We review the capability of current and scheduled satellite
observations of atmospheric methane in the shortwave infrared (SWIR) to
quantify methane emissions from the global scale down to point sources. We
cover retrieval methods, precision and accuracy requirements, inverse and
mass balance methods for inferring emissions, source detection thresholds,
and observing system completeness. We classify satellite instruments as area
flux mappers and point source imagers, with complementary attributes. Area
flux mappers are high-precision (<1 %) instruments with 0.1–10 km
pixel size designed to quantify total methane emissions on regional to
global scales. Point source imagers are fine-pixel (<60 m)
instruments designed to quantify individual point sources by imaging of the
plumes. Current area flux mappers include GOSAT (2009–present), which
provides a high-quality record for interpretation of long-term methane
trends, and TROPOMI (2018–present), which provides global continuous daily
mapping to quantify emissions on regional scales. These instruments already
provide a powerful resource to quantify national methane emissions in
support of the Paris Agreement. Current point source imagers include the
GHGSat constellation and several hyperspectral and multispectral land
imaging sensors (PRISMA, Sentinel-2, Landsat-8/9, WorldView-3), with
detection thresholds in the 100–10 000 kg h-1 range that enable
monitoring of large point sources. Future area flux mappers, including
MethaneSAT, GOSAT-GW, Sentinel-5, GeoCarb, and CO2M, will increase the
capability to quantify emissions at high resolution, and the MERLIN lidar
will improve observation of the Arctic. The averaging times required by area
flux mappers to quantify regional emissions depend on pixel size, retrieval
precision, observation density, fraction of successful retrievals, and
return times in a way that varies with the spatial resolution desired. A
similar interplay applies to point source imagers between detection
threshold, spatial coverage, and return time, defining an observing system
completeness. Expanding constellations of point source imagers including
GHGSat and Carbon Mapper over the coming years will greatly improve
observing system completeness for point sources through dense spatial
coverage and frequent return times.
Introduction
Methane is a powerful greenhouse gas that has contributed 0.6 ∘C of
global warming since pre-industrial times (Naik et al., 2021). It is emitted
by a number of anthropogenic source sectors, including livestock, oil and gas
systems, coal mining, landfills, wastewater treatment, and rice cultivation.
Wetlands are the main natural source. The main sink is oxidation by the
hydroxyl radical (OH), resulting in an atmospheric lifetime of about 9 years
(Prather et al., 2012). Because of this short lifetime, decreasing methane
emissions is a powerful lever to slow down near-term greenhouse warming
(Nisbet et al., 2020). However, methane emission estimates and the
contributions from different sectors are highly uncertain (Saunois et al.,
2020), hindering climate policy. Here we review the capability of satellite
observations of atmospheric methane to quantify emissions from the global
scale down to point sources.
Methane emission inventories are typically constructed using bottom-up
methods in which activity levels (such as number of cows) are multiplied by
emission factors (methane emitted per cow; IPCC, 2019). Bottom-up methods
relate emissions to the underlying processes, thus providing a basis for
emission control strategies. Observations of atmospheric methane provide
top-down information to improve these emission estimates by using inverse
methods to relate observed concentrations to emissions (Miller and Michalak,
2017). Satellite observations are of particular interest for this purpose
because of their high observation density and global coverage (Palmer et
al., 2021).
Satellites retrieve atmospheric methane column concentrations with near-unit
sensitivity down to the surface by measuring spectrally resolved
backscattered solar radiation in the shortwave infrared (SWIR; Jacob et
al., 2016). Global observation of methane from space began with the
SCIAMACHY instrument (2003–2014, 30×60 km2 pixels)
(Frankenberg et al., 2005) and has continued since with the TANSO-FTS
instrument aboard GOSAT (2009–present, 10 km circular pixels separated by
about 270 km; Parker et al., 2020) and the TROPOMI instrument
(2018–present, 5.5×7 km2 pixels; Lorente et al., 2021a). Many
studies have used these satellite observations to quantify methane emissions
globally (Bergamaschi et al., 2013; Alexe et al., 2015; Wang et al., 2019;
Qu et al., 2021), on continental scales (Wecht et al., 2014; Maasakkers et
al., 2021; Lu et al., 2022), on finer regional scales (Miller et al., 2019;
Zhang et al., 2020; Shen et al., 2021), and for large point sources (Pandey
et al., 2019; Sadavarte et al., 2021; Lauvaux et al., 2022; Maasakkers et
al., 2022a, b). Targeted observation of methane point sources from space began
with the 2015 Aliso Canyon blowout using the Hyperion hyperspectral sensor
(Thompson et al., 2016) and has since continued with the GHGSat instruments
(2016–present, 25×25 m2 pixels; Jervis et al., 2021).
Hyperspectral land-imaging spectrometers (measuring continuous spectra with
∼10 nm resolution in selected wavelength channels) and
multispectral land-imaging spectrometers (measuring radiances in discrete
∼100 nm channels) have also demonstrated capability to detect
large methane point sources in their SWIR bands (Cusworth et al., 2019;
Guanter et al., 2021; Varon et al., 2021; Ehret et al., 2022; Sanchez-Garcia
et al., 2022).
Better quantification of methane emissions worldwide is urgently needed to
meet the demands of climate policy. Individual countries must report their
emissions by sector to the United Nations Framework Convention on Climate
Change (UNFCCC) on a yearly basis for Annex I (developed) countries. The
enhanced transparency framework of the Paris Agreement requires all
countries to submit national sector-resolved emissions for expert review by
November 2024 as basis for setting their nationally determined contributions
to meet climate goals. Independently of the Paris Agreement, over 110
countries have now signed the Global Methane Pledge of 2021, committing them
to reduce their collective 2030 methane emissions by 30 % relative to 2020
levels. Satellites can help to quantify national emissions by sector as
baseline for setting methane reduction goals and can then monitor emissions
over time to evaluate success in achieving those goals. They provide near-real-time information on emissions, whereas bottom-up inventories typically
have latencies of a few years, and are thus a unique resource to document
rapid changes in emissions (Barré et al., 2021).
Jacob et al. (2016) previously reviewed the state of the science for
quantifying methane emissions from space. They presented observing
capabilities at the time, discussed the inverse methods for inferring
methane emissions from satellite observations, and laid out observing
requirements for future satellite missions. Since then, new satellite
instruments for measuring atmospheric methane have been launched and new
capabilities for detecting methane point sources from space have emerged.
New analytical tools have been developed to infer emissions from satellite
observations, including for point sources. Additional satellite instruments
are scheduled to be launched over the next few years that will augment
current capabilities. These new developments motivate our updated review.
Observing atmospheric methane from spaceCurrent and planned instruments
Table 1 lists current and scheduled satellite instruments with documented or
expected capability for quantifying methane emissions, and Table 2 gives
specific attributes for each. We classify the instruments as area flux
mappers or point source imagers, and Fig. 1 illustrates these two fleets.
Area flux mappers are designed to observe total emissions on global or
regional scales with 0.1–10 km pixel size. Point source imagers are
fine-pixel (<60 m) instruments designed to quantify individual
point sources by imaging the plumes. Point source imagers have much finer
spatial resolution than area flux mappers but lower precision.
Current and planned SWIR satellite instruments for observing
atmospheric methanea.
InstrumentOrganizationbLaunch dateNadir pixel sizeCoverageReturn time (d)cMethane band (mum)dSpectral resolution (nm)ePrecisionfReferenceArea flux mappersgGOSAThJAXA, MOE, NIES200910 km diameteriglobal31.65, 2.3j0.060.7 %Parker et al. (2020); Noël et al. (2022)TROPOMIESA2017k5.5×7 km2global12.30.250.8 %lLorente et al. (2021a)GOSAT-GWJAXA, MOE, NIES20231×1–10×10 km2∗global +targets31.650.060.6 %NIES (2021)MethaneSATEDF2023130×400 m2200×200 km2 targets3–41.650.30.1 %–0.2 %nRohrschneider et al. (2021)Sentinel-5ESA20247.5×7.5 km2global11.65, 2.30.250.8 %ESA (2020)GeoCarbNASA20256×3 km2N and S Americao0.52.30.20.3 %–0.6 %Moore et al. (2018)CO2MESA20252×2 km2global51.650.30.6 %Sierk et al. (2019)MERLINCNES, DLR20270.1×50 km2p global281.653×10-4q1.5 %Ehret et al. (2017)Point source imagersrLandsat-8sUSGS201330×30 m2global162.320030 %–90 %tEhret et al. (2022)WorldView-3Digital Globe20143.7×3.7 m266.5×112 km2 targets<12.3506 %–19 %tSanchez-Garcia et al. (2022)Sentinel-2ESA201520×20 m2global2–52.320030 %–90 %tVaron et al. (2021)GHGSatuGHGSat, Inc.201625×25 m212×12 km2 targets1–7 v1.650.31.5 %wJervis et al. (2021)PRISMAxASI201930×30 m230×30 km2 targets42.3103 %–9 %Guanter et al. (2021)EnMAPxDLR202230×30 m230×30 km2 targets42.3103 %–9 %Cusworth et al. (2019)
Continued.
EMITNASA202260×60 m2Dust-emitting regionsy32.392 %–6 %zCusworth et al. (2019)CarbonMapperaaCarbon Mapper and Planet202330×30, 30×60 m218 km swathsab1–7v2.361.2 %–1.5 %Duren (2021)
a This table lists shortwave infrared (SWIR) satellite instruments
currently operating or scheduled for launch that have documented
methane-observing capabilities and offer publicly accessible data (some for
purchase; see Table 2). Instruments not yet launched are in italics, and
launch dates are estimates as of this writing. All instruments are in low-elevation polar sun-synchronous orbits except for GeoCarb, which will be
in geostationary orbit over the Americas, and EMIT, which will be in an
inclined precessing orbit. All instruments measure SWIR solar radiation
backscattered from the Earth's surface except for MERLIN, which is a lidar
instrument. The Gaofen-5 series of Chinese satelliteshas capabilities
similar to PRISMA and EnMAP (Irakulis-Loitxate et al., 2021) but is not
included in the table because of the opacity of data acquisition and
distribution. A more comprehensive list of instruments, including from
private companies with proprietary data, is available from GEO, ClimateTRACE,
WGIC (2021).
b Organization abbreviations are as follows: JAXA is the Japan Aerospace Exploration Agency, MOE is the Ministry of Environment, NIES is the National Institute for Environmental
Studies, ESA is the European Space Agency, EDF is the Environmental
Defense Fund, NASA is the National Aeronautics and Space Administration,
CNES is the Centre National d'Etudes Spatiales, DLR is the Deutsches
Zentrum für Luft- und Raumfahrt, USGS is the United States Geological
Survey, and ASI is the Agenzia Spaziale Italiana.
c Time interval between successive viewings of the same scene.
d Most useful band(s) for methane retrieval. The 1.65 and 2.3 µm bands have exploitable features at 1.63–1.70 and 2.2–2.4 µm,
respectively.
e Full width at half maximum.
f Precision is reported as a percentage of the retrieved dry column
methane mixing ratio XCH4.
g Area flux mappers are primarily designed to quantify total methane
emissions on regional to global scales.
h The TANSO-FTS instrument aboard the GOSAT satellite. The instrument is
commonly referred to as GOSAT in the literature. GOSAT-2 was launched in
2018 with specifications similar to GOSAT but adding a 2.3 µm band
(Suto et al., 2021).
i Circular pixels separated by about 270 km along-track and
cross-track distance.
j The 2.3 µm band was added in GOSAT-2.
k TROPOMI was launched in October 2017, but the methane data stream
begins in May 2018.
l The TROPOMI product reports a much higher precision of
∼2 ppb, but this only includes error from the measured
radiances. Accounting for retrieval errors by validation with TCCON data
indicates a precision of 0.8 % (Schneising et al., 2019).
∗ Narrow-swath mode (1×1 to 3×3 km3 pixels)
for urban regions and wide-swath mode (10×10 km2) for global
coverage.
n For 1–5 km binned data.
o From 45∘ S to 55∘ N.
p Integrating the signal along 50 km of the lidar orbit track.
q Lidar online and offline sampling at 1645.552 and 1645.846 nm, respectively.
r Point source imagers quantify emissions from individual point sources
by imaging of the atmospheric plume.
s Landsat-9 was launched in 2021 with a similar capability to
Landsat-8.
t For favorable (bright and spectrally homogeneous) surfaces.
u Including GHGSat-D (2016), GHGSat-C1 (2020), GHGSat-C2 (2021), and GHGSat-C3–GHGSat-C5 (2022).
Plans are for six more launches in 2023.
v For the constellation. Individual satellites have return times of
about 14 d.
w For the GHGSat-C satellites. GHGSat-D has a precision of 12 %–25 %.
x Other planned hyperspectral imaging spectrometers with observing
capabilities similar to PRISMA and EnMAP include SBG and CHIME (Cusworth et
al., 2019).
y EMIT is a surface mineral dust mapper that will fly on the
International Space Station in a 51.6∘ inclined orbit and will target
arid areas.
z Based on the precision of PRISMA (Guanter et al., 2021) and the
higher spectral resolution of EMIT (Cusworth et al., 2019).
aa Carbon Mapper is expected to be a constellation of satellites with
two launches in 2023 and a goal of six launches in 2024.
ab Carbon Mapper push-broom mode has imaging strips as long as 1000 km
with 30×60 m2 pixels. Carbon Mapper target-tracking mode has
shorter imaging strips with 30×30 m2 pixels and ground-motion
compensation to achieve higher signal-to-noise ratio (lower detection
threshold).
Attributes and data availability for satellite instruments
observing atmospheric methanea.
InstrumentAttributesData availabilitybArea flux mappersGOSATLong-term record of high-quality dataL2, openTROPOMIGlobal continuous daily coverageL2, openGOSAT-GWHigh-resolution mapping of urban areasL2, openMethaneSATHigh-resolution mapping of oil/gas/agricultural source regions with imaging of large point sourcesL1, L2, and L4, opencSentinel-5Global continuous daily coverage including the 1.65 µm bandL2, openGeoCarbContinuous coverage for methane-CO2-CO over North and South America with subdaily observationsL2, openCO2MHigh-resolution global continuous coverageL2, openMERLINArctic and nighttime observationsL2Point source imagersSentinel-2, LandsatGlobal continuous data acquisition, long-term recordsL1, openWorldView-3Very high spatial resolutionL1, for purchaseGHGSatHigh sensitivity (∼100 kg h-1), established constellationL2 and L4, for purchasedPRISMA, EnMAPMedium sensitivity (100–1000 kg h-1), extensive coverageL1, free on requestEMITMedium sensitivity (100–1000 kg h-1), extensive coverage of low-latitude arid regionsL1, openeCarbon MapperHigh sensitivity (∼100 kg h-1), high observing system completenessL2 and L4, openf
a See Table 1 for the specifications of each instrument. Instruments not yet launched are in italics.
b L1 (Level 1) indicates measured radiances, L2 indicates retrieved
column dry mixing ratio XCH4, L4 indicates derived emission rates.
c L1 and L2 data will be made available upon request.
d Data may also be obtained from space agencies through agreements
negotiated with GHGSat.
e Generation of an L2 product is under discussion.
f L1 data will be available for purchase.
Satellite instruments for observation of methane in the shortwave
infrared (SWIR). Area flux mappers are designed to quantify total methane
emissions on regional to global scales. Point source imagers are designed to
quantify emissions from individual point sources by imaging the atmospheric
plumes. Specifications for each instrument are in Tables 1 and 2. Satellite
icons were obtained from https://www.gosat.nies.go.jp (last access: 22 July 2022) for
GOSAT; Wikipedia Commons for TROPOMI, EMIT (International Space Station), and
Sentinel-2; https://space.skyrocket.de (last access: 22 July 2022) for GOSAT-GW, MERLIN,
CO2M, and Carbon Mapper; https://www.methanesat.org (last access: 22 July 2022) for MethaneSAT; ESA (2020) for Sentinel-5;
https://www.ou.edu/geocarb/mission (last access: 22 July 2022) for GeoCarb;
https://www.planetek.it/ (last access: 22 July 2022) for PRISMA; https://www.ghgsat.com/ (last access: 22 July 2022)
for GHGSat; https://www.enmap.org/mission (last access: 22 July 2022) for EnMAP;
https://directory.eoportal.org (last access: 22 July 2022) for WorldView-3; and
https://www.usgs.gov/landsat-missions (last access: 22 July 2022) for Landsat.
All instruments in Table 1 except MERLIN observe methane by SWIR solar
backscatter from the Earth's surface, either at 1.63–1.70 µm (1.65 µm band) or at 2.2–2.4 µm (2.3 µm band). Atmospheric
scattering is weak in the SWIR except for clouds and large aerosol
particles. Under clear skies, methane is observed down to the surface with
near unit sensitivity (Worden et al., 2015). The retrieval may fail if the
surface is too dark, such as over water or forest canopies (Ayasse et al., 2018).
Observations over water can be made by sunglint when the sun–satellite
viewing geometry is favorable. The MERLIN lidar instrument emits its own
1.65 µm radiation and detects the reflected signal. It can observe
over water and at night, but its sensitivity and coverage are lower than for
the solar backscatter instruments. Lidar capability to observe methane from
space is currently limited by laser technology (Riris et al., 2019).
Not included in Table 1 are instruments that measure methane in the thermal
infrared (TIR) or by solar occultation. These instruments are not sensitive
to methane near the surface and are therefore not directly useful for
quantifying methane emissions. TIR instruments have been used for remote
sensing of methane plumes from aircraft (Hulley et al., 2016), but
measurements from satellites mainly sense the upper tropospheric background
(Worden et al., 2015). Solar occultation instruments such as ACE-FTS provide
sensitive measurements of stratospheric methane profiles (Koo et al., 2017)
but cloud interference prevents observations in the troposphere. TIR and
solar occultation instruments can complement SWIR data by providing
information on background methane in the upper troposphere and stratosphere
(Zhang et al., 2021; Tu et al., 2022).
The spectrally resolved SWIR backscattered solar radiation detected by
satellite under clear-sky conditions can be used to retrieve the total
atmospheric column of methane, ΩCH4 [molecules cm-2], as
will be reviewed in Sect. 2.2. To remove the variability from surface
pressure, measurements are typically reported as dry column mixing ratio
XCH4=ΩCH4/Ωa,d, where Ωa,d is the dry air column [molecules cm-2]. Normalizing to dry air rather
than total air avoids introducing dependence on water vapor.
