Forest Ecology and Management 432 (2019) 365–375

Contents lists available at ScienceDirect

Forest Ecology and Management

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

Above-ground woody biomass distribution in Amazonian floodplain forests: T

Effects of hydroperiod and substrate properties
Rafael Leandro de Assisa,b, , Florian Wittmannc, Yennie Katarina Bredinb, Jochen Schöngarta,
⁎

Carlos Alberto Nobre Quesadaa, Maria Teresa Fernandes Piedadea, Torbjørn Haugaasenb
a
National Institute of Amazonian Research, Av. André Araújo, 2936 Petrópolis, Manaus, AM 69067-375, Brazil
b
Faculty of Environmental Sciences and Natural Resource Management, Norwegian University of Life Sciences, P.O. Box 5003, 1432 Ås, Norway
c
Department of Wetland Ecology, Institute of Geography and Geoecology, Karlsruhe Institute of Technology, 76131 Karlsruhe, Germany

ARTICLE INFO ABSTRACT

Keywords: The importance of tropical forests in regulating global carbon stocks is well known. However, the role of abiotic
Above-ground woody biomass variables related to climate conditions and edaphic parameters for patterns of above-ground woody biomass
Amazonian floodplains (AGWB) are still under debate. For Amazonian forests subjected to periodic floods, these patterns are even more
Hydroperiod uncertain. This study aimed to evaluate AGWB stocks in Amazonian floodplain forest, and investigate the im-
Soil nutrients
portance of forest structure, hydroperiod and edaphic parameters for AGWB. Results are based on floristic in-
Forest structure
ventories conducted in twelve hectares of forest distributed across four floodplains. All trees ≥10 cm DBH were
tagged, identified, and had their DBH and height measured. Allometric equations were applied for calculating
AGWB. Hydroperiod was estimated for each sample plot, and soil samples were collected and chemical and
physical components analyzed. Hierarchical partitioning was applied to determine importance of forest structure
variables for AGWB, and GLMMs to evaluate the individual role of several edaphic parameters and hydroperiod
for AGWB stocks. AGWB estimates varied substantially both between and within sites, as did the proportional
contribution of forest structure variables to AGWB. Fabaceae contributed most to AGWB overall, and hydro-
period was more important than soil fertility in explaining variation in AGWB values. Amongst the edaphic
variables, Iron (Fe) was the component that influenced AGWB the most, followed by Aluminium (Al) and
Phosphorus (P). Overall, our results indicate that, on the investigated Amazonian floodplains, AGWB is mainly
driven by hydroperiod rather than edaphic properties. This occurs despite a constant input of nutrients caused by
flooding events. In addition, this is the first study to suggest that P appears to be of some importance in
Amazonian várzea and paleo-várzea floodplains, where soil fertility is generally higher than in non-flooded terra
firme forests.

1. Introduction
Reducing Emissions from Deforestation and Degradation (REDD) is
a United Nations Framework Convention on Climate Change (UNFCCC)
program to offer financial incentives to developing countries to curb
forest carbon emissions. This scheme, which has been amplified in
scope and ambition as REDD+, can also indirectly support biodiversity
conservation through reduced habitat loss and degradation. These ef-
forts currently focus on tropical forest countries for several reasons.
Tropical forests cover only ∼7% of the global land area, but store as
much as 55% of all terrestrial carbon in its vegetation (Pan et al., 2011).
Additionally, tropical forests annually process around six times more
carbon via photosynthesis and respiration than that in all fossil fuel
emissions (Malhi and Grace, 2000; Lewis et al., 2009). Deforestation
rates are also higher in tropical forests than elsewhere (Achard et al.,
2002; Santilli et al., 2005). This forest clearing and degradation account
for roughly 15% of current global greenhouse gas emissions, making it
the second most important contributor to climate change after fossil
fuel combustion (Houghton, 2005).
Tropical forests also play a pivotal role in mitigating climate change
by absorbing nearly one-fifth of the annual CO2 released by fossil fuels
around the globe (Lewis et al., 2009) and recycling nearly half of all
precipitation as evapotranspiration (Avissar and Werth, 2005). In fact,
there is mounting evidence that undisturbed old-growth tropical forests

Abbreviations: AGWB, above-ground woody biomass; GLMM, generalized linear mixed models
⁎
Corresponding author at: National Institute of Amazonian Research, Av. André Araújo, 2936 Petrópolis, Manaus, AM 69067-375, Brazil.
E-mail addresses: rafale.assis@gmail.com (R.L. de Assis), f-wittmann@web.de (F. Wittmann), yennie.bredin@nmbu.no (Y.K. Bredin),
torbjorn.haugaasen@nmbu.no (T. Haugaasen).

