Data driven discussions about gas extraction and related topics.

PA Marcellus Shale Violations by Operator and County


This post has been archived.

Earlier this month, I examined 2010 oil and gas violation data for Pennsylvania on a summary level. Now I’m going to focus just on Marcellus Shale wells, and from the perspective of trying to determine the extent that drilling operators and location have on the likelihood of increased violations for any given well.

There are a number of ways in which this could be done. Although I am interested to see if the number of wells drilled in an area or by a drilling operator has a noticeable effect on violations, I have limited both investigations to 10 or more Marcellus Shale wells, both to avoid the erratic results of small sample sizes, and because the large number of results tends to make for crowded data displays.

Violations are only for 2010, both because it is current, and with 1,544 records, it is sufficiently large to make some generalizations about. As for wells, I decided to go for all drilled Marcellus Shale wells rather than just the 2010 wells, because wells that were spudded between 2006 and 2009 might well be included on the violations list. Also, the overall number of wells is a handy way to gauge the relative weight of a given operator or county on the industry as a whole.

2010 Marcellus Shale (MS) violations per total MS well, for counties with 10 or more MS wells.

2010 MS violations per total MS well, for operators with 10 or more MS wells.

There are, of course, a number of issues once we start looking at the data in detail. One of the biggest challenges is that many of these wells change hands, or else the companies that drill them change hands or are no longer active in the Marcellus Shale drilling industry in Pennsylvania. For example, in a previous analysis, I noted that Turm Oil had a large number of violations per well, and now they had none at all. I checked to see when their most recent well was drilled, and found my answer:

List of operators with 10 or more MS wells, with 2010 violations, frequencies per well, and the date of the most recent well drilled

Turm Oil hasn’t drilled a Marcellus well in almost two years. Looking at the chart above, Turm and Dominion don’t really belong in this analysis, and Eastern American and Fortuna may not for 2011. These four operators account for everyone with 10 or more total Marcellus Shale wells and no violations in 2010 except for Consol—kudos to Consol!

Does the number of wells impact violations?

Clearly, there is a tremendous range of violations per well, both in regards to who the operator is, as well as where the well is located. My hypothesis going into this analysis was that the more wells an operator drilled or the more wells drilled in a county, the fewer violations per well there would be.

I had several reasons for suspecting this to be the case. First of all, think of what must be done for a drilling operation without violations being issued. The site must be carefully placed, in accordance with all applicable regulations. The site must be prepared, taking into account all ground disturbing regulations. Then you have to drill through thousands of feet of rock, making sure that there is a structurally sound casing and cementing job to prevent gas migrations. Then there is the horizontal drilling, and the hydraulic fracturing. All along the way, not a drop of fracing fluid, diesel fuel, brine, or really anything else can touch the ground or enter Pennsylvania’s waters in any way. And when the drilling is done, the site must be restored in a timely fashion. And those are just a few of the regulations that the drilling operators agree to when they undertake drilling operations in Pennsylvania.

Let’s face it—that’s a lot to achieve. In order to consistently come in and out of a site with a clean record has got to take some practice. I’ve heard from industry sources that Marcellus Shale wells cost about $5 million to drill, so to do it right without cutting corners clearly takes significant resources as well.

In addition to all of that, a county level map of Marcellus Shale violations per well shows that the counties which produce the most Marcellus Shale gas wells are not the ones with the highest number of violations per well.

Map showing Marcellus Shale violations per county. Counties outlined in red have 100 or more Marcellus Shale wells. Click the gray compass rose and double carat (^) to hide those menus.

So is there then any correlation between the number of wells drilled and the violations issued? Let’s take a look (1).

The equations for the trendlines were calculated by Excel, and I selected the ones with the highest R-squared values. I was surprised that the best fit for operators was a convex polynomial. In the graph above, there is indeed a cluster of operators with between 300 and 400 wells, and with about 0.5 violations per well or less, but on the other end of the spectrum, the violations per well are spread far apart. There are so many operators big and small with around 0.5 violations per well or less that it seems some other factors must be at play for all of the operators that exceed that value by a significant margin. Perhaps not everyone will achieve zero violations like Consol did in 2010, but it doesn’t seem reasonable that Williams Production Appalachia should have more than twice the amount of violations as Range Resources Appalachia, despite having only 8 percent of the number of wells as the gas extraction giant.

The correlation was a bit stronger for the counties than for the operators. I’m not entirely sure what factors are at play here, other than perhaps having a crew that is well familiar with the geology of the region and all the specific challenges associated with that. Again though, that doesn’t seem to explain why Wyoming County would have 3.44 violations for every Marcellus Shale well, while those in Washington county can only expect a violation for less than one well out of eight drilled.

[photo removed]

Again, I’m sure that the way this analysis was set up had an effect on some of these numbers, and Wyoming is certainly a smaller sample size than Washington. I also don’t want to infer that there aren’t problems when large operators drill in well established portions of the Marcellus Shale. What’s more, while over 1,500 Marcellus Shale violations in a year is a huge number, it likely doesn’t account for all of the actual incidents, just the ones that the DEP can demonstrate, apparently beyond a shadow of a doubt (2). And even in the best scenario, there are significant impacts upon the land and neighboring residents near the well site.

I am suggesting, however, that there are companies that need to do better in their efforts to comply with laws designed to protect Pennsylvanians from pollution and other deleterious effects of oil and gas drilling. Excuses that environmental regulations are too strict don’t hold water, as their competitors are closer to compliance, sometimes dramatically so.

