Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

Contents lists available at ScienceDirect

Perspectives in Plant Ecology, Evolution and Systematics

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

Longer research articles

Dispersal limitation, soil, and fire affect functional properties of tropical

secondary forests on abandoned cattle ranching landscapes
Francisco S. Álvarez a, b, *, Bryan Finegan a, Diego Delgado a, Zayra Ramos a, Luis P. Utrera a,
Vanessa Granda a
a
CATIE-Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica
b
Fundación Naturaleza El Salvador, Pasaje Francisco Campos #166, Colonia Escalón, San Salvador, El Salvador

A R T I C L E I N F O A B S T R A C T

Keywords: Forest age is a major predictor of secondary forest functional properties and through the chronosequence
Bark thickness approach continues to be a principal focus in local, landscape, and regional-scale studies of secondary succession.
Community Weighted Mean Recent work has shown that patterns of temporal change in functional properties differ markedly between wet
Fire
and seasonally dry lowland forests, suggesting that decreasing light and increasing water availability, respec­
Functional trait
Resprouting
tively, are the main drivers of successional change. Meanwhile, however, the potentially marked effects of
Wood specific gravity anthropogenic factors (especially fire), soils, climate, and dispersal limitation on the variation of forest char­
acteristics over landscapes remain poorly understood. We studied the functional properties of seasonally dry
secondary forests 5–35 years after pasture abandonment on the Nicoya Peninsula, Costa Rica. We measured 11
functional traits for 63 dominant tree and palm species in 52 plots of 0.12 ha. We used linear regression and
variation partitioning to determine the relative importance of soil, climate, site use, fire history, spatial factors,
and forest age in the determination of community weighted mean (CWM) trait values. CWM leaf trait and stem
trait spectra of the 52 plots were orthogonal in a PCA ordination. Our seasonal forest hypothesis, that forest
functional properties become more acquisitive with age as suggested by other authors, was not supported,
perhaps because our chronosequence was relatively short. Our fire tolerance hypothesis was that bark thickness,
wood specific gravity and resprouting capacity would increase with time of exposure to fires. This hypothesis was
not supported for bark thickness. Rather, our results suggest that wood specific gravity and resprouting capacity
are better predictors of fire tolerance. Also, our results suggest that >10 years of exposure to fire generates
changes in forest fire tolerance strategies from high CWM bark thickness to high wood specific gravity. Finally,
we tested our dispersal limitation hypothesis, that spatial variables, expressed as principal coordinates of
neighbor matrices (PCNM) eigenfunctions, predict variation in forest functional properties, using variation
partitioning analysis with matrices of climate, soil, anthropogenic and spatial variables. The results (overall
model R2 = 0.46) indicated that spatial variables, followed by soil (acidity, depth, and extractable Mn), are the
best overall predictors of forest functional traits values, supporting this hypothesis. Overall, fire, dispersal lim­
itation and soil characteristics explain the functional properties of these secondary forests, with no effect of age
in the 5–35 years range. This indicates a critical need for sampling designs and analytical approaches that take
into account all these factors to advance understanding of tropical seasonal forest recovery at the landscape scale.

1. Introduction Rozendaal et al., 2019). Many studies have attempted to explain the
processes of forest recovery through secondary succession in terms of
Most of the world’s forests have undergone changes as a consequence the relationship between age after site abandonment, floristics diversity,
of human activity (Chazdon, 2003; Chazdon et al., 2007; Steffen et al., and forest structure. As a result, we now know that rates and trajectories
2004) and large forest areas are recovering through natural regeneration of secondary succession are affected by soil conditions, the proximity of
on abandoned agricultural and cattle ranching land (Chazdon, 2014; seed sources (Chazdon, 2003, 2008; Chazdon et al., 2007; Smith et al.,

* Corresponding author at: CATIE-Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica.
E-mail address: francisco.alvarez@catie.ac.cr (F.S. Álvarez).

https://doi.org/10.1016/j.ppees.2021.125632
Received 24 September 2020; Received in revised form 19 July 2021; Accepted 16 August 2021
Available online 18 August 2021
1433-8319/© 2021 The Authors. Published by Elsevier GmbH. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

1997), seed dispersal capacity, and differences between species in policies and the decline in international beef prices, large areas were
growth rates, longevity, and shade-tolerance (Finegan, 1996). Although abandoned and their vegetation cover has been recovered through
functional ecology has been a predominant paradigm for some time natural regeneration (Stan and Sanchez-Azofeifa, 2019). Previous work
(McGill et al., 2006), its application to understanding tropical secondary on the seasonally dry forests to the north of our study area shows that
forest succession is relatively recent. Also, the predominant focus on age abandoned pasture areas may be colonized by slow-growing tree species
as a predictor variable, frequently through chronosequence approaches, that can dominate the forest for many decades (Finegan and Nasi, 2004;
has obscured understanding of the effects of other factors on succession, Janzen, 1988). Granda Moser et al. (2015) showed that on the Nicoya
this is especially so at the landscape scale. Peninsula, seasonally dry forests 5–35 years old, exposed to fire for
We carried out our study in seasonally dry secondary forests of the periods of up 25 years, belonged to three floristically distinct types of
Nicoya Peninsula, northwestern Costa Rica. We characterize these for­ secondary forest. The floristic variation explained by environmental
ests as seasonally dry because mean annual rainfall values <1800 mm factors, anthropogenic variables, and spatial factors (PCNM eigenfunc­
predominate across the peninsula, with a 5–6 month dry season whose tions) was low, although anthropogenic variables had an important in­
rainfall values are <100 mm per month, and because the vegetation is fluence on the floristic composition.
drought-deciduous (DRYFLOR et al., 2016; Gei et al., 2018). Recent Here, we apply a functional trait approach to investigate the drivers
traits-based studies have highlighted differences between wet and that shape the characteristics of these new forests and their variation
seasonally dry lowland secondary forest successions (e.g. Poorter et al., across this landscape. We measured eleven functional traits related to
2019). On the one hand, in wet forests, successional change may be the acquisitive-conservative plant spectrum (Reich, 2014) as well as to
driven by the gradient of decreasing light availability with forest age, so fire tolerance, for each of the 63 tree and palm species that form 80 % of
that forest functional properties become more conservative over time the basal area in 52 plots established by Granda Moser et al. (2015).
(Lohbeck et al., 2015, 2013). Succession in seasonally dry forests like Community weighted mean (CWM) trait values of each trait were
those of our study area, on the other hand, may be driven by increasing calculated using the relative basal area as the weighting variable. We
water availability with forest age, so that initially conservative forest tested three hypotheses related to forest age, community assembly
functional properties become more acquisitive over time (Buzzard et al., models, and site factors that are known to influence the characteristics of
2015; Lohbeck et al., 2015). However, at the landscape scale, soil, secondary forests. Because the study area has a strong dry season and the
topography, and the type and intensity of previous land use may forest is drought-deciduous, we first tested the hypothesis that CWM
contribute to explaining the variation of forest functional properties traits become more acquisitive with stand age as suggested by some
with succession and across the landscape (Becknell and Powers, 2014; authors (see Lohbeck et al., 2013). For example, that CWM wood specific
Craven et al., 2018; Derroire et al., 2018; González-M et al., 2019; gravity decreases, and CWM specific leaf area increases (Lohbeck et al.,
Sanaphre-Villanueva et al., 2017; Santiago-García et al., 2019; Werden 2015; Poorter et al., 2019). We call this the seasonal forest hypothesis.
et al., 2018). Despite this, the operation of these drivers at the landscape Concerning site factors, because dry season fires are common in this
scale is not well understood. Also, seasonally dry secondary forests like landscape, we tested the hypothesis that values of traits involved in fire
those of our study area, are likely to be exposed to dry season fires tolerance (CWM bark thickness, CWM wood specific gravity, and
entering the forest from adjacent pastures or starting within the forest resprouting capacity) would increase with the length of the time period
itself (Janzen and Hallwachs, 2020). Therefore, it is necessary to during which forests were exposed to fires. We call this the fire tolerance
consider the measurement of an appropriate set of traits to understand hypothesis. Because tradeoffs among stem traits associated with fire
the different response mechanisms of forests to a broad set of environ­ tolerance are less well-explored than those among, for example, leaf
mental factors and disturbances along the succession, for example, bark traits (Poorter et al., 2014), we complemented the testing of the fire
thickness, wood specific gravity, and resprouting capacity, which are tolerance hypotheses with analysis of the relationships among CWM
key traits related to fire tolerance (Brando et al., 2012; Charles-Domi­ traits. Finally, we used variation partitioning to determine the relative
nique et al., 2017; Clarke et al., 2012; Jackson et al., 1999; Pausas, 2017; importance of multiple variables that can explain the CWM traits values
Pausas et al., 2004; Poorter et al., 2019; Rosell, 2016). Therefore, the of the forest. In variation partitioning, we used matrices of climatic, soil,
variability of environmental conditions and disturbances that forests anthropogenic, and spatial variables, as well as forest age, as predictors.
face cause responses through differentiated functional strategies that We hypothesized that spatial variables, PCNM eigenfunctions generated
can involve tradeoffs between leaf and stem traits (Baraloto et al., 2010; from a matrix of geographical distances among sample plots (Borcard
Chave et al., 2009; Reich, 2014; Wright et al., 2004), and some these and Legendre, 2002), would be significantly related to forest CWM
tradeoffs can be decoupled from each other (Baraloto et al., 2010) or functional traits values, after accounting for the effects of age, climate,
highly coupled (Reich, 2014). anthropogenic, and soil factors. We call this the dispersal limitation
Our overarching question was, how do the functional properties of hypothesis.
secondary forests change over time, across gradients of environmental
factors and human influence, and in relation to spatial factors that may 2. Material and methods
suggest the operation of dispersal limitation? It is necessary to consider
all these factors as potential drivers of variation in the functional 2.1. Study area and secondary forest characteristics
properties of the forest along the succession. Spatial factors such as the
geographical distances that are involved in dispersal limitation are The study was carried out on the Nicoya Peninsula of northwestern
especially important. This is because species can vary widely in their Costa Rica (Fig. A.1). Most of the study area belongs to the tropical moist
functional characteristics (Díaz et al., 2004; Poorter et al., 2019). forest life zone (Holdridge, 1967), although the area closer to the coast
Geographical distance –spatial variables– can explain the variation of corresponds to the transition to the tropical dry forest. However,
tree species composition, perhaps through dispersal limitation and because the Holdridge life zone system does not take into account sea­
ecological drift (Legendre et al., 2009; Rosindell et al., 2011). Significant sonality, we classify these forests as seasonally dry following recent
effects of geographical distance on species composition have been literature, because mean annual rainfall values <1800 mm predominate
demonstrated for tropical secondary forests (van Breugel et al., 2019). across the peninsula, with a 5–6 month dry season whose rainfall values
Therefore, spatial variables may also be related to the variation of sec­ are <100 mm per month, and because the vegetation is
ondary forest functional traits values. drought-deciduous (DRYFLOR et al., 2016; Gei et al., 2018). The altitude
The Nicoya Peninsula, our study area, was largely deforested during on the peninsula reaches 1018 masl, though most of the landscape lies at
the 1960s due to the expansion of agriculture and especially cattle <200 m (Granda Moser et al., 2015). The altitude of our 52 sample plots
ranching (Arroyo-Mora et al., 2005). However, due to changes in public (see below) ranged between 50 to ~800 masl. WorldClim data (Fick and

