated from published works. In
eight studies in the Neotropical and Caribbean regions
reviewed by Houghton et al. (in press), belowground
(root) biomass averaged 21.0% (range 13?34%) of
aboveground biomass. Using this mean value, root
biomass in our plots would be 84 Mg ha1 on average.
This estimate is reduced slightly if only studies in
Amazonian terra-firme forests are considered (Klinge
and Rodriguez, 1973; Jordan and Uhl, 1978; Saldar-
riaga et al., 1988; Nepstad, 1989); these averaged
17.5% (range 13.4?23.5%) of aboveground biomass.
In this case, belowground biomass of roots would
average 70 Mg ha1 in our plots.
In addition to roots, Amazonian forests store large
amounts of belowground carbon in the form of soil
organic matter. Based on 1162 soil profiles throughout
the Brazilian Amazon from the RADAMBRASIL
project, Moraes et al. (1995) found that carbon density
(excluding live roots) in the upper 1 m of soil averaged
103 Mg ha1. Remarkably, Davidson and Trumbore
(1995) found an additional 200 Mg ha1 of carbon in
soil organic matter between 1 and 8 m depth in the
eastern Amazon.
Collectively, these studies illustrate the high
capacity of intact Amazonian forests for carbon
storage. Assuming that 50% of biomass is carbon,
aboveground carbon storage in our study area is nearly
200 Mg ha1, with perhaps another 35?40 Mg ha1 as
belowground biomass and roughly 300 Mg ha1 as
soil organic matter. Forest conversion is unlikely to
have major effects on deep soil organic matter, but it
can dramatically reduce carbon storage aboveground
and in the roots and upper soil layers (Kauffman et al.,
1994, 1995; Fearnside, 1997a; Fearnside and Barbosa,
1998).
4.2. Components of aboveground biomass
In our 20 plots, large (10 cm DBH) trees comprised
an average of 81.9% of TAGB, with the remainder
divided between other live plants (7.7%) and dead
material (10.4%). In tropical forests, large trees are
inevitably the dominant component of aboveground
Fig. 3. Relationship between sample size (number of 1 ha plots)
and tolerable error for estimating TAGB in central Amazonian
rainforests. The curve was fitted by a smoothing function.
H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321 317
biomass. Indeed, in some forests up to a quarter of
aboveground biomass is concentrated in the largest
canopy and emergent trees (Clark and Clark, 1995),
although in our study area nearly half of all biomass
(46.7%) was in the intermediate-size-classes (20?
50 cm DBH), with a lower proportion (13.4%)
contained in the largest trees (60 cm DBH).
Understory plant biomass appears moderately low
in central Amazonian forests. Although the forest
understory is often quite dense (with many saplings
and stemless palms), herbs, epiphytes, and climbing
vines are less abundant than in many other neotropical
forests (Guillaumet, 1987; Gentry, 1990). In our study
area, for example aboveground biomass of lianas was
lower, and comprised a smaller percentage of total
forest biomass, than in other neotropical (Klinge and
Rodriguez, 1974; Putz, 1983; Pe?rez-Salicrup, 1998),
Asian (Ogawa et al., 1965; Kato et al., 1978), and
African (Greenland and Kowai, 1960) forests. This
low density of herbs, epiphytes, and lianas may result
from extremely low soil fertility (Gentry and
Emmons, 1987), and disturbance favouring lianas
could also be limited by naturally low rates of tree
mortality (1:3 
 0:5% yr1 over the past two decades)
in our study area (Laurance et al., 2001b).
Estimates of standing stocks of wood debris are
important for validating carbon cycling models
(e.g. Chambers et al., 2000). In our study area, coarse
wood debris (including snags) averaged 31:0

7:7 Mg ha1, which is higher than other estimates
for dense forests of the central Amazon (10.5?
21.2 Mg ha1; Martius, 1997) but comparable to that
in similar undisturbed forest about 20 km southwest of
our study area (29:7 
 12:2 Mg ha1; Summers,
1998). An empirical model incorporating long-term
tree mortality and decomposition rate data predicts
that standing stocks of biomass in coarse debris should
be about 10% of that contained in large (10 cm
DBH) trees (Chambers et al., 2001a). Based on our
data, the Chambers model predicts that coarse debris
should average 32.6 Mg ha1, which is very close to
our observed mean of 31.0 Mg ha1.
