Publicação
Spatial modelling of factor analysis scores : the case of heavy metal biomonitoring in mainland Portugal
| Resumo: | The use of mosses as biomonitors operates as an indicator of their concentration in the environment, becoming a methodology which provides a significant interpretation in terms of environmental quality. The different types of pollution are variables that can not be measured directly in the environment - latent variables. Therefore, we propose the use of factor analysis to estimate these variables in order to use them for spatial modelling. On the contrary, the main aim of the commonly used principal components analysis method is to explain the variability of observed variables and it does not permit to explicitly identify the different types of environmental contamination. We propose to model the concentration of each heavy metal as a linear combination of its main sources of pollution, similar to the case of multiple regression where these latent variables are identified as covariates, though these not being observed. Moreover, through the use of geostatistical methodologies, we suggest to obtain maps of predicted values for the different sources of pollution. With this, we summarize the information acquired from the concentration measurements of the various heavy metals, and make possible to easily determine the locations that suffer from a particular source of pollution. |
|---|---|
| Autores principais: | Piairo, Helena |
| Outros Autores: | Menezes, Raquel; Sousa, Inês; Figueira, Rui; Sérgio, Cecilia |
| Assunto: | Biomonitors Heavy metals Sources of pollution Latent variables Factor analysis Geostatistical methodologies |
| Ano: | 2014 |
| País: | Portugal |
| Tipo de documento: | artigo |
| Tipo de acesso: | acesso restrito |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | The use of mosses as biomonitors operates as an indicator of their concentration in the environment, becoming a methodology which provides a significant interpretation in terms of environmental quality. The different types of pollution are variables that can not be measured directly in the environment - latent variables. Therefore, we propose the use of factor analysis to estimate these variables in order to use them for spatial modelling. On the contrary, the main aim of the commonly used principal components analysis method is to explain the variability of observed variables and it does not permit to explicitly identify the different types of environmental contamination. We propose to model the concentration of each heavy metal as a linear combination of its main sources of pollution, similar to the case of multiple regression where these latent variables are identified as covariates, though these not being observed. Moreover, through the use of geostatistical methodologies, we suggest to obtain maps of predicted values for the different sources of pollution. With this, we summarize the information acquired from the concentration measurements of the various heavy metals, and make possible to easily determine the locations that suffer from a particular source of pollution. |
|---|