Floristic Differentiation of a Deciduous Seasonal Forest Tree Stratum, Jaguari, RS, Brazil

This study aims at determining and differentiating the floristic groups of a Deciduous Seasonal Forest’s arboreal component, located on the ridge of the Southern Plateau, in Southern Brazil. Individuals with diameter at breast height ≥ 5.0 cm were sampled in sixty-two plots measuring 10 m × 10 m, which were systematically installed in the forest. Three floristic groups were found: Middle Stage, including 54 species, with Casearia sylvestris as an indicator species; Advanced Stage, with 38 species and Pilocarpus pennatifolius as the indicator species; and Altered Forest, with 27 species and Apuleia leiocarpa, Helietta apiculata, and Machaerium paraguariense as indicator species. A higher proportion of climax light-demanding individuals was reported in the Middle Stage and Altered Forest groups, contrasting with the Advanced Stage group, in which climax shade-tolerant species were predominant. In addition, the groups were differentiated according to their dispersion strategy, with specific syndromes occurring in each group (Middle Stage: zoochory; Advanced Stage: autochory; Altered Forest: anemochory).


INTRODUCTION
The replacement of forests by other forms of land occupation, such as agricultural crops and pastures, as well as the disorderly exploitation of timber and non-timber products, has diminished the natural resources available for use by the human population. The environmental functions of forests have also been adversely affected, such as maintenance of soil productive capacity and water quality in river basins.
The Deciduous Seasonal Forest of Rio Grande do Sul, the primary forest typology of that state (23.8% of the natural forests), partakes in this situation (Rio Grande do Sul, 2002). The forest's remnants, despite their range and significance, are actively threatened by rural and urban development . This condition evidences the need for studies that determine these forests' ability to support sustainable farming or the need to recover certain ecosystems.
Information on floristic composition, vegetation structure, and relationships between vegetation and the environment were contemplated by Scipioni et al. (2012) and Turchetto et al. (2017) regarding remnants of Deciduous Seasonal Forest located on the ridge of the Southern Plateau. However, existing data concerning the western portion of the Plateau is scarce, especially when considering hillside forests and floristic differentiation studies within a fragment, since hillside environments have not been the focus of research in the region.
The differentiation of communities into a forest fragment can be conducted by cluster analysis, which aims at gathering objects into homogeneous groups. The formation of distinct groups suggests the need for differentiated sustainable management in the forest. Similarly, knowledge of indicator species may foster future efforts of preservation and restoration (Marangon et al., 2016).
In this context, this study aimed at determining and differentiating floristic groups of the arboreal component of a Deciduous Seasonal Forest located in the western region of the Southern Plateau ridge of Rio Grande do Sul.

MATERIAL AND METHODS
A fragment of a Submontane Deciduous Seasonal Forest (IBGE, 2012), with an area of 10.5 ha, located in the municipality of Jaguari, Rio Grande do Sul, at coordinates 29º24'12" South and 54º38'06" West was assessed. According to the climatic Köppen classification, the region is considered of Cfa (subtropical) climate (Alvares et al., 2013), and the municipality's average annual precipitation is of 1,879 mm (Machado & Freitas, 2005).
The forest is situated in the western portion of the Southern Plateau ridge of Rio Grande do Sul, in an area with a steeply undulating relief (between 190 m and 260 m) and a predominantly western exposure, where the maximum slope is 40° (89%). Regosols and leptosols predominate in the area and exhibit lithic contact relatively close to the surface (Pedron & Dalmolin, 2011).
Five parallel sample strips were systematically installed from the base (west) to the top of the hill (east), approximately 50 m apart and with variable length (Figure 1). The sample area included the stretch of the forest fragment located inside a rural property that exhibited better conditions of preservation (3.8 ha), disregarding the section with visually more noticeable change. The second northernmost sampling range had an interruption due to the presence of rocky outcrop. The sampling strips were divided into contiguous plots with dimensions of 10 m × 10 m (100 m²), totaling 62 plots (0.62 ha of sample area). During the second semester of 2010 and the first half of 2011, all individuals with diameter at breast height (DBH) ≥ 5.0 cm in the plots were identified and measured.