2
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

Hijmans, 2017), indicate a 22.6 ◦ C–26.6 ◦ C range of mean annual CWM for the eleven functional traits, FDiversity software was used
temperature and a mean annual rainfall ranging from 1710 mm to 2458 (Casanoves et al., 2011).
mm. The area has a well-defined dry season from late December to April
and a rainy season from May to early December. Soils are mainly Alfisols 2.3. Soil, climate, and site history data
characterized by low to moderate fertility, neutral pH, low acidity, low
Al, moderate P, N, and K, and high organic matter and cations (Bertsch In each plot, soil physicochemical properties and topographical
et al., 2000; Granda Moser et al., 2015; Sanchez, 1981). The secondary variables were measured (henceforth “soil” variables). For this, three
forests are dominated by tree species adapted to the cattle ranching soil samples were collected to a depth of 30 cm and mixed to form a
landscape, with three floristic types recognized: Schizolobium-Miconia composite sample. Soil depth was measured at each of these three points
forest, Semialarium-Lonchocarpus forest, and Guazuma forest (Granda with a metal rod 1.10 m long (Sesnie et al., 2009) and recorded using a
Moser et al., 2015). scale of four categories: a) deep, > 90 cm; b) moderately deep, 50− 90
We sampled secondary forests identified by farmers as being on cm; c) surface, 25− 50 cm; d) superficial, <25 cm. Analysis of soil
abandoned pastures, with ages after abandonment in the range of 5–35 samples was carried out in the laboratory of the Tropical Agricultural
years. These forests had been exposed to fire for periods of up to 25 Centre for Research and Higher Education (CATIE). Samples were
years. We measured functional traits for the dominant tree and palm air-dried. Soil texture (percent of sand, clay, and silt) was obtained by
species of 52 plots of 0.12 ha distributed across the landscape (Granda the Bouyoucos method. Total acidity, Ca and Mg extractions were ob­
Moser et al., 2015). Plots were established on slopes < 45 ◦ , >50 m from tained in 1 M potassium chloride (KCl). Extractable K, P, Zn, Cu, Mn, and
water bodies, and in locations separated by at least 300 m. Species for Fe were obtained by modified Olsen extraction with a 0.5 N sodium
trait measurement were identified as those that contributed to approx­ bicarbonate (NaHCO3) solution at pH 8.5. The percentage of C and N
imately 80 % of plot basal area ≥5 cm diameter at breast height (DBH, were determined by dry combustion. Slope (%) was measured with a
1.3 m) (Garnier et al., 2004; Grime, 1998). Sixty-three species, all clinometer as the maximum value in each plot. The topographical po­
eudicot trees except the palm Attalea butyracea, made up the final spe­ sition of each plot was scored as 1) foot, 2) lower slope, 3) upper slope,
cies list (see Table A.1). and 4) hilltop (see Table A.3). As the topography and geographical
location of the study area suggest the existence of climate gradients, data
2.2. Functional trait measurements and Community Weighted Mean for 19 bioclimatic variables (henceforth “climate” variables) were taken
(CWM) from WorldClim climate grids (www.worldclim.org) with a spatial res­
olution of approximately 1 km2, equivalent to 30 arc-seconds (Fick and
Eleven functional traits were selected due to their relation to Hijmans, 2017). Each sample plot was assigned values of climate vari­
important ecological processes in ecosystems as well as their responses ables based on its location within a WorldClim pixel. Some plots were in
to environmental variation, such as climate and soil, and tolerance of the same pixel. Geographical information was managed with ArcGIS
fires (see Table A.2). Due to the intraspecific variability of the species 10.3 (ESRI software). Concerning the site history and anthropogenic
and to the landscape heterogeneity in our study area (Derroire et al., variables (henceforth “anthropogenic” variables), we used 28 variables
2018), we sampled individuals in plots distributed widely across the characterizing the history of the forest previous to abandonment, for
area. Widely-used protocols were used for the measurement of most example, area of the farm, cattle pastures management (burning grass,
traits (Cornelissen et al., 2003; Pérez-Harguindeguy et al., 2013; Wil­ cutting grass), number of heads of cattle at the time of site abandon­
liamson and Wiemann, 2010). For leaf traits, whole-plant traits, and ment, grazing management (cattle rotation between pasture divisions),
bark thickness, five adult trees per species were sampled with a DBH and the period of time during which the regenerating forests were
between 10–30 cm and total height between 5–20 m (see appendix for exposed to fire (henceforth, “fire period”) (see Table A.4). Anthropo­
more detail). Leaf traits measured were: leaf area (LA, mm2), specific genic variables were obtained through semi-structured interviews with
leaf area (SLA, mm2 mg− 1), leaf dry matter content (LDMC, mg g− 1), leaf farmers and forest owners [see Granda Moser et al. (2015)].
nitrogen content (LNC, mg g− 1), leaf phosphorus content (LPC, mg g− 1),
leaf density (LD, cm-3), and leaf thickness (LT, mm). In the case of stem 2.4. Statistical analyses
traits, we measured the wood specific gravity (WSG, g cm-3), and bark
thickness (BT, mm). For wood specific gravity we sampled three adult We first used a Principal Components Analysis (PCA) to explore the
individuals per species with DBH 10–30 cm and total height 5− 20 m. patterns of variation of CWM traits values across the landscape. For this,
We evaluated two qualitative traits related to the regenerative ca­ we used the CWM of ten functional traits (LA, SLA, LDMC, LNC, LPC, LD,
pacity of the plant or fire tolerance: Resprouting capacity (Re) and Fire LT, WSG, BT, and To). We selected these traits because they permit
Tolerance (To). The resprouting capacity and fire tolerance were comprehensive functional characterization of these forest stands in
established by consultations with local technicians, forest owners, and relation to their likely responses to environmental variation, including
direct observations in the field for the 63 species, also, all the informa­ fire period, across this landscape (see Table A.2). LNC, LPC, and SLA are
tion was compared with available literature (Yanes et al., 2001). In this leaf economic spectrum (LES) traits whose CWMs likely reflect variation
study, resprouting capacity is understood as a response to disturbance, in soil fertility (see Baraloto et al., 2010; Wright et al., 2004). WSG is
implying a potential for repeated vegetative regeneration through part of the global fast-slow spectrum (Reich, 2014) along with the LES
different mechanisms like aerial buds, basal buds, and belowground traits, while CWM To and CWM BT are hypothesized to represent the
buds (Clarke et al., 2012). Herein, we evaluated Re as Resprouter = 1, response to fire, as WSG may also do. LD, LT, and LDMC are relatively
none-resprouter = 0. Also, we established a fire tolerance (To) criterion, little-studied traits. We selected leaf thickness for this analysis because it
for this, we defined tolerance as the species capacity to tolerate, resist is related to the physical strength of leaves and is likely to vary in
and survive fires. Therefore, species that tolerate and survive two or relation to insolation and substrate moisture (Pérez-Harguindeguy et al.,
more fires were scored as high tolerance (To = 3), those that tolerate at 2013), which may be key factors in our landscape. LDMC and LD may
least one fire were scored as medium tolerance (To = 2), while species show similar relationships to LT as well as being correlated with flam­
that lack fire tolerance were scored as low tolerance (To = 1). Finally, mability, another key factor in our landscape (Pérez-Harguindeguy
Community Weighted Mean traits values (CWM) were determined for et al., 2013).
each of the quantitative functional traits for each of the 52 plots, using Then, we determined the relationships between each of the eleven
the basal area as the weighting variable. For the qualitative traits functional traits (CWM) as response variables and forest age as the
resprouting capacity and fire tolerance, we calculated the percent basal predictor, using general and mixed linear models in InfoStat software
area per plot for each category of the trait. For the estimation of the (Di Rienzo et al., 2011). Because the data comes from three floristic

