The source of elevated constituents is speculative, but... - Part 5

Without BTEX and methanol, the "powerful argument for continued research” Dr. Fontenot states on the the UT Arlington webpage describing their study titled "An evaluation of water quality in private drinking water wells near natural gas extraction sites in the Barnett Shale Formation" becomes less persuasive.

The paper concludes:

At a minimum, these data suggest that private wells located near natural gas wells may be at higher risk for elevated levels of constituents than those located further from natural gas wells.

No BTEX and methanol in these 92 wells would support the claim that the drilling fluids (fracturing) used are not entering into the ground water.

Based on the graphs they show:

I am in the process of looking at the data Dr. Fontenot sent.  That's going to take me a bit to get into.  Still, looking at these graphs, there are a lot of data points above the threshold (MCL), for arsenic in particular.

What does the paper say about all of this, and does the data support their conclusion?

Line 148:

Water wells [sampled] were overwhelmingly used for drinking water in rural areas without public drinking water systems (n = 82). The remaining wells were used to irrigate private lawns or provide drinking water for livestock (n = 18). To avoid contamination from pesticides, we did not sample water wells that were used for irrigating large agricultural crops.

Line 161:

Four duplicate water samples were collected in 40 mL glass vials without headspace and held at 4ºC during transport to The University of Texas at Arlington for chemical analyses. Because the objective of this study was to assess potential exposure risks of drinking water from wells in this region, we chose not to use filtration and acidification techniques. This allowed us to obtain samples representing the quality of water our participants would consume, as well as increased versatility in the number of constituents that could be probed by analytical techniques.

Okay...that could be problematic.  If you are going to compare a result to a threshold, you need to collect the sample and analyze it in a particular way.

Line 167:

We acknowledge that foregoing filtration and acidification can introduce a negative bias into metals analysis; however, this would result in a conservative underestimation of concentrations. Furthermore, the MCL values for drinking water are based on unfiltered samples that have not been acidified.

I am going to assume that the μg/L they report represent the amount of contaminant that would be consumed.  I do want to address this issue later because it brings in uncertainty which works against considering their "powerful argument."

If the MCL is the threshold they are using to denote the water is not impacted (ignoring the yellow and green coloring scheme they used), then the water should be tested the exact same way that drinking water samples are tested for these contaminants.

I am unsure what the statement "Furthermore, the MCL values for drinking water are based on unfiltered samples that have not been acidified," is based on (I see the citation, not sure why they conclude that to be true).  Here is what I know to be true about sampling arsenic in drinking water.

Source: EPA

The analytical method used by the paper is described as:

Chemical analyses were conducted using gas chromatography mass spectrometry (GC-MS), headspace-gas chromatography (HS-GC), and inductively coupled plasma-mass spectrometry (ICP-MS).

So if ICP-MS was used, EPA method 200.8 should have been used for the arsenic sample.  Footnote number 3 reads:

"Methods for the Determination of Metals in Environmental Samples - Supplement I," EPA-600/R-94-111, May 1994. Available at NTIS, PB 94-184942

In this same EPA document, they also show a table for the preservation of the sample:

Source: EPA

Here is what EPA Method 200.8 says about the sample:

This method provides procedures for determination of dissolved elements in ground waters, surface waters and drinking water.

and...

Dissolved elements are determined after suitable filtration and acid preservation.  In order to reduce potential interferences, dissolved solids should not exceed 0.2% (w/v) (Section 4.1.4).

And here is what the regulation in 40 CFR says:

I am not sure where they came to understand that "Furthermore, the MCL values for drinking water are based on unfiltered samples that have not been acidified," but that's not my take on it.

Nonetheless, they found elevated levels of arsenic in a number of the wells they sampled.  The other three metals, not so much.  I look at the data in their graph and ask "why?"

Here is what the paper says on line 259:

These constituent concentrations could be due to mechanisms other than contamination of aquifers with fluids used in natural gas extraction. For example, lowering of the water table can lead to changes in pH that cause desorption of arsenic and selenium from iron oxide complexes or mobilization of arsenic through pyrite oxidation.

The paper then goes on to say on line 267:

While the regional water table has not decreased dramatically in the last ten years, rural areas with high water withdrawal rates and/or withdrawal of large amounts of groundwater for use in hydraulic fracturing could lead to localized lowering of the water table.

...and then on line 273:

Additionally, pyrite is not found at high levels in these aquifers so it is an unlikely source of arsenic.

 In other words, the lowering of the water table could be the culprit, but the water table has not decreased, ...but it could decrease if the water is pumped for hydraulic fracturing, which would lower the table and release the arsenic from the iron pyrite.  But...there is low pyrite, so the arsenic cant be coming from that, and:

Given the low mobility of applied arsenic and the fact that none of our samples were collected from private wells in or adjacent to crop fields, we find agricultural arsenic introduction is unlikely to be the source of elevated arsenic concentrations.

Which leads them to say this on line 234:

Moreover, if agriculture were the cause of elevated arsenic levels, then concentrations in the historical data would likely have been high as well and we found no evidence of this.

So...because they find arsenic in the samples, and the historical data does not show arsenic, what changed after 1999?

This historical dataset is comprised of 330 private drinking water wells from the Trinity, Woodbine, and Nacatoch aquifers sampled over a ten year period (1989 – 1999) before natural gas activities began.

So...no arsenic in the historical dataset and arsenic in the current samples means that the arsenic is getting into the water most likely from natural gas activities.

That would be how I read their results and findings.  Seems pretty cut and dry, that is, until you look at the historical data after 1999.


Next post, Part 6


.