Floristic-structural variation of natural... FLORISTIC-STRUCTURAL VARIATION OF NATURAL REGENERATION ALONG DIFFERENT TOPOGRAPHIC POSITIONS OF AN ECOTONAL FOREST IN SANTA CATARINA, BRAZIL

1 Received on 04.08.2015 accepted for publication on 03.05.2017. 2 Universidade do Estado de Santa Catarina, Programa de Pós-Graduação em Engenharia Florestal. E-mail: <chayanesouza8@gmail.con>,<karisouza@hotmail.com> and <karol_chini@hotmail.com> . 3 Universidade do Estado de Santa Catarina, Departamento de Engenharia Florestal. E-mail: <higuchip@gmail.com> and <carol_sil4@yahoo.com.br>. 4 Universidade do Estado de Santa Catarina ,Graduando em Engenharia Florestal,Lages,SC,Brasil. E-mail: <angelica.dalla.rosa@hotmail.com>, <luiz.crj@hotmail.com> and <felipewalterr@hotmail.com>. *Corresponding author.


INTRODUCTION
The southern Brazil Atlantic forest has different formations, such as the Mixed Ombrophilous Forest (MOF) occurring at generally higher than 500 m altitudes and the Deciduous Seasonal Forest (DSF) occurring from 150 to 800 m high (Klein, 1978).In the altitudes where they overlap their contact results in the formation of ecotone areas.These ecotonal areas are not homogeneous and may be influenced by a number of local physical factors, such as edaphic and topographic variations.
A high terrain variation is observed in the ecotonal areas of the Upper Uruguay region where rivers are known for draining into "steep" valleys.In such places, there is a sequence of different types of soils spread along the landscape according to topography (Cunha et al., 2009).The organization of both adult and regenerating tree species communities developing in such environments are conditioned in a small spatial scale by environmental factors related to topographic gradients (Rodrigues et al., 2007;Scipioni et al., 2009;Budke et al., 2010;Souza et al., 2015;Braga et al., 2015;Guo et al., 2017;Mohammadi et al., 2017).According to Gandolfi (2000) and Jafari et al. (2003) this happens because declivity acts as a factor that produces a variety of environmental situations in soil particle transport and vertical canopy organization, such as moisture gradient.That explains the importance of topographic variations on the organization of tree vegetation, as reported in former studies (Rodrigues et al., 2007;Braga et al., 2015), including the adult component of the same area (Souza et al., 2015).
One of the most critical phases for the establishment of tree species in forest communities refers to their regeneration stage under the understory condition (Kitajima and Fenner, 2000;Poorter, 2007).Studies aiming to deepen the knowledge on the factors that influence this stage are therefore relevant as they allow a better understanding on the development of a forest under natural conditions.Upper Uruguay stands out among the Brazilian southern regions where studies of this nature are necessary for subsidizing both conservation and restoration strategies since its forests face a critical status of conservation due to its selective exploration and forest fragmentation history.In this sense, the main goal of this study was to investigate the floristicstructural organization of the regenerative component according to its topographic position in an ecotonal area between the Mixed Ombrophylous Forest and the Deciduous Seasonal Forest of Santa Catarina, Brazil.Considering the topographic gradient, a high floristicstructural variation is expected thus reinforcing the importance of conserving different microhabitats in a forest under such conditions.