3
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

types of forests previously identified (Granda Moser et al., 2015), forest geographical coordinates that can be directly used in hypothesis testing
type was used as a random effect in these mixed models. (see Borcard and Legendre, 2002; Dray et al., 2006). Then, we used a
We evaluated the relationship of CWM To and CWM Re with CWM forward selection analysis to select individual predictor variables for
BT and CWM WSG, to determine potential tradeoffs between traits each matrix, using those with a significant contribution (p < 0.05)
related to tolerance to fires, also using general and mixed linear models. (Blanchet et al., 2008). The individual predictor variables for each one of
Also, regression models were performed between CWM BT and CWM four predictor matrices (soil, spatial variables, climatic, and anthropo­
WSG to evaluate the relationship between the two traits, the same genic variables) chosen by the forward selection process were used in
analysis was also carried out for species-level BT and WSG values. In the VARPART analysis, which along with the forward selection, we
both cases, the models that best fit the regression (linear and quadratic) performed using QEco software (Di Rienzo et al., 2010).
were evaluated. To test the fire tolerance hypothesis, we used categor­ For general and mixed linear models, the exponential spatial corre­
ical fire period data from the forest owner survey (see Table A.4). We lation function (corExp) was used to correct possible effects of spatial
defined the fire period as the time (years range) during which forests autocorrelation in all models, and the variance identity function (VarI­
have been exposed to fires. As fires only occur in the dry season, fires dent) to optimize normality and homogeneity of variance. Values of the
may only affect the forest once per year. As the fire period is a cate­ Akaike Information Criterion (AIC) and Bayesian Information Criterion
gorical variable, the data were grouped into two categories: short (BIC) were used to select the best model for each analysis. All regression
period, 0–10 years, and long period, 11–25 years. To compare values of models, t-tests, and principal components analysis were performed using
CWM BT, CWM WSG, and CWM Re between the two fire period cate­ InfoStat software (Di Rienzo et al., 2011).
gories, a t-test was used with a significance level of p < 0.05. Finally, the
relationship between CWM BT and CWM WSG and mean annual pre­ 3. Results
cipitation was determined as some authors suggest that these traits can
be related to precipitation or water availability (Rosell, 2016; Rosell and 3.1. Overview
Olson, 2014). For this, general and mixed linear models were again used,
using precipitation variables as a fixed factor and type of forest as a The Principal Component Analysis of multi-trait forest functional
random effect. characteristics explained 75.8 % of the variance of the CWM data for the
To test the relative contribution of the predictor variables to the ten traits with the first two components (Fig. 1, Table A.5). Leaf trait and
variation of the CWM traits values, we performed a variation parti­ stem trait spectra showed orthogonal axes in the PCA ordination, sug­
tioning analysis (VARPAR) (Borcard et al., 1992; Legendre, 2008). This gesting that stand CWM trait values respond to a trade-off between leaf
allowed partitioning the variation of CWMs into the contributions of a and stem traits representing independent strategies. PC1 (52.4 %) is
set of predictor variables grouped in the matrices of soil, climate, mainly a canopy economics spectrum. Forest stands trending towards
anthropogenic variables (including fire period), and spatial variables conservatism with high CWM LD and LDMC have negative eigenvalues
(PCNM eigenfunctions). For this analysis, we used the same CWM traits on this axis, while stands with acquisitive canopies as shown by high
as for the PCA (LA, SLA, LDMC, LNC, LPC, LD, LT, WSG, To, and BT). The CWM SLA, LNC, and LPC have positive eigenvalues. PC2 (23.4 %) is a
geographic coordinates of each plot were transformed to a geographic stem trait spectrum separating stands with high CWM BT (interestingly
distance matrix by principal coordinates of neighbor matrices analysis associated with high CWM LT) with positive eigenvalues, from stands
(PCNM) to be used as spatial variables and predictors of functional with high CWM WSG with negative eigenvalues. Also, interestingly,
properties (CWMs). This analysis creates spatial components from CWM BT and CWM WSG have a negative quadratic relationship to each

Fig. 1. Principal Component Analysis (PCA) of secondary forest sampled in 52 plots on the Nicoya Peninsula, Costa Rica, characterized by Community Weighted
Means of ten functional traits: leaf area = LA, specific leaf area = SLA, leaf dry matter content = LDMC, leaf N content = LNC, leaf P content = LPC, leaf density = LD,
leaf thickness = LT, wood specific gravity = WSG, bark thickness = BT, fire tolerance = To.

