David Tyler wrote:
>From: d.tyler@mmu.ac.uk
>To: Evolution Reflector
>Subject: Re: Old Earth
>Date: Tue, 9 Apr 1996 16:07:45
>
>ABSTRACT: A look at presuppositions in historical geology - with
>reference to the use of isochrons in radiometric dating (and a brief
>comment on the work of Steve Austin).
>
>I had written: "Your comment presupposes that the presuppositions
>of these radiometric dating methods are valid. But if the
>isotopic data is better understood in terms of geochemical or
>non-radiogenic geophysical causes, the data is not meaningful in
>a chronological sense".
>
>Bill Hamilton wrote on 28th March:
>"What do you mean by this? Do you have references? I am not a
>geologist, but I understand radioactive decay and I have read a
>few references on radiometric dating methods and the rules
>geologists use to qualify samples to reduce the possibility of
>errors. As nearly as I can tell, their science is sound."
>
>In my original post to Steve, I illustrated my point with two
>examples: the theoretical distinction between "normal" and
>"abnormal" lead isotopes (which is hard to defend), and
>alternative explanations of isochrons. I will pick up on the
>latter to develop the argument, basing it on Rb/Sr dating. In
>what follows, I am not throwing stones at geochronologists -
>their methodology is sound GIVEN THEIR PREMISES. What we are
>looking at is an alternative framework for interpretation - one
>that is open to the idea that the isotopic ratios are not
>necessarily caused by the passage of time. (I should add that
>I last looked seriously at this subject about 10 years ago - but
>I'm not aware of significant developments over the last decade).
>
>1. Isochrons are not a sign of closed systems.
>
>The thought sometimes expressed is that BECAUSE a straight line
>is produced, the data must be "reliable" and the conclusions
>"accurate". This is of significance because both Sr and Rb
>isotope values are quite variable in igneous rocks. Can we be
>sure that the system is closed and isotope migration has not
>occurred?
>
>It is common to identify metamorphic or deformation events which
>disturb or reset the isotopic ratios. However, it is generally
>assumed that the system is closed outside these events - so that
>the dates obtained are dates of the disturbances.
>
>Matsua (1974) has shown that "a linear isochron can be obtained
>even for an open system". "Such open systems likely exist in
>Nature. Hence, we must be very careful to conclude the age on
>the basis of an isochron".
>
Matsuda linearizes his equations by an assumption that the diffusion
equation can be replaced by a simple first order rate equation. In my
mind this is questionable. A diffusion is not a linear process, most of
the diffusion occurring early decreasing in its effect as time goes on.
The slow diffusion late in the process when there is little original
material left in the original volume can be linearized. But this ignores
the issue entirely that early on, the diffusion is not linear and decays
exponentially. His assumption is wrong and thus produces a red herring.
>Sometimes, an identifiable disturbance is found which correlates
>with the resultant date - and the interpretation placed on the
>data is not only consistent but considered meaningful. But is
>this optimism justified? Field and Raheim (1979) published their
>data regarding "A geologically meaningless Rb-Sr total rock
>isochron". Although the rock suites used showed "evidence of
>only minor secondary alteration effects", they had to conclude
>that "An isochron so produced is geologically meaningless, that
>is, the apparent age is unrelated to any geological event".
>
But they state, "The only petrographically visible effects of this
superimposed low-grade event are (1) partial, and often variable,
alteration of orthopyroxenes to chlorite and/or serpentine, (2) minor
corrosion of pre-existing biotite, and (3) incipient turbidity within some
plagioclases. These mineral alterations were initiated along narrow (<1mm
) irregularly spaced cracks which facilitated the open system behaviour."
p. 498
The low level event was petrologically observable and so did not lead to
an acceptance of an anomalous date. It would appear to me that the system
worked as it was supposed to.
>Bell and Blenkinsop (1978) address the problem of open systems:
>"The susceptibility of fine-grained, volcanic rocks to open-
>system behaviour during regional metamorphic activity has long
>been suspected in Rb/Sr geochronology". Their paper not only
>confirms suspicions, but relates Rb/Sr ratios to bulk chemical
>composition. Their paper discusses the problem of getting any
>meaningful information from the analyses.