All instruments in Table 1 except EMIT and GeoCarb are in low-elevation
polar sun-synchronous orbit and observe globally at specific local times of
day, either morning or early afternoon. Morning has greater probability of
clear sky, while early afternoon has steadier boundary layer winds for
interpreting methane enhancements. GOSAT (2009–present) and its follow-on
GOSAT-2 (2018–present) provide global coverage every 3 d for 10 km
circular pixels spaced about 270 km apart, while TROPOMI (2018–present)
provides full global daily coverage with 5.5×7 km2 pixels.
Figure 2 shows mean TROPOMI XCH4 data for two different seasons,
illustrating the dense coverage. Future instruments GOSAT-GW (2023 launch,
10×10 km2 pixels with full global coverage every 3 d in
wide-swath mode), Sentinel-5 (2024 launch, 7.5×7.5 km2 pixels
with full global daily coverage), and CO2M (2025 launch, 2×2 km2 pixels with full global coverage every 5 d) will continue the
global observation record. MERLIN will provide day and night global coverage
along its lidar orbit track. Sentinel-2 and Landsat instruments provide full
global coverage with 20–30 m pixels every 5 d (Sentinel-2) or 16 d
(Landsat) and can detect very large point sources over bright spectrally
homogeneous surfaces. EMIT (designed to observe arid surfaces for dust
generation) will be on a 51.6∘ inclined orbit aboard the International
Space Station with variable local overpass times. GeoCarb will be in
geostationary orbit over the Americas and will provide subdaily observations
from 45∘ S to 55∘ N.
Global TROPOMI observations of methane for December 2019–February 2020 and June–August 2020. Data are from the version 2.02 product,
filtering out low-quality retrievals (qa_value <0.5)
and snow and ice surfaces diagnosed by blended albedo >0.8 (Lorente
et al., 2021a). The top panels show the mean dry methane column mixing ratios
XCH4 on a 0.1∘×0.1∘ grid. The middle panels show the
observation density as the number of successful observations per
1∘×1∘ grid cell for the 3-month periods. The bottom
panels show the mean XCH4 differences between colocated TROPOMI and
GOSAT observations plotted on a 2∘×2.5∘ grid and
adjusted upward by 10.5 ppb to account for TROPOMI being 10.5 ppb lower than
GOSAT in the global mean. GOSAT data are from the CO2 proxy retrieval
version 9.0 of Parker et al. (2020).
Several narrow-swath instruments in Table 1 are selective in their
observations to focus on specific targets and avoid cloudy conditions. The
GHGSat instruments observe selected 12×12 km2 scenes with
25×25 m2 pixel resolution and instrument pointing to increase
the signal-to-noise ratio (SNR). Carbon Mapper will observe 18 km swaths
with imaging strips as long as 1000 km in push-broom mode and shorter strips
in target-track (instrument-pointing) mode. GHGSat has six satellites in
orbit as of this writing to achieve frequent return times, and Carbon Mapper
similarly plans a constellation of satellites. WorldView-3 observes scenes
of dimensions up to 66.5×112 km2. MethaneSAT will observe
200×200 km2 targets in oil and gas systems and agricultural regions with
130×400 m2 pixel resolution, enabling high-resolution
quantification of regional emissions as well as imaging of large point
sources.
All area flux mappers in Table 1 have fine (<0.5 nm) spectral
resolution to enable precise measurements of methane concentrations, traded
against coarser (0.1–10 km) spatial resolution. GHGSat achieves a
combination of fine spatial resolution and fine spectral resolution by
instrument pointing. Most other point source imagers in Table 1 are designed
to observe land surfaces, which requires fine spatial resolution (<50 m) but less stringent spectral resolution. These instruments have
serendipitous capability to detect methane plumes in the broad 2.3 µm band, including hyperspectral sensors with ∼10 nm spectral
resolution (PRISMA, EnMAP, EMIT; Cusworth et al., 2019) and even
multispectral sensors with a single 2.3 µm channel (Sentinel-2,
Landsat; Varon et al., 2021) or a few channels (WorldView-3;
Sanchez-Garcia et al., 2022). Carbon Mapper will have 6 nm spectral
resolution, which increases precision appreciably relative to 10 nm (Cusworth et al., 2019).
All area flux mappers in Table 1 have an open data policy allowing free
access from a distribution website or from the cloud. The data are generally
provided as XCH4 retrievals (Level 2 or L2). MethaneSAT will
distribute its data publicly as inferred methane fluxes (L4), with the L1 and
L2 data also available upon request. Data access for point source imagers is
presently less straightforward. Sentinel-2 and Landsat have freely
accessible channel radiance (L1) data, but users must perform their own
methane retrievals and source rate estimates. GHGSat and WorldView-3 make
observations at the request of paying customers, with GHGSat providing
column density (L2) and source rate (L4) data and WorldView-3 providing L1
data. PRISMA and EnMAP make observations upon request from the scientific
community and stakeholders, and the resulting L1 data are then freely
accessible, but again users must perform their own methane retrievals.
Carbon Mapper will provide open L2 and L4 data.
Retrieval methods
The “full-physics” retrieval of methane columns from satellite SWIR spectra
involves inversion of the spectra with a radiative transfer model (Butz et
al., 2012; Thorpe et al., 2017). It typically solves simultaneously for the
vertical profile of methane concentration, the vertical profile of aerosol
extinction, and the surface reflectivity. Although the vertical profile of
methane may be retrieved in the inversion, there is actually no significant
information on vertical gradients, and only XCH4 is reported together with
an averaging kernel vector for sensitivity to the vertical profile (near
unity in the troposphere). The retrieval may fail if the atmosphere is hazy
or if the surface is heterogeneous or too dark. Full-physics TROPOMI
retrievals in the 2.3 µm band thus have only a 3 % global success
rate over land (Lorente et al., 2021a) with large variability depending on
location (Fig. 2). Arid areas and midlatitudes are relatively well
observed. Observations are much sparser in the wet tropics because of
extensive cloudiness and dark surfaces and in the Arctic because of
seasonal darkness, extensive cloudiness, and low sun angles. Observations at
high latitudes are very limited outside of summer, resulting in a seasonal
sampling bias.
The 1.65 µm band allows the alternative CO2 proxy retrieval,
taking advantage of the adjacent CO2 absorption band at 1.61 µm
(Frankenberg et al., 2005). In this method, ΩCH4 and ΩCO2 are retrieved simultaneously without accounting for atmospheric
scattering, and XCH4 is then derived as
XCH4=ΩCH4ΩCO2XCO2,
where XCO2 is independently specified, typically from assimilated
observations or from a global chemical transport model (Parker et al., 2020;
Palmer et al., 2021). The CO2 proxy method takes advantage of the lower
variability of CO2 than methane and of the low CO2 co-emission
from the dominant methane sources (livestock, oil and gas systems, coal mining,
landfills, wastewater treatment, rice cultivation, wetlands). It is much
faster than the full-physics retrieval, achieves similar precision and
accuracy (Buchwitz et al., 2015), and largely avoids biases associated with
surface reflectivity and aerosols because these biases tend to cancel in the
ΩCH4/ΩCO2 ratio. It is subject to errors from
unresolved variability of CO2 such as in urban regions and is also
subject to bias for sources that co-emit methane and CO2 such as
flaring and other incomplete combustion. The GOSAT instrument operating at
1.65 µm with 10 km pixels has a 24 % success rate over land using
the CO2 proxy retrieval, mainly limited by cloud cover (Parker et al.,
2020).
A limitation in using the 1.65 µm band is that it is narrower, with
fewer spectral features and weaker absorption than the 2.3 µm band,
and it therefore requires an instrument with sub-nanometer spectral resolution
(Cusworth et al., 2019; Jongaramrungruang et al., 2021). The 2.3 µm
band can be successfully sampled for a full-physics retrieval by
hyperspectral instruments with ∼10 nm spectral resolution
(Thorpe et al., 2014, 2017; Cusworth et al., 2021a; Borchardt et al., 2021;
Irakulis-Loitxate et al., 2021). Precision improves with spectral resolution
(Cusworth et al., 2019; Jongaramrungruang et al., 2021) and with spectral
positioning relative to the methane absorption lines (Scaffuto et al.,
2021). Multispectral instruments with one or several broadband channels
(∼100 nm bandwidth) do not allow a spectrally resolved
retrieval, but a simple Beer's law retrieval of the methane column
enhancement in a plume relative to background can still be achieved in the
2.3 µm band by inferring surface reflectivity from adjacent bands or
from views of the same scene when the plume is absent (Varon et al., 2021;
Sanchez-Garcia et al., 2022).
Yet another approach for retrieving methane enhancements from point sources
is the matched-filter method in which the observed spectrum is fitted to a
background spectrum convolved with a target methane absorption spectrum
capturing the 2.3 µm absorption band (Thompson et al., 2015; Foote et
al., 2020). Matched filter methods have been extensively used for mapping
methane point sources from airborne hyperspectral campaigns (Frankenberg et
al., 2016; Duren et al., 2019; Cusworth et al., 2021b) and have also been
used for satellite retrieval of point sources (Thompson et al., 2016;
Guanter et al., 2021; Irakulis-Loitxate et al., 2021). These methods
directly retrieve the methane enhancement above background and are faster
than a full-physics retrieval. They are well-suited methods for plume imaging, where
the methane enhancement above local background is the quantity of interest.
Precision and accuracy
Retrievals of XCH4 may be affected by random error (precision) and
systematic error (bias or accuracy). A uniform bias is inconsequential
because it can be simply subtracted. Random error is reducible by temporal
averaging if the observation density is high. The most pernicious error is
spatially variable bias, often called relative bias (Buchwitz et al., 2015),
which is generally caused by aliasing of surface reflectivity spectral
features into the methane retrieval. Variable bias corrupts the retrieved
concentration gradients and produces artifact features that may be wrongly
attributed to methane.
Area flux mapper instruments are generally validated by reference to the
highly accurate XCH4 measurements from the worldwide Total Carbon Column
Observing Network (TCCON) of ground-based sun-staring spectrometers (Wunch
et al., 2011). Variable bias can be estimated as the spatial standard
deviation across TCCON sites of the temporal mean bias (Buchwitz et al.,
2015). Schneising et al. (2019) inferred in this manner a global bias of
-1.3 ppb for the TROPOMI University of Bremen methane retrieval, a precision
of 14 ppb, and a variable bias of 4.3 ppb. Lorente et al. (2021a) inferred a
global mean bias of -3.4 ppb and a variable bias of 5.6 ppb for the current
TROPOMI version 2 Netherlands Institute for Space Research (SRON)
operational retrieval. Figure 3 places these values in the context of
TROPOMI observations over the Permian Basin oil field in Texas and New
Mexico. A typical single day of TROPOMI observations shows large areas of
missing and noisy data, and thus temporal averaging is necessary, which also
reduces the random error. Averaging TROPOMI observations over a month shows
full coverage of the Permian with enhancements of ∼50 ppb
over the principal areas of oil and gas production, well above the variable
bias of the instrument.
Reliance on the TCCON network to diagnose variable bias is limited by the
sparsity of network sites, almost all of which are at northern midlatitudes. An
alternative way is by reference to GOSAT. The current version 9 GOSAT
retrieval using the CO2 proxy method has a variable bias of only 2.9 ppb referenced to TCCON and is recognized as a well-calibrated measurement
(Parker et al., 2020). Spatial variability in the mean TROPOMI–GOSAT
difference provides a global assessment of TROPOMI variable bias (Qu et al.,
2021). Results in Fig. 2 (bottom panel), after correcting for a global mean
TROPOMI–GOSAT difference of -10.5 ppb (TROPOMI lower than GOSAT), show that
TROPOMI variable biases can exceed 20 ppb in some regions. The reason for
such large biases relative to GOSAT is TROPOMI's coarser spectral sampling
of the SWIR region, as well as the unavailability of the CO2 proxy
retrieval at 2.3 µm. Comparing TROPOMI and GOSAT observations for a
region of interest is good practice before interpreting TROPOMI data for
that region (Z. Chen et al., 2022).
Variable bias is also a concern for point source imagers, where it manifests
as artifact features that could be mistaken for methane plumes (Ayasse et
al., 2018). This is of particular concern for heterogeneous surfaces
(Cusworth et al., 2019). Artifacts can be screened by visual inspection of
the candidate plumes in relation to wind direction, known infrastructure,
and surface reflectivity (Guanter et al., 2021). Machine-learning methods
can also be trained to detect plumes and recognize artifact noise patterns
(Jongaramrungruang et al., 2022). Figure 3 shows illustrative observations
of point sources from Sentinel-2, PRISMA, and GHGSat in the Permian Basin.
The observations have lower precision than TROPOMI (Table 1), but the methane
enhancements are much larger because the pixels are smaller. Point source
detection thresholds and their relationship to precision are discussed in
Sect. 5.
Global, regional, and point source observations
Figure 4 classifies the satellite instruments of Table 1 in terms of their
abilities to observe methane on global and regional scales as area sources
(area flux mappers) or on the scale of individual point sources (point
source imagers). Observations of these different scales target complementary
needs for our understanding of methane, and they correspondingly have
different observing requirements. Area sources may integrate a very large
number of individually small emitters that cumulate to a large total, such
as low-production oil wells (Omara et al., 2022). A practical definition of
a methane point source for our purposes, following Duren et al. (2019), is a
single facility emitting more than 10 kg h-1 over an area less than
30×30 m2. This represents a typical limit of detection from
aircraft remote sensing combined with a typical spatial resolution for point
source imagers. With this definition of source threshold, Cusworth et al. (2022) find on average that 40 % of emissions from US oil and gas fields
originate from point sources. This emphasizes the need for characterizing
methane emissions complementarily both as area sources and as point sources.
Classification of satellite instruments by their capability to
observe atmospheric methane on global scales, on regional scales with high
resolution, and for point sources. Specifications for the satellite
instruments are listed in Table 1, and key attributes are listed in Table 2.
Point source detection thresholds are given here as orders of magnitude.
These detection thresholds are discussed in Sect. 5.2. Instruments not yet
launched are in italics.
Global and regional observations with area flux mappers
Global observation of methane targets the central question of why
atmospheric methane has almost tripled since pre-industrial times and why it
continues to increase. Ground network measurements such as from NOAA are the
reference for observing global trends because of their high accuracy
(Bruhwiler et al., 2021), and some sites include isotopic or other
information to separate contributions from different source sectors (Lan et
al., 2021). But satellites have an essential role to play because of their
dense and global coverage. They can identify the regions that drive the
global trend (Zhang et al., 2021). They have a unique capability to evaluate
the accuracy and trends of methane emissions reported by individual
countries to the UNFCCC (Janardanan et al., 2020) and thus contribute to the
transparency framework of the Paris Agreement (Deng et al., 2022; Worden et
al., 2022).
Global observation of methane from space is presently available from GOSAT
and TROPOMI. GOSAT provides a continuous and well-calibrated record going
back to 2009 (Parker et al. 2020). Inversions of GOSAT data have been used
to attribute the contributions of different source regions and sectors to
the methane increase over the past decade (Maasakkers et al., 2019; Chandra
et al., 2021; Palmer et al., 2021; Zhang et al., 2021). The TROPOMI data
stream begins in May 2018 and is much denser than GOSAT, but the ability to
use TROPOMI data in global inversions is presently limited by large variable
biases in some regions of the world (Qu et al., 2021; Fig. 2). This is
likely to improve with future retrieval versions and may be overcome with
careful data selection. Continuity of global methane observations from space
is expected over the next decade with the GOSAT series (GOSAT-2, GOSAT-GW),
Sentinel-5, and CO2M (Table 1). MERLIN could make an important contribution
toward better understanding of methane emissions in the Arctic, which is
otherwise difficult to observe from space.
There is considerable interest in using satellite observations to quantify
methane emissions with high resolution on regional scales. This is important
for reporting of emissions at the national or sub-national state level, for
monitoring oil and gas production basins, and for separating contributions from
different source sectors. Oil and gas production basins are typically a few
hundred kilometers in size and may contain thousands of point sources that are
individually small but add up to large totals and are best quantified on a
regional scale (Lyon et al., 2015). Several field campaigns using surface
and aircraft measurements have targeted oil and gas fields in North America
(Karion et al., 2015; Pétron et al., 2020; Lyon et al., 2021), but these
campaigns are necessarily short and are not practical in many parts of the
world.
TROPOMI with its 5.5×7 km2 pixel resolution and global
continuous daily coverage is presently the only satellite instrument capable
of high-resolution regional mapping of methane emissions. GOSAT data are too
sparse. TROPOMI has been used to quantify emissions from oil and gas production
fields including the Permian Basin (Zhang et al., 2020), other fields in the
US and Canada (Shen et al., 2022), and the Mexican Sureste Basin (Shen et
al., 2021), revealing large underestimates in the bottom-up inventories. It
has also been used to quantify total methane emissions from China and
to attribute them to source sectors (Z. Chen et al., 2022). The variable bias
problems that affect global TROPOMI inversions can be less problematic on
the scale of source regions where methane enhancements are large, the bias
may be less severe (Fig. 2), and bias correction is possible through
adjustment of boundary conditions in the transport model (Shen et al.,
2021). Capability for regional mapping of methane emissions is expected to
greatly expand in the future with the MethaneSAT, GOSAT-GW, Sentinel-5, and
CO2M instruments.
Point source observations with point source imagers
Monitoring large point sources is important for reporting of emissions, and
detection of unexpectedly large point sources (super-emitters) can enable
prompt corrective action. In situ sampling and remote sensing from aircraft
has been used extensively to quantify point sources (Frankenberg et al.,
2016; Lyon et al., 2016; Duren et al., 2019; Hajny et al., 2019; Y. Chen et
al., 2022; Cusworth et al., 2022) but is limited in spatial and temporal
coverage. Satellites again have an essential role to play. They have enabled
the discovery of previously unknown releases (Varon et al., 2019; Lauvaux et
al., 2022) and the quantification of time-integrated total emissions from
gas well blowouts (Cusworth et al., 2021a; Maasakkers et al., 2022a).
Observing point sources from space has unique requirements. Plumes are
typically less than 1 km in size (Frankenberg et al., 2016; Fig. 3), thus
requiring satellite pixels finer than 60 m (Ayasse et al., 2019). It is
desirable to quantify emissions from single overpasses, though temporal
averaging of plumes to improve SNR is possible with wind rotation if the
precise location of the source is known (Varon et al., 2020). The emissions
are temporally variable, motivating frequent revisit times that can be
achieved by a constellation of instruments. On the other hand, precision
requirements are less stringent than for regional or global observations
because of the larger magnitude of the concentration enhancements.