4
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

other (see below). Boukili and Chazdon, 2016; Lohbeck et al., 2015; Poorter et al., 2019).
One of the main contributions of these studies to date is the demon­
3.2. The seasonal forest hypothesis stration of fundamental differences of pattern and process between wet
and seasonally dry lowland forests, interpreted in terms of the age of
CWM LA, SLA, LDMC, LNC, LPC, LD, WSG, and To showed no sig­ forest stands and contrasting environmental drivers –gradients of light
nificant relationship with forest age (Table A.6). CWM bark thickness availability in wet forests, and water availability in seasonally dry for­
(CWM BT) and resprouting capacity (Re), were significantly related to ests (Lohbeck et al., 2015; Poorter et al., 2019). To our knowledge, ours
forest age, negatively in both cases (p = 0.0015, R2 = 0.09 and p = is one of the first studies to go beyond these biome-level models of
0.0094, R2 = 0.37, respectively; Fig. 2a, b). These responses of CWM successional change over time and determine the effect on secondary
stem traits, however, do not align with the seasonal forest hypothesis forests of the multiple drivers of spatial variation and temporal change
(see below). in a landscape, within which fire and spatial variables –potential sur­
rogates for dispersal limitation– are vitally important. Our PCA shows a
3.3. Fire tolerance hypothesis canopy economics spectrum as the main axis of variation in forest
functional properties, and an orthogonal secondary axis representing the
Tests for possible relationships and tradeoffs between fire tolerance stand-level tradeoff between CWM bark thickness and CWM WSG.
traits gave as a best model for CWM WSG and CWM BT (Fig. 2c, Neither CWM leaf trait values nor CWM WSG of forest stands were
Table A.7), a significant negative quadratic relationship between the related to forest age in our five to 35-year age range so that the seasonal
two CWM traits (p < 0.0001, R2 = 0.28). This suggests variations of forest hypothesis that we adapted from Lohbeck et al. (2015) and
CWM stem traits linked to trade-offs between BT and WSG of dominant Poorter et al. (2019) was not supported. CWM WSG was higher in forests
species. However, the relationship between these two traits at the spe­ exposed to fire for more than ten years than in forests with shorter fire
cies level is linear (Fig. 2d), BT declining as WSG increases (p < 0.0001, periods, and CWM BT lower. The increase of CWM WSG with fire period
R2 = 0.24, n = 63) so that factors other than the species-level trade-offs partially supports our fire tolerance hypothesis, but the corresponding
must be involved in shaping the CWMs. CWM WSG (positive, p < decline of CWM BT suggests that stand response to fire is mediated by
0.0001, R2 = 0.60, Fig. 2e), CWM BT (negative, p = 0.0044, R2 = 0.16, tradeoffs between these two CWM traits. Finally, using variation parti­
Fig. 2f) and CWM resprouting capacity (positive, p < 0.0001, R2 = 0.43, tioning we found that PCNM eigenfunctions had a strong relationship to
Fig. 2g) have significant relationships to CWM To. This result indicates forest functional characteristics represented by CWMs of ten traits,
that the species considered fire tolerant in the sources consulted tend to supporting our dispersal limitation hypothesis. Variation partitioning
have high WSG and resprouting capacity, while species with thick bark also highlighted gradients of soil properties and climate across the
are not considered fire tolerant. CWM BT and CWM WSG showed sig­ landscape as a contributor to variation of CWM traits. Fire and dispersal
nificant relations with precipitation (positive, p = 0.0103, R2 = 0.11, limitation are the main drivers of spatial variation in the functional
and negative, p = 0.003, R2 = 0.37, respectively, Fig. 2h). Finally, the t- properties of these seasonally dry forests, with a contribution from
test showed significant differences in CWM stem traits between the two climate and soil. We now discuss these results in more detail.
fire period categories. CWM WSG was higher in the long fire period
category, t = -2.48, p = 0.0166), and CWM BT lower in this category, t = 4.1. The seasonal forest hypothesis
2.80, p = 0.0074) (Fig. 3). CWM resprouting capacity did not differ
significantly between the two categories (t = 0.42, p = 0.6747). Given the dominance of the moist forests of our study area by typical
species of tropical dry forests, and the growing severity of the dry season
3.4. The dispersal limitation hypothesis there (Quesada-Hernández et al., 2019), our seasonal forest hypothesis
was based on the expected transition from conservative to more ac­
The Forward Selection procedure selected a total of 17 variables for quisitive forest functional with increasing forest age (Lohbeck et al.,
the four predictor matrices of the VARPART analysis: Three variables for 2013; Poorter et al., 2019). This hypothesis was not supported by our
the soil matrix (acidity, soil depth, Mn), four variables for the matrix of results. On the one hand, this can be attributed to the fact that our
anthropogenic variables (area of the farm (ha), number of cattle in the sampling design focused on capturing the effects of a range of predictor
year of pasture abandonment, principal activity on the farm –agriculture variables and lacked the bias inherent in the chronosequence approach
or livestock–, and fire period), two variables for the climate matrix (Johnson and Miyanishi, 2008). On the other hand, the 5–35 years age
(precipitation of the coldest quarter, isothermality) and eight PCNMs for range of our stands may not be sufficient for a strong test of the seasonal
the spatial variable matrix (PCNMs 9, 12, 18, 22, 23, 24, 37, 48). These forest hypothesis. It is also possible that the severity of the dry season
PCNMs represent meso- and fine-scale spatial relationships among plots. and continuing fires on the Nicoya Peninsula contribute to the persis­
Variation partitioning using these four matrices explained 46 % of tence of conservative CWM traits, delaying successional change. Finally,
the total variation of CWM functional trait values (adjusted R2 = 0.46) it is possible that edaphic factors or changes in the soil conditions along
(Table 1). All four predictor matrices were significant in the model. The the succession may be better predictors of the degree to which forest
largest contribution to the explained variation of the functional trait functional properties are conservative or acquisitive (Becknell and
values was from space, followed by soil, anthropogenic variables, and Powers, 2014; Estrada-Villegas et al., 2020; Werden et al., 2018). While
climate. However, when controlling for the effects of the other three some authors have referred to succession as an idiosyncratic process
matrices, the matrix of anthropogenic variables was not significant, and (Chazdon et al., 2007) the results of our hypothesis tests, especially as
space had the highest R2adj. This result suggests that space, soil, and regards fire, dispersal limitation, and soil, may explain some of the
climate are highly related to the functional properties of the secondary apparent idiosyncrasy.
forests of the Nicoya Peninsula. The relatively strong “pure effect” of
space, the matrix of PCNM eigenfunctions, is support for our dispersal 4.2. The fire tolerance hypothesis
limitation hypothesis.
Bark thickness plays an important role in fire tolerance (e.g. Brando
4. Discussion et al., 2012; Charles-Dominique et al., 2017; Hoffmann et al., 2003;
Pausas, 2015, 2017; Pausas et al., 2004; Poorter et al., 2014; Rosell,
The application of the principles and tools of functional ecology to 2016). At the species level, we found that BT declines linearly with
build an understanding of patterns and processes in tropical secondary increasing WSG, as did Poorter et al. (2014) in species of a Bolivian dry
forests is a relatively recent phenomenon (Becknell and Powers, 2014; forest. Despite this clear species-level tradeoff, CWM WSG and CWM BT

5
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

Fig. 2. Results of regression analysis of relationships between CWM trait values and forest age, and each other: a) CWM bark thickness (BT) and forest age; b) CWM
resprouting capacity and forest age, c) CWM WSG and CWM bark thickness, d) WSG and bark thickness at the species level, e) CWM WSG and CWM fire tolerance, f)
CWM bark thickness and CWM fire tolerance, g) CWM resprouting capacity and CWM fire tolerance, h) CWM WSG-CWM bark thickness with precipitation.

6
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

Fig. 3. Comparison of means between the fire period categories of a) CWM bark thickness (BT), b) CWM wood specific gravity (WSG) in 52 plots of Nicoya Peninsula,
Costa Rica.

other hand, the decrease in CWM bark thickness with forest age may be
Table 1 explained by the replacement of species with a higher survival rate
Results of variation partitioning. Table shows F-values, R2 values and p-values
under recurrent fires, for example, species with high wood density.
[Pr (>F)] for four matrices of predictor variables of CWM trait values of sec­
Although, the tradeoff between CWM BT and CWM WSG seems to be
ondary forest in 52 plots on the Nicoya Peninsula, Costa Rica. The general effect
represents the overall contribution of each predictor matrix and the pure effect, mediated by mean annual precipitation (MAP) as CWM BT tends to in­
the individual contribution of each matrix after removing the effects of the crease with MAP and CWM WSG to decrease. Accordingly, functional
others. See text for more details. traits of secondary forests related to fire tolerance are likely to respond
differently to environmental conditions such as recurrent fires and
Predictor matrix d. R2 R2 F Pr
(variables) f. adjusted adjusted (>F) precipitation in different stages the succession. Finally, despite a
(%) consensus on the function of the bark thickness as a fire protection trait
General Soil 3 0.210 21.00 5.58 0.001 (e.g. Brando et al., 2012; Charles-Dominique et al., 2017; Paine et al.,
effect 2010; Poorter et al., 2014; Rosell, 2016; Rosell and Olson, 2014), there
Anthropogenic 4 0.140 14.00 3.00 0.005 are other possible explanations of the variability of this trait in tropical
Climate 2 0.080 8.00 3.14 0.008 forests, such as mechanical support of the stem, photosynthesis, drought
Space 8 0.300 30.00 3.70 0.010
tolerance, phylogeny and even a relationship with soil fertility (Jager
Models All 17 0.460 46.00 3.51 0.001
Pure Soil 3 0.070 7.00 2.54 0.006 et al., 2015; Paine et al., 2010; Poorter et al., 2014; Rosell, 2016, 2019;
effect Rosell et al., 2014). These functions reflect complex interactions be­
Anthropogenic 4 0.010 1.00 1.15 0.295 tween different environmental gradients and disturbances that filter
Climate 2 0.040 4.00 2.24 0.026
stem traits and the fact that fire is a selective factor that significantly
Space 8 0.140 14.00 2.38 0.003
Residual 0.540 54.00
influences secondary forest functional properties (Brando et al., 2012;
Charles-Dominique et al., 2017; Chazdon, 2003; Pausas and Verdú,
2008; Vieira and Scariot, 2006). A large-scale study, which incorporates
have a negative quadratic relationship with each other in our landscape. all these environmental interactions including disturbances to which
Also, PCA shows that the CWM stem trait spectrum, a tradeoff between secondary forests are subjected, would clarify the variation of traits and
CWM BT and CWM WSG, is the secondary axis of multivariate trait more accurately predict the response mechanisms of forests, contrib­
variation in these forests. If BT > 1 cm is taken as a threshold at which uting better strategies within global efforts for restoring degraded areas
bark protects cambium from fire damage (Poorter et al., 2014), then in landscapes subject to fires.
CWM BT >1 cm could be taken indicating relatively fire-resistant forest
stands. By this criterion, approximately 50 % of the forest stands we 4.3. The dispersal limitation hypothesis
sampled have low fire resistance (CWM BT < 1 cm). However, CWM BT
declined with forest age but also, it was significantly lower in stands It is vital to break down the contributions of the different processes
exposed to fire for >10 years than in stands with shorter fire periods. The that can affect the composition of biodiversity (Legendre et al., 2009). In
significantly higher CWM WSG in stands in the long fire period category this way, the true, non-spatialized contributions of environmental fac­
and its positive relationship to CWM fire tolerance suggest that CWM tors to the variation of secondary forest functional properties can be
WSG, together with resprouting capacity, are stronger predictors of fire evaluated. At the same time, and perhaps more importantly, the
tolerance than CWM BT in this landscape. Our results suggest that contribution of space, potentially neutral processes, and especially
exposure to fire for more than ten years can be a driver of significant dispersal limitation, can be evaluated. The overall results of our varia­
changes in functional properties of dominant species in seasonally dry tion partitioning show that each of the four predictor matrices is related
tropical secondary forests. to variation in the functional properties of the forests in the study
Our results suggest a range of functional response strategies in the landscape. However, climate and anthropogenic matrices make negli­
face of environmental gradients and anthropogenic disturbances in the gible individual contributions to forest CWM functional traits. Our re­
Nicoya Peninsula. However, some of them, like bark thickness and sults suggest a greater contribution of space to the explanation of the
resprouting capacity, can be displaced with the forest age. It is possible functional properties of the secondary forests. This result may come
that resprouting capacity favored rapid forest regeneration after aban­ about because of spatial patterns in the availability of seed sources for
donment, but that non-resprouting species displace resprouters secondary regeneration, at different scales, and it supports our dispersal
throughout succession (Brando et al., 2012; Clarke et al., 2012). On the limitation hypothesis.