>
Wait a minute here. The classical evidence for open system behavior is a
non-linear isochron. Samples that fall off a straight line indicate that
the system is open. Bell and Blenkinsop's figure 2 clearly shows that
there is a kink in the isochron data. They related this kink to the
chemical composition of the rocks. They state,
"The evidence that we present here shows that not only can open-system
behaviour occur, but that it is partly a function of bulk chemical
composition." p. 532
The isochron method relies on the concept that the rocks came from the
same magma melt. I would question this in the Bell and Blenkinsop data.
The samples he has fall into two groups, one group has silica content
between 44 and 69% while the second group has greater than 70% silica.
The K2O values are also quite different as is the iron ratios. Given this
chemical difference in the two sets of samples one might question whether
or not the material came from the same magma melt although Bell and
Blenkinsop don't. They ascribe different behavior to the two different
chemical compositions.
Once again, the isochron system detected open system behavior and so no
one was fooled.
>Are meaningless isochrons the norm or the exception? Are
>premises relevant here?
>
>2. Isochrons can be produced by the mixing of source materials.
>
>Suppose that a two-phase model of rock magma be considered. This
>is not a way-out suggestion as bimodality is not unusual in
>igneous terrains. Whilst models of magma evolution are less
>complex, most will agree that there are many unknowns between
>generation and emplacement. In our two phases, we have
>differences in the Rb and Sr contents.
>
>The resultant rocks produced from these magmas will yield
>isochrons which are unrelated to age. The isochron is a
>geochemical signature, not a mark of the passage of time. The
>theory for this seems to be well-recognised in the literature,
>but I have not got a reference to the mathematical derivation to
>report.
>
>Dupre et al (1982) discuss isotopic variations on the island of
>Terceira. they report observations "in agreement with a model
>of mixing between two different mantle components, ..." They
>obtained isochrons using lead isotopes, but stated: "These
>straight lines must be considered as mixing lines with no
>chronometric significance".
>
But Dupre et al also say,
"The whole series of samples which seems to define a common trend of
fractional crystallization have variable isotopic ratio values and must
therefore belong to different magmatic series." (p. 621)
The lavas were emplaced on Terceira over a period of 300,000 years. In
that time at the speed of the mantle circulation (approx. 4 cm/year) the
mantle material could have traveled 12 kilometers and this could be a
factor in the change in isotopic composition over that time period.
>Shaffer and Faure (1976) looked at strontium ratios in sediments
>taken from the Ross Sea. Their conclusions are:
>"The isotopic compositions of strontium and concentrations of
>rubidium and strontium as well as those of several detrital
>minerals in the <100-mesh noncarbonate fractions vary
>systematically throughout the Ross Sea. These variations can be
>attributed to mixing of two detrital components which are
>weathering products of old sialic rocks and young basaltic
>volcanics in Antarctica. ... Linear correlations of 87Sr/86Sr and
>87Rb/86Sr ratios are due to mixing of two components and bear no
>relation to the age of the rocks in Antarctica or to the time of
>deposition of the sediment."
>
This is detrital material and should have no bearing on
age. The material ranges in age from Precambrian to Cretaceous.
The reason is that you don't know where the detritus
came from. The isochron method is supposed to work only if
you have several rocks from the same magma. You can't
guarantee that in detritus. But a thoroughly mixed
assemblage from a lot of sources probably would give you a
straight line. No one would attempt dating of the detritus
by this method.
>Again I ask: if magma mixing is a mechanism for producing
>isochrons, why do we not hear more about it? Are there ways of
>distinguishing between age-related isochrons and geochemical
>isochrons (due to mixing)? (There is one method - see below).
>Are people even asking these questions? How far are premises
>important here?
>
>3. The phenomenon of pseudoisochrons.
>
>According to Brooks et al (1976), "One serious consequence of the
>mantle isochron model is that crystallisation ages determined on
>basic igneous rocks by the Rb-Sr whole-rock technique can be
>greater than the true age by many hundreds of millions of years.
