Industries / Energy, water and waste / Environmental services / Air quality monitoring
Satellites
Satellites measure gases and particles from orbit by reading how sunlight reflected back from Earth has been absorbed on its way through the air.
6
Milestones achieved, from 1978 to 2024.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1978
Ozone mapped from orbit
NASA launched Nimbus-7 on 24 October 1978 carrying the Total Ozone Mapping Spectrometer (TOMS) and a companion ultraviolet instrument (SBUV), which began NASA's satellite record of the ozone layer. The Antarctic ozone hole was first reported from the ground: Joe Farman, Brian Gardiner and Jonathan Shanklin of the British Antarctic Survey published the springtime losses over their Halley station in Nature in 1985. In 1986 NASA scientists used TOMS and SBUV data to show that the hole covered the Antarctic region.
LessonA long record at one ground station caught the loss first, and the satellite then showed its size.
Credit
NASA Goddard Space Flight Center; Farman, Gardiner and Shanklin, British Antarctic Survey
2009
Carbon satellite lost at launch
NASA's Orbiting Carbon Observatory (OCO), built to measure carbon dioxide from orbit, was lost in 2009 when the nose cover (payload fairing) of its Taurus XL rocket failed to separate, and NASA's Glory satellite was lost the same way in 2011. A rebuilt OCO-2 launched on 2 July 2014. In 2019 NASA's investigation traced both failures to aluminum parts from the supplier Sapa Profiles, which "altered test results and provided false certifications"; NASA put the cost of the two losses at more than $700 million.
LessonNASA said it cannot retest every part and relies on suppliers' test certificates.
Credit
NASA Launch Services Program (the investigation); NASA (OCO-2)
2010
Fine particles mapped worldwide
Aaron van Donkelaar, Randall Martin, Michael Brauer and colleagues combined aerosol optical depth, a measure of how much the air dims sunlight, from NASA's MODIS and MISR instruments with a chemistry model to map ground-level fine particles (PM2.5) at about 10 km resolution for 2001 to 2006, including places with no monitors. They found a population-weighted average of 20 micrograms per cubic meter, and agreement with North American ground monitors (correlation r = 0.77 over 1,057 comparisons).
LessonThe authors put the uncertainty at about 25%, so such maps give long-term averages over areas, not readings for a single street or day.
Credit
Aaron van Donkelaar, Randall Martin, Michael Brauer and co-authors
2016
Missing polluters found from orbit
Chris McLinden of Environment and Climate Change Canada and colleagues used 2005–2014 data from the Dutch-Finnish Ozone Monitoring Instrument (OMI) on NASA's Aura satellite to find 39 major human-made sources of sulfur dioxide that were missing from emission inventories: clusters of coal power plants, smelters and oil and gas operations in the Middle East, Mexico and parts of Russia. Unreported and underreported sources made up about 12% of human-made sulfur dioxide, and in some regions reported emissions were two to three times lower than the satellite estimates.
LessonMcLinden called the work "an independent measurement of these emission sources"; inventories built from reported figures can miss whole plants.
Credit
Chris McLinden (Environment and Climate Change Canada) and co-authors, with NASA
2020
Hourly views from geostationary orbit
Every earlier ultraviolet and visible-light satellite instrument for air quality circled in low orbit and saw a place once or twice a day. Korea's Geostationary Environment Monitoring Spectrometer (GEMS), launched in February 2020 on the GEO-KOMPSAT-2 series, stays over Asia and measures ozone, nitrogen dioxide, sulfur dioxide, formaldehyde and aerosols every daylight hour. A Hong Kong University of Science and Technology team checked its 2021 hourly nitrogen dioxide against ground instruments in Xuzhou and found a close match (correlation 0.82), with GEMS reading somewhat low; a University of Bremen team in 2026 called it the longest-operating of the new geostationary instruments and used four years of its data to estimate hourly emissions.
doi.orgmeetingorganizer.copernicus.orgmeetingorganizer.copernicus.org
Credit
Jhoon Kim (Yonsei University) and the GEMS team
2024
Satellites sit a blind test
Evan Sherwin, Adam Brandt and colleagues at Stanford University made 82 metered releases of methane over two months as nine satellite systems passed overhead: the commercial GHGSat-C and WorldView-3, and the government-funded EnMAP, Gaofen 5, Ziyuan 1, Huanjing 2, PRISMA, Landsat 8 and 9, and Sentinel-2. Six analysis teams, not told the release sizes, made 134 estimates. They reported no methane where there was none, missed 41 real releases, detected releases as small as 0.03 tonnes an hour, and put 55% of their estimates within 50% of the metered rate.
LessonWithout a known source location, some true plumes are hard to tell from artifacts in the data, and clouds interfered.
Credit
Evan Sherwin, Adam Brandt and colleagues, Stanford University
Watch for
Sentinel-5's daily global maps
EUMETSAT's Metop-SGA1 satellite, launched in August 2025, carries the Copernicus Sentinel-5 spectrometer. On 26 November 2025 EUMETSAT released its first images, including maps of nitrogen dioxide, sulfur dioxide, formaldehyde and ozone, and said that once commissioning by the European Space Agency (ESA) ends it will deliver daily, high-quality data worldwide. By 16 November 2026 EUMETSAT had declared two other instruments on the satellite operational (a radio occultation sounder in May 2026 and a microwave sounder in August 2026), but not Sentinel-5.
