Introduction

Studies on arboreal vegetation in semi-evergreen and semideciduous tropical forests have mainly focused on determining their structure and diversity (^{Carreón-Santos & Valdez-Hernández, 2014}; ^{Villavicencio-Enríquez & Valdez-Hernández, 2003}; ^{Zarco-Espinosa, Valdez-Hernández, Ángeles-Pérez, & Castillo-Acosta, 2010}). In order to better understand the spatial distribution and heterogeneity of the arboreal vegetation in tropical forests, it is important to consider the influence of relief on the establishment and growth of vegetation (^{Condit et al., 2002}; ^{Valencia et al., 2004}). Some authors such as ^{Hernández, Ortiz, and Figueroa (2009)} classify reliefs based on different elevational ranges and morphological and morphometric relationships in the territorial context.

In Mexico, studies related to the variation in arboreal structure and diversity in response to relief are limited, as is the case of the tropical dry forests in the state of Guerrero (^{Galicia, Zarco-Arista, Mendoza-Robles, Palacio-Prieto, & García-Romero, 2008}) and the northern Sierra Madre de Chiapas (^{López-Pérez, Castillo-Acosta, Zavala-Cruz, & Hernández-Trejo, 2014}). In the particular case of the coastal region of Oaxaca, only floristic lists (^{Salas-Morales, Saynes-Vázquez, & Schibli, 2003}; ^{Salas-Morales, Schibli, Nava-Zafra, & Saynes-Vásquez, 2007}) and a comparative case study between the seasonally dry tropical forests of southern Honduras and Oaxaca (^{Gordon, Hawthorne, Reyes-García, Sandoval, & Barrance, 2004}) are available. Therefore, in this work the composition, vertical stratification and diameter distribution were described, by estimating the tree structural and diversity indexes by size classes, with the aim of contributing to the generation of knowledge concerning arboreal vegetation in three reliefs on the Oaxaca coast.

Materials and methods

The study area is located in the Pacific coastal plains and mountains of the municipality of Villa de Tututepec de Melchor Ocampo, Oaxaca, between coordinates 16° 02’ 43” N, 97° 38’ 48” W; 16° 11’ 18” N, 97° 29’ 17” W and 16° 05’ 58” N, 97° 24’ 04” W. The climate is warm sub-humid with average annual temperature and precipitation of 27 °C and 1,365 mm, respectively (^{Serrano-Altamirano, Silva-Serna, Cano-García, Medina-García, & Ruiz-Corral, 2005}). The predominant soils are Acrisol, Cambisol, Phaeozem, Gleysol and Regosol (^{Sánchez-Bernal, Camacho-Escobar, Rodríguez-León, & Ortega-Escobar, 2012}). The study vegetation was classified as Bravaisia integerrima forest (with flooded soils), a semi-evergreen tropical forest (^{Rzedowski, 2006}) with harvesting of the fruit of Attalea cohune Mart. until the year 1980, and a coffee agroforestry system (coffee AFS) (^{García, Valdez, Luna, & López, 2015}) since 1976. The plain (Pl), premountain (Pm) and low mountain (Lm) reliefs were considered on the basis of the characterization of the different elevational ranges and on the morphological and morphometric relationships of the territorial context of basins, intramontane valleys and coastal basins of Oaxaca (^{Hernández et al., 2009}).

Three 20 x 50 m (1,000 m²) sampling units (SU) were randomly located for the plain (0-13 m), premountain (690-726 m) and low mountain (810-846 m) reliefs. Each SU was divided into 10, 10 x 10 m (100 m²) subunits in which two 4 x 4 m (16 m²) squares and five 1 x 1 m (1 m²) ones were established, in accordance with the methodology of ^{Villavicencio-Enríquez and Valdez-Hernández (2003)}. Three size classes were surveyed in each SU: a) poles (Po > 2.5 cm in diameter at breast height [DBH) measured at 1.3 m from the ground) were recorded in the 100 m² subunits, b) saplings (Sa < 2.5 cm in DBH but > 1.5 m in height) were recorded in those of 16 m², and c) seedlings (Se < 1.5 m but > 0.3 m in height) were measured in those of 1 m² (^{Interián-Ku et al., 2009}). Diameter at breast height, total height, and major and minor crown diameter were recorded. The arboreal species were collected in triplicate, botanized, collated and deposited in the CHAPA herbarium at the Colegio de Postgraduados, Montecillo Campus. At the same time the plant database in Tropicos^® (^{Missouri Botanical Garden}), available online (http://www.tropicos.org/), as well as in ^{Pennington and Sarukhán (2005}), was consulted.

