Satellite images reveal the extent of methane leaks across the world

At the end of 2022, the press revealed that Cen­tral Asia, the Mid­dle East, and the Unit­ed States are home to around fifty indus­tri­al sites releas­ing large quan­ti­ties of methane – a green­house gas – into the atmos­phere. The pub­lic then dis­cov­ered the methane “super-emit­ters”, indus­tri­al­ists in the fos­sil ener­gy, waste treat­ment and agri­cul­tur­al sec­tors. Anthro­pogenic methane emis­sions are main­ly due to live­stock farm­ing, rice cul­ti­va­tion, the decom­po­si­tion of waste in land­fills and the exploita­tion of fos­sil fuels. The rev­e­la­tions are par­tic­u­lar­ly strik­ing for the fos­sil indus­try: these super-emis­sions are the result of leaks from instal­la­tions (wells or pipelines) or incom­plete flar­ing – or even a lack of flar­ing – in oil, coal or gas instal­la­tions, a prac­tice which nev­er­the­less pre­vents the release of methane into the atmosphere.

The dis­cov­ery was made thanks to data from the EMIT satel­lite mis­sion of NASA. “Before the satel­lite data, we sus­pect­ed that sig­nif­i­cant methane leaks were occur­ring, but we did­n’t know where, when or how much,” explains Marielle Saunois. “Satel­lites allow methane leaks to be mea­sured inde­pen­dent­ly of indus­try esti­mates.” Mea­sure­ments can be tak­en on board aero­planes, weath­er bal­loons and even drones. “These in-situ mea­sure­ments are very localised and spo­radic,” says Marielle Saunois. “How­ev­er, it is impor­tant to be able to mon­i­tor changes in methane con­cen­tra­tions, for exam­ple to check whether the leak has been repaired.”

The first satel­lite obser­va­tions of methane were made in the ear­ly 2000s. On board the Envisat satel­lite, the Euro­pean instru­ment SCIAMACHY was a spec­trom­e­ter ded­i­cat­ed to mea­sur­ing the chem­istry of the atmos­phere – an obser­va­tion instru­ment that deter­mines chem­i­cal com­po­si­tion by decom­pos­ing the light spec­trum. In 2005, the first results from SCIAMACHY made it pos­si­ble to char­ac­terise nat­ur­al and anthro­pogenic sources of methane². At the time, the instru­ment was unable to iden­ti­fy super-emit­ters due to its low spa­tial resolution.

Sub­se­quent gen­er­a­tions of satel­lites have rev­o­lu­tionised our knowl­edge of the plan­et. The Japan­ese GOSAT and Euro­pean IASI mis­sions, and espe­cial­ly the Euro­pean TROPOMI instru­ment on board the Euro­pean Sen­tinel-5P satel­lite since 2017, pro­vide data at fin­er res­o­lu­tions. Every day, TROPOMI pro­vides mea­sure­ments of methane con­cen­tra­tion across the globe at a res­o­lu­tion of 5.5 x 7 km². In addi­tion to this tech­no­log­i­cal advance, there are devel­op­ments in data pro­cess­ing, such as the use of neur­al net­works and arti­fi­cial intel­li­gence³.

In 2022, in the jour­nal Sci­ence, an inter­na­tion­al team revealed exten­sive methane emis­sions using data from TROPOMI⁴. More than 1,800 anom­alies asso­ci­at­ed with super-emit­ter sites (releas­ing more than 25 tonnes of methane per hour) were detect­ed on the plan­et over the peri­od 2019–2020. Two-thirds of the events are linked to oil and gas pro­duc­tion, and occur main­ly in Rus­sia, Turk­menistan, the Unit­ed States, the Mid­dle East and Algeria.