7
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

The operation of dispersal limitation has several potential mecha­ variables, soil, and climate on CWM trait values (Legendre et al., 2009).
nisms, which seem likely to operate together. A non-random distribution The detection of the effects on CWM traits of those three sets of variables
of seed sources is the first assumption of this discussion. Dispersal lim­ is of great interest (Becknell and Powers, 2014). Santiago-García et al.
itation, under which the location of an individual is partially determined (2019) found that a matrix of soil variables was more important than
by the location of its parent, the seed source, is a pervasive phenomenon stand age, a landscape configuration matrix, or space in the determi­
(Rosindell et al., 2011) especially to be expected in the regeneration of nation of spatial variation in secondary forest aboveground biomass in
forests on extensive pasture landscapes like the Nicoya Peninsula northeastern Costa Rica. Only one study that we are aware of has applied
(Granda Moser et al., 2015; van Breugel et al., 2019). The second this approach to the analysis of variation in secondary forest taxonomic
assumption of our discussion is that dispersal limitation occurs because composition (van Breugel et al., 2019), finding that soils and dispersal
whatever the agent of dispersal, most seeds probably fall at distances of limitation were the main drivers. The use of chronosequences to esti­
only tens of meters from fruiting trees, contributing to the creation of mate successional rates of change of composition and alpha diversity has
spatial patterns in seedling recruitment that are strongly dependent on contributed to this relative lack of knowledge of the overall suite of
the locations of parent trees (Nathan and Muller-Landau, 2000). These factors determining secondary forest ecosystem properties. Some au­
authors point out that the typical leptokurtic distributions of densities of thors have found relationships between functional properties and age in
dispersed seeds have extended tails representing rare events of secondary forests and suggest that changes in the functional properties
long-distance dispersal that must occur but are extremely difficult to of secondary forests can be predictable along succession (Boukili and
study. Spatial pattern in forest functional traits values caused by Chazdon, 2016; Lohbeck et al., 2015, 2013; Poorter et al., 2019).
dispersal limitation likely tends to be fine-grained, consistent with the However, it is necessary to understand that functional forest properties
PCNM eigenfunctions significantly related to CWM traits variation in can be influenced by many factors that work as environmental filters (e.
our study, which represent meso- and fine scale spatial relationships g. Derroire et al., 2018; Estrada-Villegas et al., 2020; Werden et al.,
among plots (Legendre et al., 2009). 2018). The availability of nutrients in the soil is an important factor
The third assumption behind our discussion is that because func­ influencing the formation of new forests in abandoned pastures (Finegan
tional trait values vary widely between species, dispersal limitation in­ and Delgado, 2000; Herrera and Finegan, 1997; Powers et al., 2009; van
fluences the CWM trait values of the secondary forests. The sources of Breugel et al., 2019). Our study area was affected by different levels of
seed for the colonization of abandoned pastures have several charac­ disturbance through different land use and continues to be affected by
teristics that likely contribute to dispersal limitation. First, some of the fires (Granda Moser et al., 2015). Therefore, understanding this broad
tree species that form 80 % of the basal area in these secondary forests set of environment variables, including the anthropogenic variables like
were probably present in the active pastures because they regenerate fire, is important to understand the functional properties of these new
naturally and are perceived as valuable by cattle farmers –Guazuma forests.
ulmifolia (Janzen, 1988) and Cordia alliodora (Finegan and Delgado,
2000) are examples. These trees become components of the regenerating 5. Conclusions
forests, disperse seeds into the forest and serve as nuclei facilitating the
dispersal of other species by flying vertebrates, creating an additional Our study shows different combinations of ecological strategies for
layer of spatial pattern (Finegan and Delgado, 2000; Galindo-González the passively restored forests of the Nicoya Peninsula. Our analysis
et al., 2000). based on functional traits, incorporating multiple explanatory variables
Second, most of the commonest tree species in our plots are dispersed and not only forest age, allowed us to identify factors affecting variation
by wind (Granda Moser et al., 2015), a dispersal mechanism well-known of traits in secondary forests. Also, spatial variables, a proxy for dispersal
to generate spatially heterogeneous seed shadows (Janzen, 1988; limitation, and soil variables, a proxy for niche partitioning, are the
Nathan and Muller-Landau, 2000) that probably contribute to the main predictors of variation in secondary forest CWM functional traits,
importance of space as a predictor in our study. while fire plays a key role. Likewise, our results suggest that forests are
Third, the dispersal of seeds by cattle in seasonally dry forests developing tradeoffs at the level of stem traits and regenerative traits
(Janzen, 1988), and from trees of species from the original forest that allow them to adapt to the incidence of fires and that ten years of
conserved by farmers during conversion to pastures, or planted by fire exposure may affect adaptation strategies of secondary forests
farmers, are additional factors likely to contribute to dispersal limita­ concerning stem traits. This research broadens the knowledge about the
tion. In the seasonally dry neotropics, farmers graze cattle non-randomly functional dynamics of restoration for the study site and provides
on their land (Godinot et al., 2020) and in one study, the density and guidelines that can be useful for decision-makers in aspects of ecosystem
diversity of regeneration of woody species in pastures was explained by management and conservation.
the density and richness of adult trees in pastures and by distance to
remnant forest patches (Esquivel et al., 2008). For vertebrate-dispersed Funding
species, birds and bats generate strongly non-random seed rains around
isolated trees in pastures (Galindo-González et al., 2000). This work was supported by the German Academic Exchange Service
Finally, the non-native species Gmelina arborea and Tectona grandis (DAAD) and Chair of Ecology in the Management of Tropical Forests,
have regenerated naturally within our secondary forests (Granda Moser CATIE.
et al., 2015). These two species are planted typically in small areas on
private farms with funding from Costa Rica’s environmental services Author contributions
payment program (Pago por Servicios Ambientales, https://onfcr.org/ps
a-2/). Besides the small areas, the distribution of the areas planted Francisco S. Álvarez and Bryan Finegan: Conceptualization, Meth­
across the Nicoya Peninsula is strongly heterogeneous (Meza Picado odology, Investigation, Data Curation, Formal analysis, Writing-Original
et al., 2019). The operation of dispersal limitation on the colonization of draft preparation, Writing-Review and Editing, Project administration,
abandoned pastures by these species is probably another contributor to Funding acquisition. Diego Delgado and Zayra Ramos: Supervision,
the spatial effect on the CWM functional traits of the secondary forests. Writing-Original draft preparation, Writing-Review, and Editing. Luis
Pedro Utrera and Vanessa Granda: Investigation, Methodology, Data
4.4. The effect of soil and climate Curation, Writing-Review, and Editing.

