Posts

Well Worker Safety and Statistics

By Samantha Malone, MPH, CPH – Manager of Science and Communications, FracTracker Alliance

The population most at risk from accidents and incidents near unconventional drilling operations are the drillers and contractors within the industry. While that statement may seem quite obvious, let’s explore some of the numbers behind how often these workers are in harm’s way and why.

O&G Risks

Oil and Gas Worker Fatalities over Time

Fig. 1. Number of oil and gas worker fatalities over time
Data Source: U.S. Bureau of Labor Statistics, U.S. Department of Labor, 2014

Drilling operations, whether conventional or unconventional (aka fracking), run 24 hours a day, 7 days a week. Workers may be on site for several hours or even days at a time. Simply the amount of time spent on the job inherently increases one’s chances of health and safety concerns. Working in the extraction field is traditionally risky business. In 2012, mining, quarrying, and oil and gas extraction jobs experienced an overall 15.9 deaths for every 100,000 workers, the second highest rate among American businesses. (Only Agriculture, forestry, fishing and hunting jobs had a higher rate.)

According to the Quarterly Census of Employment and Wages of the U.S. Bureau of Labor Statistics, the oil and gas industry employed 188,003 workers in 2012 in the U.S., a jump from 120,328 in 2003. Preliminary data indicate that the upward employment trend continued in 2013. However, between 2003 and 2012, a total of 1,077 oil and gas extraction workers were killed on the job (Fig. 1).

Causes of Injuries and Fatalities in Oil and Gas Field

Reasons for O&G Fatalities 2003-12. Aggregated from Table 1.

Fig. 2. Reasons for O&G Fatalities 2003-12. Aggregated from Table 1.

Like many industrial operations, here are some of the reasons why oil and gas workers may be hurt or killed according to OSHA:

  • Vehicle Accidents
  • Struck-By/ Caught-In/ Caught-Between Equipment
  • Explosions and Fires
  • Falls
  • Confined Spaces
  • Chemical Exposures

If you drill down to the raw fatality-cause numbers, you can see that the fatal worksite hazards vary over time and job type1 (Table 1, bottom). Supporting jobs to the O&G sector are at higher risk of fatal injuries than those within the O&G extraction job category2. The chart to the right shows aggregate data for years 2003-12. Records indicate that the primary risk of death originated from transportation incidents, followed by situations where someone came into contact with physical equipment (Fig. 2).

Silica Research

Silica-Exposed Workers

Fig. 3. Number of total silica-exposed workers and those exposed above PEL – compared across industries
Source: OSHA Directorate of Standards and Guidance

A recent NIOSH study by Esswein et al. regarding workplace safety for oil and gas workers was that the methods being employed to protect workers against respirable crystalline silica3 were not adequate. This form of silica can be found in the sand used for hydraulic fracturing operations and presents health concerns such as silicosis if inhaled over time. According to Esswein’s research, workers were being exposed to levels above the permissible exposure limit (PEL) of ~0.1 mg/m3 for pure quartz silica because of insufficient respirator use and inadequate technology controls on site. It is unclear at this time how far the dust may migrate from the well pad or sand mining site, a concern for nearby residents of the sand mines, distribution methods, and well pads. (Check out our photos of a recent frac sand mine tour.) The oil and gas industry is not the only employer that must protect people from this airborne workplace hazard. Several other classes of jobs result in exposure to silica dust above the PEL (Fig. 3).

References and Additional Resources

1. What do the job categories in the table below mean?

For the Bureau of Labor Statistics, it is important for jobs to be classified into groups to allow for better reporting/tracking. The jobs and associated numbers are assigned according to the North American Industry Classification System (NAICS).

(NAICS 21111) Oil and Gas Extraction comprises establishments primarily engaged in operating and/or developing oil and gas field properties and establishments primarily engaged in recovering liquid hydrocarbons from oil and gas field gases. Such activities may include exploration for crude petroleum and natural gas; drilling, completing, and equipping wells; operation of separators, emulsion breakers, desilting equipment, and field gathering lines for crude petroleum and natural gas; and all other activities in the preparation of oil and gas up to the point of shipment from the producing property. This industry includes the production of crude petroleum, the mining and extraction of oil from oil shale and oil sands, the production of natural gas, sulfur recovery from natural gas, and the recovery of hydrocarbon liquids from oil and gas field gases. Establishments in this industry operate oil and gas wells on their own account or for others on a contract or fee basis. Learn more

(NAICS 213111) Drilling Oil and Gas Wells comprises establishments primarily engaged in drilling oil and gas wells for others on a contract or fee basis. This industry includes contractors that specialize in spudding in, drilling in, redrilling, and directional drilling. Learn more

