The SkyTruther path: an intern’s excursion to understanding offshore oil (part I)

Practice does not guarantee perfection but it is a diligent educator. During my first week of orientation at SkyTruth, the other interns and I filtered through Sentinel 1 radar satellite imagery on the European Space Agency’s (ESA) online portal, compressed large .tiff image files using the command-line, and constructed final, publishable visuals in QGIS – a free, open source geographic information system (GIS), enabling users to create, view, edit, and analyze geospatial data. Essentially, we learned the fundamentals of finding, processing, and analyzing imagery.

Subsequent to grasping the basics, I worked on catching my first ocean offender by clicking through images, zooming in and out of rasters, and adjusting min and max values. Eventually, I spotted several shadowy slithers. Possible identification: oil leak.

Oil slicks off the coast of China.

The first image I discovered at SkyTruth.

Before hopscotching to conclusions, I checked in with mentor, Dr. Ry Covington (Doctorate of Philosophy; he’s not the medical type but he knows a thing about bodies – bodies of water, that is). Without hesitation, Ry confirmed my sighting as plausible. Three likely slicks from three unidentified sources. Mission success, phase two initiated: annotate.

When annotating an image, there are certain guidelines to follow. Most of the metadata – basic information needed to read a visual such as image credit/source, author, scale-bar, and date – is there. However, I did not include any boxed nouns or pointy things denoting the white zit-like points, or running, dark mascara streaks. Reflecting, I should have marked up this version more; I should have labeled the several pimples as unidentified sources, and measured the length of the eyelash-lacquer lines, ticketing them as slicks. Instead, I let the caption clamor over the image.

My first caption went something like this: “This image displays three leaks from several unidentified sources, off the coast of Guangzhou Province, China (near Hong Kong).”

That was all I got. New to the practice in general and unfamiliar with that latitude, I didn’t have much to say. So to boost productivity, I harvested a separate, bluer pair of eyes. My advisor with three first names, Christian James Thomas, looked over my caption. He was particularly picky with diction. One word, to be precise: leaks. Backspacing five times, Christian typed ‘slick’. Slick? Like ‘smooth’ or ‘glossy’? Or maybe like Eric Slick, the drummer of my favorite band, Dr. Dog? I wish, but certainly not.

‘Slick’ has various definitions, but to the SkyTruth team, slick typically describes flat water. Smooth surfaces on satellite radar imagery could signify oil, algae, lack of wind, or the like. What we are interested in is the accidental or purposeful release of oil or oily waste that may be a result of drilling, disposal, or disaster. Leak or spill is too specific, too assuming. I learned why this was after confusing slicks with a number of other ocean junk. When examining satellite evidence, slicks are often muddled by air and ocean current due to lag time between spill and image capture – this phenomenon also contributes to why some slicks exist without suspect in sight. Other times, slicks can be confused with false positives from weather events, natural disasters, coastal features, natural seeps, and other anomalies.

Bilge dumps off the coast of China.

This figure displays two likely slicks from intersecting bilge dumps off the coast of China. Due to their kinky shapes, these slicks are likely several days old; this image also shows the influence of time and natural forces on slick appearance.

Although I discovered how to be more transparent with terminology and make better imagery-based speculations, I did not know enough about slicks themselves. Oil naturally exists in the earth, and we harvest it to power our consumptive, energetic lifestyles. Sometimes, the oil itself leaks. In other other cases, wastewater produced during offshore drilling processes is released by us. This produced water is known as brine. Brine contains inorganic substances, toxic matter, and variably sized oil particles that must be properly disposed of or treated before release; it can be treated on platform and discharged into open water, transported to an offsite facility treatment or disposal facility, or put into beneficial reuse – for irrigation, recycled flowback fluid for other drilling operations, or as a substance for ice control (“Produced Water 101”, 2017).

