R.A.T.E., and Isochron Discordance

Isochron Discordance:
The principal investigator of discordance in isochron dating was Dr. Steven A. Austin. (If you get a chance, look him up. His abysmal track record is amazing!) The basic idea is rather simple: Measure samples from a single rock or homogenous rock unit, and develop isochrones using multiple geochronometers, such as K-Ar or Pb-Pb. Each method should then agree with the others on the age of the rock. Austin himself says in the introduction:

"Do conventional radioisotope dating methods when applied to a single rock or rock unit give concordant ages? Specifically, do the K-Ar, Rb-Sr, Sm-Nd, and Pb-Pb radioisotope pairs each give the same age, within the uncertainties allowed using radioisotope dating, for a single rock or rock unit? If concordant ages were obtained for a single rock or cogenetic suite of rocks using different radioisotope pairs, these would bee impressive evidence for the consistency of radioisotope dating and be an affirmation of the assumptions underlying these dating methods. However, if discordant ages were obtained, then some qualifications would need to be applied to the assumptions undergirding these dating methods."

Austin claims that isochron discordance has not been thoroughly described and explained. As a result, he proposes a classification system of various kinds of discordance among varying kinds of isochrons:
Category one discordance – a cogenetic suite of rocks with two or more discordant whole-rock isochron ages.
Category two discordance – a cogenetic suite of rocks that generates a whole-rock isochron age older than the associated mineral isochron ages from specific rocks.
Category three discordance – two or more discordant mineral isochron ages from the same rock.
Category four discordance – a cogenetic suite of rocks that generates a whole-rock isochron age younger than the associated mineral isochron ages.
It should be readily apparent, even to the layperson, what Austin is attempting to do here. Certain radioisotope dating methods are fine for certain kinds of rocks, and inappropriate for other types of rocks. Austin intends to juxtapose the appropriate dating technique with the inappropriate one and claim that there is too much discrepancy for any of the dating methods to be reliable. Also, Austin telescopes his intentions with his classification system. No one needs to develop such a set of classifications unless he is anticipating that he will have need of them. In other words, he’s already drawn the conclusion.
Also, note the sheer hubris of the proposal. Throughout geology, all over the world, radiometric dates have been showing themselves to be concordant with each other for decades. To be sure, not always, and sometimes not perfectly, but they do match up. And when they don’t, they usually do so in a way which teaches us something about the metamorphic history of the rock. And here comes this one, lone researcher who feels that by finding a few discordant examples that he will have unhorsed the entire thing. Now, that, my friends, is one seriously heavy set of brass balls!
For his study, Austin selected two Precambrian rock sites for collection and analysis. The first was a lower Precambrian metamorphic rock unit from the southeastern Beartooth Mountains of Wyoming, taken from a road-cut in that region. The second was the upper Precambrian diabase sill at Bass Rapids in the central Grand Canyon (Hotauta Canyon area, about 2/3 of a mile east of Shinumo Creek, amidst the Hakatai Shale). For the Beartooth sample, a single multi-kilogram andesitic amphibolite was collected. The road cut it was taken from was on U.S. highway 212, west of Beartooth Pass, 1.3 km east-northeast of Long Lake. At the Grand Canyon, eleven rock samples were taken at Bass Rapids on the north bank of the Colorado River at mile 107.6 to 108.0, a location where the diabase sill is nearly 100m thick in places. These eleven were chosen, “to represent the overall petrographic variability within the complete thickness of the sill” (Austin, et al., 2005). (This will be revisited later.) Eight of the samples were from the lower approximately 85m of the sill. The other three were from the 6m thick granophyre at the top of the sill.
The samples were prepared and tested. K-Ar analysis was performed by Geochron Laboratories, Cambridge, MA, as well as Activation Laboratories, Ancaster, Canada. Rb-Sr, Sm-Nd, and Pb-Pb analyses were performed on whole-rocks and mineral separates using a Finnigan-MAT 6-collector solid source mass spectrometer at the University of Colorado.
The results? The Beartooth sample contained various minerals, including plagioclase quartz, biotite and hornblende. The ages yielded by K-Ar dating, for the whole rock, plagioclase, biotite and hornblende respectively, were 2011 Ma, 1520 Ma, 2403 Ma, and 2620 Ma. The plagioclase quartz had far less 40Ar, but also had the lowest 36Ar, consistent with the Argon loss model. The biotite had the most 40K and 40Ar, but an isochron plot was not drawn up for it. The Rb-Sr method, using the whole rock, plus five minerals, yielded an isochron which dated the rock to be 2515±110 Ma. Sm-Nd gave an isochron age of 2886±190 Ma. And finally, Pb-Pb gave an isochron age of 2689.4±8.6 Ma. So the results were concordant with Rb-Sr, Sm-Nd, and Pb-Pb. Not perfectly, but the error margins do overlap. K-Ar, however, was another matter. Even accounting for Argon loss, most of the dates produced were too young. The one exception was the hornblende, which produced a date which matched the other results rather well.