Botanical material of unidentified species was collected for analysis and identification in the Herbarium of the Department of Forest Sciences of Universidade Federal de Santa Maria. In order to update and confirm the nomenclature of the species and their respective authors, the Flora do Brasil 2020 catalog was utilized (Forzza et al., 2018). Family delimitation followed the Angiosperm Phylogeny Group IV classification system (Byng et al., 2016).
Sampling adequacy was verified using the species accumulation curve generated with the PC-ORD 4.41 program (McCune & Mefford, 1999). The plots were grouped based on the abundance of the species, using the agglomerative hierarchical method, with Ward attachment, and the Euclidian distance as a measure of dissimilarity. The data were organized in spreadsheets with 62 lines (sampled plots) and 57 columns (sampled species), and the cluster analysis was processed using the SPSS 13.0 program for Windows (SPSS, 2004).
The formed clusters were differentiated by the occurrence of species, floristic similarity, analysis of indicator species, aggregation, successional group, and species dispersion strategy. The floristic similarity between the groups was obtained using the Jaccard index, which is based on the presence or absence of species, and ranges from 0 to 1, where 1 indicates maximum floristic similarity between the compared areas (e.g., forests, clusters, and plots). In other words, all species that belonged to a sample also belonged to another sample (Souza & Soares, 2013).
Indicator species analysis (ISA) was employed to verify the species that characterize the floristic groups. The method provides the indicator value (IV) for each species in each group, with the statistical significance of the IV verified by the Monte Carlo test, which retained 500 iterations in this study. Such analysis, performed in the PC-ORD 4.41 program, utilized a matrix containing the abundance of species in the plots in each floristic group (Kent, 2012;McCune & Mefford, 1999;Valentin, 2012).
The spatial distribution pattern (aggregation) of the species was assessed using the Payandeh index (Pi), determined in three classes (Souza & Soares, 2013): random (Pi ≤ 1.0), tend to aggregate (1.0 < Pi ≤ 1.5), and aggregated (Pi > 1.5). In this analysis, species with one sampled individual were disregarded, given they exhibited a non-aggregated (random) distribution, a procedure also adopted by Watzlawick et al. (2011) and Nascimento et al. (2001). When one species presented in more than one floristic group, the Payandeh index was calculated in the groups where the species had two or more individuals sampled.
The species were classified into successional groups, according to Swaine & Whitmore (1988) and Oliveira-Filho et al. (1994): pioneers (P), climax light-demanding (CL), and climax shadow-tolerant (CS). The former group requires full solar luminosity in regeneration, growth, development, and survival processes. Lightdemanding climatic species can germinate under shade, but young plants need abundant light to grow and reach the canopy and maturity. In turn, shadetolerant species can develop under shading, and reach maturity at the canopy or beforehand.

4/11
Floresta e Ambiente 2019; 26(4): e20150139 Andrzejewski C, Callegaro RM, Longhi SJ, Biali LJ As for the dispersion strategy, the species were classified into three categories, according to Pijl (1972): zoochorous, anemochorous, and autochorous. Zoochorous species are dispersed by animals and have attractive and/or alimentary diaspores, as well as diaspores with adhesive structures, such as hooks and bristles. Anemochorous species have diaspores with structures that provide wind dispersion (winged, feathery, or balloon-shaped). Autochorous species, on the other hand, do not fit into the two previous categories, presenting dispersion by gravity and explosion.
The successional group (P, CL, and CS) and the dispersion strategy (zoochory, anemochory, and autochory) were determined by bibliographic consultation. When the species exhibited different classification in the literature, field observations and researcher consultations were carried out.

RESULTS AND DISCUSSION
The species accumulation curve indicated that the sample was sufficient to represent the studied vegetation, given a tendency towards stabilization was observed (Figure 2). Such a trend was evidenced at the final portion of the curve, in which the 10% increase in the sample area resulted in an increment in new species of less than 5%, a condition mentioned by Kersten & Galvão (2011) as necessary to determine the minimum sample area.
