Floristic Diversity, Structure, and Carbon Storage of a Sub-Andean Forest in Southwestern Colombia: The Case of the El Mangón

1. Introduction

Sub-Andean forests constitute one of the most biodiverse and threatened ecosystems in the Neotropics. Distributed between 1,000 and 2,400 m a.s.l. in Colombia, they harbor exceptional levels of species richness and endemism, function as biological corridors among altitudinal zones, and provide critical ecosystem services, including hydrological regulation, soil retention, and carbon storage [1,2]. In Colombia, these forests have lost more than 74% of their original cover because of agricultural-frontier expansion, selective timber extraction, and land-use conversion [3]. The most recent national monitoring data indicate that 113,608 ha were deforested in Colombia during 2024 [4], and that in the department of Cauca Forest loss reached 790.34 ha that same year [5], a trend that severely affects connectivity among fragments and alters species composition in sub-Andean landscapes [6]. The Colombian Central Cordillera concentrates an especially significant fraction of this threatened diversity: its sub-Andean forests are dominated by families such as Lauraceae, Melastomataceae, Rubiaceae, and Piperaceae [7,8,9], and represent one of the systems with the greatest floristic-documentation deficit in the country [2,10].

Despite recognition of their ecological importance, knowledge of the floristic diversity and structure of the sub-Andean forests of the Colombian Central Cordillera remains fragmentary. Available studies in the region document between 62 and 431 species per site, with methodological differences that hinder direct comparison [11,12,13,14,15], and only a few simultaneously integrate floristic characterization, quantitative structural analysis, and carbon-storage estimation [16]. This gap is particularly relevant given that montane secondary forests accumulate between 40 and 70% of the carbon stored in comparable mature forests [17,18] and play an increasing role in national climate-change mitigation strategies. The development of functional indices that integrate carbon storage with individual capture efficiency and the stability associated with species spatial-distribution patterns represents a scarcely explored approach in Colombian floristic studies, with direct potential to guide conservation and restoration strategies based on functional efficiency.

The El Mangón forest remnant, located in the corregimiento of Tunía, municipality of Piendamó, department of Cauca (1,600–1,700 m a.s.l.; 24 ha), constitutes a case of special interest for the landscape ecology of fragmented areas of the Central Cordillera. With an uninterrupted history of private conservation since 1912, when the Gómez family decided to protect the forest in a context of growing agricultural pressure, this remnant represents a rare example of a sub-Andean forest with low direct intervention embedded in an intensive agricultural mosaic. The ongoing initiative to declare the property as a Natural Reserve of Civil Society (Reserva Natural de la Sociedad Civil, RNSC) confers immediate relevance on its floristic and functional characterization for private-conservation policy in Colombia. Until now, no scientific study had comprehensively documented its floristic diversity, vegetation structure, or aboveground carbon-storage capacity.

The general objective of this study was to characterize the floristic diversity, vegetation structure, and carbon storage of the El Mangón sub-Andean forest, in order to contribute to the knowledge of the sub-Andean ecosystems of the Colombian Central Cordillera and to provide baseline information for their conservation and management. Specific objectives were: (i) to inventory and analyze the taxonomic composition of the remnant through free collections across the total area (24 ha); (ii) to characterize the horizontal and vertical structure of the forest through systematic transect-based sampling; (iii) to estimate aboveground biomass and carbon storage by species and to interpret the results using the Carbon Capture Efficiency Index (CCEI) and the Carbon Sustainability Index (CSI), which integrate storage magnitude with individual efficiency and spatial stability of species; and (iv) to compare the findings with analogous studies in sub-Andean forests of southwestern Colombia and the Central Cordillera.

2. Materials and Methods

2.1. Study Area

The El Mangón forest remnant is in the corregimiento of Tunía, municipality of Piendamó, department of Cauca, Colombia ((2º40´52.2" N, 76º32´25.5" W), covering an area of 24 ha between 1,600 and 1,700 m a.s.l. The site has a mean annual temperature of 18 °C and an annual precipitation of 2,300 mm. It is crossed by three water bodies (Colcha, Espino, and Quebrada Grande streams) and contains a water spring in its interior (Figure 1). The conservation history dates to 1912, when the Gómez family decided to protect the forest in the face of growing agricultural pressure in the region, an initiative sustained until the present and oriented toward declaring the property as a Natural Reserve of Civil Society (RNSC).