(NAICS 213112) Support Activities for Oil and Gas Operations comprises establishments primarily engaged in performing support activities on a contract or fee basis for oil and gas operations (except site preparation and related construction activities). Services included are exploration (except geophysical surveying and mapping); excavating slush pits and cellars, well surveying; running, cutting, and pulling casings, tubes, and rods; cementing wells, shooting wells; perforating well casings; acidizing and chemically treating wells; and cleaning out, bailing, and swabbing wells. Learn more

2. Fifteen percent of all fatal work injuries in 2012 involved contractors. Source

3. What is respirable crystalline silica?

Respirable crystalline silica – very small particles at least 100 times smaller than ordinary sand you might encounter on beaches and playgrounds – is created during work operations involving stone, rock, concrete, brick, block, mortar, and industrial sand. Exposures to respirable crystalline silica can occur when cutting, sawing, grinding, drilling, and crushing these materials. These exposures are common in brick, concrete, and pottery manufacturing operations, as well as during operations using industrial sand products, such as in foundries, sand blasting, and hydraulic fracturing (fracking) operations in the oil and gas industry.

4. OSHA Fact Sheet: OSHA’s Proposed Crystalline Silica Rule: General Industry and Maritime. Learn more

Employee health and safety are protected under the following OSHA regulations. These standards require employers to make sure that the workplace is in due order:

Table 1. 2003-2012 U.S. fatalities in oil & gas industries by year, job category, & event/exposure
Year Oil and Gas (O&G) Industriesa Total Fatal Injuries (number)b Event or Exposurec
Violence / injuries by persons / animalsd Transportatione Fires & Explosions Falls, Slips, Trips Exposure to Harmful Substances or Environments Contact w/Objects & Equipment
2012
O&G Extraction 26 0 8 6 5 3 4
Drilling O&G Wells 39 0 10 6 8 3 10
Support Activities 77 0 46 11 5 3 10
Yearly Totals 142 0 64 23 18 9 24
2011
O&G Extraction 13 0 7 0 0 0 3
Drilling O&G Wells 41 0 15 5 4 5 12
Support Activities 58 3 29 7 4 4 11
Yearly Totals 112 3 51 12 8 9 26
2010
O&G Extraction 12 0 5 3 0 3 0
Drilling O&G Wells 47 0 8 14 7 6 12
Support Activities 48 3 28 8 0 0 8
Yearly Totals 107 3 41 25 7 9 20
2009
O&G Extraction 12 0 6 0 0 0 3
Drilling O&G Wells 29 0 9 0 0 4 13
Support Activities 27 0 12 5 0 4 5
Yearly Totals 68 0 27 5 0 8 21
2008
O&G Extraction 21 0 7 4 0 0 5
Drilling O&G Wells 30 0 6 3 4 4 13
Support Activities 69 0 36 11 4 6 12
Yearly Totals 120 0 49 18 8 10 30
2007
O&G Extraction 15 0 5 0 0 0 5
Drilling O&G Wells 42 0 12 0 4 8 16
Support Activities 65 0 33 6 0 5 19
Yearly Totals 122 0 50 6 4 13 40
2006
O&G Extraction 22 0 6 7 0 3 4
Drilling O&G Wells 36 0 11 0 5 4 14
Support Activities 67 0 2 12 0 5 21
Yearly Totals 125 0 19 19 5 12 39
2005
O&G Extraction 17 0 4 5 0 0 4
Drilling O&G Wells 34 0 9 0 7 4 10
Support Activities 47 0 21 5 0 5 13
Yearly Totals 98 0 34 10 7 9 27
2004
O&G Extraction 29 0 17 0 0 0 8
Drilling O&G Wells 30 0 6 0 6 3 11
Support Activities 39 0 22 5 0 0 10
Yearly Totals 98 0 45 5 6 3 29
2003
O&G Extraction 17 0 9 4 0 0 3
Drilling O&G Wells 26 0 5 5 0 0 13
Support Activities 42 0 17 10 0 3 10
Yearly Totals 85 0 31 19 0 3 26
2003-12 TOTAL FATALITIES 1077 6 411 142 63 85 282
a Oil and gas extraction industries include oil and gas extraction (NAICS 21111), drilling oil and gas wells (NAICS 213111), and support activities for oil and gas operations (NAICS 213112).
b Data in event or exposure categories do not always add up to total fatalities due to data gaps.
c Based on the BLS Occupational Injury and Illness Classification System (OIICS) 2.01 implemented for 2011 data forward
d Includes violence by persons, self-inflicted injury, and attacks by animals
e Includes highway, non-highway, air, water, rail fatal occupational injuries, and fatal occupational injuries resulting from being struck by a vehicle.
US Pipeline Incidents map

Pipeline Incidents Updated and Analyzed

Pipeline spill in Mayflower, AR on March 29, 2013. Photo by US EPA via Wikipedia.

