J. For. Res.

https://doi.org/10.1007/s11676-019-00994-x

ORIGINAL PAPER

Fine-scale habitat differentiation shapes the composition,

structure and aboveground biomass but not species richness
of a tropical Atlantic forest
Alice Cristina Rodrigues1 • Pedro Manuel Villa2 • Arshad Ali3 • Walnir Ferreira-Júnior4 •
Andreza Viana Neri1

Received: 16 November 2018 / Accepted: 11 March 2019

 The Author(s) 2019

Abstract Evaluating the influences of fine-scale habitat at breast height C 10 cm were tagged and identified to
heterogeneity on the composition, diversity, structure and species. In each selected area, 100 subplots of 10 9 10 m
functioning of forests is critical to understand how tropical were established. We mainly found that higher topographic
forests will respond to climate change and devise forest variability caused higher habitat differentiation with
management strategies that will enhance biodiversity changes in species composition and community structure,
conservation and aboveground biomass stock. Here, we but did not change species richness. Our habitat-scale
hypothesized that topographic and soil factors determine analyses indicated that, in the less heterogeneous area, the
fine-scale habitat differentiation, which in turn shape distribution of species was more uniform along a fine-scale
community composition, species richness, structure and topographical gradient with no variation in convexity,
aboveground biomass at the local scale in tropical forests. which induced changes in structure and aboveground bio-
To test this hypothesis, we selected two areas (each mass, but not in species richness. The nonsignificant rela-
100 9 100 m) with contrasting fine-scale topographic tionship between species richness and aboveground
conditions where all trees, palms and lianas with a diameter biomass may be attributable to species redundancy or
functional dominance. This study suggests that environ-
Project funding: This research received no external funding. mental filtering is a fundamental process for shaping
community assembly and forest functioning along a local
The online version is available at http://www.springerlink.com. topographical gradient in tropical forests.

Corresponding editor: Chai Ruihai.

Keywords Community–habitat associations  Convexity 
Electronic supplementary material The online version of this Rarefaction  Topographic variability
article (https://doi.org/10.1007/s11676-019-00994-x) contains sup-
plementary material, which is available to authorized users.

& Andreza Viana Neri

Introduction
[email protected]
Understanding the roles of environmental factors for
1
Laboratório de Ecologia e Evolução de Plantas–LEEP, determining community assembly, ecosystem functioning
Departamento de Biologia Vegetal, Universidade Federal de
Viçosa, Viçosa, Minas Gerais CEP 36570-900, Brazil
and biodiversity recovery of tropical forests is a central
2
focus in forest ecology (Ali et al. 2018a; Poorter et al.
Fundación para la Conservación de la Biodiversidad,
Mérida 7101, Mérida, Venezuela
2017; Rozendaal et al. 2019). Tropical forests harbor more
3
than half of the global biodiversity and have a major
College of Forestry, Nanjing Forestry University,
influence on the mitigation of climate change while pro-
Nanjing 210037, Jiangsu, People’s Republic of China
4
viding important ecosystem services that humans depend
Setor de Biologia e Meio Ambiente, Instituto Federal do Sul
on (Lewis et al. 2015). For instance, aboveground biomass
de Minas Gerais -IFSULDEMINAS- Campus Machado,
Highway Machado - Paraguaçú, km 3, Machado, stock, as a key ecosystem property, in tropical forests plays
MG CEP 37750000, Brazil a vital role in the global carbon cycle (Lewis et al. 2015;

123
 A. C. Rodrigues et al.

Anderson-Teixeira et al. 2016) by sequestering carbon gradients and across habitat differentiation in these Atlantic
dioxide, the main contributor to the greenhouse effect forests is crucial for understanding forest functioning.
(Anderson-Teixeira et al. 2016). Several studies have This study aimed to evaluate whether habitat differen-
shown that topographic factors (e.g., convexity, elevation tiation affects the community composition, species rich-
and slope) can determine plant growth (Moeslund et al. ness, structure and aboveground biomass along a local
2013), species distribution (Toledo et al. 2012) and diver- topographical gradient in an Atlantic forest in Minas Gerais
sity patterns (Ali et al. 2018a; Rodrigues et al. 2019). As state, southeastern Brazil. Specifically, we hypothesized
such, topography modulates the relationships between that (1) topographical factors and soil properties determine
species diversity and ecosystem functioning in tropical habitat differentiation; (2) tree community composition,
forests (Ali et al. 2019; Jucker et al. 2018). However, more species richness, structure and aboveground biomass will
research is needed to understand the influences of habitat change in different habitats; and (3) tree diversity is posi-
heterogeneity on species diversity, composition, structure tively related with aboveground biomass along a fine-scale
and aboveground biomass and to understand the main topographical gradient. This study allowed us to investigate
mechanisms underlying the fine-scale community assembly at the community scale whether habitat differentiation
in species-rich, structurally complex natural tropical increases with increasing variability in topographical and
forests. soil factors, thereby leading to increased species richness
Topography can determine resource availability (e.g., and enhanced aboveground biomass at a fine scale.
energy, soil nutrients and climatic water), which provide
different habitats that favor the differential use of resources
by tree species (McEwan and Muller 2006; Brown et al. Materials and methods
2013). The habitat differentiation is important because it
can lead to an increased differential number of coexisting Study area
species and therefore strongly shapes the community
assembly (Brown et al. 2013; Liu et al. 2014; Jucker et al. The seasonal semideciduous Atlantic forest fragment
2018). It is generally well-understood that topographic or studied in Viçosa municipality, Minas Gerais state, south-
geographical gradients in climatic factors (such as mean eastern Brazil (Fig. 1) had been used for shade coffee
annual temperature and precipitation) influence species cultivation under natural forest cover until 1926, when it
diversity, structure and function of the forests over a large became fully protected, allowing for natural regeneration
scale (Jucker et al. 2018; Ali et al. 2019), whereas edaphic to occur. Later, land-use has been reserved (Paula et al.
factors (such as soil physical and chemical properties) did 2002). The study area has a moderate humid tropical cli-
so at a local or fine scale (Chiang et al. 2016; Ali et al. mate, with a dry season from May to September and a wet
2018b). In this context, it is understandable that topo- season between December and March. The mean annual
graphic, climatic and edaphic factors determine the habitats relative humidity is ca. 80%, mean annual air temperature
differentiation in forests (i.e., Wang et al. 2016; Guo et al. is 19 C, and mean annual precipitation is 1340 mm.
2017), and hence may affect community composition, The study area is located between 620 and 820 m a.s.l.,
species richness, structure and functioning of the forests and the relief varies from strongly undulating to moun-
due to the heterogeneity of the available resources (Liu tainous. A Dystric Red-Yellow Latosol soil dominates in
et al. 2014; Guo et al. 2017; Ali et al. 2018a, b, 2019; hilltops and mountainsides, and a Cambic Red-Yellow
Rodrigues et al. 2019). Determining how habitat differen- Podzolic dominates in the upper fluvial terraces
tiation can affect tree communities can also be quite rele- (EMBRAPA 1997).
vant for the management and conservation of forests at a
fine and local scale. Forest inventory and data collection
The Atlantic Forest in Brazil is a hotspot of vascular
plant diversity (Murray-Smith et al. 2009) and has a high We selected two sampling areas with contrasting topo-
capacity for carbon storage in the standing biomass graphic conditions: a southeastern area and a northeastern
(Magnago et al. 2015). Nevertheless, it is also one of the area. Each area (100 9 100 m) was subdivided into 100
most threatened tropical forests in the world, mainly due to plots of 10 9 10 m (Fig. 1). A total of 200 plots (2 ha)
deforestation and fragmentation (Laurance 2009; Ribeiro from the two areas were sampled from December 2016 to
et al. 2011). Currently, only around 10% of the mature January 2017. Within each plot, all trees having a diameter
forests are conserved, while the remnants of native vege- at breast height (DBH) 10 cm were identified to the species
tation are restricted to ca. 20% of the original forest cover level and tagged for the measurement of tree height. All
(Scarano and Ceotto 2015). Hence, evaluating the species individuals were identified using specialized literature,
diversity and community structure along topographical through consultation with the Herbarium of Universidade