The number of species per relief and size class, with respect to the sampled area, was plotted to obtain the species-area curve (^{García et al., 2015}). The forest mensuration attributes were analyzed considering the delimitation of the vertical strata by relief for the Po size class, from the inflections of the curves generated between the number of individuals and their height (^{López-Toledo, Valdez-Hernández, Pérez-Farrera, & Cetina-Alcalá, 2012}). The diameter distribution per relief was organized by diameter category (cat), "n" classes of 10 cm of amplitude for Po: 2.5-12.5 (cat 10), 12.6-22.5 (cat 20) to 92.6-102.5 (cat 100) (^{Zarco-Espinosa et al., 2010}) and less than 2.5 cm (cat < 10) for Se and Sa.

In order to rank the ecological importance of species per relief and size class, two structural valuation indices were obtained: importance value index (IVI) and forestry value index (FVI) (^{Corella et al., 2001}):

IVI = relative dominance + relative density + relative frequency

FVI = relative DBH + relative height + relative cover

The diversity, equity and similarity of tree species per relief and size class were determined as follows:

Shannon-Wiener Index (H’): Measures the degree of uncertainty associated with the random selection of an individual in the community (^{Pla, 2006}).

The “t” method modified by Hutchenson (^{Magurran, 2004}) was applied to determine the existence of significant differences (P < 0.05) between the H´ diversity for the size classes per relief and reliefs per size class.

Equity index (E): To the extent that this index is higher means that different species tend to have the same number of individuals (^{Rodríguez, 2014}).

Sorensen coefficient (SI): Relates the number of species in common with the arithmetic mean of the species in both sites (^{Magurran, 2004}).

In the case of richness, H´ and equity, analysis of variance and comparison of means tests (Tukey, P < 0.05) were performed between reliefs per size class and between size classes per relief using the SAS statistical package (^{Statistical Analysis System [SAS], 2004}).

Results and discussion

Figure 1 shows the cumulative number of species per relief and size class on the Oaxaca coast. The cumulative number of species in the seedling size class was higher in Lm, whereas in the pole class it was higher in Pm; the cumulative number of species per relief was higher in Lm, followed by Pm and Pl. This is similar to what was reported in a coffee AFS in Veracruz, where the number of species was similar to that of Lm (^{García et al., 2015}), and also suggests a possible trend in the floristic composition of the repopulation as a result of the management (^{Carreón-Santos & Valdez-Hernández, 2014}) in the coffee AFS.

Figure 1 Species-area curves by relief and size class on the Oaxaca coast. 

Figure 2 shows the vertical stratification of poles. The three reliefs presented two strata (lower and upper); in Lm, the highest number of individuals (81 %) had heights of less than 21.0 m, while individuals of Pl and Pm (62 and 58 %, respectively) were concentrated at heights greater than 6.5 m. Something similar was reported by ^{Vázquez-Negrín, Castillo-Acosta, Valdez-Hernández, Zavala-Cruz, and Martínez-Sánchez (2011)} and ^{Zarco-Espinosa et al. (2010)} regarding the presence of two strata in Tabasco forests. The upper stratum in Pl was represented by Bravaisia integerrima (Spreng.) Standl (523 individuals∙ha^-1) with heights greater than 6.5 m (62 % of total individuals) as a result of good repopulation by regrowth in flooded areas; this species had a high importance value in the three size classes. This situation also occurs in the forests of Tacotalpa, Tabasco (^{Maldonado-Sánchez & Maldonado-Mares, 2010}), and Huatulco, Oaxaca (^{Salas-Morales et al., 2007}). It is important to mention that the species not only has a very restricted habitat, but is also included in NOM-059 with threatened status (^{Secretaría del Medio Ambiente y Recursos Naturales [SEMARNAT], 2010}).

Figure 2 Vertical stratification of poles in three reliefs on the Oaxaca coast. 

Figure 3 shows the diameter distribution of trees by size class and relief. The diameter distribution was concentrated in the cat < 10 (Se and Sa) and cat > 10 (Po); the cumulative percentage of individuals in Pl (75 %), Pm (70 %) and Lm (73 %) was similar for both categories, as reported in forests of Quintana Roo (^{Carreón-Santos & Valdez-Hernández, 2014}), Tabasco and Chiapas (^{Vázquez-Negrín et al., 2011}), where they suggest an inverted J-shaped distribution. The presence of some disturbances promotes the establishment of the repopulation (^{Rodríguez, 2014}; ^{Smith & Smith, 2004}), as observed in Pl with soils constantly saturated with water, resulting in the phenomenon of policaulescence (^{Gallardo, Meave, & Pérez, 2005}) and generating a high regeneration frequency by resprouting; in the case of Pm, the repopulation was due to the opening of clearings due to the fall of trees and branches, whereas in Lm it was due to weed control activities for the coffee AFS. ^{Martínez (1994)} named these regeneration processes repopulation routes.

