FOREST EDGE EFFECTS ON THE PHYTOSOCIOLOGICAL COMPOSITION OF AN ARAUCARIA FOREST FRAGMENT IN SOUTHERN BRAZIL

The Araucaria forests of Southern Brazil have been drastically reduced as a result of anthropic intervention and are currently found at different stages of succession. The objective of our study was to understand the influence of forest fragmentation on the phytosociological composition of its tree component. The composition of floristic groups of the tree component and their correlation with the edaphic and environmental variables in a fragment of the Araucaria Forest in the municipality of Curitibanos, state of Santa Catarina, Brazil was analyzed. A hundred and ten permanent continuous plots (10 m2) with forest edges of different origins and formation times were established for sampling. Trees with diameter ≥ 5 cm at breast height (DBH) were measured and identified. Using TWINSPAN, data were analyzed for species importance values (including indicator species) in the formation of floristic groups. Floristic similarities resulted in separating groups conditioned by the forest succession. The presence of the invasive species Pinus taeda L. and pioneer species at the forest expansion border contributed to the formation of a spatially cohesive group. The other areas did not show influence of edge effects, being in an initial secondary succession stage, dominated by Jacaranda puberula Cham. Low edaphic variation among the classes of lithosols in a flat relief had no influence on the formation of floristic groups.


INTRODUCTION
The overall forest cover by the subtropical forests with Araucaria has been drastically reduced over the last hundred years, making them one of the most threatened forest formations in Brazil.This forest type, called the Mixed Ombrophilous Forest (MOF) or Araucaria Forest, is characterized by the presence of tropical Afro-Brazilian and austral-antarctic Andean elements, with multiple plant species associations in various stages of succession.The forest canopy is made conspicuous by the tree species Araucaria angustifolia (Bertol.)Kuntze (LEITE, 2002).The Santa Catarina Forest and Floristic Inventory (IFFSC) estimates that the remaining MOF forest in this state represents only 22% of the original cover (VIBRANS et al., 2013).
Ecological succession and edge effects in the vegetation that is undergone anthropic interventions is still not well understood.The state of Santa Catarina is noticeable for predominance of secondary forests with a long history of logging, where selective tree harvesting, extensive livestock management, and opening of roads are the main internal impact factors, with agricultural crop cultivation, cattle farming, and forest plantations of such exotics as Pinus and Eucalyptus spp., external.These activities promote alterations in the tree species abundance and changes in the fauna/flora interactions, both biotic and abiotic, thus affecting ecological processes within the plant communities.Reestablishment of natural dynamics in such plant communities depends on the ability of specific functional groups to overcome adversities imposed by the local conditions and on the richness of species adaptable to development in the altered environment (REZNIK et al., 2012).
The edge effect is a direct consequence of forest fragmentation, impacting its structure by exposing the forest environment to altered microclimatic conditions and leading to changes in its composition, population structure, number of tree species (REZNIK et al., 2012), and, most importantly, local extinctions caused by invasion of exotic species that favor biodiversity loss.In such environments, the dynamic of change is more pronounced due to higher individual mortality rates at the forest edge (MURCIA, 1995).
In this context, studies on species distribution and floristic groups aid to our understanding of phytosociological relationships within fragmented forest communities, making more effective their monitoring and management with the aim of their conservation.
The objective of the present study was to analyze the phytosociology of the arboreal component in a forest fragment with distinct edges in a flat, drained terrain.The suggested hypothesis is that the fragmentation and proximity of the distinct forest edges may be directing formation of floristic groups with pioneer species.Edaphic variables were also studied to determine their role in the forest succession process.

