Phytosociology and natural subtropical grassland communities on a granitic hill in southern Brazil

Fitossociologia e comunidades de campos subtropicais naturais em um morro granítico no sul do Brasil

ORIGINAL PAPERS

Phytosociology and natural subtropical grassland communities on a granitic hill in southern Brazil

Fitossociologia e comunidades de campos subtropicais naturais em um morro granítico no sul do Brasil

Robberson Bernal Setubal^{I, 1}; Ilsi Iob Boldrini^II

^IUniversidade Federal do Rio Grande do Sul – UFRGS, Depto. Botânica, Lab. Estudos em Vegetação Campestre, Av. Bento Gonçalves 9500, prédio 43432, s. 110, 91501-970, Porto Alegre, RS

^IIUniversidade Federal do Rio Grande do Sul, Programa de Pós-graduação em Botânica. Av. Bento Gonçalves, 9500, Bloco IV, Prédio 43433, Campus do Vale, Bairro Agronomia, 91501-970, Porto Alegre, RS, Brasil

ABSTRACT

A phytosociological survey was carried out in a study area located at Serra do Sudeste, southern Brazil, where forests and grasslands are distributed in a mosaic, seeking to unravel diversity patterns in four different grassland communities. Grassland management traditionally adopted by the local population is characterized by burning practices that aim to eliminate woody species, delaying the forest expansion process that is favored by the extant climate. The number of plots distributed per community was as follows: rocky grasslands (17), dry grasslands (33), moist grasslands (15) and marshy grasslands (5). Different numbers of plots were used due to the natural conditions of these communities, with highest cover for dry grasslands, followed by rocky, moist and marshy grasslands. Data analyses consisted of calculating community indexes and parameters and exploratory multivariate analysis. We verified that c. 15% of species among the 177 registered taxa were highly dominant in the constitution of the vegetation matrix in all communities, whereas most of the species showed low frequency and cover values. Rocky and dry grasslands showed higher similarity and diversity indexes than moist and marshy grasslands. We concluded that the large number of rare or intermediate-frequency species is decisive for the high diversity found in these grasslands.

Key words: Pampa biome, diversity, rare species, fire.

RESUMO

Um levantamento fitossociológico foi realizado em uma área de vegetação campestre localizada na Serra do Sudeste, sul do Brasil, onde campos e florestas estão distribuídos em mosaico, buscando compreender padrões de diversidade em quatro comunidades. O manejo destes campos realizado pela população local é caracterizado por práticas de queimadas periódicas visando eliminar espécies lenhosas, retardando assim o processo de expansão florestal favorecido pelo clima atual. O número de parcelas por comunidade foi: campo rupestre (17), campo seco (33), campo úmido (15), campo brejoso (5). Diferentes totais de parcelas foram utilizados devido às condições naturais dessas comunidades, com uma alta cobertura de áreas de campo seco, seguido pelos campos rupestres, úmidos e brejosos. A análise de dados consistiu no cálculo de índices e parâmetros fitossociológicos e análises multivariadas exploratórias. Verificou-se que aproximadamente 15% dos 177 táxons registrados apresentaram maior dominância na composição da vegetação nas quatro comunidades, enquanto a maioria das espécies apresentou baixos valores de frequência e cobertura. Maiores índices de similaridade e diversidade foram encontrados entre campos secos e rupestres do que entre campos úmidos e brejosos. Conclui-se que o grande número de espécies raras e de frequência intermediária é determinante para os elevados índices de diversidade encontrados nestes campos.

Palavras-chave: bioma Pampa, diversidade, espécies raras, fogo.

INTRODUCTION

In southern South America, there is a grassland formation inserted in the Pampa Biome, regionally recognized as "campos grasslands", that extends as far as Uruguay, eastern Argentina and southern Brazil (Rosengurtt 1944; Burkart 1975; Bilenca & Minarro 2004). In Brazil, this biome is found exclusively in the southern half of the state of Rio Grande do Sul (RS), and shows high species richness and foraging value (Overbeck et al.2007). The plant diversity found in these grasslands is the result of a long evolutionary history and extensive habitat variation that occurs along the environmental gradient present between the coast and the continent interior (Boldrini 2007). Events of climatic fluctuation that took place during the quaternary shaped a complex vegetation mosaic, due to the interpenetration of forests in grassland areas, which makes these formations difficult to delimitate and controversial concerning about an adequate classification system (Lindman 1906; Rambo 1956; Teixeira et al.1986; Pillar & Quadros 1997; Boldrini 1997; Marchiori 2002; Boldrini 2009).