MATERIAL AND METHODS
The study was conducted in an advanced successional stage area previously inventoried (adult tree component) by Souza et al. (2015).It is an ecotonal area between the Mixed Ombrophilous Forest and the Deciduous Seasonal Forest in the municipality of Capão Alto, SC, Brazil (28º11'29"S and 50º45'34"L and altitude about 600 to 700 m).Its topography varies mainly from smooth to gentle slopes and its higher-sloped areas are near the main rivers of the region therefore forming steep valleys located in the sandstone-basalt plateau.It has a Cfa climate according to Köppen classification and its average annual rainfall varies from 1,200 to 1,900 mm well distributed along the year.According to the CRU CL 2.0 database (New et al., 2002) the average temperatures of its coldest and hottest months in the last 30 years  were 12.0ºC in July and 20.7°C in January, respectively (Grechka et al., 2016).
Data were collected in a hillside on the banks of the Barra Grande dam on Pelotas River, which reservoir is formed since 2005.Three sectors have been defined for the area: lower sector (near the generally less-sloped reservoir bank which shows moisture and areas subject to flooding), intermediate sector (highly sloped in the middle of the hillside), and upper sector (also highly sloped but at the top of the hillside).Each sector's slope was measured by Souza et al. (2015) who have observed about 14.38º (± 2.64), 19.74º (± 1.98) and 20.66º (± 1.24) for the lower, intermediate and upper sectors, respectively.
For characterizing the regenerative tree component (DBH lower than 5 cm and height higher or equal to 20 cm) we have used part of the plots allocated by Souza et al. (2015) -who have carried out a stratifiedsystematic sampling with 200 m 2 plots spaced 10 m apart from each other -for sampling the adult tree component (DBH higher than or equal to 5 cm).We have used 10 plots allocated in each of the three sectors, making 30 allocated plots (Figure 1) which were subdivided into different sized subplots according to the regenerating tree species height classification.Each smaller subplot was allocated within a larger plot according to Volpato's methodology (Volpato, 1994).Thus, the regenerating tree species class 1 has consisted of 20 cm to 1 m high plants sampled in 5 m 2 subplots; the class 2 has consisted of above 1 to 3 m high plants sampled in 10 m 2 subplots; and the class 3 has consisted of more than 3 m high plants with DBH less than 5 cm sampled in 20 m 2 subplots.
Regenerating individuals were tagged with metal tags so as to register each plant's height class.The individuals were collected and identified through both the literature and experts support.Species were classified by families arranged according to the APG III system (Angiosperm Phylogeny Group, 2009) so that their spellings were verified in the Brazilian Flora Species List website (http://floradobrasil.jbrj.gov.br/).
For data analysis, the species were arranged according to their presence and density in the sampling plots of all three sectors (lower, intermediate and upper).Abundance, Shannon diversity index (H') and Pielou's evenness index (J) were determined for each sector in order to analyze the existence of spatial heterogeneity of structural and diversity patterns due to environmental compartmentalization. Abundance values were evaluated for normality (Shapiro-Wilk test) and if data did not show normal distribution they were compared among sectors through the Kruskal-Wallis test followed by a post hoc nonparametric multiple test so the hypothesis of structural variation among sectors was tested.Richness was compared by the rarefaction technique in order to test the species number variation hypothesis according to topographic position.A multivariate ordering of non-metric multidimensional scaling (NMDS) type was performed by using Bray-Curtis as floristicstructural distance, from a species' abundance matrix related to each sector's plots, to verify the existence of floristic-structural composition variations due to topographic position.A nonparametric multivariate analysis of variance (NPMANOVA) was held through the same data matrix used in the NMDS in order to test the hypothesis of community variation in relation to topographic position.An indicator species analysis (IndVal) was carried out with the species' abundance matrix according to each sector's plots for verifying the existence of species associated to the sectors.All analyzes were held by the means of the statistical programming language R (R Development Core Team, 2015) together with the vegan (Oksanen et al., 2015) and labdsv (Roberts, 2015) packages.Figura com baixa qualidade -enviar em EPS em 300 dpi´s

RESULTS
A total of 771 individuals were sampled in 24 botanical families, 42 genera and 51 species along the different topographic positions (Table 1).One individual has not been identified for being too young for collection.The regenerating tree species abundance did not occur homogeneously along the hillside; the lower sector has shown more than half of the individuals [401 (251 from class 1 in 50 m 2 , 122 from class 2 in 100 m 2 , and 28 from class 3 in 200 m 2 )] which has significantly differed from the upper sector which has shown only 16.8% sampled individuals.
Likewise, variations in richness and diversity were observed along the topographic gradient so that the lower sector has shown the lowest diversity (2.63), evenness (0.75) and standardized richness (23.12).This pattern has reflected in the ordering produced by the NMDS, which has shown a 0.18 stress value therefore indicating that the analysis was adequate for ecological interpretations.In the ordering, one can notice that the lower sector, which has shown the lowest diversity and the highest abundance and ecological dominance, has also been the most distinct and homogeneous from the floristic-structural point of view since its plots have shown the lowest dispersion (Figure 2).These differences in the lower sector are indicated by the segregation relative to the plots of the other two sectors along the NMDS axis 2, which significance (p <0.001) has been confirmed by NPMANOVA.Some species are widely distributed along the area, such as Cupania vernalis Cambess., which has occurred in all sectors.However, nine indicator species have been observed, mainly in the lower sector (six) (Table 2).The intermediate and upper sectors had only two and one indicator species, respectively.Results are in line with the multivariate ordering thus indicating a higher floristic-structural distinction in the lower sector.

DISCUSSION
Results show that the environmental compartmentalization related to the terrain has proved to be an important source of floristic-structural heterogeneity for the regenerative component where the lower sector turned out to be the most distinct one in comparison to the other two sectors by showing the highest number of indicator species as well as the highest abundance with more than half of all 771 sampled individuals (121,400 ind.ha -1 ) -which have opposed its low richness, diversity and evenness.Variations in environmental conditions due to topographic sectors may have influenced diversity patterns in different ways, such as: i) the existence of more selective environmental conditions in the lower sector therefore causing its low richness and the dominance of a group of adapted species; (ii) on the opposite side, the lower sector may show better environmental conditions (e.g. a higher fertility) so that ecological dominance results from the competitive exclusion for species more efficient in acquiring resources; and iii) a higher environmental heterogeneity in the intermediate and upper sectors, such as the presence of gaps, resulting in a larger distribution spectrum of species from different ecological groups.
The spatial heterogeneity in the abundance distribution of tree species individuals is a common behavior in forest ecosystems (Hubbel et al., 1999;Pitman et al., 1999) and results from a complex interaction Floristic-structural variation of natural...

