Duke Univ. Marcellus Study

Meh. Not impressed with the study.
 
I had to laugh about the last quote about not wanting to deal with climate change. funny. so the use of natural gas will reverse the warming cycle of the earth the we have been experiencing for the past 10,000 years?
 
Their samples from near fracked wells were from Dimock. It tells us what we already know, there was a problem at Dimock which allowed methane to leak from a well or wells into the ground water.

A truly unbiased study would encompass samples from across various areas of the state and across multiple gas drillers. This would help determine if specific drillers had processes or materials issues. Or if there was a general issue with methane leakage. In addition it should sample wells prior to drilling and after drilling to identify changes that may have been caused by drilling. It's not as if there was a shortage of new sites where such studies could be done.

When I couple this with the Cornell report that claims coal has less impact on global warming that gas, I have to wonder if the science departments changed to political science since I went to school?
 
Methane in drinking water...
Elevated radiation in waste water from wells...
Accidents/explosions increasing as more are drilled...
Watchdog agencies unable to keep up with the volume of new wells...

In the words of the man on your avatar....
It just doesn't matter!

 
Want to see the actual study, not the MSM read out of it which doesn't allow you to see how the study was conducted.

If Franklin is right about all the "near well" sampling sites being near Dimock, then I agree 100% with him. It's just confirming something we all knew. If the results are consistent among all wells, then thats something different. But the article doesn't say one way or the other.
 
If anyone wants to read more:

http://www.bloomberg.com/news/2011-05-09/gas-in-marcellus-water-wells-is-an-explosion-risk-study-says.html
 
I'm surprised people think this some revolutionary finding. Exploration geologist have known about methane in well water for half a century or more.

One of our exercises in petroleum geochemistry involved plotting gas composition of well water on a structure map. Groups were tasked with proposing a wildcat well location and explaining why. The solution was to drill up-dip from water with high concentration, because they indicated a leak-point in the underlying reservoir. Bonus points for identifying structural closure and distinguishing the oil vs gas prone prospects. Cool thing about the exercise, its exactly how one of the largest oil fields in the Permian basin was discovered: by testing rancher's well water for dissolved gasses.

PA has different geology then W. TX of course but its generally agreed among geochemists that the source of most if not all the oil and gas in the state is from Devonian shales, Marcellus being just one of a half dozen or so. Highest gas concentrations roughly follow the natural fracture trends, which are, coincidentally, favored areas for hydraulic fracturing.

Its a fair question though. Does this new technology make things worse, about the same or better?
 
franklin wrote:
When I couple this with the Cornell report that claims coal has less impact on global warming then gas, I have to wonder if the science departments changed to political science since I went to school?

I think he was told to stop using university resources for his personal agenda or face disciplinary action. His conclusions have also been held up in the peer review process. Last I heard he was on his 9th re-write of the abstract.
 
All but one of their samples from active extraction sites was from nearby Dimock. Of the ones near Dimock all but one were concentrated in a small area, a line about 4 - 5 KM long. So of about 26 samples (it's hard to tell exactly how many, they don't say and you have to try to total up marks on a map and a graph) 24 seem to be concentrated in a very small geographic area.

Anyone want to bet that the water wells that were reported in the Gasland documentary is smack in the middle of the sample set?
 
Thanks troutbert. It's still not the actual peer reviewed article I'm looking for, but its definitely better than the cnn article.

Franklin, if what you're saying is true, then no question its a biased study. I'm just tryin to verify. Do you have a source?

Gone4Day, thanks, good info as always. I do have a question. I am aware that the gas can be chemically traced to a formation, likely even to an area within a formation. But I'm not sure how specific they can get. For instance, if 2 wells are 3 miles apart and tapping the same shale formation, could you chemically determine which exact well the gas came from?

In any case, the whole situation becomes a circular argument. If gas in aquifers is the result of natural fissures, and gas companies target these fissures, then of course there's going to be more gas in aquifers near gas wells. They chose those areas to drill because there's more gas in aquifers. For any sort of conclusion, you have to have before and after tests from the same wells.
 
You can find the abstract for the paper, and a link to the full text, here:
http://www.pnas.org/content/early/2011/05/02/1100682108
 
pcray1231 wrote:
Thanks troutbert. It's still not the actual peer reviewed article I'm looking for, but its definitely better than the cnn article.

Franklin, if what you're saying is true, then no question its a biased study. I'm just tryin to verify. Do you have a source?

Gone4Day, thanks, good info as always. I do have a question. I am aware that the gas can be chemically traced to a formation, likely even to an area within a formation. But I'm not sure how specific they can get. For instance, if 2 wells are 3 miles apart and tapping the same shale formation, could you chemically determine which exact well the gas came from?

In any case, the whole situation becomes a circular argument. If gas in aquifers is the result of natural fissures, and gas companies target these fissures, then of course there's going to be more gas in aquifers near gas wells. They chose those areas to drill because there's more gas in aquifers. For any sort of conclusion, you have to have before and after tests from the same wells.