According to Boldrini (2009), grassland vegetation at the Pampa biome, present in RS, shows eight different phytophysiognomies related to the variation of dominant species and ambient factors in different geomorphological provinces of the state. The chain of granitic hills present in Porto Alegre city (RS), which belongs to the geological formation locally known as Serra do Sudeste, presents the "Savanoid Vegetation"phytophysiognomy. In this region, grasslands and forests occur in mosaics, the first usually occupying smaller areas, restricted to hilltops and isolated by forests present in hillsides and surrounding plains. The savanoid physiognomy is characterized by the occurrence of eventual patches or isolated individuals of ligneous shrubs and pioneer tree species in grasslands. The progressive increase of ligneous species over grasslands offers conditions to the establishment and development of other tree species in the surroundings, leading to a slow process of forest advance through nucleation and border expansion dynamics (Muller & Forneck 2004; Oliveira & Pillar 2004; Behling et al.2007; Pillar et al.2009). The ongoing process of forest expansion is linked to the hot and moist climate of the last 4,000 years, and was intensified in the last thousand years (Behling 2002; Behling et al.2007). The human management of the grasslands consists on periodical burnings, traditionally used to eliminate the biomass of ligneous species and cespitose grasses that progressively dominate these formations. Since grazing activities are low in Porto Alegre region, fire is predominantly used to maintain the grassland matrix, and the renewing of pastures for animal foraging is of lesser interest. Fire also slows down the process of forest expansion, and shapes a complicated mosaic of vegetation in different successional stages (Pillar & Quadros 1997; Behling et al.2007; Muller et al.2007).

Besides the regional suggestions of classification for the grasslands of RS state and the lack of consensus among them, there are few quantitative studies with suggestions of community classification in the local scale (e.g., Focht & Pillar 2003; Ferreira & Setubal 2009). Setubal & Boldrini (2010) qualitatively investigated a natural grassland remnant in a granitic hill at Porto Alegre, Serra do Sudeste, identifying floristic groups and their association with environmental variables, suggesting a classification of four plant communities. The present study aims for the complement of this knowledge by comparatively evaluating these communities through quantitative patterns such as abundance, richness, diversity and similarity. We believe that a larger number of local studies might be the basis for a better comprehension of floristic and ecological patterns in the different phytophysiognomies that constitute the grasslands of the Pampa biome in RS state.

Material and Methods

Study area

The study area is located at Morro São Pedro (30º10'S; 51º06'W; 289 m maximum altitude), Porto Alegre municipality, RS, an elevation that is part of the chain of granitic hills that belongs to the Serra do Sudeste (or Crystalline Shield) (Rambo 1956; Suertegaray & Fujimoto 2004). Climate at the region is moist-subtropical (Cfa according to the Köppen's classification) with annual mean temperature ranging between 18 and 20ºC and annual mean rainfall between 1300 and 1500 mm, without drought season (Nimer 1990).

Geology is constituted by Viamão granite that originated Argisols (hillsides), Cambisols and Neosols (hilltops and upper hillsides), characterized by low fertility and moderate to strong acidity (Streck et al.2008; Schneider et al.2008). Vegetation comprises ca. 692 ha of forests and 440 ha of grasslands, which occur in mosaic (Güntzel et al. 1994). Grasslands predominantly occupy hilltops and hillsides, whereas forests predominate in southern hillsides, and this is the general vegetation distribution pattern at the local granitic hills (Aguiar et al.1986).

Phytosociological survey

For the quantitative sampling, we carried out a phytosociological survey between October and December 2008, in grassland patches distributed in the central and southern portions of Morro São Pedro, summing up ca. 80 ha of sampling area. Sampling was carried out in four communities that were recognized according to the classification based on Setubal & Boldrini (2010): Rg = rocky grassland; Dg = dry grassland; Mog = moist grassland; Mag = marshy grassland (Tab. 1). Four different environmental sites of each community were selected for sampling. These sites were of different area sizes due to the natural conditions of these communities, with high cover of dry grasslands, followed by rocky, moist and marshy grasslands. In the marshy grassland community we used only one sampling site, due to the lack of another similar environment, summing up a total of 13 sites sampled. Transects were established along the core areas of each site, and sampling units were systematically placed in regular intervals of 15 m. In sites where they were used more than one transect, the distance between transects was 15 m too. We performed a total of 70 sampling units of 1 × 1 m (size was defined based on local vegetation size), and the number of sampling units per community was: Dg = 33; Rg = 17; Mog = 15; Mag = 5. In each sampling unit we quantitatively evaluated composition and the abundance of each registered species (only angiosperms) through identification and cover estimating using a modification of the decimal scale of Londo (1976) (Tab. 2). Vouchers of all the species registered in the survey were deposited in the Herbarium ICN.

