Exploring the Forest Structure and Floristic Composition Patterns of the Equatorial Chocó, Esmeraldas Province

Study area

The Esmeraldas province was the area selected for a comparison of the floristic ecology and forest diversity in a diverse equatorial forest of Chocó. This province is located on the western Pacific Coast of Ecuador, with habitats ranging from evergreen lowland forests (0-300 m) to evergreen seasonal piedmont forests in the Chocó Coastal Cordillera (200-400 m) (MAE, 2013). The Canandé Reserve, Cayapas, and Loma Linda were selected as sites for the comparative analysis due to their conservation status. The Canandé Reserve, a protected area with 9000 ha of representative Chocó landscape, located in the parish of Quininde, in the Esmeraldas province of northwest Ecuador (Figure 1), was a key component of the study area. The Canandé Reserve belongs to a network of private forest reserves owned by Fundación JOCOTOCO and has been under formal protection and management since 2000. The Cayapas site is located in the western part of Cotacachi Cayapas National Park, mostly below 100 m in elevation. It has been under formal protection and management by the govern- ment since 1968. Finally, the Loma Linda site is located in an area of intact Chocó forest at an elevation of 60 m, just outside the western boundary of Cotacachi Cayapas National Park.

Figure 1: Location of the study area in Ecuador (Equatorial Chocó, Esmeraldas province, and Canandé Reserve), with the distribution of survey plots

The selected study area has a moist, tropical climate with temperatures exceeding 30 °C, an average precipitation of 2500-4000 mm per year, and a relative humidity higher than 70% (MAE, 2013). Further inland, the landscape is dominated by the Perhumid Foothills of the Andes, where the multi-stratified forest varies in height from 25 to 30 m, with emergent trees reaching up to 40 m tall. Ecologically, it exhibits a particular flora composition with elements from the Andean slopes of the coastal cordillera, the Chocó Lowlands, and the Lowland Evergreen Montane Forest of the Western Andes (Cerón, 1999).

Data collection and processing

The fieldwork was conducted using a stratified sampling model. Homogeneous plots were selected based on topographic landscape and vegetation (Mateucci & Colma, 1982). Three 1 ha plots (100 x 100 m) and four additional 50 x 50 m plots were established at different elevations (60-450 m) and spatially distributed throughout the Esmeraldas province in Ecuador. Sampling was conducted at different time scales between 1998 and 2020 (Table 1, Figure 1). All trees with a diameter ≥ 10 cm at breast height (DBH) were recorded. Three to four specimens were collected from each tree, and all the vegetative material was coded, dried, and processed in the field. The plant specimens were incorporated into the ECUAMZ herbarium collection and later transported for taxonomical identification at Pontificia Universidad Católica’s herbarium (QCA; Quito, Ecuador), the herbarium of the Ecuadorian Museum of Natural Sciences (QCNE-INABIO), and the the International Herbarium (K) at the Royal Botanic Gardens, Kew, with herbaria abbreviations according to Thiers (2022). The taxonomic review was performed based on the key treatments of the Flora of Ecuador (Holm-Nielsen, 1986), the Catalogue of Vascular Plants of Ecuador (Jorgensen & León-Yánez, 1999), and the Plant List and Tropicos database from the Missouri Botanical Garden (2016). The endemic species categories were regarded as threatened under IUCN (2022) criteria, using relevant sources such as the Red Book of Endemic Plants of Ecuador (León-Yánez et al., 2011), Magnolias of Ecuador: At Risk of Extinction (Vásquez et al., 2016), and online information from the Bioweb (2022).

Table 1: Distribution of plot surface, coordinates, altitude, and establishment date

* Vegetation map of Ecuador Project (MAATE); ** Palacios, 1998; ***Lozano et al., 2022

Statistical analysis of plant diversity

Forest structure

A matrix was constructed which included the site, family, scientific name, and DBH (cm) for the recorded individuals of each plot, excluding young individuals and vines. To analyze the forest structure, the DBH was used to calculate the basal area and dominance, as follows:

where:

BA = Basal area expressed in square meters

pi = 3.1416

DBH = Diameter at breast height (cm)

The basal area (BA) is the surface area of a cross-section of the stem or trunk of a tree at a given height above the ground, which is calculated using the DBH. Relative dominance is a measure of abundance and was calculated as a percentage of the BA of a given species compared to the total BA of all species:

where:

BA _sp = Basal area of a given species

BA _total = Sum of the basal area of all species

In addition, the importance value index (IVI) was calculated based on two parameters, i.e., the sum of the relative frequency and the relative dominance, in order to assess the relative ecological importance of each species in the study (Mostacedo & Fredericksen, 2000).

