Current featured articles for the home page

Parked Oil Trains in Berks County, PA

By
Matt Kelso, Manager of Data & Technology
Kirk Jalbert, Manager of Community Based Research & Engagement

The Risks of Crude Oil Trains

As new oil fields boomed across North America in recent years, drillers looked for ways to get the product to refineries thousands of miles away. One solution was to use the nation’s rail infrastructure to ship hundreds of thousands of barrels of crude oil per day. The flow of oil was so great that thousands of additional tanker cars were ordered to get the oil to market. And yet, this solution of transporting crude by rail brought additional problems. Shipping large quantities of highly volatile and combustible crude oil on often antiquated rail lines has resulted in numerous accidents, at times spectacular in scale. In recent months, however, thousands of these oil tankers have been sitting idle on the tracks around the country, partially due to dropping oil prices, leading refineries to opt for cheaper imported oil and less expensive ways to get the domestic product to market such as through pipelines.

Communities Along the Tracks

The interactive story map below investigates a stretch of oil trains that have been parked for months in close proximity to homes, schools, and busy intersections in Berks County, Pennsylvania. Altogether, 30,494 people live in the seven communities through which the tracks in question pass. We began this project in response to concerns from residents who contacted FracTracker for assistance in understanding why these trains were located in their community, what hazards they might pose, and to help people bring this story to the public to foster meaningful discussions about the risks of parked oil trains.

Berks_staticmap


FracTracker has covered the risks of oil trains in a series of other articles. Click here to learn more.

FracTracker map of the density of wells by U.S. state as of 2015

1.7 Million Wells in the U.S. – A 2015 Update


 

Updated National Well Data

By Matt Kelso, Manager of Data & Technology

In February 2014, the FracTracker Alliance produced our first version of a national well data file and map, showing over 1.1 million active oil and gas wells in the United States. We have now updated that data, with the total of wells up to 1,666,715 active wells accounted for.

Density by state of active oil and gas wells in the United States. Click here to access the legend, details, and full map controls. Zoom in to see summaries by county, and zoom in further to see individual well data. Texas contains state and county totals only, and North Carolina is not included in this map. 

While 1.7 million wells is a substantial increase over last year’s total of 1.1 million, it is mostly attributable to differences in how we counted wells this time around, and should not be interpreted as a huge increase in activity over the past 15 months or so. Last year, we attempted to capture those wells that seemed to be producing oil and gas, or about ready to produce. This year, we took a more inclusive definition. Primarily, the additional half-million wells can be accounted for by including wells listed as dry holes, and the inclusion of more types of injection wells. Basically anything with an API number that was not described as permanently plugged was included this time around.

Data for North Carolina are not included, because they did not respond to three email inquiries about their oil and gas data. However, in last year’s national map aggregation, we were told that there were only two active wells in the state. Similarly, we do not have individual well data for Texas, and we use a published list of well counts by county in its place. Last year, we assumed that because there was a charge for the dataset, we would be unable to republish well data. In discussions with the Railroad Commission, we have learned that the data can in fact be republished. However, technical difficulties with their datasets persist, and data that we have purchased lacked location values, despite metadata suggesting that it would be included. So in short, we still don’t have Texas well data, even though it is technically available.

Wells by Type and Status

Each state is responsible for what their oil and gas data looks like, so a simple analysis of something as ostensibly straightforward as what type of well has been drilled can be surprisingly complicated when looking across state lines. Additionally, some states combine the well type and well status into a single data field, making comparisons even more opaque.

Top 10 of 371 published well types for wells in the United States.

Top 10 of 371 published well types for wells in the United States.

Among all of the oil producing states, there are 371 different published well types. This data is “raw,” meaning that no effort has been made to combine similar entries, so “gas, oil” is counted separately from “GAS OIL,” and “Bad Data” has not been combined with “N/A,” either. Conforming data from different sources is an exercise that gets out of hand rather quickly, and utility over using the original published data is questionable, as well. We share this information, primarily to demonstrate the messy state of the data. Many states combine their well type and well status data into a single column, while others keep them separate. Unfortunately, the most frequent well type was blank, either because states did not publish well types, or they did not publish them for all of their wells.

There are no national standards for publishing oil and gas data – a serious barrier to data transparency and the most important takeaway from this exercise… 

Wells by Location

Active oil and gas wells in 2015 by state. Except for Texas, all data were aggregated published well coordinates.

Active oil and gas wells in 2015 by state. Except for Texas, all data were aggregated published well coordinates.

There are oil and gas wells in 35 of the 50 states (70%) in the United States, and 1,673 out of 3,144 (53%) of all county and county equivalent areas. The number of wells per state ranges from 57 in Maryland to 291,996 in Texas. There are 135 counties with a single well, while the highest count is in Kern County, California, host to 77,497 active wells.

With the exception of Texas, where the data are based on published lists of well county by county, the state and county well counts were determined by the location of the well coordinates. Because of this, any errors in the original well’s location data could lead to mistakes in the state and county summary files. Any wells that are offshore are not included, either. Altogether, there are about 6,000 wells (0.4%) are missing from the state and county files.

Wells by Operator

There are a staggering number of oil and gas operators in the United States. In a recent project with the National Resources Defense Council, we looked at violations across the few states that publish such data, and only for the 68 operators that were identified previously as having the largest lease acreage nationwide. Even for this task, we had to follow a spreadsheet of which companies were subsidiaries of others, and sometimes the inclusion of an entity like “Williams” on the list came down to a judgement call as to whether we had the correct company or not.

No such effort was undertaken for this analysis. So in Pennsylvania, wells drilled by the operator Exco Resources PA, Inc. are not included with those drilled by Exco Resources PA, Llc., even though they are presumably the same entity. It just isn’t feasible to systematically go through thousands of operators to determine which operators are owned by whom, so we left the data as is. Results, therefore, should be taken with a brine truck’s worth of salt.

Top 10 wells by operator in the US, excluding Texas. Unknown operators are highlighted in red.

Top 10 wells by operator in the US, excluding Texas. Unknown operators are highlighted in red.

Texas does publish wells by operator, but as with so much of their data, it’s just not worth the effort that it takes to process it. First, they process it into thirteen different files, then publish it in PDF format, requiring special software to convert the data to spreadsheet format. Suffice to say, there are thousands of operators of active oil and gas wells in the Lone Star State.