  1. In order for Excel to be able to calculate all of the various regression lines, counties and operators with zero violations per well had to be excluded from this analysis.
  2. I have personally talked to numerous residents who feel that their well water was spoiled by gas operations on neighboring lands. A common theme in their complaints is that the DEP places the burden of proof on residents that their wells were not spoiled before drilling operations began–an almost impossible situation for the residents to predict and be proactive about.

How Long Between MS Permit Issuance and Drilling in PA?

2011 Marcellus Shale Permits and Drilled Wells in PA (large)
2011 Marcellus Shale drilled wells (green circles) and permits issued (red stars). For a larger, dynamic view, please click the image.
Marcellus Shale Permits and Drilled Wells (large)
All Marcellus Shale permits issued (red circles) and drilled wells (green circles). Please zoom in for a closer look in the denser portions of the map.

Sometimes it seems like the oil and gas industry is in an awfully big hurry. They are in a hurry to get the mineral leases, presumably because if they don’t, some other operator will. They are in a hurry to get their drilling permits from the Department of Environmental Protection (DEP)–already this year, the DEP has issued 979 permits from the Marcellus Shale formation alone. And sometimes they are in a hurry to get the drill in the ground.  Sometimes, however, they are not.

This does not mean that I think the 444 Marcellus Shale wells that have been spudded (time when the drill first hits the ground) so far this year is a small number. After all, today is just the 104th day of the year, which means that on average, almost 4.3 Marcellus Shale wells are started every single day. That’s a lot of industrial activity, and yet it reflects well under half of the 9.5 Marcellus Shale permits that DEP secretary Michael Krancer signs off on every day.

The longer term trends are similar: Of the 6,092 Marcellus Shale wells with active permits(1), 2,574 have been drilled. That represents about 42 percent, meaning that the 45 percent clip for 2011 is actually running a bit ahead of schedule. All of this brings a couple questions to mind:

  • Why does the oil and gas industry get more than twice as many permits as they are able to drill?
  • What’s the lag time for drilling once the permit is in hand?

I’m still scratching my head over the first one. I have been told that the siting and permitting processes are so involved and expensive that once the permit is in hand, the industry will drill the site, but the numbers don’t seem to reflect that as being fully true. Certainly, the 107 oil and gas drilling rigs available in Pennsylvania right now is a limiting factor in how many wells are drilled, but that doesn’t explain why the permitting process is years ahead of the drilling queue.

As for how long it takes to drill once a permit has been issued, there are means of answering that question. First, I matched the permits data to the spuds data using the wells’ unique API numbers, finding 2,804 matches for 2,574 distinct wells (2)(3). The second step was to subtract the number of days between the spud date and the permit date to determine the lag time for those permits which have been drilled, and where API numbers did match up. Let’s take a look at the results:

Number of days between permit issuance and spud (initial drilling) date.

Some of the 39 wells marked as “reworked” may not have originally been Marcellus Shale wells, so they were not included in the chart above. In addition, there were two negative values, for which it would appear that well was drilled before the permit was issued. I am assuming those are attributable to clerical error, and those wells were not included in the chart above (4).

Number of days from permit issuance to spud date for Marcellus Shale wells. Please click the “i” and then a map feature for more information. Please click the gray compass rose and double carat (^) to hide those menus.

Overall, the value ranges from -86 to 2,274 days, with an average turnaround time of just over 100 days. If we omit the outliers discussed above, the values range from 1 to 566 days, with an average of just under 99 days.

After looking at these results, I am surprised by the vast range, and beyond the number of available rigs, I can only speculate as to what factors go into determining this. It also seems remarkable that there are wells that can get the equipment in place, the site prepared, and the drill in the ground the very next day after the permit was issued. And yet, for all of that celerity, sometimes it takes well over a year to start churning dirt.

  1. This data comes from the DEP’s Operators With Active Wells Inventory section of their Reports page. What I called “active permits” are actually “active wells” according to the DEP. These include all wells for which the permit has been issued but have not yet been plugged. This would include wells that hav not been drilled, thus my distinction.
  2. Both datasets had some duplication of well numbers. All records that were exact duplicates were removed, meaning that the remainder had at least slight variances in one or more columns.
  3. I should mention that the number of matches to the permits list means that there are 93 mismatches between the two datasets. In theory, all of the drilled wells should be on the permit report, but for now, let’s take the 97% match rate and move forward.
  4. All values are included in the posted dataset, and therefore the DataTool map.

Proposed Tire Fire Plant in Greenwood Twp., Crawford County, PA

In the fall of 2010 Crawford Renewable Energy, LLC (CRE) announced plans to build a “tire-fired” power plant in Greenwood Township of Crawford County. The facility is designed to produce 100 MW of energy by burning used, recycled tires in two circulating fluidized bed (CFB) boiler systems. The design of the facility includes several pollutant emission control technologies. These types of equipment remove a portion of the pollutants from the exhaust. As nice as it is to think of tires simply “disappearing” rather than being land-filled, when any hydrocarbon fuel source is burned, such as a tire or coal, a multitude of toxic and carcinogenic compounds are released. And most of these pollutants cannot be captured using control technologies, so they are emitted into the air.

The facility is planned on an 80 acre industrial park land parcel. The control equipment includes a CFB scrubber, a fabric filter baghouse, and a regenerative catalytic reactor. The flue gasses will then be emitted through a 325 foot tall stack. A CFB scrubber uses limestone to decrease sulfur emissions. The regenerative catalytic reactor is used to reduce NOX. The fabric filter baghouse is a series of screens and filters that remove the majority of the mass of particulate matter. The majority of the mass of particulate emissions are removed by capturing the coarse fraction of particles, which are particles with larger diameters and mass, but do not pose a significant health threat. Baghouses and other particulate control devices (PCDs) are not as efficient at capturing the fine and ultrafine fraction of particulate emissions, which have smaller diameters. The fine and ultrafine modes of particulates are the most hazardous, and are directly related to asthma exacerbation, chronic obstructive pulmonary disorder (COPD) and other forms of respiratory disease.

