Reconstruction of sediment-transport pathways on a modern microtidal coast by a new grain-size trend analysis method

A new method for granulometric-parameter-based reconstruction of sediment-transport pathways is proposed and is termed P-GSTA (grain-size trend analysis using principal component analysis) herein. The main advantage of this method is its applicability to depositional environments involving mixed transport processes, for instance, fluvial, tidal, and wave-influenced environments. In the P-GSTA method, a linear function with all significant granulometric parameters that are summed with different weighting factors was used to infer sediment-transport direction (sediment flux pattern); the previous grain-size trend analysis (GSTA) methods considered only three parameters (mean grain size, sorting, and skewness) with equal weighting. This study chose six parameters (namely, median grain size, coefficient of variation, skewness, kurtosis, and mud and gravel log-ratios) for calculation. First, the zero values of mud and gravel fractions are replaced, and their log-ratios are defined. Then, all values are standardized. Thereafter, principal component analysis (PCA) is conducted to determine the weighting factor of each granulometric parameter. Each principal component is then interpreted, and the function best representing a sediment flux pattern is chosen from these components. Trend vectors are calculated, solely on the basis of a map interpolated from the scores of the chosen principal component, as the two-dimensional gradient of this value. The P-GSTA method proposed in this study was applied to a modern microtidal coast (tidal sand flat along the Kushida River Delta, central Japan). Sediment-transport pathways reconstructed by this method were consistent with observed sediment-transport patterns determined by field experiments using tracer sediments and geomorphologic observation; the results of the previous GSTA method were inconsistent with the observations. The proposed method also revealed additional minor depositional processes on the sand flat, namely, the deposition of fluvial-channel lags and muddy particles. Thus, this study demonstrates that the proposed P-GSTA method is a potentially powerful tool to reconstruct sediment-transport patterns even under mixed transport processes, where the estimation of the sediment-transport function is difficult.