123
Fine-scale habitat differentiation shapes the composition, structure and aboveground biomass…

Fig. 1 Location of the study area in relation to South America (a), municipality, Minas Gerais state, southeastern Brazil (c). Location
the Minas Gerais State, Brazil (b), and the forest fragment within the of the two sample plots within the forest fragment (d)
campus of the Federal University of Viçosa (UFV), Viçosa

Federal de Viçosa, or by taxonomists. The Angiosperm The total aboveground biomass per plot was the sum of the
Phylogeny Group IV (APG IV 2016) was used for taxon aboveground biomass of all trees having DBH C 10 cm,
classification. which was then converted to megagrams per hectare
(Mg ha-1) (Ali et al. 2018b). Species-level biomass was
Estimation of aboveground biomass calculated as the sum of the biomass of all individuals from
a given species. Estimation of aboveground biomass was
The aboveground biomass of individual stems was calcu- performed using the R package BIOMASS (Réjou-
lated using the general allometric equation (Eq. 1) pro- Méchain et al. 2017).
posed by Chave et al. (2014), based on tree DBH (cm),
height (H, m) and species wood density (q, g cm-3). We Measurements of topographic variables
used data from the Global Wood Density Database (Zanne
et al. 2009; Chave et al. 2009) to obtain the wood density We measured three topographic variables (i.e., slope, ele-
of each species, using genus or family averages whenever vation, and convexity) within each plot. Topographic
species-level information was not available (e.g., Jucker variables were measured using a total station, which
et al. 2018; Ali et al. 2019). measures vertical and horizontal angles and linear dis-
AGB ¼ 0:0673ðqDBH2 HÞ0:976 ð1Þ tances and is positioned at an obstacle-free location and

123
 A. C. Rodrigues et al.

aimed at a prism. The prism sits on a metal stick and placed this study, the root node consisted of all 100 plots
over the point to be measured. The total station then emits a (10 9 10 m) from each area. Each cluster defines a species
laser beam that reflects in the prism and returns to the assemblage, and the threshold values of environmental
equipment. Using the response time of of the laser beam to variables define an associated habitat type (Larsen and
the equipment and the angle of rotation of the station’s Speckman 2004; Punchi-Manage et al. 2013; Rodrigues
bezel, the internal computer calculates the angles and dis- et al. 2019). MRT analysis was performed using the rpart
tances and stores the data in its internal memory (Kahmen package (Therneau et al. 2017). We represented the spatial
and Faig 1988). The data was then transferred to a com- distribution of topographic variables and species compo-
puter and analyzed with AutoCAD software (Autodesk sition from each area using Kriging maps with the Field
Inc., San Rafael, CA, USA). package (Nychka et al. 2019).
Elevation was calculated using the mean elevation at Species richness in the two sampled areas was evaluated
each of the four corners of the plot. The slope (measured in using both individual-based and sampled-based rarefaction
degrees) was the mean angular deviation of the horizontal and extrapolation curves, which were constructed using the
of each of the four triangular planes formed by the con- first (species richness, q = 0) Hill number (Chao et al.
nection of three of its edges (Guo et al. 2017). Convexity 2014). Extrapolations were made from the abundance data
was determined by subtracting the elevation at the centre of considering between two and three times the total sample
the quadrat from the mean elevation of the eight sur- size by habitat type (Colwell et al., 2012). Sample and
rounding plots. On edge plots, convexity was calculated as individual-based rarefaction/extrapolations were computed
the altitude of the plot of interest minus the mean altitude using the iNEXT package (Hsieh et al. 2016). Rarefaction
of the surrounding plots (Lan et al. 2011). was estimated as the mean of 100 replicate bootstrapping
runs to estimate 95% confidence intervals. Whenever the
Measurements of soil properties 95% confidence intervals did not overlap, species numbers
differed significantly at p \ 0.05 (Colwell et al. 2012).
Within each plot, a composite sample of the surface soil Differences in community composition among habitats
(0–10 cm depth) was collected. Soil properties of the in each of the two areas was evaluated using nonmetric
samples were measured in the Soil Analysis Laboratory of multidimensional scaling (NMDS) and Jaccard similarity
the Federal University of Viçosa, following standard pro- (Clarke 1993) in the R package vegan (Oksanen et al.
tocols (EMBRAPA 1997). The following soil properties 2018). We used permutation multivariate analysis of vari-
were assessed: soil organic carbon (C); total N; available P, ance (PERMANOVA, 9999 permutations) to determine
K, Ca, Mg, Fe, Zn; effective cation exchange capacity differences in species composition, also using vegan (Ok-
(CEC); exchangeable acidity (H ? Al); sum of bases (SB); sanen et al. 2018). We used axis 1 of the NMDS (Euclidian
base saturation (V); aluminum saturation (m); pH and distance) as a proxy for community composition, which
organic matter (OM). explains the highest amount of variance with absence/
presence data (Oksanen et al. 2018; Villa et al. 2018).
Data and statistical analyses Variables were tested for normal distribution by evalu-
ating the Q–Q plot, and homogeneity of variances by
All data and statistical analyses were conducted in R.3.2.2 Bartlett’s test using the dplyr package (Crawley 2012;
(R-Core-Team 2017). Soil properties were summarized Wickham et al. 2018). To compare means of the variables
through principal components analysis (PCA) on the cor- (i.e., topography and soil factors, community composition,
relation matrix, using the FactoMineR package (Husson species richness, abundance, tree height, and aboveground
et al. 2018). For this purpose, all variables were centered biomass) among habitats in each of the two areas, we used
and standardized to meet the assumptions of linearity and a one-way ANOVA (for normally distributed data) fol-
normality. After that, we performed multivariate regression lowed by a post hoc Tukey’s test (p \ 0.05), and PER-
tree (MRT) analysis (De’ath 2002; Larsen and Speckman MANOVA (for non-normally distributed data) followed by
2004) to group habitats with similar species composition a posterior pairwise Adonis test (Martinez-Arbizu 2019).
(i.e., species assemblages) according to topographic vari- All these analyses were performed with the packages stats
ables (Punchi-Manage et al. 2013; Wang et al. 2016; Guo and pairwiseadonis (Martinez-Arbizu 2019).
et al. 2017). MRT is a method of constrained clustering that We constructed a series of linear models to find the most
identifies clusters (a group of plots) that are most similar to parsimonious models explaining the main effects of pre-
each other based on a set of predefined values (De’ath dictor variables (i.e., topographic and soil factors) on the
2002). We then analyzed species dissimilarity (Euclidian response variables (i.e., community composition, species
distance) between each cluster as being defined by richness, abundance, tree height, aboveground biomass)
threshold values of topographic variables (De’ath 2002). In across the local scale topographical gradient. We also