Our test of the dispersal limitation hypothesis using variation parti­

tioning necessarily involved the removal of the effects of anthropogenic

8
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

Declaration of Competing Interest forests following major disturbances. Philos. Trans. R. Soc. Lond., B, Biol. Sci. 362,
273–289. https://doi.org/10.1098/rstb.2006.1990.
Clarke, P.J., Lawes, M., Midgley, J., Lamont, B., Ojeda, F., Burrows, G., Enright, N.,
The authors report no declarations of interest. Knox, K., 2012. Resprouting as a key functional trait: how buds, protection and
resources drive persistence after fire. New Phytol. 197, 19–35. https://doi.org/
Acknowledgments 10.1111/nph.12001.
Cornelissen, J., Lavorel, S., Garnier, E., Diaz, S., Buchmann, N., Gurvich, D., Reich, P.,
Ter Steege, H., Morgan, H., Van Der Heijden, M., 2003. A handbook of protocols for
We are grateful for the financial support of the German Academic standardised and easy measurement of plant functional traits worldwide. Aust. J.
Exchange Service (DAAD) for the development of this research. Also, we Bot. 51, 335–380. https://doi.org/10.1071/BT02124.
Craven, D., Hall, J.S., Berlyn, G.P., Ashton, M.S., van Breugel, M., 2018. Environmental
thank SINAC for the collection and scientific research permits in the filtering limits functional diversity during succession in a seasonally wet tropical
Área de Conservación de Tempisque (ACT). We thank the Universidad secondary forest. J. Veg. Sci. 29, 511–520. https://doi.org/10.1111/jvs.12632.
Estatal a Distancia (UNED; Centro Universitario Cañas) of Costa Rica for Derroire, G., Powers, J.S., Hulshof, C.M., Varela, L.E.C., Healey, J.R., 2018. Contrasting
patterns of leaf trait variation among and within species during tropical dry forest
support in the laboratory phase. We are grateful to the Centro Agrícola succession in Costa Rica. Sci. Rep. 8, 1–11. https://doi.org/10.1038/s41598-017-
Cantonal de Hojancha (CACH), mainly the support of Ademar Molina for 18525-1.
the accompaniment throughout the field phase. We are grateful to Di Rienzo, J., Pla, L., Di Rienzo, M., Vílchez, S., Casanoves, F., 2010. Qeco-Quantitative
ecology software: a collaborative approach. Revista Latinoamericana de
Xochilt Pocasangre-Orellana and Andrea Paiz for the accompaniment conservación 1, 73–75.
throughout the field phase. Finally, we thank the Chair of Ecology in the Di Rienzo, J., Casanoves, F., Balzarini, M., Gonzalez, L., Tablada, M., Robledo, C., 2011.
Management of Tropical Forests (CATIE) for their support. InfoStat versión 2011 (11 May 2015). http://www.infostat.com.ar.
Díaz, S., Hodgson, J., Thompson, K., Cabido, M., Cornelissen, J., Jalili, A., Montserrat-
Marti, G., Grime, J., Zarrinkamar, F., Asri, Y., 2004. The plant traits that drive
Appendix A. Supplementary data ecosystems: evidence from three continents. J. Veg. Sci. 15, 295–304. https://doi.
org/10.1111/j.1654-1103.2004.tb02266.x.
Supplementary material related to this article can be found, in the Dray, S., Legendre, P., Peres-Neto, P.R., 2006. Spatial modelling: a comprehensive
framework for principal coordinate analysis of neighbour matrices (PCNM). Ecol.
online version, at doi:https://doi.org/10.1016/j.ppees.2021.125632. Modell. 196, 483–493. https://doi.org/10.1016/j.ecolmodel.2006.02.015.
DRYFLOR, Banda-R, K., Delgado-Salinas, A., Dexter, K.G., Linares-Palomino, R., Oliveira-
References Filho, A., Prado, D., Pullan, M., Quintana, C., Riina, R., Rodríguez M, G.M.,
Weintritt, J., Acevedo-Rodríguez, P., Adarve, J., Álvarez, E., Aranguren B, A.,
Arteaga, J.C., Aymard, G., Castaño, A., Ceballos-Mago, N., Cogollo, Á., Cuadros, H.,
Arroyo-Mora, J.P., Sánchez-Azofeifa, G.A., Rivard, B., Calvo, J.C., Janzen, D.H., 2005. Delgado, F., Devia, W., Dueñas, H., Fajardo, L., Fernández, Á., Fernández, M.Á.,
Dynamics in landscape structure and composition for the Chorotega region, Costa Franklin, J., Freid, E.H., Galetti, L.A., Gonto, R., González-M, R., Graveson, R.,
rica from 1960 to 2000. Agric. Ecosyst. Environ. 106, 27–39. https://doi.org/ Helmer, E.H., Idárraga, Á., López, R., Marcano-Vega, H., Martínez, O.G., Maturo, H.
10.1016/j.agee.2004.07.002. M., McDonald, M., McLaren, K., Melo, O., Mijares, F., Mogni, V., Molina, D.,
Baraloto, C., Timothy Paine, C., Poorter, L., Beauchene, J., Bonal, D., Domenach, A.M., Moreno, N.d.P., Nassar, J.M., Neves, D.M., Oakley, L.J., Oatham, M., Olvera-Luna, A.
Hérault, B., Patiño, S., Roggy, J.C., Chave, J., 2010. Decoupled leaf and stem R., Pezzini, F.F., Dominguez, O.J.R., Ríos, M.E., Rivera, O., Rodríguez, N., Rojas, A.,
economics in rain forest trees. Ecol. Lett. 13, 1338–1347. https://doi.org/10.1111/ Särkinen, T., Sánchez, R., Smith, M., Vargas, C., Villanueva, B., Pennington, R.T.,
j.1461-0248.2010.01517.x. 2016. Plant diversity patterns in neotropical dry forests and their conservation
Becknell, J.M., Powers, J.S., 2014. Stand age and soils as drivers of plant functional traits implications. Science 353, 1383–1387. https://doi.org/10.1126/science.aaf5080.
and aboveground biomass in secondary tropical dry forest. Can. J. For. Res. 44, Esquivel, M.J., Harvey, C.A., Finegan, B., Casanoves, F., Skarpe, C., 2008. Effects of
604–613. https://doi.org/10.1139/cjfr-2013-0331. pasture management on the natural regeneration of neotropical trees. J. Appl. Ecol.
Bertsch, F., Alvarado, A., Henriquez, C., Mata, R., Hall, C.A., 2000. Properties, 45, 371–380. https://doi.org/10.1111/j.1365-2664.2007.01411.x.
geographic, distribution, and management of major soil orders of Costa Rica. In: Estrada-Villegas, S., Bailón, M., Hall, J.S., Schnitzer, S.A., Turner, B.L., Caughlin, T., van
Hall, C.A.S., Leon, C., Leclerc, G. (Eds.), Quantifying Sustainable Development: The Breugel, M., 2020. Edaphic factors and initial conditions influence successional
Future of Tropical Economies. Academic Press, San Diego, California, pp. 265–294. trajectories of early regenerating tropical dry forests. J. Ecol. 108, 160–174. https://
Blanchet, F.G., Legendre, P., Borcard, D., 2008. Forward selection of explanatory doi.org/10.1111/1365-2745.13263.
variables. Ecology 89, 2623–2632. https://doi.org/10.1890/07-0986.1. Fick, S.E., Hijmans, R.J., 2017. WorldClim 2: new 1-km spatial resolution climate
Borcard, D., Legendre, P., 2002. All-scale spatial analysis of ecological data by means of surfaces for global land areas. Int. J. Climatol. 37, 4302–4315. https://doi.org/
principal coordinates of neighbour matrices. Ecol. Modell. 153, 51–68. https://doi. 10.1002/joc.5086.
org/10.1016/S0304-3800(01)00501-4. Finegan, B., 1996. Pattern and process in neotropical secondary rain forests: the first 100
Borcard, D., Legendre, P., Drapeau, P., 1992. Partialling out the spatial component of years of succession. Trends Ecol. Evol. 11, 119–124. https://doi.org/10.1016/0169-
ecological variation. Ecology 73, 1045–1055. https://doi.org/10.2307/1940179. 5347(96)81090-1.
Boukili, V.K., Chazdon, R.L., 2016. Environmental filtering, local site factors and Finegan, B., Delgado, D., 2000. Structural and floristic heterogeneity in a 30-year-old
landscape context drive changes in functional trait composition during tropical Costa Rican rain forest restored on pasture through natural secondary succession.
forest succession. Perspect. Plant Ecol. Evol. Syst. 24, 37–47. https://doi.org/ Restor. Ecol. 8, 380–393. https://doi.org/10.1046/j.1526-100x.2000.80053.x.
10.1016/j.ppees.2016.11.003. Finegan, B., Nasi, R., 2004. The biodiversity and conservation potential of shifting
Brando, P.M., Nepstad, D.C., Balch, J.K., Bolker, B., Christman, M.C., Coe, M., Putz, F.E., cultivation landscapes. In: Schroth, G., da Fonseca, G.A.B., Harvey, C.A., Gascon, C.,
2012. Fire-induced tree mortality in a neotropical forest: the roles of bark traits, tree Vasconcelos, H.L., Izac, A.M.N. (Eds.), Agroforestry and Biodiversity Conservation in
size, wood density and fire behavior. Glob. Chang. Biol. 18, 630–641. https://doi. Tropical Landscapes. Island Press, Washington, DC, pp. 153–197.
org/10.1111/j.1365-2486.2011.02533.x. Galindo-González, J., Guevara, S., Sosa, V.J., 2000. Bat-and bird-generated seed rains at
Buzzard, V., Hulshof, C.M., Birt, T., Violle, C., Enquist, B.J., 2015. Re-growing a tropical isolated trees in pastures in a tropical rainforest. Conserv. Biol. 14, 1693–1703.
dry forest: functional plant trait composition and community assembly during https://doi.org/10.1111/j.1523-1739.2000.99072.x.
succession. Funct. Ecol. 30, 1006–1013. https://doi.org/10.1111/1365-2435.12579. Garnier, E., Cortez, J., Billès, G., Navas, M.-L., Roumet, C., Debussche, M., Laurent, G.,
Casanoves, F., Pla, L., Di Rienzo, J.A., Díaz, S., 2011. FDiversity: a software package for Blanchard, A., Aubry, D., Bellmann, A., 2004. Plant functional markers capture
the integrated analysis of functional diversity. Methods Ecol. Evol. 2, 233–237. ecosystem properties during secondary succession. Ecology 85, 2630–2637. https://
https://doi.org/10.1111/j.2041-210X.2010.00082.x. doi.org/10.1890/03-0799.
Charles-Dominique, T., Midgley, G., Bond, W., 2017. Fire frequency filters species by Gei, M., Rozendaal, D.M., Poorter, L., Bongers, F., Sprent, J.I., Garner, M.D., Aide, T.M.,
bark traits in a savanna-forest mosaic. J. Veg. Sci. 28, 728–735. https://doi.org/ Andrade, J.L., Balvanera, P., Becknell, J.M., 2018. Legume abundance along
10.1111/jvs.12528. successional and rainfall gradients in Neotropical forests. Nat. Ecol. Evol. 2,
Chave, J., Coomes, D., Jansen, S., Lewis, S.L., Swenson, N.G., Zanne, A.E., 2009. Towards 1104–1111. https://doi.org/10.1038/s41559-018-0559-6.
a worldwide wood economics spectrum. Ecol. Lett. 12, 351–366. https://doi.org/ Godinot, F., Somarriba, E., Finegan, B., Delgado Rodríguez, D., 2020. Secondary tropical
10.1111/j.1461-0248.2009.01285.x. dry forests are important to cattle ranchers in Northwestern Costa Rica. Rev. Cienc.
Chazdon, R.L., 2003. Tropical forest recovery: legacies of human impact and natural Ambient. 54, 20–50. https://doi.org/10.15359/rca.54-2.2.
disturbances. Perspect. Plant Ecol. Evol. Syst. 6, 51–71. https://doi.org/10.1078/ González-M, R., Norden, N., Posada, J.M., Pizano, C., García, H., Idárraga-Piedrahita, Á.,
1433-8319-00042. López-Camacho, R., Nieto, J., Rodríguez-M, G.M., Torres, A.M., 2019. Climate
Chazdon, R.L., 2008. Beyond deforestation: restoring forests and ecosystem services on severity and land-cover transformation determine plant community attributes in
degraded lands. Science 320, 1458–1460. https://doi.org/10.1126/ Colombian dry forests. Biotropica 51, 826–837. https://doi.org/10.1111/btp.12715.
science.1155365. Granda Moser, V., Finegan, B., Ramos Bendaña, Z.S., Detlefsen, G., Molina, A., 2015.
Chazdon, R.L., 2014. Second Growth: the Promise of Tropical Forest Regeneration in an Potencial de manejo de bosques restaurados por sucesión natural secundaria en
Age of Deforestation. University of Chicago Press. Guanacaste, Costa Rica: Composición, diversidad y especies maderables. Centro
Chazdon, R.L., Letcher, S.G., van Breugel, M., Martinez-Ramos, M., Bongers, F., Agronómico Tropical de Investigación y Enseñanza (CATIE), Turrialba, Costa Rica.
Finegan, B., 2007. Rates of change in tree communities of secondary Neotropical Grime, J., 1998. Benefits of plant diversity to ecosystems: immediate, filter and founder
effects. J. Ecol. 86, 902–910. https://doi.org/10.1046/j.1365-2745.1998.00306.x.