Study area
The study was a forest fragment of approximately eight ha, at the Curitibanos Campus of the Federal University of Santa Catarina (UFSC) (Figure 1).The area is constituted of tholeitic basalts of the Gramado Formation (Serra Geral Group, São Bento Supergroup, Lower Cretaceous Period) (WILDNER et al., 2014).Within the regional context, the relief is smooth-wavy to wavy; in the study area, smooth wavy to flat.The climate of the region is characterized by average annual temperature of 17 °C, ranging from 18.1 °C to 22 °C.The average temperature of the coldest month is less than or equal to 13 °C and that of the hottest month exceeds 22 °C.The precipitation is 1,600 to 1,900 mm.ano -1 .The climate is classified, according to Köppen, as humid subtropical (Cfb), with the oceanic influence in the dry season and a temperate summer (ALVARES et al., 2013).The vegetation typology of the study area is that of the Mixed Ombrophilous Forest, located at the end of a montane altitudinal gradient at 1,100 m (IBGE, 2012).The fragment has three borders: one, limited by an open road and pastures formed more than 60 years ago; another, formed by a firebreak and a forest of Pinus taeda formed in the 90's; the third, a humid area with native herbaceous vegetation.

Data sampling
For the analysis of the vegetation, 110 continuous plots of 100 m², with trees (living and dead) and lianas of a diameter at breast height (DBH -1.3 m) greater than or equal to 5.0 cm, were sampled (Figure 1).Bifurcated tree trunks were considered a single individual, with their diameter the sum of the basal area.The height of the trees was estimated by comparing the extension of the telescopic cable of a high pruning pole and by using reference trees measured with the TruPulse 200X laser rangefinder.
The canopy cover of the plots was estimated, from January to February 2013, using a spherical densiometer readings of the four cardinal points.The canopy opening variable was defined using the mean readings.Botanical material of the unidentified tree species was submitted for determination to taxonomic specialists, and botanical nomenclature was verified in the Flora List of Brazil (FLORA DO BRASIL 2020, 2016).Exsicates were incorporated in the herbarium of the Federal University of Santa Catarina, Campus Curitibanos (CTBS/UFSC).
The forest substrate was characterized by the physical and chemical properties of the soil.Surface soil sampling (0 -20 cm) was performed according to the procedures suggested by Santos et al. (2013b).Trenches at depths of 0.5 to 1.3 m were used to determine soil classes according to the Brazilian Soil Classification System (SiBCS) (SANTOS et al., 2013a) and their correspondences in the World Reference Base for Soil Resources (USS Working Group WRB) and subsequently defined by probing with Dutch augers at 0.4 -0.6 m.Soil samples were analyzed at the Soil Laboratory of the Federal University of Santa Maria (UFSM) for the following chemical parameters: pH in H2O, calcium (Ca), magnesium (Mg), aluminum (Al), potassium (K), phosphorus (P), effective cation exchange capacity (ECEC), base saturation levels (V%), Al saturation (%), and organic matter (OM) FLORESTA, Curitiba, PR, v. 48, n. 4, p. 483-492, out/dez. 2018Scipioni. M. C. et.al. ISSN eletrônico 1982-4688 DOI: 10.5380/rf.v48i4.51281485 (DONAGEMA et al., 2011).Morphological features of the soil horizons were described and the same chemical parameters used to study their composition.The granulometric composition of soil samples (DONAGEMA et al., 2011) was studied at the Chemical Laboratory of Federal University of Santa Catarina, Campus Curitibanos (CTBS/UFSC).

Data analysis
The Two-Way Indicator Species Analysis (TWINSPAN) was used to identify floristic groups, indicator species, and spatial distribution of plots on the forest edges.The species abundance matrix by plots was analyzed.In the absence of an environmental relief gradient, the group divisions were defined with the eigenvalue ≥ 0,25.The definition of groups was also based on the phytosociological composition of clusters, by observing the successional species change and using knowledge of the autoecology of the indicator species and species with the highest importance values.PC-ORD ver.6.0 (MCCUNE; MEFFORD, 2011) was used for the grouping analysis.
In order to diagnose existence of environmental differences and significant variables in the formation of floristic groups, environmental and phytosociological parameters for the soil class and floristic groups were compared using mean values for each sampled plot in the One-way ANOVA analysis, with the post-hoc Tukey's test.
Species accumulation curves (CHAO et al., 2014) and sample coverage were used to assess the sampling adequacy in the iNEXT software package (CHAO et al., 2016).For data extrapolation, the species accumulation curve twice the sample size was used.Confidence intervals of the species accumulation curve, based on the 1,000 replications bootstrap, were 95%.

