Keywords
modelling, chemometrics, principal component analysis, multivariate curve resolution
Start Date
1-7-2004 12:00 AM
Abstract
Environmental monitoring studies produce huge amounts of concentration values of chemicalsspread at distant geographical sites and during different time periods. Moreover, the content of chemicals isalso estimated at different environmental compartments (i.e. air, water, sediments, biota...). All these datavalues are difficult to cope and evaluate in a simple and fast way using simple univariate statistical tools,specially due to their large number and to their multivariate correlation. In order to discover relevant patternswithin large multivariate data sets, the application of modern chemometric methods based in statisticalmultivariate data analysis and in Factor Analysis is proposed. The basic assumption of chemometric methodsis that each of the measured parameter in a particular sample is affected by contributions coming frommultiple independent sources. Each one of these sources is characterized by a particular chemicalcomposition and is distributed among samples in an unknown way. After applying chemometric methods,point and diffuse sources of contaminants in the environment and their origin (natural, anthropogenic,industrial, agricultural...) are identified and their relative distribution among samples (geographical, temporal,among environmental compartments) evaluated. At each sampling site, relative source quantitativeapportionment is estimated allowing a global evaluation of the environmental impact, distribution andevolution of main chemical contamination sources in the environment. In this presentation, differentchemometric methods will be tested on a series of environmental data sets. In particular, the application ofprincipal component analysis and multivariate resolution methods is shown to be a powerful tool for the goalof chemometrics modelling of contamination sources in large environmental data sets acquired in monitoringstudies.
Chemometrics Modelling of Environmental Data
Environmental monitoring studies produce huge amounts of concentration values of chemicalsspread at distant geographical sites and during different time periods. Moreover, the content of chemicals isalso estimated at different environmental compartments (i.e. air, water, sediments, biota...). All these datavalues are difficult to cope and evaluate in a simple and fast way using simple univariate statistical tools,specially due to their large number and to their multivariate correlation. In order to discover relevant patternswithin large multivariate data sets, the application of modern chemometric methods based in statisticalmultivariate data analysis and in Factor Analysis is proposed. The basic assumption of chemometric methodsis that each of the measured parameter in a particular sample is affected by contributions coming frommultiple independent sources. Each one of these sources is characterized by a particular chemicalcomposition and is distributed among samples in an unknown way. After applying chemometric methods,point and diffuse sources of contaminants in the environment and their origin (natural, anthropogenic,industrial, agricultural...) are identified and their relative distribution among samples (geographical, temporal,among environmental compartments) evaluated. At each sampling site, relative source quantitativeapportionment is estimated allowing a global evaluation of the environmental impact, distribution andevolution of main chemical contamination sources in the environment. In this presentation, differentchemometric methods will be tested on a series of environmental data sets. In particular, the application ofprincipal component analysis and multivariate resolution methods is shown to be a powerful tool for the goalof chemometrics modelling of contamination sources in large environmental data sets acquired in monitoringstudies.