Unlike shoplifting or arson, oil slicks are not always a result of unlawfulness. Some slicks are consequences of legal dumping – legality depending on individual cases in regard to international and country approvals. Accidental spills and leaks also occur and must be cleaned up. However, not all slicks are legally permitted or accidental and concern arises when oil slicks appear without record. Bilge dumping is one indicator of purposeful, often illegal, offshore pollution. To relieve ship weight and space, ships release oily waste from their engine and fuel systems, flushing residual material out of their cargo holds. This is highly illegal, as noted by a case in 2016, when Caribbean Princess, a luxury cruise ship under Princess Cruises, was fined $40 million for illegally discharging thousands of gallons of bilge. Senior intern, Daniel Nicholls, spotted a similar incident with another Princess Cruises ship in late January, indicating an ongoing dilemma.

A bilge dump from the Sapphire Princess.

Nicholls’s annotated Sentinel 1 radar satellite image of possible bilge dumping by Caribbean Princess-owned cruise ship, Sapphire Princess, as it heads towards Kuala Lumpur, Malaysia. Check out the full post here.

Now, I understand slicks not just as mascara tears or eyeliner blunders across a wrinkled ocean display; but as oily remnants with purpose and disposition. This comprehension allows me to more appropriately identify and interpret oil slicks in marine environments. As valuable as this was process and realization was, I registered that the beluga colored specks, aka the potential sources of the slicks, were still unidentified. Probably boats….

Who’s to blame? The murky dilemma of oil spill accountability

As global energy consumption continues to grow, Trinidad and Tobago — a small, Caribbean nation rich in oil and gas resources — has become one of the top exporters of liquified natural gas (LNG) in the world.  But the benefits to the economy of Trinidad and Tobago have come with a cost: chronic leaks and spills from aging oil and gas infrastructure on and offshore.

In early July, an abandoned oil well off of the west coast of Trinidad ruptured, sending dangerous hydrocarbons spewing into the ocean.  Trinidad and Tobago’s state-owned oil company Petrotrin stepped in to help address the rupture, but six days after the orphaned well erupted, the Ministry of Energy was still trying to determine which private company was responsible.

Insufficient documentation and incomplete record-keeping makes response efforts more difficult.  Gary Aboud, Corporate Secretary of Trinidad and Tobago’s Fishermen and Friends of the Sea (FFOS), summed up the the deeper issue in Trinidad concisely: Who is the responsible party? Nothing has been done all week…There are literally hundreds of decades-old, capped, orphaned or abandoned wells which may not have been properly decommissioned, and are corroding.”

Map of Trinidad and Tobago's energy resources.

Energy map of Trinidad and Tobago. Source: The National Gas Company of Trinidad and Tobago.

Better documentation about ownership and better geospatial data showing oil and gas fields, pipelines, and abandoned wells would be a step in the right direction.  Some of this information is available but, as the map above shows, much of it is in a form that is very difficult to use. This energy map is pretty ‘busy’ — the various oil fields, gas fields, and pipelines depicted together make it difficult to use, especially in a crisis scenario like responding to an oil spill.

Officials need a comprehensive geospatial data set — filled with attributes like ownership or responsible party — that they can easily examine, especially during crises like this one.  One of my tasks as an intern at SkyTruth has been to pick apart the existing information (including the map above) and provide it as a robust geospatial data set that’s easy for the public to use.

A map of Trinidad and Tobago's gas fields and gas pipelines.

Gas fields and gas pipelines in Trinidad and Tobago, digitized from the map above.

I’ve digitized all of the oil and gas fields, pipelines, and existing platforms around Trinidad and Tobago, and I’m constantly adding in new fields and data that I’m collecting about these features.

Map of Trinidad and Tobago's oil fields and oil pipelines.

Oil fields and oil pipelines in Trinidad and Tobago, digitized from the map above.

I’m using these new data sets — combined with Sentinel 1 radar satellite imagery — to help monitor oil leaks and spills around Trinidad and Tobago like the one described above.  Having better geospatial data will improve not only how companies handle clean-ups, but will also provide local fisherfolk with more insight into leaks and spills from oil infrastructure as they happen.

Sediment or Oil?