The emissions and deposition pattern from this facility were modeled by the Center for Healthy Environments and Communities to assess the impact on local air quality. Several pollutant species were modeled, including sulfur dioxide (SO2), oxides of nitrogen (NOx), and both the course and the fine fractions of particulate matter, PM10 and PM2.5 respectively. Concentrations of these pollutants at ground level in ambient air were modeled using the CalPUFF non-steady state dispersion model. These will not be the only pollutants transported, rather these are efficient to model. Plumes of some of the other contaminants will most likely have similar patterns.

The mean, or average, levels of predicted ambient air concentrations are presented first for each pollutant (Figures 1, 3, 5, and 7). These maps show the average concentrations of the pollutant that are predicted to occur while the facility is operating. The concentrations are averaged over a one year period. Next, peak day concentrations of pollutants are presented (Figures 2, 4, 6, and 8). These concentrations are the highest predicted concentrations for a single day that would occur when the facility is operating normally, over the one year modeled cycle. The concentrations shown in all of the maps are only attributable to the proposed facility, and do not include any other sources of pollution or background concentrations of pollutants. These values essentially show the increases in ambient air pollutants that will occur when the proposed facility is operating.

For this 80 acre industrial park, a square “fence-line” with 570 meter sides could surround the park. Typically, exposures are expected to be very limited within the fence-line because the area is inaccessible to the public. Concern is focused on the exposures that may occur beyond the limit of the fence-line. If the smokestack is assumed to be located at the center of the park, it would be at a distance approximately 235 meters from the fence-line. Using the scales on the maps, it is evident that the even the highest concentration gradients shown in the maps would occur beyond the fence-line. When the facility is operating, it is reasonable for the surrounding communities to expect exposures to even the highest concentration gradients shown in the maps.

Figure 1.  Mean values of modeled SO2 ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.
Figure 2.  Peak day values of modeled SO2 ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.
Figure 3.  Mean values of modeled NOX ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.
Figure 4.  Peak day values of modeled NOX ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.
Figure 5.  Mean values of modeled PM10 ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.
Figure 6.  Peak day values of modeled PM10 ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.
Figure 7.  Mean values of modeled PM2.5 ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.
Figure 8.  Peak day values of modeled PM2.5 ambient air concentrations at ground level, attributable to emissions from the proposed CRE plant.

A Look at Horizontal Well Production in Virginia

Horizontal gas wells in Virginia in 2008 to 2009. Note that they are all clustered in the extreme western portion of the Commonwealth.

According to the Virginia Department of Mines, Minerals, and Energy’s (DMME) Division of Gas and Oil, there are 30 horizontal gas wells that produced gas between January 2008 and December 2009 (1). While this is not a large number of horizontal wells, the dataset is interesting, since Virginia publishes monthly production data online.

Of the 30 wells, only nine were in production for at least 12 of the 24 months that I looked at. This is, admittedly, a small sample size, but is as good an entry point as any into the discussion of how gas production changes over time.

Chart 1: Production in Thousands of Cubic Feet (Mcf) of Horizontal Gas Wells in Virginia, with at least 12 months of production between 2008 to 2009.

Many of the wells in this analysis have a spike in production within the first few months of production, followed by a gradual decline.

Chart 2: Maximum, Minimum, and Mean Production of Horizontal Gas Wells in Virginia, with at least 12 months of production between 2008 to 2009.

For wells with at least 12 months of production, the mean production value tends to be closer to the minimum value than the maximum. This is particularly true for those wells which show a significant spike in production, such as VH-520008.

Chart 3: Ratio of Most Recent Monthly Production to Peak Monthly Production of Horizontal Gas Wells in Virginia, with at least 12 months of production between 2008 to 2009.

Ratio of December 2009 production to each well’s maximum monthly production for all horizontal gas wells. Please click the gray compass rose and double carat (^) to hide those menus. Click the “i” tool then any map feature for more information.

For these nine horizontal gas wells in Virginia, the average production of the most recent month (December 2009) is slightly less than 30 percent of the peak monthly production. This figure is skewed on the one side by a well with a tremendous production spike (VH-530008, 6.93%) and on the other by a well with low but relatively steady production (VH-536927, 42.75%). When all wells are considered (as with the map) the range of values is much greater.

Each of these wells had been in production between 12 and 24 months as of December 2009, and none of them produced even half as much gas in that month as the month for their respective maximum production values. The complete production data is available on FracTracker’s DataTool.

  1. The most recent production data currently available is for January 2010, one month after the end of this analysis.

Pennsylvania 2010 Oil and Gas Violation on FT’s DataTool

2010 Oil and Gas Violations per Drilled Well (small)
All oil and gas violations issued by the PA DEP in 2010, divided by the number of wells drilled in the same time frame, by county. Please click the map for a larger, dynamic view.

Two new violation datasets are up on FracTracker’s DataTool: 2010 Oil and Gas Violations in PA and 2010 Violations by County. The first set includes the raw data from the Pennsylvania Department of Environmental Protection (DEP)(1), and the second set includes violation and other oil and gas data at the county level.

Well Violation Data

See the legend for description of well type. Please click the gray tabs with the compass rose and the double carat (^) to hide those menus. for information on specific wells, click the “i” tool then any map feature.

