Critical evaluation of spectroscopic indices for organic matter source tracing via end member mixing analysis based on two contrasting sources

• Source tracing capabilities of spectroscopic indices are critically evaluated.
• Different behaviors (linear or nonlinear) of indices with mixing are identified.
• Solution chemistry effects are systematically demonstrated for well known indices.
• An insight into field data interpretation in variable environments is provided.

Despite the wide use of absorption and fluorescence spectroscopy for tracking the sources of dissolved organic matter (DOM), there are limited studies on evaluating their source discrimination capabilities at variable solution chemistry (pH, NaCl, Ca2+, and DOM concentration). For this study, we compared the applicability of several well-known spectroscopic indices via end member mixing analysis based on two contrasting DOM sources (Suwannee River fulvic acid and an algal DOM). The absorption coefficients and the intensities of fluorescent components from parallel factor analysis (PARAFAC) showed linear relationships with increasing algal carbon fraction in the mixture of the two DOMs. In contrast, although they still behaved conservatively, spectral ratio indices such as spectral slopes, ratios of PARAFAC components, humification index, and fluorescence index changed in nonlinear patterns with the mixing ratios. The indices based on PARAFAC results exhibited strong discrimination capabilities, as indicated by high susceptibility to the changes in DOM sources relative to the analytical precision. While variable NaCl concentrations had limited effects, most fluorescence indices were considerably affected by other solution chemistry such as pH, Ca2+, and DOM level. Our study demonstrated that the applicability of the source discrimination indices should be critically examined especially in the environments with notable changes in the solution chemistry. The solution chemistry effects could be minimized by adjusting samples to a constant condition prior to the measurements or otherwise the effects should be fully taken into account in interpreting the field observations.

Figure optionsDownload high-quality image (178 K)Download as PowerPoint slide