123
Fine-scale habitat differentiation shapes the composition, structure and aboveground biomass…

tested the main effects of species richness on community The southeastern area was topographically less hetero-
structural attributes (i.e., abundance, tree height, above- geneous, with the habitat types determined by the two
ground biomass). We used a generalized linear mixed topographic factors, i.e., elevation and slope, and hence can
effects model (GLMMs) with Poisson error distribution to be divided into five habitats, i.e., (1) high plateau, (2)
investigate the effects of multiple predictors on species intermediate plateau, (3) low plateau, (4) high valley, and
richness. Generalized linear models (GLMs) with negative (5) low valley (Table 1). The northeastern area was topo-
binomial distributions to assess the effects of predictors on graphically more heterogeneous, where habitat types were
abundance. The effects of topographic factors, soil factors determined by all three topographic variables (elevation,
and species richness on aboveground biomass and com- slope, and convexity), and hence can be divided into seven
munity composition were calculated using linear mixed habitats. The northeastern area (NE) consisted of the same
effects models (LMMs) with Gaussian distributions. We types of habitats as the southeastern area (NE), but had two
assessed collinearity between selected predictor variables additional habitat types, i.e., (1) intermediate low valley
using Pearson correlation analysis, and when two variables and (2) a transition area between the high valley and low
were strongly correlated (r C 0.7) they were then included plateau (Table 1). The soil properties were not important in
into the separate models (Fig. S1. Appendix/from Elec- the habitats differentiation. Elevation determined the first
tronic Supplement Material, ESM hereafter). Predictor split of MRT for habitat types in both areas (breakpoint:
variables were grouped into three categories, i.e., topo- SE = 711, NE = 716), accounted for 21–34% of species
graphic variables, soil properties and species richness. The variance, and the second split accounted lower values
topographic variables included elevation, slope and con- (\ 10%) of species variance. In the SE, elevation was also
vexity, whereas soil properties included pH and the main factor for the second (breakpoint = 718) and third
exchangeable acidity (H ? AL). In these models, the (breakpoint = 705) split of habitat types, followed by
identity of the subplots as pseudoreplication in each area slope, which determined fourth split (breakpoint = 29).
was included as a random factor to avoid autocorrelation. Convexity was not included in the MRT for the SE but it
To select the best model, we applied a multi-model was a main factor for the second split (breakpoint = 0.37)
inference approach (Burnham and Anderson 2002) with the in the NE. Elevation determined the third (break-
MuMIn package (Barton 2017), which allows all possible point = 707) and fifth (breakpoint = 703) split and slope
combinations of the explanatory variables included in the determined the fourth (breakpoint = 28) and sixth (break-
global model (Barton 2017). To determine which of these point = 19) splits of habitat types in the NE (Fig. S2. from
variables were the most decisive in explaining changes in ESM).
community composition, species richness, abundance, The NMDS showed that tree species composition varied
community structural and aboveground biomass, we used considerably among habitats along the topographical gra-
an information theoretical approach based on the Akaike dient (Fig. 3). We found no significant difference between
information criterion with a correction for small sample the species richness patterns of the different habitats using
sizes (AICc) and AICc weights (Burnham and Anderson both individual-based and sampled-based rarefaction and
2002). We selected the best model with the lowest AICc extrapolation curves (Fig. 4). Likewise, species richness at
and all models whose difference in AICc with the best the plot level did not differ among habitats (Table 1).
model was less than four units (Burnham et al. 2011). All Community structural attributes and aboveground bio-
models were calculated in R using the packages lme4, mass did not show differences among habitats in the NE
nlme, and MASS (Pinheiro and Bates 2017; Ripley 2017; (Table 1). However, when analyzed at the area scale, basal
Bates et al. 2019). area, tree height and aboveground biomass differed sig-
nificantly between low valleys and high plateaus in the SE.
Subplots in depressions were dominated by species that had
Results high tree maximum height and aboveground biomass, but
low basal area (Table 1). Finally, soil factors differed
We observed an evident differentiation between areas in significantly among habitats only in the SE (Table 1).
relation to topographic and soil variables. The first two The multi-model comparison among habitats within
axes of the PCA explained 66.8% of the commulative areas showed that models including a single topographical
variation in the topographical and soil factors (Fig. 2). The factor (i.e., elevation) consistently explained more varia-
first axis explained 49.6% of variance and was positively tion in community composition than those with multiple
correlated with cation exchange capacity (r = 0.55, effects of topographic and soil factors (Table 2; Table S1.
p \ 0.001) and pH (r = 0.42, p \ 0.001). The second axis from ESM). At local-scale analysis, topographic factors
explained 17.2% of the variation and was positively cor- were better predictors than soil properties. However, spe-
related with elevation but negatively with a slope (Fig. 2). cies richness, abundance, community structural and

123
 A. C. Rodrigues et al.

Fig. 2 Biplot of the principal

component analysis (PCA) for
the topographic and soil factors
of the study area within the
campus of the Federal
University of Viçosa, Viçosa
municipality, Minas Gerais
state, southeastern Brazil. For
analysis, elevation (elev), slope,
convexity (convex), pH,
effective cation exchange
capacity (t), exchangeable
acidity (HAl), sum of bases
(SB), base saturation (V),
organic matter (mo) were
included

