Author: Chris Stassen
Subject: FAQ: Evaluation of ICR Grand Canyon Research Project [DRAFT]
NOTE: Austin has promised to publish all of his data at some point, and
this work cannot really be completed until that happens.
The Grand Canyon looks something like this:
___\ /__________________ <-- Paleozoic rocks
...\ \ \ \ \ \ .... <-- Angular unconformity
.....\ \ \ \ \ ......
.......\ \ \ ............
.........\ .................. <-- Metamorphic rocks
Steve Austin of the ICR wrote _Impact #224_, claiming that he has
derived a Rb/Sr isochron age for lava flows on the Uinkaret Plateau
in excess of 1.3 billion years (several of these flows spill into
the canyon). The Cardenas Formation, a tilted layer at the bottom
of the canyon (below the angular unconformity) is generally accepted
as being about 1.1 billion years in age.
Clearly, something is wrong here. Lava flows which spill
off the plateau into the canyon cannot be older than the canyon
itself. The canyon can't be older than the rocks that it is cut
into. The flat sediments can't be older than those below the
angular unconformity (as they were deposited on top).
Austin's apparent thesis: Rb/Sr dating of the feature which must
be the youngest gives a value older than that which must be the
oldest, therefore the method does not work. (Of course, then, all
radiometric results can safely be ignored.)
I received a description of the data from Steve Austin (a little
more than what was published in _Impact_, but not nearly enough
to really know what is going on, unfortunately). I forwarded the
data to Dr. G. Brent Dalrymple of the U.S. Geological Survey. He
has apparently received many questions on Austin's work, and sent
me a short article that he had written on the subject. I am in
the process of getting permission to give out copies of this report;
please send contact me (firstname.lastname@example.org) if you are interested in
getting a copy (I can probably get copying done for free, but you
may have to cover the postage).
A few excerpts:
"Trying to prove that some methodology does not work by
focusing on a small percentage of anomalous data, while
ignoring a much larger percentage of concordant data is
not a productive exercise because it will prove nothing
except that the method did not work *in those particular
instances* [underlined in original -CS]. It is even
less meaningful if one seeks to perform an experiment
under circumstances that are known beforehand to be
inappropriate, and this is precisely what Steven Austin
is doing. He has set up an experiment that will fail
because he has chosen circumstances that guarantee the
outcome. So what is being tested--radiometric dating
methods or Steve Austin's knowledge and objectivity."
Commenting on the fact that Austin has published three *different*
alleged isochron dates for the Cenozoic lava flows:
"These differences [changes in various measurements for
certain samples! -CS] explain the change in Steven's
``isochron age'' from 2.07 Ga to 1.39 Ga (or vice versa),
but we do not know why the analytical data changed so
radically from one table (and one diagram) to the other,
why one sample was omitted from the second table, or why
Steven selects just a few of the many available hawaiite
flows for his project."
"In summary, I can't really tell what Steven Austin is
doing with his western Grand Canyon data. They keep
changing and he never provides enough information to do
an independent evaluation of important things like the
isochron fit. In addition, he is ignoring what I and
others have told him about using lava flows that are
demonstrably not cogenetic, and he is ignoring Leeman's
data, which clearly indicates that the rough correlation
between the Rb and Sr isotopic ratios reflects time-
integrated radioactive decay in the source rock(s), and
not in the lava flows."
This is some additional material, which I put together at a later
David M.V. Utidjian writes, in :
> But isn't the K-Ar method of radiometric dating the apropriate method
> for dating igneous rocks like lava flows? Please correct me if I am
The K-Ar dating method is probably the most _popular_ method for dating
lava flows. Since it involves relatively few measurements, and they
are relatively straightforward, it is also one of the cheapest methods.
Jim Lippard was kind enough to look up and copy for me a technical
paper on the topic of K-Ar dating the Grand Canyon lava flows (McKee
et. al, 1968). It gives K-Ar dates for laval flows associated with
the Toroweap fault (~1.2 million years). Austin references this paper
(Austin, 1992, 1988), so it seems that these are the same flows that
he is "dating."
Despite its popularity, the K-Ar method is not the most _reliable_
dating method. Since it is not an isochron method, there is no
built-in check for contamination (that is, in the event of contamination
it will yield an incorrect date -- while an isochron method under the
same conditions will usually yield no date at all). Since argon is an
inert gas, later heating events can permit it to escape, and partially
or completely reset the "clock." Since there is excess argon associated
with the lava source, its method of cooling must be such that the initial
argon could escape.