2.2. Study Design and Component Distinction

This study comprises two methodologically differentiated components. Component I (floristic inventory) consisted in documenting total diversity through free collections across the 24 ha, recording 281 species in 209 genera and 99 families (Supplementary Material, Appendix I). Component II (structural and functional analysis) was based on systematic sampling in fixed-area transects: 35 species and 498 individuals in 0.1 ha (five 50 × 4 m transects). All alpha diversity, horizontal and vertical structure, biomass, and carbon analyses correspond exclusively to Component II, unless otherwise indicated.

2.3. Free Floristic Inventory (Component I)

Twelve field expeditions were conducted between June 2024 and June 2025, following the free-collection methodology for floristic inventories [21]. Each expedition involved a team of two to four botanical collectors, with an average duration of eight hours of sampling effort (total effort: ≈96 person-hours). Collections covered the total area of the remnant (24 ha), including previously inaccessible interior forest zones. All groups of vascular plants and bryophytes were collected, with emphasis on epiphytes, lianas, and ferns. Taxonomic identification was performed using: (a) dichotomous keys and monographs of the Flora of Colombia [2,7]; (b) reference collections of the CAUP Herbarium (Universidad del Cauca); (c) virtual collections of GBIF and COL; and (d) nomenclatural verification in World Flora Online (WFO) [40]. Reference specimens were deposited at the CAUP Herbarium under collection numbers Lelopez-4000 to Lelopez-4281.

2.4. Structural Sampling via Transects (Component II)

For the structural analysis, five 50 × 4 m transects (total area = 0.1 ha; sampling intensity = 0.42% of the remnant area) were systematically distributed in the interior of the forest, separated from each other by a minimum of 10 m. In each transect, all arboreal and shrubby individuals with circumference at breast height (CBH) ≥ 5 cm (equivalent to DBH ≥ 1.59 cm) were recorded. For each individual we measured: CBH (measuring tape ± 1 mm), total height (Ht) and stem height (Hf) using a Suunto clinometer, and relative spatial coordinates within the transect.

DBH (cm) = CBH (cm) / π

(1)

Conversion of circumference to diameter at breast height, applied uniformly to all recorded individuals.

To evaluate the completeness of the structural inventory, we performed: (a) individual-based rarefaction curves by transect, standardized to the smallest sampling unit size (n = 52 individuals; transect 1); and (b) asymptotic diversity estimation using Hill numbers (q = 0, 1 and 2) with 95% confidence intervals computed by bootstrap resampling (999 iterations) using the iNEXT package [35] in R [36].

2.5. Alpha Diversity and Horizontal Structure (Component II)

Standard alpha diversity indices were calculated for the 35 species and 498 individuals recorded in the transects:

H′ = −∑(pi · ln pi)

(2)

Shannon–Wiener; pi = proportion of individuals of species i.

DMg = (S − 1) / ln N

(3)

Margalef; S = number of species; N = number of individuals.

1 − D = 1 − ∑(pi²)

(4)

Simpson (complement); pi = proportion of individuals of species i.

J′ = H′ / ln S

(5)

Pielou's evenness; S = number of species.

d = Nmax / N

(6)

Berger–Parker; Nmax = individuals of the most abundant species; N = total individuals.

The Importance Value Index (IVI) was calculated as the sum of relative abundance, relative frequency, and relative dominance of each species, with a theoretical maximum of 300 [21]. The spatial distribution pattern was evaluated using the Pielou aggregation index:

Ga = Do / De

(7)

Ga = aggregation index; Do = observed density; De = expected density under random distribution. Ga < 1: tendency toward dispersion; 1 ≤ Ga ≤ 2: tendency toward clumping; Ga > 2: aggregated pattern.