There are a number of problems with this dataset. Altogether, there were 3,273 violations, but the total number of unique wells that represents is not known, because 271 of the violations didn’t even have a valid well API number associated with it. Since this data does not contain longitude, latitude, well type, or any indicators as to whether the violating wells were Marcellus Shale wells or horizontally drilled, none of this information can be known about these 271 violations.

2010 Oil and Gas Violations: Marcellus Shale, Other Formations, and Unknown Wells

In fact, of the remaining 3,011 violations, 665 are from wells where the API number do not match a compilation of over 40,000 permits from 1998 to 2010 which has been published on the DEP website. It’s a pity, since the rate of violations per offending well is lower than either of the other category then we must say that this value for both Marcellus Shale and non Marcellus Shale wells are overstated. We just don’t know by how much.

Violations per offending well type, January 2007 to September 2010

However, in a previous analysis over a 40 month period (including a nine month overlap with this data), the number of violations per offending wells were fairly comparable to the 2010 data. In the older dataset, offending Marcellus Shale wells were likely to have 1.47 times as many violations as their non Marcellus counterparts, and in the current data, that number is 1.44.

The most frequent violations are as follows:

Most frequently cited oil and gas violations in 2010 (2)

Here are the five wells which were issued the most citations in 2010:

Wells with most violations issued by the PA DEP in 2010

County Level Violation Data

The 2010 Violations by County dataset linked to above contains a wealth of county level oil and gas data for Pennsylvania. Also included are the number of drilled wells in 2010, July 2010 to December 2010 Marcellus Shale production data (3), as well as ratios of violations to both categories.

2010 Marcellus Shale Violations per Drilled Well (large)
2010 Marcellus Shale (MS) violations per 2010 MS well drilled. Please click the image for a dynamic view.

2010 Violations per Non Marcellus Shale Well (large)
2010 non MS violations per 2010 non MS drilled well. Please click the image for a dynamic view.

To my mind, it is notable that Washington and Green Counties in Southwestern Pennsylvania both have relatively few violations per well, despite the fact that they are both in the top five counties in terms of Marcellus Shale production.

Speaking of production, let’s take a look at that. While violations per well can give you an idea of what to expect for any new well in a geographic area, production from the Marcellus Shale is uneven. Some may argue that industry violations are more permissible in areas that yield more gas. Whether that argument holds water for you or not, violation per production amount is still a useful cost-benefit tool.

2010 MS Violations per Bcf of Gas Produced (large)
2010 MS violations per billion cubic feet (Bcf) produced by the MS between July 2010 and December 2010. Counties with at least 5 Bcf of production in that period are outlined in red. Please click the image for a dynamic view.

As was the case in Utah, a pattern is emerging where the most violations come from areas where drilling is relatively less well established or productive. None of the seven counties with at least 5 Bcf produced (outlined in red) are near the top of the violations per Bcf map.

  1. The dataset required heavy formatting to be transformed into a usable file. If you look at the original data linked above, you will note that there are boxes, in which values listed at the top apply to all boxes in that range. There are Excel tricks to allow for automatically filling these boxes, yet those could lead to significant error. There are instances where the box ends, but the spaces below are blank as well. My interpretation of this is that that values outside of the box are intended to be blank. It would be preferable if the DEP output filled in all of these cells appropriately, not only saving time, but reducing the chance for errors, and removing viewer interpretation as a factor in the dataset.
  2. The large number of “Failure to plug a well upon abandonment” for the “Unknown Formation” category may suggest that most of these wells are non Marcellus Shale, as many of those wells are older and more likely to be abandoned. In retrospect, I might have gotten more well number matches if I had used the PASDA list, which includes wells older than 1998, and last I checked, has over 120,000 wells in their database. PASDA data includes location, but no indication of whether the wells are Marcellus Shale or horizontally drilled.
  3. Unfortunately, there is no way to separate out Marcellus Shale production for the first half of the year, the data for which had been formatted to reflect a July to June fiscal year. Also, as of this writing, no production data for non Marcellus Shale wells for any part of 2010 is available.

SRBC Water Withdrawal Permits and Water Quality Monitoring

March 2011 SRBC Water Withdrawals (small)
Susquehanna River Basin Commission water withdrawal permits issued on March 10, 2011. Please click the image for a larger, more dynamic view.

Water Withdrawals

On March 24th, the Susquehanna River Basin Commission released notes about the public portion of their quarterly Commission meeting, which included a variety of water withdrawal permits. Specific locations were not included in the report, so the geographic information available on our DataTool is approximate.

March 2011 SRBC Water Withdrawals by Source Type. Please click the information tool (“i” button) then a map feature for more information. Please click on the gray compass rose and double carat (^) to hide those menus.

Water permits issued by the Susquehanna River Basin Commission at their March 2011 quarterly meeting by water source type, in millions of gallons per day.

March 2011 SRBC Water Withdrawals by Industry Type. Please click the information tool (“i” button) then a map feature for more information. Please click on the gray compass rose and double carat (^) to hide those menus.

Water permits issued by the Susquehanna River Basin Commission at their March 2011 quarterly meeting by applicant’s industry type, in millions of gallons per day.

The financial sector in the chart above is represented by Peoples Financial Services. Their own company website is almost completely useless, but the New York Times explains that they are a commercial and retail bank, primarily active in Susquehanna and Wyoming counties. There is no reason to think that a regional bank would go through a million gallons of water a day, so their permit request seems likely to be on behalf of one of their clients.

The total permitted amount approved on March 10, 2011 is 15.695 million gallons per day. According to the American Water Works Association, the average daily per capita residential water usage is 69.3 gallons, meaning that the water permits approved in the Susquehanna River Basin this month is the equivalent to the water usage of 226,479 people.