The potential for space-based land imaging spectrometers to detect methane
point sources was first demonstrated with the hyperspectral Hyperion
instrument for the Aliso Canyon blowout (Thompson et al., 2016).
Hyperspectral sensors such as PRISMA and others of similar design have since
proven capable of quantifying point sources of ∼500 kg h-1 (Cusworth et al., 2021a; Guanter et al., 2021; Irakulis-Loitxate et
al., 2021; Nesme et al., 2021). The first satellite instrument dedicated to
quantifying methane point sources was the GHGSat-D demonstration instrument
launched in 2016 with 50×50 m2 effective pixel resolution and
a precision of 12 %–25 % depending on surface type (Jervis et al., 2021).
Varon et al. (2019) demonstrated the capability of that instrument for
discovering and quantifying persistent point sources in the range 4000–40 000 kg h-1 in an oil and gas field in Turkmenistan. Five follow-up GHGSat
instruments with precisions of 1 %–2 % were subsequently launched in
2020–2022, building up to a constellation with frequent return times.
Multispectral instruments such as Sentinel-2, Landsat, and WorldView-3 are
also capable of detecting and quantifying very large point sources (Varon et
al., 2021; Ehret et al., 2022; Sanchez-Garcia et al., 2022;
Irakulis-Loitxate et al., 2022a). Sentinel-2 and Landsat provide global and
freely accessible data that could form the foundation of a global detection
system for super-emitters (Ehret et al., 2022). A large-scale survey of
point emissions across the western coast of Turkmenistan was achieved with the
combination of Sentinel-2 and Landsat (Irakulis-Loitxate et al., 2022a).
Detection of methane plumes from space has mainly been over bright land
surfaces. Observation of offshore plumes such as from oil and gas extraction
platforms is more difficult because of the low reflectance of water in the
SWIR. The signal can be enhanced by observing in the sunglint mode, in which
the sensor captures the solar radiation specularly reflected by the water.
The sunglint observation configuration can be achieved by agile platforms
able to point in the Sun-surface forward scattering direction (PRISMA,
Worldview-3, GHGSat, Carbon Mapper) or by instruments with a field-of-view
sufficiently large that part of the swath falls in the forward scattering
area (TROPOMI, Sentinel-2, Landsat). Irakulis-Loitxate et al. (2022b)
demonstrated the ability of sunglint retrievals from WorldView-3 and
Landsat-8 to detect large plumes from offshore platforms in the Gulf of
Mexico.
The capability to monitor methane point sources from space is expected to
expand rapidly in coming years through the GHGSat and Carbon Mapper
constellations as well as new hyperspectral missions (Cusworth et al.,
2019). Expanding constellations observing with frequent return times and at
different times of day will enable better understanding of the intermittency
of methane emissions. In an aircraft survey of the Permian Basin, Cusworth
et al. (2021b) found that individual point sources produced detectable
emissions only 26 % of the time on average. Similar intermittency was
observed for oil and gas facilities in California (Duren et al., 2019). Allen et
al. (2017) and Vaughn et al. (2018) point out that some emissions from the
oil and gas infrastructure are highly intermittent by design (liquids unloading,
blowdowns, and startups) and may have predictable diurnal variations.
Emissions due to equipment failure may be persistent (leaks, unlit flares),
sporadic (responding to gas pressure), or single events (accidents). An
increased frequency of observation can identify persistence of emissions to
enable corrective action, and better understanding of point sources that are
intermittent by design can lead to better quantification of time-averaged
emissions. Beyond this short-term intermittency, there is also long-term
variability related to operating practices and facility life cycle
(Cardoso-Saldaña and Allen, 2020; Johnson and Heltzel, 2021; Varon et
al., 2021; Allen et al., 2022; Ehret et al., 2022), stressing the importance
of sustained long-term monitoring.
Inferring methane emissions from satellite observations
Inferring methane emissions from satellite observations of methane columns
involves different methods for area flux mappers and point source imagers.
Area flux mappers are typically used to optimize 2-D distributions of
emissions on regional or global scales by inverse methods. Point source
imagers are used to infer individual point source rates by some form of mass
balance analysis.
Global and regional inversions with area flux mappers
Area flux mappers produce 2-D fields of methane observations from which to
optimize 2-D fields of gridded emission fluxes. The optimization involves an
atmospheric transport model (forward model) to relate emissions to the
observed concentrations. The optimal emissions are generally obtained by
Bayesian inference, fitting the observations to the forward model and
including prior estimates of emissions to regularize the solution where the
observations provide insufficient information (Brasseur and Jacob, 2017).
Optimizing temporal trends of emissions can be done as part of the solution
or sequentially using a Kalman filter (Feng et al., 2017).
The basic procedure is as follows. Given an ensemble of observations over a
domain of interest assembled in an observation vector y, the task is to
optimize the distribution of emission fluxes assembled in a state vector x of
dimension n. The relationship between x and y can be assumed linear for methane,
despite the sensitivity of OH concentrations to methane concentrations. This
is because the inversion does not significantly change the global methane
concentration, which is set by observation; furthermore, for regional
inversions, the timescale for ventilation of the regional domain is much
shorter than that for chemical loss. Global inversions often optimize OH
concentrations as part of the state vector and that relationship can also be
assumed linear. Further assuming Gaussian error probability density
functions (pdfs) for x and y, the optimal (posterior) estimate of x is obtained
by minimizing a Bayesian cost function J(x) of the form (Brasseur and Jacob,
2017):
J(x)=(x-xA)TSA-1(x-xA)+γ(y-Kx)TSO-1(y-Kx),
Here xA is the prior estimate of emissions, SA is the
corresponding prior error covariance matrix, K=∂y/∂x is the Jacobian matrix describing the sensitivity
of observations to emissions as given by the atmospheric transport model,
SO is the observational error covariance matrix including contributions
from instrument and transport model errors, and γ is a
regularization parameter that may be needed to correct overfit caused by
imperfect definition of SO (Lu et al., 2021). Since the relationship
between x and y is linear, K fully defines the atmospheric transport model for
the inversion. Jacob et al. (2016) describe alternative formulations for the
cost function such as in geostatistical inverse modeling where prior
information is provided as the relative spatial distribution of emissions
rather than emission magnitudes (Miller et al., 2020).
Specification of the error covariance matrices SA and SO strongly
affects the solution. Construction of SA can be done by intercomparing
bottom-up inventories (Maasakkers et al., 2016; Bloom et al., 2017) or by
using error estimates generated by the bottom-up inventories (Scarpelli et
al., 2020). Construction of SO can be done by the residual error method
in which the observations are compared to simulated concentrations from the
atmospheric transport model with prior emission estimates, and the residual
difference after removing the mean bias is taken to be the observational
error (Heald et al., 2004; Wecht et al., 2014). The observational error for
satellites is generally found to be dominated by the instrument retrieval
error rather than by the transport model error, whereas for in situ
observations it is dominated by the transport model error (Lu et al., 2021).
Minimization of the cost function J(x) in Eq. (2) to obtain the posterior
solution x^ and its error covariance
matrix S^ can be done either numerically or analytically
(Brasseur and Jacob, 2017). S^ and the related averaging
kernel matrix A=∂x^/∂x=In-S^SA-1 (Rodgers, 2000) determine the information content
from the observations and the ability of the inversion to improve on the
prior estimate. The diagonal terms of A ranging from 0 to 1 are called the
averaging kernel sensitivities and measure the ability of the observations
to constrain the solution for that state vector element independently of the
prior estimate (1= fully, 0= not at all). The trace of A is called the
degrees of freedom for signal (DOFS) and represents the total number of
pieces of information that can be fully constrained from the observations.
An inherent assumption is that the observations, the transport model, and
the prior information are unbiased. Although the prior estimate is in
principle unbiased since it represents our best estimate before the
observations are taken, under-accounting of SA together with incorrect
spatial distribution of prior emissions can drive bias in inversion results
(Yu et al., 2021).
Numerical solution for min(J(x)) using the adjoint of the atmospheric
transport model or other variational methods optimizes a state vector of any
dimension by avoiding explicit construction of the full Jacobian matrix K
and may use various procedures to estimate S^ (Bousserez et
al., 2015; Cho et al., 2022). Analytical solution provides a closed-form
expression for S^ but requires the computationally expensive
construction of K column-by-column with n perturbation runs of the
atmospheric transport model. This limits the dimension and hence the
resolution of the state vector that can be optimized. However, once K has
been constructed, inversion ensembles can be conducted at no significant
added computational cost to explore uncertainties in inversion parameters
or to examine the complementarity and consistency of different observation
subsets such as from different satellite instruments or from ground-based
sites (Lu et al., 2021, 2022). This includes optimization of the
regularization parameter γ so that the sum of prior terms in the
posterior cost function matches the expected value from the chi-square
distribution, JA(x^)=(x^-xA)TSA-1(x^-xA)∼n (Lu et al., 2021).
Increasing access to large computational clusters has facilitated the
construction of K as an embarrassingly parallel problem, enabling analytical
solution for state vectors with n>1000 (Maasakkers et al., 2019).
Nesser et al. (2021) show that even larger dimensions can be accessed by
approximating the Jacobian along leading patterns of information content.
Figure 5 illustrates the inversion of TROPOMI observations with a 1-month
example for the Permian Basin using an analytical solution with
0.25∘×0.3125∘ (≈25×25 km2)
resolution. This calculation was done on the Amazon Web Services (AWS) cloud
with the Integrated Methane Inversion (IMI) open-access facility for
analytical inversions of TROPOMI data, enabling users to directly access the
TROPOMI data archived on AWS and infer emissions for their selected domain
and time window of interest with pre-compiled inversion code (Varon et al.,
2022).
Integrated Methane Inversion (IMI) on the Amazon Web Services
(AWS) cloud (Varon et al., 2022). The IMI accesses the TROPOMI operational
data posted on the cloud and carries out analytical inversions for
user-selected domains and time periods. Before conducting the inversion,
users can run an IMI preview to visualize the observations, the default
prior emission estimates (to which they can substitute their own), the
expected information content of the inversion (degrees of freedom for signal
or DOFS), and the SWIR albedos for indication of data artifacts. If the
preview is satisfactory, they can then run the inversion to generate
posterior emission estimates with averaging kernel sensitivities indicating
where the observations can successfully constrain emissions. Shown here is
an example given by Varon et al. (2022) for a 1-month (May 2018) inversion
over the Permian Basin, using the prior emission estimate from the EDF
inventory (Zhang et al., 2020). The IMI is accessible at
https://imi.seas.harvard.edu (last access: 23 July 2022).
The assumption of Gaussian error pdfs for prior emission estimates in Eq. (2) may not always be appropriate. A log-normal distribution is often more
correct (Yuan et al., 2015) and can be accommodated in analytical inversions
(Maasakkers et al., 2019; Z. Chen et al., 2022). Brandt et al. (2016) show
that the log-normal distribution still underestimates the heavy tail of the
frequency distribution of point sources (the super-emitters). Application of
inverse methods to detect and quantify individual super-emitters within a
source region (such as an oil and gas field) may require a bimodal pdf for prior
estimates, and an L1 norm cost function may be better suited than the
standard L2 norm of Eq. (2) (Cusworth et al., 2018). A Markov chain
Monte Carlo (MCMC) method for the inversion as used by Western et al. (2021)
enables the specification of any prior and observational error pdfs and
returns the full posterior error pdf on emissions, but it is computationally
expensive and its cost increases rapidly as n increases.
The inversion typically optimizes a geographical 2-D array of emission
fluxes, but quantifying emissions by source sector is often of ultimate
interest. Sectoral information is generally contained in the prior
inventory. The simplest approach is to assume that the posterior-to-prior
emissions ratio for a given grid cell applies equally to all
emissions in that grid cell (Turner et al., 2015) or in a manner weighted by
the prior uncertainties of the different sectors (Shen et al., 2021). The
posterior error covariance matrix S^ and averaging kernel
matrix A on the 2-D grid can similarly be mapped to specific sectors and/or be
summed over a domain such as an individual country (Maasakkers et al.,
2019). A more general approach for sectoral attribution introduced by
Cusworth et al. (2021c) maps the (x^,
S^) solution onto any alternative state vector z (such as
sector-resolved emissions) with its own prior information (zA, ZA)
to obtain a solution z^ with posterior
error covariance matrix Z^. This specifically allows for
comparisons of results from inversions using different prior information.
Quantification of point sources with point source imagers
Quantification of point sources from satellite observations of instantaneous
plumes poses a different kind of inversion problem. In this case, a single
quantity, the point source rate Q [kg s-1], is to be inferred from a
single observation of the plume. Figure 3 showed examples of plume
observations. The morphology of the instantaneous plume is determined by
turbulent diffusion superimposed on the mean wind, with a plume boundary
(commonly called the plume mask) defined by the detection limit of the
instrument. The observation is of the total methane column and so is
relatively insensitive to vertical boundary layer mixing, which is a major
source of error in interpreting plumes from in situ aircraft observations
(Angevine et al., 2020). On the other hand, unlike for in situ aircraft
observations, there is no direct measurement of the wind speed U in the plume.
The lack of precise wind speed information is a major source of error for
interpreting satellite observations because concentrations in the plume vary
as the ratio Q/U, meaning that errors in U propagate proportionally to errors in
Q.
Figure 6 summarizes different methods for inferring point source rates from
satellite observations of instantaneous plumes. Details on these methods are
given by Krings et al. (2011), Varon et al. (2018), and Jongaramrungruang et
al. (2019, 2022). The Gaussian plume is the classic model for turbulent
diffusion from a point source, but it is valid only for a plume sampling a
representative ensemble of turbulent eddies. Methane plumes are generally
too small for this condition to be met (Jongarangmrungruang et al., 2019),
as illustrated in Fig. 3 where the plume shapes are not Gaussian. A simple
mass balance method applying the local wind speed to the methane enhancement
observed in the plume is flawed for sub-kilometer scales because ventilation is
determined by turbulent eddies more than by the mean wind (Varon et al.,
2018).
Seven different methods for inferring point source rates Q [kg s-1] from satellite observations of instantaneous plumes of methane
column enhancements ΔΩ [kg m-2] relative to background.
The methods involve (1) fit to a Gaussian plume, (2) local mass balance for
near-source pixels, (3) Gauss theorem with integration of the outward flux
along a closed contour s, (4) cross-sectional flux (CSF) integral, (5) integrated mass enhancement (IME) with independent wind speed information,
(6) IME with wind speed inferred from the plume angular width θ, and
(7) machine learning applying a convolution neural network (CNN) to the
plume image. Methods (1), (2), (4), and (5) are described by Varon et al. (2018); method (3) is described by Krings et al. (2011); method (6) is described by Jongaramrungruang
et al. (2019); and method (7) is described by Jongaramrungruang et al. (2022). In the
equations, x denotes the plume axis for transport by the mean wind and y
denotes the horizontal axis normal to the wind. The IME [kg] is the spatial
integral of the methane column enhancement ΔΩ over the plume
mask. The wind speed U is that relevant to transport of the plume, and in the
IME method (4) it is parameterized as an effective wind speed Ueff to
include the effect of turbulent diffusion. The Gauss theorem and CSF methods
require wind direction information. The IME method (4) requires a
characteristic plume size L that can be taken as the square root of the plume
area (Varon et al., 2018) or the radial plume length (Duren et al., 2019).
The empirical dispersion parameter σy [m] in the Gaussian plume
method (1) characterizes the spread of the plume. n in the Gauss
theorem method is the unit vector normal to the contour.
The Gauss theorem method, in which the source rate is calculated as the
outward flux summed along a contour surrounding the point source, is
extensively used for in situ aircraft observations where concurrent
measurements of wind vector and methane concentration are available to
calculate the local flux as the aircraft circles around the source (Hainy et
al., 2019). In the absence of in situ wind data, one can apply a single
estimate of the wind vector based on local station or assimilated data
(Krings et al., 2011). However, the calculation then does not account for
the contribution of turbulent diffusion to the outward flux. In addition,
any sources within the contour will alias into the inferred point source
rate.
Two successful methods to derive point source rates from observations of
instantaneous plumes have been the cross-sectional flux (CSF) method (White,
1976; Krings et al., 2011), in which the source rate is inferred from the
product of the methane enhancement and the wind speed integrated across the
plume width, and the integrated mass enhancement (IME) method (Frankenberg
et al., 2016; Varon et al., 2018), in which the total mass enhancement in
the plume is related to the magnitude of emission with a parameterization
dependent on wind speed. Both methods are widely applied to the retrieval of
point source rates from satellite observations and they yield consistent
results (Varon et al., 2019). The CSF method is more physically based, and
source rates can be derived from cross sections at different distances
downwind to reduce error (Fig. 6). The contribution of turbulent diffusion
to the flux can be neglected in the direction of the wind following the
slender plume approximation (Seinfeld and Pandis, 2016). However, the
dependence on wind direction is an additional source of error relative to
the IME method.
Both the CSF and IME methods require estimates of wind speed relevant to
plume transport. For the CSF method this is the mean wind speed over the
vertical depth of the plume, which can be parameterized from the 10 m wind
speed (Varon et al., 2018) or interpolated from a database of wind speed
vertical profiles (Krings et al., 2011). The effective wind speed
Ueff in the IME method accounts for the effect of turbulent diffusion
in plume dissipation and can be parameterized as a function of an
observable 10 m wind speed by using large-eddy simulations (LES) of
synthetic plumes sampled with the instrument pixel resolution, plume mask
definition, and observing time of day (Varon et al., 2018). The need for
independent information on wind speed, either from measurements at the point
source location or from a meteorological database, can dominate the error
budget in inferring source rates from the CSF and IME methods, and typically
limits the precision to 30 % (Varon et al., 2018). The error is larger for
weak winds, which tend to be more variable and smaller for strong steady
winds. However, plumes are less likely to be detectable in strong winds
because of dilution. Weak winds are thus favorable for plume detection but
can induce large error in source quantification.
