Automated counting of sand sized particles in marine records

• Published in: Sedimentology
• Main Authors: Becker, Lukas W. M., Hjelstuen, Berit O., Storen, Eivind W. N., Sejrup, Hans Petter
• Format: Text
• Language: English
• Published: Wiley
• Subjects: envir; geo; North East Atlantic
• Online Access: https://doi.org/10.1111/sed.12407; https://archimer.ifremer.fr/doc/00396/50700/90024.pdf; https://archimer.ifremer.fr/doc/00396/50700/

Content and fluxes of ice transported sand-sized and gravel-sized, lithic particles in marine sediment cores are a common tool used to reconstruct glacial variability. Ice-rafted debris datasets are currently acquired in several different and often time consuming ways, and within various grain size fractions. This paper proposes a novel workflow using an automated method to count ice-rafted debris to reduce analysis time and subjectivity. The described method is based on the instrument ‘Morphologi G3’ from Malvern Instruments Limited, and includes all pre-processing and post-processing steps. This particle characterization tool is an automated microscope combined with a proprietary software package. In the present study, the analysis was performed on the 150 to 1000 μm fraction. If desired, grain counts can be carried out on the entire sand and silt fractions, but then at the expense of considerably greater turnover time. A total of 350 sediment samples from core MD99-2283, taken on the upper continental slope at the southern part of the north-east Atlantic margin, were counted with this automated method. In addition, a total of 161 samples were counted manually as a control on the reliability of the scanning. The comparison of automated versus manually counted biogenic and lithic material shows a convincing correlation between both methods. The turnover time per automatically counted sample is around 20 minutes, the method requiring less experience and skills than manual counting. The results yield a promising, time saving new technique to achieve high-resolution ice-rafted debris counting records with acceptable error margins.