ABSTRACT Although mean size provides a useful value for characterizing sediments, it is an inadequate expression of the interaction between sediments and hydraulic regimes and many authors have explored the significance of other statistical moment measures in different parts of the particle size spectrum. Few publications, however, have described such relationships over a very wide range of particle sizes. Using a combined data base (of marine and lacustrine sediments) we have been able to demonstrate the following:

The distributions of standard deviation (and entropy) values are repeated in similar form on either side of the sand size class, describing mixtures between sands and gravel and sands and clay.

Skewness is modified by the hydraulic response of the sediment and a skewness divide is present at about 2.7 φ, where it coincides with the minimum shear velocity required to initiate particle motion. This divide may be used to distinguish between high and low energy regimes.

Kurtosis, like standard deviation, is a measure of the mixing of end member populations but it is more sensitive to changes in the tails of distribution curves. The maximum positive peak of kurtosis, at about 2.5 φ, lies close to the skewness divide (in our data).

By using skewness/kurtosis plots it is possible to define separate field distribution for sediments which, to a greater or lesser degree, appear to conform to hydraulic equilibrium.

By using sediments which are close to equilibrium (and lie at or close to our skewness/kurtosis boundary curves, it should be possible to define hydraulic shear velocities, at the time of formation. Values may be established by empirical relationships or by defining the break-point (traction and intermittent suspension) between linear components of individual sample φ size/probability distribution plots. Such relationships should be enhanced by using hydraulic equivalent diameters instead of sieve diameters.

In fine sediments, the use of fractional (silt: clay) ratios offers sensitivity at a level comparable to that of skewness/kurtosis plots and may be more easily computed.