The mixture coefficient was calculated as CM = (number of species / number of individuals) [7].

2.6. Vertical Structure and Allometric Analysis (Component II)

Vertical structural diversity was evaluated using the Pretzsch Index (A) [34]:

A = −∑(pij · ln pij) / ln(S × Z)

(8)

pij = proportion of individuals of species i in stratum j; S = number of species; Z = number of strata.

Vertical strata were defined at four levels: ground layer (< 1.5 m), shrub layer (1.5–5 m), sub-canopy (5–12 m), and canopy (> 12 m), following [12]. Class width for height and diameter distributions was determined using Scott's criterion [33]:

h = 3.5 · σ · n−1/3

(9)

h = class width; σ = standard deviation; n = number of individuals.

The Ogawa diagram [32] was constructed to visualize height distribution in relation to diameters, and the allometric relationship between Ht and Hf was evaluated using linear regression. Residual normality (Shapiro–Wilk) and homoscedasticity (Breusch–Pagan) assumptions were verified; R² and p-values are reported.

2.7. Biomass and Carbon Estimation (Component II)

Aboveground biomass (AGB) was estimated using the FAO volumetric methodology [26]:

AGB = VCC × WD × BEF

(10)

AGB = aboveground biomass (t·ha−1); VCC = stem volume with bark (m³·ha−1); WD = wood density (t·m−3); BEF = biomass expansion factor.

VCC = BA × Hf × 0.7

(11)

BA = basal area (m²·ha−1); Hf = mean stem height (m); 0.7 = form factor.

WD = 0.6 t·m−3 was used as the regional reference value for tropical Americas [26], and BEF = 1.74. Carbon content was estimated as C = AGB × 0.47 (IPCC convention [27]). Analyses were performed exclusively on Component II and results are expressed in kg·ha−1.

2.7.1. Proposed Carbon Indices (Component II)

To integrate carbon storage with the structural ecological importance of each species, three novel functional indices are proposed: the Carbon Valuation Index (CVI), the Carbon Capture Efficiency Index (CCEI), and the Carbon Sustainability Index (CSI).

CVIi = IVIi × (Ci / Ctotal)

(12)

CVI = Carbon Valuation Index; IVIi = IVI of species i; Ci = carbon stored by species i (kg·ha−1); Ctotal = total carbon in the sampling area (kg·ha−1).

CCEIi = Ci / (IVIi × Di)

(13)

CCEI = Carbon Capture Efficiency Index; Ci = carbon stored by species i (kg·ha−1); IVIi = Importance Value Index of species i; Di = density of species i (ind·ha−1).

For CSI estimation, a spatial factor (Fspatial) was assigned to each species according to its distribution pattern (Table 1):

CSIi = Ci × Fspatial,i × (1 + |H′i|)

(14)

CSIi = Carbon Sustainability Index of species i; Ci = stored carbon (kg·ha−1); Fspatial,i = spatial factor (Table 1); H′i = individual contribution of species i to the community Shannon index (absolute value).

The CSI of group g is calculated as the sum of the individual CSIs of all species in the group: CSIg = ∑CSIi for all species in group g.

2.8. Collection Permits and Ethical Declarations

Biological material collection was carried out under the Framework Collection Permit for Biological Specimens for Non-commercial Scientific Research, Permit 001040 of 30 May 2025, ANLA (Autoridad Nacional de Licencias Ambientales), in accordance with Resolution 1484 of 2014 of the Colombian Ministry of Environment and Sustainable Development (MADS). Specimens were deposited at the CAUP Herbarium of Universidad del Cauca under collection numbers Lelopez-4000 to Lelopez-4281.