aboveground biomass was not related to different predic- niche complementarity hypothesis, i.e., the occurrence of a
tors. There was no significant relationship between species positive diversity effect at the more heterogeneous area due
richness and aboveground biomass at the local scale. to an increased resource use via niche differentiation
Species composition, elevation and slope showed a non- (Tilman 1999; Pausas and Austin 2001; Poorter et al. 2017)
uniform spatial distribution in each area (Fig. 5). caused by the higher topographic variability (Liu et al.
2014; Ali, et al. 2018a). Such variation among areas is
probably due to the heterogeneous distribution of resources
Discussion such as light, water and nutrients (e.g., Katabuchi et al.
2012). We presume that the nonsignificant differences for
Our results showed that topography is an important driver species richness among habitat types in the studied areas
for habitat differentiation, thereby determining community may be due to the marked difference in elevation of the
composition and structure, but not species richness, at the topographic gradient.
local scale in tropical forests. More specifically, we found Community structural attributes including maximum
that elevation and slope were the main factors explaining tree height and basal area as well as aboveground biomass
habitat differentiation in studied areas, and hence deter- differed among habitats in the studied southeastern area
mining variation in community composition. Our results only. In our analyses, maximum tree height decreased from
corroborate the hypothesis that topographic factors deter- valleys to plateaus. Valley areas have a higher availability
mine community composition, structure and aboveground of soil water and nutrients as compared to plateau areas
biomass in tropical forests (Guo et al. 2017; Ali et al. (Moeslund et al. 2013; Liu et al. 2014). The increased tree
2018a; Jucker et al. 2018; Ali et al. 2019). The observed height observed in valleys is a typical feature of trees
habitat structuring along topographical gradients is proba- growing in areas with high resource availability, whereas
bly linked to the underlying spatial variation of light, soil the decreased height of trees from plateau habitats is
nutrients and climatic water availability, which are strongly characteristic of plants in resource-poor areas (Poorter
influenced by topographic factors (John et al. 2007; 2009; Reich 2014). These results are consistent with the
Moeslund et al. 2013; Ali et al. 2019). This study shows findings of a previous study that species distribution along
that environmental filtering is a fundamental process for the topographical gradients can be strongly influenced by
shaping community assembly in tropical forests (e.g., Liu habitat filtering, which selects attributes such as tree height,
et al. 2014), even at a local-scale indicating that species being related to resource use, such as the light niche and
composition changes along a topographical gradient. colonizing strategy (Liu et al. 2014; Ali et al. 2018a).
Our results indicated that species richness, using both While these results indicate that habitat differentiation
individual-based and sampled-based rarefaction and influences the distributions of many individual species, the
extrapolation curves, maintains a similar pattern between community-level effects of soil resource variation have not
different habitats. Although the studied northeastern area been examined extensively, which can be analyzing using
had a species richness higher than the southeastern area at soil nutrients and topographic data to examine their relative
the plot scale, the observed pattern is consistent with the

123
 Table 1 Species richness pattern, community structural attributes, topographic and soil cation exchange capacity (t), exchangeable acidity (H ? Al); sum of bases (SB); base
factors (mean ± SD) of the habitats by area within the campus of the Federal University of saturation (V); organic matter (mo). According to the MRT, the areas were divided into the
Viçosa, Viçosa municipality, Minas Gerais state, southeastern Brazil. Species diversity: following habitats: high plateau (Hp); intermediate plateau (Ip); low plateau (Lp); high valley
species richness (S), stand structural attributes: diameter at breast height (DBH), wood (Hv); low valley (Lv); intermediate low valley (Iv), and ii) a transition area between high
density (Wd); tree height (Hmax), basal area (BA), aboveground biomass (AGB). valley and low plateau (Tpv)
Topographic factors: elevation (elev), slope, convexity (convex). Soil factors: pH, effective
Variable Northeastern site Southeastern site
Hp Ip Lp Tpv Hv Iv Lv Hp Ip Lp Hv Lv

Diversity
S** 12.1 ± 2.4 11.7 ± 2.5 11.0 ± 3.5 10.9 ± 3.50 12.0 ± 3.02 11.5 ± 2.80 9.2 ± 1.69 8.42 ± 3.3 8.96 ± 2.8 8.18 ± 2.40 8.12 ± 2.9 8.37 ± 1.9
Structure
DBH* 12.0 ± 3.3 12.0 ± 2.2 10.6 ± 2.62 10.1 ± 2.43 10.5 ± 3.06 10.3 ± 2.22 11.4 ± 1.90 10.4 ± 3.2 10.5 ± 3.2 9.8 ± 3.4 10.3 ± 2.1 9.38 ± 2.5
Wd** 0.65 ± 0.1 0.65 ± 0.02 0.67 ± 0.02 0.65 ± 0.03 0.66 ± 0.04 0.67 ± 0.35 0.67 ± 0.03 0.66 ± 0.02 0.66 ± 0.03 0.66 ± 0.03 0.64 ± 0.05 0.67 ± 0.1
Hmax* 14.9 ± 1.7 16.37 ± 3.9 16.9 ± 2.8 16.4 ± 4.34 16.0 ± 3.96 16.3 ± 3.38 17.9 ± 4.08 14.7 ± 1.5b 15.7 ± 2.6b 14.6 ± 2.1b 16.9 ± 2.5ab 20.2 ± 5.7a
BA* 21.74 ± 7.3 21.94 ± 8.3 28.9 ± 16.9 33.2 ± 23.40 32.8 ± 21.35 34.5 ± 18.34 30.1 ± 21.90 33.1 ± 9.7b 42.2 ± 15.2a 31.3 ± 14.4bc 43.7 ± 19.8a 22.8 ± 9.3d
a a a ab
AGB* 119.2 ± 60.6 125.9 ± 70.7 214.1 ± 158.0 234.1 ± 220.6 244.2 ± 209.43 242.1 ± 203.6 232.3 ± 225.9 217.4 ± 117.2 205.6 ± 165.1 198.1 ± 164.4 262.3 ± 225.0 228.3 ± 78.7bc
Topography
Elev** 722.6 ± 4.4a 721.3 ± 3.4a 711.5 ± 2.1b 711.5 ± 2.9b 705.2 ± 1.2c 700.0 ± 1.7d 700.3 ± 2.3d 725.6 ± 4.4a 726.2 ± 5.8a 716.0 ± 3.3b 708.6 ± 2.3c 698.0 ± 3.7d
Slope** 21.8 ± 6.2bc 28.7 ± 6.8ab 23.7 ± 2.9b 32.8 ± 4.7a 26.1 ± 9.1b 24.4 ± 5.0b 17.5 ± 1.4c 32.9 ± 2.6a 21.8 ± 4.8c 26.4 ± 6.2b 29.4 ± 3.1a 25.8 ± 6.1b
a b a b b c c
Convex** 1.1 ± 0.5 0.1 ± 0.3 0.9 ± 0.8 0.3 ± 0.9 0.4 ± 3.2 - 1.1 ± 0.9 - 0.9 ± 1.1 2.8 ± 2.9 - 0.3 ± 4.1 0.8 ± 2.7 0.04 ± 1.5 - 0.6 ± 1.7
Soil
pH (H2O)** 5.8 ± 0.8 5.5 ± 0.9 5.6 ± 0.5 5.2 ± 0.7 5.4 ± 0.7 5.2 ± 0.4 5.2 ± 0.4 4.6 ± 0.1b 4.5 ± 0.1b 4.6 ± 0.1b 4.8 ± 0.1ab 5.2 ± 0.5a
Fine-scale habitat differentiation shapes the composition, structure and aboveground biomass…