Remote Water Quality Monitoring Network

While we are discussing the Susquehanna River Basin Commission, they have an interesting tool called the Remote Water Quality Monitoring Network, which is a collection of solar powered water monitoring stations, and provides real time data for pH, conductance, dissolved oxygen, and turbidity. In browsing this for a moment or two, the pH level for Canacadea Creek near Almond, NY stuck out. It’s value of 3.87 is acidic enough to kill most fish and macroinvertebrates. The tool also has historic data, which shows that a month and a half ago, the pH from the same location was up at 8.79 pH units.

While I certainly hope that the SRBC and other authorities in New York figure out what’s going on in Canacadea Creek, I applaud the transparency that the Remote Water Quality Monitoring Network brings to the table. In the 21st Century, residents should have access to tools of this nature to alert them to real-time environmental challenges in their own communities.

Municipal Level Census Data Now on FT’s DataTool

2010 Municipal Population in Pennsylvania (small)
Municipal level census data in Pennsylvania for the year 2010. Click the image to see a larger, dynamic snapshot.

Municipal level census data is now available to visualize or download from FracTracker’s DataTool. Categories of note include:

  • 2010 US Census count
  • 2000 US Census count
  • Net change from 2000 to 2010
  • Percent change from 2000 to 2010

Among other uses, this dataset allows for some basic explorations of how the Marcellus Shale industry affects communities throughout the Commonwealth.

Pennsylvania population and Marcellus Shale gas production by municipality. For information on a specific municipality, please zoom in and click the “i” button in the blue circle, then the map feature of your choice. Please click on the gray compass rose and double carat (^) to hide those menus.

Without doing any serious number crunching, this map shows that gas from the Marcellus Shale is being extracted in more sparsely populated areas of the state. Let’s take a closer look at Southwestern Pennsylvania.

Southwestern Pennsylvania population and Marcellus Shale gas production by municipality. For information on a specific municipality, please zoom in and click the “i” button in the blue circle, then the map feature of your choice. Please click on the gray compass rose and double carat (^) to hide those menus.

Note the ring of Marcellus Shale production around the heavily populated municipalities surrounding Pittsburgh.

We can also take a look to see whether the Marcellus Shale gas industry had any obvious effect on populations in Pennsylvania.

Population change and the Marcellus Shale in Pennsylvania. Please zoom in and click the “i” button in the blue circle, then the map feature of your choice. Please click on the gray compass rose and double carat (^) to hide those menus.

Areas with the most population loss are white, and those with the largest gains are black. In addition, municipalities with Marcellus Shale production in the last half of 2010 are outlined in red, while those without are outlined in blue. With a cursory look, it appears that the areas with Marcellus Shale production are actually more likely to lose population–a topic that merits further analysis.

The municipal spatial data is from PennDOT (via PASDA), while the population data is of course from the US Census Bureau.

Ohio River Barium Concentration Trending Upward

The Ohio River Valley Water Sanitation Commission (ORSANCO) has been testing the main stem of the Ohio River for the presence of a variety of metals for some time, with results since 1998 published at their website. Mappable versions (1) of this data from 2010 are now posted on FracTracker’s DataTool as well. Over the years, the scope of the ORSANCO sampling efforts has broadened, both in the number of sampling locations as well as sampling frequency. In recent years, there are seventeen (2) locations, from which samples are obtained every odd numbered month. Currently, the most recent data available is July 2010.
Given the rapid surge in Marcellus Shale oil and gas drilling activity within the ORSANCO drainage basin and the millions of gallons of wastewater that ultimately finds its way into the Ohio River by way of numerous treatment plants and road deicing, I wanted to see if the impact of this industrial activity was reflected in the data.

[map archived]

I decided to take a look at barium concentrations. According to the Environmental Protection Agency, background levels of barium are not especially high in this region (3), noting:

…[Barium] occurs naturally in almost all (99.4%) surface waters examined, in concentration of 2 to 340 ug/l, with an average of 43 ug/l. The drainage basins with low mean concentration of barium (15 ug/l) occur in the western Great Lakes, and the highest mean concentration of 90 ug/l is in the southwestern drainage basins of the lower Mississippi Valley. In stream water and most groundwater, only traces of the element are present.

Barium is also a signature constituent of sorts of Marcellus Shale wastewater. According to this industry report, barium values range from 2,000 to 6,500 milligrams per liter in the wastewater.

[map archived]

This gives us an idea of how concentrations vary in space, at least on this occasion. Note that each of the first four testing locations downstream from the confluence of the Allegheny and Monongahela Rivers in Pittsburgh are among the highest group, with barium values in the 56.7 to 70.8 micrograms per liter (µg/L)range. These values are at once notably above the average background level and well below the EPA drinking water standard for barium of 2 milligrams per liter (4).

But what about changes over time? Marcellus Shale drilling activity has been increasing exponentially since the first well was drilled in 2006. Could this activity have any long term effects? To investigate this point, I compiled the barium amounts since 2006, and selected the three testing locations closest to Pennsylvania: New Cumberland Locks and Dam, Pike Island Locks and Dam, and Hannibal Locks and Dam.

ORCANCO barium values at New Cumberland, Pike Island, and Hannibal testing locations. Please click here for a larger view.

Right off the bat, we can see that there are significant seasonal variances, with peaks in late summer, and troughs in the late spring. That appears to be inversely proportional to the average flow rate of the Ohio River.

Average annual flow rate of the Ohio River at Wheeling, WV. Units are in Thousands of cubic feet per second (KCBS), and represent values between 9-1-98 and 2-29-08. Detailed flow data is available here.