https://doi.org/10.1016/j.foreco.2018.09.031
Received 4 July 2018; Received in revised form 23 August 2018; Accepted 17 September 2018
Available online 28 September 2018
0378-1127/ © 2018 Elsevier B.V. All rights reserved.
R.L. de Assis et al.
may be accelerating in growth (Lewis et al., 2004) and increasing in
biomass and carbon (Malhi and Grace, 2000; Baker et al., 2004a),
possibly in response to a CO2 fertilization effect; i.e. elevated atmo-
spheric CO2 concentrations (Malhi and Grace, 2000). Yet alternative
hypotheses suggest that any evidence of biomass accumulation is
simply a local response to forest recovery from natural and human
disturbances. For example, severe El Niño events cause elevated tree
mortality (Condit et al., 1995; Laurance et al., 2001) and recent esti-
mates indicate that natural tropical forest regeneration currently re-
moves 1.6 Pg C yr−1 from the atmosphere (Pan et al., 2011).
Yet, despite their obvious importance for the global carbon budget,
the total amount of carbon stored in tropical forests is still under intense
debate. This is underlined by global values ranging from 158 to
324 Pg C (Gibbs et al., 2007; Baraloto et al., 2011). There is therefore an
urgent need to improve our understanding of how different factors af-
fect the spatial and temporal variation in above-ground woody biomass
(AGWB) in the tropics (Gullison et al., 2007; Baraloto et al., 2011).
Several studies have reported a role for climatic variables in de-
termining patterns of AGWB within tropical forests. For instance, re-
gional variation of AGWB was positively related to annual precipitation
across Amazonia, while dry season length had a negative impact on
AGWB (Chave et al., 2004; Malhi et al., 2006). However, other studies
found no effect of water availability and claim that AGWB in the
Amazon is mainly influenced by forest structure variables, such as basal
area (Baker et al., 2004b; Malhi et al., 2006; Paoli et al., 2008), number
of large trees (DeWalt and Chave, 2004; Paoli et al., 2008; Rutishauser
et al., 2010), mean tree height (Chave et al., 2005), mean wood specific
gravity (Baker et al., 2004b; DeWalt and Chave, 2004; Stegen et al.,
2009) and, maximum individual biomass (Stegen et al., 2011).
Soil properties, including drainage ability, have also been re-
peatedly reported as highly important for variation in species compo-
sition and other forest structure components at large (e.g. ter Steege
et al., 1993; Ruokolainen et al., 1997; Sollins, 1998) and local scales
(Clark et al.,1995; Condit et al., 2013). For example, changes in AGWB
have been attributed to soil fertility (DeWalt and Chave, 2004; Paoli
et al., 2008; Quesada et al., 2009). However, most studies have been
conducted at large spatial scales and few studies have analyzed the
effects of soil fertility on AGWB at local scales (Laurance et al., 1999;
Ferry et al., 2010). Moreover, a common denominator for existing
studies is that the vast majority have been conducted in upland terra
firme forest (i.e. forests not subjected to periodic flooding). A more
refined understanding of how different environmental variables affect
patterns of AGWB across different forest types in Amazonia is therefore
lacking.
In Amazonia, floodplain forests subjected to predictable, long-
lasting, and monomodal flood pulses cover vast areas (Melack and Hess,
2010; Junk et al., 2011). These floodplains are divided into different
types based on the hydro-chemical properties of the rivers that inundate
them. The two major floodplain systems in Amazonia are called várzea
and igapó. Várzea forests occupy approximately 450,000 km2 and occur
along white-water rivers rich in Tertiary/Quaternary sediments origi-
nating from the Andes or pre-Andean regions. Due to the annual de-
position of nutrient rich sediments, these forests are highly fertile (Junk
et al., 2011). Igapó forests cover an area of approximately 300,000 km2
and occur on floodplains along black- and clear-water rivers that carry
small amounts of Tertiary sediment originating from the Guyana and
Central Brazilian Shields (Sioli, 1965; Prance, 1979; Melack and Hess,
2010; Junk et al., 2011). These are low fertility systems, as they lack the
significant seasonal input of nutrient rich sediments (Junk et al., 2011).
A third, but less extensive, floodplain type was also recently supported
based on an analysis of tree species composition - paleo-várzea (Assis
et al., 2015a). It covers approximately 125,000 km2 and occurs on
fluvial Andean deposits that have been abandoned by white-water
rivers. These floodplains are inundated by small to intermediate black-
water rivers that originate in cratonic areas and carry these already
once-deposited paleo-sediments (Junk et al., 2011).
366
Forest Ecology and Management 432 (2019) 365–375
Amazonian floodplains are subjected to floods that may last for
several months during the course of the year. The effects of water-
logging include oxygen deficiency at the root level, and a reduction in
photosynthesis and tree growth during inundation (Parolin et al., 2004;
Parolin, 2009; Schöngart et al., 2002). To cope with the annual flood
pulse, tree species that inhabit the Amazonian floodplains have there-
fore developed a range of phenological, physiological, and structural
adaptations. These strategies may strongly influence the AGWB in these
ecosystems (Hawes et al., 2012). For example, tree species portray
rapid growth rates in várzea forests due to the fertile alluvial soils. This,
and the highly dynamic nature of white-water rivers, favour the pre-
sence of species with relatively low wood specific gravity (Worbes
et al., 1992; Baker et al., 2004b; Wittmann et al., 2006). In addition,
disturbance caused by inundation and related geomorphic dynamics
appears to restrict tree height, explaining why Amazonian floodplain
trees are generally lower than their terra firme counterparts (Souza and
Martins, 2005). Such variation between different forest types across
Amazonia underline the importance of habitat-specific AGWB estimates
to more adequately account for the role of Amazonian forests in the
global carbon budget.
Yet, there are few studies investigating potential changes in AGWB
along environmental gradients in the Amazon, particularly within
floodplains. One exception is that of Hawes et al. (2012), who observed
that várzea forest biomass was higher in plots subjected to longer flood
duration. However, the effects of flooding on forest parameters (in-
cluding AGWB) are known to be intimately associated with edaphic
variables (Wittmann et al., 2010; Assis et al., 2015b, Targhetta et al.,
2015). It is therefore only possible to disentangle the influence of
edaphic variables and flooding on AGWB by jointly considering these
gradients. The main objective of this study was to estimate AGWB
across four different floodplain forests in central-western Amazonia and
evaluate how AGWB is related to hydroperiod and soil properties. More
specifically, we addressed the following questions: (i) how do forest
structure and species composition on the different floodplains affect
AGWB patterns? and (ii) what is the role of edaphic variables and the
hydroperiod in determining patterns of AGWB?
2. Material and methods
2.1. Study areas
Floristic inventories were conducted in four floodplain forests along
different tributaries of the Solimões (=Amazon) river: Purus (S4°19′;
W61°52′), Tefé (S4°9′; W65°6′), Juruá (S3°14′; W66°3′) and Jutaí
(S3°22′; W67°28′; Fig. 1). All study areas are located at approximately
the same latitude (2° S), but are separated longitudinally by a minimum
and maximum distance of 120 and 600 km, respectively. The floodplain
forests at each site are affected by an annual, long-lasting (> 5 months)
flood pulse, so the trees are subjected to alternating terrestrial and
aquatic phases with a seasonal inundation of up to 230 days per year
(flood pulse, sensu Junk et al., 1989). The forests surveyed are con-
sidered pristine as they are located distant of urban centres. Annual
rainfall averages approximately 2800, 2700, 2500 and 2600 mm, for
Jutaí, Juruá, Tefé and Purus, respectively. None of these areas have
more than one month per year with rainfall < 100 mm (Sombroek,
2001).
The Purus and Juruá rivers are white-water rivers carrying large
amounts of nutrient-rich sediments from the Andes. Floodplains along
these rivers are thus classified as várzea and their substrates are highly
fertile. The Tefé and Jutaí rivers carry paleo-sediments originating from
the Andes, and are therefore classed as floodplain paleo-várzeas (Junk
et al., 2011; Assis et al., 2015a).
2.2. Floristic inventories
Floristic inventories were conducted during the 2009 and 2010 low-
R.L. de Assis et al.
Fig. 1. Map of the Amazon basin (insert) with the four study regions indicated (black dots). Black circles within each river basin represent the sample plots.
water season. In total, 12 ha of floodplain forest were inventoried across
27 forest plots (six in each of Juruá, Jutaí and Tefé and nine in Purus);
three hectares in each river basin. Inventory plots ranged from 0.25 to
1.0 ha in size in order to avoid large differences in hydroperiod within
plots. All inventory plots were located in late-successional forest and
the distance between plots within the same study area ranged from 0.5
to 10 km (Fig. 1). All live trees ≥10 cm in diameter at breast height
(DBH) were tagged, identified, and had their DBH and height measured.
Plots were divided into sub-plots measuring 25 × 25 m to facilitate
measurements. Tree heights were measured using a clinometer for three
to five trees per sub-plot, and the height of the remaining individuals
were estimated based on these measurements.
Vouchers were collected from all trees and transported to the her-
barium at the National Institute of Amazonian Research (INPA),
Manaus, where they were identified to the lowest taxonomic level
possible. Specimens not identified to species were assigned to morpho-
species. Only fertile material was permanently deposited in the INPA
herbarium.
The floristic inventories contained 7722 trees, belonging to a
minimum of 518 species, 203 genera and 55 families across the 12 ha of
floodplain forests. Species richness was highest in Purus (213 species),
followed by Tefé (192), Juruá and Jutaí (146 each). Overall, approxi-
mately 80% of all trees were identified to species, while around 20%
were only identified to genus level. Less than 1% were determined to
family level or remained unidentified. Further details on species com-
position and diversity can be found in Assis et al. (2017).
2.3. Wood specific gravity
Wood specific gravity (WSG) values were obtained from the Global
Wood Density Database (Chave et al., 2009; Zanne et al., 2009). Most of
the sampled species are included in this database. When a species in the
367
Forest Ecology and Management 432 (2019) 365–375
database had more than one reference for WSG, we used the average of
the listed values. When a species was missing from the database, we
used the average for the respective genus or family. The same criterion
was applied for individuals allocated to morpho-species. We pre-
ferentially used WSG data from specimens collected in tropical South
America. Since variation among taxa within the same family and genera
is relatively low (Baker et al., 2004b), we assume that this practice is
not significantly biasing our results.
2.4. Above-ground woody biomass (AGWB)
Basal area (in cm2) was calculated for all sampled trees using the
following formula: BA = π × (d/2)2, where d is DBH in cm. For esti-
mating biomass, we applied three allometric models for all recorded
trees:
(1) Based on Cannell (1984) equation: AGWB = 0.6 × ρ × h × π × (d/2)2
(2) Based on Chave et al. (2005): AGWB = 0.112 × (ρ × h × d2)0.916
(3) Based on Chave et al. (2005): AGWB = 0.0509 × ρ × h × d2
where: ρ is wood specific gravity; h is tree height (in m), and d is
DBH (in cm).
These models were chosen as there is no specific allometric equation
for calculating AGWB in floodplain forests. Previous studies have found
these equations to have good confidence, mainly because these models
consider WSG and tree height as predictors (Chave et al., 2005; Malhi
et al., 2006; Schöngart et al., 2010; Hawes et al., 2012). Thus, an
average of the three allometric models was used to estimate AGWB.
Subsequently, we aggregated all individual tree BA and AGWB values
within a sample plot and extrapolated to a per hectare measure
(m2 ha−1 for total basal area and Mg ha−1 for AGWB).
R.L. de Assis et al.
Table 1
Biomass-related variables averaged across each study site. Different letters between sites represent significant differences (p < 0.01) based on Mann-Whitney U tests.
BAL = basal area of large trees.
Site Tree density (stems ha−1) Total basal area (m2 ha−1)
Juruá 615.3b 36.88a
Jutaí 748a 37.46a
Purus 626.3b 35.74a
Tefé 583.3b 26.59b
2.5. Substrate and flooding
Three soil samples (from 0 to 20 cm depth) were collected at ran-
domly chosen points within each plot with a soil auger. These samples
were subsequently mixed to produce one composite sample per plot
(1 ha). Three samples per plot are assumed to be sufficient for the soil
analysis. As most nutrients and soil particles in these forests arrive from
the adjacent river during flooding, the soil characteristics are very si-
milar across large areas, but may differ depending on local topography
(i.e. flooding level; Wittmann et al., 2004).
All samples were analyzed at the EMBRAPA Soil Laboratory in
Manaus. In total, eighteen edaphic components were considered: three
textural (contents of clay, sand and silt) and thirteen chemical (C, N, P,
K, Na, Fe, Zn, Mn, Cu, Ca, Mg, Al, CEC - Cation Exchange Capacity and
pH). Details of laboratory methods used for soil analyses can be found
in EMBRAPA (1997) and Fearnside and Leal-Filho (2001).
Flooding period was estimated following Wittmann and Junk
(2003) and Schöngart and Junk (2007). This method consists of mea-
suring the maximum flood level of the last high-water period for each
tree, evidenced by a pronounced water mark on the tree trunks. These
measurements were coupled with data from the nearest river gauge
stations, all regulated by the Brazilian National Water Agency (ANA,
2010). They contain information on historical river levels that have
been recorded for one to four decades. Details of gauge stations used are
presented in the supplementary material (Table S1). All references to
flooding hereafter refer to mean flood duration (days year−1).
2.6. Analyses
In order to improve homoscedasticity, all environmental variables
were transformed to zero skewness (cf. Sokal and Rohlf, 1995), i.e. after
the transformation, all variables had values ranging from 0 to 1. We
used Mann-Whitney nonparametric U-tests to evaluate the significance
(at p < 0.01) of forest structure variables (i.e. tree density, total basal
area, mean tree height and mean WSG) on AGWB among study areas, as
several variables were not normally distributed. We used Kendall’s
nonparametric correlation t tests (Kendall, 1938) to analyze for corre-
lations between predictor variables (i.e. the different edaphic properties
and hydroperiod). We applied hierarchical partitioning at site level to
determine the relative importance of each forest structure variable for
patterns of AGWB.
For the analyses, we used Generalized Linear Mixed Models
(GLMMs; Breslow and Clayton, 1993; Lee and Nelder, 1996), assuming
autocorrelation between the sample units. The GLMMs are re-
commended, because they fit overall fixed effects when the sample
units are spatially correlated (nested). In the current study, each study
area (river basin) was used as random effect, and the relative support
for each model was determined through Akaike’s Information Criterion
(AICc), considering both the relative likelihood (L) of each model and
its complexity (i.e. number of explanatory variables included in the
model; Akaike, 1981). The response variables were tested for normal
distribution by applying the Shapiro–Wilk test of normality. When not
normally distributed, we applied the goodness-of-fit (GOF) test to verify
which distribution pattern (e.g. Poisson, Zero-inflated Poisson, Nega-
tive binomial) provided a better fit.
Forest Ecology and Management 432 (2019) 365–375
Average WSG Average tree height (m) BAL (%) AGWB (Mg ha−1)
0.57 ± 0.15b 17.91 ± 5.96a 49.5a 282.07a
0.63 ± 0.13a 15.44 ± 7.43b 38.8b 309.4a
0.58 ± 0.17b 13.03 ± 4.33d 39.1b 188.2b
0.65 ± 0.16a 14.09 ± 3.39c 39.3b 178.2b
We adopted two strategies to evaluate the influence of edaphic
variables and hydroperiod on AGWB. First, we performed a Principal
Component Analysis (PCA) to describe major gradients in hydroperiod
and soil data, using a subset of 19 variables (hydroperiod plus the 18
edaphic components) that were recorded in all 27 plots. The gradients
were thus summarized into four main PCA axes (PCA 1–4), and their
importance for AGWB were tested using GLMMs. We applied the PCA
axes to ensure that the models do not suffer from collinearity effects,
because the ordination axes are statistically independent.
Second, we tested the individual effect of the predictor variables on
patterns of AGWB. As predictors for the GLMMs, we used the largest set
of weakly correlated variables (t ≤ |0.4|: sand, silt, N, P, Fe, Mn, Al and
pH (see Supplementary Information Table S2)). For model validation,
we plotted the residuals against the fitted values for all the GLMMs.
All statistical analyses were performed in R 3.4.1 (R Core Team
2013), with packages “hier.part” (Walsh and Mac Nally, 2013) for the
hierarchical partitioning analysis, “vegan” (Oksanen et al., 2015) for
the ordination analysis (PCA), “MASS” (Venables and Ripley, 2002),
“VGAM” (Thomas, 2010) for GOF, and “nlme” (Pinheiro et al., 2018)
for the GLMMs.
3. Results
3.1. Forest structure and AGWB
All forest structure variables and AGWB estimates varied sub-
stantially both between and within the sites surveyed (Table 1, Fig. 2).
Total basal area and basal area of large trees (> 40 cm DBH) were
highest in Juruá, whereas Jutaí had the highest tree densities. Average
WSG was higher in Jutaí and Tefé compared to Purus and Juruá, and
AGWB had significantly higher estimates for Juruá and Jutaí compared
to Purus and Tefé. Average tree height was the only variable that was
significantly different in all survey sites.
Overall, sites differed substantially in relation to the contribution of
structural parameters (total basal area, mean height, WSG average and
tree density) to AGWB. Basal area was the most important factor con-
tributing to AGWB values at all sites, except for Tefé (Fig. 3A–D). In
Tefé, WSG was most important. Tree height was important in Jutaí and
Purus, and all variables appear to have a similar contribution to AGWB
in Jutaí.
The influence of different DBH classes on total AGWB was similar
across sites. Tefé was the site where trees > 70 cm DBH contributed the
least, only 6%, to the total AGWB. At the other sites, trees of this dia-
meter class comprised on average 12–13% of the total AGWB
(Fig. 3A–D, in bars).
3.2. Importance of floristic composition for AGWB
Two families, Fabaceae and Lecythidaceae, concentrated about one
third (33%) of the total AGWB across the four study sites. Fabaceae
supported the highest AGWB values for Juruá and Purus, whereas
Lecythidaceae was the most important contributor to AGWB in Jutaí
and Tefé (Table 2). The ten most important families combined corre-
sponded to ≥80% of the total AGWB for Jutaí, Juruá and Tefé, whereas
the contribution was somewhat lower for Purus (67.3%). Juruá had the
368
R.L. de Assis et al. Forest Ecology and Management 432 (2019) 365–375