Not counting Texas, there are 39,693 different operators listed in the United States. However, many of those listed are some version of “we don’t know whose well this is.” Sorting the operators by the number of wells that they are listed as having, we see four of the top ten operators are in fact unknown, including the top three positions.

Summary

The state of oil and gas data in the United States is clearly in shambles. As long as there are no national standards for data transparency, we can expect this trend to continue. The data that we looked for in this file is what we consider to be bare bones: well name, well type, well status, slant (directional, vertical, or horizontal), operator, and location. In none of these categories can we say that we have a satisfactory sense of what is going on nationally.

Click on the above button to download the three sets of data we used to make the dynamic map (once you are zoomed in to a state level). The full dataset was broken into three parts due to the large file sizes.

Utica Drilling in Pennsylvania

In Pennsylvania, the vast majority of unconventional oil and gas activity is focused on the Marcellus Shale formation, a Devonian period deposit of black shale with a high hydrocarbon content, which requires horizontal drilling and large scale hydraulic fracturing to produce enough oil and gas to make the drilling economically viable.  This formation was created about 390 million years ago, when organic-rich deposits accumulated in what is now the Appalachian Mountains, but was at that time a shallow sea.  Down below the base of the Marcellus lies the Utica Shale, an Ordovician period formation, with almost the same geographic extent as the Marcellus, but the deposits were placed there about 65 million years earlier.


Utica permits and violations in Pennsylvania. Click here to access the legend and other map tools.

In neighboring Ohio, it is the Utica that gets most of the attention, with 937 permitted wells, as opposed to just 20 for the Marcellus.  In Pennsylvania, the reverse is true:  there are 16,110 permitted Marcellus wells, but only 279 permits for Utica wells.  Part of the reason for this is because the subsurface characteristics of these formations vary widely, especially in terms of thickness and depth.  With changes in depth come changes in temperature and pressure, which are key criteria in hydrocarbon formation.  In other words, the same formation that produces considerable quantities of gas and valuable liquid hydrocarbons in eastern Ohio may be economically unviable just a county or two over in western Pennsylvania.

Utica shale permits, drilled wells, violations, and violations per well for Pennsylvania, through June 19, 2015.

Utica shale permits, drilled wells, violations, and violations per well for Pennsylvania, through June 19, 2015.

Utica drilling permits have been issued in 19 different counties in Pennsylvania, with wells having been drilled in 15 of those.  The violations per well (VpW) score for Utica wells in the Keystone State is 0.9, meaning that there are nine violations issued for every 10 wells that have been drilled.  It is worth noting, however, that only 36 of the 114 drilled wells have received violations, meaning that some wells have been cited on multiple occasions.

Of particular note is Bradford county, the site of only one Utica well, but 19 items on the compliance report.  The problematic Bayles 1 well was run by three different operators before being permanently plugged.  This well also has two “Drill Deeper” permits, and as a result, it is likely that the first six violations assessed to this well were issued before it was associated with the Utica Shale, as they precede the most recent spud date for the well in June, 2005.  Most of the violations for this well seem to be for pit violations and discharges to the ground and nearby stream.

Wells drilled into the Utica Formation in Pennsylvania, by year and current status.

Wells drilled into the Utica Formation in Pennsylvania, by year and current status.

In terms of drilling activity, it appears to have peaked in 2012, calling into question whether the industry considers the formation to be economically viable in Pennsylvania.  Of the 28 wells drilled since the beginning of 2014, Tioga County has seen the most activity with 11 wells drilled, followed by five wells in Butler County, then three in Lawrence County.  If we think of drilling activity as a sort of positive feedback from the industry – meaning that they like what they see and want to keep exploring – then only Tioga County seems to be holding the attention of the various operators who have been active in the Utica Shale.  Given the Utica activity in Ohio, one might have thought that counties on the western edge of the state – especially Beaver, Lawrence, and Mercer – would have shown the most promise, but this appears not to be the case.

Frac

Fracking’s Most Wanted – An NRDC Issue Paper

Lifting the Veil on Oil & Gas Company Spills & Violations

NRDC Issue Paper • April 2015

Today Natural Resources Defense Council (NRDC) released a report in conjunction with work by those of us at FracTracker Alliance.

We launched this investigation to determine what information about oil and gas company violations is publicly available on the Internet, how accessible it is, and whether it provides an adequate understanding about the practices of different companies.

This report highlights the information gaps about the frequency and nature of oil and gas company violations; such data is only publically accessible in 3 states – even though 36 states have active oil and gas development.

To take the review one step further, we analyzed the data that was available from these states – Pennsylvania, Colorado, and West Virginia. The results show that companies have been issued a series of violations, some of which were quite severe.

Of these companies, the following 10 had the most violations overall, in order of most to least:

  1. Chesapeake Energy (669)
  2. Cabot Oil and Gas (565)
  3. Talisman Energy (362))
  4. Range Resources (281)
  5. EXCO Resources (249)
  6. ExxonMobil (246)
  7. EQT Corporation (245)
  8. Anadarko Petroleum Corporation (235)
  9. Shell (223)
  10. Penn Virginia Corporation (186)

Find out more information, including the top violators in PA, CO, and WV, on NRDC’s website or by reading the full report (PDF)

Contact: Kate Slusark Kiely, 212-727-4592 or kkiely@nrdc.org

 

Danger Around the Bend

The Threat of Oil Trains in Pennsylvania

A PennEnvironment Report – Read Full Report (PDF)

On the heels of the West Virginia oil train explosion, this new study and interactive map show populations living in the evacuation zone of a potential oil train crash.

PA Oil Train Routes Map

This dynamic map shows the population estimates in Pennsylvania that are within a half-mile of train tracks – the recommended evacuation distance in the event of a crude oil rail car explosion. Zoom in for further detail or view fullscreen.

Danger Around the Bend Summary

The increasingly common practice of transporting Bakken Formation crude oil by rail from North Dakota to points across the nation—including Pennsylvania—poses a significant risk to the health, well-being, and safety of our communities.

