Ecological Drivers of Standing Volume and Carbon Stocks in Contrasting Tropical Forests of Mexico and Colombia

Abstract

Tropical forests differ widely in floristic composition, stand structure, standing volume, and carbon storage, yet comparative evidence across contrasting tropical forest types remains limited. This study examined whether variation in standing volume and carbon stocks among contrasting tropical forests was more closely associated with structural attributes or with diversity-related patterns. Two tropical wet forests in Colombia and one tropical semi-deciduous forest in Mexico were evaluated using 40 circular plots of 500 m² established within a 100 ha reference area in each forest, where all trees with DBH > 10 cm were measured. Floristic composition, ecological dominance, diversity, dendrometric attributes, standing volume, biomass, and carbon stocks were estimated using a common analytical framework. The two wet forests showed higher effective diversity, broader taxonomic dominance, greater basal area, mean height, standing volume, biomass, and carbon stocks than the tropical semi-deciduous forest. In contrast, the semi-deciduous forest showed stronger dominance concentrated in fewer taxa, especially Euphorbiaceae, a pattern that may reflect the ecological suitability of this family under more seasonal and water-limited conditions. At the family level, standing volume, biomass, and carbon were distributed more evenly among dominant families in the wet forests, whereas they were more concentrated in fewer lineages in the semi-deciduous forest. Basal area showed the strongest association with standing volume, total biomass, and total carbon, followed by mean height and mean DBH. Overall, the results indicate that, under the conditions evaluated, structural organization was more closely associated with standing volume and carbon storage than diversity alone, while diversity acted as a complementary correlate.

1. Introduction

Tropical forests play a central role in the global carbon cycle because they concentrate large carbon pools, sustain high biological diversity, and regulate ecological processes closely linked to climate and hydrology; yet, these ecosystems are increasingly exposed to warming, shifting precipitation regimes, and more frequent extreme events, all of which can alter tree growth, mortality, biomass accumulation, and long-term carbon residence time; under this context, understanding how ecological organization influences forest standing volume and carbon storage has become essential for improving both ecological interpretation and climate-change mitigation strategies in tropical landscapes [1,2].

Current evidence suggests that the relationships among diversity, standing volume, and carbon storage in tropical forests are not linear and should not be interpreted through simple climatic gradients alone; across the tropical biome, carbon stocks are shaped not only by species richness, but also by dominance patterns, stand density, tree-size inequality, and structural complexity, while forest standing volume is strongly associated with stand structure and the way vertical and horizontal organization regulates resource capture and biomass accumulation. This indicates that the main ecological controls of forest functioning emerge from the interaction between composition and structure, rather than from any single attribute considered in isolation [1,3].

This perspective becomes especially relevant when tropical forests differ markedly in water availability and seasonality. Recent studies across the Neotropics have shown that forest structure, biomass accumulation, and carbon storage respond not only to climatic conditions, but also to variation in stand attributes such as basal area, tree-size distribution, and taxonomic dominance. In humid tropical forests, biomass and carbon dynamics are strongly influenced by structural disturbance, degradation, and canopy development, whereas in more seasonal tropical forests the distribution of tree sizes, local dominance patterns, and allometric variation can strongly affect standing volume and carbon estimates. Together, these studies indicate that the ecological interpretation of carbon storage and woody accumulation requires an integrated view of both forest composition and stand structure across contrasting tropical environments [4,5,6].

Although these studies have advanced the understanding of tropical forest carbon and standing volume, most have focused on single forest types or regional contexts, which limits broader ecological interpretation across contrasting tropical systems. In this sense, comparing the Tropical Wet Forests (TWF) of Colombia with the Tropical Semi-Deciduous Forest (TSDF) of Mexico provides a useful framework for examining whether differences in standing volume and carbon stocks are primarily associated with forest type itself or with the internal ecological organization of each stand. Here we propose that forests under stronger climatic seasonality may partially offset environmental constraints when stand structure is organized efficiently, particularly through structural configurations that favor space occupation, tree-size distribution, standing volume accumulation, and carbon storage. This expectation is based on the idea that forest functioning depends not only on climate, but also on the way stand structure regulates resource capture and biomass distribution under contrasting environmental conditions. Accordingly, the objectives of this study were to characterize the floristic composition and stand structure of TWF and TSDF, estimate standing volume and carbon stocks from forest inventory data, and evaluate whether variation among these contrasting tropical forests was more closely associated with structural attributes or with diversity-related patterns. In this way, the study specifically addresses the central ecological question of whether forest structure or diversity shows a stronger association with standing volume and carbon storage under contrasting tropical conditions [1,3].

2. Materials and Methods

2.1. Study Areas

This study was conducted in three contrasting tropical forest sites located in Colombia and Mexico [4,7]. Two sites were situated in the Department of Chocó, Colombia, in the municipalities of Medio Baudó (5°11′ N, 76°57′ W) and Nóvita (4°57′ N, 76°37′ W), and were grouped as TWF [4]. The third site was located in the municipality of Tomatlán, Jalisco, Mexico (19°56′ N, 105°15′ W), and was classified as TSDF (Figure 1) [8,9].

For bioclimatic interpretation, the study areas were framed using the Holdridge life-zone system, which relates vegetation patterns to mean biotemperature, annual precipitation, and potential evapotranspiration [7]. Under this framework, the two sites in Colombia correspond to warm and highly humid tropical conditions and are located in the Chocó biogeographic region, which is characterized by persistently warm conditions, very high annual rainfall, and low climatic seasonality. These conditions support dense tropical forest cover and favor the development of structurally complex stands with high woody accumulation. By contrast, the site in Tomatlán, Jalisco, corresponds to a warm subhumid tropical setting with stronger rainfall seasonality, where a pronounced dry period influences canopy dynamics, water availability, and stand development [7,8,10]. This broader climatic contrast provided the environmental basis for comparing floristic organization, standing volume, and carbon storage among the three forests. This classification was used as an ecological reference to support comparisons among sites rather than as a strict vegetation map [7].

The sites in Colombia are located within the Chocó biogeographic region, one of the wettest tropical areas in the world, where persistent rainfall and warm conditions sustain dense forest cover and high aboveground biomass [4,10]. In these forests, climatic conditions strongly influence stand structure, species composition, and carbon accumulation, making them suitable for evaluating ecological controls on forest standing volume under very humid tropical conditions [4,10]. Previous studies in the region have also shown that biomass variation is closely associated with structural disturbance and forest degradation, highlighting the importance of site-level inventories for understanding carbon dynamics in these forests [4].

By contrast, the site in Mexico is located in the tropical Pacific region of western Mexico, where climatic seasonality is more pronounced and warm subhumid conditions predominate [8,9]. These environmental characteristics favor semi-deciduous tropical vegetation, in which part of the canopy loses foliage during the dry season and forest functioning reflects greater temporal variation in water availability [8,11]. This condition makes TSDF an appropriate counterpart to TWF for examining whether variation in standing volume and carbon stocks is driven mainly by broad climatic setting or by the internal organization of the stand [7,9].

2.2. Sampling Design and Forest Inventory

Field sampling was conducted in three contrasting tropical forest sites, including two TWFs in Colombia and one TSDF in Mexico. In each forest, a reference area of 100 ha was established for inventory purposes, and a sampling intensity of 2% was applied, resulting in a total sampled area of 2.0 ha per site. All study sites were located within an altitudinal range of 0–200 m a.s.l., corresponding to lowland tropical forest conditions.

Within each 100-ha site, 40 circular plots of 500 m² were established, each with a radius of 12.62 m. This plot size was selected to provide a consistent representation of stand structure while maintaining operational efficiency during field measurements. The same sampling design was applied across the three forests to ensure methodological comparability among sites.

Within each plot, all trees with a diameter at breast height (DBH) > 10 cm were measured. DBH was recorded at 1.3 m above ground level using a Forestry Suppliers diameter tape (Jackson, MS, USA). Total tree height (H) was measured with a Haga hypsometer (Haga GmbH & Co. KG, Nuremberg, Germany). These measurements were used to characterize forest structure and to derive the dendrometric variables required for the estimation of standing volume and carbon stocks.

The use of a common inclusion threshold and standardized field procedures across the three forests allowed direct comparison of structural attributes among sites. This design provided a robust basis for evaluating how stand organization varied between TWF and TSDF and how such variation was related to differences in standing volume and carbon storage. The sampling design was intended to support site-level comparison among contrasting tropical forest sites under their current field conditions, rather than broad regional generalization or comparison among stands of equivalent age or developmental phase.

2.3. Floristic and Structural Analysis

Floristic composition and horizontal structure were evaluated from the complete inventory of trees with DBH > 10 cm recorded in the sampling plots of each forest. For each species, absolute density, absolute frequency, and absolute dominance were calculated and then expressed as relative density (DR), relative frequency (FR), and relative dominance (DoR), respectively. These parameters were integrated into the Importance Value Index (IVI), which was used to identify the species with the greatest ecological and structural relevance within each forest, an approach widely applied in tropical forest studies because it combines the main components of stand representation into a single synthetic measure [12,13].