The debate over the Keystone XL pipeline expansion project has grabbed a lot of headlines, but it is just one of several proposed major pipeline projects in the United States. As much of the discussion revolves around potential impacts of the pipeline system, a review of known incidents is relevant to the discussion.

A year ago, the FracTracker Alliance calculated that there was an average of 1.6 pipeline incidents per day in the United Sates.  That figure remains accurate, with 2,452 recorded incidents between January 1, 2010 and March 3, 2014, a span of 1,522 days.

The Pipeline and Hazardous Materials Safety Administration (PHMSA) classifies the incidents into three categories:

  • Gas transmission and gathering:  Gathering lines take natural gas from the wells to midstream infrastructure.  Transmission lines transport natural gas from the regions in which it is produced to other locations, often thousands of miles away.  Since 2010, there have been 486 incidents on these types of lines, resulting in 10 fatalities, 71 injuries, and $620 million in property damage.
  • Oil and hazardous liquid:  This includes all materials overseen by PHMSA other than natural gas, predominantly crude and refined petroleum products.  Liquified natural gas is included in this category.  There were 1,511 incidents during the reporting period for these pipelines, causing 6 deaths and 15 injuries, and $1.8 billion in property damage.
  • Gas distribution:  These pipelines are used by utilities to get natural gas to consumers.  In just over 40 months, there were 455 incidents, resulting in 42 people getting killed, 183 reported injuries, and $86 million in property damage.

Curiously, while incidents on distribution lines accounted for 72 percent of fatalities and 67 percent of all injuries, the property damage in these cases were only responsible for just over 3 percent of $2.5 billion in total property damage from pipeline spills since 2010.  A reasonable hypothesis accounting for the deaths and injuries is that distribution lines are much more common in densely populated areas than are the other types of pipelines; an incident that might be fatal in an urban area might go unnoticed for days in more remote locations, for example.  However, as the built environment is also much more densely located in urban areas, it does seem surprising that reported property damage isn’t closer to being in line with physical impacts on humans.

How accurate are the data?

In the wake of the events of September 11, 2001, governmental agency data suddenly became much more opaque.  In terms of pipelines, public access to the pipeline data that had been mapped to that point was removed.  It was later restored, with limitations.  As it stands now, most pipeline data in the United States, including the link to the pipeline proposal map above, are intentionally generalized to the point where pipelines might not even be rendered in the appropriate township, let alone street.

There are some exceptions, though.  If you would like to know where pipelines are in US waters in the Gulf of Mexcio, for example, the Bureau of Ocean Energy Management makes that data not only accessible to view, but available for download on data.gov, a site dedicated to data transparency.  While the PHMSA will not do the same with terrestrial pipelines, the do release location data along with their incident data.


Pipeline incidents from 1/1/2010 through 3/3/2014. To access details, legend, and other map controls, please click the expanding arrows icon in the top-right corner of the map.

This fatal pipeline incident was in Allentown, PA, but was given coordinates in Greenland.

This fatal pipeline incident was in Allentown, PA, but was given coordinates in Greenland.

Unfortunately, we see evidence that the data are not well vetted, at least in terms of location.  One of the most serious incidents in the timeframe, an explosion in Allentown, Pennsylvania that killed five people and injured three more, was given coordinates that render in the middle of Greenland.  Another incident leading to fatalities was given location data that put it in Manatoba, well outside of the reach of the US agency that publishes the data.  Still another incident appears to be in the Pacific Ocean, 1,300 miles west-southwest of Mexico.  There are many more examples as well, but the majority of incidents seem to be reasonably well located.

Fuzzy data: are national security concerns justified?

Anyone who watches the news on a regular basis knows that there are people out there who mean others harm. However, a closer look at the incident data shows that pipelines are not a common means of accomplishing such an end.

Causes of pipeline incidents from 1/1/10 to 3/3/14, with counts.

Causes of pipeline incidents from 1/1/10 to 3/3/14, with counts.

