EM 1110-2-1100 (Part III)
30 Apr 02
normally determined by the mesh size of a sieve that will just allow the grain to pass. This is defined as a
particle's sieve diameter. When performed in a standard manner, sieving provides repeatable results,
although there is some uncertainty about how the size of a sieve opening relates to the physical size of the
particle passing through the opening. (See page 58 of Blatt, Middleton, and Murray (1980) for further
(2) Another way to define a grain's diameter is by its fall velocity. A grain's sedimentation diameter is
the diameter of a sphere having the same density and fall velocity. This definition has the advantage of
relating a grain's diameter to its fluid behavior, which is the usual ultimate reason for needing to determine
the diameter. However, a settling tube analysis is somewhat less reproducible than a sieve analysis, and
testing procedures have not been standardized. Other diameter definitions that have been occasionally used
include the nominal diameter, the diameter of a sphere having the same volume as the particle; and the axial
diameter, the length of one of the grain's principal axes, or some combination of these axes.
(3) For nearly spherical particles, as many sand grains are (but most shell and shell fragments are not),
there is little difference in these definitions. When reporting the results of an analysis, it is always appropriate
to define the diameter (or describe the measurement procedure), particularly if the sieve diameter is not being
(4) Usually, there is a need to characterize an appropriate diameter for an aggregation of particles, rather
than the diameter of a single particle. Even the best-sorted natural sediments have a range of grain sizes.
Most sediment samples found in nature have a few relatively large particles covering a wide range of
diameters and many small particles within a small range of diameters. That is, most natural sediment samples
have a highly skewed distribution, in an absolute sense. However, if size classes are based upon a logarithmic
(power of 2) scale and if the contents of each size fraction are considered by weight (not by the number of
particles), then most typical sediment samples will have a distribution that is near Gaussian (or normal).
When such sediment samples are plotted as a percentage of the total weight of the sample being sieved, the
sediment size distribution that results usually approximates a straight line on a log-normal graph. This line
is known as the log-normal distribution. Meaningful descriptions of the distributions of this type of data can
be made using standard statistical parameters.
(5) Sediment size data normally come from the weight of sediment that accumulates on each sieve in a
nest of graduated sieves. This can be plotted on semilog paper as shown in Figure III-1-1 (ENG form 2087)
or on log-normal paper as shown in Figure III-1-2.
b. Sediment size classifications.
(1) The division of sediment sizes into classes such as cobbles, sand, silt, etc., is arbitrary, and many
schemes have been proposed. However, two classification systems are in general use today by coastal
engineers. Both have been adopted from other fields.
(2) The first is the Modified Wentworth Classification, which is generally used in geologic work.
Geologists have been particularly active in sediment size research because they have long been interested in
interpreting slight differences in size as indicating particular processes or events in the geologic past. A
summary of geological work applicable to coastal engineering is found in Chapter 3 of the book by Blatt,
Middleton, and Murray (2nd ed., 1980).
Coastal Sediment Properties