Jongaramrungruang et al. (2019) showed that the morphology of an observed
plume contains information on wind speed, as long slender plumes are
associated with high wind speeds while short stubby plumes are associated
with low wind speeds. By using the plume angular width as a measure of wind
speed, they were able to infer source rates without independent wind
information. Jongaramrungruang et al. (2022) developed that idea further
with a convolutional neural network (CNN) approach trained on LES plume
images to learn the source rate from the 2-D plume structure. Application to
synthetic plumes, as would be sampled by the AVIRIS-NG aircraft instrument at
1–5 m pixel resolution, showed a mean precision of 17 % and a detection
threshold of 50 kg h-1 over spectrally homogeneous surfaces. This
method has not yet been applied to satellite observations where coarser
pixels would result in lower sensitivity and where retrievals are more
subject to artifacts.
Detection thresholdsArea sources
Here we examine the ability of area flux mappers to detect total methane
emission fluxes from a target domain with a desired spatial resolution. This
can involve repeated observations of the domain over multiple passes to
increase precision and observation density, as illustrated in Fig. 3. The
observation time required to detect a desired flux threshold at a desired
spatial resolution then depends on the instrument precision, the spatial
coverage, the fraction of successful retrievals, the pixel size, the
variability of emissions, and the return time.
Following the conceptual model of Jacob et al. (2016), the methane column
enhancement ΔX [ppb] resulting from a uniform emission flux E [kg km-2 h-1] over a square domain of dimension W [km] is given by
ΔX=αEW,
with a scaling coefficient α=(Ma/MCH4)g/pU where Ma and
MCH4 are the molecular weights of dry air and methane, g is the
acceleration of gravity, p is the surface pressure, and U is the wind speed for
ventilation of the domain. With the units above and assuming p=1000 hPa
and U=5 km h-1, we have α=4.0×10-2 ppb km h kg-1. An instrument with pixel-level precision σI [ppb]
can detect this emission flux with a single measurement if ΔX>>σI, but this is often not the case.
Spatial and temporal averaging of observations improves the effective
precision, and this improvement goes as the square root of the number of
observations if the error is random, uncorrelated, and representatively
sampled (IID conditions). The time required for detecting the mean emission
flux E over a domain of dimension W with a signal-to-noise ratio of 2 is then
given by
t=tRmax1,1FNmax1,2σΔX2,
where tR is the return time of the instrument (time interval between
successive passes), N is the number of observations within the domain per
individual pass for instrument pixel sizes D smaller than W (for continuous
mapping and square pixels, we have N=(W/D)2), F is the fraction of
successful retrievals, and σ [ppb] is the variability that results
from both the instrument precision and the spatial variability σX (D,W) of the enhancement ΔX sampled by the pixels within the
domain:
σ=σI2+σX(D,W)2.
Equations (3)–(5) provide a simple conceptual framework for evaluating the
ability of area flux mappers to detect regional emissions of a certain
magnitude and with a desired spatial resolution. For illustration purposes,
consider an objective to detect emissions at either 100 or 10 km
resolution. In the gridded version of the methane emission inventory from
the US Environmental Protection Agency (Maasakkers et al., 2016), 75 % of
total national anthropogenic emissions are contributed by
0.1∘×0.1∘ (≈10×10 km2) grid
cells with emission flux E>0.5 kg km-2 h-1, and 30 %
are contributed by grid cells with E>5 kg km-2 h-1
(Jacob et al., 2016). Shen et al. (2022) find a mean emission of 0.18 Tg yr-1 for 12 major oil and gas production basins in the US EPA inventory,
which for a typical basin scale of 200×200 km2 corresponds to
a mean emission flux of 0.5 kg km-2 h-1. Taking E=0.5 kg km-2 h-1 as a desired flux detection threshold on a 100 km scale,
or alternatively E=5 kg km-2 h-1 as a desired flux detection
threshold on a 10 km scale, we find from Eq. (3) a mean enhancement
ΔX=2.0 ppb. Instrument precisions for the flux mappers in Table 1
are in the range 3–15 ppb, and we assume that σX is small in
comparison. We further assume F=0.24 for instruments operating at 1.65 µm
by analogy with GOSAT using the CO2 proxy method (mainly
limited by cloud cover) and F=0.03 for instruments operating at 2.3 µm by analogy with TROPOMI (limited by both cloud cover and spectrally
inhomogeneous surfaces). Other instrument properties are taken from Table 1.
Table 3 shows the results of this illustrative calculation. In the 100 km
resolution case we find that TROPOMI requires a 4-week averaging period,
limited by the small fraction of successful retrievals. GOSAT-GW requires 18 d in global viewing mode, as the greater fraction of successful
retrievals is offset by coarser pixels and 3 d return time. It requires only one
pass in target mode. MethaneSAT requires a single pass and is limited by its
3 d return time. Sentinel-5 requires 5 d, much shorter than TROPOMI
despite coarser pixels because it uses the 1.65 µm band. GeoCarb
requires only 3 d because of its twice-daily observations. CO2M requires
only a single pass and is limited by its 5 d return time. In the 10 km
resolution case, we find that only MethaneSAT has an averaging time of less
than a week, with GOSAT-GW requiring 18 d in target mode (limited by its
lower instrument precision) and other instruments requiring several months
or more. However, both MethaneSAT and GOSAT-GW in target mode only cover
limited domains (200×200 km2 for MethaneSAT).
Averaging time requirements for regional source detection by area
flux mappersa.
InstrumentAveraging timeAveraging timeE=0.5 kg km-2 h-1, 100×100 km2E=5 kg km-2 h-1, 10×10 km2TROPOMI28 d>1 yearGOSAT-GW18 d (global), 3 d (target)18 d (target)MethaneSAT3 d5 dSentinel-55 d>1 yearGeoCarb3 d1 yearCO2M5 d120 d
a Illustrative calculation using the conceptual model of Eqs. (3)–(5) applied to the detection of an emission flux averaging 0.5 kg km-2 h-1 over a desired spatial resolution of 100×100 km2, or 5 kg km-2 h-1 over a desired spatial resolution of
10×10 km2. See the text for details and Table 1 for the
specifications of the different instruments. Results for GOSAT-GW are given
for both global and target viewing modes. Instruments not yet launched are
in italics.
The above conceptual model is crude and overoptimistic, assuming ideal
reduction of errors and uncorrelated retrieval success across instrument
pixels, ignoring variable bias, and taking instrument specifications from
Table 1 at face value, but it is useful for intercomparing instruments and
it highlights critical variables determining detection thresholds for
different applications. The advantage of the 1.65 µm band is readily
apparent because it achieves a much higher success rate through the CO2
proxy retrieval. The MethaneSAT instrument with high precision and small
pixels is most useful for quantifying fluxes at high spatial resolution. For
coarser resolutions, return time and spatial coverage can be more important
considerations.
Point sources
In the case of point source imagers, the detection threshold applies to
single-pass observations of the plumes. Table 4 lists point source detection
thresholds reported in the literature for different instruments. Detection
thresholds are defined by the ability to determine the plume mask against a
noisy background and to retrieve the corresponding emissions. The detection
thresholds for a given instrument depend strongly on surface type and are
lowest for bright, spectrally homogeneous surfaces. They also depend on wind
speed, which complicates the definition of detection threshold because weak
winds facilitate detection but cause large error in quantification (Varon et
al., 2018). The best range of wind speeds to allow both detection and
quantification is 2–5 m s-1 (Varon et al., 2018). Sherwin et al. (2022) conducted a series of controlled release experiments under those
favorable surface and wind conditions and confirmed the ability of GHGSat to
quantify emissions down to 200 kg h-1 and Sentinel-2, Landsat-8,
PRISMA, and WorldView-3 to quantify emissions down to the 1400–4000 kg h-1 range.
Point source detection thresholds for different satellite
instrumentsa.
InstrumentDetection threshold (kg h-1)ReferenceTROPOMI25 000bLauvaux et al. (2022)Sentinel-2, Landsat-8/91800–25 000cVaron et al. (2021); Ehret et al. (2022); Irakulis-Loitxate et al. (2022a)PRISMA500–2000dGuanter et al. (2021)MethaneSAT500Christopher Chan Miller, Harvard University, personal communication, 2022.GHGSat-D1000–3000Jervis et al. (2021)GHGSat-C1, C2100–200eGauthier (2021)Carbon Mapper50–200fDuren (2021)WorldView-3<100Sanchez-Garcia et al. (2022)AVIRIS-NG (aircraft)g2–10hDuren et al. (2019)
a The detection thresholds are as reported in the references and are
generally for favorable winds (<5 m s-1) and favorable
surfaces (bright and spectrally homogeneous) unless otherwise indicated. As
pointed out in the text, weak winds are favorable for detection but not for
quantification, and this places some ambiguity in the definition of detection
threshold. Specifications for each instrument are in Table 1. Instruments
that are yet to be launched are in italics.
b From an ensemble of 1800 observed detections for TROPOMI
5.5×7 km2 pixels. The pixels may contain multiple point
sources.
c Observations over surfaces ranging from bright and homogeneous
(Sahara) to highly heterogeneous (farmland).
d From LES synthetic plumes and observations over surfaces ranging from
Sahara (bright homogeneous surfaces) to Shanxi Province in China (darker
more heterogeneous surfaces with significant terrain).
e Verified by controlled releases (MacLean, 2021; Sherwin et al.,
2022).
f 50 kg h-1 in target mode with pointing and 200 kg h-1 in
push-broom mode.
g Airborne imaging spectrometer with spectral resolution of 5 nm and
pixel resolution of 1–8 m depending on aircraft altitude (Thorpe et al.,
2017).
h Observations in California with range determined by surface
brightness.
For a given surface and wind speed, the main instrument predictors of point
source detection threshold are spatial resolution, spectral resolution, and
precision. Finer spatial resolution decreases the dilution of the plume
enhancements over the pixel area, thus increasing the magnitude of the
enhancements within plume pixels and facilitating detection. An airborne
imaging spectrometer observing from low altitude such as AVIRIS-NG (with
spatial resolution of 1–8 m depending on aircraft altitude) is thus much
more sensitive than satellite instruments with similar spectral resolution.
Higher spectral resolution increases precision and reduces the aliasing of
surface spectral features into the methane retrieval (Cusworth et al., 2019;
Jongaramrungruang et al., 2021). For hyperspectral and multispectral
instruments, the spectral positioning of the bands relative to the methane
absorption lines is also important (Scaffuto et al., 2021; Sanchez-Garcia et
al., 2022). Precision depends on other instrument properties beyond spectral
resolution and positioning, including the capability of pointing to specific
targets to increase the SNR through longer sample collection. Pointing is
how GHGSat achieves a combination of high spatial and spectral resolution.
The detection thresholds in Table 4 are not strictly comparable between
instruments because they reflect different levels of evidence. One may still
usefully classify the instruments by order-of-magnitude thresholds of
∼100, ∼500, and
∼1000–10 000 kg h-1 (Fig. 4). Instruments in the
∼100 kg h-1 class include GHGSat, WorldView-3, and
Carbon Mapper. A typical point source imager with spatial resolution
∼30 m requires spectral resolution of 5 nm or better to fit
into this class (Cusworth et al., 2019), though WorldView-3 can achieve this
class for bright spectrally homogeneous surfaces through its combination of
very high spatial resolution (3.7×3.7 m2) and favorable
spectral positioning (Sanchez-Garcia et al., 2022).
Instruments in the ∼500 kg h-1 class include the land
hyperspectral sensors (PRISMA, EnMAP, EMIT) and MethaneSAT. The land
hyperspectral sensors have ∼30 m spatial resolution and
achieve this class with 10 nm spectral resolution in the 2.3 µm band,
enabling either a full-physics or matched filter retrieval. MethaneSAT will
have coarser 130×400 m2 spatial resolution but higher
precision enabled by 0.3 nm spectral resolution in the 1.65 µm band,
with the added benefit of allowing a CO2 proxy retrieval to minimize
artifacts.
Instruments in the 1000–10 000 kg h-1 class include the multispectral
land sensors Sentinel-2 and Landsat with 20–30 nm spatial resolution and a
single measurement in the 2.3 µm band to allow a simple Beer's law
retrieval. TROPOMI can detect extremely large point sources or clusters of
sources (>25000 kg h-1) over its 5.5×7 km2
pixels (Lauvaux et al., 2022), though coarse spatial resolution hinders
source identification.
The relevance of measuring individual point sources at these different
thresholds can be assessed by considering their contributions to total
emissions. Cusworth et al. (2022) find on average that 40 % of emissions
from US oil and gas fields originate from point sources >10 kg h-1 detectable by AVIRIS-NG. Figure 7 shows the cumulative frequency
distributions (CFDs) by number and total emission of point sources larger
than 10 kg h-1 sampled by airborne remote sensing over California and
over US oil and gas fields (Duren et al., 2019; Cusworth et al., 2022). Results
are shown for individual campaigns and for the combined CFD with equal
weighting between campaigns. A satellite instrument with detection threshold
of 100 kg h-1 could detect 50 %–95 % of point sources depending on the
region (80 % in the combined data set), contributing 75 %–99 % of point
source emissions (95 % for the combined data set). An instrument with
detection threshold of 1000 kg h-1 could detect 0 %–15 % of point
sources (5 % for the combined data set), contributing 0 %–55 % of point
source emissions (30 % in the combined data set). Brandt et al. (2016)
find that sources in the 10–100 kg h-1 range contribute 20 % of
emissions from point sources >10 kg h-1 in their survey of
emissions from US oil and gas fields. The data set of Fig. 7 includes only a few
emitters in the ∼ 10 000 kg h-1 range. Global statistics
of aircraft and satellite data suggest a power law frequency distribution of
point source emissions with ∼100× fewer sources at
10 000 kg h-1 than at 1000 kg h-1 (Ehret et al., 2022; Lauvaux et
al., 2022). These so-called ultra-emitters could still contribute
significantly to total emissions in some regions.
Cumulative frequency distributions (CFDs) of point source rates
above 10 kg h-1 for 3879 point sources detected by airborne remote
sensing in California and in US oil and gas basins by Duren et al. (2019) and
Cusworth et al. (2022). Many of the individual point sources were detected
multiple times, and the values entered in the frequency distributions are
the averages of these detections not including non-detection events; they
thus represent the average emission from the source when on, as is relevant
to the definition of the instrument detection threshold CD in Eq. (8). The colored curves are for individual campaigns, and the black curve is
the combined CFD for all regions with equal weighting per campaign. The top
panel gives the cumulative fraction of emissions contributed by detected
point sources above a given rate, and the bottom panel gives the cumulative
fraction of the number of point sources. For example, a satellite instrument
with detection threshold of 100 kg h-1 could detect 80 % of the point
sources in the combined CFD, contributing 95 % of total point source
emissions. An instrument with detection threshold of 1000 kg h-1 could
detect 5 % of the point sources in the combined CFD, contributing 30 %
of total point source emissions.
Observing system completeness
Here we introduce the concept of observing system completeness as the
capability of an instrument (or ensemble of instruments) to fully quantify
their target emissions within a selected domain and time window. For area
flux mappers the target would be the total methane emissions within the
domain at a desired spatial resolution, while for point source imagers the
target would be the total emissions within the domain contributed by point
sources larger than 10 kg h-1.
Observing system completeness for area flux mappers
Observations from area flux mappers are generally used to infer 2-D
distributions of total emissions over a regional domain of interest by
Bayesian inference. The observing system completeness is then defined by the
DOFS (Sect. 4.1 and Fig. 5). Given n state vector elements of emissions on the
2-D grid, the DOFS tell us how many of those elements are quantified by the
observations, and the averaging kernel sensitivities (diagonal terms of the
averaging kernel matrix, adding up to the DOFS) give that information for
the individual state vector elements.
As pointed out by Nesser et al. (2021) and Varon et al. (2022), it is
possible to roughly estimate the DOFS of an observing system for a selected
domain and time period without doing any actual forward model calculations.
Consider a domain divided into n emission state vector elements of individual
dimension W [km], sampled with an instrument providing m successful observations
over the domain in the selected time period. Let σA be the mean
prior error standard deviation for the individual state vector elements and
σO the mean observational error standard deviation. The DOFS
can then be estimated as
DOFS=nσA2σA2+(σO/k)2m,
where k=ΔX/E [ppb km2 h kg-1] is the Jacobian matrix
element that relates the column mixing ratio enhancement ΔX [ppb]
over a state vector element to the emission flux E [kg km-2 h-1] for
that element. Following Nesser et al. (2021), we can approximate k with a
simple mass balance model as
k=ηMaMCH4WgUp,
where η is a coefficient to account for turbulent diffusion. Nesser et
al. (2021) and Varon et al. (2022a) find that η=0.4 is a suitable
value for W in the range 25–100 km. Further assuming U=5 km h-1 and p=1000 hPa, we obtain k=1.4×1010W [ppb km2 h kg-1]. The mean prior error standard deviation can be estimated as
σA=fQA/(nW2), where QA is the total prior estimate
of emission in the domain [kg h-1] and f is the fractional error (such
as 50 %). For the example of Fig. 5 with a 1-month inversion of TROPOMI
observations over the Permian Basin, Varon et al. (2022) find that this rough
estimate prior to doing the inversions yields a DOFS of 11.7, close to the
value of 10.8 found in the actual inversion.
The simple estimate of DOFS in Eq. (6) yields basic insights into the
factors affecting observing system completeness for an area flux mapper.
Instrument precision and number of observations (or observation density for
a given area) are critical. The bar for the observations to improve on the
prior estimate depends on the estimated error for that prior estimate
(smaller prior error means a higher bar for the observations). Increasing
the requirement on spatial resolution (large n, small W) leads to smaller
absolute prior errors for individual state vector elements and in
turn raises the requirement on the precision and number of observations.
Observing system completeness for point source imagers
Observing system completeness for a point source imager (or a constellation)
can be defined as its ability to quantify total emissions from point sources
larger than 10 kg h-1 over a selected domain and time window. Such
completeness in observation of point sources is important not only for
complementing the information from area flux mappers but also for leak
detection and repair (LDAR) programs where regularly surveying point sources
in a region can enable prompt action to fix malfunctioning equipment (Kemp
et al., 2016; Fox et al., 2021). Current LDAR programs rely on a combination
of ground surveys, drones, and aircraft, but we will see that satellites
have an important role to play.
Let C∈[0,1] denote the observing system completeness for point sources
as the fraction of total point source emissions larger than 10 kg h-1
within a domain and time window that can be detected by a given instrument
(or constellation of instruments). C is limited by a combination of the
instrument detection threshold (CD), spatial coverage (CS), and
temporal sampling (CT):
C=CD×CS×CT.
