Total organic carbon and organic matter quality estimation based on hyperspectral imaging and Rock-Eval analysis

The estimation of Total Organic Carbon (TOC) and the degradation states of organic matter in water-laid sediment are most often characterized via laboratory analyses, whereas infrared spectroscopic methods have proven their effectiveness in this domain for soils. Thus, hyperspectral imaging appears to be a promising technique for the study of TOC in wetlands, lagoons and lakes. Eight sites with different environments (lake, pond, and lagoon), and therefore various origins / degradation states of organic matter and mineralogies were selected for this study. Short Wave Infrared (SWIR) hyperspectral imaging was used to analyze the sediment cores, which were also analyzed via Rock-Eval pyrolysis. Partial Least Squares regression (PLS) was used to model TOC, Hydrogen and Oxygen Indexes (HI and OI) with SWIR spectra. The models for the three parameters are quantitative, with a predictive R²p above 0.7, or even 0.8 for TOC and HI of the parameters. These models are therefore very promising for studying TOC and organic matter (OM) degradation states, but they do have limitations. Indeed, due to the use of external testing sediment cores, the texture and moisture parameters of the cores are essential for obtaining robust predictions, as is the type of OM, as detrital OM does not seem to be modeled owing to low HI values (CH bond absorptions are weak), or the presence of certain minerals at high concentrations, such as carbonate shells. Such approaches open promising prospects for large-scale studies of the carbon cycle in sedimentary environments, particularly for assessing the storage potential of organic matter and monitoring its state of degradation in the context of global change.