The IVI was calculated as follows (1) [12]:

IVI = DR + DoR + FR,

(1)

where DR is the relative density of each species, DoR is its relative dominance, and FR is its relative frequency. Relative density was estimated as the proportion of individuals of a given species with respect to the total number of individuals recorded in the forest. Relative frequency was obtained from the proportion of plots in which each species occurred relative to the sum of frequencies of all species. Relative dominance was calculated from the contribution of each species to total basal area, since basal area is one of the most informative descriptors of tree size occupation and stand structural weight in forest inventories [12,14].

In addition to the species-level assessment, a family-level importance value was also calculated to evaluate the structural contribution of the major taxonomic groups represented in each forest. For this purpose, the relative density, relative frequency, and relative dominance of all species belonging to the same family were summed, and the family importance value index (F-IVI) was obtained as follows (2):

F-IVI = F-DR + F-DoR + F-FR,

(2)

where F-DR, F-DoR, and F-FR correspond to the relative density, relative dominance, and relative frequency of each family, respectively. This aggregation was useful for identifying dominant lineages associated with stand organization and biomass distribution, particularly in forests where species turnover is high but structural dominance may be concentrated in a limited number of families [15,16].

To complement the floristic analysis, alpha diversity was quantified using Shannon’s diversity index (H′), Simpson’s index (1 − D), and Pielou’s evenness (J′), because together these metrics provide information on richness, dominance concentration, and the degree of equitability in species abundances within each forest [17,18,19,20]. These indices were calculated as (3)–(5):

H′ = −∑ pi ln(pi),

(3)

D = ∑ pi²,

(4)

J′ = H′/ln(S),

(5)

where pi is the proportional abundance of species i and S is the total number of species recorded in the forest. Shannon’s index emphasizes the contribution of both common and rare species, Simpson’s index is more sensitive to dominance by abundant taxa, and Pielou’s evenness expresses how evenly individuals are distributed among species, which together provide a more robust basis for comparing floristic organization among contrasting forests than richness alone [20,21].

Because direct comparisons of diversity can be biased by unequal sample completeness, diversity was also evaluated using Hill numbers, where q0 corresponds to species richness, q1 to the exponential of Shannon’s entropy, and q2 to the inverse Simpson concentration. This framework expresses diversity as the effective number of species and allows more consistent comparisons among assemblages with different abundance structures [22,23]. To standardize comparisons among forests, sample-size-based rarefaction and extrapolation curves were generated up to twice the reference sample size, with 95% confidence intervals, in order to evaluate inventory completeness and compare expected diversity under a common sampling basis [21,24].

These floristic and structural attributes were selected because they allow comparisons among the three forests beyond simple species listing. In particular, IVI and F-IVI identify the taxa and lineages that dominate stand organization, while diversity indices and Hill numbers provide complementary information on richness, dominance, and evenness. Together, these metrics offer an ecological basis for examining whether differences in standing volume and carbon stocks are associated with forest composition alone or with broader patterns of structural organization and taxonomic dominance across TWF and TSDF [21,23].

2.4. Dendrometric Attributes and Standing Volume

Dendrometric analysis was based on the tree measurements collected in the sampling plots, specifically DBH and total height, because these variables provide the structural foundation for evaluating stand organization and estimating wood volume in tropical forests [25,26]. At the individual level, basal area was calculated from DBH, and this variable was later aggregated at the plot and site levels to characterize the horizontal occupation of the stand and the structural contribution of trees of different sizes [25].

Basal area was calculated as follows (6):

BA = π/4 × DBH²,

(6)

where BA is the basal area of the tree (m²) and DBH is the diameter at breast height (m). From these values, stand density (trees ha⁻¹), basal area (m² ha⁻¹), mean DBH, mean height, and diameter-class distribution were determined for each forest, since these variables are widely used to describe stand structure and to support comparisons of forest standing volume under contrasting ecological conditions [25,26].

Standing volume was evaluated from individual tree volume estimated with diameter and height measurements, using a morphometric approach based on basal area, total height, and a form factor, which is commonly applied when destructive sampling is not feasible and when inventories are intended to compare standing woody production among forest sites [15,26]. Individual tree volume was estimated as (7)

V = BA × H × f,

(7)

where V is the individual stem volume (m³), BA is the basal area (m²), H is the total tree height (m), and f (0.5) is the form factor. The form factor of 0.5 was adopted as a generalized morphometric coefficient commonly used in forest mensuration when direct measurements of stem form are not available. In this study, it was applied uniformly across the three forests to preserve methodological consistency and comparability among sites, although its use may introduce uncertainty in the absolute standing volume estimates. This procedure allowed volume to be estimated consistently across the three forests using the same field variables and the same computational framework [15,25,27].

At the stand level, standing volume was expressed on an area basis by summing individual tree volume within each plot and then converting the result to m³ ha⁻¹. Because this estimate represents a stock variable rather than a periodic increment, it was interpreted as standing volume and used as a comparative descriptor of woody accumulation among forests at the time of measurement. This approach allowed consistent comparisons among forests with contrasting structural organization and floristic composition, while avoiding conceptual confusion with productivity in its temporal sense [26,28].

To strengthen interpretation, standing volume was examined jointly with stand density, basal area, mean height, and diameter structure, because variation in standing volume may reflect not only differences in tree size but also differences in packing, dominance, and size hierarchy within the stand. This integrated perspective is especially relevant in contrasting tropical forests, where similar volume values may emerge from different structural pathways and ecological configurations [25,28].

2.5. Above- and Belowground Biomass and Carbon Estimation

Aboveground biomass (AGB) was estimated from tree diameter and total height using a generalized allometric model developed for tropical forests, which does not require species-specific wood density values and is therefore suitable for inventories that include a large number of taxa with incomplete functional information. For each tree, AGB was calculated as follows [29] (8):

AGB = exp[−2.977 + ln(DBH² × H)],

(8)

where AGB is the aboveground biomass (kg tree⁻¹), DBH is the diameter at breast height (cm), and H is the total tree height (m). Biomass estimates obtained at the individual level were summed within plots and later expressed on a per-hectare basis to enable comparisons among forests [29].

Species-specific wood density data were not available for a large proportion of the recorded taxa; therefore, a generalized model based on DBH and total height was used to maintain consistency across the three forests.

Belowground biomass (BGB) was estimated indirectly from AGB using a root-to-shoot ratio approach, which is widely applied in tropical forest studies when destructive root sampling is not feasible [30,31]. Following this approach, BGB was calculated as a fixed proportion of AGB (9):

BGB = 0.24 × AGB,

(9)

where 0.24 represents the average belowground-to-aboveground biomass ratio commonly used for tropical forests under generalized estimation frameworks [30,31]. Total biomass (TB) was then obtained as

TB = AGB + BGB,

(10)

Carbon stocks were estimated by applying a conversion factor of 0.47 to both above- and belowground biomass, as recommended by international guidelines for forest carbon accounting [1]. Accordingly, aboveground carbon (AGC), belowground carbon (BGC), and total carbon stock (TC) were calculated as (11)–(13)

AGC = AGB × 0.47,

(11)

BGC = BGB × 0.47,

(12)

TC = TB × 0.47,

(13)

The resulting values were expressed in Mg ha⁻¹ and summarized at the plot, species, family, and site levels. This procedure made it possible to identify the principal taxonomic contributors to biomass and carbon storage and to compare the relative contribution of floristic and structural organization to ecosystem carbon accumulation across TWF and TSDF [29,31].

This approach was selected because it provides a consistent estimation framework when wood density data are unavailable for a large proportion of species, while still preserving comparability among contrasting tropical forests. Although generalized models may be less precise than species-specific equations, they remain useful for broad ecological comparisons and are widely accepted in regional assessments of tropical forest biomass and carbon stocks [29,30,31].

2.6. Statistical Analysis

All analyses were performed at the plot level in order to preserve the sampling structure and allow direct comparisons among the three forests. Structural, floristic, volumetric, biomass, and carbon variables were summarized as mean ± standard error.

Before inferential analyses, data distribution was evaluated using normal probability plots (P–P plots), and homogeneity of variances was verified to determine whether parametric procedures were appropriate; differences among forests were tested using one-way analysis of variance (ANOVA), and when significant differences were detected, Tukey’s multiple comparison test was applied to separate means among forests.

When variables did not satisfy the assumptions required for parametric analysis, differences among forests were evaluated using the Kruskal–Wallis test [32]. This analytical framework was applied to the main response variables, including density, basal area, mean DBH, mean height, standing volume, aboveground biomass, belowground biomass, total biomass, and carbon stocks, in order to determine whether variation among forests was associated with differences in floristic and structural organization, as proposed in the study hypothesis. To strengthen the evaluation of plot-level associations, both Pearson and Spearman correlation coefficients were calculated between the main ecological attributes and the response variables. Pearson correlation was used to describe linear associations under the parametric framework, whereas Spearman correlation was used as a complementary non-parametric approach to verify whether the main patterns remained consistent across correlation metrics. These analyses were interpreted as exploratory associations and not as evidence of direct causal relationships. All analyses were performed using SAS software v9.4, and statistical significance was evaluated at α = 0.05 [33].

