Chemometrics applied to environmental systems

Abstract

To understand the processes that give rise to the observed concentrations of chemical species in environmental samples, it is often useful to apply chemometric methods to help convert chemical data into environmental information. In this review, various chemometical methods including pattern recognition, mixture resolution methods, and probability estimation will be described, and their applications to a variety of environmental systems are presented. However, there are important differences between applying chemometrical methods in the laboratory as compared to environmental data. The environmental system can be monitored, but in general, it is not possible to plan experiments and use known distributional properties of the system under study. Given the complexity of these systems, large data sets are typically needed to provide the basis for obtaining a good understanding of its functioning. Expanding computer capabilities coupled with improved analytical methods have allowed the production and analysis of larger and larger data sets with more and more computationally intensive sophisticated algorithms that offer the promise of even greater information retrieval in the future.