RESULTS
The relief of the study forest fragment was flat, characterized by absence of a topographic gradient, rock outcrops, and significant variation in the canopy opening between the soils classes.The soil texture ranged from clayey to very clayey.The stoniness of the soil was considered very stony, and the rockiness as rocky.The soils were classified as Neossolo Litólico (SiBCS)/Epileptic Regosol (Clayic) (WRB) and Cambissolo Húmico FLORESTA,Curitiba,PR,v. 48,n. 4,out (SiBCS)/Endoleptic Umbrisols (Clayic, Profundihumic) (WRB).There was no significant variation (α = 0,05) in the chemical and granulometric composition, density, and basal area between the soil classes in the sampled plots, except for higher magnesium values in Cambisols (Table 1).
Table 1.Chemical and soil variables (sand, silt, and clay) (0 -20 cm), canopy cover, tree density, and basal area of 110 plots in the study fragment of the Araucaria Forest, municipality of Curitibanos, state of Santa Catarina, Brazil.Tabela 1. Variáveis químicas e granulométricas do solo (0 -20 cm), abertura de dossel, densidade de árvores e área basal de 110 parcelas alocadas em um fragmento de Floresta Ombrófila Mista, em Curitibanos, Santa Catarina, Brasil.Values are means ± standard deviation of n samples of each soil class.In situations where the F and P test indicated significant differences between the soil classes, the averages followed by the same letter did not differ significantly from each other in the Tukey test (α = 0,05), and the values without letters were not significant.
The largest basal areas were those of the dominant canopy species: Araucaria angustifolia, Ocotea bicolor, and Matayba elaeagnoides Radlk (11,9% basal area/ha).The Shannon diversity and Pielou equability indices were, respectively, H'= 3.00 and J' = 0.72.The high ecological dominance indicated by the Pielou index was influenced by the high values of relative density (RD) of Jacaranda puberula (Table 2).This species presented the highest importance value (IV), followed by Ocotea bicolor, Araucaria angustifolia, dead standing trees, Calyptranthes concinna, Vernonanthura discolor (Spreng.)H.Rob., Casearia decandra, Matayba elaeagnoides, and Ocotea puberula (Rich.)Nees, which, together, amounted to 64.8% of the cover value and 61.2% of the IV.The first three species were present in more than half of the plots.
In the TWINSPAN analysis, the floristic composition of the study area was divided into three groups (Figure 3).In the first division, it segregated the secondary forest area (Group 1) and the border area with predominance of pioneering species next to the pine (Pinus taeda) forest plantation (Group 2).In the second division, two groups within the secondary forest with low floristic dissimilarity, with Jacaranda puberula of the greatest importance value, emerged (Table 3): Group 1.1 had Casearia decandra as the predominant indicator species in the shaded lower stratum, whereas Group 1.2 had such pioneer indicator species as Zanthoxylum rhoifolium Lam. and Vernonanthura discolor.In terms of spatial distribution, Group 2 was limited to the border area, where an expansion of pioneer forest vegetation, favored by the space between the forest and the pine forest plantation, was possible.The other groups (2.1 and 2.2) were limited by the edge areas: the roadside and forest border.The soils did not show influence on the spatialization of the floristic groups due to the low pedological variation among the sampled plots (Figure 3).Table 3. Phytosociological parameters of the main ecological groups and their totals, with values for the most important species, responsible for approximately 50% of the IV.Tabela 3. Parâmetros fitossociológicos total e dos principais grupos ecológicos, com a descrição dos valores das espécies de maior importância, que representam, aproximadamente, 50% do VI total.Canopy opening and basal area were consistent parameters for differentiation of the groups in the secondary (Groups 1.1 and 1.2) and pioneer vegetation (Group 2), respectively, with higher values for the basal area and lesser values for the canopy opening.The mean diameter and height values showed no differences among the groups, whereas, the density values were distinctive for the three groups, with high values for Group 1.2, in the intermediate process of succession, and a smaller number of trees in Group 2, dominated by shrubby vegetation (Table 4).Values are means ± standard deviation of n samples of each floristic group.In cases when the F and p tests indicated significant differences between the groups, averages followed by the same letter did not differ significantly from each other in the Tukey test (α = 0,05), while values without letters were not significant.