Here CD is the fraction of point source emissions that can be detected
on the basis of the instrument's detection threshold, as inferred for
example from Fig. 7. CS is the fraction of the domain that the
instrument observes at least once within the time window. If there is full
spatial coverage within the time window, CS=1.CT=1-1-FpN is the probability for an observed source to be actually detected
within the time window given the number N≥1 of observations in the
window, the source persistence p (fraction of time that the source is emitting
above the detection threshold), and the fraction F of successful retrievals,
taken here as the fraction of clear-sky observations. For example, an
intermittent source with p=0.2 that is observed with a 1-week return time
and 30 % clear skies would have CT=0.96 for 1 year of observations
but CT=0.23 for 1 month. If spatial coverage and observing frequency
are sufficient, C is limited by the instrument's detection threshold
(CD). If they are not (and depending on source persistence and cloud
cover), CS and CT may limit observation system completeness
rather than CD.
Figure 8 shows the frequency distribution of persistence (p) for 2500 oil and
gas point sources detected and quantified by the airborne AVIRIS-NG and
Global Airborne Observatory instruments in US field campaigns (Cusworth et
al., 2022). The left panel shows the frequency distribution of mean
emissions from individual point sources for each persistence bin. From there
we can estimate the observing system completeness for any instrument on the
basis of its detection threshold, spatial coverage, and return time. The
right panel plots the resulting cumulative observing system completeness for
the ensemble of 2500 point sources as achieved by either (1) an airborne
instrument with 10 kg h-1 detection threshold and bi-monthly (60 d)
sampling interval or (2) a satellite instrument with 100 kg h-1
detection threshold and bi-weekly (14 d) sampling interval. The
calculation is done for a 1-year time window with 30 % clear skies,
assuming CS=0.95 in both cases, and the cumulative results are shown
across the range of persistence bins. We see in this example that the two
observing systems have comparable success for persistent sources (p>0.5) by trading CD for CT, but the satellite system is better for
intermittent sources (p<0.5), despite its higher detection
threshold, because of the greater benefit from frequent observations.
Point source rates, persistence, and observing system completeness
for an ensemble of 2500 oil and gas point sources sampled by aircraft remote
sensing in five US oil and gas basins (Cusworth et al., 2022). The left panel
shows the frequency distribution of mean point sources rates for different
persistence bins (p, fraction of the time that the source is detected), where
the mean is computed by assuming zero emission when no plume is detected.
Boxes and whiskers indicate 10th, 25th, 50th, 75th, and
90th percentiles. The right panel shows the percentage of total point
source emissions contributed by different persistence bins. Also shown in
that panel is the cumulative observing system completeness C=CD×CS×CT (Eq. 8) for 1 year of observations under 30 % clear-sky
conditions and two observing systems, one with 100 kg h-1 detection
threshold and bi-weekly sampling (green line) and one with 10 kg h-1
and bi-monthly sampling (red line). We assume spatial coverage CS=0.95 for both. The observing system completeness is computed individually
for each basin and then averaged. Both observing systems have comparable
performance for sources with high persistence (p>0.5) but the
biweekly observing system performs better for sources with low persistence
despite its higher detection threshold.
Figure 9 further illustrates the trade space between detection threshold and
return time for determining observing system completeness. Results are for
the ensemble of 2500 point sources with statistics given in Fig. 8. We see
from Fig. 9 that an observing system completeness of 0.6 can be achieved by
an instrument with a detection threshold of 300 kg h-1 sampling weekly.
Such an instrument performs as well as one with low detection threshold but
sampling only every 2 months. Achieving an observing system completeness
higher than 0.8 requires an instrument with detection threshold better than
150 kg h-1 that samples at least biweekly.
Observing system completeness of a point source imager as a
function of detection threshold and return time. The calculation is for the
ensemble of point sources in Fig. 8. Observing system completeness for a
point source imager is defined here as the ability to quantify emissions
from all point sources larger than 10 kg h-1.
Our calculation of CT as presented above assumes that a point source
follows a binary emission frequency distribution (on/off) with constant
emissions when on. Actual sources have more complex variability (Allen et
al., 2022; Zimmerle et al., 2022). Similar to the analysis of Sect. 5.1,
a simple analysis can be done by assuming Gaussian statistics following Hill
and Nassar (2019) to estimate the number N of observations needed to quantify
a mean point source emission rate (1±δ)Q with relative
precision of δ defined by the 95 % relative confidence interval:
9N=1Fp(1.96σδ)2,10σ=σI2+σS2.
Here σ is the standard deviation of individual measurements
determined by instrument precision (σI) and variability in the
source (σS). Using statistics from airborne surveys in the
Permian Basin, we find that 71 observations per year (roughly 5 d return
time, assuming 30 % clear skies) would be required to estimate annual
point source emissions from that highly intermittent population within
50 % (p=0.24, σI=36 %, σS=45 %;
Cusworth et al., 2021b). Increasing the required annual emission precision
to 35 % would require 145 observations per year (2 d return time). For a
less intermittent population (p=0.5), we find N=43 (8 d return time)
to achieve 50 % precision and N=87 (4 d return time) to achieve 35 %
precision. These observing frequencies can be achieved with a satellite
constellation but would be challenging for an airborne program.
The tails of the pdfs for point source emissions are a particular challenge
to sample representatively. The pdfs are generally heavy-tailed, resulting
in a low estimate of mean emissions (Zimmerle et al., 2022), which may be
addressed with very dense sampling (Y. Chen et al., 2022) or with supporting
observations from area flux mappers. Persistence is defined in the
observations by the frequency of occurrence of emissions above the detection
threshold, but non-detection could represent the low tail of the pdf rather
than an on/off switch. The definition of persistence may thus depend on the
detection threshold, increasing the importance of that threshold as a
measure of observing system completeness. Further complicating matters is
that the instrument detection threshold is variable, depending notably on
the wind speed at the time of observation. This calls for better
characterization of the full pdf of emissions from point sources as a means
to extrapolate the observations (Allen et al., 2022).
Concluding remarks
Satellite observations of atmospheric methane in the shortwave infrared
(SWIR) provide an increasingly powerful system for continuous monitoring of
emissions from the global scale down to point sources. We reviewed the
current and scheduled fleet of instruments including area flux mappers to
quantify total emissions on regional scales and point source imagers to
quantify individual source rates. We discussed retrieval methods to infer
concentrations from measured radiances, precision and accuracy requirements,
inverse methods to infer emissions from observed concentrations, emission
detection thresholds, and observing system completeness.
Synergy between different satellite instruments is important to exploit.
Area flux mappers can constrain total emissions, while point source imagers
provide specific facility-level attribution. Detection of coarse-resolution
hotspots by area flux mappers can direct targeted observation by point
source imagers to identify the causes (Maasakkers et al., 2022b). Point
source observations with adequate completeness can improve the bottom-up
estimates used as prior information in inversions of area flux mapper data.
Constellations of point source imagers can achieve high observing system
completeness in support of point source mapping and leak detection
and repair (LDAR) programs.
Synergy with suborbital (ground-based and airborne) platforms is essential
for a multi-tiered observing strategy (Cusworth et al., 2020). Suborbital
observations have a unique role to complement the intrinsic limitations of
satellites in terms of spatial resolution, return time, cloud cover, dark
surfaces, and nighttime. Surface measurements are typically 10 times more
sensitive to local emissions than satellite observations (Cusworth et al.,
2018). They can also include correlative chemical information such as
isotopes, ethane, and ammonia concentrations (Yuan et al., 2015; Ganesan et
al., 2019; Graven et al., 2019; Pétron et al., 2020; Yang et al., 2020).
Correlative chemical information available from satellites needs to be
better exploited. Concurrent satellite observations of CO and methane have
been used to quantify methane emissions from open fires (Worden et al.,
2013) and from cities (Plant et al., 2022) by reference to CO emissions,
although this is contingent on an accurate CO emission inventory, and errors
in these inventories are often large. GeoCarb will measure methane,
CO2, and CO, offering further application of this method, including the
use of methane / CO2 enhancement ratios. Concurrent enhancements of
CO2 and methane in oil and gas fields observed by the PRISMA instrument,
together with nighttime flare data from the VIIRS instrument, have been used
to identify flaring point sources and quantify flaring efficiency (Cusworth
et al., 2021a). Measurements of ammonia from space (Van Damme et al., 2018)
have the potential to identify livestock sources but have not yet been used
in combination with methane.
Some methane sources are intrinsically difficult to observe from space,
including those over water, the wet tropics, and the Arctic. Potentially large
methane sources over water include offshore oil and gas facilities, wastewater
facilities, hydroelectric and agricultural reservoirs, and estuaries. They
can be observed in the sunglint mode or by lidar (Kiemle et al., 2017;
Irakulis-Loitxate et al., 2022b). The wet tropics and the Arctic are a
challenge because of persistent cloudiness, compounded in the Arctic by high
solar zenith angles and polar darkness and by the collocation of oil and gas emissions with
wetland emissions. The MERLIN lidar instrument will provide a unique
observation capability for the Arctic. The GeoCarb geostationary instrument
will increase data density over tropical South America. The tropics are
thought to be the principal driver for the recent methane increase (Chandra
et al., 2021; Yin et al., 2021; Zhang et al., 2021), and there would be
considerable value in dedicated geostationary or inclined-orbit satellite
observations of the tropics with high pixel resolution.
The ultimate goal of top-down methane emission estimates is to improve
bottom-up estimates, as the latter provide the information needed for
climate action by relating emissions to processes. This calls for
partnerships where discrepancies identified by satellite for a particular
sector motivate work to improve bottom-up estimates for that sector. The
International Methane Emissions Observatory (IMEO; United Nations
Environmental Program, 2021) aims to facilitate this infusion of top-down
information into the improvement of bottom-up inventories on a global scale
in support of the Paris agreement, and initiatives in the oil and gas
industry aim to achieve the same at the level of oil and gas production fields
and individual facilities (Cooper et al., 2022).
The capability is thus emerging for satellite observations to anchor a
global methane monitoring system delivering global information on emissions
in near real time, from the global scale down to point sources, to support
climate policy and to guide corrective action. The basic framework for
building such a facility is already here and will be rapidly augmented in
coming years with the launch of new instruments.
Data availability
The GOSAT methane data in Fig. 2 are available at https://www.leos.le.ac.uk/data/GHG/GOSAT/v9.0/CH4_GOS_OCPR_v9.0_final_nceo_2009_2021.tar.gz (last access: 22 July 2022).
The SRON S5P-RemoTeC scientific TROPOMI methane data in Figs. 2 and 3 are available at 10.5281/zenodo.4447228 (Lorente et al., 2021b).
The Sentinel-2 data in Fig. 3 are available at 10.5270/S2_-d8we2fl (European Space Agency, 2021).
The PRISMA and GHGSat data in Fig. 3 are available for non-commercial uses upon request to the corresponding author. The data in Figs. 7 and 8 are available at 10.1038/s41586-019-1720-3 (Duren et al., 2019) for California in 2016–2017,
10.1021/acs.estlett.1c00173 (Cusworth et al., 2021b) for the Permian Basin in 2019,
and 10.5281/zenodo.5606120 (Cusworth et al., 2021d) for the AVIRIS-NG/GAO campaigns in 2020–2021.
Author contributions
DJJ wrote the manuscript with contributions from DJV, DHC, PED, CF, RG, LG, JK, JMK, LEO, BP, ZQ, AKT, JRW, and RMD. DJV, DHC, JK, and ZQ produced the figures. RMD and DHC wrote the
initial draft of Sect. 6.2. JK led the CAMS project that produced this
paper.
Competing interests
The contact author has declared that none of the authors has any competing interests.
Disclaimer
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Acknowledgements
We thank Felipe J. Cardoso-Saldaña and Cynthia Randles
of ExxonMobil Technology and Engineering Company, Yasjka Meijer and Ben Veihelmann of the ESA, Ilse Aben of SRON, and Robert Parker of U. Leicester for their
valuable comments. We thank Halina Dodd of the Halo Agency, LLC, for producing
Fig. 1. Riley M. Duren and Daniel H. Cusworth
acknowledge additional support from Carbon Mapper's philanthropic donors.
Portions of this research was carried out at the Jet Propulsion Laboratory,
California Institute of Technology, under a contract with the National
Aeronautics and Space Administration (grant no. 80NM0018D0004). Philip E. Dennison acknowledges
funding from NASA Carbon Monitoring System (grant no. 80NSSC20K0244).
Financial support
This research has been supported by the Collaboratory to Advance Methane Science (CAMS) and the National Aeronautics and Space Administration, Earth Sciences Division (grant no. NNH20ZDA001N-CMS).
Review statement
This paper was edited by Jason West and reviewed by two anonymous referees.
ReferencesAlexe, M., Bergamaschi, P., Segers, A., Detmers, R., Butz, A., Hasekamp, O., Guerlet, S., Parker, R., Boesch, H., Frankenberg, C., Scheepmaker, R. A., Dlugokencky, E., Sweeney, C., Wofsy, S. C., and Kort, E. A.: Inverse modelling of CH4 emissions for 2010–2011 using different satellite retrieval products from GOSAT and SCIAMACHY, Atmos. Chem. Phys., 15, 113–133, 10.5194/acp-15-113-2015, 2015.
Allen, D. T., Cardoso-Saldaña, F., and Kimura, Y.: Variability in
spatially and temporally resolved emissions and hydrocarbon source
fingerprints for oil and gas sources in shale gas production regions,
Environ. Sci. Technol., 51, 12016–12026,
https://pubs.acs.org/doi/10.1021/acs.est.7b02202, 2017.Allen, D. T., Cardoso-Saldaña, F. J., Kimura, Y., Chen, Q., Xiang, Z.,
Zimmerle, D., Bell, C., Lute, C., Duggan, J., and Harrison, M.: A Methane
Emission Estimation Tool (MEET) for predictions of emissions from upstream
oil and gas well sites with fine scale temporal and spatial resolution:
Model structure and applications, Sci. Total Environ., 829, 154277,
10.1016/j.scitotenv.2022.154277, 2022.Angevine, W. M., Peischl, J., Crawford, A., Loughner, C. P., Pollack, I. B., and Thompson, C. R.: Errors in top-down estimates of emissions using a known source, Atmos. Chem. Phys., 20, 11855–11868, 10.5194/acp-20-11855-2020, 2020.Ayasse, A. K., Thorpe, A. K., Roberts, D. A., Funk, C. C., Dennison, P. E.,
Frankenberg, C., Steffke, A., and Aubrey, A. D.: Evaluating the effects of
surface properties on methane retrievals using a synthetic airborne
visible/infrared imaging spectrometer next generation (AVIRIS-NG) image,
Remote Sens. Environ., 215, 386–397,
10.1016/j.rse.2018.06.018, 2018.Ayasse, A. K., Dennison, P. E., Foote, M., Thorpe, A. K., Joshi, S., Green,
R. O., Duren, R. M., Thompson, D. R., and Roberts, D. A.: Methane mapping with
future satellite imaging spectrometers, Remote Sensing, 11, 3054, 10.3390/rs11243054, 2019.Barré, J., Aben, I., Agustí-Panareda, A., Balsamo, G., Bousserez, N., Dueben, P., Engelen, R., Inness, A., Lorente, A., McNorton, J., Peuch, V.-H., Radnoti, G., and Ribas, R.: Systematic detection of local CH4 anomalies by combining satellite measurements with high-resolution forecasts, Atmos. Chem. Phys., 21, 5117–5136, 10.5194/acp-21-5117-2021, 2021.Bergamaschi, P., Houweling, S., Segers, A., Krol, M., Frankenberg, C.,
Scheepmaker, R. A., Dlugokencky, E., Wofsy, S. C., Kort, E. A., Sweeney, C.,
Schuck, T., Brenninkmeijer, C., Chen, H., Beck, V., and Gerbig, C.:
Atmospheric CH4 in the first decade of the 21st century: Inverse
modeling analysis using SCIAMACHY satellite retrievals and NOAA surface
measurements, J. Geophys. Res.-Atmos., 118, 7350–7369,
10.1002/jgrd.50480, 2013.Bloom, A. A., Bowman, K. W., Lee, M., Turner, A. J., Schroeder, R., Worden, J. R., Weidner, R., McDonald, K. C., and Jacob, D. J.: A global wetland methane emissions and uncertainty dataset for atmospheric chemical transport models (WetCHARTs version 1.0), Geosci. Model Dev., 10, 2141–2156, 10.5194/gmd-10-2141-2017, 2017.Borchardt, J., Gerilowski, K., Krautwurst, S., Bovensmann, H., Thorpe, A. K., Thompson, D. R., Frankenberg, C., Miller, C. E., Duren, R. M., and Burrows, J. P.: Detection and quantification of CH4 plumes using the WFM-DOAS retrieval on AVIRIS-NG hyperspectral data, Atmos. Meas. Tech., 14, 1267–1291, 10.5194/amt-14-1267-2021, 2021.Bousserez, N., Henze, D.K., Perkins, A., Bowman, K.W., Lee, M., Liu, J.,
Deng, F., and Jones, D. B. A.: Improved analysis error covariance matrix for
high-dimensional variational inversions: application to source estimation
using a 3D atmospheric transport model, Q. J. Roy. Meteor. Soc., 141, 1906–1921,
10.1002/qj.2495, 2015.Brandt, A. R., Heath, G. R., and Cooley, D.: Methane leaks from natural gas
systems follow extreme distributions, Environ. Sci. Technol., 50,
12512–12520, 10.1021/acs.est.6b04303, 2016.Brasseur, G. P. and Jacob, D. J.: Modeling of Atmospheric Chemistry, Cambridge
University Press, Cambridge, UK,
10.1017/9781316544754, 2017.Bruhwiler, L., Basu, S., Butler, J. H., Chatterjee, A., Dlugokencky, E.,
Kenney, M. A., McComiskey, A., Montzka, S. A., and Stanitski, D.: Observations
of greenhouse gases as climate indicators, Clim. Change, 165, 12,
10.1007/s10584-021-03001-7, 2021.Buchwitz, M., Reuter, M., Schneising, O., Boesch, H., Guerlet, S., Dils, B.,
Aben, I., Armante, R., Bergamaschi, P., Blumenstock, T., Bovensmann, H.,
Brunner, D., Buchmann, B., Burrows, J., Butz, A., Chédin, A.,
Chevallier, F., Crevoisier, C., Deutscher, N., Frankenberg, C., Hase, F.,
Hasekamp, O., Heymann, J., Kaminski, T., Laeng, A., Lichtenberg, G.,
Mazière, M. D., Noël, S., Notholt, J., Orphal, J., Popp, C., Parker,
R., Scholze, M., Sussmann, R., Stiller, G., Warneke, T., Zehner, C.,
Bril, A., Crisp, D., Griffith, D., Kuze, A., O'Dell, C., Oshchepkov, S.,
Sherlock, V., Suto, H.,Wennberg, P.,Wunch, D., Yokota,
T., and Yoshida, Y.: The Greenhouse Gas Climate Change Initiative (GHG-CCI):
Comparison and quality assessment of near-surface-sensitive
satellite-derived CO2 and CH4 global data sets, Remote Sens.