A B

100

5
Total basal area (m2/m2)
80
Number of stems

4
60

3
40

2
20

1
b a b b a a a b
Juruá
Juruá Jutaí
Jutaí Purus
Purus Tefé
Tefé Juruá
Juruá Jutaí
Jutaí Purus
Purus Tefé
Tefé

C D
80

22
Basal area of large trees (%)

20
Tree height average (m)
60

18
16
40

14
20

12
10
0

a b b b a b d c
8

Juruá
Juruá Jutaí
Jutaí Purus
Purus Tefé
Tefé Juruá
Juruá Jutaí
Jutaí Purus
Purus Tefé
Tefé

E F
40
0.7

30
AGWB (Mg/m2)
Average WSG
0.6

20
0.5

10
0.4

b a b a a a b b
0

Juruá
Juruá Jutaí
Jutaí Purus
Purus Tefé
Tefé Juruá
Juruá Jutaí
Jutaí Purus
Purus Tefé
Tefé

Fig. 2. Boxplots of (A) number of stems, (B) total basal area, (C) basal area of large trees, (D) average tree height, (E) average WSG and (F) AGWB across sample plots
(25 × 25 m) in each study site. The boxes represent the middle 50% of the data, the line in the middle is the median, and the edges are the 25th and 75th percentiles.
Different letters (lower case) within the same sub-figure represent significant differences at p < 0.01 (Mann-Whitney U tests).

highest concentration of AGWB in a single family (Fabaceae, 25%;
Table 2).
Individual species contribution to total AGWB was more variable.
Only four species, Eschweilera albiflora, Eschweilera ovalifolia, Licania
micrantha and Virola surinamensis, were among the top ten most im-
portant contributors to AGWB in more than one site (Table 3). V.
surinamensis was the only species present among the ten most important
contributors in all four study areas. However, E. albiflora contributed
the most to total AGWB values when all four sites were combined. The
ten species contributing most to AGWB values exceeded 50% for Jutaí
and Juruá, whereas for Purus and Tefé the corresponding values were
40–41% (Table 3).

369
R.L. de Assis et al. Forest Ecology and Management 432 (2019) 365–375

A) Jutaí B) Juruá
Height
Basal area Basal area
Height 10-20 cm, 15% 10-20 cm, 13%

20-40 cm, 35% 20-40 cm, 36%

WSG
WSG
40-70 cm, 37% 40-70 cm, 38%
Tree Tree
density >70 cm, 12% density >70 cm, 12%

C) Purus D) Tefé Height

Basal area
Basal area
10-20 cm, 16%
10-20 cm, 21%
Height
20-40 cm, 41% WSG
20-40 cm, 36%

WSG 40-70 cm, 29%

40-70 cm, 37%
>70 cm, 13% Tree
density >70 cm, 6%
Tree
density
Fig. 3. The proportional contribution of mean tree height (m; black), average WSG (dark gray), mean tree density (stems ha−1; light gray) and total basal area
(m2 ha−1; white) to AGWB estimates in the four study areas. The contribution of different DBH classes are presented separately for each site.

3.3. Influence of environmental factors on AGWB
The PCA ordination of the full set of environmental variables shows
that the first axis (PCA1) explained almost 61% of the total variation in
the dataset. This axis was correlated with most of the edaphic para-
meters, including sand, clay, N, K, and CEC (Fig. 4). PCA2 was also
correlated with a number of edaphic variables (clay, pH, C, and Al), but
was the only PCA axis significantly correlated with hydroperiod. The
subsequent PCA axes were mainly correlated with Fe and P (PCA3), and
sand and silt (PCA4). Plots from Jutaí and Tefé (paleo-várzea sites) are
dominated by sandy soils and are loosely clustered to the right side of
the ordination diagram, whereas plots from Juruá and Purus (várzea
sites) are dominated by clayish substrates and are clustered to the left
side of the ordination diagram (Fig. 4).
The GLMMs show that the null model best describes the observed
patterns of AGWB, explaining approximately 46% of the observed
variation (Table 4). PCA3 alone explained 21% of the AGWB variation
and PCA3 is listed as the best predictor in five of the ten best models.

Table 2
Total above-ground woody biomass (AGWB, in Mg ha−1) and the percentage contribution of
the ten most important families for the study areas in four southern tributaries of the Solimões
(=Amazon) river, central Amazonia, Brazil. Families are listed in descending order of their
contribution to AGWB. Values presented in light gray are the sum of AGWB from the 10 most
important families for each study area. The dark gray row represents the summed AGWB of the
remaining families combined.

