dc.contributor.author | Kurchaidt, Sonia Maria | pt_BR |
dc.contributor.other | Koehler, Henrique Soares, 1953- | pt_BR |
dc.contributor.other | Watzlawick, Luciano Farinha | pt_BR |
dc.contributor.other | Universidade Federal do Paraná. Setor de Ciências Agrárias. Programa de Pós-Graduação em Engenharia Florestal | pt_BR |
dc.date.accessioned | 2014-12-02T11:52:01Z | |
dc.date.available | 2014-12-02T11:52:01Z | |
dc.date.issued | 2014 | pt_BR |
dc.identifier.uri | http://hdl.handle.net/1884/36802 | |
dc.description | Orientador : Prof. Dr. Henrique Soares Koehler | pt_BR |
dc.description | Co-orientador : Prof. Dr. Luciano F. Watzlawick | pt_BR |
dc.description | Tese (doutorado) - Universidade Federal do Paraná, Setor de Ciências Agrárias, Programa de Pós-Graduação em Engenharia Florestal. Defesa: Curitiba, 06/06/2014 | pt_BR |
dc.description | Inclui referências | pt_BR |
dc.description | Área de concentração : Manejo florestal | pt_BR |
dc.description.abstract | Resumo: Em estudos de ecologia, inúmeras são as variáveis que afetam o estabelecimento,
crescimento e permanência de espécies arbóreas. A análise multivariada de dados,
sendo um conjunto de ferramentas de redução, classificação ou simplesmente
utilizada como técnica exploratória torna-se útil e pertinente nestas situações. Este
trabalho foi realizado com dados de quatro parcelas permanentes coletadas
anualmente pelo "Programa Ecológico de Longa Duração" (PELD), localizado em um
fragmento de Floresta Ombrófila Mista em São João do Triunfo, PR. Teve como
objetivo principal aplicar técnicas multivariadas neste fragmento de Floresta, nos
anos de 1995 a 2010, com o intuito de verificar o comportamento da floresta no
período. Foi ajustada para cada uma das quatro parcelas, em cada um dos anos
estudados, uma curva de acumulação espécie-área segundo um polinômio de 2º
grau e testadas quanto a sua identidade. Os resultados revelaram que as equações
diferem estatisticamente entre si e que não é possível utilizar uma única equação
para representar a curva espécie/área no fragmento estudado. Foram verificadas se
as intensidades amostrais empregadas (n=350) foram suficientes nos anos de 1995
e 2010, e todos os resultados obtidos mostraram que a intensidade efetuada foi
maior do que a recomendada. Para os dados de ambos os anos foram obtidas as
estimativas dos parâmetros fitossociológicos Densidade, Frequência, Dominância,
Valor de Importância e Valor de Cobertura. Por meio da análise fatorial verificou-se
que nenhum dos fatores, isolados ou em conjunto, destacou-se, uma vez que todos
apresentaram altas cargas nos fatores. Foram identificadas as dez espécies que
obtiveram maior valor de importância, nas quatro parcelas, com o objetivo de
verificar se existem diferenças entre elas, por meio da utilização da MANOVA.
Detectada a diferença entre as parcelas, a análise discriminante realizada
selecionou três combinações lineares, para ambos os anos. Foram selecionadas as
20 espécies com maior Valor de Importância para cada uma das parcelas e foi
realizada uma análise de agrupamentos, visando encontrar grupos similares,
obtendo-se em cada um dos anos três grupos. Com os grupos formados foi
realizada uma análise discriminante, que forneceu, para os anos em estudo, duas
funções discriminantes, verificando-se que no ano de 1995, tanto na primeira como
na segunda função discriminante, o grupo dois discrimina os demais, enquanto que
para o ano de 2010, em ambas as funções o grupo um é que discrimina os demais.
O grupo dois no ano de 1995 é formado predominantemente por subunidades
pertencentes à parcela Rio, e para o ano de 2010, no grupo um predominam
subunidades da parcela Araucária. A análise discriminante apontou as espécies
Araucaria angustifolia, Nectandra grandiflora, Ilex paraguariensis e Casearia
decandra como espécies predominantes na área estudada. Pode-se afirmar, com
base nos resultados obtidos, que as parcelas Araucária, Rio, Fogo e Imbuia são
pertencentes a uma única tipologia florestal, caracterizada como Floresta Ombrófila
Mista.
Palavras chave: Floresta Ombrófila Mista, componentes principais, Manova, análise
de agrupamentos; análise discriminante. | pt_BR |
dc.description.abstract | Abstract: Ecological studies usually involve several variables which affect establishment,
growth and permanence of three species. Multivariate analysis, used as a tool to
reduce, classify or explore the data is very useful in these situations. The study was
conducted using four permanent plots collected annually from the forest inventory of
the "Long Term Ecological Program" (PELD), in São João do Triunfo Experimental
Station of the Federal University of Paraná. The main objective was to apply
multivariate techniques in a fragment of Araucaria forest, in 1995 and 2010, in order
to verify the behavior and dynamics of the forest in the period. Initially, a second
degree polynomial function was fitted to the species-area accumulation data, to each
of the four plots, in each of the years studied. The analysis showed that the
equations were statistically different and could not be replaced by a single equation
to represent the type/area curve in the whole fragment. It was found that the sample
intensities employed (n = 350) were sufficient in both years, with the results showing
a sampling size higher than the recommended. For the data from both years, it was
obtained estimates of the phytosociological parameters Density, Frequency,
Dominance, Importance Value and Coverage Value. Factor Analysis showed that
none of the factors, alone or combined, stood out, since all presented high loading
factors. The ten species with the greatest amount of importance, in all four plots,
were identified with the objective of determining if there were differences among the
plots, through the use of multivariate analysis of variance. Once differences among
the plots were detected, a discriminant analysis selected three linear combinations of
species, for both years. Three discrimination functions were found. For both years,
the 20 species with the highest importance value were selected from each of the
plots and a cluster analysis was performed, aiming to find similar groups among
them. Three groups were founded, in each of the years. Using the groups found by
the discriminant analysis, that provided two discriminant functions, for the both years
under study, it was found that in 1995, both the first and the second groups of
subunits is discriminating the others, while for the year 2010, in both functions the
group one is discriminating against others. Group two in 1995, is predominantly
formed by subunits belonging to the Rio plot, and for the year 2010, in group one
predominates the subunits from the Araucaria plot. The discriminant analysis showed
the species Araucaria angustifolia, Nectandra grandiflora, Ilex paraguariensis and
Casearia decandra as the predominant species in the studied area. It can be stated,
based on the results obtained, that the Araucaria, Rio, Fogo and Imbuia plots
belongs to a single forest type, characterized as Araucaria Forest .
Key words: Araucaria Forest , principal components, Manova, cluster analysis,
discriminant analysis. | pt_BR |
dc.format.extent | 106f. : il. algumas color.,grafs., tabs. | pt_BR |
dc.format.mimetype | application/pdf | pt_BR |
dc.language | Português | pt_BR |
dc.relation | Disponível em formato digital | pt_BR |
dc.subject | Teses | pt_BR |
dc.title | Uso de métodos estatíticos multivariados em fitossociologia florestal | pt_BR |
dc.type | Tese | pt_BR |