Nature, origin and transfers of mineral, organic, and biological particles : a new methodological approach of SPM transfers analysis in hydrosystems by morphogranulometry

International audience; Source to sink is one of the main concepts in Earth Sciences for a better knowledge of hydrosystems dynamics. Regarding this issue, the present day challenge consists in the characterization by in-situ measurements of the nature and the origin of suspended particles matters (SPM). Few methods can fully cover such requirements and among them, the methodology using the form of particles deserves to be developed. Indeed, morphometry of particles is widely used in sedimentology to identify different sedimentary stocks, source-to-sink transport and sedimentation mechanisms. Currently, morphometry analyses are carried out by scanning electron microscope coupled to image analysis to measure various size and shape descriptors on particles like flatness, elongation, circularity, sphericity, bluntness, fractal dimension. However, complexity and time of analysis are the main limitations of this technique for a long-term monitoring of SPM transfers. Here we present an experimental morphometric approach using a morphogranulometer (a CCD camera coupled to a peristaltic pump). The camera takes pictures while the sample is circulating through a flow cell, leading to the analysis of numerous particles in a short time. The image analysis provides size and shape information discriminating various particles stocks according to their nature and origin by statistical analyses. Measurements were carried out on standard samples of particles commonly found in natural waters. The size and morphological distributions of the different mineral fractions (clay, sand, oxides etc), biologic (microalgae, pollen, etc) and organic (peat, coal, soil organic matter, etc) samples are statistically independent and can be discriminated on a 4D graph. Next step will be on field in situ measurements in a sink-spring network to understand the transfers of the particles stocks inside this simple karstic network. Such a development would be promising for the characterisation of natural hydrosystems.