Environ., 162, 344–362,
10.1016/j.rse.2013.04.024, 2015.Butz, A., Galli, A., Hasekamp, O., Landgraf, J., Tol, P., and Aben, I.:
TROPOMI aboard Precursor Sentinel-5 Precursor: Prospective performance of
CH4 retrievals for aerosol and cirrus loaded atmospheres, Remote Sens.
Environ., 120, 267–276,
10.1016/j.rse.2011.05.030, 2012.Cardoso-Saldaña, F. J., and Allen, D. T.: Projecting the temporal
evolution of methane emissions from oil and gas production sites, Environ.
Sci. Technol., 54, 14172–14181, 10.1021/acs.est.0c03049, 2020.Chandra, N., Patra, P.K., Bisht, J. S. H., Ito, A., Umezawa, T., Saigusa, N.,
Morimoto, S., Aoki, S., Janssens-Maenhout, G., Fujita, R., Takigawa, M.,
Watanabe, and Saitoh, N.: Emissions from the oil and gas sectors, coal
mining and ruminant farming drive methane growth over the past three
decades, J. Met. Soc. Jpn., 99, 309–337,
10.2151/jmsj.2021-015, 2021.Chen, Y., Sherwin, E.D., Berman, E. S. F., Jones, B. B., Gordon, M. P.,
Wetherley, E. B., Kort, E. A., and Brandt, A. R.: Quantifying regional methane
emissions in the New MexicoPermian Basin with a comprehensive aerial survey,
Environ. Sci. Technol.,
10.1021/acs.est.1c06458, 2022.Chen, Z., Jacob, D., Nesser, H., Sulprizio, M., Lorente, A., Varon, D., Lu, X., Shen, L., Qu, Z., Penn, E., and Yu, X.: Methane emissions from China: a high-resolution inversion of TROPOMI satellite observations, Atmos. Chem. Phys. Discuss. [preprint], 10.5194/acp-2022-303, in review, 2022.Cho, T., Chung, J., Miller, S. M., and Saibaba, A. K.: Computationally efficient methods for large-scale atmospheric inverse modeling, Geosci. Model Dev., 15, 5547–5565, 10.5194/gmd-15-5547-2022, 2022.Cooper, J., Dubey, L., and Hawkes, A.: Methane detection and quantification
in the upstream oil and gas sector: the role of satellites in emissions
detection, reconciling and reporting, Environ. Sci. Atmos.,
2, 9–23, 10.1039/D1EA00046B, 2022.Cusworth, D. H., Jacob, D. J., Sheng, J.-X., Benmergui, J., Turner, A. J., Brandman, J., White, L., and Randles, C. A.: Detecting high-emitting methane sources in oil/gas fields using satellite observations, Atmos. Chem. Phys., 18, 16885–16896, 10.5194/acp-18-16885-2018, 2018.Cusworth, D. H., Jacob, D. J., Varon, D. J., Chan Miller, C., Liu, X., Chance, K., Thorpe, A. K., Duren, R. M., Miller, C. E., Thompson, D. R., Frankenberg, C., Guanter, L., and Randles, C. A.: Potential of next-generation imaging spectrometers to detect and quantify methane point sources from space, Atmos. Meas. Tech., 12, 5655–5668, 10.5194/amt-12-5655-2019, 2019.Cusworth, D. H., Duren, R. M., Yadav, V., Thorpe, A. K., Verhulst, K.,
Sander, S., Hopkins, F., Raiq, T., and Miller, C. E.: Synthesis of methane
observations across scales: Strategies for deploying a multitiered observing
network, Geophys. Res. Lett., 47, e2020GL087869,
10.1029/2020GL087869, 2020.Cusworth, D. H., Duren, R., Thorpe, A. K., Pandey, S., Aben, I., Jervis, D.,
Varon, D., Jacob, D. J., Randles, C. A., Gautam, R., Omara, M., Schade, G.,
Dennison, P. E., Frankenberg, C., Gordon, D., Lopinto, E., and Miller, C. E.:
Multi-satellite imaging of a gas well blowout provides new insights for
methane monitoring, Geophys. Res. Lett., 48, e2020GL090864, 10.1029/2020GL090864, 2021a.Cusworth, D. H., Duren, R. M., Thorpe, A. K., Olson-Duvall, W., Heckler, J.,
Chapman, J. W., Eastwod, M. L., Helmlinger, M. C., Green, R. O., Asner, G. P.,
Dennison, P. E., and Miller, C. E.: Intermittency of large methane emitters in
the Permian Basin, Environ. Sci. Technol. Lett., 8, 567–573,
10.1021/acs.estlett.1c00173, 2021b.Cusworth, D. H., Bloom, A. A., Ma, S., Miller, C. E., Bowman, K., Yin, Y.,
Maasakkers, J. D., Zhang, Y., Scarpelli, T. R., Qu, Z., Jacob, D. J. and
Worden, J. R.: A Bayesian framework for deriving sector-based methane
emissions from top-down fluxes, Nat. Commun. Earth Environ., 2,
242, 10.1038/s43247-021-00312-6, 2021c.Cusworth, D. H., Duren, R., Ayasse, A., Thorpe, A., Asner, G., and Green, R.: Methane plumes from airborne surveys 2020–2021 (1.0), Zenodo [data set], 10.5281/zenodo.5606120, 2021d.Cusworth, D. H., Thorpe, A. K., Ayasse, A. K., Stepp, D., Heckler, J., Asner,
G. P., Miller, C. E., Chapman, J. W., Eastwood, M. L., Green, R. O., Hmiel, B.,
Lyon, D., and Duren, R. M.: Strong methane point sources contribute a
disproportionate fraction of total emissions across multiple basins in the
U.S., Earth ArXiv, 10.31223/X53P88, 2022.Deng, Z., Ciais, P., Tzompa-Sosa, Z. A., Saunois, M., Qiu, C., Tan, C., Sun, T., Ke, P., Cui, Y., Tanaka, K., Lin, X., Thompson, R. L., Tian, H., Yao, Y., Huang, Y., Lauerwald, R., Jain, A. K., Xu, X., Bastos, A., Sitch, S., Palmer, P. I., Lauvaux, T., d'Aspremont, A., Giron, C., Benoit, A., Poulter, B., Chang, J., Petrescu, A. M. R., Davis, S. J., Liu, Z., Grassi, G., Albergel, C., Tubiello, F. N., Perugini, L., Peters, W., and Chevallier, F.: Comparing national greenhouse gas budgets reported in UNFCCC inventories against atmospheric inversions, Earth Syst. Sci. Data, 14, 1639–1675, 10.5194/essd-14-1639-2022, 2022.Duren, R. M., Thorpe, A. K., Foster, K. T., Rafiq, T., Hopkins, F. M.,
Yadav, V., Bue, B. D., Thompson, D. R., Conley, S., Colombi, N. K.,
Frankenberg, C., McCubbin, I.B., Eastwood, M. L., Falk, M., Herner, J. D.,
Croes, B. E., Green, R. O., and Miller, C. E.: California's methane
super-emitters, Nature, 575,
180–184, 10.1038/s41586-019-1720-3, 2019.Duren, R. M.: Carbon Mapper: on-orbit performance predictions and airborne
prototyping, American Geophysical Union Fall Meeting, New Orleans, 13–17 December 2021, A53F-05,
https://agu.confex.com/agu/fm21/meetingapp.cgi/Paper/985888 (last access: 23 July 2022),
2021.Ehret, T., Bousquet, P., Pierangelo, C., Alpers M., Millet, B., Abshire, J.B., Bovensmann, H., Burrows, J.P., Chevallier, F., Ciais, P., Crevoisier, C., Fix, A., Flamant, P., Frankenberg, C., Gibert, F., Heim, B., Heimann, M., Houweling, S., Hubberten, H.W., Jöckel, P., Law, K., Löw, A., Marshall, J., Agusti-Panareda, A., Payan, S., Prigent, C., Rairoux, P., Sachs, T., Scholze, M., and Wirth, M.: MERLIN: A French-German
space lidar mission dedicated to atmospheric methane, Remote Sens., 9,
1052, 10.3390/rs9101052, 2017.Ehret, T., De Truchis, A., Mazzolini, M., Morel, J.-M., d'Aspremont, A.,
Lauvaux, T., and Facciolo, G.: Global tracking and quantification of oil and
gas methane leaks from recurrent Sentinel-2 imagery, arXiv, 2110.11832v1,
physics.ao-ph, 10.48550/arXiv.2110.11832, 2022.European Space Agency: Sentinel-5, ESA SP-1336, ISBN 978-92-9221-434-0, https://esamultimedia.esa.int/docs/EarthObservation/SP_1336_Sentinel-5_web.pdf (last access: 23 July 2022),
2020.European Space Agency: Copernicus Sentinel-2 (processed by ESA), MSI Level-1C TOA Reflectance Product, Collection 0, European Space Agency [data set], 10.5270/S2_-d8we2fl, 2021.Feng, L., Palmer, P. I., Bösch, H., Parker, R. J., Webb, A. J., Correia, C. S. C., Deutscher, N. M., Domingues, L. G., Feist, D. G., Gatti, L. V., Gloor, E., Hase, F., Kivi, R., Liu, Y., Miller, J. B., Morino, I., Sussmann, R., Strong, K., Uchino, O., Wang, J., and Zahn, A.: Consistent regional fluxes of CH4 and CO2 inferred from GOSAT proxy XCH4 : XCO2 retrievals, 2010–2014, Atmos. Chem. Phys., 17, 4781–4797, 10.5194/acp-17-4781-2017, 2017.Foote, M. D., Dennison, P. E., Thorpe, A. K., Thompson, D. R.,
Jongaramrungruang, S., Frankenberg, C., and Joshi, S. C.: Fast and accurate
retrieval of methane concentration from imaging spectrometer data using
sparsity prior, IEEE Trans. Geosci. Remote Sens., 58, 6480–6492, 10.1109/TGRS.2020.2976888, 2020.Fox, T. A., Gao, M., Barchyn, T. E., Jamin, Y. L., and Hugenholtz, C. H.: An
agent-based model for estimating emissions reduction equivalence among leak
detection and repair programs, J. Cleaner Production, 282, 125237,
10.1016/j.jclepro.2020.125237, 2021.Frankenberg, C., Platt, U., and Wagner, T.: Iterative maximum a posteriori
(IMAP)-DOAS for retrieval of strongly absorbing trace gases: Model studies
for CH4 and CO2 retrieval from near infrared spectra of SCIAMACHY onboard ENVISAT, Atmos. Chem. Phys., 5, 9–22, 10.5194/acp-5-9-2005, 2005.Frankenberg, C., Thorpe, A. K., Thompson, D. R., Hulley, G., Kort, E. A.,
Vance, N., Borchardt, J., Krings, T., Gerilowski, K., Sweeney, C., Conley, S., Bue, B. D., Aubrey, A. D., Hook, S., Green, R. O.: Airborne methane remote measurements reveal heavy-tail
flux distribution in Four Corners region, P. Natl. Acad. Sci. USA, 113, 9734–9739,
10.1073/pnas.1605617113, 2016.Ganesan, A. L., Schwietzke, S., Poulter, B., Arnold, T., Lan, X., Rigby, M.,
Vogel, F. R., van der Werf, G. R., Janssens‐Maenhout, G., Boesch, H., Pandey, S., Manning, A. J., Jackson, R. B.,Nisbet, E. G., and
Manning, M. R.: Advancing scientific understanding of the global methane budget in
support of the Paris Agreement, Global Biogeochem. Cy., 33,
1475–1512, 10.1029/2018GB006065, 2019.Gauthier, J.-F.: The importance of matching needs to satellite system
capability when monitoring methane emissions from space, IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Brussels, 11–16 July 2021,
10.1109/IGARSS47720.2021.9555123, 2021.GEO, ClimateTRACE, WGIC: GHG Monitoring from Space: A mapping of
capabilities across public, private, and hybrid satellite missions,
https://wgicouncil.org/publication/reports/industry-reports/ (last access: 23 July 2022),
2021.Graven, H., Hocking, T., and Zazzeri, G.: Detection of fossil and biogenic
methane at regional scales using atmospheric radiocarbon, Earth's Future, 7,
283–299, 10.1029/2018EF001064, 2019.Guanter, L. Irakulis-Loitxate, I., Gorroño, J., Sánchez-García,
E., Cusworth, D. H., Varon, D. J., Cogliati, S., and Colombo, R.: Mapping
methane point emissions with the PRISMA spaceborne imaging spectrometer,
Remote Sens. Environ., 265, 112671,
10.1016/j.rse.2021.112671, 2021.Hajny, K. D., Salmon, O. E., Rudek, J., Lyon, D. R., Stuff, A. A., Stirm, B. H.,
Kaeser, R., Floerchinger, C. R., Conley, S., Smith, M. L., and Shepson, P. B.:
Observations of methane emissions from natural gas-fired power plants,
Environ. Sci. Technol., 15, 8976–8984,
10.1021/acs.est.9b01875, 2019.Heald, C. L., Jacob, D. J., Jones, D. B. A., Palmer, P. I., Logan, J. A., Streets,
D. G., Sachse, G. W., Gille, J. C., Hoffman, R. N., and Nehrkorn, T.:
Comparative inverse analysis of satellite (MOPITT) and aircraft (TRACE-P)
observations to estimate Asian sources of carbon monoxide, J. Geophys. Res.,
109, D23306, 10.1029/2004JD005185, 2004.Hill, T. and Nassar, R.: Pixel size and revisit rate requirements for
monitoring power plant CO2 emissions from space, Remote Sens., 11,
1608, 10.3390/rs11131608, 2019.Hulley, G. C., Duren, R. M., Hopkins, F. M., Hook, S. J., Vance, N., Guillevic, P., Johnson, W. R., Eng, B. T., Mihaly, J. M., Jovanovic, V. M., Chazanoff, S. L., Staniszewski, Z. K., Kuai, L., Worden, J., Frankenberg, C., Rivera, G., Aubrey, A. D., Miller, C. E., Malakar, N. K., Sánchez Tomás, J. M., and Holmes, K. T.: High spatial resolution imaging of methane and other trace gases with the airborne Hyperspectral Thermal Emission Spectrometer (HyTES), Atmos. Meas. Tech., 9, 2393–2408, 10.5194/amt-9-2393-2016, 2016.IPCC: 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories, edited by: Calvo Buendia, E., Tanabe, K., Kranjc, A., Jamsranjav, B., Fukuda,
M., Ngarize, S., Osako, A., Pyrozhenko, Y., Shermanau, P., and Federici, S., IPCC, Switzerland, https://www.ipcc.ch/report/2019-refinement-to-the-2006-ipcc-guidelines (last access: 23 July 2022), 2019.Irakulis-Loitxate, I., Guanter, L., Liu, Y.-N., Varon, D. J., Maasakkers,
J. D., Zhang, Y., Thorpe, A. K., Duren, R. M., Frankenberg, C., Lyon, D.,
Hmiel, B., Cusworth, D. H., Zhang, Y., Segl, K., Gorron, J., Sanchez-Garcıa, E., Sulprizio, M. P., Cao, K., Zhu, H., Liang, J., Li, X., Aben, I., and
Jacob, D. J.: Satellite-based survey of extreme methane emissions in the
Permian Basin , Sci. Adv., 7, eabf4507, 10.1126/sciadv.abf4507,
2021.Irakulis-Loitxate, I., Guanter, L., Maasakkers, J. D., Zavala-Araiza, D., and
Aben, I.: Satellites detect abatable super-emissions in one of the world's
largest methane hotspot regions, Environ. Sci. Technol. 56, 4, 2143–2152,
10.1021/acs.est.1c04873, 2022a.Irakulis-Loitxate, I., Gorrono, J., Zavala-Araiza, D., and Guanter, L.:
Satellites detect a methane ultra-emission event from an offshore platform
in the Gulf of Mexico, Environ. Sci. Technol. Lett., 9, 522-525,
10.1021/acs.estlett.2c00225, 2022b.Jacob, D. J., Turner, A. J., Maasakkers, J. D., Sheng, J., Sun, K., Liu, X., Chance, K., Aben, I., McKeever, J., and Frankenberg, C.: Satellite observations of atmospheric methane and their value for quantifying methane emissions, Atmos. Chem. Phys., 16, 14371–14396, 10.5194/acp-16-14371-2016, 2016.Janardanan, R., Maksyutov, S., Tsuruta, A., Wang, F., Tiwari, Y. K.,
Valsala, V., Ito, A., Yoshida, Y., Kaiser, J. W., Janssens-Maenhout, G., Arshinov, M., Sasakawa, M., Tohjima, Y., Worthy, D. E. J.,
Dlugokencky, E. J., Ramonet, M., Arduini, J.,
Lavric, J. V., Piacentino, S., Krummel, P. B., Langenfelds, R. L.,
Mammarella, I., and Matsunaga, T.: Country-scale analysis of methane
emissions with a high-tesolution inverse model using GOSAT and surface
observations, Remote Sens., 12, 375, 10.3390/rs12030375,
2020.Jervis, D., McKeever, J., Durak, B. O. A., Sloan, J. J., Gains, D., Varon, D. J., Ramier, A., Strupler, M., and Tarrant, E.: The GHGSat-D imaging spectrometer, Atmos. Meas. Tech., 14, 2127–2140, 10.5194/amt-14-2127-2021, 2021.Johnson, D. and Heltzel, R.: On the long-term temporal variations in
methane emissions from an unconventional natural gas well site, ACS Omega,
6, 14200–14207, 10.1021/acsomega.1c00874, 2021.Jongaramrungruang, S., Frankenberg, C., Matheou, G., Thorpe, A. K., Thompson, D. R., Kuai, L., and Duren, R. M.: Towards accurate methane point-source quantification from high-resolution 2-D plume imagery, Atmos. Meas. Tech., 12, 6667–6681, 10.5194/amt-12-6667-2019, 2019.Jongaramrungruang, S., Matheou, G., Thorpe, A. K., Zeng, Z.-C., and Frankenberg, C.: Remote sensing of methane plumes: instrument tradeoff analysis for detecting and quantifying local sources at global scale, Atmos. Meas. Tech., 14, 7999–8017, 10.5194/amt-14-7999-2021, 2021.Jongaramrungruang, S., Matheou, G., Thorpe, A. K., Zeng, Z.-C., and
Frankenberg, C.: MethaNet – An AI-driven approach to quantifying methane
point-source emission from high-resolution 2-D plume imagery, Remote Sens.