H ? Al** 4.9 ± 2.5 5.3 ± 2.9 5.1 ± 1.4 5.2 ± 1.9 5.8 ± 2.1 5.6 ± 1.2 5.9 ± 1.0 10.9 ± 0.9ab 12.0 ± 1.1a 10.6 ± 1.2ab 10.4 ± 1.2bc 8.4 ± 2.4c
SB** 6.0 ± 3.2 4.8 ± 3.6 5.4 ± 1.8 3.9 ± 2.5 4.4 ± 2.6 3.7 ± 1.5 3.5 ± 1.5 1.1 ± 0.2c 1.1 ± 0.2c 1.2 ± 0.3c 1.7 ± 0.4b 4.3 ± 4.0a
(t)** 6.3 ± 2.7 5.3 ± 3.1 5.6 ± 1.6 4.4 ± 2.1 4.8 ± 2.2 4.1 ± 1.1 3.9 ± 1.1 2.8 ± 0.2b 3.1 ± 0.4b 2.8 ± 0.2b 3.2 ± 0.4b 5.1 ± 3.6a
d d cd b
V** 55.5 ± 28.7 44.9 ± 32.6 50.5 ± 16.6 36.9 ± 22.0 41.3 ± 23.3 35.6 ± 13.8 32.8 ± 13.7 9.3 ± 2.1 8.7 ± 1.7 10.6 ± 3.1 12.4 ± 1.7 30.8 ± 22.8a
b a b b
mo** 5.7 ± 0.7 5.4 ± 0.8 5.6 ± 0.6 5.3 ± 0.7 5.3 ± 0.6 5.2 ± 0.1 5.2 ± 0.2 6.3 ± 0.9 7.4 ± 1.2 6.2 ± 0.8 6.3 ± 0.5 5.8 ± 0.4bc

Different letters indicate significant differences (P \ 0.05) between habitats in post hoc Tukey’s test (*), and pairwise Adonis test (**)

123
 A. C. Rodrigues et al.

Fig. 3 Nonmetric
multidimensional scaling based
on species composition
according to habitats (point
colors) by study area,
southeastern (a) and
northeastern (b) within the
campus of the Federal
University of Viçosa, Viçosa
municipality, Minas Gerais
state, southeastern Brazil.
According to the MRT, the
areas were divided into the
following habitats: (high plateau
(Hp); intermediate plateau (Ip);
low plateau (Lp); high valley
(Hv); low valley (Lv);
intermediate low valley (Iv),
and transition area between the
high valley and low plateau
(Tpv)

contributions to diversity and aboveground biomass (e.g., a strong correlation between community structure attri-
Chiang et al. 2016; Ali et al. 2018b). butes and abiotic factors (Cornwell and Ackerly 2010) as a
The difference in tree height observed among habitats in trade-off to better withstand conditions of strong winds,
the studied southeastern area correlated with differences in nutrient-poor soils or low soil water availability (Poorter
basal area, whereas habitats with lower tree heights had 2009). We also found a significant decrease in above-
higher basal areas, but the habitat with the tallest tree, i.e., ground biomass among subplots in the studied southeastern
the low valley, had the lowest basal areas. Our data support area, from valleys to plateaus, probably due to the differ-
the results of a previous report that the conditions at high- ences in tree height and basal area among habitats (e.g.,
elevation habitats are apparently well suited for slow- Leuschner et al. 2007). Thus, the local topography is
growing canopy tree species and have higher tree densities related to community structural attributes, with flatter
and basal areas per unit area, compared to low-elevation habitats harboring larger trees with higher biomass and
habitats (Punchi-Manage et al. 2013). Furthermore, there is lower basal area than steeper areas. This fact may be

123
Fine-scale habitat differentiation shapes the composition, structure and aboveground biomass…

Fig. 4 Sample-based and individual-based rarefaction (solid line) plateau (Hp); intermediate plateau (Ip); low plateau (Lp); high valley
and extrapolation curves (dashed lines) of species richness for (Hv); low valley (Lv); intermediate low valley (Iv), and transition
different habitats by northeastern (a, c) and southeastern (b, d) areas area between the high valley and low plateau (Tpv). Rarefaction and
within the campus of the Federal University of Viçosa, Viçosa extrapolation curves present the lines that represent the mean values
municipality, Minas Gerais state, southeastern Brazil. According to and the bands the standard deviation with 95% confidence intervals
the MRT, the areas were divided into the following habitats: high

Table 2 Subset of models

Distribution Response variable Predictors AICc DAICc AICcwt
predicting the species
composition (linear mixed Gaussian lme Species composition Null model 195.17 0 0.53*
effect model). The result of
information-theoretic–based LMMs * Elev 196.14 0.97 0.23*
model selection is indicated. * Elev ? slope 196.7 1.53 0.18*
Only models with DAICc \ 2 * Elev ? convex 199.72 4.55 0.04*
are shown
*Slope 202.35 7.18 0.01
*Convex 207.62 12.45 0.01
The Akaike information criterion corrected for small samples (AICc), difference between one estimated
AICc and the lowest AICc the best model (DAICc), and model weights (AICcwt). *Models that explain
significant main effects

123
 A. C. Rodrigues et al.

Fig. 5 Kriging maps of

topographic variables (elevation
and slope) and species
composition along the
environmental gradient in each
area (100 9 100 m),
northeastern (figures on the left)
and southeastern (figures on the
right) areas within the campus
of the Federal University of
Viçosa, Viçosa municipality,
Minas Gerais state, southeastern
Brazil. Each grid represents a
community weighted variable
value of a 10 9 10 m subplot.
Elevational contours are
indicated by black lines

correlated with the existence of a marked small-scale richness and aboveground biomass, which might be
edaphic gradient since we detected significant differences attributable to functional redundancy or functional domi-
in soil properties among habitats in the southeastern area. nance (Ali et al. 2019). The positive relationship between
Previous studies on forests have reported positive rela- species richness and biomass reported in the aforemen-
tionships between species richness and biomass (Paquette tioned studies is consistent with the sampling, niche com-
and Messier 2011; Poorter et al. 2017). However, our plementarity and facilitation effects (e.g., Poorter et al.
results show a nonsignificant relationship between species 2015; Tilman 1999). On the other hand, the negative