Data analyses

For each species we calculated the following phytosociological parameters: absolute and relative frequency (FA and FR), absolute and relative cover (CA and CR) and index of importance value (IVI) (Boldrini & Miotto 1987). We also calculated the Shannon-Wiener diversity index and the equitability based on the values of relative cover of each species (Krebs 1999). Sample sufficiency was estimated by the rarefaction curve with bootstrap resampling with 1000 iterations and confidence interval of 95% using the software PAST (Hammer et al. 2001) and by the non-parametric estimator of CHAO (Chao 1984). We also performed a multivariate variance analysis (MANOVA) and an ordination (PCoA) with scatter diagrams using software MULTIV (Pillar 2004), evaluating the significance of contrasts among communities based on composition and abundance of sampled species (resemblance measure: chord distance). Analyses of indicator species were carried out with software PC-ORD (McCune & Mefford 1997). The floristic similarity index of Jaccard (Krebs 1999) was calculated between our sampled communities and between our study and three other surveys of grassland vegetation encompassing the chain of granitic hills at Porto Alegre region. The species arrangement in families followed the APG III (2009).

Results

Phytosociological survey

In a sample of 70 sampling units we registered 177 angiosperm taxa, distributed in 32 families and 106 genera (^{Appendix 1}). Among these species, only one (Maytenus cassineformis Reissek (Celastraceae)) is a pioneer forest species, so that the remaining species are native to the local grasslands. The five families with higher species richness were Asteraceae (39), Poaceae (34), Fabaceae (22), Cyperaceae (12) and Apiaceae (10), summing up 66.1% of the species sampled in the survey (the first three families correspond to 53% of this total) (Fig. 1).

The rarefaction curve registered a significant increase on its first third part (70% of the species appeared in the first 30% of the sampling units), followed by a progressive decline in the rate of species increase, despite the lack of total stabilization (Fig. 2). The non-parametric diversity estimator of CHAO pointed out 181 expected species for our survey. The actual number of species (97% of the value estimated), coupled with the rarefaction curve and the total sampled area validates the survey's representativeness, showing a significant sampling effort considering the general knowledge of the sampled vegetation.

Six families (Poaceae, Asteraceae, Fabaceae, Apiaceae, Cyperaceae and Iridaceae) condensed the larger sum of phytosociological parameters (FR = 80%; CR = 88,5%; IVI = 84%), but showed two groups with distinct sociological behavior (Tab. 3). Poaceae, Apiaceae and Cyperaceae registered higher cover values in comparison to their respective frequencies, and were responsible for the composition of the vegetation matrix. Out of these three families, six species concentrated ca. 26% of the CR parameter (position considering IVI rating in parentheses): Sorghastrum albescens (E. Fourn.) Beetle (1ª), Eryngium pristis Cham. & Schltdl. (2ª), Andropogon lateralis Nees (5ª), Scleria balansae Maury (6ª), Schizachyrium tenerum Nees (7ª) and Ischaemum minus J. Presl (13ª). Among the 15 grasses present in the first 30 IVI-rated species, 10 showed higher cover values in comparison to frequency. Among them, I. minus registered the higher proportional value between cover and frequency.

On the other hand, Asteraceae, Fabaceae and Iridaceae presented higher values of relative frequency in comparison to cover, evidencing species that are common in the study area, but of lesser contribution for the vegetation matrix. Examples of these species are (position considering IVI rating in parentheses): Aspilia montevidensis (Spreng.) Kuntze (3ª), Chrysolaena flexuosa (Sims) H.Rob. (4ª), Sisyrinchium palmifolium L. (14ª) and Macroptilium prostratum (Benth.) Urb. (24ª). Only five species (three Asteraceae and two Fabaceae) among the first 30 were subshrubs or ligneous shrubs, and all showed higher values of frequency in comparison to cover: Baccharis cognata DC. (8ª), Disynaphia ligulifolia (Hook. & Arn.) R.M. King & H. Rob. (9ª), Chromolaena congesta (Hook. & Arn.) R.M. King & H. Rob. (12ª), Collaea stenophylla (Hook. & Arn.) Benth. (17ª), Desmanthus virgatus (L.) Willd. (18ª) and Centrosema virginianum (L.) Benth. (22ª). Rubiaceae also stands out due to the frequency values of the eight species registered, all showing usually low cover values.

The pattern of vegetation heterogeneity evidenced that most of the species showed low values of frequency and cover (Figs. 3-4). Among the 177 sampled species, 66% occurred in up to 10% of the sampling units, and 50 species (20%) occurred in only one or two sampling units, and can be considered rare species, whereas the remaining species showed intermediate frequency values. Among the 1448 registries of cover per species, 95% ranged between cover classes 0.1 and 2.0 (64% only in class 0.5) demonstrating the low cover values shown by most species.

Among the species with highest frequency values, only 19 (11%) were registered in 20 or more sampling units, and only six (3.4%) were registered in more than 30 sampling units (A. montevidensis, C. flexuosa, E. pristis, S. albescens, C. congesta and A. lateralis). Only five families presented cover values per species in class 3.0 or superior: Poaceae (16 species), Cyperaceae, Apiaceae, Asteraceae and Fabaceae (3 species). However, between classes 6.0 to 10.0, besides Poaceae, only Apiaceae (class 6 – E. pristis) and Cyperaceae (classes 7 and 8 – S. balansae) were represented. Only three species showed cover values between classes 9.0 and 10.0, and all were grasses: I. minus, S. tenerum and S. albescens.