Since barium values are clearly lower when there is more water in the river, it seems likely that such fluctuations would be due to dilution of pollution rather than natural circumstances.

Seasonal differences aside, the dashed trendlines of barium concentration show another story. Barium values are going up at all three locations. Significantly.

Approximate start and end values for the trendlines representing barium content in micrograms per liter at three testing locations on the Ohio River.

So while the recorded values themselves in the main stem of the northeastern portion of the Ohio River are not alarming, the fact that they are increasing so rapidly is a concern. It is worth bearing in mind that the values in some tributaries might be much higher, and that barium is only one of many pollutants associated with Marcellus Shale wastewater disposal.

Of course, none of this amounts to establishing causation between the Marcellus Shale industry and the elevated barium levels, but the circumstantial evidence is strong: barium values are very high in the wastewater, which is finding its way in large amounts into the Ohio River, where barium values are rising sharply.

  1. Locations were found with Google Maps, based on location description. In most cases, samples were taken from specific locks and dam structures, allowing for a fairly exact location. Some other locations are designated by the name of a small town, in which case, the mapped locations may be off by a mile or so.
  2. There is now an eighteenth testing location, McAlpine, 0.2 miles downstream of the Louisville testing station.
  3. While surface water is typically low in barium here, well water can be a significant issue:

    The drinking water of many communities in Illinois, Kentucky, Pennsylvania, & New Mexico contains concentrations of barium that may be 10 times higher than the drinking water standard. The source of these supplies is usually well water. Currently 60 ground water supplies and 1 surface water supply exceeds 1000 ug/l.

  • While these numbers are not alarming, it is worth noting that they are measured at an extremely well mixed area (locks and dams) of a massive river; at the time of this writing, the flow at Wheeling, WV was 134,700 cubic feet per second. Barium values on some tributaries could be much higher.

The Devil’s Details about Radioisotopes and Other Toxic Contaminants in Marcellus Shale Flowback Fluids

and Their Appearance in Surface Water Sources and Threats to Recreationalists, Private Well Water Users, and Municipal Water Supplies

By Conrad Dan Volz, DrPH, MPH – Director, Center for Healthy Environments and Communities, University of Pittsburgh Graduate School of Public Health

In yesterday’s FracTracker post, CHEC’s data manager Matt Kelso told the tale of two stories regarding radionuclides in Marcellus Shale flowback water and in river water as sampled by the PA DEP. As he said “the devil is in the details” and here are the “devil’s details” that put both stories into their proper public health context.
There are without doubt higher levels of radioisotopes in Marcellus Shaleflowback fluids than in the fracking fluids, which are injected under highpressure to fracture the shale layer. And in general problems related tonaturally occurring radioisotope buildup in the oil and gas industry arewell documented. The following is a passage from my expert testimony in theMatter of Delaware River Basin Commission Consolidated AdministrativeAdjudicatory Hearing on Natural Gas Exploratory Wells; Filed November23, 2010:

Elevated concentrations of naturally occurring radioactive materials(NORM), including 238U, 232Th and their progeny, are found inunderground geologic deposits and are often encountered during drillingfor oil and gas deposits (Rajaretnam G, and Spitz HB., 2000). Drill cuttingsfrom the Marcellus may be enriched in radium radionuclides and off-gas the radioelement radon. Also, the activity levels and/or availability ofnaturally occurring radionuclides can be significantly altered by processesin the oil, gas and mineral mining industries (B. Heaton and J. Lambley,2000). Scales in drilling and process equipment may become enrichedin radionuclides producing technologically enhanced naturally occurringradioactive materials (TENORM). Exposure to TENORM in drillingequipment may exceed OSHA and other regulatory authority standardsfor the protection of both human and ecological health. The occurrenceof TENORM concentrated through anthropogenic processes in soils atoil and gas wells and facilities represents one of the most challengingissues facing the Canadian and US oil and gas industry today (Saint-Fortet al., 2007). The risk of contamination of surface water and ground waterby TENORM accompanies the risk of soil contamination, as TENORMgenerated may runoff of drilling equipment during rain events or if onthe soil surface into surface water sources and/or enter groundwater bytransport through the unsaturated zone.

In a review article in Environmental Science and Technology (ES&T),authors Karbo, Wilhelm and Campbell (EPA Region III leads and Office ofRadiation and Indoor Air) stated:

New York’s Department of Environmental Conservation (NYDEC) reportedthat thirteen samples of wastewater from Marcellus Shale gas extractioncontained levels of radium-226 (226Ra) as high as 267 times the safedisposal limit and thousands of times the limit safe for people to drink.The New York Department of Health (NYDOH) analyzed three MarcellusShale production brine samples and found elevated gross alpha, grossbeta, and 226Ra in the production brine. Devonian-age shales containnaturally occurring radioactive material (NORM), such as uranium (U)and thorium (Th) and their daughter products, 226Ra and 228Ra. TheMarcellus Shale is considered to have elevated levels of NORMs. NORMsthat have been concentrated or exposed to the accessible environmentas a result of human activities, such as mineral extraction, are defined bythe EPA as technologically enhanced NORM (TENORM). TENORM maybe concentrated because of (1) temperature and pressure changes duringoil and gas production, (2) 226Ra and 228Ra in produced waters reactingwith barium sulfate (BaSO4) to form a scale in well tubulars and surfaceequipment, (3) 226Ra and 228Ra occurring in sludge that accumulates inpits and tanks, and (4) NORM occurring as radon (Rn) gas in the naturalgas stream.