2.9. Statistical Analysis

Data were organized in Microsoft Excel 2019. Statistical and ecological analyses were performed in R v4.3.0 [36], with packages vegan v2.6-4 [37], ggplot2 v3.4.0 [38], dplyr v1.1.2 [39], and iNEXT [35]. Hierarchical clustering for carbon functional groups was performed using Ward's method (Euclidean distance). Principal Component Analysis (PCA) was applied to the matrix of carbon, biomass, and structure variables with standardized variables (mean = 0, standard deviation = 1). PC1 explained 64.6% of the total variance and PC2 explained 26.3% (cumulative variance: 90.9%), indicating that both components adequately summarize the structural, ecological, and functional variation among the analyzed species. Comparison of CSI values among spatial-pattern groups was performed using the non-parametric Kruskal–Wallis test, with post hoc pairwise Wilcoxon comparisons and Benjamini–Hochberg correction. Statistical significance was set at α = 0.05.

3. Results

Results are presented in two differentiated blocks reflecting the two methodological components of the study (see Section 2.2). Component I corresponds to the free floristic inventory (24 ha): 281 species, 209 genera, 99 families. Component II corresponds to the structural sampling in five 50 × 4 m transects (0.1 ha): 35 species and 498 individuals. Unless otherwise indicated, all numerical values for IVI, diversity indices, biomass, and carbon refer to Component II.

3.1. Taxonomic Diversity and Floristic Richness (Component I — 24 ha)

A total of 281 species distributed in 209 genera and 99 families were recorded in the El Mangón forest remnant (Supplementary Material, Appendix I). The species/genus ratio (1.34), species/family ratio (2.84), and genus/family ratio (2.11) indicate a balanced taxonomic structure. The most diverse families were Poaceae (20 spp.), Asteraceae (13 spp.), and Orchidaceae (13 spp.) — 16.4% of total richness — which together with Polypodiaceae and Rubiaceae (10 spp. each), Piperaceae (9 spp.), Fabaceae and Melastomataceae (8 spp. each), and Bromeliaceae and Bryaceae (7 spp. each), accumulated 38.4% of species diversity (Table 2). The class distribution showed a predominance of Magnoliopsida (117 spp., 41.64%), followed by Liliopsida (53 spp., 18.86%) and Bryopsida (48 spp., 17.08%). Lecanoromycetes and Polypodiopsida contributed 24 spp. (8.54%) and 25 spp. (8.90%), respectively.

3.1.1. Growth Forms — Free Floristic Inventory (Component I)

Analysis of growth forms showed that epiphytes constituted the most diverse group with 124 spp. (44.13%), followed by herbs with 97 spp. (34.52%), shrubs with 31 spp. (11.03%), and trees with 25 spp. (8.90%). Hemiparasites and lianas were the least represented, with 2 spp. each (0.71%). The high proportion of epiphytes exceeded the typical values of 25–35% reported for Neotropical sub-Andean forests.

3.2. Alpha Diversity Indices (Component II — Transects, 35 spp., 498 ind.)

Species diversity analysis of Component II revealed a community of 35 species and 498 individuals. The Shannon–Wiener index (H′ = 2.55) indicated intermediate diversity characteristic of forests in intermediate succession; the Margalef index (DMg = 5.47) evidenced high species richness relative to sample size. Community structure exhibited low dominance (Simpson 1−D = 0.89; Pielou J′ = 0.72; Berger–Parker d = 0.17), confirming a relatively homogeneous distribution of abundances among the 35 sampled species (Figure 2).

3.2.1. Structural Sampling Completeness: Rarefaction and Asymptotic Estimation

Individual-based rarefaction analysis revealed differences in standardized richness among the five transects. Standardizing to n = 52 individuals (smallest transect size: transect 1), rarefied richness ranged from 10.85 spp. (transect 2; SE = 1.35) to 15.00 spp. (transect 1; reference), with intermediate values for transect 3 (13.46 ± 1.51), transect 5 (13.11 ± 1.40), and transect 4 (11.31 ± 1.19) (Figure 3).