Forest diversity

In this work, the Biodiversity R package (Kindt & Coe, 2005), a graphical user interface for statistical diversity analysis, was employed. To measure the alpha diversity, richness was calculated as the total number of species. In addition, Simpson's reciprocal diversity index was determined as follows (Jost, 2006):

where:

The Simpson inverse index is a measure of the alpha diversity present in a given area. The lower the index, the greater the diversity. The abundance ranking curve was used to analyze patterns of diversity in terms of evenness.

Species accumulation curves are used when comparing species richness in the data when the sample sizes of the subsets are different. To evaluate the sampling efficiency, a sample-based species accumulation curve was used, following Kindt and Coe (2005) and applying rarefaction. This was done in order to determine the differences in the mean number of species per site for various random subsets of the data, seeking to account for the unbalanced sample size. Figure 3 shows the cumulative average of the number of species and the total number of sampled plots.

Non-metric multidimensional scaling (NMDS) was used to measure the beta diversity. This type of analysis has proven to be robust for analyzing plant community data (Minchin, 1987; Martin & Goebel, 2011), as well as for handling data collected through discontinuous sampling (Pyke et al., 2001; Stropp, 2011). NMDS determines a spatial arrangement of samples, such that the distances between the samples in the plots best correspond to dissimilarities in species composition (Lepš & Šmilauer, 1999). This was done based on indices that consider abundance, such as the Bray-Curtis index.

Floristic composition and endemism

The abundance per hectare.plot^-1 ranges from 441 to 624 individual trees, and the number of species of trees ranges between 95 and 112 per hectare. Inland, in the ¼ ha plot (Table 2), the abundance varies between 112 and 160 individual trees, and the number of species between 63 and 80 per hectare. The data show significant differences in relation to plot-size. The palm family (Wettinia quinaria, Iriartea deltoidei, and Socratea exorrhiza), as well as the dicotyledon families Moraceae (Perebea xantochyma), Myristicaceae (Virola reidii), Fabaceae (Ecuadendron acosta-solisianum and Swartzia haughtii), Meliaceae (Guarea polymera, Carapa longiptala), and Malvaceae (Matisia grandiflora), among others, exhibit the highest number of individuals and species (Table 3).

Table 2: Synoptic structural and diversity inputs, richness, and Simpson index for the studied plots

Table 3: List of the ten species with the greatest relative ecological importance in the study according to the IVI

Eight species of trees endemic to Ecuador were recorded, out of which three are critically endangered (CR) (Ecuadendron acosta-solisianum, Magnolia canandeana, and Simira standleyi), four are endangered (EN) (Matisia palenquiana, Carapa longipetala, Sorocea sarcocarpa, and Virola aequatorialis), and one is vulnerable (VU) (Inga silanchensis). Within their distribution, most of the Chocoan Ecuadorian species are mainly threatened (Cornejo, 2011) by habitat loss due to logging and land use changes (Lozano et. al., 2022), leaving only isolated populations in the few remnants of vegetation present in the Coastal Equatorial Chocó region (MAATE, 2024). Ten tree species reported as endemic from the Colombian southern Chocó were recorded in this study. Malvaceae has three endemic species: Herrania balaensis (P. Preuss), Matisia Castano (Triana), and Matisia grandiflora (Little). Another seven species were represented by different families, i.e., Boraginaceae: Tournefortia gigantifolia (Killip ex J.S.Mill); Clusiaceae: Clusia dixonii (Little); Fabaceae: Swartzia haughtii (R.S. Cowan); Melastomataceae: Miconia dapsiliflora (Wurdack); Meliaceae: Guarea cartaguenya (Cuatrec. P. Jorg & C. Ulloa); Olacaceae: Hesiteria pacifica (Ulloa); and Violaceae: Gloeospermum grandifolium (Hekking). These findings are representative of the high endemism of the Chocó Bioregion and corroborate the theory of isolated evolution.