3. Results

3.1. Taxonomic Composition and Ecological Dominance

Taxonomic composition varied among the three forests. Both wet forests contained a larger number of recorded species than TSDF, while the diversity curves also showed a clear separation between the humid and seasonal systems. In Figure 2, TWFmb and TWFn maintained higher values of q0, q1, and q2 across the observed and extrapolated sample sizes, whereas TSDF consistently showed lower effective diversity. This pattern was especially evident for q1 and q2, indicating that the lower diversity in TSDF was not only related to species richness, but also to a stronger concentration of individuals in a smaller set of dominant taxa.

The family-level analysis revealed contrasting patterns of ecological dominance among forests. In TWFmb and TWFn, structural importance was distributed among several major lineages, including Fabaceae, Moraceae, Malvaceae, Myristicaceae, Sapotaceae, and Lecythidaceae, indicating a broader family-level contribution to stand organization. By contrast, TSDF was more unevenly structured, with Euphorbiaceae clearly exceeding all other families in density, basal area, and F-IVI, followed at a considerable distance by Malvaceae, Fabaceae, Bignoniaceae, and Fagaceae. This contrast indicates that family-level dominance was more broadly distributed in the two wet forests, whereas in TSDF it was concentrated more strongly in a reduced number of lineages (

Table 1. Family-level structural and ecological attributes across contrasting tropical forests.

Family	Variable	TWFmb	TWFn	TSDF
Euphorbiaceae	Genera (n)	–	–	4
	Species (n)	–	–	4
	Density (ind ha−1)	–	–	74.5
	Basal area (m2 ha−1)	–	–	4.97
	F-IVI	–	–	105.03
Fabaceae	Genera (n)	5	4	7
	Species (n)	5	4	9
	Density (ind ha−1)	58.0	28.0	16.5
	Basal area (m2 ha−1)	4.79	2.36	1.49
	F-IVI	47.48	25.72	32.72
Malvaceae	Genera (n)	5	4	5
	Species (n)	5	5	5
	Density (ind ha−1)	32.0	29.0	27.5
	Basal area (m2 ha−1)	2.72	2.64	0.99
	F-IVI	31.03	28.20	35.17
Moraceae	Genera (n)	2	3	2
	Species (n)	3	3	2
	Density (ind ha−1)	35.0	35.0	3.0
	Basal area (m2 ha−1)	3.04	3.04	1.78
	F-IVI	31.18	31.04	18.35
Sapotaceae	Genera (n)	2	2	–
	Species (n)	2	2	–
	Density (ind ha−1)	29.0	29.0	–
	Basal area (m2 ha−1)	3.26	3.26	–
	F-IVI	24.77	24.66	–
Myristicaceae	Genera (n)	3	3	–
	Species (n)	3	3	–
	Density (ind ha−1)	35.5	35.5	–
	Basal area (m2 ha−1)	2.17	2.17	–
	F-IVI	27.79	27.68	–
Lecythidaceae	Genera (n)	2	3	–
	Species (n)	2	3	–
	Density (ind ha−1)	30.0	30.0	–
	Basal area (m2 ha−1)	1.81	1.81	–
	F-IVI	22.13	23.20	–
Chrysobalanaceae	Genera (n)	–	1	–
	Species (n)	–	1	–
	Density (ind ha−1)	–	30.0	–
	Basal area (m2 ha−1)	–	2.38	–
	F-IVI	–	21.40	–
Burseraceae	Genera (n)	2	2	1
	Species (n)	2	3	1
	Density (ind ha−1)	11.0	11.5	7.5
	Basal area (m2 ha−1)	0.99	1.05	0.42
	F-IVI	11.90	12.62	11.54
Fagaceae	Genera (n)	–	–	1
	Species (n)	–	–	2
	Density (ind ha−1)	–	–	15.5
	Basal area (m2 ha−1)	–	–	1.11
	F-IVI	–	–	18.92

TWFmb = Tropical Wet Forest of Medio Baudó; TWFn = Tropical Wet Forest of Nóvita; TSDF = Tropical Semi-Deciduous Forest of Tomatlán. Density values are expressed as individuals per hectare, and basal area as m² ha⁻¹. F-IVI corresponds to the family importance value index. A dash (–) indicates absence of the family in the corresponding forest. The complete family-level dataset is provided in Table S1.

and Table S1).

At the species level, the same contrast was evident. In TWFmb, the highest IVI values corresponded to Chrysophyllum cainito L., Pithecellobium foreroi C.Barbosa, Brosimum utile (Kunth) Oken, Virola sebifera Aubl., and Eschweilera sclerophylla Cuatrec. In TWFn, the most important species were Pouteria caimito Radlk., Chrysophyllum cainito L., Licania micrantha Miq., Brosimum utile (Kunth) Oken, and Virola reidii Little. In TSDF, dominance was more strongly concentrated in Hura polyandra Baill., followed by Sapium appendiculatum Pax & K.Hoffm., Tabebuia rosea (Bertol.) DC., Guazuma ulmifolia Lam., and Quercus glaucescens Bonpl. (

Table 2. Species-level ecological importance across contrasting tropical forests.

Species	Family	Variable	TWFmb	TWFn	TSDF
Hura polyandra Baill.	Euphorbiaceae	Density (ind ha−1)	–	–	48.50
		Basal area (m2 ha−1)	–	–	4.06
		IVI	–	–	69.31
Sapium appendiculatum Pax & K.Hoffm.		Density (ind ha−1)	–	–	21.50
		Basal area (m2 ha−1)	–	–	0.84
		IVI	–	–	29.05
Chrysophyllum cainito L.	Sapotaceae	Density (ind ha−1)	28.50	28.50	–
		Basal area (m2 ha−1)	3.23	3.23	–
		IVI	24.10	24.00	–
Pouteria caimito Radlk.		Density (ind ha−1)	0.50	23.00	–
		Basal area (m2 ha−1)	0.03	2.40	–
		IVI	0.73	24.00	–
Tabebuia rosea (Bertol.) DC.	Bignoniaceae	Density (ind ha−1)	–	–	16.50
		Basal area (m2 ha−1)	–	–	0.76
		IVI	–	–	23.30
Pithecellobium foreroi C.Barbosa	Fabaceae	Density (ind ha−1)	30.00	–	–
		Basal area (m2 ha−1)	2.38	–	–
		IVI	21.49	–	–
Licania micrantha Miq.	Chrysobalanaceae	Density (ind ha−1)	–	30.00	–
		Basal area (m2 ha−1)	–	2.38	–
		IVI	–	21.40	–
Guazuma ulmifolia Lam.	Malvaceae	Density (ind ha−1)	–	–	19.50
		Basal area (m2 ha−1)	–	–	0.41
		IVI	–	–	20.24
Brosimum utile (Kunth) Oken	Moraceae	Density (ind ha−1)	22.50	22.50	–
		Basal area (m2 ha−1)	2.20	2.20	–
		IVI	18.57	18.49	–
Virola sebifera Aubl.	Myristicaceae	Density (ind ha−1)	26.50	17.50	–
		Basal area (m2 ha−1)	1.72	0.73	–
		IVI	18.04	11.84	–
Virola reidii Little		Density (ind ha−1)	–	26.50	–
		Basal area (m2 ha−1)	–	1.72	–
		IVI	–	17.96	–
Quercus glaucescens Bonpl.	Fagaceae	Density (ind ha−1)	–	–	14.50
		Basal area (m2 ha−1)	–	–	0.85
		IVI	–	–	15.83
Eschweilera sclerophylla Cuatrec.	Lecythidaceae	Density (ind ha−1)	21.50	19.50	–
		Basal area (m2 ha−1)	1.42	1.32	–
		IVI	15.44	14.44	–
Ficus cotinifolia Kunth	Moraceae	Density (ind ha−1)	–	–	2.00
		Basal area (m2 ha−1)	–	–	1.61
		IVI	–	–	15.25
Pourouma chocoana Standl.	Urticaceae	Density (ind ha−1)	22.00	22.00	–
		Basal area (m2 ha−1)	1.23	1.23	–
		IVI	14.54	14.48	–
Inga edulis Mart.	Fabaceae	Density (ind ha−1)	17.00	–	–
		Basal area (m2 ha−1)	1.19	–	–
		IVI	13.26	–	–
Inga nobilis Willd.		Density (ind ha−1)	–	17.00	–
		Basal area (m2 ha−1)	–	1.19	–
		IVI	–	13.21	–
Sterculia colombiana Sprague	Malvaceae	Density (ind ha−1)	–	17.00	–
		Basal area (m2 ha−1)	–	1.28	–
		IVI	–	13.16	–
Sterculia apetala (Jacq.) H.Karst.		Density (ind ha−1)	16.50	–	–
		Basal area (m2 ha−1)	1.20	–	–
		IVI	12.78	–	–
Bursera excelsa (Kunth) Engl.	Burseraceae	Density (ind ha−1)	–	–	7.50
		Basal area (m2 ha−1)	–	–	0.42
		IVI	–	–	11.54
Annona purpurea Moc. & Sessé ex Dunal	Annonaceae	Density (ind ha−1)	–	–	9.00
		Basal area (m2 ha−1)	–	–	0.24
		IVI	–	–	11.59
Enterolobium cyclocarpum (Jacq.) Griseb.	Fabaceae	Density (ind ha−1)	–	–	2.50
		Basal area (m2 ha−1)	–	–	0.93
		IVI	–	–	10.97
Simaba guianensis Aubl.	Simaroubaceae	Density (ind ha−1)	–	11.50	–
		Basal area (m2 ha−1)	–	1.21	–
		IVI	–	10.86	–
Homalolepis cedron (Planch.) Devecchi & Pirani		Density (ind ha−1)	10.00	–	–
		Basal area (m2 ha−1)	1.12	–	–
		IVI	10.15	–	–
Minquartia guianensis Aubl.	Olacaceae	Density (ind ha−1)	9.50	9.50	–
		Basal area (m2 ha−1)	0.79	0.79	–
		IVI	9.57	9.53	–