It is interesting that we found no significant
correlations between large-tree biomass and the other
live and dead biomass components. One might expect,
for example that large-tree biomass would be
negatively correlated with liana biomass (cf. Laurance
et al., 2001b) because lianas tend to suppress tree
growth and survival (Putz, 1984), or with wood debris
and snags, because disturbed sites might have fewer
large trees and more debris. The absence of such
correlations might have resulted because tree mortal-
ity rates were generally low in our plots and because
large-tree biomass did not span a particularly large
range of values. If fragmented or logged forests were
included in this study, tree mortality would have been
much higher and correlations among various biomass
components might well have become evident.
4.3. Developing correction factors
Our data suggest that TAGB of central Amazonian
terra-firme forests can be effectively approximated by
increasing biomass estimates for large (10 cm DBH)
trees by 22.2%. Such correction factors have been
used by many investigators (e.g. Brown and Lugo,
1984, 1992; Fearnside, 1992; Brown et al., 1995), as it
is far easier to measure only large trees than all live
and dead components of biomass. For our 20 plots, the
mean difference between estimated and measured
aboveground biomass was just 2.1% (Fig. 2).
Our correction factor incorporates aboveground
mass estimates for all significant dead and live
material aside from large trees. Estimates of dead
material have varied considerably among studies,
ranging from 2.3 to 16.6% of aboveground live
biomass at 15 neotropical sites (Houghton et al., in
press), although the overall mean (8.7%) did not differ
greatly from that in this study (10.4%). As noted
above, understory plant biomass is relatively low in
central Amazonia; small trees (<10 cm DBH), for
example comprised 12% of large-tree biomass in the
Rio Negro region of Venezuela (Jordan and Uhl, 1978)
but only 6.5% in this study.
4.4. Sampling biomass in Amazonian forests
Our results suggest that, for a landscape such as
ours, three randomly positioned plots will provide a
reasonable degree of sample accuracy for estimating
TAGB (Fig. 3). With just three plots, tolerable errors
for the estimated mean (using 95% confidence
intervals) were 35 Mg ha1, or under 9% in magni-
tude. If TAGB is estimated using large (10 cm DBH)
trees only, then four plots would be advisable to limit
overall sampling errors to ca. 10% in magnitude. This
318 H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321
result is fortunate, in the sense that a limited number of
plots appears to provide a reasonably good estimate of
average aboveground biomass at a landscape scale.
This implies that efforts to quantify Amazon forest
biomass should be extensive rather than intensive;
researchers should sample a large number of geo-
graphically separate areas with a few plots each, rather
than attempting to sample a small number of areas
more intensively.
Our recommendation that 3?4 random samples will
normally be sufficient for sampling Amazon forest
biomass is based on the assumption that 1 ha plots
will be used; for smaller plots, sample variances
would undoubtedly be higher (Clark and Clark, 2000),
especially for large trees (Brown et al., 1995),
meaning that more plots would be needed. We also
assume that biomass in other Amazonian landscapes is
not more variable spatially than that in our study area.
Although few comparable data are available (cf. Clark
and Clark, 2000; for a study at a 600 ha scale in Costa
Rica), we believe this assumption is reasonable except
where forest disturbances, edaphic features, or climate
(e.g. at forest-savanna ecotones) vary strongly across a
particular study area. Finally, methodological differ-
ences among investigators will undoubtedly compli-
cate efforts to quantify the spatial distribution of forest
biomass (cf. Houghton et al., in press). Because most
investigators work in a particular area, such differ-
ences will tend to overestimate among-site differ-
ences. Clearly, efforts to standardise biomass-
estimation methods (e.g. Brown, 1997; Clark and
Clark, 2000) will be just as important as expanding the
geographic distribution of study sites across the
Amazon basin.
4.5. Urgency of the problem
Amazon forests contain a very large stock of
terrestrial carbon (Fearnside, 1997a; Houghton et al.,
in press) and are probably an important carbon sink
(Grace et al., 1995; Phillips et al., 1998; Malhi et al.,
2000; Chambers et al., 2001a). Thus, understanding
the spatial distribution of biomass in these forests is
an urgent priority. Our detailed survey indicates that,
at least in the central Amazon, carbon storage in
undisturbed forests is very high, suggesting that forest
conversion is likely to be an even larger source of
greenhouse gas emissions than previously anticipated.