Environ., 269, 112809,
10.1016/j.rse.2021.112809, 2022.Karion, A., Sweeney, C., Kort, E. A., Shepson, P. B., Brewer, A., Cambaliza,
M., Conley, S. A., Davis, K., Deng, A., Hardesty, M., Herndon, S. C.,
Lauvaux, T., Lavoie, T., Lyon, D., Newberger, T., Petron, G., Rella, C.,
Smith, M., Wolter, S., Yacovitch, T. I., and Tans, P.: Aircraft-Based
Estimate of Total Methane Emissions from the Barnett Shale Region, Environ.
Sci. Technol., 49, 8124–8131, 10.1021/acs.est.5b00217,
2015.Kemp, C. E., Ravikumar, A. P., and Brandt, A. R.: Comparing natural gas leakage
detection technologies using an open-source “virtual gas field” simulator,
Environ. Sci. Technol., 50, 4546–4553,
10.1021/acs.est.5b06068, 2016.Kiemle C., Ehret, G., Amediek, A., Fix, A., Quatrevalet, M., and Wirth, M.:
Potential of spaceborne lidar measurements of carbon dioxide and methane
emissions from strong point sources, Remote Sens., 9, 1137,
10.3390/rs9111137, 2017.Koo, J.-H., Walker, K. A., Jones, A., Sheese, P. E., Boone, C. D., Bernath,
P. F., and Manney, G. L.: Global climatology based on the ACE-FTS version
3.5 dataset: Addition of mesospheric levels and carbon-containing species in
the UTLS, J. Quant. Spectrosc. Ra., 186, 52–62, 10.1016/j.jqsrt.2016.07.003, 2017.Krings, T., Gerilowski, K., Buchwitz, M., Reuter, M., Tretner, A., Erzinger, J., Heinze, D., Pflüger, U., Burrows, J. P., and Bovensmann, H.: MAMAP – a new spectrometer system for column-averaged methane and carbon dioxide observations from aircraft: retrieval algorithm and first inversions for point source emission rates, Atmos. Meas. Tech., 4, 1735–1758, 10.5194/amt-4-1735-2011, 2011.Lan, X., Basu, S., Schwietzke, S., Bruhwiler, L. M. P., Dlugokencky, E. J.,
Michel, S. E., et al.: Improved constraints on global
methane emissions and sinks using δ13C-CH4, Glob. Biogeochem.
Cy., 35, e2021GB007000, 10.1029/2021GB007000, 2021.Lauvaux, T., Giron, C., Mazzolini, M., d'Aspremont, A., Duren, R., Cusworth,
D., Shindell, D., and Ciais, P.: Global assessment of oil and gas methane
ultra-emitters, Science, 375, 557–561, 10.31223/X5NS54, 2022.Lorente, A., Borsdorff, T., Butz, A., Hasekamp, O., aan de Brugh, J., Schneider, A., Wu, L., Hase, F., Kivi, R., Wunch, D., Pollard, D. F., Shiomi, K., Deutscher, N. M., Velazco, V. A., Roehl, C. M., Wennberg, P. O., Warneke, T., and Landgraf, J.: Methane retrieved from TROPOMI: improvement of the data product and validation of the first 2 years of measurements, Atmos. Meas. Tech., 14, 665–684, 10.5194/amt-14-665-2021, 2021a.Lorente, A., Borsdorff, T., aan de Brugh, J., Landgraf, J., and Hasekamp, O.: SRON S5P – RemoTeC scientific TROPOMI XCH4 dataset, Zenodo [data set], 10.5281/zenodo.4447228, 2021b.Lu, X., Jacob, D. J., Zhang, Y., Maasakkers, J. D., Sulprizio, M. P., Shen, L., Qu, Z., Scarpelli, T. R., Nesser, H., Yantosca, R. M., Sheng, J., Andrews, A., Parker, R. J., Boesch, H., Bloom, A. A., and Ma, S.: Global methane budget and trend, 2010–2017: complementarity of inverse analyses using in situ (GLOBALVIEWplus CH4 ObsPack) and satellite (GOSAT) observations, Atmos. Chem. Phys., 21, 4637–4657, 10.5194/acp-21-4637-2021, 2021.Lu, X., Jacob, D. J., Wang, H., Maasakkers, J. D., Zhang, Y., Scarpelli, T. R., Shen, L., Qu, Z., Sulprizio, M. P., Nesser, H., Bloom, A. A., Ma, S., Worden, J. R., Fan, S., Parker, R. J., Boesch, H., Gautam, R., Gordon, D., Moran, M. D., Reuland, F., Villasana, C. A. O., and Andrews, A.: Methane emissions in the United States, Canada, and Mexico: evaluation of national methane emission inventories and 2010–2017 sectoral trends by inverse analysis of in situ (GLOBALVIEWplus CH4 ObsPack) and satellite (GOSAT) atmospheric observations, Atmos. Chem. Phys., 22, 395–418, 10.5194/acp-22-395-2022, 2022.Lyon, D. R., Zavala-Araiza, D., Alvarez, R. A., Harriss, R., Palacios, V.,
Lan, X., Talbot, R., Lavoie, T., Shepson, P., Yacovitch, T. I., Herndon, S.
C., Marchese, A. J., Zimmerle, D., Robinson, A. L., and Hamburg, S. P.:
Constructing a spatially
resolved methane emission inventory for the Barnett Shale region, Environ.
Sci. Technol., 49, 8147–8157, 10.1021/es506359c, 2015.Lyon, D. R., Alvarez, R. A., Zavala-Araiza, D., Brandt, A. R., Jackson, R. B.,
and Hamburg, S. P.: Aerial surveys of elevated hydrocarbon emissions from oil
and gas production sites, Environ. Sci. Technol., 50, 4877–4886,
10.1021/acs.est.6b00705, 2016.Lyon, D. R., Hmiel, B., Gautam, R., Omara, M., Roberts, K. A., Barkley, Z. R., Davis, K. J., Miles, N. L., Monteiro, V. C., Richardson, S. J., Conley, S., Smith, M. L., Jacob, D. J., Shen, L., Varon, D. J., Deng, A., Rudelis, X., Sharma, N., Story, K. T., Brandt, A. R., Kang, M., Kort, E. A., Marchese, A. J., and Hamburg, S. P.: Concurrent variation in oil and gas methane emissions and oil price during the COVID-19 pandemic, Atmos. Chem. Phys., 21, 6605–6626, 10.5194/acp-21-6605-2021, 2021.Maasakkers, J. D., Jacob, D. J., Sulprizio, M. P., Turner, A. J., Weitz, M.,
Wirth, T., Hight, C., DeFigueiredo, M., Desai, M.,
Schmeltz, R., Hockstad, L., Bloom, A. A., Bowman, K. W., Jeong, S., and
Fischer, M. L.: Gridded national inventory
of U. S. methane emissions, Environ. Sci. Technol., 50, 13123–13133,
10.1021/acs.est.6b02878, 2016.Maasakkers, J. D., Jacob, D. J., Sulprizio, M. P., Scarpelli, T. R., Nesser, H., Sheng, J.-X., Zhang, Y., Hersher, M., Bloom, A. A., Bowman, K. W., Worden, J. R., Janssens-Maenhout, G., and Parker, R. J.: Global distribution of methane emissions, emission trends, and OH concentrations and trends inferred from an inversion of GOSAT satellite data for 2010–2015, Atmos. Chem. Phys., 19, 7859–7881, 10.5194/acp-19-7859-2019, 2019.Maasakkers, J. D., Jacob, D. J., Sulprizio, M. P., Scarpelli, T. R., Nesser, H., Sheng, J., Zhang, Y., Lu, X., Bloom, A. A., Bowman, K. W., Worden, J. R., and Parker, R. J.: 2010–2015 North American methane emissions, sectoral contributions, and trends: a high-resolution inversion of GOSAT observations of atmospheric methane, Atmos. Chem. Phys., 21, 4339–4356, 10.5194/acp-21-4339-2021, 2021.Maasakkers, J. D., Omara, M., Gautam, R., Lorente, A., Pandey, S., Toi, P.,
Borsdorff, T., Houweling, S., and Aben, I.: Reconstructing and quantifying
methane emissions from the full duration of a 38-day natural gas well
blowout using space-based observations, Remote Sens. Environ., 270,
112755, 10.1016/j.rse.2021.112755, 2022a.Maasakkers, J. D., Varon, D. J., Elfarsdottir, A., McKeever, J. D., Jervis, D.,
Mahapatra, G., Pandey, S., Lorente, A., Borsdorff, T., Foorthuis, L. R.,
Schuit, B. J., Tol, P., van Kempen, T. A., van Hees, R., and Aben, I.: Using
satellites to uncover large methane emissions from landfills, arXiv
[preprint], 10.31223/X5N33G,
2022b.MacLean, J.-P.: Detecting and quantifying methane emissions with the
high-resolution GHGSat satellite constellation, American Geophysical Union
Fall meeting, New Orleans, 13–17 December 2021, A54F-01,
https://agu.confex.com/agu/fm21/meetingapp.cgi/Paper/946992 (last access: 23 July 2022), 2021.Miller, S. M. and Michalak, A. M.: Constraining sector-specific CO2 and
CH4 emissions in the US, Atmos. Chem. Phys., 17, 3963–3985, 10.5194/acp-17-3963-2017, 2017.Miller, S. M., Michalak, A. M., Detmers, R. G., Hasekamp, O. P., Bruhwiler,
L. M. P., and Schwietzke, S.: China's coal mine methane regulations have not
curbed growing emissions, Nat. Commun., 10, 303,
10.1038/s41467-018-07891-7, 865, 2019.Miller, S. M., Saibaba, A. K., Trudeau, M. E., Mountain, M. E., and Andrews, A. E.: Geostatistical inverse modeling with very large datasets: an example from the Orbiting Carbon Observatory 2 (OCO-2) satellite, Geosci. Model Dev., 13, 1771–1785, 10.5194/gmd-13-1771-2020, 2020.Moore, B., Crowell, S. M. R., Rayner, P. J., Kumer, J., O'Dell, C. W., O'Brien,
D., Utembe, S., Polonsky, I., Schimel, D., and Lemen, J.: The Potential of
the Geostationary Carbon Cycle Observatory (GeoCarb) to provide multi-scale
constraints on the carbon cycle in the Americas, Front. Environ. Sci., 6, 109,
10.3389/fenvs.2018.00109, 2018.Naik, V., Szopa, S., Adhikary, B., Artaxo, P., Berntsen, T., Collins, W. D., Fuzzi, S., Gallardo, L., Kiendler Scharr, A., Klimont, Z., Liao, H., Unger, N., and Zanis, P.: Short-Lived Climate Forcers, in:
Climate Change 2021: The Physical Science Basis. Contribution of Working
Group I to the Sixth Assessment Report of the Intergovernmental Panel on
Climate Change, edited by: Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S. L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M. L., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J. B. R., Maycock, T. K., Waterfield, T., Yelekçi, O., Yu, R., and Zhou, B.,
Cambridge University Press, https://www.ipcc.ch/report/ar6/wg1/downloads/report/IPCC_AR6_WGI_Chapter06.pdf (last access: 23 July 2022), 2021.Nesme N, Marion, R., Lezeaux, O., Doz, S., Camy-Peyret, C., and Foucher,
P.-Y.: Joint use of in-scene background radiance estimation and optimal
estimation methods for quantifying methane emissions using PRISMA
hyperspectral satellite data: application to the Korpezhe industrial site,
Remote Sens., 13, 4992, 10.3390/rs13244992,
2021.Nesser, H., Jacob, D. J., Maasakkers, J. D., Scarpelli, T. R., Sulprizio, M. P., Zhang, Y., and Rycroft, C. H.: Reduced-cost construction of Jacobian matrices for high-resolution inversions of satellite observations of atmospheric composition, Atmos. Meas. Tech., 14, 5521–5534, 10.5194/amt-14-5521-2021, 2021.NIES: GOSAT-GW mission: background, aims, and mission requirements,
https://gosat-gw.global-atmos-chem-lab.jp/en/collaboration/ (last access: 23 July 2022),
2021.Nisbet, E. G., Fisher, R. E., Lowry, D., France, J. L., Allen, G.,
Bakkaloglu, S., Broderick, T. J., Cain, M., Coleman, M., Fernandez, J., Forster, G., Griffiths, P. T., Iverach, C. P., Kelly, B. F. J., Manning, M. R., Nisbet-Jones, P. B. R., Pyle, J. A., Townsend-Small, A., al-Shalaan, A., Warwick, N., and Zazzeri, G.: Methane mitigation:
methods to reduce emissions, on the path to the Paris agreement, Rev.
Geophys., 58, e2019RG000675, 10.1029/2019RG000675, 2020.Noël, S., Reuter, M., Buchwitz, M., Borchardt, J., Hilker, M., Schneising, O., Bovensmann, H., Burrows, J. P., Di Noia, A., Parker, R. J., Suto, H., Yoshida, Y., Buschmann, M., Deutscher, N. M., Feist, D. G., Griffith, D. W. T., Hase, F., Kivi, R., Liu, C., Morino, I., Notholt, J., Oh, Y.-S., Ohyama, H., Petri, C., Pollard, D. F., Rettinger, M., Roehl, C., Rousogenous, C., Sha, M. K., Shiomi, K., Strong, K., Sussmann, R., Té, Y., Velazco, V. A., Vrekoussis, M., and Warneke, T.: Retrieval of greenhouse gases from GOSAT and GOSAT-2 using the FOCAL algorithm, Atmos. Meas. Tech., 15, 3401–3437, 10.5194/amt-15-3401-2022, 2022.Omara, M., Zavala-Araiza, D., Lyon, D. R., Hmiel, B., Roberts, K. A., and
Hamburg, S. P.: Methane emissions from US low production oil and natural gas
well sites, Nat. Commun., 13, 2085,
10.1038/s41467-022-29709-3, 2022.Palmer, P. I., Feng, L., Lunt, M. F., Parker, R. J., Bosch, H., Lan, X.,
Lorente, A., and Borsdorff, T.: The added value of satellite observations of
methane for understanding the contemporary methane budget, Phil. Trans. R.
Soc. A, 379, 20210106, 10.1098/rsta.2021.0106,
2021.Pandey, S., Gautam, R., Houweling, S., Denier van der Gon, H., Sadavarte,
P., Borsdorff, T., Hasekamp, O., Landgraf, J., Tol, P., van Kempen, T., Hoogeveen, R., van Hees, R., Hamburg, S. P., Maasakkers, J. D., and Aben, I.: Satellite observations reveal extreme methane
leakage from a natural gas well blowout, P. Natl. Acad. Sci. USA, 116, 23676–23681, 10.1073/pnas.1908712116, 2019.Parker, R. J., Webb, A., Boesch, H., Somkuti, P., Barrio Guillo, R., Di Noia, A., Kalaitzi, N., Anand, J. S., Bergamaschi, P., Chevallier, F., Palmer, P. I., Feng, L., Deutscher, N. M., Feist, D. G., Griffith, D. W. T., Hase, F., Kivi, R., Morino, I., Notholt, J., Oh, Y.-S., Ohyama, H., Petri, C., Pollard, D. F., Roehl, C., Sha, M. K., Shiomi, K., Strong, K., Sussmann, R., Té, Y., Velazco, V. A., Warneke, T., Wennberg, P. O., and Wunch, D.: A decade of GOSAT Proxy satellite CH4 observations, Earth Syst. Sci. Data, 12, 3383–3412, 10.5194/essd-12-3383-2020, 2020.Pétron, G., Miller, B., Vaughn, B., Thorley, E., Kofler, J., Mielke-Maday, I., Sherwood, O., Dlugokencky, E., Hall, B., Schwietzke, S., Conley, S., Peischl, J., Lang, P., Moglia, E., Crotwell, M., Crotwell, A., Sweeney, C., Newberger, T., Wolter, S., Kitzis, D., Bianco, L.,
King, C., Coleman, T., White, A., Rhodes, M., Tans, P., and Schnell, R.: Investigating large methane
enhancements in the U.S. San Juan Basin, Elem. Sci. Anth, 8, 038,
10.1525/elementa.038, 2020.Plant, G., Kort, E. A., Murray, L. T., Maasakkers, J. D., and Aben, I.:
Evaluating urban methane emissions from space using TROPOMI methane and
carbon monoxide observations, Remote Sens. Environ., 268, 112756,
10.1016/j.rse.2021.112756, 2022.Prather, M. J., Holmes, C. D., and Hsu, J.: Reactive greenhouse gas
scenarios: Systematic exploration of uncertainties and the role of
atmospheric chemistry, Geophys. Res. Lett., 39, L09803,
10.1029/2012gl051440, 2012.Qu, Z., Jacob, D. J., Shen, L., Lu, X., Zhang, Y., Scarpelli, T. R., Nesser, H., Sulprizio, M. P., Maasakkers, J. D., Bloom, A. A., Worden, J. R., Parker, R. J., and Delgado, A. L.: Global distribution of methane emissions: a comparative inverse analysis of observations from the TROPOMI and GOSAT satellite instruments, Atmos. Chem. Phys., 21, 14159–14175, 10.5194/acp-21-14159-2021, 2021.Riris, H., Numata, K., Wu, S., and Fahey, M.: The challenges of measuring
methane from space with a LIDAR, CEAS Space J., 11, 475–483,
10.1007/s12567-019-00274-8, 2019.
Rodgers, C. D.: Inverse Methods for Atmospheric Sounding: Theory and
Practice, World Scientific, River Edge, USA, ISBN 978-9810227401, 2000.