Asymptotic extrapolation curves (Hill numbers, q = 0) showed that none of the five transects reached inventory saturation at the individual sampling unit scale: estimated asymptotic richness exceeded observed richness in all cases. Asymptotic estimates ranged from 20.1 spp. (transect 3; observed: 18; 95% CI: 18.0–34.1) to 32.9 spp. (transect 2; observed: 15; 95% CI: 15.0–58.7). Transect 3 was closest to saturation (relative completeness ≈ 90%), while transect 1 and transect 2 showed the largest gaps (completeness ≈ 49% and ≈ 46%, respectively; Table 3, Figure 4). These results confirm that sampling by individual transect is incomplete; the five transects combined allow adequate estimation of community patterns for the objectives of the present study.

3.3. Horizontal Structure of the Forest Remnant (Component II)

Analysis of horizontal structure revealed marked heterogeneity in the ecological importance of species, with IVI values ranging from 1.460 to 45.124 (scale 0–300). P. crocea (Rubiaceae) was the ecologically dominant species (IVI = 45.124; relative abundance = 45.12%; relative dominance = 22.21%), followed by Lacistema aggregatum (IVI = 22.181) and Heliconia griggsiana (IVI = 19.601). The fifteen most important species concentrated approximately 80% of total IVI, with the top ten accumulating 66.4% (Figure 5). Variation in stem size by species is summarized in Figure 6, where Myrcianthes hallii, Myrcia popayanensis, and Palicourea heterochroma displayed the largest median DBH values, while shrubby and palm species (Geonoma pinnatifrons, Coffea arabica, Citrus reticulata) presented the smallest size classes.

Of the species, 57.14% (20 spp.) exhibited Ga < 1 (low-aggregation distribution), while 25.71% (9 spp.) showed an aggregated pattern (Ga > 2). Olmedia aspera showed the highest degree of aggregation (Ga = 15.28), followed by P. crocea (Ga = 3.60) and G. pinnatifrons (Ga = 2.82). The mixture coefficient (CM = 0.070; 1/CM = 14.23) classified the ecosystem as structurally homogeneous.

3.4. Vertical Structure of the Forest Remnant (Component II)

The vertical structure revealed four clearly differentiated strata: lower stratum (0–8 m), middle stratum (8.1–12 m), upper stratum (12.1–18 m), and emergent stratum (18.1–33 m). The Pretzsch Index reached A = 3.21, equivalent to 64.93% of the theoretical maximum (4.94), indicating notable vertical structural complexity for a forest in active regeneration.

The Ogawa diagram (Figure 7) showed an asymmetric height distribution, with a strong concentration of individuals in lower and intermediate classes and a small number of dominant canopy trees. Mean height was 4.34 m (± 4.09 m). The allometric relationship between Ht and Hf was moderate but significant (R² = 0.506, p < 0.001). The lower stratum concentrated 91.77% of individuals (457 ind. of 32 spp.); the middle stratum 5.62% (28 ind. of 12 spp.); and the upper and emergent strata 2.61% (13 ind. of 8 spp.). The species-level distribution of total height (Figure 8) reveals that the tallest medians correspond to canopy trees such as Alchornea latifolia, Cinchona pubescens, and M. popayanensis, while the bulk of the assemblage exhibits modal heights below 5 m, consistent with an inverse-J pattern typical of secondary forests in active regeneration.

Volumetric calculations yielded 5.49 m³·ha−1 of stem (commercial) volume and 24.00 m³·ha−1 of total volume with bark, with P. crocea contributing the greatest volume (1.71 m³·ha−1 commercial; 6.70 m³·ha−1 total).

3.5. Biomass and Carbon of the Remnant (Component II)

Total aboveground carbon storage (Component II, 0.1 ha) was 2,866.23 kg·ha−1. P. crocea (Rubiaceae) showed the highest Carbon Valuation Index (CVI = 14.08) and the highest storage per hectare (894.41 kg·ha−1, equivalent to 31.2% of total carbon), followed by H. griggsiana (Heliconiaceae; CVI = 3.69) and Myrsine guianensis (Primulaceae; CVI = 1.63) (Figure 9). However, when efficiency rather than absolute storage is considered, C. arabica (Rubiaceae, introduced species of African origin) exhibited the highest Carbon Capture Efficiency Index (CCEI = 12.46), together with Myrsine coriacea (Primulaceae) and Ocotea oblonga (Lauraceae) (Figure 10).