This is important because plans are well underway to
expand networks of highways, railroads, gas lines, and
other major infrastructure projects in the Brazilian
Amazon. By opening large expanses of the basin?s
remote interior to exploitative activities, these projects
are predicted to sharply accelerate the pace of forest
conversion and degradation (Carvalho et al., 2001;
Laurance et al., 2001a). Agricultural practices in
the Amazon are limited by very low soil fertility
(Kauffman et al., 1994; Fearnside and Leal-Filho,
2001), and the economic value of intact forests for
carbon storage, biodiversity conservation, and other
natural ecosystem services may ultimately be far
greater than that provided by forest conversion.
Acknowledgements
We thank Boone Kauffman for very helpful discus-
sion and Jeff Chambers and two anonymous referees
for comments on the manuscript. The NASA-LBA
program, Mellon Foundation, National Institute for
Amazonian Research (INPA), and Smithsonian Insti-
tution provided research support. This is publication
number 356 in the BDFFP technical series.
References
Brown, J.K., 1974. Handbook for Inventorying Downed Woody
Material. USDA Forest Service, Ogden, UT, 25 pp.
Brown, S., 1997. Estimating Biomass and Biomass Change of
Tropical Forests: A Primer. Forestry Paper 134, FAO, Rome.
Brown, S., Lugo, A.E., 1984. Biomass of tropical forests: a new
estimate based on forest volumes. Science 223, 1290?1293.
Brown, S., Lugo, A.E., 1990. Biomass estimates for Brazil?s
Amazonian moist forests. In: Forest?90: Annals of the First
International Symposium on Environmental Studies on Tropical
Rain Forests, Manaus, Brazil, pp. 46?52.
Brown, S., Lugo, A.E., 1992. Aboveground biomass estimates for
tropical moist forests of the Brazilian Amazon. Interciencia 17,
8?18.
Brown, S., Lugo, A.E., Iverson, L.R., 1992a. Processes and lands
for sequestering carbon in the tropical forest landscape. Water
Air Soil Pollut. 64, 139?155.
Brown, I.F., Nepstad, D.C., Pires, O., Luz, L.M., Alechandre, A.S.,
1992b. Carbon storage and land-use in extractive reserves,
Acre, Brazil. Environ. Conserv. 19, 307?315.
Brown, I.F., Martinelli, L.A., Thomas, W.W., Moreira, M.Z.,
Ferreira, C.A., Victoria, R.A., 1995. Uncertainty in the biomass
of Amazonian forests: an example from Rondo?nia, Brazil.
Forest Ecol. Mgmt. 75, 175?189.
H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321 319
Carvalho, G., Barros, A.C., Moutinho, P., Nepstad, D.C., 2001.
Sensitive development could protect the Amazon instead of
destroying it. Nature 409, 131.
Chambers, J.Q., Higuchi, N., Schimel, J.P., Ferreira, L.V., Melack,
J.M., 2000. Decomposition and carbon cycling of dead trees in
tropical forests of the central Amazon. Oecologia 122, 380?
388.
Chambers, J.Q., Higuchi, N., Tribuzy, E.S., Trumbore, S.E., 2001a.
Carbon sink for a century. Nature 410, 429.
Chambers, J.Q., dos Santos, J., Ribeiro, R.J., Higuchi, N., 2001b.
Tree damage, allometric relationships, and above-ground net
primary production in central Amazon forest. Forest Ecol.
Mgmt., in press.
Chauvel, A., Lucas, Y., Boulet, R., 1987. On the genesis of the soil
mantle of the region of Manaus, Central Amazonia, Brazil.
Experientia 43, 234?240.
Clark, D.B., Clark, D.A., 1995. Abundance, growth and mortality
of very large trees in neotropical lowland rain forest. Forest
Ecol. Mgmt. 80, 235?244.
Clark, D.B., Clark, D.A., 2000. Landscape-scale variation in forest
structure and biomass in a tropical rain forest. Forest Ecol.
Mgmt. 137, 185?198.
Cochrane, M.A., Alencar, A., Schulze, M.D., Souza, C.M.,
Nepstad, D.C., Lefebvre, P., Davidson, E., 1999. Positive
feedbacks in the fire dynamics of closed canopy tropical
forests. Science 284, 1832?1835.
Cummings, D.L., 1998. Total aboveground biomass and structure
of tropical forest delineated by Projeto RADAMBRASIL in
northern Rondo?nia, Brazil. M.Sc. Thesis. Oregon State
University, Corvallis, OR.