Figure 3 Diametric distribution of trees by size class (< 10 cm = seedlings and saplings; > 10 cm = poles) and reliefs on the Oaxaca coast. 

Tables 1, 2 and 3 report the species with the highest IVI and FVI by size class in each relief. The species with the greatest structural importance in Pl in the three size classes were B. integerrima, Pithecellobium dulce (Roxb.) Benth and Nectandra ambigens (S. F. Blake), considering this relief as having semideciduous tropical forest vegetation; however, because Pl is found in low-lying areas and flooded soils, and in the presence of B. integerrima as a dominant species, ^{Pennington and Sarukhán (2005)} consider that the vegetation type is B. integerrima forest (canacoital). For the case of Pm, the species with the highest structural value were F. occidentalis, which is present in early serial conditions (^{Maldonado-Sánchez & Maldonado-Mares, 2010}) and M. zapota, which can be found in both early (^{Carreón-Santos & Valdez-Hernández, 2014}) and late serial conditions (^{Sánchez, Islebe, & Valdez-Hernández, 2007}). It has been reported that Faramea occidentalis (L.) A. Rich., Brosimum alicastrum Sw., Manilkara zapota (L.) P. Royen, Guarea glabra Vahl and Diospyros sp. are part of the semi-evergreen tropical forests of Veracruz (^{Godínez-Ibarra & López-Mata, 2002}) and Chiapas (^{López-Pérez et al., 2014}; ^{Soto-Pinto, Romero-Alvarado, Caballero-Nieto, & Segura, 2001}). In Lm, the species with the highest values in the structural index were Swartzia cubensis (Britton & P. Wilson) Standl, Nectandra globosa (Aubl.) Mez and Guarea glabra Vahl, species that are part of the medium-elevation semi-evergreen forests, and Cecropia obtusifolia Bertol., Dendropanax arboreus (L.) Decne. & Planch. and Cupania dentata DC, which are early serials, indicative of structural changes (^{Niembro, Vázquez, & Sánchez, 2010}; ^{Villavicencio-Enríquez & Valdez-Hernández, 2003}).

Differences in species diversity were significant (P < 0.05) between Po, Sa and Se for low mountain, pre-mountain and plain (Table 4). This is supported by the results obtained in the Sorensen coefficient values, where Po, Sa and Se shared on average less than 20 % of the species among the three reliefs (Table 5). Also, the total value of the Shannon index (H´) for low mountain (H’ = 3.24) was higher than for plain (H’ = 2.07) and premountain (H’ = 1.76), just as in Atoyac, Veracruz (H´ = 2.47; ^{García et al., 2015}) and in Hampolol, Campeche (H´ = 2.28; ^{Gutiérrez, Zamora-Crescencio, & Puc-Garrido, 2013}), but it was similar to that in San Miguel (H´ = 3.16; ^{Villavicencio-Enríquez & Valdez-Hernández, 2003}) and Vega de Alatorre, Veracruz (H´ = 3.3; ^{Godínez-Ibarra & López Mata, 2002}). This confirms that coffee AFS have potential for the conservation of tree species diversity (^{García et al., 2015}). This may be due to the low impact or traditional management of coffee that helps to preserve the structure and diversity of the species (^{Manson, Contreras, & López-Barrera, 2008}). Similarly, the establishment of arboreal species of commercial interest is promoted (^{Bandeira, Martorell, Meave, & Caballero, 2005}; ^{Soto-Pinto et al., 2001}).

Conclusions

The tree species with the highest structural importance values were different in two reliefs for the three size classes: in plain, Bravaisia integerrima for poles and Pithecellobium dulce for saplings and seedlings; in low mountain, Cecropia obtusifolia and Swartzia cubensis for poles, Cupania dentata for saplings and Nectandra globosa for seedlings. The exception occurred in the premountain relief, where Faramea occidentalis was the most important species in the three size classes. The richness and diversity of tree species were different in each size class in at least one relief against the other two. The exception arose in species richness in saplings that was equal in the plain, premountain and low mountain reliefs. The seedling size class obtained the highest richness and diversity values in plain and low mountain. The richness and diversity of tree species was different within each relief in at least one size class against the other two. The exception was species richness in premountain that was equal among poles, saplings and seedlings. The low mountain was the relief that showed the highest richness and diversity values in the three size classes.