Note on C. arabica: this cultivated species, native to Ethiopia and introduced to Colombia during the colonial period [23], occurs in the transects as remnant individuals from a coffee plantation with more than 50 years of history at the site, currently abandoned. C. arabica is therefore not a native component of the Colombian sub-Andean ecosystem. Its high CCEI value reflects specific physiological characteristics (photosynthetic efficiency under shade and high C/N ratio) rather than a central ecological function in the natural ecosystem; ecological implications are discussed in Section 4.5.

Functional-group analysis identified four clusters through hierarchical clustering: Group 1 (4 spp.) with mean storage of 343.56 kg·ha−1; Group 2 (represented by P. crocea) with the highest storage (894.41 kg·ha−1) and highest density (166 ind.); Group 3 (5 spp.) with low storage (25.29 kg·ha−1) but high mean density (121.6 ind.); and Group 4 with intermediate storage and density. PCA showed that PC1 (64.6% of variance) was influenced primarily by IVI, Simpson index, density, and relative abundance, while PC2 (26.3% of variance; cumulative variance: 90.9%) was associated with relative dominance, biomass, and carbon (Figure 11).

3.6. Carbon Sustainability Index (CSI) by Spatial Pattern (Component II)

Evaluation of CSI by spatial pattern revealed differences in cumulative values among distribution groups (Figure 12). The cumulative CSI of the low-aggregation group (20 spp.) was 1,618.24, dominated by H. griggsiana (560.15), M. coriacea (345.15), and C. arabica (230.84). The aggregated-pattern group (9 spp.) showed a cumulative CSI of 811.17, with P. crocea as the main contributor (580.75).

The Kruskal–Wallis test did not detect statistically significant differences in individual CSI values among the three spatial-pattern groups (H = 4.8579, df = 2, p = 0.088). Post hoc pairwise comparisons (Wilcoxon with Benjamini–Hochberg correction) confirmed that none of the group pairs differed significantly (low aggregation vs. aggregated: p = 0.21; low aggregation vs. tendency toward clumping: p = 0.12; aggregated vs. tendency toward clumping: p = 0.61). The higher cumulative CSI of the low-aggregation group reflects, in part, the larger number of species in that group (20 spp.) compared with the aggregated group (9 spp.); interpretation of cumulative totals must therefore be made with caution. The mean CSI per species was: 80.91 spp.−1 (low aggregation) and 90.13 spp.−1 (aggregated).

4. Discussion

4.1. Taxonomic Diversity and Floristic Richness

The floristic richness of the El Mangón remnant (281 spp., 209 gen., 99 fam.) is comparable to that reported for El Peñol, Antioquia (285 spp.; [1]) and surpasses records from other Cauca remnants such as Timbío, Hacienda Hato Viejo (151 spp.; [11]) and Popayán, Reserva Forestal Cajete (164 spp.; [3]). This level of richness confirms the patterns of high diversity characteristic of the sub-Andean forests of the Colombian Central Cordillera described by [7,9].

The predominance of Poaceae (20 spp.), Asteraceae (13 spp.), and Orchidaceae (13 spp.) partially departs from the general pattern of Colombian sub-Andean forests, where Rubiaceae and Melastomataceae typically dominate floristic composition [7,9]. This peculiarity may be explained by the specific ecological conditions of the site — especially the surrounding agricultural matrix and the disturbance history — which favour the establishment of grasses and composites typical of forest edges and intermediate successional stages [8]. The dominance of Rubiaceae in the structural analyses (driven mainly by P. crocea) indicates that the family is indeed important at the structural level, although not in terms of species diversity in the free inventory.

The high representation of epiphytes (44.13%) greatly exceeds the typical values of 25–35% reported for Neotropical sub-Andean forests [31]. This exceptional epiphytic richness can be attributed to: the permanence of forest cover since 1912 without clear-cutting, the presence of four well-differentiated vertical strata that increase microhabitat availability, and microclimate stability provided by the three water bodies crossing the remnant. The association between canopy vertical structural heterogeneity and epiphytic diversity has been documented in tropical Andean montane forests [10,30].