9
F.S. Álvarez et al. Perspectives in Plant Ecology, Evolution and Systematics 52 (2021) 125632

Herrera, B., Finegan, B., 1997. Substrate conditions, foliar nutrients and the distributions Mora, F., de Souza Moreno, V., Müller, S.C., Muñoz, R., Muscarella, R., de Oliveira
of two canopy tree species in a Costa Rican secondary rain forest. Plant Soil 191, Neto, S.N., Nunes, Y.R.F., Ochoa-Gaona, S., Paz, H., Peña-Claros, M., Piotto, D.,
259–267. https://doi.org/10.1023/A:1004209915530. Ruíz, J., Sanaphre-Villanueva, L., Sanchez-Azofeifa, A., Schwartz, N.B.,
Hoffmann, W.A., Orthen, B., Nascimento, P.K.V.d., 2003. Comparative fire ecology of Steininger, M.K., Thomas, W.W., Toledo, M., Uriarte, M., Utrera, L.P., van
tropical savanna and forest trees. Funct. Ecol. 17, 720–726. https://doi.org/ Breugel, M., van der Sande, M.T., van der Wal, H., Veloso, M.D.M., Vester, H.F.M.,
10.1111/j.1365-2435.2003.00796.x. Vieira, I.C.G., Villa, P.M., Williamson, G.B., Wright, S.J., Zanini, K.J., Zimmerman, J.
Holdridge, L.R., 1967. Life Zone Ecology. Tropical Science Center, San José, Costa Rica. K., Westoby, M., 2019. Wet and dry tropical forests show opposite successional
Jackson, J.F., Adams, D.C., Jackson, U.B., 1999. Allometry of constitutive defense: a pathways in wood density but converge over time. Nat. Ecol. Evol. 3, 928–934.
model and a comparative test with tree bark and fire regime. Am. Nat. 153, 614–632. https://doi.org/10.1038/s41559-019-0882-6.
Jager, M.M., Richardson, S.J., Bellingham, P.J., Clearwater, M.J., Laughlin, D.C., 2015. Powers, J.S., Becknell, J.M., Irving, J., Pèrez-Aviles, D., 2009. Diversity and structure of
Soil fertility induces coordinated responses of multiple independent functional traits. regenerating tropical dry forests in Costa Rica: geographic patterns and
J. Ecol. 103, 374–385. https://doi.org/10.1111/1365-2745.12366. environmental drivers. For. Ecol. Manage. 258, 959–970. https://doi.org/10.1016/j.
Janzen, D.H., 1988. Management of habitat fragments in a tropical dry forest: growth. foreco.2008.10.036.
Ann. Mo. Bot. Gard. 75, 105–116. Quesada-Hernández, L.E., Calvo-Solano, O.D., Hidalgo, H.G., Pérez-Briceño, P.M.,
Janzen, D.H., Hallwachs, W., 2020. Área de Conservación Guanacaste, northwestern Alfaro, E.J., 2019. Dynamical delimitation of the Central American Dry Corridor
Costa Rica: converting a tropical national park to conservation via biodevelopment. (CADC) using drought indices and aridity values. Prog. Phys. Geogr. 43, 627–642.
Biotropica 52, 1017–1029. https://doi.org/10.1111/btp.12755. https://doi.org/10.1177/0309133319860224.
Johnson, E.A., Miyanishi, K., 2008. Testing the assumptions of chronosequences in Reich, P.B., 2014. The world-wide ‘fast–slow’plant economics spectrum: a traits
succession. Ecol. Lett. 11, 419–431. https://doi.org/10.1111/j.1461- manifesto. J. Ecol. 102, 275–301. https://doi.org/10.1111/1365-2745.12211.
0248.2008.01173.x. Rosell, J.A., 2016. Bark thickness across the angiosperms: more than just fire. New
Legendre, P., 2008. Studying beta diversity: ecological variation partitioning by multiple Phytol. 211, 90–102. https://doi.org/10.1111/nph.13889.
regression and canonical analysis. J. Plant Ecol. 1, 3–8. https://doi.org/10.1093/ Rosell, J.A., 2019. Bark in woody plants: understanding the diversity of a multifunctional
jpe/rtm001. structure. Integr. Comp. Biol. 59, 535–547. https://doi.org/10.1093/icb/icz057.
Legendre, P., Mi, X., Ren, H., Ma, K., Yu, M., Sun, I.-F., He, F., 2009. Partitioning beta Rosell, J.A., Olson, M.E., 2014. The evolution of bark mechanics and storage across
diversity in a subtropical broad-leaved forest of China. Ecology 90, 663–674. habitats in a clade of tropical trees. Am. J. Bot. 101, 764–777. https://doi.org/
https://doi.org/10.1890/07-1880.1. 10.3732/ajb.1400109.
Lohbeck, M., Poorter, L., Lebrija-Trejos, E., Martínez-Ramos, M., Meave, J.A., Paz, H., Rosell, J.A., Gleason, S., Méndez-Alonzo, R., Chang, Y., Westoby, M., 2014. Bark
Pérez-García, E.A., Romero-Pérez, I.E., Tauro, A., Bongers, F., 2013. Successional functional ecology: evidence for tradeoffs, functional coordination, and environment
changes in functional composition contrast for dry and wet tropical forest. Ecology producing bark diversity. New Phytol. 201, 486–497. https://doi.org/10.1111/
94, 1211–1216. https://doi.org/10.1890/12-1850.1. nph.12541.
Lohbeck, M., Lebrija-Trejos, E., Martínez-Ramos, M., Meave, J.A., Poorter, L., Rosindell, J., Hubbell, S.P., Etienne, R.S., 2011. The unified neutral theory of biodiversity
Bongers, F., 2015. Functional trait strategies of trees in dry and wet tropical forests and biogeography at age ten. Trends Ecol. Evol. 26, 340–348. https://doi.org/
are similar but differ in their consequences for succession. PLoS One 10, e0123741. 10.1016/j.tree.2011.03.024.
https://doi.org/10.1371/journal.pone.0123741. Rozendaal, D.M., Bongers, F., Aide, T.M., Alvarez-Dávila, E., Ascarrunz, N.,
McGill, B.J., Enquist, B.J., Weiher, E., Westoby, M., 2006. Rebuilding community Balvanera, P., Becknell, J.M., Bentos, T.V., Brancalion, P.H., Cabral, G.A., 2019.
ecology from functional traits. Trends Ecol. Evol. 21, 178–185. https://doi.org/ Biodiversity recovery of Neotropical secondary forests. Sci. Adv. 5, eaau3114.
10.1016/j.tree.2006.02.002. https://doi.org/10.1126/sciadv.aau3114.
Meza Picado, V., Alfaro Jiménez, K., Bedoya Arrieta, R., Romero Mora, M., Valerio Sanaphre-Villanueva, L., Dupuy, J.M., Andrade, J.L., Reyes-García, C., Jackson, P.C.,