You may recall we posted about a slick emerging from an unidentified platform off the coast of the Democratic Republic of the Congo on June 4th. At the time, we noted that the slick was most likely directed by the strong currents from the nearby entrance to the Congo River as it wasn’t in line with the wind direction. In this image from June 28, we now see a second slick alongside the first.

Sentinel 1 imagery showing the slicks visible with Synthetic Aperture Radar.

This could be a sign of new construction in the area. We also noticed a slick closer to shore which led us to check Sentinel 2 imagery which allows us to see in the visual spectrum. In the inset image, from June 8th, we can see that there are long, brown trails coming from the platforms, usually a sign of sediment being kicked up by wake turbulence from strong currents hitting the structures.

Detailed view of one of the trails in Sentinel 2 imagery.

This raises the possibility that the slicks we are seeing on the radar images are not from oil but from sediment plumes. Turbidity and sediment in the water can dampen wind-driven wavelets, just like an oil slick, making a dark slick on a radar image. The fact that the wind was very low in these images, between 0-5 knots, could possibly emphasize the sediment plumes against the slack water, making them more visible than usual.

The original slick we reported on in June.

However, the way that these slicks remain coherent over 50km lends weight to them being comprised of an oily substance, especially the feathering pattern seen in the middle. This is consistent with what we expect from wind and currents pulling an oily slick in different directions.  So another possibility is that we’re seeing the intentional discharge of drilling fluids and/or “produced water” that includes residual amounts of oil.

In the end, we cannot say with certainty what we are seeing in these images. There is evidence supporting chronic leaking or discharge from the platforms, but there is also support for these being trails of sediment, kicked up by the strong currents coming from the Congo River. It’s times like this that we need some ground truth to help solve the mystery.

The Search for Sanchi

On January 6th, a tanker named the Sanchi collided with a cargo ship called the CF Crystal in the East China Sea causing a fire which killed nearly all of the crew and eventually sank the Sanchi. While the CF Crystal (which survived the collision) was only carrying grain, the Sanchi was carrying natural-gas condensate. This ultra-light oil is highly flammable which no doubt contributed to the blaze that prevented any rescue of the crew. Though there was originally hope it would evaporate quickly, there have been reports of it approaching the Japanese coastline. More persistent heavy bunker oil from the ship’s fuel tanks might also be leaking, compounding the problem.

Usually, we use radar imagery collected by the European Space Agency’s Sentinel 1 satellite to track and monitor oil slicks, but, in this case, the area is not completely covered by Sentinel 1, and what imagery we have seen has been washed out by strong winds that make it difficult to see slicks. We’ve been relying on multispectral imagery from Sentinel 2, but heavy cloud cover in the area has made it difficult to locate the slick and monitor the cleanup and salvage operations.

These Sentinel 2 images do not show the slick as clearly as radar images would. Because we are working in the visible spectrum, we can only see a faint difference between the ocean and the lighter-than-usual slick. We’ve done our best to boost the contrast to highlight the slick, so the color of the water might seem a little brighter than usual.

Sentinel 2 image taken on January 18, showing vessels and slick around site of Sanchi wreck. We inferred the location of Sanchi based on the movements of response vessels, reconstructed from their AIS tracking broadcasts.

We can see two vessels which appear to be either spraying chemicals to disperse the slick or deploying oil-skimming gear, from booms extending from either side, as shown in this zoomed image:

Closeup view of the previous image, showing cleanup vessel in greater detail.

This Planet image, also taken on January 18, showing part of a larger area of slick east of the Sanchi.

Thanks to Planet and their fleet of Dove satellites, we can see that the slick extends further to the east. We are also able to see the vessels in more detail:

This collection of close-up shows views of oil spill response vessels in the area from the previous image.

We have been following the ships in the area via their Automatic Identification System (AIS) broadcasts, and have seen a variety of Chinese and Japanese vessels come and go, including the Koyo Maru and Koshiki, Japanese patrol boats; the Dong Lei 6, a cleanup tanker; the Shen Qian Hao, a Chinese diving vessel; the Hai Xun 01, a Chinese Patrol Boat; and the Dong Hai Jiu 101, a Chinese Search and Rescue boat.  Based on the movements of these vessels, we’ve inferred the location where the Sanchi likely sank and is the source of this ongoing spill.