>This problem of inherited age is more serious for younger rocks,
>and there are well-documented instances of conflicts between
>stratigraphic age and Rb-Sr age in the literature." This
>phenomenon of "inherited age" is the theme of their paper.
>
>They distinguish a pseudoisochron from a simple isochron plot in
>the following way. The simple plot is of current-day values, but
>the pseudoisochron used values corrected back to the time of
>crystallisation of the magma. However, in practice, "For the
>most part the pseudoisochrons come from young volcanic terranes
>in which no age correction of the measured present-day Sr
>isotopic composition is necessary."
>
>The paper documents a whole series of "anomalous dates" from
>tertiary volcanics. Of particular interest are two reported by
>Leeman (1974). Hawaiites from the Western Grand Canyon have an
>apparent age of 1300 million years, and alkali basalts from the
>same area are dated as 1100 million years. I'll comment further
>on these below. The authors suggest three ways of getting
>pseudoisochrons:
>1. Mixing of heterogeneous magmas.
>2. Selective melting of heterogenous mantle material of the same
>age or different ages.
>3. Disequilibrium melting of homogeneous mantle in which
>individual mineral phases are in isotopic disequilibrium.
>"Distinguishing between these possibilities is difficult as none
>of the proposed mechanisms leads to a uniquely identifiable
>compositional property in the rocks, other than that of Pb and
>Sr isotopic variation ..."
>"One means of evaluating pseudoisochrons is to determine whether
>they involve identifiable mixing-line chemistry. Rubidium-
>Strontium variation diagrams and plots of 87Sr/86Sr against 1/Sr
>can be used for this purpose. Examination of such plots for the
>data [used in this study] reveals that a good pseudoisochron is
>commonly accompanied by a good positive correlation between
>87Sr/86Sr and 1/Sr, but no correlation at all between Rb and Sr.
>The absence of an Rb-Sr correlation is inconsistent with simple
>mixing. Thus, while we cannot discount mixing of heterogeneous
>mantle material, the data in general seem to preclude simple two-
>component magma mixing as an important factor in most
>pseudoisochrons."
>
>I do have a problem with the last point above. I would say that
>a correlation of 87Sr/86Sr and 1/Sr is indicative of mixing. Are
>the 87Sr isotopes due to radioactive decay of 87Rb, or are they
>dependent on the concentration of Sr in the rock? The positive
>correlations reported indicate that the strontium isotope
>concentrations are directly dependent on the strontium present
>in the rock. At the moment, I do not follow Brooks argument -
>if anyone can see the logic behind it, I will be glad to hear
>from you.
>
>The Grand Canyon lavas mentioned above are the ones studied by
>Steve Austin - some discussion took place on this last year.
>Austin has obtained values of 1.3 billion years for the Tertiary
>basalts of Grand Canyon: something for which he is using to
>question the appropriateness of using Rb/Sr dating. The
>important point here is that his findings are is close agreement
>with Leeman (1974). To nit-pick at possible selective use of
>data seems to be an unwise response to Austin. The argument of
>Brooks et al (1976) is that Austin has found a pseusoisochron -
>that conveys information about the earth's mantle and mechanisms
>of petrogenesis. At the moment, unless I can get my question
>(previous paragraph) answered, I think the magma mixing
>explanation is to be considered the most likely - which implies
>that the isochrons have geochemical, but not geochronological,
>information.
>
This whole thing with the bad Rb/Sr dates ignores the fact that we know
they are wrong because other radiometric dates give a more reasonable age.
The dating of the lava dam (which gives a billion years age with Rb/Sr)
gives 1.16 Myr with K-Ar. This date also agrees with the date derived
from stratigraphy. This always seems to be ignored. The bad Rb/Sr dates
at Grand Canyon are always spoken of as if this is the only dating method
applicable to them.
One of the explanations given by Brooks et al is that some of the
lithosphere has become trapped below the continent accounting for the
problematic Rb/Sr dates. If this is correct then it is quite possible
that the magma melt which gave rise to the Cardenas Basalt also gave rise
to the late Cenozoic volcanism. This would explain the discrepancy
between the K/Ar date and the Rb/Sr date.