This does not mean that the K-Ar method is useless; it only
means that it must be used with caution -- on samples that have the
right sort of history and properties. In fact, this case (plateau
lava flows) appears to be one where K-Ar yields the right age and
proper use of Rb/Sr dating yields no age at all.
If we plot all of the Rb/Sr data which Austin has ever published
(in three different places) for "Western Grand Canyon" lava flows, you
will see why I say "no age at all":
0.707 + (4) (f)
0.706 + (3)
0.705 + (2) (e)
| (1) (d)
0.704 + (b)
0.703 + (Aa)
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Lowercase letters (e.g., "(a)") are from Austin 1988. Uppercase letters
(e.g., "(A)") are from Austin, No Date. Numbers (e.g., "(1)") are from
Austin, 1992. Warning: All data points were "eyeballed" from published
diagrams. ["(Aa)" is "(A)" and "(a)" relatively close together.]
In Austin's diagrams, A-B-C-D fall relatively close to a line.
a-b-c-d-e-f also fall relatively close to a line. 1-2-3-4-5-6 fit very
well to a single line. However, you can see that there is quite a bit
of scatter when all of the data is plotted together.
I've been told that Austin has in the past used a computer program that
discards data that do not lie near a line. If that is what he is doing
here (it has that appearance, but I really don't have enough information
to be sure), it is no surprise that he is ending up with data points
that fall on or near a line.
It is also no surprise that the resulting "age" is not meaningful.
In ignoring data points away from the line, one would be ignoring data
points which indicate that the "line" does not represent an isochron.
Here are two reasons why there is the appearance that this is what is
In Austin 1988 (lowercase letters above), he presents data from Leeman,
1974. He uses samples "U-33, U-35, U-59, U-24, U-43, and U-58." He does
not explain why he chooses only these data points. Leeman's diagram
shows considerable scatter to the data. From Leeman's diagram, it is
clear that no Rb/Sr age can be calculated. Austin has chosen only a
few of the data points, and these happen to fall near a line (however,
the appearance is that even these are not close enough to a line to
be considered a valid isochron date).
For another example, the most recent data (Austin, 1992) appears in
two forms. In an as-yet-unpublished paper (maybe a chapter from a
forthcoming book), Austin produces an isochron that gives an "age"
of 2.07 Ga, from samples "QU-1, QU-2, QU-3, and QU-5."
In a report on the progress of the "Grand Canyon Research
Project," Austin notes that there was a problem with some of the
measurements, that an incorrect value resulted, but that the sample
is now correctly dated. (The project is running a little behind for
The _Impact_ article uses the some of the same sample numbers
as the other isochron, however (as Dalrymple notes) some of the
isotope concentration values have changed. (I suspect that this is
a result of the correction of the lab results.) However, the _Impact_
isochron uses samples "QU-1, QU-2, QU-5 and QU-14." Sample QU-3 is
missing. Its previous position would be well off the line.
Only ICR folks know why sample QU-3 was dropped, why sample
QU-4 has never been used, and where samples QU-6 through QU-13 are.
Until they publish _all_ of the lab results and _all_ of the data
points, it will be difficult to assess exactly what has been going
The appearance at the moment is that their results are not the major
blow against isochron methods that they are claiming. Until they
are more open with their results (ALL of their results), that will
have to suffice.
> It also appears that K-Ar methods would be more accurate for the
> apparent ages of these samples than the Rb-Sr method because the of
> their respective halflives of 1.25 x 10^9 for K-Ar and 4.88 x 10^10
> for Rb-Sr, respectively. That is to say the error bars for the Rb-Sr
> method may be an order of magnitude larger than the ones for the K-Ar
> method in my amateur evaluation of the two techniques.
It is true that the decay rate has some influence on the size of the
errors. However, the disagreement in this case cannot be explained
merely by measurement error. It is still important to understand where
the two methods agree or do not agree (and why).
Austin, Steve, 1992, "Excessively Old ``Ages'' For Grand Canyon Lava
Flows", _Impact_, Number 224, February.
Austin, Steve, 1988, "Grand Canyon lava flows: A survey of isotope
dating methods", _Impact_, Number 178, April.
Austin, Steve, No Date, "Grand Canyon dating project: A proposal for
Leeman, 1974, "Late Cenozoic Alkali-Rich Basalt from the Western Grand
Canyon Area, Utah and Arizona: Isotopic Composition of Strontium", in
_Geological Society of America Bulletin_, Vol. 85, November, pp. 1691-1696.
McKee, E., D., et al., 1968, "K-Ar Age of Lava Dam in Grand Canyon",
in _Geological Society of America Bulletin_, Vol. 79, January, pp. 133-136.