If this flowback-produced water with elevated TENORM is disposed ofin sewage treatment facilities or other ineffective wastewater disposalprocesses – then the TENORM level in surface water (the receiving streamor river) will be largely determined by dilution offered by fluid flows withinthe waste plant and dilution offered by the water flows themselves in theriver or stream.

So, it is entirely possible that Marcellus Shale flowback and produced fluids(yes I hesitate to call it water because it is contaminated fluid – with manyidentified toxic contaminants; if this were coming from other industries itwould be a hazardous liquid waste) will have elevated levels of TENORM and many other contaminants (see explanation in Appendix 1 below) butlevels of TENORM in the surface water it is going into will not exceedbackground levels, seen in the stream or river system, once it is completelymixed in the stream or river.

But here is the devil in the details as Matt said in his article.Recreationalists fish and boat around these outfalls (this is documentedby CHEC in the Allegheny River Stewardship Project and the PittsburghFish Consumption Project), and we have no idea of the levels of TENORM(or other contaminants) in receiving water near the outfalls before fullriver mixing occurs. Additionally we have no idea of the level of long term bioaccumulation of TENORM (and other contaminants) in fish and otheraquatic resources that may frequent or live in areas where this material isdisposed of in.

Concentrations of TENORM and the many other contaminants in the effluent from treatment of oil and gas flowback fluids will vary in receivingstreams and rivers according to the flow of water in the receiving streamor river and their concentrations in the flowback fluids. Therefore, levels ofTENORM in receiving streams and rivers will reach a peak (everything elsebeing equal) during times of low flow – such as a drought or long periodswithout rain or snowmelt – and peak levels will be higher in the surface waternear the outfall then downstream in the river after it is mixed completelywith water flow from the stream or river. The PA DEP river water samplesfor radium were not taken during periods of low flow but during the fallseason when rain was more plentiful. Furthermore, they were not taken near outfallsof plants accepting oil and gas waste fluids for treatment, before completemixing occurs—therefore, peak levels in these areas were not captured bytheir sampling plan.

Additionally, levels of TENORM (and other contaminants) from sewagetreatment plants and inefficient brine treatment plants will be higher inlow volume streams (such as 10 Mile Creek in Greene and WashingtonCounties and Blacklick Creek in Indiana County) than in large volume riversystems like the Monongahela River. We simply don’t know what levels ofTENORM are like at peak levels in low volume streams during periods oflow flow or in areas just downstream of effluent outfalls before completemixing takes place.

CHEC has data showing that levels of bromides, barium and strontiumexiting the McKeesport POTW (a sewage treatment plant) vary over a day’s sampling; they aredependent on when the slug of produced-flowback brine is introduced intothe system and the slug’s rate of entry into the treatment system. At theMcKeesport POTW, it is customary that the slug of oil and gas waste fluidis introduced into the treatment system at 7pm. One sees that the levels ofthese contaminants in outfall effluent raises sharply over a short period oftime and then falls back to baseline (See CHEC figures 1, 2 and 3), whenthe slug is through the system. Any TENORM, in the oil and gas waste fluidbeing treated, not taken out by the treatment system will reasonably followthe same pattern. That is it will come and go quickly and we have no ideaof peak levels of TENORM or any other contaminants in the stream or rivernear the treatment plant outfall.

What is the solution to all this? Are we to sample continuously – at alltreatment plant outfalls, in river and stream segments between treatmentplant outfalls and water intakes, at all water intakes and in all finisheddrinking water (and I might add in private well water systems that may pullin contaminants from nearby streams and rivers) across the entire areaMarcellus Shale waste fluids are being disposed of? (This would includePennsylvania, New York, Ohio, and West Virginia). This is exactly whatis necessary to be done to assure protection of drinking water supplies,recreationalists, and the health of aquatic resources if we continue to allowoil and gas flowback water to be disposed of in sewage treatment andinefficient brine treatment plants.

NO – this would be cost prohibitive and impractical to do on the scale thatis necessary to protect public health and aquatic resources. We must usethe precautionary principal here and insist that sewage treatment plants notaccept oil and gas wastewater, period. Batches of oil and gas wastewaterneed to be tested continuously for levels of TENORM and all other possiblecontaminants so that a determination can be made of where the fluids canbe adequately and safely disposed of. Fluids that are determined to behazardous and/or toxic should be transported only by certified haulers andloads need to be properly manifested so there is an accurate accounting ofthe volumes of waste and where it is being sent for ultimate treatment. Thetechnical capabilities and acceptance of brine fluids, of and by, oil and gaswaste fluid treatment facilities must be matched exactly to the realities oflevels of contaminants in the brine fluids.

The intent of the Resource Conservation and Recovery Act (RCRA) wasto ensure that there is a “cradle to grave” system to document, handleand dispose of all hazardous and toxic waste from all industries and evenmunicipal authorities in a safe and effective manner. RCRA is basically anextension of the environmental public health precautionary principal – andif implemented and enforced thoughtfully and comprehensively preventsthe formation of new Superfund sites and will assure that the publicand environmental receptors are protected from contaminants in oil andgas waste fluids- be they called flowback or produced water, or brine oranything else.

Figure 1, Time-plot of Barium concentration in effluent from the McKeesportPOTW, sampled beginning 10/19/2010. Hour 1 begins at 19:00 (7:00 PM).A sample was taken on the hour, every hour, for a period of 24 hours. (To zoom in, click on the image.)

Figure 2, Time-plot of Strontium concentration in effluent from the McKeesport POTW, sampled beginning 10/19/2010. Hour 1 begins at 19:00 (7:00 PM). A sample was taken on the hour, every hour, for a period of 24 hours. (To zoom in, click on the image.)