Davidson, E.A., Trumbore, S.E., 1995. Gas diffusivity and the
production of CO2 in deep soils of the eastern Amazon. Tellus
B 47, 550?565.
Delaney, M., Brown, S., Lugo, A.E., Torres-Lezama, A., Quintero,
N.B., 1998. The quantity and turnover of dead wood in
permanent forest plots in six life zones of Venezuela.
Biotropica 30, 2?11.
de Oliveira, A.A., Mori, S.A., 1999. A central Amazonian terra
firme forest. I. High tree species richness on poor soils.
Biodivers. Conserv. 8, 1219?1244.
dos Santos, J., 1996. Ana?lise de modelos de regeressa?o para estimar
a fitomassa de floresta tropical u?mida de terra-firme da
Amazo?nia Brasileira. Ph.D. Thesis. Universidade Federal de
Vic?osa, Vic?osa, Minas Gerais, Brazil.
Fearnside, P.M., 1985. Brazil?s Amazon forest and the global
carbon problem. Interciencia 10, 179?186.
Fearnside, P.M., 1986. Brazil?s Amazon forest and the global
carbon problem: reply to Lugo and Brown. Interciencia 11, 58?
64.
Fearnside, P.M., 1992. Forest biomass in Brazilian Amazonia:
comments on the estimate by Brown and Lugo. Interciencia 17,
19?27.
Fearnside, P.M., 1997a. Greenhouse gases from deforestation in
Brazilian Amazonia: net committed emissions. Climatic
Change 35, 321?360.
Fearnside, P.M., 1997b. Wood density for estimating forest biomass
in Brazilian Amazonia. Forest Ecol. Mgmt. 90, 59?87.
Fearnside, P.M., 2000. Global warming and tropical land-use
change: greenhouse gas emissions from biomass burning,
decomposition and soils in forest conversion, shifting cultiva-
tion and secondary vegetation. Climatic Change 46, 115?158.
Fearnside, P.M., Barbosa, R.I., 1998. Soil carbon changes from
conversion of forest to pasture in Brazilian Amazonia. Forest
Ecol. Mgmt. 108, 147?166.
Fearnside, P.M., Leal-Filho, N., 2001. Soil and development in
Amazonia: lessons from the biological dynamics of forest
fragments project. In: Bierregaard, R.O., Gascon, C., Lovejoy,
T.E., Mesquita, R. (Eds.), Lessons from Amazonia: The
Ecology and Conservation of a Fragmented Forest. Yale
University Press, New Haven, CT, in press.
Gentry, A.H., 1990. Floristic similarities and differences between
southern central America and upper and central Amazonia. In:
Gentry, A.H. (Ed.), Four Neotropical Rainforests. Yale
University Press, New Haven, CT, pp. 141?157.
Gentry, A.H., Emmons, L.H., 1987. Geographic variation in
fertility and composition of the understory of neotropical
forests. Biotropica 19, 216?227.
Grac?a, P.M.L.A., Fearnside, P.M., Cerri, C.C., 1999. Burning of
Amazonian forest in Ariquemes, Rondo?nia, Brazil: biomass,
charcoal formation and burning efficiency. Forest Ecol. Mgmt.
120, 179?191.
Grace, J., Lloyd, J., McIntyre, J., Miranda, A., Meir, P., Miranda,
H., Nobre, C., Moncrieff, J., Massheder, J., Malhi, Y., Wright,
I., Gash, J., 1995. Carbon dioxide uptake by an undisturbed
tropical rain forest in southwest Amazonia, 1992 to 1993.
Science 270, 778?780.
Greenland, D.J., Kowai, J.M.L., 1960. Nutrient content of the moist
tropical forest of Ghana. Plant Soil 12, 154?174.
Guillaumet, J.-L., 1987. Some structural and floristic aspects of the
forest. Experientia 43, 241?251.
Houghton, R.A., 1991. Tropical deforestation and atmospheric
carbon dioxide. Climatic Change 19, 99?118.
Houghton, R.A., Skole, D.L., Nobre, C.A., Hackler, J.L.,
Lawrence, K.T., Chomentowski, W.H., 2000. Annual fluxes
of carbon from deforestation and regrowth in the Brazilian
Amazon. Nature 403, 301?304.
Houghton, R.A., Lawrence, K.T., Hackler, J.L., Brown, S., in press.