370
R.L. de Assis et al.
Table 3
Total above-ground woody biomass (AGWB, in Mg ha−1) and the percentage contribution of
the ten most important species for the study areas in four southern tributaries of the Solimões
(=Amazon) river, central Amazonia, Brazil. Species are listed in descending order of their
contribution to AGWB. Values presented in light gray are the sum of AGWB from the 10 most
important species for each study area. The dark gray row represents the summed AGWB of the
remaining species combined.
Fig. 4. Ordination diagram containing the PCA plot scores from PCA1 and
PCA2, including the edaphic variables and hydroperiod. The orientation of
arrows indicates the direction of the correlation and arrow length is related to
the strength of the correlation.
Hydroperiod was the second best predictor (explaining approximately
8% of the variation in AGWB), followed by PCA2, PCA1 and PCA4,
respectively.
The null model also best explained the variation in AGWB when
considering the set of uncorrelated substrate properties as explanatory
variables. After the null model, Fe was the best predictor. Fe was highly
associated with PCA3 and was the only variable that appeared three
times in the list of most important predictors in the GLMMs, always
negatively influencing AGWB. Aluminum was the third best predictor,
and explained approximately 3.5% of the AGWB variation. Among the
371
Forest Ecology and Management 432 (2019) 365–375
thirteen most reliable models presented in Table 5, the models that
considered any of the edaphic variables alone or combined explained
only 32.6% of the variation in AGWB.
4. Discussion
4.1. Distribution of AGWB across river basins
In the current study, AGWB values in floodplain forests were highly
variable across the four river basins. For example, AGWB was about
twice as high in Juruá and Jutaí compared to Purus and Tefé. This result
is in line with previous studies at large spatial scales in the Amazon,
where a similar variability in AGWB has been observed (e.g. DeWalt
and Chave, 2004; Malhi et al., 2006; Saatchi et al., 2009; Quesada et al.,
2010; Baraloto et al., 2011). However, the common denominator for
these studies is that they focused on terra firme forest. The current
study therefore suggests that a similar tendency can be observed for the
floodplains, particularly at a regional scale. In addition, values pre-
sented here show that AGWB in Amazonian várzea and paleo-várzea
forests may be comparable to terra firme forest estimates at least in
some regions (DeWalt and Chave, 2004; Baraloto et al., 2011).
Large trees contribute substantially to the AGWB in most tropical
forests worldwide (Slik et al., 2013). However, large trees were not an
important contributor to AGWB estimates in the current study. The
Juruá plots had a higher number of large individuals than the other
three sites, yet differences in the contribution to AGWB between larger
and smaller stems was not significant across study sites. In addition,
Jutaí, Purus and Tefé supported very similar numbers of large in-
dividuals, but were highly different in AGWB. Large floodplain trees
have a lower stature than their terra firme counterparts (Campbell
et al., 1986; Palminteri et al., 2012), which may explain their reduced
importance for the AGWB values. The canopy heights observed in the
present study are similar to other studies from Amazonian floodplains
where mature forest canopy height averaged 15–19 m (e.g. Schöngart
et al., 2010) - substantially lower than in terra firme.
The importance of the different forest structure variables used to
R.L. de Assis et al. Forest Ecology and Management 432 (2019) 365–375

Table 4
The 10 best Generalized Linear Mixed Models (GLMMs) ranked according to their importance. GLMMs included all possible combinations of hydroperiod and PCA
axes (1–4) as explanatory variables and AGWB as the response variable. df = degrees of freedom.
Rank Intercept Hydroperiod PCA1 PCA2 PCA3 PCA4 df logLik AICc delta weight

1 0.430 3 −1.50 10.1 0 0.46

2 0.428 0.23 4 −0.91 11.6 1.59 0.21
3 0.349 0.16 4 −1.91 13.6 3.58 0.08
4 0.429 0.09 4 −2.38 14.6 4.52 0.05
5 0.436 −0.09 4 −2.44 14.7 4.65 0.04
6 0.429 0.05 4 −2.48 14.8 4.73 0.04
7 0.416 0.02 0.23 5 −1.64 16.1 6.08 0.02
8 0.428 0.23 0.04 5 −1.99 16.8 6.78 0.01
9 0.429 0.04 0.22 5 −2.03 16.9 6.87 0.01
10 0.433 −0.06 0.22 5 −2.09 17 6.98 0.01

Table 5
Main results from the Generalized Linear Mixed Models (GLMMs), which included all possible combinations of nine substrate properties as explanatory variables and
AGWB as response variable. df = degrees of freedom.
Rank Intercept Sand Silt N P Fe Mn Al pH d.f. AICc Delta Weight (%)

1 −0.89 2 10.2 0 9.9

2 −0.35 −1.1 3 11.9 1.71 4.2
3 −1.25 0.7 3 12.3 2.10 3.4
4 −1.16 0.5 3 12.5 2.30 3.1
5 −0.71 −0.4 3 12.6 2.37 3.0
6 −1.03 0.2 3 12.7 2.51 2.8
7 −0.99 0.2 3 12.7 2.51 2.8
8 −0.94 0.08 3 12.8 2.54 2.8
9 −0.93 0.1 3 12.8 2.54 2.8
10 −0.6 −1.8 1 4 14.1 3.90 1.4
11 −0.7 −1.2 0.7 4 14.2 3.93 1.4
12 0.11 −0.7 −1.4 4 14.2 3.98 1.4

predict AGWB were highly variable across river basins. Only total basal
area was relatively important at all investigated sites. In Jutaí, the
contribution of the different forest structure variables to AGWB was
almost equal. Coincidentally, it was also Jutaí that supported the
highest AGWB values. This may indicate that an equal importance
across the forest structure variables (i.e. similar percentage contribution
of WSG and height classes), correspond to higher AGWB. In addition,
these results highlight the importance of accounting for as many forest
structure variables as possible (e.g. DBH, height, and WSG) during
forest inventories to achieve reliable AGWB estimates.
Across Amazonian terra firme forests, WSG is a major factor de-
termining patterns of AGWB (Baker et al., 2004b; Malhi et al., 2006).
Since average WSG for a forest site is largely driven by species com-
position, it is expected that the floristic composition accounts for a large
fraction of WSG variations at large spatial scales in Amazonia. Yet, WSG
was of minor importance in the current study. Instead, total basal area
contributed most to AGWB at all forest sites. Species composition
therefore appears to be of minor importance for AGWB in the floodplain
forests studied. This is reinforced by the fact that the floristic similarity
across forest sites with similar substrates was not reflected in AGWB
patterns. Juruá and Purus (várzea forests) are floristically similar and so
are Jutaí and Tefé (paleo-várzea forests), but the two floodplain forest
types differ substantially in species composition (Assis et al., 2017).
AGWB was highest in Jutaí (paleo-várzea), followed by Juruá (várzea),
Purus (várzea) and Tefé (paleo-várzea), respectively. A similar trend
was observed for patterns of alpha-diversity, with substantially higher
alpha-diversity in Tefé (α = 29.5), followed by Purus (α = 28.4), Juruá
(α = 26.2) and Jutaí (α = 18.7). This indicates that high measures of
alpha- or beta-diversity may not necessarily reflect high biomass stocks
in Amazonian várzea and paleo-várzea forests.
Fabaceae was the most dominant family and accounted for at least
15% of the total AGWB values at all study sites. This reinforces the
importance of Fabaceae, which reportedly is the most abundant family
both in Amazonian floodplains (e.g. Ayres, 1993; Keel and Prance,
1979; Ferreira, 1997; Haugaasen and Peres, 2006; Wittmann et al.,
2010; Assis et al., 2017) and in terra firme (Gentry 1988, 1992;
Terborgh and Andresen, 1998; ter Steege et al., 2000, 2006). Lecythi-
daceae, a family that comprises species of large trees with relatively
high WSG was the most important family in paleo-várzea, and the
second most important family for AGWB estimates overall. Other fa-
milies with similar characteristics, such as Sapotaceae and Myr-
isticaceae, were also important across the study areas. These results
indicate that, although of minor importance, the floristic component
plays a small role for AGWB stocks on Amazonian várzea and paleo-
várzea floodplains.
We mainly found late-successional species with high WSG (e.g. from
the genera Eschweilera, Licania, Pouteria, Tapura and Caryocar) among
the ten most important species for AGWB in Jutaí and Tefé.
Contrastingly, Juruá and Purus plots contained several early-succes-
sional species often characterized by low WSG among the most im-
portant species (e.g. Lueheopsis rosea, Leonia crassa, Pseudobombax
munguba and Luehea cymulosa). All four study sites were located in late-
successional forest, but the presence of species typical of earlier suc-
cessional stages may indicate that forest turnover in Juruá and Purus
are higher than in Jutaí and Tefé (Schöngart et al., 2010). This is in line
with the higher geomorphological dynamics (erosion and deposition) in
várzea floodplains compared to paleo-várzea areas (Irion et al., 2010;
Assis et al., 2015a).
Studies in terra firme support the hypothesis that forests with high
turnover rates have lower biomass (Phillips et al., 1998; Baker et al.,
2004a). This is due the fact that larger, long-lived trees with higher
WSG are associated with late-successional forests. Moreover, the ac-
cumulation of large trees drives an increase in the local biomass pools
(Quesada et al., 2010). In the current study, we observed the highest
average WSG in Tefé. However, Tefé also had the lowest number of
individuals, probably because disturbance processes are rarer and
turnover rates are lower than at the other investigated sites. Jutaí was
intermediate with a high concentration of AGWB in species with high

