Estuarial Fingerprinting through Multidimensional Fluorescence and Multivariate Analysis

Abstract

As part of a strategy for preventing the introduction of aquatic nuisance species (ANS) to U.S. estuaries, ballast water exchange (BWE) regulations have been imposed. Enforcing these regulations requires a reliable method for determining the port of origin of water in the ballast tanks of ships entering U.S. waters. This study shows that a three-dimensional fluorescence fingerprinting technique, excitation emission matrix (EEM) spectroscopy, holds great promise as a ballast water analysis tool. In our technique, EEMs are analyzed by multivariate classification and curve resolution methods, such as N-way partial least squares Regression-discriminant analysis (NPLS-DA) and parallel factor analysis (PARAFAC). We demonstrate that classification techniques can be used to discriminate among sampling sites less than 10 miles apart, encompassing Boston Harbor and two tributaries in the Mystic River Watershed. To our knowledge, this work is the first to use multivariate analysis to classify water as to location of origin. Furthermore, it is shown that curve resolution can show seasonal features within the multidimensional fluorescence data sets, which correlate with difficulty in classification.

Study of the pH influence on the optical properties of dissolved organic matter using fluorescence excitation–emission matrix and parallel factor analysis

“…Note that the dominant EEM peak location of Peak A observed in this work was in domain III, and Peaks B and C were located in domain V, these peaks can be categorized as fulvic acid-like substances (Peak A) and humic acid-like substances (Peak B and C) according to literature reports [2,21,[32][33][34][35].…”

Cross-connection detection in Australian dual reticulation systems by monitoring inherent fluorescent organic matter

“…Fluorescence EEM spectroscopy has been successfully utilized to monitor natural organic matter in a number of freshwater applications, such as for the characterization and classification of spatially separated rivers and other surface waters [96,97], identification of DOM variability in estuaries [98], and the monitoring of organic fluxes through groundwater [99,100] amongst others. …”

“…These multivariate approaches have been applied to fluorescence-based water research to detect the presence and quantify the underlying fluorescence characteristics of complex mixtures of DOM [98,125]. A number of studies have been carried out on fluorescent DOM with multivariate analysis techniques in marine water [126], estuarine water [96,127,128], fresh water [86,129] glacial systems [130,131] and, more recently, in drinking water [114,132] and sewage treatment [133], including process performance [134] as well as to elucidate changes in fluorescence throughout the treatment trains of a number of water recycling plants [135]. Another approach to the interpretation of EEM data is by fluorescence regional integration (FRI) -a method where the EEM is divided into separate regions as appropriate and the intensity area of each region is integrated for comparison [136].…”

Characterizing dissolved organic matter fluorescence with parallel factor analysis: a tutorial

“…An excellent example is parallel factor analysis (PARAFAC) which can decompose the fluorescence signal into underlying individual fluorescent phenomena (Bro 1997). This is a valuable tool for characterizing and quantifying changes in DOM fluorescence enabling the tracing of different fractions in the natural environment (Cory and McKnight 2005;Hall et al 2005;Stedmon and Markager 2005a,b;Murphy et al 2006).…”