Density values are expressed as individuals per hectare, basal area as m² ha⁻¹, and IVI as the importance value index. A dash (–) indicates absence of the species in the corresponding forest. The complete species-level dataset is provided in Table S2.

and Table S2). Taken together, these results indicate that TWFmb and TWFn were characterized by broader taxonomic representation and a more even distribution of structural importance among families and species, whereas TSDF showed lower effective diversity and a more concentrated pattern of ecological dominance. In TSDF, the Euphorbiaceae pattern was supported mainly by Hura polyandra and Sapium appendiculatum, while family-level dominance was completed by two additional species recorded in the forest (Table 1, Table 2 and Table S2).

3.2. Structural Variation Across Contrasting Tropical Forests

Dendrometric structure differed among forests, although this contrast was not expressed uniformly across all attributes. Tree density, basal area, mean height, and standing volume varied significantly among forests, whereas mean DBH did not differ significantly (p = 0.705). In all significant variables, TWFmb and TWFn remained statistically similar to each other, while TSDF differed from both wet forests. Relative to TWFmb and TWFn, TSDF showed reductions of approximately 41% in stem density, 52% in basal area, 18% in mean height, and 56% in standing volume, indicating a structurally less developed stand condition despite the absence of significant differences in mean DBH (

Table 3. Dendrometric attributes and standing volume across contrasting tropical forests.

Variable	TWFmb	TWFn	TSDF
Density (trees ha−1)	334.0 ± 22.9 a	335.0 ± 26.9 a	196.0 ± 18.3 b
Basal area (m2 ha−1)	27.55 ± 2.04 a	27.71 ± 2.32 a	13.14 ± 1.69 b
Mean DBH (cm)	29.43 ± 0.27 a	29.53 ± 0.56 a	27.07 ± 2.30 a
Mean height (m)	15.31 ± 0.16 a	15.33 ± 0.24 a	12.62 ± 0.29 b
Standing volume (m3 ha−1)	239.35 ± 18.31 a	240.63 ± 20.29 a	104.64 ± 17.30 b

Values are presented as mean ± standard error. Different lowercase letters within rows indicate significant differences among forests at p ≤ 0.05. TWFmb = Tropical Wet Forest of Medio Baudó; TWFn = Tropical Wet Forest of Nóvita; TSDF = Tropical Semi-Deciduous Forest of Tomatlán.

). This apparent contrast is explained by the fact that stand-level basal area reflects not only mean DBH, but also stem density and the distribution of individuals across diameter classes, particularly the reduced representation of intermediate and large trees in TSDF.

The diameter structure also showed a clear separation among forests. In TWFmb and TWFn, the greatest concentration of stems occurred in the 20–29.9 cm class, which represented 31.3% of all individuals in both forests, followed by the 10–19.9 cm and 30–39.9 cm classes, each contributing about one fifth of the total inventory. In TSDF, by contrast, nearly half of all stems (48.7%) were concentrated in the 10–19.9 cm class, while the relative contribution of intermediate classes declined sharply. This shift in stem distribution was reflected in significant differences from the 20–29.9 cm class onward, where TWFmb and TWFn consistently exceeded TSDF in stem density (p ≤ 0.001 for the 20–29.9, 30–39.9, 40–49.9, 50–59.9, and 60–69.9 cm classes;

Table 4. Diameter-class distribution expressed as tree density per hectare across contrasting tropical forests.

Diameter Class (cm)	TWFmb	TWFn	TSDF
10–19.9	68.5 ± 6.4 a	68.5 ± 7.9 a	95.5 ± 12.2 a
20–29.9	104.5 ± 10.7 a	105.0 ± 12.2 a	52.0 ± 7.4 b
30–39.9	68.0 ± 6.7 a	68.0 ± 6.3 a	27.0 ± 5.9 b
40–49.9	54.0 ± 4.1 a	54.0 ± 5.6 a	11.5 ± 2.4 b
50–59.9	25.0 ± 3.6 a	25.0 ± 3.9 a	3.5 ± 1.4 b
60–69.9	14.0 ± 2.7 a	14.5 ± 2.8 a	3.5 ± 1.2 b
70–79.9	0.0 ± 0.0 a	0.0 ± 0.0 a	0.5 ± 0.5 a
80–89.9	0.0 ± 0.0 a	0.0 ± 0.0 a	0.0 ± 0.0 a
90–99.9	0.0 ± 0.0 a	0.0 ± 0.0 a	1.0 ± 0.7 a
100–109.9	0.0 ± 0.0 a	0.0 ± 0.0 a	0.0 ± 0.0 a
110–129.9	0.0 ± 0.0 a	0.0 ± 0.0 a	1.0 ± 0.7 a
130–159.9	0.0 ± 0.0 a	0.0 ± 0.0 a	0.5 ± 0.5 a

Values are presented as mean ± standard error. Different lowercase letters within rows indicate significant differences among forests at p ≤ 0.05. TWFmb = Tropical Wet Forest of Medio Baudó; TWFn = Tropical Wet Forest of Nóvita; TSDF = Tropical Semi-Deciduous Forest of Tomatlán.

).

The strongest structural contrast was observed in the intermediate diameter range. In the 30–39.9 cm class, TWFmb and TWFn retained about 20% of their stems, whereas TSDF contained only 13.8%. The same pattern became more pronounced in the 40–49.9 cm class, which represented 16.2% and 16.1% of stems in TWFmb and TWFn, respectively, but only 5.9% in TSDF. Above 50 cm DBH, the proportional contribution of stems declined in all forests, although the reduction was steeper in TSDF, where the combined contribution of classes above 50 cm remained below 5% of the total inventory (Table 4).

At the lower end of the distribution, no significant differences were detected in the 10–19.9 cm class (p = 0.237), indicating that recruitment-sized and small stems were abundant in all three forests. However, the structural role of this class differed among systems. In TSDF, the concentration of nearly half of all stems in this class defined a narrower size structure and was accompanied by a marked reduction in the representation of intermediate and large trees. As a result, a substantial fraction of the stand remained concentrated in lower-size categories, which helps explain the lower basal area, standing volume, biomass, and carbon stocks observed in this forest relative to the two wet forests. In contrast, in TWFmb and TWFn this class represented only about one fifth of individuals, reflecting a broader distribution of stems across intermediate sizes and a more structurally developed stand profile, with greater contribution of these classes to woody accumulation. This difference in diameter allocation was consistent with the higher basal area and standing volume observed in the two wet forests (Table 3 and Table 4).

The contribution of taxonomic groups to standing volume differed significantly among forests, although the magnitude of these differences varied among families (

Table 5. Family-level contribution to standing volume across contrasting tropical forests.

Family	TWFmb (m3 ha−1)	TWFn (m3 ha−1)	TSDF (m3 ha−1)
Euphorbiaceae	0.00 ± 0.00 b	4.08 ± 2.72 b	37.90 ± 6.60 a
Fabaceae	41.27 ± 7.33 a	19.62 ± 4.30 a	12.76 ± 5.98 b
Malvaceae	23.76 ± 5.87 a	23.34 ± 7.14 a	6.83 ± 2.61 b
Moraceae	27.61 ± 3.81 a	27.61 ± 4.69 a	18.15 ± 16.72 a
Sapotaceae	28.15 ± 2.84 a	28.15 ± 3.19 a	0.00 ± 0.00 b
Myristicaceae	17.04 ± 3.22 a	17.04 ± 3.84 a	0.00 ± 0.00 b
Lecythidaceae	16.50 ± 3.65 a	16.11 ± 3.46 a	0.00 ± 0.00 b
Chrysobalanaceae	0.00 ± 0.00 b	21.26 ± 3.07 a	2.99 ± 2.38 b
Burseraceae	8.78 ± 2.49 a	9.29 ± 2.66 a	2.98 ± 1.61 a
Fagaceae	0.00 ± 0.00 a	0.00 ± 0.00 a	8.48 ± 4.67 a

Values are presented as mean ± standard error. Different lowercase letters within rows indicate significant differences among forests at p ≤ 0.05. TWFmb = Tropical Wet Forest of Medio Baudó; TWFn = Tropical Wet Forest of Nóvita; TSDF = Tropical Semi-Deciduous Forest of Tomatlán.

and Table S3). In TSDF, standing volume was strongly concentrated in Euphorbiaceae, which accounted for 36.2% of total standing volume and differed significantly from both wet forests (p < 0.001). In contrast, TWFmb and TWFn distributed standing volume more broadly among several families, with no single lineage reaching such a disproportionate contribution.