This risk is due to a confluence of dangerous factors including, but not limited to:

  1. Bakken Formation crude oil is far more volatile and combustible than typical crude, making it an incredibly dangerous commodity to transport, especially over the nation’s antiquated rail lines.
  2. The routes for these trains often travel through highly populated cities, counties and neighborhoods — as well as near major drinking water sources.
  3. Bakken Formation crude is often shipped in massive amounts — often more than 100 cars, or over 3 million gallons per train.
  4. The nation’s existing laws to protect and inform the public, first responders, and decision makers are woefully inadequate to avert derailments and worst-case accidents from occurring.
Lac-Mégantic derailment. Source: http://en.wikipedia.org/wiki/Lac-M%C3%A9gantic_derailment

Lac-Mégantic derailment, July 2013. Source

In the past few years, production of Bakken crude oil has dramatically increased, resulting in greater quantities of this dangerous fuel being transported through our communities and across the nation every day. This increase has led to more derailments, accidents, and disasters involving oil trains and putting local com- munities at risk. In the past 2 years, there have been major disasters in Casselton, North Dakota; Lynchburg, Virginia; Pickens County, Alabama; and most recently, Mount Carbon, West Virginia. The worst of these was the town of Lac-Mégantic, in Canada’s Quebec Province. This catastrophic oil train accident took place on July 6, 2013, killing 47 people and leveling half the town.

Oil train accidents have not just taken place in other states, they have also happened closer to home. Pennsylvania has had three near misses in the last two years alone — one near Pittsburgh and two in Philadelphia. In all three cases, trains carrying this highly volatile Bakken crude derailed in densely populated areas, and in the derailment outside of Pittsburgh, 10,000 gallons of crude oil spilled. Fortunately these oil train accidents did not lead to explosions or fires.

All of these incidents point to one fact: that unless we take action to curb the growing threat of oil trains, the next time a derailment occurs an unsuspecting community may not be so lucky.

Bakken oil train routes often travel through high-density cities and neighborhoods, increasing the risk of a catastrophic accident for Pennsylvania’s residents. Reviewing GIS data and statewide rail routes from Oak Ridge National Laboratory, research by FracTracker and PennEnvironment show that millions of Pennsylvanians live within the potential evacuation zone (typically a half-mile radius around the train explosion ). Our findings include:

  • Over 3.9 million Pennsylvania residents live within a possible evacuation zone for an oil train accident.
  • These trains travel near homes, schools, and day cares, putting Pennsylvania’s youngest residents at risk. All told, more than 860,000 Pennsylvania children under the age of 18 live within the 1⁄2 mile potential evacuation zone for an oil train accident.
  • Philadelphia County has the highest at-risk population — Almost 710,000 people live within the half-mile evacuation zone. These areas include neighborhoods from the suburbs to Center City.
  • 16 of the 25 zip codes with the most people at risk — the top percentile in the state — are located in the city of Philadelphia.
  • The top five Pennsylvania cities with the most residents at risk are:
    • Philadelphia (709869, residents),
    • Pittsburgh (183,456 residents),
    • Reading (70,012 residents),
    • Scranton (61,004 residents), and
    • Erie (over 51,058 residents).

 

Bakken Crude Oil

How we get it and why we ship it

Bakken crude oil comes from drilling in the Bakken Formation, located in North Dakota. It contains deposits of both oil and natural gas, which can be accessed by hydraulic fracturing, or “fracking.” Until recent technological developments, the oil contained in the formation was too difficult to access to yield large production. But advances in this extraction technology since 2007 have transformed the area into a major oil producer — North Dakota now ranks second in the U.S. for oil production. The vast expansion of wells over the last 4 years (from 470 wells to over 3,300 today) means that there is more oil to transport to the market, both domestically and abroad. This increase is especially concerning considering that the U.S. Department of Transportation stated in early 2014 that Bakken crude oil may be more flammable than traditional crude, therefore making it more dangerous to transport by rail.

For More Information

Pennsylvania Data Discrepancies

By Matt Kelso, Manager of Data & Technology

The Pennsylvania Department of Environmental Protection (PADEP) publishes oil and gas well data in two different places: on their own website’s Spud Data Report, and in the Oil and Gas Locations file published on the PA Spatial Data Access repository, also known as PASDA. Because these two sources are both ultimately published by PADEP, it would stand to reason that the data sources would match up. Unfortunately, that is not the case. Learn more about the data discrepancies we uncovered:


This map shows those wells in Pennsylvania that only show up on one of the two data sources. Pink dots show wells that appear on PASDA but not the PADEP site, while the reverse is true for blue wells. Click here for the full screen view with additional map tools.

Methodology

Both of these data sources have existed for years. When FracTracker does analyses of PA, we usually use data directly from the PADEP site, because it includes far more information about the wells, such as the spud date, county, municipality, well configuration, and whether or not the well is classified as unconventional. Even though it has less information about each well, the data on PASDA is useful for expediently mapping the inventory of wells in the Keystone State. In this current analysis, we looked at both sources, and found significant discrepancies between the two.

Individual oil and gas wells have been given unique API numbers since the 1950’s. The overwhelming majority of items on both lists that we examined have these numbers, and those that do not have other numeric identifiers in their place. The uniqueness of the data in these columns is what we used to determine the number of wells on both lists. These columns in both data sources were then tested against one another using Microsoft Excel in order to determine which wells were included on both lists.

The data on PASDA is described as “Oil and Gas Locations,” and nothing in available metadata made it clear as to whether wells that were permitted but not yet drilled might be included in this or not. Additionally, we are mostly interested in wells that are still operational, assuming that there might be accuracy issues for historical wells in an industry that has been operational in the state since before the Civil War. We did, however, include orphaned and abandoned wells, as they remain a source of impact throughout the state.

Summary

PADEP_PASDA_descrep

Number of wells in PA in various categories. For brevity, “Total wells – Drilled and not plugged” is shown as “TW-DnP.”

We found 3,315 records of drilled, unplugged wells with location information on the PASDA dataset that are not on the PADEP search tool, and 96 such wells on the PADEP site that aren’t found on PASDA. Additionally, there are 35,434 drilled and unplugged wells in the PADEP data that lack location data, although six of these wells are actually on the PASDA site, meaning that there is some location data for them somewhere at PADEP.

For those of you who might be looking for discrepancies in our discrepancy table, one might expect the number of both wells that appear on both lists (the second to last row on the chart) to be identical. The biggest reason that they are not is that some wells appear in the PASDA dataset multiple times. There are 6,997 fewer unique wells than there are entries on the full file, or a 95.74% match rate. In comparison, the PADEP spud report only has 19 duplicates for over 204,000 wells, a 99.99% match between the number of wells and the number of records. Indeed, when we filter for unique wells, the difference between the two lists shrinks to only 40 records, which might be explained by differences is well statuses that were used to shape our analysis.