Forest structure

According to our analysis of the relative dominance per BA, the most dominant family is Moraceae (20%). Approximately 90% of the dominance stems from 18 families, seven of which account for over 60%, i.e., Moraceae (20%), Meliaceae (8.99%), Myristicaceae (8.22%), Fabaceae (7.98%), Euphorbiaceae (7.18%), Arecaceae (6.94%), and Burseraceae (5.49%). However, Arecaceae is the most abundant family, particularly in the northeastern part of the study area (Loma Linda and Cotachachi Cayapas). Here, 438 individuals of Wettinia quinaria were recorded among more than 2000 registered individuals, accounting for approximately 7% of the dominance (Figure 2). Out of the 17 species contributing over 50% of the dominance, Brosimum utile contributed 11.64% alone and was recorded in four out of the seven plots. It is worth noting that two individuals of this species also exhibited the highest recorded DBH (116.1 cm).

Figure 2: Relative dominance based on the BA of the recorded families in all the plots of the study area

The species accumulation curve showed a linear trend rather than a clear asymptote (Figure 3). This could be due to the high diversity recorded in the study area. The curve suggests that further sampling is needed to fully represent the diversity of the study area, and that there may be new species yet to be discovered.

Figure 3: Accumulation curve of the species sampled in all the plots of the study area

The rank abundance curve indicates a high level of species richness (Figure 4), albeit with a concentration of highly abundant species, with the palm Wettinia quinaria (Arecaceae) being the most abundant. The curve shows that the distribution of individuals per species is not uniform. It is important to note that 76% of the recorded species are represented by singleton or doubleton species, which contribute significantly to the observed diversity.

Figure 4: Rank abundance curve of all the species recorded in the study area

The observed IVI corresponds to 39 out of 360 species, covering 63% of the total index value. Table 3 lists the ten most important species.

The inverse Simpson diversity index values indicated that, at the local level, the Loma Linda and Cotacachi Cayapas plots exhibit high species diversity (Table 2), as evidenced by their high number of individuals and species richness (Table 3). Among the plots within the Canandé Reserve, there was a remarkable similarity in terms of species diversity, with the Ecuadendron plot, in the northernmost part of the reserve, showing the highest species diversity.

The results of the NMDS analysis, depicted in Figure 5, indicate a high beta diversity in the plots, with several shared species and unique differences in terms of composition. Three distinct groups of similar plots were observed, with Loma Linda and Cotacachi Cayapas situated close to each other and belonging to the lowland group (below 100 m in altitude). The second group included MAATE02, MAATE03, and MAATE04 (altitudes between 337 and 388 m), and the third group consisted of the Ecuadendron and MAATE01 plots (400-450 m in altitude).

Figure 5: NMDS ordination of the forest plots in the study area.

The results of the NMDS show an association related to the altitude of the plots, showing proximity of those located in lowlands and on the other hand of those located within the Canandé ecological reserve just near to the Andean foothill, remarking the difference in the floristic diversity of the plots according altitude, also considering the fact that a greater degree of isolation from anthropic pressures can mean a higher state of conservation.

Spatial scale changes in forest structure and diversity

The Chocoan Forest of Ecuador, within the Canandé Reserve and its surrounding forests, showed changes in both abundance and diversity. In 1 ha plots, the abundance ranged from 441 to 624 individuals (≥10 cm) and the species diversity ranged from 101 to 115. Meanwhile, in ¼ ha plots, the species diversity ranged from 63 to 80, and the number of individuals from 112 to 160. The data on species diversity in the Río Palenque Reserve, in the central inland area of the Coastal Chocó Rainforest of Ecuador, show 360 species from all life forms and size classes (Gentry & Dodson, 1987). Gentry’s findings (1986) suggest that lower altitudes have higher abundance and diversity, which is associated with more favorable climatic conditions for plant growth and diversity. In the Canandé 1 ha plot (≥10 cm), a record 39 families, 78 genera, and 102 species was found, which is a higher number of species compared to other Chocoan forests such as Río Palenque in Ecuador (32 species) (Gentry, 1986), La Selva in Costa Rica (43 species) (Gentry, 1995), and Bajo Calima in Colombia (77 species) (Gentry, 1986).