The Bass Rapids results were somewhat different. K-Ar dates were a complete mess, with ages ranging everywhere from 656 Ma to 1053 Ma. For the other dating methods, Austin shows us the results of two of his eleven samples, labeled DI-13 and DI-15. The DI-13 sample yielded a mineral isochron age of 1060±24 Ma, while DI-15 yielded an age of 1075±34 Ma. For Sm-Nd, a whole rock analysis using all eleven samples yielded an isochron age of 1379±140 Ma, which matched very well with the eight diabase samples, but the three samples taken from the granophyre were widely scattered. This, Austin suggests, is suggestive of contamination of feldspar from the overlying hornfels wall rock, incorporating less radiogenic Nd. A mineral isochron analysis of DI-13 yielded the exact same age of 1379±140 Ma. Whole rock Pb-Pb analysis gave an isochron age of 1250±130 Ma, and this is the same as the mineral analysis age achieved on sample DI-13 as well.
So what are we to make of all this? Clearly, the dates do not all agree, and Austin thinks he knows why. He notes that the α-decay isotopes consistently showed an age much older than the β-decay isotopes, and he believes that this indicates that α-decay took place much faster in the past, and that this is the cause of the discrepancy.
So where did Austin go wrong? Well, there are a number of areas. First and foremost is his methodology in collecting the samples. One simply cannot help but notice that Austin himself did the sampling. A great degree of variation can be achieved by selecting samples that look non-homogenous while claiming that they are homogenous on paper. To make it a single-blind test, some other geologist should have collected the samples to prevent any bias regarding which samples were selected, and have been told only that a true age of the collected sample was desired (i.e. avoid any xenoliths and phenocrysts). Second, the rocks selected were hardly the ideal kinds of rocks for K-Ar or Rb-Sr dating. The road-cut along highway 212 had probably been there for some time, allowing plenty of water to have access to the rock sample collected there. Now, water leeching is a major component of error for Rb-Sr dating, as is surface outgassing for K-Ar. We were only told that the rock taken from along that highway was “multi-kilogram.” How multi? Was it thick and round, or long and flat? Based on his results, I would be willing to guess long and flat is more likely. Further fishy sampling can be seen in the initial proposal, which stated that the sample collected would be a “flood gabbro” from the Grand Canyon. This, of course, means Noah’s flood. This is specious in and of itself, because no creationist has ever definitively stated which geologic formations mark the onset of The Flood, and Austin makes no effort to resolve that issue. And while technically one can do such dates on any volcanic rock, gabbro isn’t the best choice. Potassium-bearing minerals are best for K-Ar dating, and a gabbro, or in this case, a diabase, contains few such minerals. Also, the diabase bifurcates the Hakatai shale – a porous rock. This again makes water a factor in the Rb-Sr dates, to say nothing of exposure to the occasionally wet, eroding environment of the Grand Canyon itself. Third, the Grand Canyon samples were taken from a very wide area of exposed sill – nearly half a kilometer long, and nearly 100 meters thick! A very nice picture illustrates where Austin collected his samples – vertically. But he does not show horizontally where he collected them! Might that be because he gathered them over the entire length of the exposed sill? Over such a large area, there is little guarantee that the rock samples are of the same radiogenic age, or even the same volcanic event. Remember earlier when Austin was quoted, saying that his eleven samples were chosen “to represent the overall petrographic variability within the complete thickness of the sill.” But if you’re doing radiometric dating, you don’t want variability! You want homogeneity! This alone is enough to invalidate his results.
Given the millions of potential erroneous samples one could choose from, one could always find something which will yield the kind of errant radiometric date desired. I’m surprised Austin’s results weren’t more discordant! One can only assume he couldn’t resist the temptation to tag the Grand Canyon as containing creationist evidence.
The end of Austin’s chapter in RATE’s second book is devoted to answering back those who would point to certain factors involved in the radiometric dating process. He addresses mineral inheritance, slow cooling, and post-magmatic loss of Sr, after which there isn’t much of his argument left, though champions of creationism will see this honesty as something which bolsters his conclusions rather than weakens them (oh, brilliant politics!). Here’s what he has to say about Strontium loss:

"If we accept the Pb-Pb and Sm-Nd whole rock and mineral isochrons as true ages, then the Rb-Sr isochrons would be interpreted as significantly altered with a specific bias to give significantly younger but spurious ages. We need to imagine some type of Sr loss process that removed significant 87Sr and/or added significant 87Rb over geologic time. Less likely, but also making Rb-Sr ages much younger, would be addition of Rb, a process not considered further here (Vardiman, et al, 2005)."

Note how artfully he plays it! He acknowledges the possibility of Strontium loss, then acts like Rubinium would also need to be added in order to make the likelihood of such a thing seem nearly impossible. The truth of course is that these are trace elements, and only a little Strontium need be removed to achieve a massively discordant date!
Of all the RATE team’s research projects, Austin’s is the most insidious, the most difficult to follow, the one most deeply buried in technicalia. As a result, it will likely hold more weight among creationists than the others, and for many years to come, sadly.