123
Fine-scale habitat differentiation shapes the composition, structure and aboveground biomass…

relationship between species richness and aboveground topographic factors to community composition, structure
biomass indicates that communities having high above- and aboveground biomass along topographical gradients in
ground biomass may exclude weak competitors (Ali et al. a tropical Atlantic forest.
2016), as predicted by competitive exclusion hypothesis
(Grime 1973). Moreover, the direct and indirect effects of Acknowledgements We acknowledge the Conselho Nacional de
Desenvolvimento Cientı́fico e Tecnológico (CNPq) for concession the
environmental factors such as soil and topographic factors scholarship of the first and, PDJ to the second author. We also thank
could also explain such weak or negative relationship Dr. Ferry Slik for his valuable comments on the early version of this
between species richness and aboveground biomass (Jucker article. AA is financially supported by China Postdoctoral Science
et al. 2018; Ali et al. 2019). Foundation (Grant No. 2018M643117) for ecological research at
South China Normal University. We acknowledge the support of
In this study, we found nonsignificant relationship Northeast Forestry University for this publication.
between species richness and aboveground biomass, which
may be attributable to the species redundancy or niche Open Access This article is distributed under the terms of the
overlap in the studied northeastern area, which means that Creative Commons Attribution 4.0 International License (http://crea
tivecommons.org/licenses/by/4.0/), which permits unrestricted use,
despite an increase in species richness, biomass remains distribution, and reproduction in any medium, provided you give
relatively constant without significant variation (Williams appropriate credit to the original author(s) and the source, provide a
and Houseman 2014; Muledi et al. 2017). For instance, in link to the Creative Commons license, and indicate if changes were
the northeastern area, there were fewer stems and more made.
species, probably due to the higher topographic hetero-
geneity, which increases species coexistence. In the
southeastern area, on the other hand, the inverse situation
References
was observed, with the presence of a higher number of
stems and lower number of species, which could analo- Ali A, Yan ER, Chen HYH, Chang SX, Zhao YT, Yang XD, Xu MS
gously be caused by the lower topographic heterogeneity (2016) Stand structural diversity rather than species diversity
(Ali et al. 2018a; Jucker et al. 2018). As such, a recent enhances aboveground carbon storage in secondary subtropical
study has found a weak positive relationship between forests in Eastern China. Biogeosciences 13:4627–4635
Ali A, Lin SL, He JK, Kong FM, Yu JH, Jiang HS (2018a) Climatic
species diversity and aboveground biomass at larger spatial water availability is the main limiting factor of biotic attributes
scales due to species redundancy (Poorter et al. 2015). across large-scale elevational gradients in tropical forests. Sci
Furthermore, other studies in tropical forests have reported Total Environ 647:1211–1221
that functionally dissimilar species may be playing equiv- Ali A, Lohbeck M, Yan E-R (2018b) Forest strata-dependent
functional evenness explains whole-community aboveground
alent roles in the ecosystem processes, e.g., biomass pro- biomass through opposing mechanisms. For Ecol Manag
duction (Lohbeck et al. 2016; Poorter et al. 2017). 424:439–447
Presumably, the higher species richness in the northeastern Ali A, Lin SL, He JK, Kong FM, Yu JH, Jiang HS (2019) Elucidating
area compensates for the aboveground biomass produced space, climate, edaphic and biodiversity effects on aboveground
biomass in tropical forests. Land Degrad Dev. https://doi.org/10.
by the higher number of stems in the southeastern area. 1002/ldr.3278
This result indicated that stocks of aboveground biomass Anderson-Teixeira KJ, Wang MMH, Mcgarvey JC, LeBauer DS
were equivalent between the two studied areas, despite the (2016) Carbon dynamics of mature and regrowth tropical forests
northeastern area having 48% more species and 20% fewer derived from a pantropical database (TropForC-db). Glob Chang
Biol 22:1690–1709
stems than in the southeastern area. Apg IV (2016) An update of the Angiosperm Group classification for
the orders and families of flowering plants: APG IV. Biol J
Linnean Soc 141:399–436
Conclusions Barton K (2017) ‘MuMIn’: multi-model inference. R package version
1.40.0. https://cran.r-project.org/web/packages/MuMIn/MuMIn.
pdf. Accessed 15 Feb 2018
This study reveals that higher variability of topographic Bates D, Maechler M, Ben Bolker B, Walker S, Christensen RHB,
factors determines habitat differentiation with changes in Singmann H, Dai B, Scheipl F, Grothendieck G, Green P, Fox J
tree community composition, structure and aboveground (2019) ‘lme4’: linear mixed-effects models using ‘Eigen’ and
S4. R package version 1.1-21. https://cran.r-project.org/web/
biomass, but not in species richness. Our local-scale anal- packages/lme4/lme4.pdf. Accessed 31 May 2019
yses indicate that in the less heterogeneous area there is a Brown C, Burslem DFRP, Illian JB, Bao L, Brockelman W, Cao M,
more uniform distribution pattern of elevation along with Chang LW, Dattaraja HS, Davies S, Gunatilleke CVS,
no variation in convexity across the gradients, which Gunatilleke IAUN, Huang J, Kassim AR, LaFrankie JV, Lian
J, Lin L, Ma K, Mi X, Nathalang A, Noor S, Ong P, Sukumar R,
induces changes in community structural attributes. On the Su SH, Sun IF, Suresh HS, Tan S, Thompson J, Uriarte M,
other hand, the more heterogeneous area shows no differ- Valencia R, Yap SL, Ye W, Law R (2013) Multispecies
ence among habitats along the topographical gradients. We coexistence of trees in tropical forests: spatial signals of
provide the first indication of the relative importance of