We are doing our best to monitor this area as the clean-up continues.

Monitoring Update: Oil Rocks In The Caspian Sea

The Oil Rocks (Neft Daşları)  is a massive offshore oil complex in the Caspian Sea. The complex was constructed in the late 1940’s by the Soviet Union and has been producing oil consistently since 1951. The area around the Oil Rocks has experienced catastrophe in the past, when a fire at a nearby platform was responsible for the death of 32 workers and a particularly nasty oil spill in December 2015.

As part of SkyTruth’s Watchdog program, we keep an eye on locations such as this. Over the past 2 months, we estimate that over 380,000 gallons of oil have leaked into the Caspian Sea, based on our assumption that the slicks we are observing are 1 micron (1/1000th of a millimeter) thick.

Above: The European Space Agency’s Sentinel 1 radar satellite captured this image on December 9th revealing a 306 square kilometer oil slick surrounding the Oil Rocks complex.

Above: Sentinel 1 collected this image of the Oil Rocks with a much smaller oil slick (23 square kilometers) on December 21, 2017.

Wind speeds in the Caspian Sea were as strong as 35 knots toward the south on December 21st and may have dispersed an additional volume of oil on the water’s surface.

Above: Sentinel 1 imagery from January 7, 2018 reveals the Oil Rocks leaking oil. The slicks cover a total area of 34 square kilometers.

Wind speeds were very low (between 0-15 knots) on January 7th heading southward, allowing the oil to form slicks around the complex.

And on January 13th, they were between 20-30 knots also heading southward. Similar to the image from December 21st, the high wind speeds may have contributed to dispersing the oil.

Above: The most recent Sentinel 1 image collected on January 19, 2018 reveals a massive oil slick emanating from the Oil Rocks complex, covering an area of 1094 square kilometers and containing at least 288,940 gallons of oil.

For context, 50,000 gallons of oil leaked from the SOCAR#10 platform during a fatal fire in 2015 mentioned above. And this massive Azerbaijani complex has a consistent leaky history on satellite imagery. Azerbaijan, Iran, Kazakhstan, Russia and Turkmenistan, the five countries surrounding the Caspian, all have efforts to tap into the Sea’s 44 billion barrel reserve. But this most recent satellite image from January 19th suggests a troubling future for the environment of the Caspian Sea.

Sentinel 1 radar satellite image showing small slick in the Gulf of Mexico on December 23, 2017, indicating an apparent leak or spill from an oil platform. Image courtesy of the European Space Agency (ESA).

A Look at the Risks of Offshore Drilling as Trump Proposes Largest Leasing Expansion Ever

Last week, the Trump Administration proposed to open up nearly all Federal waters for oil and gas drilling, reversing decades-long protection of areas in the Atlantic, Pacific and Arctic Oceans and in the eastern Gulf of Mexico.

On December 23rd, the European Space Agency’s Sentinel 1 satellite collected an image of the Heater Plf platform which appears to be leaking oil in the Gulf of Mexico. The Heater Plf is located approximately 13 kilometers north of the Delta National Wildlife Refuge in the Mississippi Delta. Based on our conservative analysis, the slick detected on this image holds at least 220 gallons of oil.

Sentinel 1 radar satellite image showing small slick in the Gulf of Mexico on December 23, 2017, indicating an apparent leak or spill from an oil platform. Image courtesy of the European Space Agency (ESA).

The image above reveals a small slick indicating a minor spill. But it’s the latest illustration of a much larger problem that doesn’t get much attention: the steady drumbeat of leaks and spills that accompany offshore drilling. For more examples illustrating this stealthy issue, see our analysis calculating the volume of the ongoing Taylor leak, our map of 10,000 spills since the BP spill and our work with the Gulf Monitoring Consortium.