At lower altitudes, monocots are the dominant group of plant families, as observed in the northern Chocó areas (Galeano, 2022). For example, 1177 individuals of this plant group were recorded in the area by Galeano et al. (1998). In our study of the Loma Linda area, the species Wettinia quinaria was found to have the highest density, with a record 438 individuals (IVI=23.79), which is higher compared to other northern Chocoan forests such as Amargal, where 260 individuals of this species were recorded (Galeano et al., 1998). However, other factors such as mountain topography also influence species composition and abundance. For instance, in the plots around Canandé (MAATE 1-4) and the Ecuadendron plot (337-450 m in elevation), we found a higher presence of dicots than in lower-lying zones. In the Ecuadendron plot, the density of W. quinaria was only 13 individuals, while that of Iriartea deltoideia was 25 and that of Bactris gasipaes was 1, suggesting a change in landscape and a decrease in monocot presence at this higher altitude.

The Ecuadorian northern Chocó shows an average DBH range of 20.9-25 cm. Here, the BA varies from 24.6 m² in the Ecuadendron plots to 34.56 m² in the MAATE plots (summed over MAATE 1-4) and 34.3 m² in the Cotacachi Cayapas and Lomalinda plots. Plant species dominance (60%) is mainly found in the following woody families: Moraceae, Meliaceae, Myristicaceae, Fabaceae, Euphorbiaceae, Areacaceae, and Burseraceae. The most dominant species is Brosimum utile, represented by two individuals with DBH ≥ 116 cm. This species seems to be characteristic of this forest type, which has a lower presence of big trees compared to other forests (Gentry & Terborgh, 1990).

The sample-based species accumulation curve suggests that a larger sample size is needed to better understand the species diversity in the study area. Despite this, the high alpha diversity indices found in the Loma Linda and Cotacachi Cayapas plots are consistent with the fact that the lowland regions of Chocó have some of the highest diversity values in the region (Dinerstein et al., 1995). The high beta diversity observed also shows the significant variation within the Equatorial Chocó.

The Map of Ecosystems of Continental Ecuador recognizes nine different ecosystems in this region (MAE, 2013):

1. Evergreen lowland forest (0-300 m in altitude)

2. Evergreen seasonal lowland forest (0-300 m)

3. Flooded hill forest in intertidal zone (0-50 m)

4. Flooded alluvial forest on hillslopes (50-200 m)

5. Flooded lowland herbaceous riverine system (0-300 m)

6. Flooded herbaceous lake of Chocó (50-300 m)

7. Mangrove (0-20 m)

8. Evergreen montane piedmont forest from the Coastal Cordillera of Chocó (>400 m)

9. Evergreen seasonal piedmont forest from the Coastal Cordillera of Chocó (200-400 m)

The Loma Linda and Cotacachi Cayapas plots are located in ecosystem no. 1, while the Canandé plot corresponds to no. 8 and the MAATE plots 1-4 to numbers 2 and 8.

The lowland plots have a distinct composition compared to those at higher altitudes. These differences are primarily due to differences in species composition. Our study provides a general overview of the northern Ecuadorian Chocó, but the altitudinal limits of the MAATE 1-4 and Ecuadendron plots are at the boundary with the geomorphological definition of the Piedmont of the Western Cordillera of the Andes, as per MAE (2013), which is considered a transitional zone. However, the results of the NMDS show great similarity within these plots, in that their diversity is lower but distinct from the lowland plots. These plots exhibit an intermediate or transitional character (ecotone), retaining a Chocó composition with influences from the Andean zone.

The eight endemic trees from Ecuador were all recorded in plots around the Canandé Reserve (Andean foothills between 337 and 450 m in altitude). The other ten that share records for the southern Chocó in Colombia were recorded in the lowland plots (60-85 m).

Our findings highlight the importance of preserving the remaining forests in this area. In this vein, it is crucial to establish new plots along the altitudinal gradient and create a network of continuous research programs focused on biodiversity and ecosystem conservation in the Equatorial Chocó.