Analysis of multivariate variance found high significance of contrasts (p = 0.001) among the four pairs of communities in comparison to each other. The exploratory ordination analysis grouped together the sampling units in relatively clear sets of communities, arranged in a gradient (Fig. 5), but registered low explanation values for axes 1, 2 and 3 (14.4%, 11.2% and 7.4% respectively). Out of the 20 species that were more correlated with the axes, only six obtained eingenvalues higher than 0.4 (S. albescens, E. pristis, Trachypogon montufarii (Kunth) Nees var. montufarii, Sisyrinchium vaginatum Spreng., I. minus and S. balansae).

The analysis of indicator species registered 18 significant taxa (p<0.01) to recognize the four plant communities we evaluated. Moist grassland communities showed a larger proportion of indicator species/total species in comparison to communities present in drier areas (Mog = 9; Mag = 4; Rg = 4; Dg = 1) (Annex 2 and Tab. 4). Annex 3 shows a comparison between the sets of the 16 species with the highest IVI values registered at the four communities. Species with higher frequency values showed a generalist behavior, occurring in more than one community, such as A. lateralis that was present in rocky, moist and marshy grasslands. The vegetation's pattern of heterogeneity in all communities followed the general vegetation pattern, that is, a small group of species condensed the larger portion of the calculated parameters, whereas most species were less expressive in the survey.

The diversity values obtained for the survey as a whole were 4.42 nats (H') and 0.85 (J'). Mean number of species per plot was 20.7, and minimum and maximum values were six (Mag) and 41 species (Rg). The highest values of floristic similarity were found between the well-drained rocky and dry grasslands (Tab. 4). On the other hand, moderately-drained and poorly-drained communities (Mog and Mag) obtained distinct resemblance values. While moist grasslands registered similarity values ranging between 40 and 50% considering the well-drained areas, marshy grasslands obtained lower values when compared to all other communities. Despite the higher species richness found in dry grasslands, the highest diversity indexes were registered at rocky grasslands, which also presented highest family richness, total exclusive families and mean number of species per plot (Tab. 5). Generally speaking, dry grasslands present the highest species richness considering the most expressive families in our survey. Marshy grasslands, on the other hand, presented a larger proportion of exclusive species (56%), despite the lower richness and diversity values in comparison to the remaining communities, and the predominance of the families Poaceae and Cyperaceae.

The analysis of floristic similarity between our study and three other phytosociological surveys of grassland vegetation at Porto Alegre's granitic hills (Boldrini et al.1998; Overbeck et al.2006; Ferreira et al. 2010) registered similar values (ca. 27%) in all pairwise comparisons. A list of the 37 common species registered in the four studies is presented in ^{Appendix 4}. Families with highest number of shared species among the hills were once again Poaceae (12), Asteraceae (10) and Fabaceae (4), highlighting the shared herbaceous and shrub species with high frequency and/or cover in the composition of the regional grassland vegetation.

Discussion

Floristic and phytosociological surveys of the grassland vegetation present at the chain of granitic hills of Porto Alegre indicate the occurrence of ecosystems with high species richness. Overbeck et al.(2006) and Setubal & Boldrini (2010) registered the occurrence of ca. 500 species at Morro Santana (220 ha) and Morro São Pedro (440 ha), respectively. Ferreira et al. (2010) registered diversity indexes of 4.51 nats (H') and 0.86 (J') in the grasslands at Morro do Osso, which were very similar to the ones found in our study. Preliminary evaluations regarding a complete checklist of the grassland species that occur in Porto Alegre municipality point out 737 species (Setubal et al., in press), which represents ca. 35% of the 2.150 grassland species registered in the Pampa biome in RS state (I.I. Boldrini, personal communication). Total numbers of mosses and ferns that occur in the grasslands of the same region remain unknown. The high number of species coupled with the small area this vegetation occupies in Porto Alegre (7.43% according to Hasenack et al.2008) characterizes environments with high plant diversity, currently isolated in islands at hilltops due to the occupation of forests in hillsides and surrounding plains, and they are insufficiently represented in local conservation units (only 0.45%) (Setubal & Boldrini 2010). In spite of the lack of studies that point out the percent contribution of the local species' phytogeographic origin, dominant species present estival behavior, being derived from the altitude grasslands at northern RS state (Brazilian Rainforest biome). However, the occurrence of endemism of pampean species in the hills of the region (Ferreira 2010), points out the occurrence of transitional ecosystems between the two biomes, fact that still demands further evaluation.

The sociological behavior registered between the two groups shaped by the six most expressive families in the survey (i.e., with predominance in species' cover or frequency values) reflect the general pattern of heterogeneity and diversity of the studied grassland vegetation. While a small number of species with high cover values represents the larger portion of plant biomass in these grasslands, highly frequent species, but with low cover values, are intermingled in the vegetation matrix, contributing to increase diversity. The occurrence of a large number of rare and intermediately-frequent species significantly increases diversity, resulting in the elevated patterns of species richness presently found in these grasslands.

