Solving methane mysteries with satellite imagery
By combining two types of remote sensing data, we can see hidden leaks from ships
The methane menace
Methane emissions from the fossil fuel industry are a key challenge in the global energy transition. It’s now well understood that deliberate venting and accidental leaking of methane from oil and gas infrastructure across the world is a significant contributor to global warming, with methane in general being responsible for as much as 30% of the rise in global temperatures since the Industrial Revolution.
The international community has reacted. In 2021, the United States and European Union announced the Global Methane Pledge and the International Energy Agency (IEA) now publishes an annual Global Methane Tracker report. At least three major satellite programmes now exist — NASA EMIT, GHGSat and MethaneSAT — and the data that they’re sending back to Earth is making darkness visible, revealing huge plumes that have in some cases been flowing for decades.
A growing body of scientific work has emerged around this new data, using visual satellite imagery and maps of global oil and gas infrastructure to help identify and attribute the methane plumes observed from space. This is also the approach taken by the industry-led Oil and Gas Climate Initiative (OGCI), which published the results of a satellite monitoring campaign covering Kazakhstan, Algeria and Egypt this year.
This approach works beautifully for static oil and gas assets like wells and compressor stations. But what about things that move? One of the key points of disagreement in the global methane debate is around liquefied natural gas (LNG) shipping — that is, the transformation of methane into a liquid, its transportation around the world in super-chilled tankers and its subsequent ‘regasification’ at import terminals.
If you believe the industry, increased use of LNG is vital to global decarbonisation, particularly for displacing dirtier coal-fired electricity generation in Asia over the next few decades. By contrast, some scientists — notably Robert Howarth at Cornell University — have argued that when fugitive emissions from tankers are added to the significant upstream emissions from gas extraction, and depending on the length of the journey taken and the fuel used, LNG may be responsible for more CO2-equivalent emissions than coal.
A closer look at shipping
Over the past few months, we’ve been doing an increasing amount of work on LNG — our latest briefing (also available in Japanese), produced for Jubilee Australia, looks at how Japanese companies are re-selling Australian LNG to neighbouring countries in Southeast Asia, at the same time as funding new import terminals and gas-fired generation capacity in those same countries. To carry out this type of analysis, we pay for a live feed of LNG tanker positions from the London Stock Exchange Group (LSEG).
What if we overlaid this data on methane plumes observed by satellites? Could we find examples of significant emissions directly from tankers?
To test this idea out, we asked our friends at Global Energy Monitor (GEM) to take a look at their Global Methane Emitters Tracker (GMET), which associates plumes observed in publicly available sources with data on oil and gas infrastructure locations from GEM’s broader datasets. Sarah, GMET’s project manager, sent us a list of all methane plumes in their database observed in or around LNG import and export terminals in the United States, where the most comprehensive measurement campaigns have so far taken place.
We took this data and examined each plume in detail, paying special attention to the date and time and whether or not an LNG tanker was in the vicinity, as well as looking at the tankers’ recorded behaviour in LSEG data — were they loading, unloading or neither at the time of the plume?
Of the 12 plumes identified by GEM, the majority could be attributed to flaring — when excess gas is burned off at the terminal for a variety of technical and process reasons. But one plume was different, appearing to come directly from the Diamond Gas Rose tanker (IMO 9779238) while AIS data showed it was docked at the Cameron LNG terminal in Louisiana. By overlaying a TIFF image of the plume from Carbon Mapper with a Planet satellite image of the same spot taken just 83 minutes earlier — a lucky find! — we were able to create some striking composite images showing the leak in action.
While this emissions event is now relatively old, having occurred on 28 April 2021, it’s the first time we’ve seen a plume from an LNG tanker identified in public remote sensing data. I particularly like the way that the two images — taken less than an hour and a half apart — complement each other: the bright colours of the detected methane plume, invisible to the naked eye, shade into clouds visible in the underlying optical image. These may have been caused by the reaction between methane and oxygen in the air, which produces water vapour, or by other substances present in the plume. Update, 8 October: Ignore me, they’re just clouds.
To try to understand the plume’s exact source, we contacted the authors of a January 2024 study by the International Council on Clean Transportation (ICCT), which looks in detail at fugitive and unburned methane emissions from ships. Drawing on their experience of measuring emissions directly on board LNG-powered ships and at export terminals, they pointed to a number of potential sources, ranging from engine slip (unburned methane in exhaust fumes) from the vessel’s auxiliary engine to direct leaks from the terminal’s LNG loading arms. They also cautioned that wind conditions can make it hard to attribute plumes directly to sources on the ground — an important consideration, as according to historical data 28 April 2021 was a relatively windy day in Louisiana.
What next?
According to Carbon Mapper’s estimates, the plume we’re looking at here was emitting methane at a rate of around 3.4 tonnes per hour. This is as much as some of the largest plumes observed from static oil and gas infrastructure in countries notorious for their methane-intensive upstream oil and gas production — for example, Algeria and Turkmenistan.
But ships are different from oil wells. Rather than emitting methane constantly, the largest emissions from LNG tankers appear to happen at specific times, during cargo loading and unloading or during bunkering (refuelling). Unlike oil wells, ships move around, and current satellites have a hard time detecting methane plumes over water, meaning that researchers without direct access to the ships are left poring over chance images like the one we’ve found. All this makes it especially challenging to come up with reliable estimates for the LNG shipping industry’s emissions as a whole.
There is hope, however. As more and more satellite-based methane emissions data comes online, repeating this analysis will likely yield more and more observations linked to LNG terminals around the world — and, in some cases, plumes emerging directly from LNG tankers. This data will feed into projects like GEM’s Global Methane Emitters Tracker and eventually into the methodologies behind major initiatives like ClimateTRACE, where efforts are currently focussed on estimating emissions from the global shipping industry.
Speaking on a recent episode of the Cleaning Up podcast, Bryony Worthington described an advertising campaign around pollution that she’d been involved with in the 2000s: “We had a reasonable budget, and we took some stock footage, and we coloured in the gases so you could see a luminescent colour that we added to all the exhaust of the cars and the chimneys, and it all built up in the atmosphere. And the tagline was, ‘if you could see the blanket of gases, you’d do something about it.’” With methane emissions becoming more visible by the day, we hope this maxim holds true.