Input Characterization of Sedimentary Organic Contaminants and Molecular Markers in the Northwestern Mediterranean Sea by Exploratory Data Analysis

Abstract

Deposition zones of the NW Mediterranean were characterized according to the source of organic pollutants (i.e., UCM, PAHs, PCBs, DDTs) and lipidic compounds (i.e., alkanes and sterols) identified in surface sediments (31 samples) by principal component analysis (PCA) and hierarchical cluster analysis (HCA). Score plots of the two main principal components showed a cluster comprising the off-shore Barcelona and Rhône prodelta samples corresponding to the most polluted samples, while the remaining samples were clustered together. Loading plots revealed that most of the compounds were present in the first component except benzo[ghi]fluoranthene, the major DDT metabolites (i.e., DDE and DDD), and perylene, which was probably of diagenetic origin. In order to define further the cluster containing the most samples, a second data base that excluded the Rhône and offshore Barcelona samples was constructed. Score plot of the two principal components showed that three different depositional environments could be clearly defined, namely the Gulf of Lions, the Ebro prodelta, and the deep sea basin. Similar clustering was confirmed by HCA. The loading plots enabled riverine-transported compounds such as n-alkanes, PCBs, DDTs, sterols, and perylene (first component) to be distinguished from pyrolytic PAHs (second component). Furthermore, in order to obtain an apportionment of the inputs received to each station, a recently developed factor analysis multivariate curve resolution (MCR) method based on the alternating least squares (ALS) positive factorization of a data matrix was carried out for the first time on marine sediment samples. The ALS positive matrix factorization method enabled the apportionment of the environmental source of the main components of the compounds in the area of study.

Source apportionment of submicron organic aerosol at an urban background and a road site in Barcelona (Spain) during SAPUSS

“…MCR-ALS method has been successfully used in the analysis of environmental data sets, both in contamination studies of surface waters (Terrado et al, 2009) and in air source apportionment studies (Salau et al, 1997;Tauler et al, 2009). One of the main advantages of MCR-ALS is that it decomposes the data matrix by applying more natural constraints than PCA, such as non-negativity, and, therefore, interpretation of results is more straightforward.…”

“…PMF integrates natural constraints like non-negativity and uncertainty estimations in a rigorous non-linear optimization of the distribution and composition source profiles. In this work we deploy the multivariate curve resolutionalternating least squares (MCR-ALS) method Tauler, 1995;Jaumot et al, 2005) that has previously been applied for environmental source apportionment (Salau et al, 1997;Terrado et al, 2009). MCR-ALS is based on an alternating linear least squares optimization under nonnegativity constraints which produces physically better profiles than PCA, and it has been shown to produce analogous results to PMF Staminirova et al, 2011).…”

Multivariate Correlation between Concentrations of Selected Herbicides and Derivatives in Outflows from Selected U.S. Midwestern Reservoirs

“…In this paper, a different approach is proposed. The new approach has already been applied to the input characterization of sedimentary, organic chemical markers in the north eastern Mediterranean Sea (12). To limit the number of possible solutions in the matrix decomposition proposed in eq 3 and to find solutions that were more easily interpretable from a physical point of view, the multivariate curve resolution (MCR) method, initially developed for the mixture analysis of spectrometric evolutionary processes (9,10), is proposed.…”

Neural Computation Methods in the Determination of Fungicides

“…Chemometric methods are an important alternative to the purely instrumental ones (Olivieri, 2008), finding a high application of the same in the determination, not only of fungicides but also of pesticides in general, applying one or a combination with the instrumental ones or the neural ones. Among these we point out the MCR-ALS method (multivariate curve resolution-alternating-least-squares) which has been applied to the study of the contamination of sediments and waters and the measurement of air quality (Salau, 1997) and parallel factor analysis or PARAFAC (Bro, 1997). Other methods needing mention are those based on partial least squares approach (PLS) in the setting of environmental analysis (Piccirilli, 2006).…”