Anoth­er rev­e­la­tion of the satel­lite images was that the offi­cial data great­ly under­es­ti­mate methane emis­sions. In 2018, based on ground mea­sure­ments, a team already point­ed out that its esti­mate of emis­sions from the oil and gas sec­tor in the Unit­ed States was 60% high­er than that of the Envi­ron­men­tal Pro­tec­tion Agency inven­to­ry⁵. In 2023, anoth­er team (Shen et al., 2023, in the fig­ure below) used satel­lite data to quan­ti­fy nation­al methane emis­sions from hydro­car­bon exploita­tion⁶. The result: total emis­sions are 30% high­er than those offi­cial­ly report­ed by states under the Unit­ed Nations Frame­work Con­ven­tion on Cli­mate Change, main­ly due to under-report­ing by the four high­est-emit­ting coun­tries (Unit­ed States, Rus­sia, Venezuela and Turk­menistan). Methane emis­sions amount to 62 mil­lion tonnes per year for the oil and gas sec­tor, and 32 mil­lion tonnes per year for the coal sec­tor. For the cli­mate, this rep­re­sents the equiv­a­lent of 20 years of CO₂ emis­sions from nat­ur­al gas com­bus­tion, accord­ing to the authors.

Oth­er esti­mates have been made by var­i­ous insti­tu­tions: the Glob­al Methane Bud­get sci­en­tif­ic con­sor­tium and the Inter­na­tion­al Ener­gy Agency (IEA). All are high­er than the data sub­mit­ted by the States and are much high­er than those pro­vid­ed by the indus­tries them­selves. “Methane is the sec­ond most preva­lent green­house gas emit­ted by human activ­i­ty: it is cru­cial to have a good under­stand­ing of its emis­sions in order to reduce them and mit­i­gate cli­mate change,” points out Marielle Saunois. In Octo­ber 2024, aver­age methane con­cen­tra­tions in the atmos­phere reached 1,943 parts per mil­lion, 2.6 times high­er than before the indus­tri­al era⁷. Anthro­pogenic emis­sions (those linked to human activ­i­ty) account for around two-thirds of methane releas­es, with nat­ur­al sources main­ly being wet­lands and inland fresh­wa­ters⁸. While the IEA believes that reduc­ing emis­sions from fos­sil fuel indus­tries is “one of the most prag­mat­ic and cost-effec­tive options for reduc­ing green­house gas emis­sions,” satel­lites are valu­able allies in iden­ti­fy­ing the most effec­tive levers.

“In recent years, the num­ber of satel­lite mis­sions ded­i­cat­ed to methane has been increas­ing, both for sci­en­tif­ic rea­sons and to expand the range of ser­vices avail­able to indus­try,” analy­ses Marielle Saunois. Many pri­vate com­pa­nies are launch­ing their own satel­lites to help indus­try iden­ti­fy methane leaks on their sites. The detec­tion thresh­olds for methane plumes – i.e. the low­est con­cen­tra­tion at which the satel­lite can detect them – are being low­ered all the time: this could be use­ful for waste indus­tries, whose emis­sions are less con­cen­trat­ed. Anoth­er ini­tia­tive: at COP27, an inter­na­tion­al alert sys­tem, called MARS, was launched by the Unit­ed Nations Envi­ron­ment Pro­gramme. By com­pil­ing all pub­lic satel­lite data, it detects major leaks and alerts the coun­tries and indus­tries con­cerned. Dur­ing the first 9 months of its oper­a­tion, the sys­tem detect­ed 500 events.

So, are satel­lites the Holy Grail for track­ing our methane emis­sions? In an edi­to­r­i­al pub­lished in Nature¹¹, astro­physi­cist Lor­na Fin­man denounces an exag­ger­at­ed enthu­si­asm: “Their effec­tive­ness is often ham­pered by lim­it­ed spa­tial res­o­lu­tion, atmos­pher­ic inter­fer­ence and the chal­lenge of pre­cise­ly iden­ti­fy­ing spe­cif­ic emis­sion sources.” She calls for the devel­op­ment of ground-based and aer­i­al obser­va­tions to improve the accu­ra­cy of methane mon­i­tor­ing. Marielle Saunois replies: “It is of course nec­es­sary to step up our efforts to improve our esti­mates of methane emis­sions sec­tor by sec­tor. Satel­lite data can­not detect every­thing: there is a revis­it delay, mea­sure­ments are obstruct­ed by clouds, releas­es into the sea are more dif­fi­cult to detect and satel­lite mea­sure­ments face cer­tain bias­es. But methane leaks from the fos­sil fuel indus­try or mega-land­fills could be eas­i­ly avoid­ed – unlike those from agri­cul­ture. It is cru­cial to encour­age indus­tri­al­ists to improve their infra­struc­tures, and satel­lites help to iden­ti­fy the sites respon­si­ble for these major leaks and to mon­i­tor them.” 

Anaïs Marechal