This chart shows the number of wells drilled per year in Susquehanna County, through 2/11/15.

Number of wells drilled per year in Susquehanna Co., through 2/11/15.

Undoubtedly, it will take some effort to get the two datasets to reflect the full set of wells in PA, but that is certainly a task than can be accomplished. The wells lacking location data are likely to be much more of a challenge. If we include all status types, there are 75,508 wells on the spud report that lack latitude and longitude values altogether, leaving us with only the county and municipality to determine where these wells are located. Hopefully, this crucial data exists somewhere in the PADEP inventory, and these wells are not in fact lost.

Finally, there are a couple of things to note about dates. Since the PASDA dataset does not include spud dates, it is impossible to determine the age of the majority of the mismatched wells. Looking at the pink dots on the interactive map above, though, it is clear that a large number of these mismatched PASDA wells are in the northeastern corner of the state that has been booming since the recent development of the Marcellus, but saw little to no development before that time – at least according to the spud report.

Of the 96 wells that are on the spud report but not PASDA, 67 are given the date “1/1/1800,” which seems to be a default date; over 94,000 wells on the report have this listed as the spud date. Most of the other wells that don’t match are relatively old wells, with spud dates ranging between 1960 and 1984. One of these wells was drilled on May 6, 1999 though, and four more were drilled on August 19, 2014.

The mismatched data can be accessed here for those who are interested.

Updated PA Data and Trends

By Matt Kelso, Manager of Data and Technology

The FracTracker Alliance periodically takes a deeper look into the unconventional oil and gas data in Pennsylvania, in order to provide updates for some frequently requested statistics on the industry. Here we provide updated PA data and trends as of December 4, 2014. Since unconventional drilling began in the Commonwealth permits have been issued to drill 15,573 unconventional wells, according to data from the Pennsylvania DEP. Many – 8,696 (56%) – of those permits have actually been drilled. In terms of violations, there have been 5,983 entries on the statewide Compliance Report for unconventional wells throughout the state, which are attributed to 1,790 distinct wells.

Pennsylvania Shale Viewer Map


Please click here for the full screen version, with additional map tools and controls.

Additional Stats

The number of permits, wells, and violations vary significantly from month to month, but each category is well off of its peak. The largest number of unconventional permits issued in a single month was 402, which was in December 2010, more than twice as many as were issued last month. In that year, there were six months with 300 or more permits issued, whereas there has only been one such month to date in 2014.

PA unconventional O&G activity per month from Jan. 2009 to Nov. 2014.  Source:  PADEP

PA unconventional O&G activity per month from Jan. 2009 to Nov. 2014. Source: PADEP

The 210 wells spudded (drilled) in August 2011 represents the high water mark, and is more than two times the amount of wells drilled last month. In the 28 months between March 2010 and June 2012, the industry failed to spud 100 wells only once, reaching 98 in April 2011. In the first 11 months of 2014, that plateau was missed three times, with a low of 58 spuds in February.

There was a significant spike in violations appearing on the compliance report from December 2009 through August 2011. More than 100 violations were issued in 17 out of 21 months, including 196 in March 2010. The number of violations issued has slowed down considerably since then, with November 2014 being the 34th straight month with fewer than 100 violations. Only 14 violations were issued in June 2014.

Violations per Well (VpW)

Unconventional violations per well by county in PA, showing the 10 counties with the largest number of violations.  Counties with an above average Violations per Well (VpW) score are highlighted in red.

Unconventional violations per well by county in PA, showing the 10 counties with the largest number of violations. Counties with an above average Violations per Well (VpW) score are highlighted in red.

We often ask whether drilling is more problematic in some areas than others. Since the number of wells varies depending upon the location, we must approach this question by looking at the number of violations issued per well drilled (VpW). However, there is an important caveat to consider. Put simply, what is a violation? The Pennsylvania DEP publishes a Compliance Report for unconventional wells, which has 5,983 incidents listed from 2000 through December 4, 2014. However, it used to be common for the DEP to lump several incidents into the same Violation ID number, although this is not the case for more recent infractions. When the DEP counts violations issued, they look at the total number of unique Violation ID numbers that have been issued, not the total number of incidents on the report. Here, we include the more inclusive list of items on the compliance report.

Of the 10 counties with the largest number of violations issued, only 3 counties have a violations per well mark below the statewide average. Notably, each of those three counties are located in Southwestern Pennsylvania. It is unclear from these numbers what is going on in Potter County, but clearly there is a significant problem in that location – with almost three violations issued per well drilled, Potter County has a VpW score 4.3 times the statewide average.

Operator Trends

Before we look at the operators with the most violations, there is an additional caveat to consider: It is relatively common for wells to change hands over their operational lifetimes. This characteristic could be due to one company buying another out, or simply transferring some of their assets. Still, wells changing from one operator to another is a normal aspect of the oil and gas industry. Such a fact matters for this analysis because while violations issued always stick with the responsible party in the DEP data, the name of the operator changes on the Spud Report to the current operator.

Unconventional violations per well by operator in PA, showing the 10 operators with the largest number of violations.  Operators with an above average Violations per Well (VpW) score are highlighted in red.

Unconventional violations per well by operator in PA, showing the 10 operators with the largest number of violations. Operators with an above average Violations per Well (VpW) score are highlighted in red.

Because of how these datasets are maintained, we see that East Resources has 261 violations for zero wells, which is of course an impossibly large ratio. That is because East sold off its stake in the Marcellus to Royal Dutch Shell, which does business as SWEPI in Pennsylvania. SWEPI, by the way, is 13th on the list of violations in its own right, with 154 violations for 675 wells, resulting in a 0.23 VpW. If the legacy violations for the old East wells were included, the result would be a 0.61 ViW score, which is almost three times as high, but still below the statewide average. FracTracker doesn’t do the analysis that way, both because it is unfair to the new operator to charge them with violations that they had nothing to do with, as well as being nearly impossible to keep track of the various transactions that result in wells changing hands over the years.


Cover image by Pete Stern, 2013.

In-depth Review of the Statoil Well Pad Fire

Commentary on Shale Gas Operations: First in a Series of Articles
By Bill Hughes, Community Liaison, FracTracker Alliance
Statoil Well Pad Fire: June 28-29, 2014

The early riser residents along Long Ridge Road in Monroe County are among the first in Ohio to see the sun coming up over the West Virginia hills.  It rose about 6:00 am on the morning of June 28th.  Everyone assumed that this would be a normal Saturday morning.  Well, at least as normal as it had been for the better part of two years since the site preparation and drilling started.