I would also like to point out that Steve is quite selective in the way he
handles contradictory evidence. He cites (p. 123 of his Grand
Canyon:Monument to Catastrophe) a 117 million year old K/Ar date on the
Vulcan Throne Basalt in the text but only informs the reader at the end of
the footnote buried at the end of the chapter that olivine is not a
suitable mineral for K/Ar dating. Olivine was what gave that 117 Myr
date. He did a good job of burying that piece of info.
He then cites the 1.2 Myr date I mention above but never discusses their
significance. Argon being a gas is able to mostly leave the mineral when
the lava erupts. It's date fits with local geologic evidence and Steve
simply mentions the age and then moves on. He also cites a 10,000 year
K/Ar date for one of the upper lavas and a 230,000 year date for another.
Because of all this Brooks et al's explanation seems quite reasonable.
>Even if Brooks et al are right, the number of variables involved
>becomes very large. We have a situation where the isotope
>variations cannot be interpreted without a context:
>"Effective use of the mantle isochron concept requires knowledge
>of actual crystallisation ages (so that the measured isochron can
>be divided into its pre- and post-crystallisation components) and
>determination of isochron parameters on rocks that have been
>subjected to minimal postmelting processes (fractional
>crystallisation, wall rock contamination, and so forth)..."
>The net effect is to "fit" the data into an accepted paradigm.
>
>Conclusion
>
>I have discussed (1) the problem of open systems, (2) the
>alternative explanation of magma mixing, and (3) pseudoisochrons.
>The picture emerging is one of paradigm-dependency. The
>presuppositions are essential for the method to deliver results.
>The objectivity of radiometric dating (at least affecting Rb/Sr
>dating) is an illusion. My personal conclusion (stated in a
>previous post) is that all dating methods are subject to
>presupposition problems of this kind - and that we really have
>no valid ways to establish a chronology of earth history.
>
This last sentence is quite incorrect IMO. The relative chronology of
earth history can be worked out by studying the sequence of rock layers.
The principle of superposition is the basis of all geochronology. The
rock layer on the bottom is the oldest. Some sedimentary rocks preserve
the orbital cyclicities of the earth (20,000, 40,000 and 100,000 and
400,000 year cycles). These cyclicities allow one to date the time it
took for those sediments to be deposited. The 13 million layers in Lake
Gosuite of the Green River Formation preserve these types of cyclicities.
It is believed that this deposit took about 6.5 million years to be
deposited and it represents only 7% of the sedimentary column. It is also
one of the most rapidly deposited formations.
glenn
>References
>
>Bell, K. and Blenkinsop, J. 1978. Reset Rb/Sr whole-rock
>systems and chemical control. Nature, 273(15 June), 532-534.
>
>Brooks, C., James, D.E. and Hart, S.R. 1976. Ancient
>lithosphere: its role in young continental volcanism. Science.
>193(17 September), 1086- 1094.
>
>Dupre, B., Lambret, B. and Allegre, A.J. 1982. Isotopic
>variations within a single oceanic island: the Terceira case.
>Nature, 299(14 October), 620-622.
>
>Field, D. and Raheim. 1979. A geologically meaningless Rb-Sr
>total rock isochron. Nature, 282(29 November), 497-499.
>
>Leeman, W.P. 1974. Late Cenozoic alkali-rich basalt from the
>Western Grand Canyon Area, Utah and Arizona: isotopic composition
>of strontium. Geological Society of America Bulletin.
>85(November), 1691-1696.
>
>Matsuda, J. 1974. A virtual Rb-Sr isochron for an open system.
>Geochemical Journal. 8, 153-155.
>
>Shaffer, N.R. and Faure, G. 1976. Regional variations of
>87Sr/86Sr ratios and mineral compositions of sediment from the
>Ross Sea, Antarctica. Geological Society of America Bulletin.
>87(October), 1491-1500.
>------------------------
>
>Best wishes,
>
>*** From David J. Tyler, CDT Department, Hollings Faculty,
> Manchester Metropolitan University, UK.
> Telephone: 0161-247-2636 ***
Foundation,Fall and Flood
http://members.gnn.com/GRMorton/dmd.htm