DISCUSSION
Comparison of our study fragment and other surveys of the MOF in the advanced secondary stage showed the same floristic patterns for the richest families, Myrtaceae and Lauraceae (SCHAAF et al., 2006;SILVA et al., 2012;HIGUCHI et al, 2013).However, our study also showed that the forest edge, with the Asteraceae standing out as pioneer species, was characterized by an increase in the number of species typical of open areas (GUIDINI et al., 2014).The presence of this family at forest edges is conditioned by a greater canopy opening and luminosity in the understory, associated with the succession stage of the pioneer tree component with the low basal area (SIMINSKI et al., 2013).Guidini et al. (2014), evaluating the forest edge effects, also confirmed the presence of Asteraceae and Pinus taeda in that sector, occupying space of the initial succession species and those classified as pioneers.
Information on forest edge effects of forest fragments in contact with commercial Pinus plantations is of great importance, especially for the subtropical regions, such as southern and southeastern Brazil, where areas of commercial forest plantations dominate natural habitats (ROSA et al., 2015).The occurrence of Pinus taeda within the MOF forest fragments was low.The colonization capacity of Pinus taeda, impeded by hydromorphic soils and shade environments, is not restricted by the low fertility of litholical soils.But its clustering with pioneer species at the edge of forest fragments, highlights its potential to substitute native species in the process of regeneration, by colonizing areas with greater luminosity and open canopies, such as clearings and areas in the initial stages of succession (GUIDINI et al., 2014).
Analyses of differences in species diversity of the regenerative component in the OMF in relation to distance from the forest edge showed, according to Rosa et al. (2015), a greater diversity in the first 20 meters.In our observations, the pioneer arboreal component, composed for the most part by the genus Baccharis (Asteraceae), is restricted to the first 10 m of the forest fragment associated with Pinus taeda.
The diversity values for the MOF domain in our study, compared to those cited by Higuchi et al. (2016), may be considered as intermediary.In those studies, values higher than H'= 3,5 and J= 0,80 indicate, respectively, greater diversity and lower ecological dominance, with the lower values showing a degree of anthropogenic intervention.Silva et al. (2012) recorded 80 species in the MOF areas they sampled, with the greater species richness due to a great environmental heterogeneity of their study area along an environmental gradient.Their site was a large forest fragment in an advanced succession stage that included a spring with a stream outflow, and characterized by humidity variations associated with its topography.Our study site, on the other hand, was a small continuous forest fragment with no relief gradient, thus providing no significant variation in the environmental conditions for a greater species richness, even when accounting for the exclusive border tree species becoming important in the formation of a floristic group by changing the light gradient as a result of the border effect.Generally speaking, the smaller the biologically altered forest fragments, the lower is their floristic and environmental heterogeneity, with the isolation being a defining factor for the reproductive conditions of their

Figure 3 .
Figure 3. Dendrogram of the classification of ecological groups and indicator species.The first three letters represent name of the genus and specific epithet, as shown in Table 2. Figura 3. Dendrograma da classificação dos grupos ecológicos e espécies indicadoras.As três primeiras letras representam o nome do gênero e epíteto específico das espécies arbóreas, conforme Tabela 2.

Table 2 .
List of tree species in the study fragment of the Araucaria Forest, municipality of Curitibanos, state of Santa Catarina, Brazil, with their phytosociological parameters and ranked by the importance value (IV).Tabela 2. Espécies arbóreas e parâmetros fitossociológicos de um fragmentode Floresta Ombrófila Mista  Montana, em Curitibanos, Santa Catarina, Brasil.

Table 4 .
Phytosociological parameters and canopy cover of floristic groups in the study fragment of the Araucaria Forest, municipality of Curitibanos, state of Santa Catarina, Brazil.Tabela 4. Parâmetros fitossociológicos e abertura de dossel de grupos florístico em um fragmentode Floresta  Ombrófila Mista, em Curitibanos, Santa Catarina, Brasil.