For each category showing causation, there are numerous subcategories. While we don’t need to look into all of those here, it is worth pointing out that there is a subcategory of, “other outside force damage” that is designated as, “intentional damage.”  Of the 2,452 total incidents, nine incidents fall into this subcategory.  These subcategories are further broken down, and while there is an option to express that the incident is a result of terrorism, none have been designated that way in this dataset .  Five of the nine incidents are listed as acts of vandalism, however. To be thorough, and because it provides a fascinating insight into work in the field, let’s take a look at the narrative description for each incident that are labeled as intentional in origin:

  • Approximately 2 bbls of crude oil were released when an unknown person(s) removed the threaded pressure warning device on the scraper trap’s closure door. As a result of the absence of the 1/2 inch pressure warning device crude oil was able to flow from the open port upon start up of the pipeline and pressurization of the scraper trap. Once this was discovered the 1/2 inch pressure warning device was properly put back into the scaper trap.
  • Aboveground piping intentionally shot by unknown party. Installed stoppall on line at 176+73 (7 146′) upstream of damaged aboveground piping. Cut and capped pipeline.
  • Friday october 18th at approximately 6:00 p.m. we were notified of a gas line break at Kayenta Mobile Home Park. The Navajo Police responded to an emergency call about vandals in one of the parks alley ways kicking at meters. Upon arrival they found the broke meter riser at the mobile home park and expediently used the emergency shutdown system to remedy the situation. This immediately cut service to 118 customers in the park. [Names removed] responded to the call. we arrived on site at approximately 9:30 p.m. We located the damage and fixed the system at approximately 1:30 a.m. i called the Amerigas emergency call center and informed them that we would be restarting the system the following morning and to tell our customers they would need to be home in order to restore service. We then started the procedure of shutting every valve off to all customers before restarting the system. We started the system back up at 9:30a.m. 10/19/2013. Once the system was up to full pressure and all systems were normal we began putting customers back into service. The completion of re-establishing service to all customers on the system was completed on 10/23/2013.
  • A service tech was called at 1:15 am Sunday morning to respond to the Marlboro Fire Department at an apparent explosion and house fire. The tech arrived and called for additional resources. He then began to check for migrating gas in the surrounding buildings along the service to the house and in the street. no gas readings were detected. The distribution and service on call personnel arrived and began calling in additional company resources to assist in the response effort and controlling the incident. A distribution crew was called in to shut off and cut the service. Additional service techs were called in to assist in checking the surrounding buildings and in the streets at catch basins and manholes around the entire block. Gas supply personnel were called in and dispatched to take odorant samples in the houses directly across from 15 Grant Ct. that had active gas service. Gas survey crews were called in to survey Grant St. and the two parallel streets McEnelly St. and Washington Ct. along with the portion of Washington st. in between these streets. The meter and meter bar assembly were taken by the investigators as evidence. The service was pressure tested to the riser which was witnessed by a representative of the DPI. The service was cut off at the main. After the investigators completed gathering evidence at the scene they gave permission to begin cleaning up the site. There was a tenant home at the time of the explosion who was conscious and walking around when the fire department arrived. He was taken to the hospital and reports are that he sustained 2nd and 3rd degree burns on portions of his body.
  • On Friday, September 7, 2012 PSE&G responded to a gas emergency call involving a gas ignition. The initial call came in from the Orange Fire Department at 17:09 as a house fire at 272 Reock Ave Orange; the fire chief stated gas was not involved and the fire was caused by squatters. Subsequent investigation of the incident revealed that the fire was caused when one of the squatters lit a match which ignited leaking gas originating from gas piping removed from the head of an inside meter set. The gas meter inlet valve and associated piping were all removed by an unknown person on an unknown date prior to the fire. An appliance service tech responded and shut the gas off at the curb at 17:40 on September 7 2012. A street crew was dispatched and the gas service to 272 reock ave was cut at the curb at 19:00. Two people (names unknown) squatters were injured one by the fire one was injured jumping out a window to escape the fire. The home in question was vacated by the owner and the injured parties were trespassing on the property at the time of the incident. PSE&G has been unable to confirm any information on the status of their injuries due to patient confidentiality laws.
  • The homeowner tampered with company piping by removing 3/4″ steel end cap with a 3/4″ steel nipple on the tee was removed which caused the gas leak in the basement and resulted in a flash fire. The most likely source of ignition was the water heater. The homeowner died in the incident.
  • A structure fire involved an unoccupied hardware store and a small commercial 12-meter manifold. There were no meters on the manifold and no customers lost service. The heat from the structure fire melted a regulator on the manifold which in turn released gas and contributed to the fire. The cause is officially undetermined; however according to the fire department the cause appears to be arson with the fire starting in the back of the building and not from PG&E facilities. PG&E was notified of this incident by the fire department at 1802 hours. The gas service representative arrived on scene at 1830 hours. The fire department stopped the flow of gas by closing the service valve and the fire was extinguished at approximately 1900 hours. this incident was determined to be reportable due to damages to the building exceeding $50,000. There were no fatalities and no injuries as a result of this incident. Local news media was on-site but no major media was present.
  • A house explosion and fire occurred at approximately 0208 hours on 2/7/10. The fire department called at PG&E at 0213 hours. PG&E personnel arrived at 0245 hours. The fire department had shut off the service valve and removed the meter before PG&E arrived. The house was unoccupied at the time of the explosion. The gas service account was active and the gas service was on (contrary to initial report). The cause of the explosion is undetermined at the time of this report but the fire department has indicated the cause appears to be arson. After the explosion, PG&E performed a leak survey of the service the services on both sides of this address and the gas main in the front of all three of these addresses. No indication of gas was found. PG&E also performed bar hole tests over the service at 3944 17th Avenue and found no indication of gas. The gas service was cut off at the main and will be re-connected when the customer is ready for service.
  • On Monday, January 25, 2010 at approximately 2:30pm a single-family home at 2022 west 63rd Street Cleveland OH (Cuyahoga County) was involved in an explosion/fire. The gas service line was shut-off at approximately 4:30pm. A leak survey of the main lines and service lines on W. 83rd between Madison and Lorain revealed no indications of gas near the structure. A service leak at 2131 West 83rd Street was detected during the leak survey. This service line was replaced upon discovery. On Tuesday, January 26th, 2010 the service line at 2022 W. 83rd was air tested at operating pressure with no pressure loss. An odor test was conducted at 2028 West 83rd Street. The results of this odor test revealed odor levels well within dot compliance levels. Our investigation revealed an odor complaint at this residence on January 18th. Dominion personnel responded to the call and met with the Cleveland Fire Department. Dominion found the meter disconnected and the meter shut-off valve in the half open position. The shut-off valve was closed by the Dominion technician and secured with a locking device. The technician placed a 3/4 inch plug in the open end of the valve. The technician also attempted to close the curb-slop valve but could not. The service line was then bar hole tested utilizing a combustible gas indicator from the street to the structure. As a result, no leakage was discovered. A second attempt to close the curb box valve on January 19th ended when blockage was discovered in the valve box. The valve box was in the process of being scheduled for excevatlon and shut off by a construction crew at the time of the incident. An investigation of the incident site determined the cause to be arson as approximately 6 inches of service line and the meter shut-off valve (with locking device still intact) detached from the service line were recovered inside the structure.