The recurrence of periodical burning events, in intervals of 2–3 years in the study area (I.I. Boldrini, personal observation), also common in other granitic hills at the region (Overbeck et al.2005), could be a key factor associated to the maintenance of diversity in these grasslands, since grazing activity is low in the region. Ferreira et al.(2010) concluded that a recent burning event in the grasslands of Morro do Osso contributed to the reduction of dominance of cespitose grasses, allowing rare species to establish and raising overall equability. According to Overbeck et al.(2005), after this diversity peak, in the intermediate to long-term species composition tends to return to the former condition, that is, the domination of cespitose grasses. However, due to new burning events, this domination is not a stable state, resulting in the alternation of different successional stages in contiguous grassland patches, so that each stage is characterized by different groups of species and development stages, contributing to raise diversity. Besides that, burning events also slow down the advance of forest vegetation over grasslands (Muller et al. 2007).

The low explanation in both axes obtained in the ordination analysis is probably related to the large number of variables registered in the survey, especially species with low cover and frequency values. Overbeck et al.(2006) found a similar problem when performing multivariate analyses in a composition and abundance data matrix with 170 species registered in sampling units placed along a gradient between forest borderlines and grasslands at Morro Santana. When data were condensed in twelve species groups (biological forms), the explanatory power of axes 1 and 2 raised from 13.8% and 11.0% to 46.7% and 20.7% respectively. Moreover, one needs to critically consider the influence of supposed distortions resulting from comparisons between data generated in sets with distinct number of sampling units per community used in this survey. This incompatibility is the result of natural conditions such as the natural area these communities occupy, with high predominance of dry areas and scarcity of moist areas, which hampers a uniform and representative sampling effort in each environment. However, we believe that the species-level evaluation we presented significantly contributes to a better comprehension of taxa's occupation and distribution patterns in the environment, unraveling patterns of similarity and floristic variability between and within the four sampled communities. For example, rocky grasslands and dry grasslands presented higher levels of dispersion and interface between their sampling units due to shared generalist species, whereas moist and marshy grasslands were grouped and with smaller dispersion area in the graphic, revealing environments with lower species richness and higher floristic exclusiveness. We believe to have fulfilled other important objectives to the advance of knowledge concerning this vegetation, such as documenting the occurrence of grasslands in moist environments at the regional hills and evaluating the significance of different communities as discrete and measurable floristic units, that may be comparatively analyzed at the species level, facts so far absent in vegetation surveys carried out in the region.

The analysis of indicator species was also satisfactory concerning a better characterization of the studied communities at Morro São Pedro. We chose a value of p = 0.01 aiming to exclude species with generalist behavior that did not help to recognize the habitats described in our study area and in other environmental sites present in the Pampa and Atlantic Rainforest biomes. Boldrini et al.(1998), in a study carried out at Morro da Polícia, Porto Alegre, verified positive relations between high altitude areas, with higher solar exposition (known for the occurrence of rock outcrops) with species such as Axonopus suffultus, Axonopus siccus (Nees) Kuhlm. and T. montufarii, all of which were rocky grasslands indicator species at Morro São Pedro. However, caution is needed when doing comparative evaluations, and further studies are needed concerning species' behavior to confirm their higher or lower exclusiveness in the occupation of similar habitats and to ultimately consider them as good indicator species. The evaluation of a larger cover and frequency database of south Brazilian grassland species in the intermediate and long-term may complement this knowledge.

The influence of environmental factors in species' distribution lead to structural variations among the different communities. The lowest diversity values obtained in dry, moist and marshy grasslands, which predominantly occupy areas with deeper soils, are the result of the high cover values of dominant cespitose grass species such as S. albescens, S. tenerum and I. minus respectively, which grant higher homogeneity and density to the vegetation due to large leaf biomass cover. In rocky grasslands, on the other hand, relief favors the occurrence of shallower soils (hilltops and upper hillsides), influencing lower values of vegetation height, which facilitates light entry, and lower competitive dominance of cespitose grasses. Another factor that probably contributes positively to a higher diversity in rocky grasslands is the presence of rocks of various sizes, which grants higher structural heterogeneity to the habitat, creating a higher number of ecological niches and attenuating the intensity of fire, thus allowing the occurrence of different juxtaposed succession stages and rising diversity and similarity indexes in these environments.

The larger proportion of species that are exclusive to moist environments, in comparison to dry areas, may be related to the hypothesis of the higher selectivity to colonization of these environments due to the occurrence of variable periods of hydric stress due to soil saturation. In marshy grasslands, the higher number of exclusive species and lower similarity in comparison to the remaining communities indicates that most of the regional species do not tolerate long periods of hydric stress, fact that restricts local species colonization in this environment. Thus, we hypothesize that the circular pattern, shaped as a gradient, obtained in the graphic generated through ordination analysis reflects groups that are predominantly influenced by environmental factors that define species' distribution patterns. We believe that the main local factors are hydromorphic levels, relief position and soil types that vary along topographic gradients, which could play the role of filters for colonization of these environments. The two first factors were considered by Focht & Pillar (2003) as preponderant for the comprehension of species distribution in a study of detailed vegetation ecology of a natural grassland area pertaining to the phytophysiognomy of the Campos do Centro do Estado (Boldrini 2009), relatively close to Porto Alegre region and with similar climatic conditions.