The spatial distribution of forest biomass in the Brazilian
Amazon: a comparison of estimates. Glob. Change Biol.
Hughes, R.F., 1997. Effects of deforestation and land use on biomass,
carbon and nutrient polls in the Los Tuxtlas region, Mexico.
Ph.D. Dissertation. Oregon State University, Corvallis, OR.
INPE, 2000. Deforestation estimates in the Brazilian Amazon.
Instituto Nacional de Pesquisas Espaciais, Sa?o Jose? dos
Campos, Brazil.
Jordan, C., Uhl, C., 1978. Biomass of a terra firme forest of the
Amazon Basin. Oecologia Plantarum 13, 387?400.
Kato, R., Tadaki, Y., Ogawa, H., 1978. Plant biomass and growth
increment studies in Pasoh Forest. Malaysian Nature J. 30,
211?224.
Kauffman, J.B., Sanford, R.L., Cummings, D.L., Salcedo, I.H.,
Sampaio, E.V.S.B., 1994. Biomass and nutrient dynamics
associated with slash fires in neotropical dry forests. Ecology
74, 140?151.
320 H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321
Kauffman, J.B., Cummings, D.L., Ward, D.E., Babbit, R., 1995.
Fire in the Brazilian Amazon. I. Biomass, nutrient pools, and
losses in slashed primary forests. Oecologia 104, 397?408.
Klinge, H., Rodriguez, W., 1973. Biomass estimation in central
Amazonian rain forest. Acta Cient. Venez. 24, 225?237.
Klinge, H., Rodriguez, W., 1974. Phytomass estimation in a central
Amazonian rain forest. In: Proceedings of the IUFRO Congress
on Forest Biomass Studies, Vol. 15, Rome.
Laurance, W.F., 1998. A crisis in the making: responses of
Amazonian forests to land use and climate change. Trends
Ecol. Evol. 13, 411?415.
Laurance, W.F., Laurance, S.G., Ferreira, L.V., Rankin-de Merona,
J.M., Gascon, C., Lovejoy, T.E., 1997. Biomass collapse in
Amazonian forest fragments. Science 278, 1117?1118.
Laurance, W.F., Ferreira, L.V., Rankin-de Merona, J.M., Laurance,
S.G., 1998a. Rain forest fragmentation and the dynamics of
Amazonian tree communities. Ecology 69, 2032?2040.
Laurance, W.F., Laurance, S.G., Delamonica, P., 1998b. Tropical
forest fragmentation and greenhouse gas emissions. Forest
Ecol. Mgmt. 110, 173?180.
Laurance, W.F., Fearnside, P.M., Laurance, S.G., Delamonica, P.,
Lovejoy, T.E., Rankin-de Merona, J.M., Chambers, J.Q.,
Gascon, C., 1999. Relationship between soils and Amazon
forest biomass: a landscape-scale study. Forest Ecol. Mgmt.
118, 127?138.
Laurance, W.F., Delamonica, P., Laurance, S.G., Vasconcelos, H.L.,
Lovejoy, T.E., 2000. Rainforest fragmentation kills big trees.
Nature 404, 836.
Laurance, W.F., Cochrane, M.A., Bergen, S., Fearnside, P.M.,
Delamonica, P., Barber, C., D?Angelo, S., Fernandes, T., 2001a.
The future of the Brazilian Amazon. Science 291, 438?439.
Laurance, W.F., Perez-Salicrup, D., Delamonica, P., Fearnside,
P.M., D?Angelo, S., Jerozolinski, A., Pohl, L., Lovejoy, T.E.,
2001b. Rain forest fragmentation and the structure of
Amazonian liana communities. Ecology 82, 105?116.
Lovejoy, T.E., Bierregaard, R.O., Rylands, A.B., Malcolm, J.R.,
Quintela, C.E., Harper, L.H., Brown, K.S., Powell, A.H.,
Powell, G.V.N., Schubart, H.O.R., Hays, M.B., 1986. Edge and
other effects of isolation on Amazon forest fragments. In:
Soule?, M.E. (Ed.), Conservation Biology: The Science of
Scarcity and Diversity. Sinauer, Sunderland, MA, pp. 257?285.
Malhi, Y., Nobre, A.D., Grace, J., Kruijt, B., Pereira, M., Culf, A.,
Scott, S., 2000. Carbon dioxide transfer over a central
Amazonian rain forest. J. Geophys. Res. 57, 34534?34547.