372
R.L. de Assis et al.
WSG, but also had several species typical of early-successional forest
stages. Schöngart et al. (2010) reported similar observations for
floodplain forests in the Mamirauá Reserve. They found slightly higher
AGWB in old secondary forests (50 years old) compared to late-suc-
cessional forests (240 years old). Intermediate turnover rates may
therefore promote higher AGWB in the floodplain types surveyed. This
differs substantially from observations in terra firme forests (Phillips
et al., 1998; Baker et al., 2004a; Quesada et al., 2010).
4.2. Influence of edaphic parameters and hydroperiod on AGWB
Hydroperiod was more important than soil fertility and had a slight
positive influence on AGWB at the study sites. This is in agreement with
previous studies showing that mature várzea forests with the longest in-
undation period have higher AGWB (Wittmann et al., 2006; Hawes et al.,
2012). Small differences in the topography may therefore be more im-
portant for AGWB on the floodplains than in terra firme forest (Quesada
et al., 2010). This is because even small changes in topography may cor-
respond to substantial changes in the hydroperiod. Several studies show
that slight variations in topography and the associated differences in hy-
droperiod are important for species richness and composition (Campbell
et al., 1986; Balslev et al., 1987; Worbes et al., 1992; Duivenvoorden, 1996;
Wittmann et al., 2004, 2010; Assis et al., 2015b). These studies all report a
decrease in species richness with increasing flooding, which is opposite to
the tendency for AGWB in the current study. This may occur because trees
tend to grow slower when subjected to prolonged floods and thus in-
corporate more biomass in more constrained tree rings. Consequently, this
results in higher WSG across individual tree species, since tree rings are
more compact (Wittmann et al., 2006; Hawes et al., 2012).
Patterns of AGWB observed across the different floodplain forests ap-
peared to be unrelated to differences in soil fertility. This is evidenced by
lower AGWB in Purus where substrates are more nutrient rich than in Jutaí,
where AGWB was higher. Tefé and Purus had very similar AGWB values,
despite very distinct soil fertility. To some extent, this corresponds to the
findings of Schöngart et al. (2010), which also did not find differences in
AGWB between mature igapó and várzea forests.
Yet, some individual soil elements had an impact on AGWB. For
example, apart from the null model, the best GLMM was the model only
containing Fe (Table 5). The negative relation between iron and AGWB
could be related to the phytotoxic properties of divalent Fe, which is
caused by oxygen deficiency in waterlogged soils (Kozlowski, 1984;
Snowdon and Wheeler, 1995; De Simone et al., 2002). This is in ac-
cordance with Assis et al. (2015b), who found that alpha-diversity is
likely to decrease with increasing substrate Fe concentrations. Iron is
therefore an important factor for patterns of species composition in
floodplain forests, despite reports from experimental studies showing
that some Amazonian floodplain tree species are relatively insensitive
to Fe toxicity (De Simone et al., 2002).
Moreover, Fe is known to play an important role for the nutrient
availability in soils. Iron oxides absorb P, thus making this element un-
available to plant roots (Alexander, 1989). This may explain why the as-
sociation of AGWB and Fe was negative in the models, while P had a po-
sitive effect on AGWB. Phosphorus has been repeatedly shown to be the
main limiting soil nutrient for forest productivity worldwide (Vitousek,
1982, 1984; Cuevas and Medina, 1986; Vitousek and Sanford, 1986; Silver,
1994; Reich et al., 1995; McGrath et al., 2001; Paoli and Curran, 2007;
Quesada et al., 2010). This pattern is even stronger in the tropics, because
forest soils undergo more intense weathering processes that reduce P pools
available to plants – mainly through leaching and/or occlusion of P by Fe
and Al oxides that alter the chemical state of both organic and inorganic P
(Walker and Syers, 1976; Quesada et al., 2010). Since P is an essential
element for the carboxylation in photosynthesis and thus for primary pro-
duction (Raaimakers et al., 1995; Crews et al., 1995; Herbert and Fownes,
1995; Raich et al., 1996; Kitayama et al., 2004), biomass accumulation as a
product of forest net primary productivity (NPP) is highly correlated to P
availability (Paoli and Curran, 2007; Quesada et al., 2010). The importance
373
Forest Ecology and Management 432 (2019) 365–375
of P for NPP and biomass in the tropics has so far only been reported for
unflooded tropical forests. The current study is therefore the first to suggest
that this element also appears to be of some importance in várzea and
paleo-várzea forests, where soil fertility is generally higher than in non-
flooded terra firme forests.
Another individual soil element that influenced the observed AGWB
patterns in this study was Aluminum (Al). Several studies show a ne-
gative impact of Al on plant physiology and attribute this to the toxicity
that Al may have for plants (Taylor, 1988; Bennet and Breen, 1991;
Ryan et al., 1993). In the current study, however, there was a positive
relationship between Al and AGWB. For Amazonian terra firme forests,
Laurance et al. (1999) observed similar associations between Al and
biomass. This may occur because Al is associated with clayey soils,
which in turn are positively related to soil fertility (Lenthe, 1991). This
explains why biomass tended to be higher in sites richer in clay,
whereas the models applied showed a negative association between
AGWB and sand. However, soil texture only had a minor effect on the
biomass patterns that we observed. The same was the case for pH.
The importance of Nitrogen for forest biomass in rainforests is still un-
certain. Some studies present N as a key limiting factor for tree growth in
lowland forests (Biot et al., 1997; Laurance et al., 1999), whereas others
have shown that N rarely limits tree growth (Sollins, 1998). In the present
study, N was of minor importance for AGWB. Similarly, Assis et al. (2015b)
reported that N only played a secondary role for tree alpha-diversity and
tree species distribution. This is probably because tropical forest substrates
generally have high N concentrations (Bourgeois et al., 1972; Sanchez,
1976; Hedin et al., 2009), and its impact on species distribution is thus
usually low (Huston, 1980). This is especially true for the Amazonian várzea
floodplains, which are generally rich in N and this element therefore is
unlikely to limit tree growth.
In sum, our results revealed that hydroperiod is more important
than edaphic characteristics for AGWB patterns, despite a constant
input of nutrients to the floodplains by the rivers during the high-water
periods. Yet, some individual edaphic elements, such as P availability
and Al toxicity also appear to influence AGWB stocks on Amazonian
floodplains along várzea or paleo-várzea rivers. In addition, AGWB
stocks in floodplain forests do not appear to be determined by an
abundance of large trees and/or species with high WSG in the same way
as in terra firme forests. Instead, our observations suggest that a balance
between several forest structure variables (e.g. tree abundance, average
WSG, average tree heights, etc.) is more important for biomass stocks.
5. Conclusions
The present study highlights that Amazonian forests subjected to
predictable and long-lasting flooding contain a substantial amount of
AGWB. This is of concern since floodplain forests, particularly várzea,
are among the most threatened ecosystems in Amazonia due to direct
anthropogenic influences, such as selective logging, agricultural activ-
ities and hydropower development. In addition, these forests are likely
to be disproportionately affected by climate change (Junk et al., 2010).
Moreover, Brazil has among the highest carbon emissions in the world,
mainly due to deforestation and habitat degradation (IPCC, 2013).
Therefore, policies that better protect forest ecosystems are funda-
mental for reducing greenhouse gas emissions. Data on tree growth,
either obtained through dendrochronological studies or by using