123
 A. C. Rodrigues et al.

topographic niche differentiation increase with environmental Katabuchi M, Kurokawa H, Davies SJ, Tan S, Nakashizuka T (2012)
heterogeneity. Proc R Soc Lond Ser B Biol sci 280:1764 Soil resource availability shapes community trait structure in a
Burnham KP, Anderson DR (2002) Model selection and multimodel species-rich dipterocarp forest. J Ecol 100:643–665
inference: a practical information-theoretic approach. Springer, Lan GY, Hu YH, Cao M, Zhu H (2011) Topography related spatial
New York distribution of dominant tree species in a tropical seasonal rain
Burnham KP, Anderson DR, Huyvaert KP (2011) AIC model forest in China. For Ecol Manag 262:1507–1513
selection and multimodel inference in behavioral ecology: some Larsen DR, Speckman PL (2004) Multivariate regression trees for
background, observations, and comparisons. Behav Ecol Socio- analysis of abundance data. Biometrics 60:543–549
biol 65:23–35 Laurance WF (2009) Conserving the hottest of the hotspots. Biol
Chao A, Gotelli NJ, Hsieh TC, Sander EL, Ma KH, Colwell RK, Conserv 142:1137
Ellison AM (2014) Rarefaction and extrapolation with Hill Leuschner C, Moser G, Bertsch C, Röderstein M, Hertel D (2007)
numbers: a framework for sampling and estimation in species Large altitudinal increase in tree root/shoot ratio in tropical
diversity studies. Ecol Monogr 84:45–67 mountain forests of Ecuador. Basic Appl Ecol 8:219–230
Chave J, Coomes D, Jansen S, Lewis SL, Swenson NG, Zanne AE Lewis SL, Edwards DP, Galbraith D (2015) Increasing human
(2009) Towards a worldwide wood economics spectrum. Ecol dominance of tropical forests. Science 349:827–832
Lett 12:351–366 Liu J, Yunhong T, Slik JWF (2014) Topography related habitat
Chave J, Réjou-Méchain M, Búrquez A et al (2014) Improved associations of tree species traits, composition and diversity in a
allometric models to estimate the aboveground biomass of Chinese tropical forest. For Ecol Manag 330:75–81
tropical trees. Glob Chang Biol 20:3177–3190 Lohbeck M, Bongers F, Martinez-Ramos M, Poorter L (2016) The
Chiang J-M, Spasojevic MJ, Muller-Landau HC, Sun I-F, Lin Y, Su importance of biodiversity and dominance for multiple ecosys-
S-H, Chen Z-S, Chen C-T, Swenson NG, McEwan RW (2016) tem functions in a human-modified tropical landscape. Ecology
Functional composition drives ecosystem function through 97:2772–2779
multiple mechanisms in a broadleaved subtropical forest. Magnago LFS, Magrach A, Laurance WF, Martins SV, Meira-Neto
Oecologia 182:829–840 JA, Simonelli M, Edwards DP (2015) Would protecting tropical
Clarke KR (1993) Non-parametric multivariate analyses of changes in forest fragments provide carbon and biodiversity cobenefits
community structure. Aust J Ecol 18:117–143 under REDD?? Glob Chang Biol 21:3455–3468
Colwell RK, Chao A, Gotelli NJ, Lin SY, Mao CH, Chazdon RL, Martinez-Arbizu P (2019) pairwiseAdonis: pairwise multilevel
Longino JT (2012) Models and estimators linking individual- comparison using adonis. R package version 2.4-5. https://
based and sample based rarefaction, extrapolation and compar- github.com/pmartinezarbizu/pairwiseAdonis
ison of assemblages. J Plant Ecol 5:3–21 McEwan RW, Muller RN (2006) Spatial and temporal dynamics in
Cornwell WK, Ackerly DD (2010) A link between plant traits and canopy dominance of an old-growth mixed mesophytic forest.
abundance: evidence from coastal California woody plants. Can J For Res 36:1536–1550
J Ecol 98:814–821 Moeslund JE, Arge L, Bøcher PK, Dalgaard T, Svenning J-C (2013)
Crawley MJ (2012) The R book, 2nd edn. Wiley, London Topography as a driver of local terrestrial vascular plant
De’ath G (2002) Multivariate regression trees: a new technique for diversity patterns. Nord J Bot 31:129–144
modeling species-environment relationships. Ecology Muledi JI, Bauman D, Drouet T, Vleminckx J, Jacobs A, Lejoly J,
83:1105–1117 Meerts P, Shutcha MN (2017) Fine-scale habitats influence tree
EMBRAPA (1997) Manual de métodos de análises de solo, 2nd edn. species assemblage in a Miombo forest. J Plant Ecol 10:958–969
Centro Nacional de Pesquisa de Solos, Empresa Brasileira de Murray-Smith C, Brummitt NA, Oliveira-Filho AT, Bachman S,
Pesquisa Agropecuária, Rio de Janeiro Moat J, Lughadha EM, Lucas EJ (2009) Plant diversity hotspots
Grime JP (1973) Competitive exclusion in herbaceous vegetation. in the Atlantic Coastal Forests of Brazil. Conserv Biol
Nature 242:344–347 23:151–163
Guo Y, Wang B, Mallik AU, Huang F, Xiang W, Ding T, Wen S, Lu Nychka D, Furrer R, Paige J, Sain S (2019). ‘fields’ package: tools for
S, Li D, He Y, Li X (2017) Topographic species-habitat spatial data. R package version 2.4-5. https://cran.r-project.org/
associations of tree species in a heterogeneous tropical forest. web/packages/fields/fields.pdf. Accessed 10 Feb 2018
J Plant Ecol 57:1–10 Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn
Hsieh TC, Ma KH, Chao A. (2016) ‘iNEXT’: iNterpolation and D, Minchin PR, O’Hara RB, Simpson GL, Solymos P, Stevens
EXTrapolation for species diversity. R package version 2.0.12.. MHH, Szoecs E, Wagner E (2018) ‘Vegan’: community ecology
https://cran.r-project.org/web/packages/iNEXT/iNEXT.pdf package. R package version 2.4-6. https://cran.r-project.org/web/
Accessed 10 Dec 2017 packages/vegan/vegan.pdf. Accessed 5 Feb 2018
Husson F, Josse J, Le S, Mazet J (2018) ‘‘FactoMineR’’ package Paquette A, Messier C (2011) The effect of biodiversity on tree
multivariate: exploratory data analysis and data mining. RStudio productivity: from temperate to boreal forests. Glob Ecol Biog
package version 1.0.14. https://cran.r-project.org/web/packages/ 20:170–180
FactoMineR/FactoMineR.pdf Paula A, Silva AF, Souza AL, Santos FAM (2002) Alterações
John R, Dalling JW, Harms KE, Yavitt JB, Stallard RF, Mirabello M, florı́sticas ocorridas num perı́odo de quatorze anos na vegetação
Hubbell SP, Valencia R, Navarrete H, Vallejo M, Foster RB arbórea de uma floresta estacional semidecidual em Viçosa-MG.
(2007) Soil nutrients influence spatial distributions of tropical Rev Árvore 26:743–749
tree species. Proc Natl Acad Sci USA 104:864–869 Pausas JG, Austin MP (2001) Patterns of plant species richness in
Jucker T, Bongalov B, Burslem DFRP, Nilus R, Dalponte M, Lewis relation to different environments: an appraisal. J Veg Sci
SL, Phillips OL, Qie L, Coomes DA (2018) Topography shapes 12:153–166
the structure, composition and function of tropical forest Pinheiro J, Bates D (2017) ‘nlme’: linear and nonlinear mixed effects
landscapes. Ecol Lett 21:989–1000 models. R package version 3.1-131. https://cran.r-project.org/
Kahmen H, Faig W (1988) Surveying. Walter Gruyter e Co, Berlin, web/packages/nlme/nlme.pdf. Accessed 20 Jan 2018
p 578 Poorter L (2009) Leaf traits show different relationships with shade
tolerance in moist versus dry tropical forests. New Phytol
181:890–900