A total of 1,107 individuals were sampled in the tree stratum, pertaining to 57 species and 25 botanical families (Table 1). The dendrogram generated by the cluster analysis ( Figure 3) indicated the presence of three groups: Group 1 (36 plots), Group 2 (18 plots), and Group 3 (8 plots). The Group 1 was referred to as the Middle Stage, while Group 2 was denominated the Advanced Stage, and Group 3 the Altered Forest. These designations were applied to approximate the reality of the clusters with the floristic properties of forests in different stages or preservation conditions. Therefore, the aspects observed in the study area and the characteristics of the indicator species determined by ISA were considered.
In the ISA, the species that retained a significant observed indicator value (OIV) in the Monte Carlo test    The species Casearia sylvestris is considered climax light-demanding and generalist regarding environmental conditions, occurring in several types of soils, from low to high chemical fertility, wet or dry, and from sandy to clayey texture (Carvalho, 2007;Scipioni et al., 2013). In addition, Reitz et al. (1988) emphasized that the species also develops in rocky and shallow soils. Thus, its occurrence is explained as being an indicator of the Middle Stage group, which included plots with different environmental conditions.
In the Advanced Stage group, the occurrence of Pilocarpus pennatifolius (tolerant to shade) as an indicator species evidences that the forest structure provided shading conditions that were favorable to the development of a dense population, with an incidence of the species in all plots of the group. Also, the preference of this species is associated to the group by its form of dispersion (autochory), which, according to Giehl et al. (2007), is effective only for diaspore displacement at short distances and, therefore, hinders the spread of propagules to larger areas.
The preference of the species Apuleia leiocarpa, Helietta apiculata, and Machaerium paraguariense regarding the Altered Forest group is due to environmental characteristics, such as proximity to the edge of the forest, which favored the emergence of these species, classified as climax light-demanding. Moreover, the significant surface stoniness verified in the area possibly acted as a limiting ecological factor for the establishment of other species, but was tolerated by the three indicator species of the group. Such conditions were mentioned by Lorenzi (2008) when describing the environment where the species Machaerium paraguariense preferably occurs: in secondary forests and almost always in high rocky terrains, where drainage is swift.
The indicator species of the Altered Forest group are classified as anemochorous, which partially explains their high abundance. This syndrome is favored in more open forests that lack a continuous canopy (Howe & Smallwood, 1982), as observed in some points in the area. According to Morellato & Leitão Filho (1992), the presence of emergent arboreal individuals associated with forest deciduousness may enhance the dispersion of anemochorous diaspores, influencing forest structure.   The Jaccard similarity index indicated that the Middle and Advanced Stage groups were more similar among each other (Jaccard = 0.60; 34 species in common) and that the lowest floristic similarity was verified between the Middle Stage and Altered Forest groups (0.48). The Advanced Stage and Altered Forest groups retained an index of 0.55. The difference in similarity is related to the presence or absence of the species in the groups. Twentytwo species, including Helietta apiculata, Machaerium paraguariense, Myrocarpus frondosus, Pilocarpus 8/11 Floresta e Ambiente 2019; 26(4): e20150139 Andrzejewski C, Callegaro RM, Longhi SJ, Biali LJ pennatifolius, and Sebastiania brasiliensis, were found in all three groups. In contrast, Banara tomentosa and Ruprechtia laxiflora and Cedrela fissilis and Sideroxylon obtusifolium occurred only in the Middle and Advanced Stages, respectively, while species exclusivity was not observed in the Altered Forest group.
In order to determine the degree of species aggregation, by means of the Payandeh index, 42, 24, and 14 species were considered for the Middle Stage, Advanced Stage and Altered Forest groups, respectively (Figure 4). An aggregated pattern was predominant in the arboreal stratum, indicating that most species tend to form population densities in specific areas.
The species Gymnanthes klotzschiana (8.29), Annona neosalicifolia (7.38), Myrocarpus frondosus (6.38), and Eugenia uniflora (5.50) were significantly aggregated in the Middle Stage group, and, in the Advanced Stage group, Schaefferia argentinensis (3.67) exhibited a higher aggregation index. In turn, the Altered Forest group retained higher values of aggregation of Machaerium paraguariense (11.68) and Sebastiania brasiliensis (4.32). The more significant aggregation of these species may be influenced by more favorable habitat conditions, such as landform and availability of light and water, species autoecology, and different rates of mortality and recruitment (Capretz et al., 2012;Crawley, 1986;Silva et al., 2012).