Fabaceae was the principal volumetric component in TWFmb, where it represented 17.2% of total standing volume, whereas its contribution declined to 8.2% in TWFn and 12.2% in TSDF. This reduction was significant in TSDF relative to the two wet forests (p = 0.029). A similar pattern was observed for Malvaceae, whose contribution remained close to 10% in both wet forests but declined to 6.5% in TSDF (p = 0.006). Together, these results indicate that a substantial fraction of the standing volume in TWFmb and TWFn was supported by families that combined relatively high abundance with greater structural development.

Other families showed more restricted but still relevant contributions. Sapotaceae accounted for nearly 12% of total standing volume in both wet forests, but was absent from TSDF, resulting in a highly significant contrast among forests (p < 0.001). The same tendency was recorded for Myristicaceae and Lecythidaceae, both of which contributed appreciably to standing volume in TWFmb and TWFn but were not represented in TSDF (p < 0.001 in both cases). In TWFn, Chrysobalanaceae represented 8.8% of total standing volume and differed significantly from the other two forests (p < 0.001), reinforcing the distinctive volumetric composition of that forest.

Not all families showed statistical separation among forests. Moraceae contributed between 11.5% and 17.3% of total standing volume across the three systems and did not differ significantly among them (p = 0.903), despite its lower density in TSDF. Burseraceae also showed a relatively stable contribution, ranging from 2.9% to 3.9%, with no significant differences among forests (p = 0.059). In TSDF, Fagaceae represented 8.1% of total standing volume, but this contribution was not significantly different from the absence of this family in the two wet forests (p = 0.351).

Overall, standing volume in TWFmb and TWFn was distributed among a broader set of families, whereas TSDF showed a more concentrated volumetric pattern dominated by Euphorbiaceae. This contrast indicates that the higher standing volume observed in the wet forests was supported by a more even family-level contribution, while in TSDF a substantial fraction of the total volume was concentrated in a smaller number of dominant lineages.

3.3. Biomass Allocation and Carbon Storage

Biomass allocation differed significantly among forests for both aboveground and belowground components (ANOVA, p = 5.76 × 10⁻⁶ in both cases). TWFmb and TWFn did not differ significantly from one another, whereas TSDF showed significantly lower values for both fractions (Figure 3a). In both wet forests, aboveground and belowground biomass were more than twice those recorded in TSDF, indicating substantially greater biomass accumulation under the more humid conditions represented by these sites. Carbon stocks followed the same pattern, with no significant differences between TWFmb and TWFn and significantly lower values in TSDF for both aboveground and belowground pools (Figure 3b). Overall, the greater biomass accumulation observed in the two wet forests was directly reflected in higher carbon storage, whereas TSDF maintained a markedly lower biomass–carbon balance.

At the family level, biomass and carbon allocation (

Table 6. Family-level biomass and carbon allocation across contrasting tropical forests.

Family	Variable	TWFmb	TWFn	TSDF
Euphorbiaceae	Aboveground biomass (Mg ha−1)	0.00 ± 0.00 b	5.30 ± 3.53 b	49.17 ± 8.57 a
	Belowground biomass (Mg ha−1)	0.00 ± 0.00 b	1.27 ± 0.85 b	11.80 ± 2.06 a
	Total biomass (Mg ha−1)	0.00 ± 0.00 b	6.57 ± 4.38 b	60.98 ± 10.62 a
	Aboveground carbon (Mg C ha−1)	0.00 ± 0.00 b	2.49 ± 1.66 b	23.11 ± 4.03 a
	Belowground carbon (Mg C ha−1)	0.00 ± 0.00 b	0.60 ± 0.40 b	5.55 ± 0.97 a
	Total carbon (Mg C ha−1)	0.00 ± 0.00 b	3.09 ± 2.06 b	28.66 ± 4.99 a
Fabaceae	Aboveground biomass (Mg ha−1)	53.54 ± 9.51 a	25.45 ± 5.58 a	16.55 ± 7.75 b
	Belowground biomass (Mg ha−1)	12.85 ± 2.28 a	6.11 ± 1.34 a	3.97 ± 1.86 b
	Total biomass (Mg ha−1)	66.39 ± 11.79 a	31.56 ± 6.93 a	20.52 ± 9.62 b
	Aboveground carbon (Mg C ha−1)	25.16 ± 4.47 a	11.96 ± 2.62 a	7.78 ± 3.64 b
	Belowground carbon (Mg C ha−1)	6.04 ± 1.07 a	2.87 ± 0.63 a	1.87 ± 0.87 b
	Total carbon (Mg C ha−1)	31.20 ± 5.54 a	14.83 ± 3.25 a	9.64 ± 4.52 b
Malvaceae	Aboveground biomass (Mg ha−1)	30.82 ± 7.62 a	30.27 ± 9.26 a	8.85 ± 3.38 b
	Belowground biomass (Mg ha−1)	7.40 ± 1.83 a	7.27 ± 2.22 a	2.13 ± 0.81 b
	Total biomass (Mg ha−1)	38.22 ± 9.45 a	37.54 ± 11.48 a	10.98 ± 4.19 b
	Aboveground carbon (Mg C ha−1)	14.49 ± 3.58 a	14.23 ± 4.35 a	4.16 ± 1.59 b
	Belowground carbon (Mg C ha−1)	3.48 ± 0.86 a	3.41 ± 1.04 a	1.00 ± 0.38 b
	Total carbon (Mg C ha−1)	17.96 ± 4.44 a	17.64 ± 5.40 a	5.16 ± 1.97 b
Moraceae	Aboveground biomass (Mg ha−1)	35.81 ± 4.94 a	35.81 ± 6.08 a	23.54 ± 21.69 a
	Belowground biomass (Mg ha−1)	8.60 ± 1.18 a	8.60 ± 1.46 a	5.65 ± 5.21 a
	Total biomass (Mg ha−1)	44.41 ± 6.12 a	44.41 ± 7.54 a	29.19 ± 26.90 a
	Aboveground carbon (Mg C ha−1)	16.83 ± 2.32 a	16.83 ± 2.86 a	11.06 ± 10.19 a
	Belowground carbon (Mg C ha−1)	4.04 ± 0.56 a	4.04 ± 0.69 a	2.66 ± 2.45 a
	Total carbon (Mg C ha−1)	20.87 ± 2.88 a	20.87 ± 3.54 a	13.72 ± 12.64 a
Sapotaceae	Aboveground biomass (Mg ha−1)	36.52 ± 3.69 a	36.52 ± 4.14 a	0.00 ± 0.00 b
	Belowground biomass (Mg ha−1)	8.76 ± 0.89 a	8.76 ± 0.99 a	0.00 ± 0.00 b
	Total biomass (Mg ha−1)	45.28 ± 4.58 a	45.28 ± 5.14 a	0.00 ± 0.00 b
	Aboveground carbon (Mg C ha−1)	17.16 ± 1.73 a	17.16 ± 1.95 a	0.00 ± 0.00 b
	Belowground carbon (Mg C ha−1)	4.12 ± 0.42 a	4.12 ± 0.47 a	0.00 ± 0.00 b
	Total carbon (Mg C ha−1)	21.28 ± 2.15 a	21.28 ± 2.41 a	0.00 ± 0.00 b
Myristicaceae	Aboveground biomass (Mg ha−1)	22.11 ± 4.17 a	22.11 ± 4.98 a	0.00 ± 0.00 b
	Belowground biomass (Mg ha−1)	5.31 ± 1.00 a	5.31 ± 1.20 a	0.00 ± 0.00 b
	Total biomass (Mg ha−1)	27.42 ± 5.18 a	27.42 ± 6.18 a	0.00 ± 0.00 b
	Aboveground carbon (Mg C ha−1)	10.39 ± 1.96 a	10.39 ± 2.34 a	0.00 ± 0.00 b
	Belowground carbon (Mg C ha−1)	2.49 ± 0.47 a	2.49 ± 0.56 a	0.00 ± 0.00 b
	Total carbon (Mg C ha−1)	12.89 ± 2.43 a	12.89 ± 2.90 a	0.00 ± 0.00 b
Lecythidaceae	Aboveground biomass (Mg ha−1)	21.40 ± 4.73 a	20.90 ± 4.49 a	0.00 ± 0.00 b
	Belowground biomass (Mg ha−1)	5.14 ± 1.14 a	5.01 ± 1.08 a	0.00 ± 0.00 b
	Total biomass (Mg ha−1)	26.54 ± 5.87 a	25.91 ± 5.56 a	0.00 ± 0.00 b
	Aboveground carbon (Mg C ha−1)	10.06 ± 2.22 a	9.82 ± 2.11 a	0.00 ± 0.00 b
	Belowground carbon (Mg C ha−1)	2.41 ± 0.53 a	2.36 ± 0.51 a	0.00 ± 0.00 b
	Total carbon (Mg C ha−1)	12.47 ± 2.76 a	12.18 ± 2.61 a	0.00 ± 0.00 b
Chrysobalanaceae	Aboveground biomass (Mg ha−1)	0.00 ± 0.00 b	27.58 ± 3.98 a	3.87 ± 3.09 b
	Belowground biomass (Mg ha−1)	0.00 ± 0.00 b	6.62 ± 0.96 a	0.93 ± 0.74 b
	Total biomass (Mg ha−1)	0.00 ± 0.00 b	34.20 ± 4.94 a	4.80 ± 3.83 b
	Aboveground carbon (Mg C ha−1)	0.00 ± 0.00 b	12.96 ± 1.87 a	1.82 ± 1.45 b
	Belowground carbon (Mg C ha−1)	0.00 ± 0.00 b	3.11 ± 0.45 a	0.44 ± 0.35 b
	Total carbon (Mg C ha−1)	0.00 ± 0.00 b	16.08 ± 2.32 a	2.26 ± 1.80 b
Burseraceae	Aboveground biomass (Mg ha−1)	11.40 ± 3.23 a	12.05 ± 3.44 a	3.87 ± 2.09 a
	Belowground biomass (Mg ha−1)	2.74 ± 0.78 a	2.89 ± 0.83 a	0.93 ± 0.50 a
	Total biomass (Mg ha−1)	14.13 ± 4.01 a	14.94 ± 4.27 a	4.79 ± 2.59 a
	Aboveground carbon (Mg C ha−1)	5.36 ± 1.52 a	5.66 ± 1.62 a	1.82 ± 0.98 a
	Belowground carbon (Mg C ha−1)	1.29 ± 0.36 a	1.36 ± 0.39 a	0.44 ± 0.24 a
	Total carbon (Mg C ha−1)	6.64 ± 1.88 a	7.02 ± 2.01 a	2.25 ± 1.22 a
Fagaceae	Aboveground biomass (Mg ha−1)	0.00 ± 0.00 a	0.00 ± 0.00 a	11.00 ± 6.06 a
	Belowground biomass (Mg ha−1)	0.00 ± 0.00 a	0.00 ± 0.00 a	2.64 ± 1.45 a
	Total biomass (Mg ha−1)	0.00 ± 0.00 a	0.00 ± 0.00 a	13.64 ± 7.51 a
	Aboveground carbon (Mg C ha−1)	0.00 ± 0.00 a	0.00 ± 0.00 a	5.17 ± 2.85 a
	Belowground carbon (Mg C ha−1)	0.00 ± 0.00 a	0.00 ± 0.00 a	1.24 ± 0.68 a
	Total carbon (Mg C ha−1)	0.00 ± 0.00 a	0.00 ± 0.00 a	6.41 ± 3.53 a
Total	Aboveground biomass (Mg ha−1)	310.52 ± 23.76 a	312.17 ± 26.33 a	135.76 ± 22.44 b
	Belowground biomass (Mg ha−1)	74.52 ± 5.70 a	74.92 ± 6.32 a	32.58 ± 5.39 b
	Total biomass (Mg ha−1)	385.04 ± 29.46 a	387.09 ± 32.65 a	168.34 ± 27.83 b
	Aboveground carbon (Mg C ha−1)	145.94 ± 11.17 a	146.72 ± 12.37 a	63.81 ± 10.55 b
	Belowground carbon (Mg C ha−1)	35.03 ± 2.68 a	35.21 ± 2.97 a	15.31 ± 2.53 b
	Total carbon (Mg C ha−1)	180.97 ± 13.85 a	181.93 ± 15.34 a	79.12 ± 13.08 b