For those residents on Long Ridge who were not early risers, the blaring sirens, the smell of acrid smoke, and the presence of fire trucks and other emergency vehicles shortly after 9:00 am must surely have made them wonder if they were in the midst of a nightmare. A quick glance outside toward the Statoil Eisenbarth well pad and they would have seen this view:

Statoil 1

Figure 1. View from the southeast, as the fire spread on Sat. June 28th

The image in Fig. 1 would be enough to make most folks feel somewhat panicky and consider evacuating the neighborhood. That is exactly what soon happened – definitely not the start of a normal Saturday morning.

Adjusting to the New Normal

The traffic in the area had been a problem ever since site preparation started on the nearby well pad. The State expected the drillers to keep up the road. Crews also provided lead escort vehicles to help the many big trucks negotiate the narrow road way and to clear the residential traffic. Access to the well site required trucks to climb a two-mile hill up to the ridge top.

Statoil 2

Fig. 2. Neighbors’ views of the fire

Until June 28th, most folks had become accustomed to the extra noise, diesel fumes, and congestion and delays that always come with any shale gas well exploration and development in the Marcellus shale gas active area. Most of the neighbors had gotten used to the new normal and reluctantly tolerated it. Even that was about to change, dramatically.  As the sun got higher in the eastern sky over WV, around 9:00 AM, suddenly the sky started to turn dark. Very dark. Sirens wailed. Red trucks started a frenzied rush down Long Ridge from all directions. There was a fire on the well pad. Soon it became a very large, all consuming fire.  Smoke, fire, bitter fumes, and no one seemed to know yet exactly what had happened, and what was likely to happen soon.

This gas well location, called the Eisenbarth pad, recently changed operators. In January 2013, the well pad property and its existing well and equipment were bought out by Statoil, a company based in Norway.  Statoil had since drilled seven more wells, and even more were planned.  The original single well was in production.  Now in late spring and early summer of 2014 the new wells were to be “fracked.”  That means they were ready to be hydraulically fractured, a procedure that follows the completion of the drilling process.

Statoil hired as their fracturing sub-contractor Halliburton. All of the fracturing pump trucks, sand kings, Sand Castles, and control equipment were owned and operated by Halliburton.  The fracturing process had been ongoing for some weeks when the fire started. The eastern Ohio neighbors now watched ~$25 million worth of equipment go up in smoke and flames (Fig. 2). The billowing smoke was visible for over 10 miles.

Industrial accidents are not rare in the Ohio Valley

Many of the residents nearby had worked in the coal mining industry, aluminum plants, chemical plants, or the coal fired power plant that were up and down the Ohio River. Many had since retired and had their own industrial accident stories to tell. These were frequently private stories, however, which mostly just their co-workers knew about. In an industrial plant, the common four walls and a roof kept the dangerous processes confined and enabled a trained response to the accidents. The traditional, industrial workplace had well-proven, customized workplace safety standards.  Professional maintenance personnel were always nearby.  In stark contrast, unconventional gas well pads located in our rural communities are very different. They are put in our hayfields, near our homes, in our pastures and just down the road. You cannot hide a community accident like this.

Sept 2014 Update: Video of the fire, Copyright Ed Wade, Jr.

Print Media Coverage of the Fire

Within days, many newspapers were covering the well pad fire story. The two nearby weekly newspapers, one in Monroe County, Ohio and the other in Wetzel County, West Virginia both had detailed, long articles the following week.

Statoil 3

Fig. 3. View from the east as the fire started

The Monroe County Beacon on July 2, 2014 said that the fire spread quickly from the small original fire which was totally surrounded within the tangled complex of equipment and high pressure piping.  Early Saturday morning, the first responder would likely have seen a rather small somewhat localized fire as shown in Fig. 2. The photo to the right (Fig. 3) is the view from the east, where the access road is on Long Ridge road. This point is the only access into the Statoil well pad. The view below, showing some still intact tanker trucks in the foreground, is looking west toward the well location. Pay attention to the couple of trucks still visible.

The Monroe County emergency director said it was his understanding that the fire began with a ruptured hydraulic hose. The fluid then ignited on a hot surface. He said, “…by 9:10 AM the fire had spread to other pumps on the location and was spreading rapidly over the well pad.”   Emergency responders needed water now, lots of it. There is only one narrow public road to the site at the top of a very long, steep hill and only one narrow entrance to the densely congested equipment on the pad.  Many Volunteer Fire Departments from both Ohio and West Virginia responded.  A series of tanker trucks began to haul as much water to the site as possible.  The combined efforts of all the fire departments were at best able to control or contain but not extinguish the powerful, intensely hot and growing blaze.  The Volunteer firemen did all they could. The EMS director and Statoil were very grateful for the service of the Volunteer Fire Departments. There was a major loss of most equipment, but none of the 45-50 workers on site were injured.

Statoil 4

Fig. 4. Well pad entrance

The article from the Wetzel Chronicle also praised the coordinated effort of all the many fire departments. At first they attempted to fight the fire, and then prudently focused on just trying to limit the damage and hoping it did not spread to the well heads and off the well pad itself. The New Martinsville fire chief also said that,  “… the abundance of chemicals and explosives on the site, made attempts to halt the fire challenging, if not nearly impossible… Numerous plans to attack the fire were thwarted each time by the fires and numerous explosions…”  The intense heat ignited anything nearby that was at all combustible. There was not much choice but to let the fire burn out.

Eventually the view at the well pad entrance as seen from the east (Fig. 3) would soon look like the overhead view (Fig. 5). This aerial imagery shows what little remained after the fire was out – just some aluminum scrap melted into the decking is left of the original, white Hydrochloric Acid tanker truck. Everything near it is has almost vaporized.

Statoil 5

Figure 5. Post-fire equipment identification

Efforts to Limit the Fire

Statoil 6

Fig. 6. Protected white trailer

An excellent example of VFD’s successfully limiting the spread of the fire and controlling the extreme heat can be seen in the photo to the right (Fig. 6). This white storage trailer sure seems to be a most favored, protected, special and valuable container. It was.

