Growth of Present-Day Concepts, 1978-1988
'
E
2
05
Grom dmmeter ,d {mm)
a basis for subsurface environmental interpretations
(e.g., Pirson 1977; Berg 1986). Commonly, these were
identifi ed not on the basis oflithologic logs in cores,
but fr om the profiles displayed by electric logs - the
classic "funnel-shaped" and "bell-shaped" gammaray, spontaneous potential, and resistivity logs that
had first been identified during the 1950s (e.g., Nanz
1954).
The value of a facies model in providing a synthesis of a diverse array of fa cts is offs et by the danger
that uncritical use of the model may lead to a loss of
information or to misinterpretation, because it is
tempting to observe strata in terms of a preconceived model. Objective field measurements commonly are difficult to make, particularly in areas of
poor exposure. One of the trends in sedimentological research in the 1970s was t'o quantify observations of lithological successions in order to enable
model sequences or cycles to be defined statistically.
Markov chain analysis was commonly used for
this purpose (e.g., Allen 1970b; Miall 1973, 1977;
Cant and Walker 1976). Although this approach allows more precision in defining the model, it can
cause valuable observational detail to be ignored.
Collinson (1978, p. 579) made this point about the
Battery Point Formation, a highly varied, low-sinuosity-fluvial deposit from which Cant and Walker
(1976) extracted a single "summary" cyclic sequence
(Fig. 2.24a).
39
Fig. 2.23. Bedform phase diagram showing how
bedforms may be classified according to the physical
processes which control their size limits and stability.
(Leeder 1983)
Amongst the problems that had been recognized
with the Markov chain method was the fa ct that it
contained no information on the nature of facies
contacts. This is, of course, of critical importance,
fo llowiug Walther's law. Cant and Walker (1976)
attempted to address the problem by erecting a separate facies state for scour surfaces that, in their field
case, were typically accompanied by a distinctive
poorly-sorted sandstone. Miall and Gibling (1978)
took a different approach, documenting the number
of erosional and gradational contacts between each
fa cies state and developing a !
tabular form. Although this was a systematic approach to the problem, the method has not been used
by other workers because attention at this time began to turn away from vertical profiles. As shown by
Godin (1991), cyclic relationships can be developed
at several nested levels within fluvial systems, and
simple one-step Markov methods are not suitable for
analyzing these complexities (Chaps. 8, 10).
A somewhat different approach was taken by
Friend et al. (1976), who measured many lengthy
sections through the Devonian nonmarine sediments of East Greenland. The sections were divided
into 10-m segments, and the abundance of a range of
fades attributes (e.g., grain size, sedimentary structures) was determined for each segment. The data
were then subjected to fa ctor analysis to determine
'
E
2
05
Grom dmmeter ,d {mm)
a basis for subsurface environmental interpretations
(e.g., Pirson 1977; Berg 1986). Commonly, these were
identifi ed not on the basis oflithologic logs in cores,
but fr om the profiles displayed by electric logs - the
classic "funnel-shaped" and "bell-shaped" gammaray, spontaneous potential, and resistivity logs that
had first been identified during the 1950s (e.g., Nanz
1954).
The value of a facies model in providing a synthesis of a diverse array of fa cts is offs et by the danger
that uncritical use of the model may lead to a loss of
information or to misinterpretation, because it is
tempting to observe strata in terms of a preconceived model. Objective field measurements commonly are difficult to make, particularly in areas of
poor exposure. One of the trends in sedimentological research in the 1970s was t'o quantify observations of lithological successions in order to enable
model sequences or cycles to be defined statistically.
Markov chain analysis was commonly used for
this purpose (e.g., Allen 1970b; Miall 1973, 1977;
Cant and Walker 1976). Although this approach allows more precision in defining the model, it can
cause valuable observational detail to be ignored.
Collinson (1978, p. 579) made this point about the
Battery Point Formation, a highly varied, low-sinuosity-fluvial deposit from which Cant and Walker
(1976) extracted a single "summary" cyclic sequence
(Fig. 2.24a).
39
Fig. 2.23. Bedform phase diagram showing how
bedforms may be classified according to the physical
processes which control their size limits and stability.
(Leeder 1983)
Amongst the problems that had been recognized
with the Markov chain method was the fa ct that it
contained no information on the nature of facies
contacts. This is, of course, of critical importance,
fo llowiug Walther's law. Cant and Walker (1976)
attempted to address the problem by erecting a separate facies state for scour surfaces that, in their field
case, were typically accompanied by a distinctive
poorly-sorted sandstone. Miall and Gibling (1978)
took a different approach, documenting the number
of erosional and gradational contacts between each
fa cies state and developing a !
tabular form. Although this was a systematic approach to the problem, the method has not been used
by other workers because attention at this time began to turn away from vertical profiles. As shown by
Godin (1991), cyclic relationships can be developed
at several nested levels within fluvial systems, and
simple one-step Markov methods are not suitable for
analyzing these complexities (Chaps. 8, 10).
A somewhat different approach was taken by
Friend et al. (1976), who measured many lengthy
sections through the Devonian nonmarine sediments of East Greenland. The sections were divided
into 10-m segments, and the abundance of a range of
fades attributes (e.g., grain size, sedimentary structures) was determined for each segment. The data
were then subjected to fa ctor analysis to determine