The spatial pattern of some species was distinct between the groups. For example, Gymnanthes klotzschiana (Middle Stage group = 8.29 -aggregated; Advanced Stage group = 0.94 -not aggregated) and Machaerium paraguariense (Middle Stage = 1.57grouped; Advanced Stage = 1.35 -tend to aggregate; Altered Forest group = 11.68 -aggregated) occurred in a different manner. The strongly aggregated pattern of Gymnanthes klotzschiana in the Middle Stage group may be related to water influence, due to the presence of water bodies adjacent to some plots. According to Silva et al. (2012), such pattern occurs due to the greater tolerance of the species to soils with water excess, which favors recruitment.
Considering the succession categories, it was verified, regarding the totality of the forest, that the CL species occurred in a greater number of individuals (57% of the total forest individuals) and species (63% of the total   species of the forest), followed by CS (34% and 23%, respectively) and P (5% and 11%) species (Table 2).
Regarding the number of species, the three groups had a forest-like pattern, with greater CL richness, followed by CS and P. Similar results were reported by Scipioni et al. (2013) and Callegaro et al. (2014) in the same forest type, where light-demanding climactic species retained larger proportions of individuals and species. The higher floristic representativity of these species was also observed by Abreu et al. (2014), in a Semi-Deciduous Seasonal Forest located in the Chapada dos Guimarães National Park, who attributed this condition to canopy discontinuity due to the valley's rugged relief.
The Middle Stage and Altered Forest groups had a more substantial number of individuals pertaining to the successional CL category, contrasting with the Advanced Stage, where more CS individuals occurred. Such condition can be explained by the abundance of the species Pilocarpus pennatifolius, which is a known indicator of the Advanced Stage group, classified as climax shadow-tolerant.
The analyzed forest was primarily composed of species with zoochorous dispersion (67%), a result similar to other forests surveyed in several regions of Brazil, such as the Seasonal Forest in Querência, MT (Stefanello et al., 2010), the Mixed Ombrophylous Forest in Tijucas do Sul, PR (Liebsch & Acra, 2007), and the Dense Ombrophylous Forest in Rio de Janeiro (Carvalho, 2010). Similarly to the forest fragment as a whole, the three groups retained a more considerable richness of zoochorous species, followed by anemochorous and autochorous individuals.
The groups showed discrepancies when the number of individuals was analyzed: the Middle Stage had zoochory as its main syndrome (54%), the Advanced Stage retained autochory (54%), and the Altered Forest, anemochory (76%). The elevated proportion of individuals of the predominant dispersion syndromes was due to the abundance of some species, namely Pilocarpus pennatifolius (autochorous), with 164 species in the Advanced Stage, Helietta apiculata and Machaerium paraguariense (anemochorous), with 108 individuals in the Altered Forest, and Casearia sylvestris and Annona neosalicifolia (zoochorous), with a total of 94 individuals in the Middle Stage. The numerical predominance of different dispersion strategies in each floristic group evidences that this ecological characteristic was determinant of the differentiation of arboreal communities in the analyzed fragment. Moreover, this result corroborates that of Howe & Smallwood (1982), who stated that the prevalence of a specific dispersal mechanism in a given habitat indicates that dispersion agent pressures and physical conditions influenced species selection.

CONCLUSIONS
The presence of three floristic groups with distinct indicator species was detected: Middle Stage (Casearia sylvestris); Advanced Stage (Pilocarpus pennatifolius), and Altered Forest (Apuleia leiocarpa, Helietta apiculata, and Machaerium paraguariense). The clusters had a more significant richness of zoochorous species, while the incidence of trees with each dispersion strategy was distinct between the groups, predominating anemochory in the Altered Forest, zoochory in the Middle Stage, and autochory in the Advanced Stage.
The light-demanding climatic species predominated in terms of richness and abundance in the Middle Stage and Altered Forest, while in the Advanced Stage a greater proportion of individuals belonging to the shade-tolerant climatic species was observed. Both aspects used for cluster differentiation showed that the two are floristically and structurally distinct.