dendrometer bands, are needed to understand how shifts in nutrient
availability influence biomass accumulation and carbon stocks Ama-
zonian floodplains over time.
Acknowledgments
This research was supported by the INPA/Max Planck Project
Manaus, the Brazilian Council of Science and Technology [Universal
479599/2008-4], and PRONEX CNPq-FAPEAM, Áreas Úmidas, MAUA.
RLA was supported by a doctoral scholarship from the Norwegian State
R.L. de Assis et al.
Educational Loan Fund. We are grateful to Sr. José Ramos at the
National Institute for Amazonian Research (INPA) Herbarium for as-
sisting with species identifications. Special thanks to Natália Castro,
Celso Rabelo Costa, José Lima, Bruno Garcia Luize, Thiago Ilnicki,
Jackson de Castro, and all the field assistants who contributed to the
present study. Thank you to Alan Filipe de Souza Oliveira and to
Giulliana Appel for helping to prepare the figure containing the map.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.foreco.2018.09.031.
References
Achard, F., Eva, H.D., Stibig, H.J., Mayaux, P., Gallego, J., Richards, T., Malingreau, J.P.,
2002. Determination of deforestation rates of the world’s humid tropical forests.
Science 297, 999–1002.
Akaike, H., 1981. Modern development of statistical methods. In: Eykhoff, R. (Ed.),
Trends and Progress in System Identification. Pergamon Press, London, pp. 169–184.
Alexander, I.J., 1989. Mycorrhizas in tropical forests. In: Proctor, J. (Ed.), Mineral
Nutrients in Tropical Forest and Savanna Ecosystems. Blackwell Scientific
Publications, Oxford, pp. 169–188.
ANA-Agência Nacional de Águas - ANA, 2010. Hydrological database. <http://hidroweb.
ana.gov.br> (accessed 12 March 2012).
Assis, R.L., Haugaasen, T., Schöngart, J., Montero, J.C., Piedade, M.T.F., Wittmann, F.,
2015a. Patterns of tree diversity and composition in Amazonian floodplain paleo-
várzea forest. J. Veg. Sci. 26, 312–322.
Assis, R.L., Wittmann, F., Piedade, M.T.F., Haugaasen, T., 2015b. Effects of hydroperiod
and substrate properties on tree alpha diversity and composition in Amazonian
floodplain forests. Plant Ecol. 216, 41–54.
Assis, R.L., Wittmann, F., Luize, B.G., Haugaasen, T., 2017. Patterns of floristic diversity
and composition in floodplain forests across four Southern Amazon river tributaries,
Brazil. Flora 229, 124–140.
Avissar, R., Werth, D., 2005. Global hydroclimatological teleconnections resulting from
tropical deforestation. J. Hydrometeorol. 6, 134–145.
Ayres, J.M., 1993. As matas de várzea do Mamirauá. CNPq: Sociedade Civil Mamirauá,
Brasília, DF, Brazil.
Baker, T.R., Phillips, O.L., Malhi, Y., Almeida, S., Arroyo, L., Di Fiore, A., Erwin, T.,
Higuchi, N., Killeen, T.J., Laurance, S.G., Laurance, W.F., Lewis, S.L., Monteagudo,
A., Neill, D.A., Nunez-Vargas, P., Pitman, N.C.A., Silva, J.N.M., Vasquez- Martinez,
R., 2004a. Increasing biomass in Amazonian forest plots. Philos. Trans. R. Soc. Lond.
B 359, 353–365.
Baker, T.R., Phillips, O.L., Malhi, Y., Almeida, S., Arroyo, L., Di Fiore, A., Erwin, T.,
Killeen, T.J., Laurance, S.G., Laurance, W.F., Lewis, S.L., Lloyd, J., Monteagudo, A.,
Neill, D.A., Patiño, S., Pitman, N.C.A., Silva, J.N.M., Vasquez-Martinez, R., 2004b.
Variation in wood density determines spatial patterns in Amazonian forest biomass.
Glob. Chang. Biol. 10, 545–562.
Balslev, H., Lutteyn, J., Yllgaard, B., Holm-Nielsen, L., 1987. Composition and structure of
adjacent unflooded and floodplain forest in Amazonian Ecuador. Opera Bot. 92,
37–57.
Baraloto, C., Rabaud, S., Molto, Q., Blanc, L., Fortunel, C., Hérault, B., Dávila, N.,
Mesones, I., Rios, M., Valderrama, E., Fine, P., 2011. Disentangling stand and en-
vironmental correlates of aboveground biomass in Amazonian forests. Glob. Chang.
Biol. 17, 2677–2688.
Bennet, R.J., Breen, C.M., 1991. The aluminium signal: new dimensions to mechanisms of
aluminium tolerance. Plant Soil 134, 153–166.
Biot, Y., Brilhante, J., Veloso, J., Ferraz, J., Leal-Filho, N., Higuchi, N., Ferreira, S.,
Dejardins, T., 1997. INFORM - O modelo florestal do INPA. In: Higuchi, N., Ferraz, J.,
Antony, L., Luizao, F., Biot, Y., Hunter, I., Proctor, J., Ross, S. (Eds.), Bionte: Biomassa
e Nutrientes Florestais: Relatório Final. National Institute for Research in the Amazon
(INPA), Manaus, Brazil, pp. 271–318.
Bourgeois, W.W., Cole, D.W., Reikerk, H., Gessel, S.P., 1972. Geology and soils of com-
parative ecosystem study areas, Costa Rica. Contribution 11, Tropical Forestry Series,
Institute of Forest Products, University of Washington, Seattle.
Breslow, N.E., Clayton, D.G., 1993. Approximate inference in generalized linear mixed
models. J. Am. Stat. Assoc. 88, 9–25.
Campbell, D.G., Daly, D.C., Prance, G.T., Maciel, U.N., 1986. Quantitative ecological
inventory of terra firme and várzea tropical forest on the Rio Xingu, Brazilian
Amazon. Brittonia 38, 369–393.
Cannell, M.G.R., 1984. Woody biomass of forest stands. Forest Ecol. Manage. 8, 299–312.
Chave, J., Condit, R., Aguilar, S., Hernandez, A., Lao, S., Perez, R., 2004. Error propa-
gation and scaling for tropical forest biomass estimates. Philos. Trans. R. Soc. Lond. B
359, 409–420.
Chave, J., Andalo, C., Brown, S., Cairns, M.A., Chambers, J.Q., Eames, D., Folster, H.,
Formard, F., Higuchi, N., Kira, T., Lescure, J., Nelson, B.W., Ogawa, H., Puig, H.,
Riera, B., Yamakura, T., 2005. Tree allometry and improved estimation of carbon
stocks and balance in tropical forests. Oecologia 145, 87–99.
Chave, J., Coomes, D., Jansen, S., Lewis, S.L., Swenson, N.G., Zanne, A.E., 2009. Towards
a worldwide wood economics spectrum. Ecol. Lett. 12, 351–366.
Clark, D.A., Clark, D.B., Sandoval, R.M., Castro, M.V., 1995. Edaphic and human effects
on landscape-scale distributions of tropical rain forest palms. Ecology 76, 2581–2594.
Condit, R., Hubbell, S.P., Foster, R.B., 1995. Mortality rates of 205 neotropical tree and
shrub species and the impact of a severe drought. Ecol. Monogr. 65, 419–439.
Condit, R., Engelbrecht, B.M.J., Pino, D., Pérez, R., Turner, B.L., 2013. Species
374
Forest Ecology and Management 432 (2019) 365–375
distributions in response to individual soil nutrients and seasonal drought across a
community of tropical trees. Proc. Natl. Acad. Sci. USA 110, 5064–5068.
Crews, T.E., Kitayama, K., Fownes, J.H., Riley, R.H., Herbert, D.A., Mueller-Dombois, D.,
Vitousek, P.M., 1995. Changes in soil phosphorus fractions and ecosystem dynamics
across a long chronosequence in Hawaii. Ecology 76, 1407–1424.
Cuevas, E., Medina, E., 1986. Nutrient dynamics within Amazonian forests: I. Nutrient
flux in fine litter fall and efficiency of nutrient utilization. Oecologia 68, 466–472.
De Simone, O., Müller, E., Junk, W.J., Schmidt, W., 2002. Adaptations of Central Amazon
tree species to prolonged flooding: root morphology and leaf longevity. Plant Biol. 4,
515–522.
DeWalt, S.J., Chave, J., 2004. Structure and biomass of four lowland Neotropical forests.
Biotropica 36, 7–19.
Duivenvoorden, J.F., 1996. Patterns of tree species richness in rain forests of the middle
Caquetá area, Colombia, NW Amazonia. Biotropica 28, 142–158.
EMBRAPA, 1997. Centro Nacional de Pesquisa de Solos. Manual de métodos de análise de
solos, second ed. Rio de Janeiro, Brazil.
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, pp.
291–312.
Ferreira, L.V., 1997. Effects of the duration of flooding on species richness and floristic
composition in three hectares in the Jaú National Park in floodplain forests in central
Amazonia. Biodivers. Conserv. 6, 1353–1363.
Ferry, B., Morneau, F., Bontemps, J.D., Blanc, L., Freycon, V., 2010. Higher treefall rates
on slopes and waterlogged soils result in lower stand biomass and productivity in a
tropical rain forest. J. Ecol. 98, 106–116.
Gentry, A.H., 1988. Tree species richness of upper Amazonian forests. Proc. Natl. Acad.
Sci. USA 85, 156–159.
Gentry, A.H., 1992. Tropical forest biodiversity distributional patterns and their con-
servational significance. Oikos 63, 19–28.
Gibbs, H.K., Brown, S., Niles, J.O., Foley, J.A., 2007. Monitoring and estimating tropical
forest carbon stocks: making REDD a reality. Environ. Res. Lett. 2, 1–13.
Gullison, R.E., Frumhoff, P.C., Canadell, J.G., Field, C.B., Nepstad, D.C., Hayhoe, K.,
Avissar, R., Curran, L.M., Friedlingstein, P., Jones, C.D., Nobre, C., 2007. Tropical
