BIBLIOS

Tipo
Documentos Académicos

Tipo de Documento
Dissertação de Mestrado

Título
Análise estatística Multivariada aplicada a dados hidrogeológicos

Participantes na publicação
Gomes, Vânia Sofia Pires Simões (Author)
Diamantino, Maria Fernanda Nunes (Supervisor)
Dep. Estatística e Investigação Operacional
Silva, Maria Catarina Rosalino da (Co-Supervisor)
Dep. Geologia
Dep. Geologia
IDL

Resumo
For the present study we used chemical analyzes of 48 groundwater samples of Lisbon in which was selected 13 variables related to dissolved species (anions and cations) and physico-chemical parameters. This study aims to evaluate whether there is a relationship between the chemical composition of the groundwater and the geological formations where it circulated, this is, it was intended to identify the dissolved species that characterize each geological formation, as well as to identify what distinguishes each geological formations. Thus, initially we proceeded with an exploratory analysis in order to characterize the sample and partial samples (corresponding to each geological formation), and in a final stage, we applied a multivariate analysis, where techniques were used to reduce the dimensionality (principal component analysis and factor analysis) and techniques to data clustering (cluster analysis). It was concluded that the samples from the Formação de Benfica and Complexo Vulcânico de Lisboa had identical chemical composition. For the other geological formations, there was no homogeneity among samples from the same formation, or between samples from different formations. Therefore, the characterization of the geologic formations has become very complex. Since the values of the Pearson correlations among the variables were low, the application of reduce the dimensionality techniques do not have the desired effect.


Instituição
FACULDADE DE CIÊNCIAS DA UNIVERSIDADE DE LISBOA

Identificadores da Publicação


Identificadores do Documento

Keywords
Água subterrânea Análise exploratória Análise em componentes principais Análise fatorial Análise de clusters Teses de mestrado - 2013