Madrigal, A., Montenegro Salas, P., 2019. Reforestación comercial en Costa Rica: Paz, H., 2017. Patterns of plant functional variation and specialization along
Regiones Huetar Atlántica, Huetar Norte y Chorotega, 1 ed. UNA/ INISEFOR, secondary succession and topography in a tropical dry forest. Environ. Res. Lett. 12,
Heredia, Costa Rica. 055004 https://doi.org/10.1088/1748-9326/aa6baa.
Nathan, R., Muller-Landau, H.C., 2000. Spatial patterns of seed dispersal, their Sanchez, P.A., 1981. Suelos del trópico: características y manejo. IICA Biblioteca
determinants and consequences for recruitment. Trends Ecol. Evol. 15, 278–285. Venezuela.
https://doi.org/10.1016/S0169-5347(00)01874-7. Santiago-García, R.J., Finegan, B., Bosque-Pérez, N.A., 2019. Soil is the main predictor of
Paine, C.E.T., Stahl, C., Courtois, E.A., Patiño, S., Sarmiento, C., Baraloto, C., 2010. secondary rain forest estimated aboveground biomass across a Neotropical
Functional explanations for variation in bark thickness in tropical rain forest trees. landscape. Biotropica 51, 10–17. https://doi.org/10.1111/btp.12621.
Funct. Ecol. 24, 1202–1210. https://doi.org/10.1111/j.1365-2435.2010.01736.x. Sesnie, S.E., Finegan, B., Gessler, P.E., Ramos, Z., 2009. Landscape-scale environmental
Pausas, J.G., 2015. Bark thickness and fire regime. Funct. Ecol. 29, 315–327. https://doi. and floristic variation in Costa Rican old-growth rain forest remnants. Biotropica 41,
org/10.1111/1365-2435.12372. 16–26. https://doi.org/10.1111/j.1744-7429.2008.00451.x.
Pausas, J.G., 2017. Bark thickness and fire regime: another twist. New Phytol. 213, Smith, J., Sabogal, C., De Jong, W., Kaimowitz, D., 1997. Bosques secundarios como
13–15. https://doi.org/10.1111/nph.14277. recurso para el desarrollo rural y la conservacion ambiental en los tropicos de
Pausas, J.G., Verdú, M., 2008. Fire reduces morphospace occupation in plant America Latina. CIFOR, Bogor, Indonesia.
communities. Ecology 89, 2181–2186. https://doi.org/10.1890/07-1737.1. Stan, K., Sanchez-Azofeifa, A., 2019. Deforestation and secondary growth in Costa Rica
Pausas, J.G., Bradstock, R.A., Keith, D.A., Keeley, J.E., 2004. Plant functional traits in along the path of development. Reg. Environ. Change 19, 587–597. https://doi.org/
relation to fire in crown-fire ecosystems. Ecology 85, 1085–1100. https://doi.org/ 10.1007/s10113-018-1432-5.
10.1890/02-4094. Steffen, W., Sanderson, R.A., Tyson, P.D., Jäger, J., Matson, P.A., Moore III, B.,
Pérez-Harguindeguy, N., Díaz, S., Garnier, E., Lavorel, S., Poorter, H., Jaureguiberry, P., Oldfield, F., Richardson, K., Schellnhuber, H.J., Turner, B.L., 2004. Global Change
Bret-Harte, M.S., Cornwell, W.K., Craine, J.M., Gurvich, D.E., Urcelay, C., and the Earth System: a Planet under Pressure. Springer, Berlin.
Veneklaas, E.J., Reich, P.B., Poorter, L., Wright, I.J., Ray, P., Enrico, L., Pausas, J.G., van Breugel, M., Craven, D., Lai, H.R., Baillon, M., Turner, B.L., Hall, J.S., 2019. Soil
de Vos, A.C., Buchmann, N., Funes, G., Quétier, F., Hodgson, J.G., Thompson, K., nutrients and dispersal limitation shape compositional variation in secondary
Morgan, H.D., ter Steege, H., van der Heijden, M.G.A., Sack, L., Blonder, B., tropical forests across multiple scales. J. Ecol. 107, 566–581. https://doi.org/
Poschlod, P., Vaieretti, M.V., Conti, G., Staver, A.C., Aquino, S., Cornelissen, J.H.C., 10.1111/1365-2745.13126.
2013. New handbook for standardised measurement of plant functional traits Vieira, D.L., Scariot, A., 2006. Principles of natural regeneration of tropical dry forests
worldwide. Aust. J. Bot. 61, 167–234. https://doi.org/10.1071/BT12225. for restoration. Restor. Ecol. 14, 11–20. https://doi.org/10.1111/j.1526-
Poorter, L., McNeil, A., Hurtado, V.H., Prins, H.H., Putz, F.E., 2014. Bark traits and life- 100X.2006.00100.x.
history strategies of tropical dry-and moist forest trees. Funct. Ecol. 28, 232–242. Werden, L.K., Becknell, J.M., Powers, J.S., 2018. Edaphic factors, successional status and
https://doi.org/10.1111/1365-2435.12158. functional traits drive habitat associations of trees in naturally regenerating tropical
Poorter, L., Rozendaal, D.M.A., Bongers, F., de Almeida-Cortez, J.S., Almeyda dry forests. Funct. Ecol. 32, 2766–2776. https://doi.org/10.1111/1365-2435.13206.
Zambrano, A.M., Álvarez, F.S., Andrade, J.L., Villa, L.F.A., Balvanera, P., Becknell, J. Williamson, G.B., Wiemann, M.C., 2010. Measuring wood specific gravity… correctly.
M., Bentos, T.V., Bhaskar, R., Boukili, V., Brancalion, P.H.S., Broadbent, E.N., Am. J. Bot. 97, 519–524. https://doi.org/10.3732/ajb.0900243.
César, R.G., Chave, J., Chazdon, R.L., Colletta, G.D., Craven, D., de Jong, B.H.J., Wright, I.J., Reich, P.B., Westoby, M., Ackerly, D.D., Baruch, Z., Bongers, F., Cavender-
Denslow, J.S., Dent, D.H., DeWalt, S.J., García, E.D., Dupuy, J.M., Durán, S.M., Bares, J., Chapin, T., Cornelissen, J.H., Diemer, M., 2004. The worldwide leaf
Espírito Santo, M.M., Fandiño, M.C., Fernandes, G.W., Finegan, B., Moser, V.G., economics spectrum. Nature 428, 821–827. https://doi.org/10.1038/nature02403.
Hall, J.S., Hernández-Stefanoni, J.L., Jakovac, C.C., Junqueira, A.B., Kennard, D., Yanes, C.V., Muñoz, A.I.B., Alcocer, M.I., Silvia, M.G.Dy.C., Dirzo, S., 2001. Árboles y
Lebrija-Trejos, E., Letcher, S.G., Lohbeck, M., Lopez, O.R., Marín-Spiotta, E., arbustos nativos potencialmente valiosos para la restauración ecológica y la
Martínez-Ramos, M., Martins, S.V., Massoca, P.E.S., Meave, J.A., Mesquita, R., reforestación. Instituto de Ecología, UNAM, México.

10