Geology Reference
In-Depth Information
the C-M diagram typify deposits from traction currents,
from beaches or from various forms of suspension
(Rizzini 1968; Fabri et al. 1986). Different environ-
ments can be distinguished by examining the patterns
produced by plotting a number of samples from the
same unit on C-M or L-M diagrams (Fig. 6.6). The
Passega diagram
allows the type of sediment trans-
port to be differentiated preferentially for sediments
with C < 1 mm (Gromoll 1992). The diagram has been
successfully used both in the interpretation of modern
sediments and ancient sedimentary rocks (Faugères
1974).
Other authors have suggested that the depositional
environments and different transport dynamics can be
interpreted on the basis of the
shapes of grain-size cu-
mulative curves
plotted on probability paper (Visher
1969; Glaister and Nelson 1974). Several
multivariate
statistical techniques
(factor analysis, discriminant
function analysis, log-hyperbolic distribution, entropy
analysis) have been successfully applied to the inter-
pretation of grain-size distribution (Klovan 1967;
Vincent 1986; Forrest and Clark 1989; Merta 1991).
However, contrary to the information provided by ecol-
ogy or sedimentary structures, the grain-size evidence
provided has the advantage of being universal.
6.1.1.3 GrainSize Analysis in Thin Sections
In sandstones and limestones grain sizes are usually
measured in thin sections, peels or microphotographs.
Rocks should be sectioned in a consistent direction.
Sections parallel to the bedding plane would be more
appropriate but most students use vertical or non-ori-
ented thin sections. Peels can be made of much larger
surfaces than conventional thin sections, allowing a
large amount of grains to be measured (Gutteridge
1985). Grain size analysis of thin sections requires sta-
tistical grain selection according to a well-defined sam-
pling plan. Techniques used include point-, line-,
area-, and ribbon-counting (Sect. 6.2.1). Point count-
ing (grains selected at grid points) appears to be the
most appropriate method for sandstones, ribbon-count-
ing as well as point-counting are commonly used for
limestones. The number of grains measured ranges be-
tween 150 and 500. Measurement of 300 grains is of-
ten regarded as statistically sufficent for sandstones.
Calcarenitic and calciruditic carbonates must be handled
differently (Sect. 6.2).
The value of nearly all these techniques for identi-
fying depositional environments has been questioned
in parts or as a whole (e.g. Vandenberghe 1975; Tucker
and Vacher 1980; Sengupta et al. 1991). The main rea-
sons for these difficulties are related to the variability
of depositional conditions within the major environ-
mental settings and the complexity of depositional pro-
cesses:
•
Measurement errors:
Grain sizes derived from thin
sections are apparent, not true grain sizes. The prob-
ability that a grain will be cut by a random section is
proportional to the size of the grain. Diameters mea-
sured in thin sections, therefore are usually smaller than
the true grain diameter. Random sections lead to an in-
crease of measurements of small grains.
Hydrodynamic conditions within the sedimentation
area can considerably vary, depending on the size of
the sedimentary basin.
•
Spatial and temporal grain size distribution within a
given physiographic setting are highly variable, and
demand critical statistical analysis (Piller and
Mansour 1990, Medina et al. 1994).
Conversion of grain-size data:
Basic concepts of
grain-size analysis were developed by studying loose
sediments or disintegrated sedimentary rocks. Because
the commonly used sieve method can not be applied to
indurated rocks, a conversion of granulometric data ob-
tained by thin section analysis into sieve data appears
necessary if statistical parameters are also to be evalu-
ated for sandstones and carbonate rocks. This conver-
sion has been based on different regression models
(Friedman 1958, 1962; Harrell and Erikson 1979; Merta
1991).
Müller (1964) developed a nomogram from which
the sieve-size distributions can be directly derived. A
comparison of sieve and thin-section methods used for
siliciclastic rocks indicates that the two techniques must
be regarded as mathematically independent methods
(Burger and Skala 1979).
•
Frequency and cumulative curves reflect grain-size
subpopulations, distinguished by rolling, saltation
or suspension transport (Smolka 1990).
•
Primary grain-size distributions can be changed by
bioturbation and episodic high-energy events, e.g.
strong storms, as well as by variations in sediment
influx caused by sea-level fluctuations.
In summary, grain-size distributions primarily re-
flect transport and depositional processes which are not
necessarily unique to a particular environment. Never-
theless, in the context of facies analyses, grain-size data
assist in environmental interpretation if the data are in-
corporated in the framework of textural and composi-
tional criteria and all lines of evidence are regarded.
