Improving the quantification of sediment source contributions using different mathematical models and spectral preprocessing techniques for individual or combined spectra of ultraviolet–visible, near- and middle-infrared spectroscopy

Abstract

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.Improving the quantification of sediment source contributions using different mathematical models and spectral preprocessing techniques for individual or combined spectra of ultraviolet-visible, near-and middle-infrared spectroscopy

Terrain analysis, erosion simulations, and sediment fingerprinting: a case study assessing the erosion sensitivity of agricultural catchments in the border of the volcanic plateau of Southern Brazil

“…The origin of riverbed sediment was determined by a fingerprinting method based on NIR spectroscopy (Verheyen et al 2014;Tiecher et al 2016;Tiecher et al 2021) and modeling with Support Vector Machine (SVM) to build spectroscopic models of sediment estimation (Tiecher et al 2021). Two source groupings (by land use and by sub catchment) were considered as potential end-members to analyze the spatial variability and the main erosion processes to explain the sediment yield observed at the outlets.…”

“…The kernel function seeks to establish correlations between the reflectance value and the target variable, in which the model seeks to identify a interpolation function between the variables and creates support vectors, a robust procedure in statistical learning models (Ivanciuc 2007). This type of model was selected due to the occurrence of non-linear correlations between the organo-mineral components of soils/sediments and the spectral variables (Viscarra Rossel and Behrens 2010) and given the acquisition of more accurate estimates using this model in comparison, for example, to parametric models Partial Least Square Regression (PLSR), by Tiecher et al (2021).…”

“…There is an interest in understanding erosion processes to propose better management practices. These studies require long term monitoring data (Silva et al 2021), which can be costly to acquire. Yet, there are tools to rapidly analyze an environment's fragility.…”

“…This approach provides a quantification of source contributions to sediment through the analysis of conservative bio-physico-chemical properties in potential source and target material. For this purpose, cost-and time-intensive conventional tracing properties (e.g., radionuclides, elemental and isotopic geochemistry), or simple and alternative methods as those relying on near-infrared spectroscopy coupled to statistical modelling were shown to be efficient in estimating land use source contributions to sediment yield (Verheyen et al 2014;Tiecher et al 2021).…”

Salinity Monitoring at Saline Sites with Visible–Near-Infrared Spectral Data

“…[24]), baseline correction (first-order derivative, continuous wavelet transform, etc. [25]), scattering correction (multiplicative scattering correction (MSC), standard normal transform, etc. [23,26]), and smoothing (Savitzky-Golay (SG) smoothing, etc.…”