forests and climate policy. Science 316, 985–986.
Haugaasen, T., Peres, C., 2006. Floristic, edaphic and structural characteristics of flooded
and unflooded forests in the lower Rio Purús region of central Amazonia, Brazil. Acta
Amaz. 36, 25–36.
Hawes, E.H., Peres, C., Riley, L.B., Hess, L.L., 2012. Landscape-scale variation in structure
and biomass of Amazonian seasonally flooded and unflooded forests. For. Ecol.
Manag. 281, 163–176.
Hedin, L.O., Brookshire, E.N.J., Menge, D.N.L., Barron, A.R., 2009. The nitrogen paradox
in tropical forest ecosystems. Annu. Rev. Ecol. Evol. Syst. 40, 613–635.
Herbert, D.A., Fownes, J.H., 1995. Phosphorus limitation of forest leaf-area and net
primary production on a highly weathered soil. Biogeochemistry 29, 223–235.
Houghton, R.A., 2005. Tropical deforestation as a source of greenhouse gases. In:
Moutinho, P., Schwartzman, S. (Eds.), Tropical Deforestation and Climate Change.
Instituto de Pesquisa Ambiental da Amazônia (IPAM) e Environmental Defense (ED),
Belém, PA, Brazil, pp. 13–21.
Huston, M., 1980. Soil nutrients and tree species richness in Costa Rican forests. J.
Biogeogr. 7, 147–157.
IPCC, 2013. Climate Change 2013: The Physical Science Basis. Contribution of Working Group
I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change.
Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Irion, G., de Mello, J.A.S.N., Morais, J., Piedade, M.T.F., Junk, W.J., Garming, L., 2010.
Development of the Amazon valley during the middle to late Quaternary: sedi-
mentological and climatological observations. In: Junk, W.J., Piedade, M.T.F.,
Wittmann, F., Schöngart, J., Parolin, P. (Eds.), Central Amazonian Floodplain Forests:
Ecophysiology, Biodiversity and SUSTAINABLE management. Springer, Berlin, pp.
27–42.
Junk, W.J., 1989. Flood tolerance and tree distribution in central Amazonian floodplains.
In: Holm-Nielsen, L.B., Nielsen, I.C., Balslev, H. (Eds.), Tropical Forests: Botanical
Dynamics, Speciation and Diversity. Academic Press, London, pp. 47–64.
Junk, W.J., Piedade, M.T.F., Schöngart, J., Cohn-haft, M., Adeney, J.M., Wittmann, F.,
2011. A classification of major naturally occurring Amazonian lowland wetlands.
Wetlands 31, 623–640.
Junk, W.J., Piedade, M.T.F., Wittmann, F., Schöngart, J., 2010. The role of floodplain forests in
an integrated sustainable management concept of the natural resources of the central
Amazonian Várzea. In: Junk, W.J., Piedade, M.T.F., Wittmann, F., Schöngart, J., Parolin, P.
(Eds.), Central Amazonian Floodplain Forests: Ecophysiology, Biodiversity and Sustainable
Management. Springer, Berlin, pp. 485–509.
Keel, S.H., Prance, G.T., 1979. Studies of the vegetation of a white-sand black-water igapó
(Rio Negro, Brasil). Acta Amaz. 9, 645–655.
Kendall, M.G., 1938. A new measure of rank correlation. Biometrik 30, 81–93.
Kitayama, K., Suzuki, S., Hori, M., Takyu, M., Aiba, S.-I., MajalapLee, N., Kikuzawa, K.,
2004. On the relationship between leaf-litter lignin and net primary productivity in
tropical rain forests. Oecologia 140, 335–339.
Kozlowski, T.T., 1984. Responses of woody plants to flooding. In: Kozlowski, Y.T. (Ed.),
Flooding and Plant Growth. Academic Press, Orlando, pp. 129–164.
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. Manag. 118, 127–138.
Laurance, W., Cochrane, M., Bergen, S., Fearnside, P., Delamonica, P., Barber, C., D’Angelo, S.,
Fernandes, T., 2001. The future of the Brazilian Amazon. Science 291, 438–439.
Lee, Y., Nelder, J.A., 1996. Hierarchical generalized linear models. J. Roy. Stat. Soc. B 58,
619–678.
Lenthe, H.R., 1991. Methods for monitoring organic matter in soil. In: Studies on the
Utilization and Conservation of Soil in the Eastern Amazon Region. Deutsche
Gesellschaft fur Technische Zusammenarbeit, Eschborn, Germany, pp. 119–129.
R.L. de Assis et al.
Lewis, S.L., Phillips, O.L., Baker, T.R., Lloyd, J., Malhi, Y., Almeida, S., Higuchi, N.,
Laurance, W.F., Neill, D.A., Silva, J.N.M., Terborgh, J., Torres Lezama, A., Vásquez
Martinez, R., Brown, S., Chave, J., Kuebler, C., Núñez Vargas, P., Vinceti, B., 2004.
Concerted changes in tropical forest structure and dynamics: evidence from 50 South
American long-term plots. Philos. Trans. R. Soc. B 359, 421–436.
Lewis, S.L., Lopez Gonzalez, G., Sonké, B., Affum-Baffoe, K., Baker, T.R., Ojo, L.O.,
Phillips, O.L., Reitsma, J.M., White, L., Comiskey, J.A., Marie-Noël Djuikouo, K.,
Ewango, C.E.N., Feldpausch, T.R., Hamilton, A.C., Gloor, M., Hart, T., Hladik, A.,
Lloyd, J., Lovett, J.C., Makana, J.R., Malhi, Y., Mbago, F., Ndangalasi, H., Peacock, J.,
Peh, K.S.H., Sheil, D., Sunderland, T., Swaine, M.D., Taplin, J., Taylor, D., Thomas,
S.C., Votere, R., Wöll, H., 2009. Increasing carbon storage in intact African tropical
forests. Nature 457, 1003–1006.
Malhi, Y., Grace, J., 2000. Tropical forests and atmospheric carbon dioxide. Trends Ecol.
Evol. 15, 332–337.
Malhi, Y., Wood, D., Baker, T.R., Wright, J., Phillips, O.L., Cochrane, T., Meir, P., Chave,
J., Almeida, S., Arroyo, L., Higuchi, N., Killeen, T.J., Laurance, S.G., Laurance, W.F.,
Lewis, S.L., Monteagudo, A., Neill, D.A., Vargas, P.N., Pitman, N.C.A., Quesada, C.A.,
Salomao, R., Silva, J.N.M., Lezama, A.T., Terborgh, J.W., Martinez, R.V., Vinceti, B.,
2006. The regional variation of aboveground live biomass in old-growth Amazonian
forests. Glob. Chang. Biol. 12, 1107–1138.
McGrath, D.A., Smith, C.K., Gholz, H.L., Oliveira, F.A., 2001. Effects of land-use change
on soil nutrient dynamics in Amazônia. Ecosystems 4, 625–645.
Melack, J.M., Hess, L.L., 2010. Remote sensing of the distribution and extent of wetlands
in the Amazon basin. In: Junk, W.J., Piedade, M.T.F., Wittmann, F., Schöngart, J.,
Parolin, P. (Eds.), Central Amazonian Floodplain Forests: Ecophysiology, Biodiversity
and Sustainable Management. Springer, Berlin, pp. 43–59.
Oksanen, J., Kindt, R., Legendre, P., O’Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.
H.H., Wagner, H., 2015. Vegan: community ecology package. R Package Version 1.
15-4. <http://CRAN.R-project.org/package=vegan> (accessed Nov 2015).
Palminteri, S., Powell, G.V.N., Asner, G.P., Peres, C.A., 2012. LiDAR measurements of
canopy structure predict spatial distribution of a tropical mature forest primate.
Remote Sens. Environ. 127, 98–105.
Pan, Y., Birdsey, R.A., Fang, J., Houghton, R., Kauppi, P.E., Kurz, W.A., Phillips, O.L.,
Shvidenko, A., Lewis, S.L., Canadell, J.G., Ciais, P., Jackson, R.B., Pacala, S.W.,
McGuire, A.D., Piao, S., Rautiainen, A., Sitch, S., Hayes, D., 2011. A large and per-
sistent carbon sink in the world's forests. Science 333, 988–993.
Paoli, G.D., Curran, L.M., 2007. Soil nutrients limit fine litter production and tree growth
in lowland tropical rain forest of southwestern Borneo. Ecosystems 10, 503–518.
Paoli, G.D., Curran, L.M., Slik, J.W.F., 2008. Soil nutrients affect spatial patterns of
aboveground biomass and emergent tree density in southwestern Borneo. Oecologia
155, 287–299.
Parolin, P., 2009. Submerged in darkness: Adaptations to prolonged submergence by
woody species of the Amazonian floodplains. Ann. Bot. 103, 359–376.
Parolin, P., De Simone, O., Haase, K., Waldhoff, D., Rottenberger, S., Kuhn, U.,
Kesselmeier, J., Schmidt, W., Piedade, M.T.F., Junk, W.J., 2004. Central Amazon
floodplain forests: tree survival in a pulsing system. Bot. Rev. 70, 357–380.
Phillips, O.L., Malhi, Y., Higuchi, N., Laurance, W.F., Nuñez, V.P., Vásquez, M.R.,
Laurance, S.G., Ferreira, L.V., Stern, M., Brown, S., Grace, J., 1998. Changes in the
carbon balance of tropical forest: evidence from long-term plots. Science 282,
439–442.
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., R Core Team, 2018. nlme: Linear and
Nonlinear Mixed Effects Models. R package version 3.1-137, <https://CRAN.R-
project.org/package=nlme> (accessed 15 April 2018).
Prance, G.T., 1979. Notes on the vegetation of Amazonia III. The terminology of
Amazonian forest types subject to inundation. Brittonia 31, 26–38.
Quesada, C.A., Lloyd, J., Schwarz, M., Patiño, S., Baker, T.R., Czimczik, C., Fyllas, N.M.,
Martinelli, L., Nardoto, G.B., Schmerler, J., Santos, A.J.B., Hodnett, M.G., Herrera, R.,
Luizão, F.J., Arneth, A., Lloyd, G., Dezzeo, N., Hilke, I., Kuhlmann, I., Raessler, M.,