It was filled with some particularly dangerous inventory. The first EPA report explains it thus:

A water curtain was maintained, using pump lines on site, to prevent the fire from spreading to a trailer containing 1,100 pounds of SP Breaker (an oxidizer), 200 pounds of soda ash and compressed gas cylinders of oxygen (3-2000 lb.), acetylene (2-2000 lb.), propane (6-20 lb.), among miscellaneous aerosol cans.

Statoil 7

Fig. 7. Post-fire pad layout

Yes, this trailer got special treatment, as it should. It contained some hazardous material.  It was also at the far southwest corner of the well pad with minimal combustibles near it.  That was also the closest corner to the nearby holding pond, which early on might have held fresh water. Now the holding pond is surely very contaminated from flowback and runoff.

The trailer location can be seen in the picture to the right in the red box (Fig. 7), which also shows the complete well pad and surrounding area. However, in comparison to the one white storage trailer, the remainder of the well pad did not fare so well. It was all toast, and very burned toast at that.

Columbus Dispatch and the Fish Kill

Besides the two local newspapers, and Wheeling Jesuit researchers, the Columbus Dispatch also covered the story and provided more details on the 3- to 5-mile long fish kill in the stream below the well pad. Additional facts were added by the two EPA reports:

Those reports list in some detail many of the chemicals, explosives, and radiological components on the well pad.  Reader note: Get out your chemical dictionary, or fire up your Google search. A few excerpts from the first EPA report are provided below.

…Materials present on the Pad included but was not limited to: diesel fuel, hydraulic oil, motor oil, hydrochloric acid, cesium-137 sources, hydrotreated light petroleum distillates, terpenes, terpenoids, isoproponal, ethylene glycol, paraffinic solvents, sodium persulfate, tributyl tetradecyl phosphonium chloride and proprietary components… The fire and explosion that occurred on the Eisenbarth Well Pad involved more than 25,000 gallons of various products that were staged and/or in use on the site… uncontained run-off was exiting the site and entering an unnamed tributary of Opossum Creek to the south and west and flowback water from the Eisenbarth Well #7 was spilling onto the well pad.

Reader Warning:  If you found the above list overly alarming, you might choose to skip the next equally disturbing list. Especially since you now know that this all eventually flowed into our Ohio River.

The EPA report continues with more specific chemical products involved in the fire:

Initial reports identified the following products were involved and lost in the fire: ~250 gallons of hydrochloric acid (28%), ~7,040 gallons of GasPerm 1000 (terpenes, terpenoids, isopropanol, citrus extract, proprietary components), ~330 gallons of LCA-1 (paraffinic solvents), ~ 1900 gallons of LGC-36 UC (hydrotreated light petroleum distillate, guar gum), ~1000 gallons of BC-140 (monoethanolamine borate, ethylene glycol), ~3300 gallons of BE-9 (tributyl tetradecyl phosphonium chloride), ~30,000 gallons of WG-36 (polysaccharide gel), ~1,000 gallons of FR-66 (hydrotreated light petroleum distillate), ~9000 gallons of diesel fuel, ~300 gallons of motor and hydraulic oil.

Even more details of the incident and the on-site chemicals are given in the required Statoil 30-day report (PDF).

The EPA reports detail the “sheet” flow of unrestricted contaminated liquids off of the well pad during and after the fire. They refer to the west and south sides. The below Google Earth-based map (Fig. 8) shows the approximate flow from the well pad. The two unnamed tributaries join to form Opossum Creek, which then flows into the Ohio River four miles away.

Statoil 8

Figure 8. Map showing path of unrestricted flow off of the Statoil well pad due to a lack of berm

After describing some of the known chemicals on the well pad, the EPA report discusses the construction of a new berm, and where the liquid components flowed. Below is a selection of many excerpts strung together, from many days, taken directly from the EPA reports:

…unknown quantities of products on the well pad left the Site and entered an unnamed tributary of Opossum Creek that ultimately discharges to the Ohio River. Runoff left the pad at various locations via sheet flow….Initial inspections in the early hours of June 29, 2014 of Opossum Creek approximately 3.5 miles downstream of the site identified dead fish in the creek…. Equipment was mobilized to begin constructing an earthen berm to contain runoff and to flood the pad to extinguish remaining fires…. Once fires were extinguished, construction of a berm near the pad was begun to contain spilled liquids and future runoff from the well pad… Statoil continued construction of the containment berm currently 80% complete. (6-30-14)… Assessment of chemicals remaining on the well pad was completed. The earthen berm around the pad was completed,  (7-2-14)… ODNR Division of Wildlife completed their in stream assessment of the fish kill and reported an estimated 70,000 dead fish from an approximately 5 mile stretch extending from the unnamed tributary just west of the Eisenbarth Well Pad to Opossum Creek just before its confluence with the Ohio River… Fish collection was completed. In total, 11,116 dead fish were collected (20 different species), 3,519 crustaceans, 7 frogs and 20 salamanders.

The overall conclusion is clear. Large quantities of various chemicals, mixed with very large amounts of already contaminated water, when flooding a well pad that had no berms around it, resulted in a significant fish kill over several miles. After the fire Statoil then constructed a berm around the well pad. If there had been a pre-existing berm – just 12 inches high and level – around the well pad, it could have held over 600,000 gallons of runoff. That amount is twice the estimated quantity of water used to fight the fire.  (Note: my old 35 HP farm tractor and a single bottom plow can provide a 12-inch high mound of dirt in one pass.)

The significance for safe, potable drinking water, is that all the chemicals and petroleum products on the well pad either burned and went up in a toxic plume of black smoke, or were released in liquid form down into the well pad or flowed off of it. Since the original liner on the well pad also completely burned and there was no overall berm on the well pad, there was nothing to restrict the flow of polluted liquid. Therefore, it all seeped into the ground and/or ran off of the pad with the 300,000 gallons of water that was estimated to have been sprayed onto the burning equipment fire.

Follow Up Questions

Since this fire happened over 6 weeks ago, there have been many opportunities for nearby citizens and neighbors to meet and discuss their many concerns.  Many of the question have revolved around the overall lack of information about the process of shale gas fracturing, the equipment used, and the degree of risk that it all may present to our communities. These communities include the nearby residents, the travelling public, and all of the first responders. Unless someone has a well pad on or near their property and they are able to actively follow the process, it is usually difficult to find out the details of a specific gas operation. (We have even known of operators that have told landowners to get off of their own property both during drilling and fracturing operations and afterwards.)

