Determination of the nutritional state of soils by means of the phytosociological method and different statistical techniques (Bayesian statistics and decision trees) in Spain

• Autor(a) principal: Cano-Ortiz, Ana
• Data de Publicação: 2009
• Outros Autores: Pinto-Gomes, Carlos, Esteban, Francisco, Cano, Eusébio
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10174/6853
• topic: pasturelands; association; statistics; networks; artificial intelligence

The paper deals with the nitrophilous Hordeum leporinum pasturelands peculiar to the olive groves of Jaén, Spain. The study first selects physiognomically homogeneous land plots and subsequently samples are taken where the relevant species is dominant. Simultaneously, an edaphic sample is taken and the different edaphic parameters are analysed. By means of phytosociological and statistical methods (cluster analysis) and co-location networks, we describe a new association, namely Anacyclo clavati-Hordeetum leporinum. With the aid of advanced statistical techniques and Bayesian networks, we carry out different analyses in order to find out plant-soil correlations, in particular, for Filago lutescens and Hordeum leporinum. The paper also provides a decision tree for Hordeum leporinum. In this decision tree the values of the different edaphic parameters are checked against the values of the abundance-dominance index.