In spite of the low values of floristic similarity found among the different phytosociological studies carried out in the regional grassland vegetation, we suggest that they shape a floristic unit due to shared key species. Ferreira et al. (2010) have calculated floristic similarity values of ca. 33% between their study and two other phytosociologic surveys (Boldrini et al.1998; Overbeck et al.2006) carried out in the regional chain of granitic hills. The similarity value of ca. 27% calculated between Morro São Pedro and the three other studies, albeit low, confirms the 37 shared herbaceous and shrub species that are relatively frequent and with good participation in cover composition of the regional grassland vegetation. Besides the species registered in these four surveys, other common and rare species are also shared between two or more hills, fact already mentioned by Aguiar et al.(1986). Some factors that might be associated to the low values of similarity registered in the mentioned surveys may also be related to the difficulty to uniformly evaluate this vegetation due to its high richness of rare species (i.e., low-frequency species) and the use of different methods and sampling efforts in different phenological and successional stages of the vegetation. The larger distance and isolation between Morro São Pedro and the other hills that were already studied (Morro do Osso, Morro da Policia and Morro Santana), all of which are located in the central region of the municipality and connected by ridges, could possibly be responsible for the lower levels of similarity we found. Even considering species' distribution patterns that still need to be better interpreted, we believe that, besides the floristic unit, these grasslands share other features (e.g., similar groups of functional types of plants), since all previous surveys indicate similar structural features in the constitution of this vegetation. The occurrence of a similar anthropogenic management in all grassland-covered areas of the region also contributes to that, and represents a key aspect for the conservation of these formations that still demands further investigation, concomitantly with the establishment of Conservation Units that encompass the grassland-forest mosaic, due to its small area within the municipality. Besides that, the establishment of Conservation Units in grassland-covered areas must take into account long-term research that evaluate the use o burning practices for the conservation of grasslands and its efficiency as a maintaining agent of vegetation and its diversity. As pointed out by Nabinger et al.(2000), it is suggested that other conservationist management practices are comparatively studied in these researches, such as mowing and grazing with adequate animal density, aiming to evaluate costs and benefits of each treatment.

Acknowledgements

Thanks to Leda Charczuk Viana and Felipe C. Viana for supporting our field work; to CAPES (Coordenação de Aperfeiçoamento do Ensino Superior) for a scholarship granted to the first author; to the researching team at the Laboratório de Estudos em Vegetação Campestre (LEVCamp - UFRGS) for general support; to the Programa de Pós-Graduação em Botânica (UFRGS) for administrative support and physical infrastructure; and to Rodrigo Bergamim and Pedro Maria de Abreu Ferreira for collaboration in conducting the data analysis.

Artigo recebido em 08/04/2011.

Aceito para publicação em 02/10/2011.

This paper has additional data published in its electronic version.