Questions that follow incidents like this one typically look like this:

  1. Why was there no perimeter berm?
  2. Why could the fire not be put out quickly and easily? What all was lost? What did this site look like in the beginning?
  3. Why was there so much equipment onsite? Is this typical? What is it all called and how is it used?

1. Lack of Berm

The first and somewhat unanswered question concerns the absence of a simple containment berm around the completed well pad. Statoil must not have thought one would be very helpful, and/or the State of Ohio must not require them.

However, I had raised concern over this very topic more than a year ago from WV. In response, I received a letter in September 2013 from Statoil North America to the WVDEP. It provides some insight into Statoil thinking. Based on my interpretation of that letter, the official position of Statoil last year was that berms around the well pad do not help and are not needed. Given the recent fire, perhaps that position has changed. All we know for sure now is that at least their Eisenbarth well pad now does have a complete perimeter berm. We now have empirical proof, if any was ever needed, that in the presence of spills the absence of berms makes for greater and more expensive downstream problems.

2. An Obstinate Fire

Setting aside the berm problem, I will attempt to address the next set of questions: Why could the fire not be put out quickly and easily? What all was lost ? What did this site look like in the beginning?

The simplest way to start on such questions is to look at other hydraulic fracturing sites to identify what is there and why, and then to compare those with the charred remains on the Statoil Eisenbarth well pad in Monroe County.  Since Statoil’s contractor was Halliburton, it would help to look at their equipment when in process elsewhere.  In Figure 9 below is a clean, bright red and grey Halliburton fracking fleet.

Statoil 9

Figure 9. Example of Halliburton fracking fleet

It needs to be stated up front that I consider Halliburton to be among one of the more reputable, experienced, and dependable fracturing companies. We have seen way worse here in Wetzel County over the past seven years. Halliburton has good equipment and well-trained, safety-conscious employees. It seems to be a well-run operation. If so, then how did this massive fire happen? It simply seems that it is the nature of the beast; there are many inherent dangers to such operations. Plus there is an enormous amount of equipment on site, close coupled and stuffed into a small amount of real estate. Not to mention, the whole setup is temporary – with a lot of fuel and ignition sources. Therefore, many of the available engineered-in safeguards that would normally be installed in an industrial, fixed, permanent location, just cannot be incorporated on my neighbor’s hay field, creek bottom, or farmland.

The whole process has many risks, and many of them cannot be eliminated, just minimized. I do not think that anyone could have predicted a weak hydraulic hose. Some accidents are just that — unpreventable accidents. This is why we need to be very careful with how close we allow these sites in residential areas.

3. Serious Equipment

In Figure 10 below is a wide-angle composite photo of a Halliburton fracturing project in process. Given the shallow angle viewpoint, not all equipment is visible or numbered. The photo is still very representative of frac sites in general and equivalent to what can be seen in the scorched remains on the Statoil Eisenbarth site. The major qualification on the fracturing pumps above and the ones below, is that they are a newer generation of Halliburton dual fuel pumps. They can run on natural gas.

Statoil 10

Figure 10. Halliburton fracturing project in process

Just about everything seen in the above bright red and grey hardware can be seen in Figure 11’s charred leftovers on the Statoil site from July 5, 2014 below (six days after the fire). It is also all Halliburton equipment. The quantities and arrangement are different, but the equipment and process are the same. The numbers on the provided legend or chart should help identify the specific pieces of equipment. The newly constructed containment berm is also clearly visible here.

Statoil 11

Figure 11. Statoil site post-fire equipment identification

The above or a similar photo has been seen by many neighbors both in OH and WV. Hardly anyone can recognize what they are looking at. Even those people who are somewhat familiar with general hydraulic fracturing operations are puzzled. Nothing is obvious when viewing charred remains of burned iron, steel, and melted aluminum. All tires (over 400 of them) have been burned off the rims. Every bit of rubber, foam, composites, plastics and fiberglass truck cabs has been consumed – which is what made the black plume of smoke potentially so dangerous.

Statoil 12

Fig. 12. 16 fracturing pumps

Statoil 13

Fig. 13. 18-wheeler

What might not be so obvious is why the fire could not be extinguished.

If we look at a close-up of a small section of the well pad (Fig. 12) it is easy to see how crowded the well pad is during fracturing. The 16 fracturing pumps are all the size of a full-length 18-wheel tractor trailer (Fig. 13). Note the three fuel tanks.

The fire began between the blender-mixer trucks and the 16 hydraulic fracturing pumps. The blenders were between the fracturing pumps and the sand kings. Halliburton always keeps fire extinguishers available at every truck. They are put on the ground in front of every pump truck. Everyone knows where to find them. However, on any fracking project that location is also the most congested area. The fracturing pumps are usually parked no more than two feet apart. It is just enough room for an operator or maintenance fellow to get between them. With high pressure fluid spraying and the fire already started and now spreading, there is precious little room to maneuver or to work. It is a plumbing nightmare with the dozens of high pressure pipes connecting all the pumps together and then to a manifold. In those conditions, in the face of multiple fuel sources, then the many small explosions, prudence and self-preservation dictates a swift retreat.

To their credit, Halliburton employees knew when to retreat. No one was injured. We just burned up some trucks (and killed some fish). All the employees and all the first responders were able to go home safely, uninjured, to their families and friends. They survived a very dangerous situation to come back again in the service of their employer or their community. We wish them well.

Some Observations and Conclusions

  1. The hydraulic fracturing process is dangerous, even when done properly.
  2. Environmental and employee safeguards must be in place because “accidents will happen.”
  3. Setbacks from personal farm and residential buildings must be great enough to protect all.
  4. Setbacks from streams and creeks and rivers must be taken very seriously, especially when private or municipal water supply systems are downstream.
  5. Our communities must know what all chemicals are being used so that correct lab protocols are established ahead of time to test for contamination.

This now ends this first article addressing the Statoil Fire, its burned fracturing equipment, and the resulting water contamination. Later, I will show many examples of the quantity of equipment used on fracturing sites and why it is there. You patient readers thought this would never end. You now know more about Statoil, well pad fires, and fracturing hardware than you ever wanted to know. We will soon address the more generic questions of fracturing equipment.

Oil Transportation and Accidents by Rail

Lac-Mégantic train explosion on July 6, 2013.  Photo by TSB of Canada.

Lac-Mégantic train explosion on July 6, 2013. Photo by Transportation Safety Board of Canada.