Values are presented as mean ± standard error. Different lowercase letters within rows indicate significant differences among forests at p ≤ 0.05. Comparisons were based on the occurrence of each family across the available sampling units in each forest; therefore, some families were represented in only one or two forests. TWFmb = Tropical Wet Forest of Medio Baudó; TWFn = Tropical Wet Forest of Nóvita; TSDF = Tropical Semi-Deciduous Forest of Tomatlán.

) also differed among forests, although the magnitude and direction of these differences varied among taxonomic groups. In TSDF, Euphorbiaceae dominated both biomass and carbon storage, accounting for 36.2% of total standing volume and concentrating the largest share of aboveground and belowground biomass among all families. This dominance was reflected in significantly higher values than those recorded in TWFmb and TWFn for all biomass and carbon components (p < 0.001), indicating that a substantial proportion of ecosystem biomass in TSDF was supported by a single lineage.

In contrast, biomass and carbon in the two wet forests were distributed more evenly among several dominant families. In TWFmb, Fabaceae contributed the largest share of total biomass and carbon, followed by Sapotaceae, Moraceae, and Malvaceae, whereas in TWFn Chrysobalanaceae also emerged as an important contributor. By comparison, the contribution of Fabaceae and Malvaceae declined significantly in TSDF (p = 0.029 and p = 0.006, respectively). Some families were strongly associated with the wet forests. Sapotaceae, Myristicaceae, and Lecythidaceae made substantial contributions to biomass and carbon in TWFmb and TWFn, but were not represented in TSDF, resulting in highly significant contrasts among forests (p < 0.001 in all cases). Chrysobalanaceae showed the opposite pattern, with its highest contribution in TWFn and much lower values in the other two forests (p < 0.001), reinforcing the distinct family-level allocation pattern of that forest.

Overall, the two wet forests stored biomass and carbon through a broader set of dominant families, whereas TSDF concentrated a larger fraction of both pools in fewer lineages, especially Euphorbiaceae. This pattern was consistent with the stronger structural concentration previously observed in TSDF and with the more even distribution of taxonomic importance in TWFmb and TWFn.

3.4. Ecological Attributes Associated with Standing Volume and Carbon Stocks

Ecological attributes showed contrasting levels of association with standing volume and carbon storage, but the strongest relationships were consistently linked to stand structure rather than to diversity alone. Basal area showed the highest correlation with standing volume, total biomass, and total carbon (r = 0.991, p < 0.001), indicating that horizontal occupation of the stand was the variable most closely associated with the accumulation of woody volume and carbon pools under the analytical framework used in this study. Mean height also showed a strong positive association with these response variables (r = 0.828, p < 0.001), followed by mean DBH (r = 0.693, p < 0.001), reinforcing the importance of tree size and structural development in explaining differences among sampling units. The same variables also showed the strongest associations in the complementary Spearman analysis, confirming that the main correlation pattern remained stable across both parametric and non-parametric approaches (

Table 7. Pearson correlation coefficients between ecological attributes and standing volume, total biomass, and total carbon across sampling units.

Ecological Attribute	Standing Volume	Total Biomass	Total Carbon
Density	r = 0.498, p < 0.001	r = 0.498, p < 0.001	r = 0.498, p < 0.001
Basal area	r = 0.991, p < 0.001	r = 0.991, p < 0.001	r = 0.991, p < 0.001
Mean DBH	r = 0.693, p < 0.001	r = 0.693, p < 0.001	r = 0.693, p < 0.001
Mean height	r = 0.828, p < 0.001	r = 0.828, p < 0.001	r = 0.828, p < 0.001
Richness	r = 0.683, p < 0.001	r = 0.683, p < 0.001	r = 0.683, p < 0.001
Shannon	r = 0.637, p < 0.001	r = 0.637, p < 0.001	r = 0.637, p < 0.001
Simpson	r = 0.646, p < 0.001	r = 0.646, p < 0.001	r = 0.646, p < 0.001
Pielou	r = 0.307, p = 0.030	r = 0.307, p = 0.030	r = 0.307, p = 0.030
q1	r = 0.671, p < 0.001	r = 0.671, p < 0.001	r = 0.671, p < 0.001
q2	r = 0.660, p < 0.001	r = 0.660, p < 0.001	r = 0.660, p < 0.001

Pearson correlations were calculated at the sampling-unit level. The complementary Spearman correlation analysis is presented in Table S5. DBH = diameter at breast height; q1 = exponential Shannon diversity; q2 = inverse Simpson diversity.

and Table S5).

Diversity-related variables were also positively associated with standing volume and carbon, although with lower coefficients than those observed for structural attributes. Species richness showed a moderate to strong correlation with standing volume, total biomass, and total carbon (r = 0.683, p < 0.001). A similar pattern was found for Shannon diversity, Simpson diversity, and Hill numbers q1 and q2, all of which were significantly and positively related to the three response variables (p < 0.001 in all cases). These associations indicate that sampling units with greater taxonomic diversity also tended to store more volume and carbon, although the strength of this relationship remained below that of basal area and tree dimensions.

Among the ecological attributes evaluated, Pielou’s evenness showed the weakest association with standing volume and carbon storage (r = 0.307, p = 0.030). Although still significant, this lower coefficient suggests that the degree of equitability in species abundances had less influence on volume and carbon accumulation than structural attributes such as basal area, height, and DBH. Overall, the correlation pattern indicates that standing volume and carbon stocks were more closely related to stand occupation and tree size than to the evenness of species distribution.