The paper goes into details on how they conclude if methane in samples came from shallow or deep deposits. Most of the paper discusses the sample analysis not any detailed statistics relating to sample sites. I'd like to see the actual site data. How many different gas wells are associated with water sample sites? Why wasn't the testing conducted across a wider geographic area?

If you look at their chart on methane concentration for active extraction sites vs non-active maybe 60% of the active sites are of concern and the rest have similar methane concentrations as non-active sites. Which of the locations were the high concentration sites and which are the lower ones? They don't detail. What if all the high concentration sites are centered in one small location that they sampled? I'd like to know.
 
Here are some quotes from the Bloomberg article I linked:

"Water wells within 1 kilometer (0.6 miles) of gas drilling had 17 times more methane than more distant wells, according to the findings."

"The researchers tested 68 sites across five counties in Northeastern Pennsylvania and New York. Gases such as methane, ethane and propane were found in the water at 85 percent of the sites tested. The concentrations were higher the closer the samples were taken to gas drilling sites.

Researchers found that some of the gas sampled from the water wells originated from the same deep rock tapped by gas drillers, as opposed to the gas that naturally seeps from shallower formations. Those samples showed “specifically matching natural gas geochemistry from local gas wells,” the researchers said in their paper."
 
Thank you troutbert, thats what I was looking for.

At first glance, from Figure 1, about half of the "near well" sampling sites were near Dimock. And from Figure 3, about half of the "near well" sites were elevated with methane, with the others showing no elevated methane.

Gong a little deeper, they break it down a lot by formation. Comparing Figure 2 and Figure 1, and the text, it looks to me like the Lockhaven formation wells are in Bradford County, the Genessee formation is in New York, and the Catskill formation is in Susquehanna, Wayne, and Lackawanna Counties. Well over half of the Catskill formation wells are located very close to Dimock.

Table 1 is very telling. In the Genesee formation, they tested both active and nonactive sites, and found no methane increase in active sites. However, there is only 1 data point from active sites, so you don't know how to take that.

Lockhaven formation methane levels were fairly high, but they tested no "non active" data points there, so you have no idea if this is enriched or not, could be just natural. Thus, throughout the article, any reader should simply ignore the value of the Lockhaven formation.

The only place where there's any real result is, predicably, in the Catskill formation. It shows enrichment of methane in "active" sites was the Catskill formation. As this is an average, and they don't give data for individual sites, it is very possible/likely that ALL of this enrichment is due to the Dimock sites.

Further, in Figure 4A, examining the Catskill sites, there is a clear group of wells in the Catskill formation enriched with gas, and a clear group that is not enriched. Remember, again, that the "active" wells in the Catskill formation were a group of wells near Dimock, and then a few outside of that area. Figure 4B splits em up by type of gas, and there is a clear cluster of the Catskill sites, which indicates that these increases likely come from one location. Then there are some random scatter, which likely comes from the unenriched sites farther from Dimock.

It is impossible to tell if they found any clearly enriched sites outside of the Dimock area. But a reader must conclude from the data that ALL of the enrichment may have come from a small area near Dimock. This group of sites skews the average, so when averages are taken, the "active" sites are enriched when compared with the "non active" sites.

It is pretty much proof that in Dimock, the contamination is likely from drilling activity, which I think we all knew anyway. But I'd take nothing from it, for good or bad, regarding other wells.
 
pcray1231 wrote:It is pretty much proof that in Dimock, the contamination is likely from drilling activity, which I think we all knew anyway.

Don't forget that only recently the industry was denying that their drilling caused the problems at Dimock.

Now it's "we already knew that."

The Duke study isn't likely to be perfect, or the last word because there will probably be a lot more studies in the near future. But I wouldn't be so quick to say that it is just a "political" document, as some have said.

Duke is a major reseearch university, the paper is peer reviewed, and it was published by the National Academy of Sciences. It's not just someone typing on a blog.
 
Don't forget that only recently the industry was denying that their drilling caused the problems at Dimock.

True, but I think every 3rd party observer knew the truth on that one. Maybe not the method, but that it was related to drilling. Other than the industry reps, noone really defended them on it. I think even the courts, weren't they forced to pay damages there?

The Duke study isn't likely to be perfect, or the last word because there will probably be a lot more studies in the near future. But I wouldn't be so quick to say that it is just a "political" document, as some have said.

Agreed, and even the researchers said as much. Thats why I hate MSM articles about science. Issues like this, every study gives you a snippet of info, but only after 50 or 100 such studies does the picture start to get a little clearer. These studies are not final conclusions, rather, they're a method of conversation among scientists. You can't say a single study is garbage, its just a very small part of the whole discussion. But the media and political sides tend to take a single study and run with it as the gospel, when thats never what the scientists intended.

