We have done three previous posts on this blog pointing out the gargantuan flaws in Marvin Resnikoff’s work.  You can read them here, here and here.  We have noted his work is not only seriously blemished with numerous defects, but his testimony has been rejected again and again in courts of scientific research as well as law, qualifying him as the perfect “anti-expert.”  Well, it turns out the scientific community, may have had it with him, too.  The U.S. Geological Survey (USGS) has just released a report that all but him accuses him of being an eccentric crank.  Here’s how what they said in their cover letter distributing the report:

In response to concerns over human exposure to radon in natural gas supplies from the Appalachian Basin, the USGS has released a preliminary dataset providing radon-222 concentrations in natural gas samples derived from the Marcellus Shale and Upper Devonian sandstone reservoirs.  This preliminary dataset has been summarized in a short report authored by my colleagues Liz Rowan and Tom Kraemer, which can be downloaded from the USGS Publications Warehouse: http://pubs.usgs.gov/of/2012/1159/

And, here’s what they say in the report itself (emphasis added):

A recent report by Resnikoff (2012) has led to increased interest in possible human exposure to radon as a component of natural gas in household settings.  The report, however, relied on theoretical calculations utilizing limited data from geologic analogs.  A decision was made to release our small and preliminary dataset because, to the authors’ knowledge, measurements of radon in natural gas at the wellhead have not previously been published for the Appalachian Basin.

This is polite agency talk for “we’re not going to let this guy keep getting away with distorting our previous work based on bunch of hokey speculations.”  Their report effectively confirms the criticisms of Resnikoff’s work by Ralph Johnson and Lynn R. Anspaugh, Ph.D, the relevant details of which have been shared here, but there’s more.

Radon is no simple subject, but both Johnson and Anspaugh have noted Resnikoff’s numbers are wildly unrealistic, starting with his most basic assumptions, which rely upon bad arithmetic and speculation to extrapolate old data into new doomsday threats of radon entering metro area homes in high concentrations.  Readers of Resnikoff’s hyperbolic report will recall he said this about radon concentrations in the Marcellus Shale (emphasis added):

But, now comes the USGS with actual wellhead data, which is detailed in the following table:

Excerpt from USGS Report Entitled "Radon-222 Content of Natural Gas Samples from Upper and Middle Devonian Sandstone and Shale Reservoirs in Pennsylvania: Preliminary Data"

Notice the range of findings.  It is as low as 1 pCi/L and never reads higher than 79 pCi/L.  That’s a far cry from Resnikoff’s speculation the Marcellus Shale could produce radon at the wellhead as high as 2,576 pCi/L.  The USGS summarized its findings as follows (emphasis added):

Samples of natural gas were collected as part of a study of formation water chemistry in oil and gas reservoirs in the Appalachian Basin. Nineteen samples (plus two duplicates) were collected from 11 wells producing gas from Upper Devonian sandstones and the Middle Devonian Marcellus Shale in Pennsylvania.  The samples were collected from valves located between the wellhead and the gas-water separator.  Analyses of the radon content of the gas indicated 222Rn (radon-222) activities ranging from 1 to 79 picocuries per liter (pCi/L) with an overall median of 37 pCi/L.  The radon activities of the Upper Devonian sandstone samples overlap to a large degree with the activities of the Marcellus Shale samples.

Resnikoff’s range, in other words, is 32 to 37 times that the USGS found in testing Marcellus Shale wells, which, incidentally, produced radon levels indistinguishable from other Upper Devonian formations.  Worse, here is Resnikoff’s conclusion:

The potential environmental and public health impact of radon in natural gas from the Marcellus Shale formation is enormous. This paper has calculated the number of lung cancers in New York State as ranging between 1,182 and 30,448. This calculation is based on reasonable assumptions for a gas well in the Marcellus Shale, including the concentration of radon at the wellheads, the transit time between wellheads and homes, the dilution expected in a typical household, and reasonable risk factors drawn from studies by the US Environmental Protection Agency.

And, here is how he calculated that WAG, although he was obviously trying for a SWAG:

Notice Resnikoff STARTS his range in this instance with 100 pCi/L or almost three times the 37 pCi/L median figure observed by the USGS and 100 times the lower end of the USGS range.  So, even assuming his numbers with respect to risk were correct, and they are anything but as noted in Lynn Alspaugh’s critique, he starts out with a completely false assumption regarding radon levels.

Obviously, this guy has never read the instructions about converting a WAG into a SWAG or he would have noticed the first rule is “start your estimate with the  shortest or lowest estimate that is reasonable (and explain) why it isn’t reasonable to be shorter/lower.”  Resnikoff, in fact, starts his range of risk calculations at nearly three times what his own report estimates is the low end of the range of wellhead radon levels.  What kind of expert does this?  Well, the anti-expert, of course.

The USGS report, of course, doesn’t offer enough data to draw final conclusions, a fact duly noted in their write-up, but it does demonstrate one thing; the Resnikoff report is as baseless as a politician’s promise and that explains a lot, because his work is a political one, not a scientific one.  It confirms, too, the pipeline readings analyzed by Anspaugh, which ranged from 16.9 to 44.1 pCi/L and we’ve learned something else about those readings since reporting earlier.

All the samples other than at Mahwah and Lambertville were taken directly off the supply lines from Marcellus wells.  This is to say there was no mixing with conventional pipeline gas for those samples.  This means the range for Marcellus Shale gas was 27.6 to 44.1 pCi/L, right in line with USGS.  Moreover, the “mixed” samples were taken at Mahwah and Lambertville and on those days, approximately 60% of the throughput through Lambertville was from Marcellus Shale supplies.  This resulted in even lower concentrations due to the mixing and some further decay, just as one would expect based on the real science and not some WAG or SWAG.

There doesn’t seem to be much of anything left to Resnikoff’s work at this point.  When the courts, real experts and government agencies don’t accept your work, it’s hard to claim the mantel of “expert.”  But, he’s got that “anti-expert” thing nailed to the wall.

Follow us on Facebook and Twitter!