Achieved whenEUMETSAT declares Sentinel-5 operational and releases validated products that outside teams have checked against ground instruments.
Claimed by
EUMETSAT and ESA, November 2025
Three gases, one instrument
Japan's GOSAT-GW satellite, launched on 29 June 2025, carries TANSO-3, an instrument that measures carbon dioxide, methane and nitrogen dioxide together. In January 2026 Japan's National Institute for Environmental Studies (NIES) published a first analysis of all three gases over Los Angeles from uncalibrated data. Calibrated radiance data (Level 1B) went public on 23 July 2026, later than the spring 2026 plan; the gas products (Level 2) are planned for spring 2027.
Achieved whenthe Level 2 products are public and an outside team has validated them against ground measurements.
Claimed by
NIES, Japan's Ministry of the Environment and JAXA, January 2026
Open
01
Stack meters meet GEMS
A plant-by-plant comparison of the hourly nitrogen oxides that Korean power plants' own stack monitors report with the hourly nitrogen dioxide the GEMS satellite sees over the same plants, to test how well a geostationary satellite can follow one plant's daily rise and fall.
The briefCredit · unclaimed02
Check India's scrubbers from orbit
A before-and-after satellite check of sulfur dioxide at every Indian coal plant reported to have a flue-gas desulfurization unit (FGD, a scrubber that strips sulfur from the exhaust), using each unit's commissioning date.
The briefCredit · unclaimed
Open 01 · the brief
Stack meters meet GEMS
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A plant-by-plant comparison of the hourly nitrogen oxides that Korean power plants' own stack monitors report with the hourly nitrogen dioxide the GEMS satellite sees over the same plants, to test how well a geostationary satellite can follow one plant's daily rise and fall.
Why it is still open
Since 3 April 2020 Korea has published real-time readings from every stack fitted with a tele-monitoring system (TMS), through Korea Environment Corporation's CleanSYS open API. GEMS studies so far have estimated hourly emissions for cities and power-plant regions and compared them with inventories, not with individual stack records (Park and colleagues, 2025; Lange and colleagues at the University of Bremen, 2026). The 2026 preprints that test a geostationary satellite against power-plant monitors use NASA's TEMPO over the United States.
data.go.krntrs.nasa.govmeetingorganizer.copernicus.orgdoi.org
Built with what exists
Korea Environment Corporation's CleanSYS real-time API on data.go.kr (free key, JSON); GEMS tropospheric nitrogen dioxide data; ERA5 wind fields from the Copernicus Climate Data Store; Python with xarray; a flux method for point sources such as the one Chen and colleagues published for India in 2026.
First step
Register for a data.go.kr key and start logging hourly CleanSYS readings for ten large plants, since the API serves current readings and the record has to start now.
Done looks like
For at least ten coal or gas plants, a year of hourly satellite-based nitrogen oxide estimates set against the plants' hourly stack readings, with each plant's correlation and the smallest change in output GEMS can detect, published with the code.
Open 02 · the brief
Check India's scrubbers from orbit
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A before-and-after satellite check of sulfur dioxide at every Indian coal plant reported to have a flue-gas desulfurization unit (FGD, a scrubber that strips sulfur from the exhaust), using each unit's commissioning date.
Why it is still open
India's power ministry told Parliament in August 2025 that 57 units had FGD installed; a month earlier, on 11 July 2025, the environment ministry exempted Category C plants, about 78% of units by Health Policy Watch's count, from the sulfur dioxide limit. Chen and colleagues (Atmospheric Measurement Techniques, April 2026) mapped sulfur dioxide emissions from the TROPOMI satellite for 2018 to 2023 and detected about 80% of coal plants above 100 MW at their correct locations, but their abstract reports no before-and-after test of scrubbers. The Centre for Research on Energy and Clean Air (CREA) reported in November 2024 that satellite data showed sulfur dioxide rising from 2019 to 2023 around coal plant clusters, but its plant figures are estimates built from coal use and assumed stack emissions, and its post-scrubber cuts are projections, not measurements.
outlookbusiness.comenergyandcleanair.orghealthpolicy-watch.newsamt.copernicus.org
Built with what exists
TROPOMI sulfur dioxide data from the Copernicus Data Space; the sharpened flux-divergence method of Chen and colleagues; Global Energy Monitor's coal plant tracker for unit locations; parliamentary answers and company annual reports for commissioning dates, read with a model built for long documents (see the Note "The jobs agents repeat"); Python.
First step
Collect commissioning dates for the 57 reported units from Rajya Sabha answers and company filings, and pick five with clear dates to test first.
Done looks like
For each unit with a commissioning date found, sulfur dioxide estimates for the 12 months before and the 12 months after, with uncertainty, in a public table with the code, and a list of plants with no detectable drop sent to their owners for comment.