So, along with the NYT's big piece on radioactivity in drilling wastewater, they provide a spreadsheet of well results that they reviewed. The above shows a screenshot of a portion of it. All that I have done is sort the data by "Gross Alpha". The rows represent different wells (mainly in Pennsylvania), and the "Gross Alpha" column is a measure of the amount of the radioactivity in the water, measured in Picocuries/Liter.
For background here, a Curie is
So then, the worst well on which "Gross alpha" was measured was Sampson #147-2H, and it had 40,880 picoCuries/Liter (ie 1513 individual atomic nuclei decay and release an alpha particle, per liter of water, per second of time).
Here is the EPA standard for gross alpha in drinking water:
So that standard is 15 piC/L, and we have many wells with tens of thousands of piC/L. The average of all the wells in the spreadsheet with a measurement for gross alpha is 4857piC/L, which is 323 times the EPA standard for drinking water.
So then the question is whether it's plausible that the drilling waste going into public waterways in Pennsylvania is diluted by a factor of more than a few hundred.
According to the Times article, 1.3 billion gallons of wastewater was produced in Pennsylvania in three years. Presumably more was produced in 2010 than 2008, so let's just take 1 billion gallons/year as a round number. Now, it's very unclear whether the 200 and some wells the Times provides data on are a random sample, or are in effect the worst cases out of thousands. This is a key issue and it's frustrating that the Times doesn't document it better (journalists!):
For proof that radioactive elements in drilling waste are not a concern, industry spokesmen and regulators often point to the results of wastewater tests from a 2009 draft report conducted by New York State and a 1995 report by Pennsylvania that found that radioactivity in drilling waste was not a threat. These two reports were based on samples from roughly 13 gas wells in New York and 29 in Pennsylvania.I cannot tell from this what the sampling properties of these ~200 wells are. Let's, for now, assume the worst: that the data are a random sample, and that the notional 1 billion gallons/year was all at 5000 piC/L on average (rounding up the 4857 piC/L from the spreadsheet). Now according to the Times, most of this went into two rivers:
But a review by The Times of more than 30,000 pages of federal, state and company records relating to more than 200 gas wells in Pennsylvania, 40 in West Virginia and 20 public and private wastewater treatment plants offers a fuller picture of the wastewater such wells produce and the threat it poses.
Most of the information was drawn from drilling reports from the last three years, obtained by visiting regional offices throughout Pennsylvania, and from documents or databases provided by state and federal regulators in response to records requests.
In Pennsylvania, these treatment plants discharged waste into some of the state’s major river basins. Greater amounts of the wastewater went to the Monongahela River, which provides drinking water to more than 800,000 people in the western part of the state, including Pittsburgh, and to the Susquehanna River, which feeds into Chesapeake Bay and provides drinking water to more than six million people, including some in Harrisburg and Baltimore.Here's a map of Pennsylvania rivers, for context:
data for the flow of the Susquehanna at Marietta, PA which is close to the Maryland border:
So ballpark around 30,000 cubic feet per second in this river in the winter. It will be more in spring and less in summer - let's take 20,000 cfs as a very ballpark number for an annual average.
Here's the flow of the Monongahela at Elizabeth, PA (which is just upstream of Pittsburgh):
So let's take that to be about 5000 cfs as an annual average (these graphs will only go back for 120 days, so we can only roughly project the annual picture - but I'm just trying to get the right order of magnitude here).
So, to summarize, the ballpark model here is 1 billion gallons/year of radioactive wastewater at 5000 pCi/L going into two rivers with a combined total flow of 25000 cubic feet per seconds, on average. So now it's just unit math. There are 365.25x24x3600 = 31557600 seconds in a year. There are 7.5 gallons in a cubic foot. So overall, the wastewater stream, as a fraction of the river flow, is 1,000,000,000/7.5/25,000/31,557,600 = 0.017%. So diluted by the full river flows, the 5000 piC/L becomes 0.85 piC/L - which we may as well call 1 piC/L (given the limited accuracy of this calculation).
So on an average basis, this says the residents of Baltimore, etc, downstream on the Susquehanna are probably ok (the standard, recall, was 15 piC/L). However, it would certainly seem like there would be huge potential for localities in Pennsylvania to exceed the safe standard. In particular, the Mononghela purportedly had more of the waste, but only has a fifth the flow of the Susquehanna, so appears to have much higher potential for problems. The major caveat here is the possibility that the NYT's reported spreadsheet of wells may be highly unrepresentative. If they are the cherry-picked worst cases, the overall problem will be much less in scale than the above calculation indicates.
Certainly, it seems that Pennsylvania drinking water treatments plants should all be doing continuous monitoring of radioactivity levels henceforth. Geiger counters are not expensive. This is something citizen groups could probably aspire to figure out, rather than waiting for the government to get around to it.