DIFFERENCE IN GRAIN-SIZE PARAMETERS OF TIDAL DEPOSITS DERIVED FORM THE GRAPHIC AND ITS POTENTIAL CAUSES

Grain-size parameters of 395 samples from separately sandy or muddy layers of the intertidal-flat deposits in the Qiangtang Estuary were calculated using Fork-Ward graphic method (GM), and moment methods of Friedman-Johnson (MM_Fr) and McManus (MM_Mc), respectively. Comparative studies indicate that the parametric relationships are quite complex among the different methods especially for the higher order moments, only with an exception of mean size, in that GM mean size almost equates to that of MM. Physical meaning of MM_Mc skewness and kurtosis has never been well expressed due to its non-traditional statistical methodology, so unique application of MM_Fr formula is strongly recommended to calculate moment parameters for environmental interpretations and comparison. The parametric difference is notable between sandy and muddy layers, which are composed of separate coarser and finer dynamic populations in response to their different depositional processes on the basis of the numerical partitioning analyses (inverse modeling). It is therefore extrapolated that the sampling unit for grain size analyses should be strictly deposited under similar hydrodynamic conditions. A numerical modeling of the mixtures of two log-normal populations (forward modeling) was successfully applied to simulate the complex relationships of GM and MM parameters in terms of skewness and kurtosis, which are majorly controlled by the difference of percentiles and modes between the major and secondary populations. As the finer (secondary) population percentiles decrease, the value of MM skewness and kurtosis increases sensitively to the detail change in grain-size distribution pattern; while the value of graphic skewness and kurtosis increases before reaching their maxima and decreases after those critical points, mainly resulting from finite statistics of graphic method on a few eigenvalues and neglecting the tail (<5%) components.The critical value for graphic skewness to change from increasing into decreasing trend is 35% for the secondary population percentiles, hypothetically related with an additional expression in the grahic skewness formula to stress the percentiles (16 and 84) on the grain-size distribution, which are not included in the graphic kurtosis formula. The both methods have their own advantage and disadvantage, but the moment method has a priority in establishing a uniform standard in the future, typically for physical interpretation of grain-size parameters, considering that it can elaborately and consistently reflect changes in secondary population tail features in comparison with the failure of the graphic method.