Appendix 1

^{Click to enlarge}

Appendix 2

^{Click to enlarge}

Appendix 3

^{Click to enlarge}

Appendix 4

^{Click to enlarge}

• Aguiar, L.W.; Martau, L.; Soares, Z.F.; Bueno, O.L.; Mariath, J.E. & Klein, R.M. 1986. Estudo preliminar da flora e vegetação de morros graníticos da região da Grande Porto Alegre, Rio Grande do Sul, Brasil. Iheringia, Série Botânica 34: 3-38.
• APG III. 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Botanical Journal of the Linnean Society 161: 105-121.
• Behling, H. 2002. South and southeast Brazilian grassland during Late Quaternary times: a synthesis. Palaegeography, Palaeclimatology, Palaeoecology 177: 19-27.
• Behling, H; Pillar, V.D.; Müller, S.C. & Overbeck, G.E. 2007. Late-Holocene fire history in a forest-grassland mosaic in southern Brasil: Implications for conservation. Applied Vegetation Science 10:81-90.
• Bilenca, D.N. & Minarro, F.O. 2004. Identificacion de áreas valiosas de pastizal (AVPa) en las pampas de Argentina, Uruguay y sur de Brasil. 1Ş ed. Fundación Vida Silvestre Argentina, Buenos Aires. 323p.
• Boldrini, I.B. & Miotto, S.T.S. 1987. Levantamento fitossociológico de um campo limpo da Estação Experimental Agronomica, UFRGS, Guaíba, RS - 1Ş. etapa. Acta Botânica Brasílica 1: 49-56.
• Boldrini, I.I. 1997. Campos do Rio Grande do Sul: caracterização fisionômica e problemática ocupacional. Boletim do Instituto de Biociências/UFRGS 56.
• Boldrini, I.I.; Miotto, S.T.S.; Longhi-Wagner, H.M.; Pillar, V.P. & Marzall, K. 1998. Aspectos florísticos e ecológicos da vegetação campestre do Morro da Polícia, Porto Alegre, RS, Brasil. Acta Botânica Brasílica 12: 89-100.
• Boldrini, I.I. 2007. Formações campestres no Sul do Brasil: origem, histórico e modificadores. Pp. 7-13 In: Dall'agnol, M.; Nabinger, C.; Sant'anna, D.M. & Santos, R.J. (eds.). Anais do II Simpósio de forrageiras e produção animal. Departamento de Plantas Forrageiras e Agrometeorologia, UFRGS. Porto Alegre.
• Boldrini, I.I. 2009. A flora dos campos do Rio Grande do Sul. In: Pillar, V.D.P.; Müller, S.C.; Castilhos, Z.M.C. & Jacques, A.V.A. (eds.). Campos Sulinos conservação e uso sustentável da biodiversidade. MMA, Brasília. Pp. 63-77.
• Burkart, A. 1975. Evolution of grasses and grasslands in South America. Taxon 24: 53-66.
• Chao, A. 1984. Nonparametric estimation of the numbers of classes in a population. Scandinavian Journal of Statistics 11: 265-270.
• Ferreira, P.M.A. & Setubal, R.B. 2009. Florística e fitossociologia de um campo natural no município de Santo Antônio da Patrulha, Rio Grande do Sul, Brasil. Revista Brasileira de Biociências 7: 195-204.
• Ferreira, P.M.A.; Müller, S.C.; Boldrini, I.I. & Eggers, L. 2010. Floristic and vegetation structure of a granitic grassland in Southern Brazil. Revista Brasileira de Botânica 33: 21-36.
• Ferreira, P.M.A. 2010. Flora campestre, endêmica e ameaçada dos morros graníticos de Porto Alegre, Rio Grande do Sul, Brasil. Dissertação de Mestrado. Universidade Federal do Rio Grande do Sul, Porto Alegre. 132p.
• Focht, T. & Pillar, V. D. 2003. Spatial patterns and relations with site factors in a campos grassland under grazing. Brazilian Journal Biology 63: 423-436.
• Güntzel, A.; Freitas, A.E.; Tedesco, C.; Schirmer, C; Mondin, C.; Pinheiro, C.; Vélez, E.; Landau, E.C.; Leite, F.; Becker, F.; Rodrigues, G.; Meira, J.R.; Konrath, J.; Copertino, M.; Bendati, M.M.; Marczwski, M.; Haas, S. & Prochnow, T.R. 1994. Avaliação dos morros do município de Porto Alegre, RS, com base no uso do solo. Trabalho final para a disciplina de Estágio Integrado. PPG-ECO/UFRGS. Porto Alegre.
• Hammer, Ø.; Harper, D.A.T. & P. D. Ryan, 2001. PAST: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontologia Electronica 4. Available in <http://palaeo-electronica.org/2001_1/past/issue1_01.htm>
• Hasenack, H.; Cordeiro, J.L.P.; Boldrini, I.I.; Trevisan, R.; Brack, P. & Weber, E.J. 2008. Vegetação/Ocupação. In: Porto Alegre. Diagnóstico Ambiental de Porto Alegre. Secretaria Municipal de Meio Ambiente. Pp. 56-71.
• Krebs, C.J. 1999. Ecological methodology. 2nd. ed. Addison Wesley Longman, Menlo Park. 620p.
• Lindman, C.A.M. 1906. A vegetação no Rio Grande do Sul. Ed. Universal, Porto Alegre.
• Londo, G. 1976. The decimal scale for releves of permanent quadrats. Vegetatio 33: 61-64
• Marchiori, J.N.C. 2002. Considerações terminológicas sobre os campos sulinos. Ciência & Ambiente 24: 139-150.
