Interpretation of sedimentary subpopulations extracted from grain size distributions in loess deposits at the Sea of Azov, Russia

Loess-palaeosol sequences in Eastern Europe, especially those in the Azov region, are among the most sensitive terrestrial archives for identification of past aeolian dynamics and Quaternary palaeoclimatic reconstruction. Grain size analyses of loess sediments are widely used to interpret these transport mechanisms and palaeoclimatic changes based on granulometric parameters and statistical decomposition methods. This topic is therefore of growing interest in Earth sciences and has been a major focus of sedimentary studies. Here, we present the results of unmixing grain size distributions from a loess-palaeosol section at the Sea of Azov, Russia, by jointly applying the standard deviation method and end-member modelling. The results indicate that the two methods can produce similar grain size decompositions but that end-member modelling has advantages in terms of quantitative and objective characteristics. In addition, three main loess subpopulations or end-members (EMs) with mode sizes of 8 mu m, 18 mu m and 32 mu m, which represent distinct aerodynamic environments, are identified from the grain size distribution in the Azov region. Thereinto, EM1 with a mode size of 8 mu m is the integrated result of combining atmospheric circulation with other environmental processes. EM2 with a mode size of 18 mu m is inferred to represent continuous background dust under non-dust storm conditions. EM3 with a mode size of 32 mu m is a fraction transported in short-term, low-altitude suspension clouds during dust storm outbreaks. Of the three EMs, EM1 and EM2 have multiple origins due to their complex formations, whereas EM3 is primarily derived from the alluvial plains of different rivers flowing into the Sea of Azov.