Moraes Filho, J., Paiva, F., Araujo Filho, R., Chaves, E., Cruz, O. Junior, Pimentel,
T.P., Paiva, R., 2009. Chemical and physical properties of Amazon forest soils in
relation to their genesis. Biogeosci. Discuss. 6, 3923–3992.
Quesada, C.A., Lloyd, J., Schwarz, M., Patiño, S., Baker, T.R., Czimczik, C., Fyllas, N.M.,
Martinelli, L., Nardoto, G.B., Schmerler, J., Santos, A.J.B., Hodnett, M.G., Herrera, R.,
Luizão, F.J., Arneth, A., Lloyd, G., Dezzeo, N., Hilke, I., Kuhlmann, I., Raessler, M.,
Brand, W.A., Geilmann, H., Moraes Filho, J.O., Carvalho, F.P., Araujo Filho, R.N.,
Chaves, J.E., Cruz Junior, O.F., Pimentel, T.P., Paiva, R., 2010. Variations in chemical
and physical properties of Amazon Forest soils in relation to their genesis.
Biogeosciences 7, 1515–1541.
Raaimakers, D., Boot, R.G.A., Dijkstra, P., Pot, S., Pons, T., 1995. Photosynthetic rates in
relation to leaf phosphorus content in pioneer versus climax tropical rainforest trees.
Oecologia 102, 120–125.
Raich, J.W., Russell, A.E., Crews, T.E., Farrington, H., Vitousek, P.M., 1996. Both nitrogen
and phosphorus limit plant production on young Hawaiian lava flows.
Biogeochemistry 32, 1–14.
Reich, P.B., Ellsworth, D.S., Uhl, C., 1995. Leaf carbon and nutrient assimilation and
conservation in species of differing successional status in an oligotrophic Amazonian
forest. Funct. Ecol. 9, 65–76.
Ruokolainen, K., Linna, A., Tuomisto, H., 1997. Use of Melastomataceae and pter-
idophytes for revealing phyto-geographic patterns in Amazonian rain forests. J. Trop.
Ecol. 13, 243–256.
Rutishauser, E., Wagner, F., Herault, B., Nicolini, E., Blanc, L., 2010. Contrasting
aboveground biomass balance in a Neotropical rainforest. J. Veg. Sci. 21, 672–682.
Ryan, P.R., DiTomaso, J.M., Kochian, L.V., 1993. Aluminium toxicity in roots: an in-
vestigation of spatial sensitivity and the role of the root cap. J. Exp. Bot. 44, 437–446.
Saatchi, S., Malhi, Y., Zutta, B., Buermann, W., Anderson, L.O., Araujo, A.M., Phillips,
O.L., Peacock, J., ter Steege, H., Lopez Gonzalez, G., Baker, T., Arroyo, L., Almeida,
S., Higuchi, N., Killeen, T., Monteagudo, A., Neill, D., Pitman, N., Prieto, A., Salomão,
R., Silva, N., Vásquez Martínez, R., Laurance, W., Ramírez, H.A., 2009. Mapping
landscape scale variations of forest structure, biomass, and productivity in Amazonia.
375
Forest Ecology and Management 432 (2019) 365–375
Biogeosci. Discuss. 6, 5461–5505.
Sanchez, P.A., 1976. Properties and Management of Soils in the Tropics. Wiley, New York.
Santilli, M., Moutinho, P., Schwartzman, S., Depstad, D., Curran, L., Nobre, C., 2005.
Tropical deforestation and the kyoto protocol. Clim. Chang. 71, 267–276.
Schongart, J., Piedade, M.T.F., Ludwigshausen, S., Horna, V., Worbes, M., 2002.
Phenology and stem-growth periodicity of tree species in Amazonian floodplain
forests. J. Trop. Ecol. 18, 581–597.
Schöngart, J., Wittmann, F., Worbes, M., 2010. Biomass and net primary production of
central Amazonian floodplain forests. In: Junk, W.J., Piedade, M.T.F., Wittmann, F.,
Schöngart, J., Parolin, P. (Eds.), Central Amazonian Floodplain Forests:
Ecophysiology, Biodiversity and Sustainable Management. Springer Verlag, Berlin,
pp. 347–388.
Schöngart, J., Junk, W.J., 2007. Forecasting the flood pulse in central Amazonia by
ENSO-indices. J. Hydrol. 335, 124–132.
Silver, W.L., 1994. Is nutrient availability related to plant nutrient use in humid tropical
forests? Oecologia 98, 336–343.
Sioli, H., 1965. A limnologia e a sua importância em pesquisas da Amazônia. Amazoniana
1, 11–35.
Slik, J.W.F., Paoli, G., McGuire, K., Amaral, I., Barroso, J., 2013. Large trees drive forest
aboveground biomass variation in moist lowland forests across the tropics. Global
Ecol. Biogeogr. 22, 1261–1271.
Snowdon, R.E.D., Wheeler, B.D., 1995. Chemical changes in selected wetland species with
increasing Fe supply, with specific reference to root precipitates and Fe tolerance.
New Phytol. 131, 503–520.
Sokal, R.R., Rohlf, F.J., 1995. Biometry, third ed. Freeman, New York.
Sollins, P., 1998. Factors influencing species composition in tropical lowland rain forest:
does soil matter? Ecology 79, 23–30.
Sombroek, W., 2001. Spatial and temporal patterns of Amazon rainfall: Conse- quences
for the planning of agricultural occupation and the protection of primary forests.
Ambio 30, 388–396.
Souza, A.F., Martins, F.R., 2005. Spatial variation and dynamics of flooding, canopy
openness, and structure in a Neotropical swamp forest. Plant Ecol. 180, 161–173.
Stegen, J.C., Enquist, B.J., Ferriere, R., 2009. Advancing the metabolic theory of biodi-
versity. Ecol. Lett. 12, 1001–1015.
Stegen, J.C., Swenson, N.G., Enquist, B.J., White, E.P., Phillips, O.L., Jørgensen, P.M.,
Weiser, M.D., Monteagudo, A., Vargas, P.M., 2011. Variation in aboveground forest
biomass across broad climatic gradients. Glob. Ecol. Biogeogr. 20, 744–754.
Targhetta, N., Kesselmeier, J., Wittmann, F., 2015. Effects of the hydroedaphic gradient
on tree species composition and aboveground wood biomass of oligotrophic forest
ecosystems in the central Amazon basin. Folia Geobot. 50, 185–205.
Taylor, G.J., 1988. The physiology of aluminum phytotoxicity. In: In: Sigel, H., Sigel, A.
(Eds.), Metal Ions in Biological Systems, vol. 24. Marcel Dekker, New York, pp.
123–163.
ter Steege, H., Jetten, V.G., Polak, A.M., Werger, M.J.A., 1993. Tropical rain forest types
and soil factors in a water-shed area in Guyana. J. Veg. Sci. 4, 705–716.
ter Steege, H., Sabatier, S., Castellanos, H., van Andel, T., Duivenvoorden, J., de Oliveira,
A.A., Ek, R.C., Lilwah, R., Maas, P.J.M., Mori, S.A., 2000. An analysis of Amazonian
floristic composition, including those of the Guiana Shield. J. Trop. Ecol. 16,
801–828.
ter Steege, H., Pitman, N.C.A., Phillips, O.L., Chave, J., Sabatier, D., Duque, A., Molino,
J.F., Prévost, M.F., Spichiger, R., Castellanos, H., von Hildebrand, P., Vásquez, R.,
2006. Continental-scale patterns of canopy tree composition and function across
Amazonia. Nature 443, 444–447.
Terborgh, J., Andresen, E., 1998. The composition of Amazonian forests: patterns at local
and regional scales. J. Trop. Ecol. 14, 645–664.
Thomas, W.Y., 2010. The VGAM package for categorical data analysis. J. Stat. Softw. 32,
1–34.
Venables, W.N., Ripley, B.D., 2002. Modern Applied Statistics with S, fourth ed. Springer,
New York.
Vitousek, P., 1982. Nutrient cycling and nutrient use efficiency. Am. Nat. 119, 553–572.
Vitousek, P.M., 1984. Litterfall, nutrient cycling, and nutrient limitations in tropical
forests. Ecology 65, 285–298.
Vitousek, P.M., Sanford Jr., R.L., 1986. Nutrient cycling in moist tropical forest. Annu.
Rev. Ecol. Syst. 17, 137–167.
Walker, T.W., Syers, J.K., 1976. The fate of phosphorus during pedogenesis. Geoderma
15, 1–19.
Walsh, C., Mac Nally, R., 2013. hier.part: Hierarchical Partitioning. R package version 1.
0-4. <http://CRAN.R-project.org/package=hier.p?art> (accessed 15 October
2015).
Wittmann, F., Junk, W.J., 2003. Sapling communities in Amazonian white-water forests.
J. Biogeogr. 30, 1533–1544.
Wittmann, F., Junk, W.J., Piedade, M.T.F., 2004. The várzea forests in Amazonia:
flooding and the highly dynamic geomorphology interact with natural forest suc-
cession. Forest Ecol. Manag. 196, 199–212.
Wittmann, F., Schöngart, J., Montero, J.C., Motzer, T., Junk, W.J., Piedade, M.T.F.,
Queiroz, H.L.D., Worbes, M., 2006. Tree species composition and diversity gradients
in white-water forests across the Amazon Basin. J. Biogeogr. 33, 1334–1347.
Wittmann, F., Schöngart, J., Junk, W.J., 2010. Phytogeography, species diversity, com-
munity structure and dynamics of Amazonian floodplain forests. In: Junk, W.J.,
Piedade, M.T.F., Wittmann, F., Schöngart, J., Parolin, P. (Eds.), Central Amazonian
Floodplain Forests: Ecophysiology, Biodiversity and Sustainable Management.
Springer Verlag, Berlin, pp. 61–104.
Worbes, M., Klinge, H., Revilla, J.D., Martius, C., 1992. On the dynamics, floristic sub-
division and geographical distribution of várzea forests in Central Amazonia. J. Veg.
Sci. 3, 553–564.
Zanne, A.E., Lopez-Gonzalez, G., Coomes, D.A., Ilic, D.A., Jansen, S., Lewis, S.L., Miller,
R.B., Swenson, N.G., Wiemann, M.C., Chave, J., 2009. Data from: towards a world-
wide wood economics spectrum. Dryad Digital Reposit. https://doi.org/10.5061/
dryad.234.