123
Fine-scale habitat differentiation shapes the composition, structure and aboveground biomass…

Poorter L, van der Sande MT, Thompson J, Arets EJMM, Alarcón A, Oliveira A, Denslow JS, Dent DH, DeWalt SJ, Dupuy JM, Durán
Álvarez-Sánchez J, Ascarrunz N, Balvanera P, Barajas-Guzmán SM, Dutrieux LP, Espı́rito-Santo MM, Fandino MC, Fernandes
G, Boit A, Bongers F, Carvalho FA, Casanoves F, Cornejo- GW, Finegan B, Garcı́a H, Gonzalez N, Moser VG, Hall JS,
Tenorio G, Costa FRC, de Castilho CV, Duivenvoorden JF, Hernández-Stefanoni JL, Hubbell S, Jakovac CC, Hernández AJ,
Dutrieux LP, Enquist BJ, Fernández-Méndez F, Finegan B, Junqueira AB, Kennard D, Larpin D, Letcher SG, Licona J-C,
Gormley LHL, Healey JR, Hoosbeek MR, Ibarra-Manrı́quez G, Lebrija-Trejos E, Marı́n-Spiotta E, Martı́nez-Ramos M, Massoca
Junqueira AB, Levis C, Licona JC, Lisboa LS, Magnusson WE, PES, Meave JA, Mesquita RCG, Mora F, Müller SC, Muñoz R,
Martı́nez-Ramos M, Martı́nez-Yrizar A, Martorano LG, Maskell de Oliveira Neto SN, Norden N, Nunes YRF, Ochoa-Gaona S,
LC, Mazzei L, Meave JA, Mora F, Muñoz R, Nytch C, Ortiz-Malavassi E, Ostertag R, Peña-Claros M, Pérez-Garcı́a
Pansonato MP, Parr TW, Paz H, Pérez-Garcı́a EA, Renterı́a LY, EA, Piotto D, Powers JS, Aguilar-Cano J, Rodriguez-Buritica S,
Rodrı́guez-Velazquez J, Rozendaal DMA, Ruschel AR, Rodrı́guez-Velázquez J, Romero-Romero MA, Ruı́z Jorge,
Sakschewski B, Salgado-Negret B, Schietti J, Simões M, Sinclair Sanchez-Azofeifa A, de Almeida AS, Silver WL, Schwartz
FL, Souza PF, Souza FC, Stropp J, ter Steege H, Swenson NG, NB, Thomas WW, Toledo M, Uriarte M, Sampaio EVS, van
Thonicke K, Toledo M, Uriarte M, van der Hout P, Walker P, Breugel M, van der Wal H, Martins SV, Veloso MDM, Vester
Zamora N, Peña-Claros M (2015) Diversity enhances carbon HFM, Vicentini A, Vieira ICG, Villa PM, Williamson GB,
storage in tropical forests. Glob Ecol Biogeog 24:1314–1328 Zanini KJ, Zimmerman J, Poorter L (2019) Biodiversity
Poorter L, van der Sande MT, Arets EJMM, Ascarrunz N, Enquist BJ, recovery of Neotropical secondary forests. Sci Adv 5:3114
Finegan B, Licona JC, Martı́nez-Ramos M, Mazzei L, Meave Scarano FR, Ceotto P (2015) Brazilian Atlantic forest: impact,
JA, Muñoz R, Nytch CJ, de Oliveira AA, Pérez-Garcı́a EA, vulnerability, and adaptation to climate change. Biodivers
Prado-Junior J, Rodrı́guez-Velázques J, Ruschel AR, Salgado- Conserv 24:2319
Negret B, Schiavini I, Swenson NG, Tenorio EA, Thompson J, Therneau T, Atkinson B, Ripley B (2017) ‘rpart’: recursive
Toledo M, Uriarte M, van der Hout P, Zimmerman JK, Peña- partitioning and regression trees. R package version 4.1-11.
Claros M (2017) Biodiversity and climate determine the https://CRAN.R-project.org/package=rpart. Accessed 4 March
functioning of Neotropical forests. Glob Ecol Biogeogr 2018
26:1423–1434 Tilman D (1999) The ecological consequences of changes in
Punchi-Manage R, Getzin S, Wiegand T, Kanagaraj R, Savitri biodiversity: a search for general principles. Ecology
Gunatilleke CV, Nimal Gunatilleke IA, Wiegand K, Huth A 80:1455–1474
(2013) Effects of topography on structuring local species Toledo M, Peña-Claros M, Bongers F, Alarcón A, Balcázar J,
assemblages in a Sri Lankan mixed dipterocarp forest. J Ecol Chuviña J, Leaño C, Licona JC, Poorter L (2012) Distribution
101:149–160 patterns of tropical woody species in response to climatic and
R Core Team (2017) R: A language and environment for statistical edaphic gradients. J Ecol 100:253–263
computing. R Foundation for Statistical Computing, Vienna, Villa PM, Martins SV, Oliveira Neto SN, Rodrigues AC, Martorano
Austria. https://www.R-project.org/. Accessed 15 Feb 2018 L, Delgado L, Cancio NM, Gastauerg M (2018) Intensification
Reich PB (2014) The world-wide ‘fast-slow’ plant economics of shifting cultivation reduces forest resilience in the northern
spectrum: a traits manifesto. J Ecol 102:275–301 Amazon. For Ecol Manag 430:312–320
Réjou-Méchain M, Tanguy A, Piponiot C, Chave J, Hérault B (2017) Wang Q, Punchi-Manage R, Lu Z, Franklin SB, Wang Z, Li Y, Chi X,
biomass: an r package for estimating above-ground biomass and Bao D, Guo Y, Xu JLY, Qiao X, Jiang M (2016) Effects of
its uncertainty in tropical forests. Methods Ecol Evol topography on structuring species assemblages in a subtropical
8:1163–1167 forest. J Plant Ecol 10:440–449
Ribeiro MC, Martensen AC, Metzger JP, Tabarelli M, Scarano F, Wickham H, François R, Henry L, Müller K (2018). ‘dplyr’ package:
Fortin M-J (2011) The Brazilian Atlantic Forest: a shrinking a grammar of data manipulation. R package version 7.3-48.
biodiversity hotspot. In: Zachos F, Habel J (eds) Biodiversity https://cran.r-project.org/web/packages/dplyr/dplyr.pdf. Acces-
hotspots. Springer, Berlin sed 15 Feb 2018
Ripley B (2017) ‘MASS’: support functions and datasets for venables Williams BM, Houseman GR (2014) Experimental evidence that soil
and RIPLEY’s MASS. R package version 7.3-48. https://cran.r- heterogeneity enhances plant diversity during community
project.org/web/packages/MASS/MASS.pdf. Accessed 1 Feb assembly. J Plant Ecol 7:461–469
2018 Zanne AE, Lopez-Gonzalez G, Coomes DA, Ilic J, Jansen S, Lewis
Rodrigues AC, Villa PM, Viana NA (2019) Fine-scale topography SL, Miller RB, Swenson NG, Wiemann MC, Chave J (2009)
shape richness, community composition, stem and biomass Data from: towards a worldwide wood economics spectrum.
hyperdominant species in Brazilian Atlantic forest. Ecol Indic Dryad digital repository. https://doi.org/10.5061/dryad.234.
102:208–217 Accessed 3 Dec 2017
Rozendaal DMA, Bongers F, Aide TM, Alvarez-Dávila E, Ascarrunz
N, Balvanera P, Becknell JM, Bentos TV, Brancalion PHS, Publisher’s Note Springer Nature remains neutral with regard to
Cabral GAL, Calvo-Rodriguez S, Chave J, César RG, Chazdon jurisdictional claims in published maps and institutional affiliations.
RL, Condit R, Dallinga JS, de Almeida-Cortez JS, Jong B,

123