干旱区生态环境知识资源中心

Loess-paleosol sequences (LPS) are sensitive terrestrial archives of past aeolian dynamics and paleoclimatic changes within the Quaternary. Grain size (GS) analysis is commonly used to interpret aeolian dynamics and climate influences on LPS, based on granulometric parameters such as specific GS classes, ratios of GS classes and statistical manipulation of GS data. However, the GS distribution of a loess sample is not solely a function of aeolian dynamics; rather complex polygenetic depositional and post-depositional processes must be taken into account.

This study assesses the reliability of fixed GS ratios as proxies for past sedimentation dynamics using the case study of Remizovka in southeast Kazakhstan. Continuous sampling of the upper 8 m of the profile, which shows extremely weak pedogenic alteration and is therefore dominated by primary aeolian activity, indicates that fixed GS ratios do not adequately serve as proxies for loess sedimentation dynamics. We find through the calculation of single value parameters, that "true" variations within sensitive GS classes are masked by relative changes of the more frequent classes. Heatmap signatures provide the visualization of GS variability within LPS without significant data loss within the measured classes of a sample, or across all measured samples. We also examine the effect of two different commonly used laser diffraction devices on GS ratio calculation by duplicate measurements, the Beckman Coulter (LS13320) and a Malvern Mastersizer Hydro (MM2000), as well as the applicability and significance of the so-called "twin peak ratio" previously developed on samples from the same section. The LS13320 provides higher resolution results than the MM2000, nevertheless the GS ratios related to variations in the silt-sized fraction were comparable. However, we could not detect a twin peak within the coarse silt as detected in the original study using the same device. Our GS measurements differ from previous works at Remizovka in several instances, calling into question the interpretation of paleoclimatic implications using GS data alone.