While several of these narratives do make it seem as if the incidents in question were deliberate, these seem to have been caused by people on the ground, not by some GIS-powered remote effort. Seven of the nine incidents were on distribution lines, which tend to occur in populated areas, where contact with gas infrastructure is in fact commonplace, and six out of those seven incidents occurred inside houses or other structures.

On the other hand, there is a real danger in not knowing where pipelines are located. 237 accidents were due to excavation activities, and 86 others were caused by boats, cars, or other vehicles unrelated to excavation activity. Better knowledge of the location of these pipelines could reduce these numbers significantly.

US Pipelines Incidents Are a Daily Occurrence

Recently, there has been a lot of attention focused on the Mayflower, Arkansas pipeline failure that resulted in a massive oil spill, particularly as it comes at a time when discussions of the controversial Keystone XL Pipeline project are once again heating up.  However, the situation is far from unusual.  In fact, according to data downloaded from the Pipeline and Hazardous Materials Safety Administration (PHMSA), there were 1,887 incidents in the nation’s gathering and transmission, distribution, and hazardous liquids pipelines between January 1, 2010 and March 29, 2013, or an average of 1.6 incidents per day.

Pipeline incidents from 1/1/2010 through 3/29/2013.

Pipeline incidents from 1/1/2010 through 3/29/2013. Data Source: PHMSA.

Obviously, not all of these failures are on par with the massive spill in Mayflower, and it should be noted that there are a variety of reasons for these lines to fail.  Some of these reasons, such as excavation activity in the vicinity of a pipeline, are not necessarily the fault of the pipeline’s operator.  The fact that these incidents are commonplace, however, is not one that can be dismissed.


Pipeline incidents in the United States from 1/1/2010 through 3/29/2013. Source: PHMSA.  Red Triangles represent incidents leading to fatalities, and yellow triangles represent those leading to injuries.  To access the legend and other controls, click the “Fullscreen” icon at the top-right corner of the map.

It is clear from the map that there a few data entry errors, as a few of the data points draw in locations that aren’t even in the jurisdiction of the United States. However, each entry also contains a city and state that the incident is associated with, and for the most part, the data seem to be fairly reliable.