Application of Empirical Orthogonal Function (EOF) method to detect the spatial grain-size fractionation of aeolian sediment and comparison with other methods in case studies from northwestern China

A grain-size fraction/population of sediment can reflect a specific transport process. Use of this potential grain-size fraction is more meaningful than that of a single grain-size value for analysis of grain-size fractionation over space, so identifying the potential grain-size fraction is key in analyzes of grain-size fractionation. Our aim was to test the applicability of Empirical Orthogonal Function (EOF) analysis for sediment grain-size fractionation. We performed EOF analysis on grain-size data of aeolian sediment to identify the potential grain-size fraction and detect its grain-size fractionation over space. Two other methods, the grain-size-class standard deviation (GSC-std) method and end-member analysis (EMA) were conducted for comparison. The results show that the EOF method is not only able to identify the potential grain-size fraction, but also to simultaneously determine the spatial change of the potential grain-size fraction. The GSC-std method provides the potential grain-size fraction, but fails to exhibit its spatial change at the same time. EMA end members resemble a grain-size distribution that differs from the potential grain-size fraction; thus, the EMA method fails to identify the potential grain-size fraction directly. Our results suggest that EOF analysis is suitable for detecting grain-size fractionation over space.