This proposal to open Federal waters comes hand-in-hand with an announcement made last week, in which the Administration proposed reducing safety regulations on oil and gas drilling in the outer continental shelf. Currently, only the coastal waters of Texas, Louisiana, Mississippi, and Alabama in the Gulf of Mexico can be leased for drilling. These proposed reductions nullify safety rules which were put in place after the fatal and costly BP / Deepwater Horizon disaster of 2010.

The Heater Plf platform is owned by New Century Exploration LLC, and according to information provided by the Bureau of Ocean Energy Management (BOEM) the platform is no longer producing oil but has not been removed. In this 2014 legal brief, New Century and Champion Exploration LLC state that the new requirements under the Outer Continental Shelf Lands Act made it “infeasible” for them to prove their financial capability to handle oil spill preparedness. After the Deepwater Horizon spill, drilling companies were required to estimate the cost to clean up a “worst-case” scenario. New Century could not demonstrate they had the $1.8 billion required to clean up a worst-case spill, so they ceased activity at this site.

With the administration pushing to open new areas to drilling, it seems reckless to be walking back the safety rules put in place to help prevent the next catastrophic oil spill. More than seven years later, the Gulf is still feeling the effects of the largest accidental oil spill in history, while the drilling industry argues these regulations are burdensome to stakeholders and unnecessary. Can we afford to relax our safety standards regarding oil and gas drilling? Can the Gulf of Mexico survive another worst-case scenario like the Deepwater Horizon spill? Can we risk these disasters in the Gulf and elsewhere? Do we want the onus of recovery from these incidents to be shouldered by the taxpayers?

It’s an essential part of our democracy to voice your opinion about decisions that jeopardize public safety and the health of our public lands and waters. Submit your comment on the proposed rollback of the offshore drilling safety rules to the Federal Register by January 29, 2018.

Landsat 8 image from June 21, 2014 showing the oil slick from the Taylor Energy site.

Taylor Energy (Site 23051) Cumulative Spill Report – 2017 Update

With President Trump preparing to open the Atlantic coastline to offshore drilling, we thought it would be a good time to revisit the cautionary tale of Site 23051 — Taylor Energy’s 13-year old continuous oil leak in the Gulf of Mexico.

 

We’ve estimated the cumulative amount of oil that has leaked from the Taylor Energy site since 2004, finding:

  1. Crude oil has been leaking continuously from this site for more than 13 years; and
  2. The estimated cumulative volume of crude oil spilled into the Gulf of Mexico from this chronic leak over the period 2004 – 2017 now stands between 855,421 and 3,991,963 gallons.

 

BACKGROUND

 

The Taylor Energy site perfectly captures the dysfunction of offshore oil development: In 2004, an underwater mudslide caused by Hurricane Ivan toppled one of the company’s platforms and buried the damaged wells attached to it on the seafloor.  Reports of oil on the surface at the site of the wreckage followed shortly after and a secretive clean-up effort ensued.  

 

In 2008, after several failed attempts to stop the leaks and Taylor Energy’s decision to sell off all of its income-generating oil and gas assets in the Gulf, federal regulators ordered the company to post a $666.3 million security bond to ensure there was enough money to plug the wells and clean up remaining pollution.  

 

In  2010 and 2011, Taylor Energy used a leased drill rig called the Ocean Saratoga to slowly find and plug some of the damaged wells (only 9 of the 25 wells at the site have been plugged).  Additionally, three underwater containment domes and an underwater collection and containment system were put in place at the wellhead area to try and capture any remaining oil.

 

Taylor Energy’s next step was to sue the government to try and recover more than $400 million from the trust they had set up previously.  The lawsuit is in limbo amid negotiations over the company’s remaining responsibility and the feasibility of further clean-up. Documents filed by the Justice Department on December 15th revealed new evidence of two plumes of oil and gas resulting in an “ongoing oil release,” bringing some renewed hope Taylor Energy will be held accountable for its mess.

