Do hyperspectral transmissometers allow us to go one step further on the analysis of particulate matter characteristics of marine samples?

Several studies have been carried out to investigate the correlation between the spectral shape features of the beam attenuation coefficient and the particulate matter characteristics in seawater, but little attention has been paid to the spectral resolution of these measurements. For this reason, the potential of the new hyperspectral transmissometer VIPER (TriOS GmbH), with 1.7 nm spectral resolution, has been evaluated in this study and compared with lower resolution and multispectral based approaches (e.g. ac-9 or ac-s-with 4 nm resolution - from WETLabs Inc.) in order to evaluate whether any additional information about water composition can be retrieved from a spectral shape-based assessment. In this way, this study proposes a statistical-based method - a Hierarchical Cluster Analysis (HCA) using the cosine distance as similarity value - which allows discriminating suspended sediment samples with different particle size distribution (PSD) based on the attenuation spectral shape features. Finally, the effects of both particle size and concentration on the spectral shape have been analyzed separately. The results confirmed that the beam attenuation spectral features are in first-order driven by particle concentration, which means that a prior knowledge of particulate matter concentration is required in order to classify sediment samples according to their particle size. This approach based on hyperspectral attenuation measurements to characterize the PSD has been demonstrated a potential alternative compared to the traditional methods such as Coulter Counter or the particle size analyzer LISST 100X, which are much more expensive and time-consuming approaches.