Trichilia elegans Intermediary
Annona rugulosa Intermediary Floristic-structural variation of natural... between several ecological factors, which is related to both the species characteristics and the habitat's quality.For instance, the high density in the lower sector suggests that the species occurring in this environment, besides producing a high amount of propagules, have found favorable environmental conditions for germinating seeds and forming a bank of young plants.This is supported by the fact that the lower sector's class 1 has shown a higher number of individuals than larger size plant classes.On the other hand, the regeneration process seems to be more limited in the upper sectors due to limiting factors, such as low propagules production and/or not favorable environmental conditions for germinating and establishing young plants.

Maytenus aquifolia Upper
Some species were common to all sectors, demonstrating the high plasticity in habitats colonization, such as the hygrophilous species Cupania vernalis, which is however capable of supporting high slopes.Its higher ecological plasticity is reinforced by Meyer et al. (2013a,b) who have listed it among the most important species in the regenerative component both in areas of Mixed Ombrophilous Forest and Deciduous Seasonal Forest in Santa Catarina, Brazil.According to Via et al. (1995), high plasticity species adapt to unstable environments therefore increasing its environmental tolerance.
There were nine indicator species, most of which in the lower sector, therefore reinforcing its higher distinction.The indicator species from this sector can be considered as tolerant to water stress since some parts of the area are subject to flooding.In fact, some of these species, such as Campomanesia xanthocarpa (Mart.)O.Berg, Matayba elaeagnoides Radlk.and Eugenia uniflora L., were considered by other authors (Reitz et al., 1983;Lorenzi, 2000;Silva et al., 2007;Narvaes et al., 2008) to be tolerant to soils with excess water.Of the two indicator species from the intermediate sector, Trichilia elegans A.Juss.germinates in shaded places, thriving on wet and hillside soils (Schupp et al., 2002), which explains its place of highest abundance.According to a study by Scipioni et al. (2009) held in FED, Santa Maria, RS, Brazil, Trichila elegans also belongs to the sectors influenced by high slopes.Yet Annona rugulosa (Schltdl.)H. Rainer is considered as an early secondary to late species (Lorenzi, 2000).The representative species from the upper sector, Maytenus aquifolia Mart., usually occurs in well-drained soils typical of hillside tops (Lorenzi, 2000).

CONCLUSION
It is concluded that the terrain's position is an important source of floristic-structural heterogeneity for the regenerative tree component as far as the small spatial scale under analysis is concerned, where the lower sector has proved to be the most distinct among sectors.Topographic variations probably make room for the existence of micro-habitats due to edaphic and microclimatic factors quality.The presence of indicator tree species in each sector reinforces the idea of considering the floristic groups' terrain position when indicating them in programs for recovering degraded areas.

Figure 1 -
Figure 1 -Sample units' spatial distribution in the three sectors (lower, intermediate and upper) in a fragment of a transitional area between the Mixed Ombrophilous Forest and the Seasonal Deciduous Forest in the municipality of Capão Alto, Santa Catarina, Brazil.Figura 1 -Distribuição espacial das unidades amostrais nos três setores (inferior, intermediário e superior) em um fragmento em área de transição entre Floresta Ombrófila Mista e Floresta Estacional Decidual no município de Capão Alto, em Santa Catarina.

Figure 2 -
Figure 2 -Non-metric multidimensional scaling (NMDS) for species and sectors in a fragment of a transitional area between Mixed Ombrophilous Forest and Seasonal Deciduous Forest in the municipality of Capão Alto, Santa Catarina, Brazil.Figura 2 -Escalonamento Multidimensional Não-Métrico (NMDS) para as espécies e setores de um fragmento em área de transição entre Floresta Ombrófila Mista e Floresta Estacional Decidual no município de Capão Alto, em Santa Catarina.

Table 2 -
Indicator species of the regenerative component for each topographic sector in a fragment of a transitional area between the Mixed Ombrophilous Forest and the Seasonal Deciduous Forest in the municipality of Capão Alto, Santa Catarina, Brazil.Tabela 2 -Espécies indicadoras do componente regenerativo para cada setor topográfico, em um fragmento em área de transição entre Floresta Ombrófila Mista e Floresta Estacional Decidual no município de Capão Alto, em Santa Catarina. .
* Exotic species.The values of total individuals by sector followed by different letters indicate significant differences by the nonparametric multiple test applied after the Kruskal-Wallis test.