4.2. Diversity Indices and Community Structure

The Shannon–Wiener index (H′ = 2.55) places El Mangón in an intermediate range relative to other Colombian sub-Andean forests, below values reported for Puracé (H′ = 3.0; [12]) and Santander de Quilichao (H′ = 3.0; [13]), but above records from Timbío (H′ = 2.0; [14]) and Nariño (H′ = 2.0; [6]). Pielou's evenness (J′ = 0.72) indicates a moderately uniform abundance distribution, consistent with an ecosystem in successional reorganization.

Patterns of intermediate diversity and low dominance (Simpson 1−D = 0.89; Berger–Parker d = 0.17) are consistent with post-disturbance recovery dynamics described by [15] for tropical Andean forests in intermediate succession. The high Margalef index (DMg = 5.47) reflects high species richness relative to the structural sample size (0.1 ha).

4.3. Horizontal Structure and Spatial Distribution Patterns

The marked dominance of P. crocea (IVI = 45.124) constitutes an atypical pattern for sub-Andean forests. The concentration of 66.4% of total IVI in ten species coincides with patterns recorded in fragmented sub-Andean remnants, where the first twelve species typically accumulate ≈72% of IVI [16,17]. This pattern may be partly attributable to the placement of transects in areas of higher density of P. crocea (Ga = 3.60, aggregated pattern), an aspect that constitutes a methodological limitation discussed in Section 4.6.

The predominance of low-aggregation patterns (57.14% of species with Ga < 1) contrasts with the predominantly aggregated pattern of mature tropical forests and suggests active interspecific competition and seed-dispersal limitations, processes typical of intermediate successional stages [15,18]. The high aggregation of O. aspera (Ga = 15.28) and P. crocea (Ga = 3.60) indicates specific reproductive strategies and marked microenvironmental preferences.

4.4. Vertical Structure and Successional Position

The presence of four well-differentiated vertical strata and a Pretzsch Index of 3.21 (64.93% of the theoretical maximum) indicates notable vertical structural complexity for a forest in active regeneration. This value is consistent with the three to four well-defined strata reported for mature secondary forests of the Colombian Andes in advanced passive restoration [3,19], where gradual canopy differentiation generates light gradients that diversify microhabitats for recovering species.

The concentration of 91.77% of individuals in the lower stratum (0–8 m) reflects typical dynamics of recovering forests, where active regeneration generates high densities of young individuals [15,20]. The inverse-J pattern in diameter and height distributions is characteristic of communities with continuous recruitment and density-dependent mortality, suggesting a successional trajectory toward a forest with greater canopy dominance. The Ht–Hf allometric relationship (R² = 0.51, p < 0.001) indicates the architectural heterogeneity expected in secondary forests with high diversity of growth forms [20].

4.5. Carbon Storage and Sustainability

Carbon-storage patterns revealed marked functional heterogeneity among species. P. crocea showed the highest absolute storage (894.41 kg·ha−1, 31.2% of total: 2,866.23 kg·ha−1), while species with lower structural dominance exhibited higher per-individual capture efficiency (CCEI). Total storage is consistent with values reported for young secondary forests: recent studies show that secondary forests 25–30 years old accumulate between 40 and 70% of the carbon stored in comparable mature forests [17,18].

The presence of C. arabica (L.) among the species with the highest CCEI (12.46) requires careful interpretation. The presence of this species in the transects corresponds to remnant individuals from a coffee plantation with more than 50 years of history at the site, currently abandoned. C. arabica, native to the Ethiopian highlands and introduced to Colombia during the colonial period [23], is not a native component of the Colombian sub-Andean ecosystem. Its maintenance under the canopy of the remnant — without reproducing or expanding as part of the native assemblage — and its high CCEI may reflect specific physiological characteristics (photosynthetic efficiency under shade conditions; high C/N ratio) rather than a central ecological function in ecosystem carbon storage. This distinction is relevant both for the interpretation of the carbon analysis and for site management within the framework of the RNSC declaration, where the status of introduced species must be formally evaluated.