4. Discussion

4.1. Floristic Turnover and Ecological Dominance Across Contrasting Tropical Forests

The floristic differentiation observed among the three forests is consistent with the strong environmental and biogeographic contrast between humid tropical forests and seasonally dry tropical forests in the Neotropics. Recent evidence shows that floristic composition across Neotropical forests is strongly structured by environmental filtering, geographic location, water availability, and seasonality, which together promote marked compositional distinctiveness among forest types. Under that framework, the lower effective diversity recorded in TSDF and the broader taxonomic representation observed in TWFmb and TWFn are ecologically coherent with the contrast between wetter and more seasonal tropical systems [34].

The more even distribution of ecological importance among families in the two wet forests also agrees with patterns described for humid tropical forests, where stand organization is often supported by several dominant lineages rather than by a single overwhelmingly dominant family. In evergreen tropical forests, families such as Moraceae, Fabaceae, Sapotaceae, and related canopy groups frequently play major structural roles, although the identity and relative contribution of dominant species may vary among sites. In this study, TWFmb and TWFn followed that general pattern, with ecological dominance distributed across several families instead of being strongly concentrated in one lineage [35,36].

By contrast, the stronger concentration of ecological dominance in TSDF is also consistent with recent work on Neotropical dry forests. Tropical dry forests commonly show high compositional turnover, strong local dominance, and a greater tendency for a relatively small set of families to account for a substantial proportion of abundance and structural importance. Studies from Mesoamerica and Mexico have shown that dry forests may sustain considerable local diversity, but that this diversity is often organized under stronger environmental filtering and sharper variation in dominance among patches and vegetation types than in wetter forests. Our results fit that pattern, particularly because Euphorbiaceae assumed a disproportionately large structural role in TSDF, while Malvaceae and Fabaceae contributed secondarily [34,37].

The family-level results also suggest that floristic turnover among the three forests was expressed not only through species replacement, but through changes in the taxonomic basis of ecological dominance. Fabaceae and Malvaceae remained relevant across more than one forest, but their structural role was comparatively greater in the wet forests, whereas Euphorbiaceae became especially important in TSDF. This distinction is important because recent large-scale analyses have shown that Neotropical forest assemblages can remain floristically distinct even when they share widespread genera or functional groups, precisely because the balance of taxonomic dominance shifts among forest types and environmental contexts [36].

The diversity results point in the same direction. Higher q1 and q2 values in TWFmb and TWFn indicate that the wet forests did not differ from TSDF only in richness, but also in the way abundance was distributed among taxa. In TSDF, the lower effective diversity reflected a greater concentration of abundance in fewer dominant taxa, whereas the wet forests showed broader taxonomic participation in stand organization. This interpretation is compatible with recent studies indicating that dry forests can maintain substantial biodiversity value while still exhibiting stronger local dominance and sharper floristic partitioning than wetter tropical formations [37,38].

4.2. Structural Organization as a Basis for Differences in Standing Volume

The structural differences observed among the three forests indicate that standing volume was more closely associated with stand development than with stem abundance alone. In our study, TWFmb and TWFn showed higher basal area, mean height, and standing volume than TSDF, while mean DBH did not differ significantly among forests. This apparent contrast is explained by the fact that stand-level basal area depends not only on mean DBH, but also on stem density and the distribution of individuals across diameter classes, particularly the reduced representation of intermediate and large tree stems in TSDF. This pattern agrees with recent evidence showing that forest structure, particularly basal area and height-related attributes, is a stronger predictor of aboveground biomass and standing volume than other ecological dimensions in tropical forests [39,40].

The higher standing volume recorded in the two wet forests is also consistent with the broader structural profile observed in their diameter distributions. TWFmb and TWFn maintained a greater proportion of stems in intermediate classes, whereas TSDF concentrated a larger fraction of individuals in the smallest size class and showed a sharper decline toward larger diameters. Recent work in seasonally dry tropical forests has similarly shown that these forests often display strong concentration in smaller tree stems and greater structural dynamism, even when basal area increases through time, reflecting a distinct balance among recruitment, mortality, and stand reorganization [41]. Under that perspective, the narrower diameter structure of TSDF in our study helps explain its lower volumetric accumulation relative to the two wet forests.

Basal area appears especially important in this context. The very strong relationship found in our data between basal area and standing volume supports the interpretation that horizontal occupation of the stand was the main structural basis of standing volume differences among forests. This interpretation is in line with recent studies showing that forest structural attributes consistently explain variation in biomass and standing volume better than diversity or trait-based metrics alone, particularly when forests differ in canopy development and stem packing [39,40]. In practical terms, the results suggest that the larger standing volume in TWFmb and TWFn was sustained by a more developed structural organization rather than by a simple increase in stem number.

The results from TSDF are also consistent with recent studies from tropical dry forests in Mexico, where carbon density and biomass variation have been linked strongly to structural richness and other stand-level biophysical attributes. In those systems, structural attributes often outweigh purely taxonomic descriptors when explaining spatial variation in aboveground carbon or woody accumulation [42]. Our results support that same tendency, since TSDF combined lower basal area, lower mean height, and lower standing volume, even though mean DBH was not significantly different. This suggests that lower standing volume in TSDF was expressed through a more restricted structural arrangement rather than through uniformly smaller tree diameters.

Taken together, these patterns indicate that standing volume in the three forests was determined primarily by structural organization, especially by the extent of basal occupation and the representation of intermediate and large trees within the stand. The two wet forests shared a more expanded structural profile and, consequently, greater standing volume, whereas TSDF showed a more compressed size structure and lower volumetric accumulation. In this sense, the contrast among forests was not simply one of tree abundance, but of how stand structure translated that abundance into woody volume.

4.3. Taxonomic Control of Biomass Accumulation and Carbon Storage

The biomass and carbon patterns observed in this study indicate that differences among forests were not expressed only in total magnitude, but also in the taxonomic groups supporting those pools. In the two wet forests, biomass and carbon were distributed among several dominant families, whereas in TSDF, a substantial fraction of both pools was concentrated in fewer lineages, especially Euphorbiaceae. This contrast is consistent with recent evidence showing that carbon storage in tropical forests is often controlled by a relatively small subset of dominant taxa, particularly when forest structure is uneven or when environmental filtering favors a reduced number of lineages with strong structural influence [37]. In the case of TSDF, the dominance of Euphorbiaceae may also reflect the ecological suitability of this family under more seasonal and water-limited conditions, where persistence under climatic stress and effective occupation of available growing space may favor its structural contribution within the stand. Although functional traits were not measured directly in this study, this interpretation is consistent with the marked concentration of biomass and carbon observed in this family.

This pattern should not be interpreted as an isolated anomaly. In Mexican seasonally dry and semi-deciduous forests, strong local dominance by a reduced set of taxa has been reported as part of the marked floristic heterogeneity of these systems. In this context, the dominance of Euphorbiaceae in the TSDF is ecologically plausible rather than exceptional. Moreover, Hura polyandra has been described as one of the representative trees of Mexican semi-deciduous forests, and Sapium appendiculatum has been treated in the Mexican floristic literature as Sebastiania appendiculata, an endemic taxon. Given the limited number of sites included in this study, however, the TSDF should be interpreted as a site-based example of this broader ecological tendency rather than as a unique condition.

The strong contribution of Fabaceae, Sapotaceae, Moraceae, and Myristicaceae in the two wet forests is ecologically plausible. These families are frequently associated with large or medium-to-large trees, persistent canopy occupation, and high contribution to woody biomass in humid tropical forests. In our study, the two wet forests showed not only larger total biomass and carbon stocks, but also a broader taxonomic distribution of those stocks across families. That pattern agrees with recent work showing that forests with greater structural development tend to distribute biomass among multiple dominant groups rather than concentrating it in a single lineage [39,43].

By contrast, TSDF showed a narrower taxonomic basis for biomass and carbon storage. Euphorbiaceae accounted for the largest share of both components, while Fabaceae, Malvaceae, Moraceae, and Fagaceae contributed secondarily. This concentration is consistent with the behavior of many seasonally dry tropical forests, where climatic seasonality and water limitation favor a smaller set of structurally dominant lineages. Recent work in tropical dry forests of Mexico has also shown that carbon storage may remain strongly associated with a few taxonomic groups when stand structure is less evenly distributed and large trees are less abundant across the community as a whole [42,44].

An important point is that taxonomic dominance did not translate into the same biomass-carbon configuration in all forests. In the wet forests, families such as Sapotaceae and Myristicaceae made substantial contributions to biomass and carbon but were absent from TSDF, whereas Chrysobalanaceae became especially important in TWFn and Fagaceae contributed only in TSDF. This indicates that the taxonomic control of biomass and carbon was forest-specific, even when some broadly distributed families such as Fabaceae and Moraceae retained relevance across more than one system. Similar patterns have been reported in recent tropical forest studies, where differences in biomass distribution among families reflect local combinations of stand structure, species turnover, and environmental conditions rather than a single regional rule [37,45]. These family-level patterns should be interpreted cautiously, since large tropical families are not ecologically uniform groups, and their structural relevance across forests depends on the identity and contribution of the particular species represented in each system.