• McCune, B. & Mefford, M.J. 2011. PC-ORD: Multivariate analysis of ecological data. Version 6. MjM Software Design, Gleneden Beach.
• Müller, S. C. & Forneck, E. D. 2004. Forest-grassland mosaics in the hills of Porto Alegre: a study case of forest expansion patterns in Santana hill, Rio Grande do Sul, Brazil. In: Porto, M.L. (ed.). Workshop proteção e manejo da vegetação natural da região de Porto Alegre com base em pesquisas de padrões e dinâmica da vegetação. UFRGS, PPG Ecologia, Porto Alegre. Pp. 29-37.
• Müller, S.C.; Overbeck, G.E.; Pfadenhauer, J. & Pillar, V.D. 2007. Plant fuctional types of woody species related to fire disturbance in forest-grassland ecotones. Plant Ecology 189: 1-14.
• Nabinger, C.; De Moraes, A.; Maraschin, G.E. 2000. Campos in southern Brazil. In: Lelaire, G.; Hodgson, J.; De Moraes, A.; Carvalho, P.C.F.; Nabinger, C. (ed.). Grassland ecophysiology and grazzing ecology. CABI Publishing, Cambridge. Pp. 355-376.
• Nimer, E. 1990. Clima. Pp. 151-187. In: IBGE (Instituto Brasileiro de Geografia e Estatística). Geografia do Brasil: Região Sul. Ed. IBGE, Rio de Janeiro.
• Oliveira, J.M. & V.D. Pillar. 2004. Vegetation dynamics on mosaics of Campos and Araucaria forest between 1974 and 1999 in Southern Brazil. Community Ecology 5: 197-202.
• Overbeck, G.E.; Müller, S.C.; Pillar, V.D. & Pfadenhauer, J. 2005. Fine-scale post-fire dynamics in southern Brazilian subtropical grassland. Journal of Vegetation Science 16: 655-664.
• Overbeck, G.E., Müller, S.C., Pillar, V.D. & Pfadenhauer, J. 2006. Floristic composition, environmental variation and species distribution patterns in burned grassland in southern Brazil. Brazilian Journal Biology 66: 1073-1090.
• Overbeck, G.E.; Muller, S.C.; Fidelis, A.; Pfadenhauer, J.; Pillar, V.D.; Blanco, C.C.; Boldrini, I.I.; Both, R. & Forneck, E.D. 2007. Brazil's neglected biome: The South Brazilian Campos. Perspective Plant Ecology, Evolution and Systematics 9: 101-116.
• Pillar, V.D. & Quadros, F.L.F. 1997. Grassland-forest boundaries in Southern Brazil. Coenoses 12: 119-126.
• Pillar, V.D. 2004. MULTIV: Multivariate exploratory analysis, randomization testing and bootstrap resampling. User's guide. v. 2.3.10. Departamento de Ecologia, UFRGS, Porto Alegre, RS. Available in <http://ecoqua.ecologia.ufrgs.br>. Access on 10 Dec 2009.
• Pillar, V.D.; Müller, S.C.; Oliveira, J.M.; Machado, R.E. 2009. Mosaicos de campos e florestas com araucária: dilemas para a conservação. In: Fonseca, C.R.; Souza, A.F.; Leal-Zanchet, A.M.; Dutra, T.L.; Backes, A. & Ganade, G. Floresta com araucária: ecologia, conservação e desenvolvimento sustentável. Holos Editora, Ribeirão Preto. Pp. 273-286.
• Rambo, B. 1956. A fisionomia do Rio Grande do Sul. 2. ed. Ed. Selbach, Porto Alegre. 472p.
• Rosengurtt, B. 1944. Estudios sobre praderas naturales en Uruguay. 4Ş Contribución. Las formaciones campestres y herbáceas del Uruguay. Agros 134. Montevideo. 45p.
• Schneider, P.; Klamt, E.; Kampf, N. & Giasson, E. 2008. Solos. In: Diagnóstico Ambiental de Porto Alegre. Secretaria Municipal de Meio Ambiente, Porto Alegre. Pp. 28-43.
• Setubal, R.B. & Boldrini, I.I. 2010. Floristic and characterization of grassland vegetation at a granitic hill in Southern Brazil. Revista Brasileira de Biociências 8: 85-111.
• Setubal, R.B.; Boldrini, I.I.; Ferreira, P.M.A.; Trevisan, R.; Schneider, A.A.; Grings, M.; Mello, A.S.; Buzatto, C.R. in press. Checklist da flora campestre. In: Setubal, R.B.; Boldrini, I.I. & Ferreira, P.M.A (org.). Campos dos morros de Porto Alegre. Igré Associação Sócio-Ambientalista, Porto Alegre.
• Streck, E.V.; Kämpf, N.; Dalmolin, R.S.D.; Klamt, E.; Nascimento, P.C.; Schneider, P.; Giasson, E. & Pinto. L.F.S. 2008. Solos do Rio Grande do Sul. 2.ed. UFRGS, Porto Alegre.
• Suertegaray, D.M.A. & Fujimoto, N.S.V.M. 2004. Morfogênese do relevo do Rio Grande do Sul. In: Verdum, R.; Basso, L.A.; Suertegaray, D.M.A. (orgs.). Rio Grande do Sul: paisagens e territórios em transformação. Ed. UFRGS, Porto Alegre. Pp. 11-26.
• Teixeira, M.B.; Coura Neto, A.B.; Pastore, U. & Rangel Filho, A.L.R. 1986. Vegetação. In: Levantamento dos recursos naturais. Vol. 33. Ed. IBGE, Rio de Janeiro. Pp. 541-632.

Appendix 1

Appendix 4

Publication Dates

• Publication in this collection
  07 Mar 2013
• Date of issue
  Sept 2012

History

• Received
  08 Apr 2011
• Accepted
  02 Oct 2011