Figure 3
, Time-plot of bromides concentration in effluent from the McKeesport POTW, sampled beginning 10/19/2010. Hour 1 begins at 19:00 (7:00 PM). A sample was taken on the hour, every hour, for a period of 24 hours. (To zoom in, click on the image.)

Appendix 1, Background Information

Hydraulic fracturing (HF) of shale gas deposits uses considerable masses of chemicals, for a variety of purposes to open and keep open pathways through which natural gas, oil and other production gases and liquids can flow to the well head. HF, also known as slick-water fracturing, introduces large volumes of amended water at high pressure into the gas bearing shale where it is in close contact with formation materials that are enriched in organic compounds, heavy metals and other elements, salts and radionuclides. Typically, about 1 million gallons and from 3-5 million gallons of amended water are needed to fracture a vertical well and horizontal well, respectively (Hayes, T; 2009, Vidic, R.; 2011). Fluids recovered from these wells can represent from 25% to 100% of the injected amended water solution (Vidic R., 2011) and are called “flowback” or “produced” water depending on the time period of their return.

Flowback and produced water contain high levels of total dissolved solids, chloride, heavy metals and elements as well as enriched levels of organic chemicals, bromide and radionuclides – in addition to the frac chemicals used to make the water slick-water. Levels of contaminants in flowback water generally increase with increasing time in contact with formation materials. There is abundant evidence that fluids recovered from this operation have high levels of total dissolved solids, barium and strontium, chlorides and bromides

While there is at present considerable scientific inquiry and even controversy regarding the potential of vertical or horizontal fracturing of shale gas reservoirs to contaminate shallow or confined groundwater aquifers (thus exposing municipal or private well water users to chemicals used in the hydrofracturing process and/or toxic elements, organic compounds, and radionuclides that exist in the formation materials); disposal of oil and gas wastewater/ Marcellus shale brine water in sewage treatment plants or inefficient brine wastewater treatment facilities is a direct exposure threat to public health through ingestion, inhalation and dermal absorption exposure pathways.

PA Marcellus Shale Production by Municipality

Average 6 month MS Well Production by Municipality (small)Marcellus Shale production by municipality. The darker red municipalities have higher production, illustrating that gas production in these gas wells comes in “hot spots”, particularly in the northeast and in the southwest.

It is no secret that there is money to be made in the natural gas industry, not only for the industry, but for those leasing their mineral rights as well. Pennsylvania law requires that a royalty of at least one eighth of the wellhead price of gas be paid to the owner of the land’s mineral rights.

And yet, we continue to hear stories, such as the one about Ron Gulla, who leased his land, which was subsequently damaged by drilling operations, all for apparently no money. How can this be? According to Mr. Gulla, it was because his gas was “wet gas” which needs to be processed. The DEP website makes no such distinction. Just to be sure, I called the Harrisburg office of the Bureau of Oil and Gas Management. Their response was that wet or dry, the drilling operators are required to pay at least one eighth of the wellhead price of gas as a royalty fee.

Is it possible that after all the drilling and hydraulic fracturing and dead fish and ruined farm that there was just no gas produced from that well?

First of all, let’s find Mr. Gulla’s former community. We know from the story referenced above that he was from Hickory in Washington County. In terms of this map, that places him in the middle of Mount Pleasant Township, so let’s take a closer look at what’s going on in that area.

Average six month gas production for Mount Pleasant Township in Washington County, PA, by municipality. Production values are from 7-1-10 to 12-31-10. Click the gray compass rose and double carat(^) to hide those menus.

If you click on the “i” button in the blue circle, then the red shape in the middle of the screen, we can learn quite a bit about gas production in Mount Pleasant Township. For example, we know that there were 94 wells in the township, each producing an average of 57.8 million cubic feet of gas in the six month period of July to December 2010, for an estimated minimum royalty check of just over $30,000. That’s a lot of gas and a lot of money. So where was Mr. Gulla’s?

Average six month gas production for Mount Pleasant Township in Washington County, PA, by municipality and by well. Production values are from 7-1-10 to 12-31-10. Click the gray compass rose and double carat(^) to hide those menus.

If the statewide trend is one of hotspots, at the township level, we are now looking at hotspots within hotspots. While many wells in Mount Pleasant Township produced over 300 million cubic feet of gas in the six month period, many others produced very little. And if Mount Pleasant is a moderately high producer of Marcellus Shale gas, and Chartiers Township to the southeast is a heavyweight, it makes it all the more curious that Cecil Township to the northeast and Smith Township to the northwest have no Marcellus Shale activity at all.

So maybe you have a neighbor who hit the jackpot with the Marcellus Shale gas boom, but does that mean that you will?

Average six month minimum royalty fees by township. Note the large number of municipalities with low or no royalty averages, and the very high dollar amounts in some other communities. Click on the gray compass rose and double carat (^) to hide those menus.

From this map, you can get an idea of what the average six month well royalty check might be for a well in your community. The figures for this map are based on the production values, above, times 0.125 times the average wellhead price of gas in 2010. But as we’ve seen in Mount Pleasant, there are some holes where the gas just doesn’t flow.

After taking this to another level of complexity, the lesson is pretty much the same as before: There is money to be made in the Marcellus Shale gas extraction industry–sometimes. As Mr. Gulla’s story reminds us, there are hardships as well. The DEP issued 9,370 oil and gas violations in a period of less than four years. Things can and sometimes do go wrong, and even when they don’t, around the clock industrial action for months on end in your backyard may be at odds with your bucolic lifestyle, or that of your neighbors.

So if you leased your land, would you cash in? At best, you can look at the numbers and play the odds, but there’s only one way to find out for sure.