Rohrschneider, R. R., Wofsy, S, Franklin, J. E., Benmergui, J., Soto, J.,
and Davis, S. B.: The MethaneSAT Mission, 35th Annual Small Satellite
Conference Proceedings, virtual, 7–12 August 2021, SSC21-II-05, 1–7, 2021.Sadavarte, P., Pandey, S., Maasakkers, J. D., Lorente, A., Gon, H. D., van
der Houweling, S., and Aben, I.: Methane emissions from super-emitting coal
mines in Australia quantified using TROPOMI satellite observations, Environ.
Sci. Technol., 55, 16573–16580, 10.1021/acs.est.1c03976,
2021.Sánchez-García, E., Gorroño, J., Irakulis-Loitxate, I., Varon, D. J., and Guanter, L.: Mapping methane plumes at very high spatial resolution with the WorldView-3 satellite, Atmos. Meas. Tech., 15, 1657–1674, 10.5194/amt-15-1657-2022, 2022.Saunois, M., Stavert, A. R., Poulter, B., Bousquet, P., Canadell, J. G., Jackson, R. B., Raymond, P. A., Dlugokencky, E. J., Houweling, S., Patra, P. K., Ciais, P., Arora, V. K., Bastviken, D., Bergamaschi, P., Blake, D. R., Brailsford, G., Bruhwiler, L., Carlson, K. M., Carrol, M., Castaldi, S., Chandra, N., Crevoisier, C., Crill, P. M., Covey, K., Curry, C. L., Etiope, G., Frankenberg, C., Gedney, N., Hegglin, M. I., Höglund-Isaksson, L., Hugelius, G., Ishizawa, M., Ito, A., Janssens-Maenhout, G., Jensen, K. M., Joos, F., Kleinen, T., Krummel, P. B., Langenfelds, R. L., Laruelle, G. G., Liu, L., Machida, T., Maksyutov, S., McDonald, K. C., McNorton, J., Miller, P. A., Melton, J. R., Morino, I., Müller, J., Murguia-Flores, F., Naik, V., Niwa, Y., Noce, S., O'Doherty, S., Parker, R. J., Peng, C., Peng, S., Peters, G. P., Prigent, C., Prinn, R., Ramonet, M., Regnier, P., Riley, W. J., Rosentreter, J. A., Segers, A., Simpson, I. J., Shi, H., Smith, S. J., Steele, L. P., Thornton, B. F., Tian, H., Tohjima, Y., Tubiello, F. N., Tsuruta, A., Viovy, N., Voulgarakis, A., Weber, T. S., van Weele, M., van der Werf, G. R., Weiss, R. F., Worthy, D., Wunch, D., Yin, Y., Yoshida, Y., Zhang, W., Zhang, Z., Zhao, Y., Zheng, B., Zhu, Q., Zhu, Q., and Zhuang, Q.: The Global Methane Budget 2000–2017, Earth Syst. Sci. Data, 12, 1561–1623, 10.5194/essd-12-1561-2020, 2020.Scarfutto, R. D. M., van der Werf, H., Bakker, W. H., van der Meer, F., and de
Souza Filho, C. R.: An evaluation of airborne SWIR imaging spectrometers for
CH4 mapping: Implications of band positioning, spectral sampling and
noise, Int. J. Appl. Earth Obs., 94, 102233,
10.1016/j.jag.2020.102233, 2021.Scarpelli, T. R., Jacob, D. J., Maasakkers, J. D., Sulprizio, M. P., Sheng, J.-X., Rose, K., Romeo, L., Worden, J. R., and Janssens-Maenhout, G.: A global gridded (0.1∘×0.1∘) inventory of methane emissions from oil, gas, and coal exploitation based on national reports to the United Nations Framework Convention on Climate Change, Earth Syst. Sci. Data, 12, 563–575, 10.5194/essd-12-563-2020, 2020.Schneising, O., Buchwitz, M., Reuter, M., Bovensmann, H., Burrows, J. P., Borsdorff, T., Deutscher, N. M., Feist, D. G., Griffith, D. W. T., Hase, F., Hermans, C., Iraci, L. T., Kivi, R., Landgraf, J., Morino, I., Notholt, J., Petri, C., Pollard, D. F., Roche, S., Shiomi, K., Strong, K., Sussmann, R., Velazco, V. A., Warneke, T., and Wunch, D.: A scientific algorithm to simultaneously retrieve carbon monoxide and methane from TROPOMI onboard Sentinel-5 Precursor, Atmos. Meas. Tech., 12, 6771–6802, 10.5194/amt-12-6771-2019, 2019.
Seinfeld, J. H and Pandis, S.: Atmospheric Chemistry and Physics, 3rd
edn., Wiley, ISBN 9781118947401, 2016.Shen, L., Zavala-Araiza, D., Gautam, R., Omara, M., Scarpelli, T., Sheng,
J., Sulprizio, M. P., Zhuang, J., Zhang, Y., Lu, X., Hamburg, S. P., and
Jacob, D. J.: Unravelling a large methane emission discrepancy in Mexico
using satellite observations, Remote Sens. Environ., 260, 112461,
10.1016/j.rse.2021.112461, 2021.Shen, L., Gautam, R., Omara, M., Zavala-Araiza, D., Maasakkers, J., Scarpelli, T., Lorente, A., Lyon, D., Sheng, J., Varon, D., Nesser, H., Qu, Z., Lu, X., Sulprizio, M., Hamburg, S., and Jacob, D.: Satellite quantification of oil and natural gas methane emissions in the US and Canada including contributions from individual basins, Atmos. Chem. Phys. Discuss. [preprint], 10.5194/acp-2022-155, in review, 2022.Sherwin, A. D., Rutherford, J. S., Chen, Y., Kort, E. A., Jackson, R. B., and
Brandt, A. R.: Single-blind validation of space-based point-source methane
emissions detection and quantification, EarthArXiv,
10.31223/X5DH09, 2022.Sierk, B., Bézy, J.-L., L¨scher, A., and Meijer, Y.: The
European CO2 Monitoring Mission: observing anthropogenic greenhouse gas
emissions from space, International Conference on Space
Optics – ICSO 2018, Platanias, Greece, 12 July 2019, Proc. SPIE 11180,
10.1117/12.2535941, 2019.Suto, H., Kataoka, F., Kikuchi, N., Knuteson, R. O., Butz, A., Haun, M., Buijs, H., Shiomi, K., Imai, H., and Kuze, A.: Thermal and near-infrared sensor for carbon observation Fourier transform spectrometer-2 (TANSO-FTS-2) on the Greenhouse gases Observing SATellite-2 (GOSAT-2) during its first year in orbit, Atmos. Meas. Tech., 14, 2013–2039, 10.5194/amt-14-2013-2021, 2021.Thompson, D. R., Leifer, I., Bovensmann, H., Eastwood, M., Fladeland, M., Frankenberg, C., Gerilowski, K., Green, R. O., Kratwurst, S., Krings, T., Luna, B., and Thorpe, A. K.: Real-time remote detection and measurement for airborne imaging spectroscopy: a case study with methane, Atmos. Meas. Tech., 8, 4383–4397, 10.5194/amt-8-4383-2015, 2015.Thompson, D. R., Thorpe, A. K., Frankenberg, C., Green, R. O., Duren, R.,
Guanter, L., Hollstein, A., Middleton, E., Ong, L., and Ungar, S.:
Space-based remote imaging spectroscopy of the Aliso Canyon CH4
superemitter, Geophys. Res. Lett., 43, 6571–6578,
10.1002/2016GL069079, 2016.Thorpe, A. K., Frankenberg, C., and Roberts, D. A.: Retrieval techniques for airborne imaging of methane concentrations using high spatial and moderate spectral resolution: application to AVIRIS, Atmos. Meas. Tech., 7, 491–506, 10.5194/amt-7-491-2014, 2014.Thorpe, A. K., Frankenberg, C., Thompson, D. R., Duren, R. M., Aubrey, A. D., Bue, B. D., Green, R. O., Gerilowski, K., Krings, T., Borchardt, J., Kort, E. A., Sweeney, C., Conley, S., Roberts, D. A., and Dennison, P. E.: Airborne DOAS retrievals of methane, carbon dioxide, and water vapor concentrations at high spatial resolution: application to AVIRIS-NG, Atmos. Meas. Tech., 10, 3833–3850, 10.5194/amt-10-3833-2017, 2017.Tu, Q., Hase, F., Schneider, M., García, O., Blumenstock, T., Borsdorff, T., Frey, M., Khosrawi, F., Lorente, A., Alberti, C., Bustos, J. J., Butz, A., Carreño, V., Cuevas, E., Curcoll, R., Diekmann, C. J., Dubravica, D., Ertl, B., Estruch, C., León-Luis, S. F., Marrero, C., Morgui, J.-A., Ramos, R., Scharun, C., Schneider, C., Sepúlveda, E., Toledano, C., and Torres, C.: Quantification of CH4 emissions from waste disposal sites near the city of Madrid using ground- and space-based observations of COCCON, TROPOMI and IASI, Atmos. Chem. Phys., 22, 295–317, 10.5194/acp-22-295-2022, 2022.Turner, A. J., Jacob, D. J., Wecht, K. J., Maasakkers, J. D., Lundgren, E., Andrews, A. E., Biraud, S. C., Boesch, H., Bowman, K. W., Deutscher, N. M., Dubey, M. K., Griffith, D. W. T., Hase, F., Kuze, A., Notholt, J., Ohyama, H., Parker, R., Payne, V. H., Sussmann, R., Sweeney, C., Velazco, V. A., Warneke, T., Wennberg, P. O., and Wunch, D.: Estimating global and North American methane emissions with high spatial resolution using GOSAT satellite data, Atmos. Chem. Phys., 15, 7049–7069, 10.5194/acp-15-7049-2015, 2015.
United Nations Environment Programme: An Eye on Methane: International
Methane Emissions Observatory, Nairobi, ISBN 978-92-807-3893-3, 2021.Van Damme, M., Clarisse, L., Whitburn, M., Hadji-Lazaro, J., Hurtmans, D.,
Clerbaux, C., and Coheur, P.-F.: Industrial and agricultural ammonia point
sources exposed, Nature, 564, 99–103,
10.1038/s41586-018-0747-1, 2018.Varon, D. J., Jacob, D. J., McKeever, J., Jervis, D., Durak, B. O. A., Xia, Y., and Huang, Y.: Quantifying methane point sources from fine-scale satellite observations of atmospheric methane plumes, Atmos. Meas. Tech., 11, 5673–5686, 10.5194/amt-11-5673-2018, 2018.Varon, D. J., McKeever, J., Jervis, D., Maasakkers, J. D., Pandey, S.,
Houweling, S., Aben, I., Scarpelli, T. R., and Jacob, D. J.: Satellite
discovery of anomalously large methane point sources from oil/gas
production, Geophys. Res. Lett., 46, 13507–13516, 10.1029/2019GL083798,
2019.Varon, D. J., Jacob, D. J., Jervis, D., and McKeever, J.: Quantifying
time-averaged methane emissions from individual coal mine vents with
GHGSat-D satellite observations, Environ. Sci. Technol., 54, 10246–10253,
10.1021/acs.est.0c01213, 2020.Varon, D. J., Jervis, D., McKeever, J., Spence, I., Gains, D., and Jacob, D. J.: High-frequency monitoring of anomalous methane point sources with multispectral Sentinel-2 satellite observations, Atmos. Meas. Tech., 14, 2771–2785, 10.5194/amt-14-2771-2021, 2021.Varon, D. J., Jacob, D. J., Sulprizio, M., Estrada, L., Downs, W. B., Shen, L., Hancock, S. E., Nesser, H., Qu, Z., Penn, E., Chen, Z., Lu, X., Lorente, A., Tewari, A., and Randles, C. A.: Integrated Methane Inversion (IMI 1.0): A user-friendly, cloud-based facility for inferring high-resolution methane emissions from TROPOMI satellite observations, Geosci. Model Dev. Discuss. [preprint], 10.5194/gmd-2022-45, in review, 2022.Vaughn, T. L., Bell. C. S., Pickering, C. K, and Nummedal, D.: Temporal
variability largely explains top-down/bottom-up difference in methane
emission estimates from a natural gas production region, P. Natl. Acad. Sci. USA, 115,
11712–11717, 10.1073/pnas.1805687115,
2018.Wang, F., Maksyutov, S., Tsuruta, A., Janardanan, R., Ito, A., Sasakawa, M., Machida, T., Morino, I., Yoshida, Y., Kaiser, J. W., Janssens-Maenhout, G., Dlugokencky, E. J., Mammarella, I., Lavric, J. V., and Matsunaga, T.: Methane emission estimates by
the global high-resolution inverse model using national inventories, Remote
Sens., 11, 2489, 10.3390/rs11212489, 2019.Wecht, K. J., Jacob, D. J., Frankenberg, C., Jiang, Z., and Blake, D. R.:
Mapping of North America methane emissions with high spatial resolution by
inversion of SCIAMACHY satellite data, J. Geophys. Res., 119, 7741–7756,
10.1002/2014JD021551, 2014.Western, L. M., Ramsden, A. E., Ganesan, A. L., Boesch, H., Parker, R. J.,
Scarpelli, T. R., Tunnicliffe, R. L., and Rigby, M.: Estimates of North
African methane emissions from 2010 to 2017 using GOSAT observations,
Environ. Sci. Technol. Lett., 8, 626–632,
10.1021/acs.estlett.1c00327, 2021.
White, W. H., Anderson, J. A., Blumenthal, D. L., Husar, R. B., Gillani, N. V.,
Husar, J. D., and Wilson, W. E.: Formation and transport of secondary air
pollutants: ozone and aerosols in the St. Louis urban plume, Science, 194,
187–189, 1976.Worden, J., Wecht, K., Frankenberg, C., Alvarado, M., Bowman, K., Kort, E.,
Kulawik, S., Lee, M., Payne, V., and Worden, H.: CH4 and CO distributions over tropical fires during October 2006 as observed by the Aura TES satellite instrument and modeled by GEOS-Chem, Atmos. Chem. Phys., 13, 3679–3692, 10.5194/acp-13-3679-2013, 2013.Worden, J. R., Turner, A. J., Bloom, A., Kulawik, S. S., Liu, J., Lee, M., Weidner, R., Bowman, K., Frankenberg, C., Parker, R., and Payne, V. H.: Quantifying lower tropospheric methane concentrations using GOSAT near-IR and TES thermal IR measurements, Atmos. Meas. Tech., 8, 3433–3445, 10.5194/amt-8-3433-2015, 2015.Worden, J. R., Cusworth, D. H., Qu, Z., Yin, Y., Zhang, Y., Bloom, A. A., Ma, S., Byrne, B. K., Scarpelli, T., Maasakkers, J. D., Crisp, D., Duren, R., and Jacob, D. J.: The 2019 methane budget and uncertainties at 1∘ resolution and each country through Bayesian integration Of GOSAT total column methane data and a priori inventory estimates, Atmos. Chem. Phys., 22, 6811–6841, 10.5194/acp-22-6811-2022, 2022.
Wunch, D., Toon, G. C., Blavier, J.-F. L., Washenfelder, R. A., Notholt, J.,
Connor, B. J., Griffith, D. W. T., Sherlock,V., and Wennberg, P. O.: The
Total Carbon Column Observing Network, Philos. T. R. Soc. A, 369,
2087–2112, 2011.Yang, S., Lan, X., Talbot, R., and Liu, T.: Characterizing anthropogenic
methane sources in the Houston and Barnett Shale areas of Texas using the
isotopic signature δ13C in CH4, Sci. Total. Environ., 696,
133856, 10.1016/j.scitotenv.2019.133856, 2019.Yin, Y., Chevallier, F., Ciais, P., Bousquet, P., Saunois, M., Zheng, B., Worden, J., Bloom, A. A., Parker, R. J., Jacob, D. J., Dlugokencky, E. J., and Frankenberg, C.: Accelerating methane growth rate from 2010 to 2017: leading contributions from the tropics and East Asia, Atmos. Chem. Phys., 21, 12631–12647, 10.5194/acp-21-12631-2021, 2021.Yu, X., Millet, D. B., and Henze, D. K.: How well can inverse analyses of high-resolution satellite data resolve heterogeneous methane fluxes? Observing system simulation experiments with the GEOS-Chem adjoint model (v35), Geosci. Model Dev., 14, 7775–7793, 10.5194/gmd-14-7775-2021, 2021.
Yuan, B., Kaser, L., Karl, T., Graus, M., Peischl, J., Campos, T. L., Shertz, S., Apel, E. C., Hornbrook, R. S., Hills, A., Gilman, J. B., Lerner, B. M., Warneke, C., Flocke, F. M., Ryerson, T. B., Guenther, A. B., and de Gouw, J. A.: Airborne flux measurements of methane
and volatile organic compounds over the Haynesville and Marcellus shale gas
production regions, J. Geophys. Res.-Atmos., 120, 6271–6289,
10.1002/2015JD023242, 2015.Zhang, Y., Gautam, R., Pandey, S., Omara, M., Maasakkers, J. D., Sadavarte,
P., Lyon, D., Nesser, H., Sulprizio, M. P., Varon, D. J., Zhang, R.,
Houweling, S., Zavala-Araiza, D., Alvarez, R. A., Lorente, A., Hamburg, S.
P., Aben, I., and Jacob, D. J.: Quantifying methane emissions from the
largest oil-producing basin in the United States from space, Sci. Adv., 6,
17, 10.1126/sciadv.aaz5120, 2020.Zhang, Y., Jacob, D. J., Lu, X., Maasakkers, J. D., Scarpelli, T. R., Sheng, J.-X., Shen, L., Qu, Z., Sulprizio, M. P., Chang, J., Bloom, A. A., Ma, S., Worden, J., Parker, R. J., and Boesch, H.: Attribution of the accelerating increase in atmospheric methane during 2010–2018 by inverse analysis of GOSAT observations, Atmos. Chem. Phys., 21, 3643–3666, 10.5194/acp-21-3643-2021, 2021.Zimmerle, D., Duggan, G., Vaughn, T., Bell, C., Lute, C., Bennett, K.,
Kimura, Y., Cardoso-Saldaña, F. J., and Allen, D. T.: Modeling air
emissions from complex facilities at detailed temporal and spatial
resolution: The Methane Emission Estimation Tool (MEET), Sci. Total
Environ., 824, 153653, 10.1016/j.scitotenv.2022.153653, 2022.