The close parallel between biomass and carbon allocation was also expected, since carbon estimates were derived directly from biomass pools. However, the biological interpretation remains relevant: the families that contributed most strongly to biomass were also those that sustained the main carbon reservoirs in each forest. In that sense, the results suggest that the contrast among forests was not only one of total stock, but of how taxonomic dominance shaped the partitioning of biomass and carbon within each community. Under the conditions evaluated here, the two wet forests were characterized by broader taxonomic support for biomass and carbon storage, whereas TSDF relied more heavily on fewer dominant families.

4.4. Ecological Correlates of Standing Volume and Carbon Storage

The correlation analysis showed that standing volume and carbon storage were associated more strongly with stand structure than with diversity alone. Basal area was by far the attribute most closely related to standing volume, total biomass, and total carbon, followed by mean height and mean DBH. This pattern is consistent with recent studies showing that forest structural attributes, especially basal area and tree size, are often stronger predictors of biomass and carbon stocks than taxonomic or functional descriptors considered in isolation [39,46]. This interpretation was further supported by the complementary non-parametric analysis, which preserved the same general ranking of ecological correlates and again identified basal area, mean height, and mean DBH as the variables most strongly associated with standing volume, biomass, and carbon storage.

The strong relationship between basal area and the three response variables is ecologically expected because basal area integrates the cumulative horizontal occupation of the stand and reflects the presence of larger stems with greater contribution to woody volume and biomass. Recent work has similarly shown that large trees and structurally developed stands disproportionately influence carbon accumulation, particularly in tropical forests where a relatively small fraction of stems can sustain a large share of total carbon stock [39,47]. Under this interpretation, the higher standing volume and carbon storage of TWFmb and TWFn were primarily associated with their greater structural development rather than with richness alone.

Tree height and mean DBH also showed strong positive associations with standing volume and carbon storage, reinforcing the idea that stand maturity and size hierarchy were key determinants of ecosystem functioning in the three forests. Similar results have been reported in recent tropical forest studies, where mean stem size, canopy development, and vertical stratification explain a substantial proportion of the variation in biomass and carbon stocks across forest types and disturbance contexts [46,48]. In our case, these relationships indicate that the differences among forests were expressed not only in taxonomic composition, but also in the degree to which structural development translated into greater woody accumulation.

Diversity-related metrics were also positively associated with standing volume, biomass, and carbon, although their coefficients were consistently lower than those of structural variables. Richness, Shannon diversity, Simpson diversity, and Hill numbers q1 and q2 all showed significant positive relationships, suggesting that sampling units with greater taxonomic diversity also tended to accumulate more woody volume and carbon. This agrees with recent evidence indicating that species diversity can contribute positively to carbon storage, although its effect often becomes secondary when structural attributes such as basal area, tree height, or the presence of large trees are accounted for simultaneously [47,49]. This association should not be interpreted as direct evidence of causality, since wetter and structurally more developed forests may simultaneously support greater carbon storage and higher taxonomic diversity under more favorable environmental conditions.

Among the ecological attributes evaluated, Pielou’s evenness showed the weakest association with standing volume and carbon storage. Although the relationship remained significant, its lower coefficient suggests that the uniformity of species abundances contributed less to volume and carbon accumulation than structural occupancy and tree dimensions. This result is important because it indicates that not all dimensions of diversity have the same ecological weight in relation to standing volume. Recent studies have reached similar conclusions, showing that different components of diversity may be positively related to carbon storage, but that structural attributes usually retain stronger explanatory power at the stand level [48,49]. The consistency of this pattern across both parametric and non-parametric analyses indicates that the main ecological interpretation was not dependent on a single correlation metric, but reflected a stable relationship between stand structure and the response variables evaluated.

4.5. Scope of Inference, Methodological Constraints, and Ecological Implications

The novelty of this study lies in the use of a common comparative framework to evaluate how stand structure, taxonomic dominance, and biomass distribution are associated with standing volume and carbon storage across contrasting tropical forests of Mexico and Colombia.

The comparative framework adopted in this study allowed consistent evaluation of floristic composition, stand structure, standing volume, biomass, and carbon across three contrasting tropical forests. However, the interpretation of these results should remain within the scope of the sampled sites and the estimation procedures applied. The present study should therefore be interpreted as a comparative evaluation of contrasting forest sites under their current field conditions, rather than as a comparison among stands of equivalent age or developmental phase. The study was based on inventory data from three forests with different environmental settings and used generalized approaches to estimate standing volume, biomass, and carbon. In tropical forest research, the use of generalized allometric equations is common when destructive sampling is not feasible, but recent studies also emphasize that this choice introduces uncertainty because allometric relationships may vary with forest type, species composition, height allometry, and local site conditions [50,51,52].

This is especially relevant in studies that compare humid and seasonally dry forests, where differences in architecture, stem form, and biomass allocation can affect the precision of generic models. Recent work has shown that species-specific or site-calibrated allometric models can reduce uncertainty relative to generalized equations, particularly in tropical forests with strong structural or climatic contrasts [53,54]. The same consideration applies to belowground biomass, which in many comparative studies is derived from fixed root-to-shoot ratios rather than direct measurement, providing a practical but simplified estimate of root biomass and carbon pools [51,53]. An additional source of uncertainty is that the biomass model used in this study did not include species-specific wood density, which may affect estimation accuracy, particularly in forests with contrasting taxonomic composition and stem architecture.

Despite these limitations, the main ecological patterns observed here were internally consistent across independent lines of evidence. The wet forests showed higher basal area, greater standing volume, larger biomass and carbon stocks, and broader taxonomic support of these pools than the seasonally dry forest. In turn, the correlation analysis indicated that basal area, mean height, and mean DBH were more strongly associated with standing volume and carbon storage than diversity metrics. This pattern is consistent with previous studies showing that structural attributes often retain stronger explanatory power than diversity-related variables in tropical forests [39,49]. Under this perspective, the present results are useful not as a basis for broad continental extrapolation, but as a robust comparative assessment of how contrasting tropical forest types differ in the structural and taxonomic organization of standing volume and carbon storage [48].

An additional implication of the study is that differences among forests were not expressed only in total stock, but also in the taxonomic pathways through which those stocks were sustained. In the two wet forests, standing volume and carbon were distributed among a broader set of dominant families, whereas in the seasonally dry forest a larger fraction of biomass and carbon was concentrated in fewer lineages. That distinction is ecologically relevant because it suggests that forest functioning may differ even where some dominant families are shared, and it highlights the importance of evaluating family-level contributions in comparative tropical forest studies. Recent work on tropical forest carbon storage has similarly emphasized that structural dominance and taxonomic composition interact in shaping biomass and carbon distribution across forest types [37,55].

The evidence supports a cautious but clear inference: under the conditions represented by these three forests, structural organization showed a stronger association with standing volume, biomass, and carbon than diversity-related attributes. From a theoretical perspective, this pattern is consistent with ecological interpretations in which dominant structural components exert a stronger influence on ecosystem functioning than diversity alone, while diversity-related attributes act as complementary correlates under contrasting forest conditions. The strength of the study lies in the use of a common comparative framework across contrasting tropical systems in Mexico and Colombia; its main limitation lies in the restricted number of forests evaluated and in the use of generalized estimation models. Accordingly, the results should be interpreted as a site-based comparative contribution to tropical forest ecology rather than as a universal characterization of humid and seasonally dry forests in the Neotropics [50,52].

5. Conclusions

Under the conditions evaluated in this study, the two tropical wet forests showed greater stand development, standing volume, biomass, and carbon stocks than the tropical semi-deciduous forest. These differences were associated primarily with structural attributes, particularly basal area, mean height, and mean DBH, which showed the strongest relationships with standing volume and carbon storage. Although diversity-related variables were also positively associated with these response variables, their contribution was secondary relative to stand structure.

The results also showed that standing volume and carbon accumulation were not supported by the same taxonomic configuration in all forests. In the wet forests, volume, biomass, and carbon were distributed among a broader set of dominant families, whereas in the semi-deciduous forest, these components were concentrated in fewer lineages, especially Euphorbiaceae. This indicates that differences among forests were expressed not only in total magnitude but also in the way structural and taxonomic dominance contributed to ecosystem functioning. These findings should be interpreted within the scope of the sampled sites and the estimation framework used in this study. Because the study was based on inventory data from three contrasting forests and on generalized equations for volume, biomass, and carbon estimation, the results are most useful as a comparative ecological assessment rather than as a basis for broad regional generalization. Even so, the evidence consistently indicates that, in these forests, structural organization was more closely associated with standing volume and carbon stocks than diversity alone. Thus, the hypothesis was only partially supported: although stand structure showed a strong association with standing volume and carbon storage, the more seasonal forest did not compensate sufficiently to reach the levels observed in the two wet forests. Further evaluation across a larger number of sites will be necessary to determine the extent to which the patterns observed here are maintained across broader tropical forest gradients. Accordingly, the results should be interpreted as a site-based comparative contribution rather than as a basis for broad regional generalization.