Martius, C., 1997. Decomposition of wood. In: Junk, W.J. (Ed.),
Decomposition of Wood. Springer, New York, pp. 267?276.
Moraes, J.L., Cerri, C.C., Mellilo, J., Kicklighter, D., Neill, C.,
Skole, D., Steudler, P., 1995. Soil carbon stocks of the Brazilian
Amazon basin. Soil Sci. Soc. Am. J. 59, 244?247.
Nelson, B.W., Mesquita, R., Pereira, J.L.G., Souza, S.G.A., Batista,
G.T., Couto, L.B., 1999. Allometric regressions for improved
estimates of secondary forest biomass in the central Amazon.
Forest Ecol. Mgmt. 117, 149?167.
Nepstad, D.C., 1989. Forest regrowth on abandoned pastures in
eastern Amazonia: limitations to tree seedling survival and
growth. Ph.D. Thesis. Yale University Press, New Haven, CT.
Nepstad, D.C., Verissimo, A., Alencar, A., Nobre, C., Lima, E.,
Lefebre, P., Schlesinger, P., Potter, C., Moutinho, P., Mendoza,
E., Cochrane, M., Brooks, V., 1999. Large-scale impoverish-
ment of Amazonian forests by logging and fire. Nature 398,
505?508.
Ogawa, K., Yoda, K., Ogino, K., Kira, T., 1965. Comparative
ecological studies on three main types of forest vegetation in
Thailand. II. Plant biomass. Nature Life Southeast Asia 4, 49?80.
Ott, L., Mendenhall, W., 1990. Understanding Statistics, 5th
Edition, PWS-Kent Publishing, Boston, MA.
Pe?rez-Salicrup, D.R., 1998. Effects of liana cutting on trees and
tree seedlings in a tropical forest in Bolivia. Ph.D. Thesis.
University of Missouri, St. Louis, MO.
Phillips, O.L., Malhi, Y., Higuchi, N., Laurance, W.F., Nunez, P.,
Vasquez, R., Laurance, S.G., Ferreira, L., Stern, M., Brown, S.,
Grace, J., 1998. Changes in the carbon balance of tropical
forests: evidence from long-term plots. Science 282, 439?442.
Putz, F.E., 1983. Liana biomass and leaf area of a terra firme forest
in the Rio Negro Basin, Venezuela. Biotropica 15, 185?189.
Putz, F.E., 1984. The natural history of lianas on Barro Colorado
Island, Panama. Ecology 65, 1713?1724.
Richter, D.D., Babbar, L.I., 1991. Soil diversity in the tropics. Adv.
Ecol. Res. 21, 315?389.
Saldarriaga, J.G., West, D.C., Tharp, M.L., Uhl, C., 1988. Long-
term chronosequence of forest succession in the Upper Rio
Negro of Colombia and Venezuela. J. Ecol. 76, 938?958.
Saloma?o, R.P., Nepstad, D.C., Vieira, I.C., 1998. Biomassa e
estoque de carbono de florestas tropicais primarias e secundar-
ias. In: Gascon, C., Moutinho, P. (Eds.), Floresta Amazo?nica:
Dinamica, Regenerac?a?o e Manejo. National Institute for
Amazonian Research (INPA), Manaus, Brazil, pp. 99?119.
Skole, D., Tucker, C.J., 1993. Tropical deforestation and habitat
fragmentation in the Amazon: satellite data from 1978 to 1988.
Science 260, 1905?1910.
Summers, P.M., 1998. Estoque, decomposic?a?o, e nutrientes da
liteira grossa em floresta de terra-firme, na Amazo?nia central.
M.Sc. Thesis, National Institute for Amazonian Research
(INPA), Manaus, Brazil.
Uhl, C., Jordan, C.F., 1984. Vegetation and nutrient dynamics during
five years of succession following forest cutting and burning in
the Rio Negro Region of Amazonia. Ecology 65, 1976?1990.
Uhl, C., Buschbacher, R., Serrao, E.A.S., 1988. Abandoned
pastures in eastern Amazonia. I. Patterns of plant succession.
J. Ecol. 76, 633?681.
Van Wagner, C.E., 1968. The line-intersect method in forest fuel
sampling. Forest Sci. 14, 20?26.
H.E.M. Nascimento, W.F. Laurance / Forest Ecology and Management 168 (2002) 311?321 321