Functional-group analyses and PCA revealed clear trade-offs between diversity and carbon storage: the group dominated by P. crocea maximizes total storage but reduces structural diversity, while groups with a greater number of species generate greater functional heterogeneity. CSI results are consistent with the hypothesis that low-aggregation distribution patterns — by involving more species and a more homogeneous carbon distribution — may favour the long-term stability of stored carbon [24,25]. However, the Kruskal–Wallis test did not detect statistically significant differences in individual CSI values among groups (p = 0.088); this interpretation must therefore be considered tentative given the limited sampling scale (0.1 ha).

4.6. Study Limitations

This study presents limitations that must be considered in the interpretation of results. First, the structural sampling intensity (0.1 ha out of 24 ha, equivalent to 0.42%) is low for a forest with the reported spatial heterogeneity. The high abundance of P. crocea (IVI = 45.124) may partly reflect a bias of the transects toward areas of higher density of this species. Future studies with greater sampling intensity or stratified sampling designs will allow evaluation of the representativeness of the structural analysis.

Second, biomass estimation with a uniform WD value (0.6 t·m−3) may introduce biases in the comparative estimation of carbon by species; use of the Global Wood Density Database [27] and updated allometric models such as Chave et al. [28] would reduce this uncertainty. Third, the rarefaction analysis indicated that sampling by individual transect was incomplete (relative completeness between ≈46% and ≈90%), reinforcing the need to expand the structural inventory. Finally, the CVI, CCEI, and CSI indices proposed here are novel tools that require comparative validation at additional sites before generalising their applicability as indicators of forest carbon sustainability.

5. Conclusions

The El Mangón forest remnant represents one of the floristically richest sub-Andean fragments documented in the department of Cauca and in southwestern Colombia, with a species diversity that reflects the conservation potential of secondary forests with a prolonged history of private protection. The composition of the free inventory — with predominance of epiphytes (44.13%) and significant cryptogamic contributions — evidences the high degree of microclimate maturity and structural complexity accumulated over decades of passive restoration, positioning the site as a functional biodiversity refuge in an intensive agricultural landscape.

Structural analysis reveals a forest in an intermediate–advanced successional stage, with four well-differentiated vertical strata, high vertical structural diversity (Pretzsch Index = 3.21), and an inverse-J pattern in height and diameter distributions that confirms the viability of the active regeneration process. The structural dominance of P. crocea and the predominance of low-aggregation spatial-distribution patterns are indicators of a community in active competitive reorganization, compatible with a trajectory toward stages of greater forest complexity.

The proposed carbon indices (CVI, CCEI, and CSI) constitute complementary tools to standard floristic and structural analysis by integrating the magnitude of carbon storage with individual species efficiency and the stability associated with their spatial distribution patterns. Results are consistent with the hypothesis that spatial heterogeneity of the tree community may contribute to the long-term stability of stored carbon, although differences among spatial groups were not statistically significant (p = 0.088); this hypothesis requires validation with greater sampling intensity and temporal monitoring.

The potential declaration of the property as a Natural Reserve of Civil Society has a solid ecological basis: the documented floristic richness, the complex vertical structure, the carbon-storage potential, and the uninterrupted history of private conservation justify inclusion of the remnant in national biodiversity-protection instruments. Technical and legal support for the Corporación Autónoma Regional del Cauca is recommended for the declaration process as a landscape-conservation strategy in the municipality of Piendamó.

This contribution helps fill a floristic gap in the Colombian Central Cordillera and provides a quantitative baseline for long-term monitoring. For future research, we recommend: (i) increasing the intensity of structural sampling through stratified designs; (ii) updating the biomass estimation model with species-specific wood densities; (iii) formally documenting the status of introduced species present in the remnant; and (iv) evaluating the connectivity potential of the remnant with other forest fragments in the Río Piendamó corridor.