 

CUMULATIVE SPILL ESTIMATES

SkyTruth became aware of the chronic leak from the Taylor Energy site in 2010 while analyzing satellite imagery of the BP / Deepwater Horizon disaster.  We’ve reported on slicks coming from the Taylor Energy site dozens of times in the years since, and in 2012 we released a cumulative spill report estimating that between 300,000 and 1.4 million gallons of oil had leaked from the site since 2004.  But with offshore drilling in the Atlantic looming once again, we thought now would be a good time to revisit those calculations and reconsider the risks that offshore drilling poses for coastal communities.

 

Our initial report estimated the cumulative amount of oil that had leaked from the Taylor Energy site over the period 2004-2011. We’ve updated those calculations to include years 2012-2017, finding that:

  1. Crude oil has been leaking continuously from this site for more than 13 years; and
  2. The estimated cumulative volume of crude oil spilled into the Gulf of Mexico from this chronic leak over the period 2004 – 2017 now stands between 855,421 and 3,991,963 gallons.

 

Our 2017 update uses the same methods outlined in our 2012 cumulative spill report. Our update analyzes the information contained in 2,719 public pollution reports filed with the National Response Center. Most reports were likely filed by Taylor Energy or their contractors covering 2,275 out of 4,852 days (just 47%) from the first report of oil at the site on September 17, 2004, through December 12, 2017.  We computed an ‘estimated average daily slick extent,’ and from that, we derived an ‘estimated average daily flow rate’ for each calendar year since the spill began.  Multiply the daily flow rate by the number of days the site has been leaking, and you have a rough estimate of the cumulative volume of the spill. For more on the methods, see our original report.  The data and analysis are accessible here.

 

In addition to our reliance on the accuracy of the pollution reports submitted by Taylor Energy, there are two assumptions we used to compute the average daily flow rate:

  • the average oil thickness in observable slicks; and
  • the average rate of degradation of an oil slick expressed as a half-life.

For average thickness, we used our conservative standard of 1 micron (1 millionth of a meter); we also computed everything using an even more conservative estimate of 0.5 microns to reflect the possibility that this slick is thinner than most.  For degradation half-life, we assumed that one half of a given amount of a thin slick of oil on the surface of the ocean would degrade in 3-7 days. We believe this range is a very conservative assumption because the longer the assumed lifetime of oil on the surface of the water, the lower the implied daily flow rate will be.

Combining all our data on slick extent with the high and low values for each of the key assumptions, we get four values for estimated cumulative oil spilled:

Half-life (Days) Thickness (Microns) Estimate (Gallons)
3 1.0 3,991,963
3 0.5 1,995,981
7 1.0 1,710,841
7 0.5 855,421

 

There is another potentially troubling trend in the data: since 2015, the average daily reported sheen extent has been significantly larger than in the past, while the number of pollution reports submitted to the NRC has come down.  

 

Average Daily Reported Sheen Extent
Year 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008
# of reports 192 176 371 346 361 323 130 167 381 272
# of days with reports 161 147 328 314 302 309 128 164 260 162
Average daily reported sheen extent (sq. mi) 12.845 14.351 15.330 4.423 1.572 0.337 1.10 1.70 5.83 2.73

 

On the one hand, these numbers could be the result of more diligent and accurate measurements made during routine monitoring and overflights, spurred in part by the public scrutiny this chronic leak has come under due to the work of SkyTruth and our partners in the Gulf Monitoring Consortium.  On the other hand, they could be the result of some qualitative change on the seafloor, in the damaged wells, or in the subsea reservoir that is allowing larger amounts of oil to leak out into the Gulf.

NEXT STEPS

The slight decrease in average reported sheen size over the past three years is somewhat encouraging: if the significant jump in 2015 was indeed due to more accurate reporting by Taylor Energy, then this recent downward trend could indicate the leaks are finally slowing.  But we are hampered by our dependence on observations and reports submitted by the responsible party, Taylor Energy.  These reports have been proven inaccurate, systematically underestimating the size of the slick by more than an order of magnitude compared with independent measurements based on direct observation of the slick on satellite imagery.  Direct, regular measurement and observation of the leak by a neutral party is crucial to understanding what is happening and predicting the likely future at this site. For this reason, we will continue our monitoring work.