Likewise, as a scientist myself, I'm not going to sit here and say that there was intentional bias or anything here. It is what it is, these guys set out to test for contamination near gas wells. They started too late to get "before" measurements, they probably didn't get permission to test in as many places as they'd like. They ended up getting valid measurements on 4 or 5 wells. And they found that 1 of them was indeed contaminated. Not only that, but they essentially proved that the contamination was due to drilling activity. Not bad for one paper.

1 in 4 or 5 isn't enough to make a statistical conclusion when there are 1000's of wells out there. From the data, an overall rate of 1 in 2 or 1 in 10,000 are still possible. But getting actual valid statistical conclusions in a situation like this is a job way too big for 1 study. Another group will copy the study on 4 or 5 more wells, then another on 4 or 5 more, and eventually you collectively have a decent amount of data where maybe you can make some statistical conclusions.

 
pcray1231 wrote:
Gone4Day, thanks, good info as always. I do have a question. I am aware that the gas can be chemically traced to a formation, likely even to an area within a formation. But I'm not sure how specific they can get. For instance, if 2 wells are 3 miles apart and tapping the same shale formation, could you chemically determine which exact well the gas came from?

Just trying to inform the discussion. Short answer is no. The longer answer...

There are two types of gas found in nature - biogenic, formed from the decay of organic matter and thermogenic, formed from the thermal cracking of organic matter. The former is almost all methane while the latter contains longer chain hydrocarbons that can only be formed through thermal cracking. These two are easy to distinguish from their gas chromatograms. Shallow gas can be one or the other or a mixture of the two.

With themogenic gas, it is often possible to correlate them to their source rocks or each other using biomarkers and/or isotopic composition. However, the Devonian shales all have very similar organic matter and there is substantial mixing making it nearly impossible to correlate them with a specific shale layer. So basically all the thermogenic gas looks about the same and it would therefore be unlikely that you could distinguish gas of one well from another nearby. The best you can do is constrain it by depth based on maturity.

Looking at the study, its actually from the Nicholas School of the Environment. The Duke guy is the last author mentioned, which usually means he had the least to do with it. He probably runs the geochemistry lab at Duke. Generally these studies are attributed to the lead author, not the last contributor. So that's a little bit misleading.

Since they had to resort to comparing isotope ratios, it suggests that most if not all their samples are of mixed origin. All they are really showing is more thermogenic gas near gas wells. Its a chicken or egg argument. Do they drill because of higher gas content or is higher gas content because they drilled? Without before and after data, there's no way to tell.

The simple way to test a well for leakage is by pressure testing. If it can't maintain pressure, its leaking.
 
pcray1231 wrote:

They ended up getting valid measurements on 4 or 5 wells. And they found that 1 of them was indeed contaminated.

Can you explain further what you mean? They measured far more wells than that.

You're saying the great majority of the study was invalid. How so?
 
Not invalid. Just a small amount of info. What's invalid are the media conclusions, not the researchers'. The research conclusions are merely incomplete, which is going to happen nearly 100% of the time if you rely on a single study. I'll explain the best I can. The writers aren't real forthcoming with the data, you have to back it out, which I'm doing to the best of my ability. I'd prefer if they had a table with each test site, location, and results. But they only give you that in graphical form, without exact location, and for overall conclusions they rely on averages.

Active site = near gas well. Non active site = far from gas wells.

Genesee formation - located in NY. They tested 1 "active site", and 8 non active sites. There was no increase in methane at the active site, the non active sites actually contained more methane. But the active site is only 1 data point, which is hard to make conclusions from.

Loyalsock formation - located in Bradford County. They tested 7 active sites and zero non active sites. The active sites averaged fairly high in methane, but because they tested no nonactive sites, you can't take this as "enriched". There is no baseline comparison.

Catskill formation - all of the "conclusions" are drawn from here, the test sites are in Susquehanna, Lackawanna, and Wayne counties. Per Table I, they have 13 baseline "active" test sites, and 5 non-active that they used for data. The map in Figure 1 shows more data points than that, so I assume that means they have thrown out some invalid data from a few of the sites, which is normal (assuming the data was truly invalid). But also keep in mind, thats 13 test sites near active wells, not 13 different active wells.

From the map on Figure 1, it appears that 8 of the 13 "active" sites in the Catskill formation are from right around the Dimock well, and 5 are from other wells. For the Table I results, what they have done is averaged these results in with the other (unenriched) active well sites, and yes, shown enrichment of methane in "active" sites. But surprise surprise, if you look at Figure 4A, of the "active" Catskill data, there are exactly 8 tests that were enriched with methane, and 5 that fall along the baseline with methane concentrations around the same levels as the non-active sites.

Thus, it is not clear, but it is possible, even likely, that the ONLY sites that were enriched with methane due to drilling activity in this study were located very close to Dimock. By averaging those results in with other "active" sites in the Catskill, as well as active sites in the Loyalsock (high methane but no baseline comparison) and Genesee (active well not enriched), the average is indeed higher for active sites than non-active sites, but this all may or may not be attributed solely to a single gas well in Dimock.

 
Back
Top