Fluorescence indices of organic matter in groundwater and other aquatic environments : proposal for improvement

Fluorescence indices are particularly well adapted to characterize dissolved organic matter. In environmental studies, the three most often used fluorescence indices are the Fluorescence index (FI), the Humification index (HIX) and the Biological index (BIX). They are based on ratios of specific emission intensities to excitation wavelengths from 2D spectra that are assumed to correspond to specific families of organic compounds. Three main remarks can be made regarding the use of these indices. First, they are often applied in aquatic environments different from those where they were developed. However, the specific emission wavelengths used to calculate fluorescence indices are based on the location of the maximal intensity of fluorescent compounds families which is likely to be blue-shifted (shifted towards shorter emission wavelength) along the water cycle (from soil, surface, ground, to marine waters). Second, they are defined based on 2D spectras, while excitation-emission matrices (EEM - 3D matrices) are becoming more widespread and allow, thanks to PARAFAC modeling, the identification of specific maximal fluorescence intensities related to fluorescent compounds families. EEM may thus have the potential to better assess the relative concentrations of organic matter compounds families, and thus the ratios between these compounds. Third, these indices were developed based on a conceptual model of organic matter decomposition, which has now been reconsidered. The present study addresses the above-mentioned methodological issues. A list of points to pay particular attention to when calculating fluorescence indices is proposed, highlighting the need for knowledge of fluorescent compounds sources in the water samples studied. We propose to update fluorescence indices calculation by identifying specific ranges of excitation and emission wavelengths for each aquatic environment (freshwaters, soil and marine waters), or using fluorescence indices derived from PARAFAC components. The results of this study warn against unsupervised calculation and interpretation of most usual fluorescence indices. It also greatly encourages the rethinking of fluorescence indices, particularly by considering the PARAFAC components loadings or weights instead of peaks.