On July 5, 2013, the lone engineer of a Montreal, Maine, and Atlantic (MMA) train arrived in Nantes, Quebec, set both the hand and air brakes, finished up his paperwork. He then left the train parked on the main line for the night, unattended atop a long grade. Five locomotives were pulling 72 tanker cars of oil, each containing 30,000 gallons of volatile crude from North Dakota’s Bakken Formation. During the night, the lead locomotive caught fire, so the emergency responders cut off the engine, as per protocol.  However, that action led to a loss of pressure of the air brakes.  The hand brakes (which were supposed to have been sufficient by themselves) failed, and the train began to run away. By the time it reached Lac-Mégantic early the next morning, the unattended cars were traveling 65 mph.  When the train reached the center of town, 63 tank cars derailed and many of those exploded, tragically killing 47 people in a blaze that took over two days to extinguish.

With that event came a heightened awareness of the risks of transporting volatile petroleum products by rail.  A derailment happened on a BNSF line near Casselton, North Dakota on December 30, 2013. This train was then struck by a train on an adjacent track, igniting another huge fireball, although this one was luckily just outside of town.  On April 30, 2014, a CSX train derailed in Lynchburg, Virginia, setting the James River on fire, narrowly avoiding the dense downtown area of the city of 75,000 people.


North American petroleum transportation by rail. Click on the expanding arrows icon in the top-right corner to access the full screen map with additional tools and description.

Regulators in the US and Canada are scrambling to keep up.  DOT-111 tank cars were involved in all of these incidents, and regulators seek to phase them out over the next two years. These cars account for 69% of the fleet of tank cars in the US, however, and up to 80% in Canada.  Replacing these cars will be a tough task in the midst of the oil booms in the Bakken and Eagle Ford plays, which have seen crude by rail shipments increase from less than 5,000 cars in 2006 to over 400,000 cars in 2013.

This article is the first of several reports by the FracTracker Alliance highlighting safety and environmental concerns about shipping petroleum and related products by rail. The impacts of the oil and gas extraction industry do not end at the wellhead, but are a part of a larger system of refineries, power plants, and terminals that span the continent.

DEP, downloaded 7/9/2014.">

What’s in PA Senate Bill 1378?

State Senator Joseph Scarnati III, from north-central Pennsylvania, has introduced a bill that would redefine the distinction between conventional and unconventional oil and gas wells throughout the state.  In Section 1 of the bill, the sponsors try to establish the purpose of the legislation,  making the case that:

  1. Conventional oil and gas development has a benign impact on the Commonwealth
  2. Many of the wells currently classified as conventional are developed by small businesses
  3. Oil and gas regulations, “must permit the optimal development of oil and gas resources,” as well as protect the citizens and environment.
  4. Previous legislation already does, and should, treat conventional and unconventional wells differently
This diagram shows geologic stata in Pennsylvania.  The Elk Sandstone is between the Huron and Rhinestreet shale deposits from the Upper Devonian period.

This diagram shows geologic stata in Pennsylvania. The Elk Group is between the Huron and Rhinestreet shale deposits from the Upper Devonian period. Click on the image to see the full version. Source: DCNR

Certainly, robust debate surrounds each of these points, but they are introductory in nature, not the meat and potatoes of Senate Bill 1378.  What this bill does is re-categorize some of the state’s unconventional wells to the less restrictive conventional category, including:

  1. All oil wells
  2. All natural gas wells not drilled in shale formations
  3. All shale wells above (shallower than) the base of the Elk Group or equivalent
  4. All shale wells below the Elk Group from a formation that can be economically drilled without the use of hydraulic fracturing or multi-lateral bore holes
  5. All wells drilled into any formation where the purpose is not production, including waste disposal and other injection wells

The current distinction is in fact muddled, with one DEP source indicating that the difference is entirely due to whether or not the formation being drilled into is above or below the Elk Group, and another DEP source indicates that the difference is much more nuanced, and really depends on whether the volumes of hydraulic fracturing fluid required to profitably drill into a given formation are generally high or low.

This table shows the number of wells in each formation in Pennsylvania that has both conventional and unconventional wells drilled into it.  Data source:  DEP, downloaded 7/9/2014.

This table shows the number of distinct wells in each producing formation in Pennsylvania that has both conventional and unconventional wells drilled into it. Data source: DEP, downloaded 7/9/2014.

As one might expect, this ambiguity is represented in the data. The chart at the left shows the number of distinct number of wells by formation, for each producing formation that has both conventional and unconventional wells in the dataset.  Certainly, there could be some data entry errors involved, as the vast majority of Bradford wells are conventional, and almost all of the Marcellus wells are unconventional.  But there seems to be some real confusion with regards to the Oriskany, for example, which is not only deeper than the Elk Group, but the Marcellus formation as well.

While an adjustment to the distinction of conventional and unconventional wells in Pennsylvania is called for, one wonders if the definitions proposed in SB 1378 is the right way to handle it.  If the idea of separating the two is based on the relative impact of the drilling operation, then a much more straightforward metric might be useful, such as providing a cutoff in the amount of hydraulic fracturing fluid used to drill a well.  Further, each of the five parts of the proposed definition serve to make the definition of unconventional wells less inclusive, meaning that additional wells would be subject to the less stringent regulations, and that the state would collect less money from the impact fees that were a part of Act 13 of 2012.

Instead, it is worth checking to see whether the definition of unconventional is inclusive enough.  In May of this year, FracTracker posted a blog about conventional wells that were drilled horizontally in Pennsylvania.


Conventional, non-vertical wells in Pennsylvania. Please click the expanding arrows icon at the top-right corner to access the legend and other map controls. Please zoom in to access data for each location.

These wells require large amounts of hydraulic fracturing fluids, and are already being drilled at depths of only 3,000 feet, and could go as shallow as 1,000 feet.  It’s pretty easy to argue that due to the shallow nature of the wells, and the close proximity to drinking water aquifers, these wells are deserving of even more rigorous scrutiny than those drilled into the Marcellus Shale, which generally ranges from 5,000 to 9,000 feet deep throughout the state.

A summary of the different regulations regarding conventional and unconventional wells can be found from PennFuture.  In general, unconventional wells must be further away from water sources and structures than their conventional counterparts, and the radius of presumptive liability for the contamination of water supplies is 2,500 feet instead of 1,000.

SB 1378 has been re-referred to the Appropriations Committee.