In the thermal infrared (wavelength greater than 4 μm), the radiance emitted by the atmosphere is very much a function of its temperature. On the other hand, it is also function of its composition.

The radiance I(λ) measured at the top of the atmosphere at wavelength λ can be written:

In Eq. (3) the first term is the surface contribution which is proportional to the surface emissivity ɛ(λ). The second term is the atmospheric contribution. It is the integral of the Planck function over the atmospheric column weighted by the derivative of the transmission. By selecting the wavelength, one can choose different levels of absorption, from transparent channels (τ(P_surf) ≈ 1 ) up to highly absorbing channels(τ(P_surf) ≈ 0 ). For absorbing channels, the weighting function in the atmospheric term is close to zero both in the boundary layer and in the highest atmospheric levels. There is a level in the atmosphere where it takes its maximum value. The channel “sees” preferentially this particular level. The weighting function ∂τ/∂p extends over several kilometres around the level where it takes its maximal value. By selecting different channels with different absorbing levels, one can probe different levels in the atmosphere.

For a given wavelength, the absorption K_P(λ,T,p) depends on the atmosphere composition (concentration of the absorbing gases), and on temperature and pressure. If the concentration of the absorbing gases is known, the radiance measurements may be used for an estimate of the atmospheric temperature profile. This measurement concept has been used for nearly three decades to retrieve atmospheric temperature profiles from the TOVS instrument package onboard the operational NOAA satellites. It is now used for the same purpose with the instruments IASI and AIRS. Note that some of the temperature profile estimates make use of CO₂ absorption bands and the retrieval technique assumes in that case that CO₂ concentration is known.

Conversely, if the atmospheric temperature profile is known, one may deduce from the measurements the absorbing gas concentration from the inferred K. Basically, as the gas concentration gets larger, the channel of interest sees higher in the atmosphere and, for a troposphere peaking channel, the measured radiance corresponds to a colder temperature. By adjusting the measured and modelled radiance using the known temperature profile, the absorbing gas concentration can be estimated. The main uncertainty is the knowledge on the temperature profile. Note also that, although some vertical profiling information can be obtained by using different channels, the vertical resolution is not better than a few kilometres. Finally, the technique usually requires channels with a significant absorption so that there is negligible contribution from the surface. As a consequence, the measurement sensitivity to the lower atmospheric layers is comparatively small (Engelen and Stephens, 2004).

In the emission spectroscopy technique, the wavelength range is in the thermal infrared. In the thermal infrared, the radiance emitted by the Earth (surface + atmosphere) is significantly larger than that emitted by the sun and reflected by the Earth. At shorter wavelengths (λ < 3 μm), the opposite is true. In this spectral range, there is no significant emission, neither by the surface nor by the atmosphere (we exclude here marginal cases of city light and fires). Earth observing satellite instruments measure the light emitted by the Sun and reflected either at the surface or through atmospheric scattering. The reflected light shows some spectral variations that reflect radiative transfer processes. For so-called “atmospheric window” channels, the atmosphere is nearly transparent and the radiance at the top of the atmosphere varies as a function of surface reflectance. On the other hand, at other wavelengths, the atmosphere is absorbing and the measurement is also a function of the amount of absorbing material in the atmosphere. Assuming an Earth sensing satellite, the measured radiance seen by this satellite is equal to (first order approximation):

The differential absorption technique requires clear sky along the line of sight. It also requires sunlight with a sun sufficiently high above the horizon to limit scattering in the atmosphere. These requirements limit the time-space sampling capabilities. In particular, no night measurement is possible, and high latitudes are not sampled during the winter season. This is a limitation to quantify processes such as winter ecosystem respiration, or CH₄ emissions by boreal and arctic regions. Another constraint is that the surface reflectance must be large enough as it has a direct impact on the signal-to-noise. Over land, natural surfaces are relatively dark at the wavelengths of interest, in particular when they are snow-covered. However, the problem is mostly over water that is extremely dark, and therefore not suitable for the measuring technique, except in the direction of the glint. As a consequence, measurements over the oceans are only possible when the instrument aims at the sun glint, which limits the spatial coverage and puts some constraints on the instrument design.

Active sensing is similar in its principle to differential absorption technique, replacing the sun by an artificial source. The emitted light is transmitted in the atmosphere and scattered back to the satellite either by surface reflectance or through scattering in the atmosphere. At least two spectrally close channels are needed to retrieve both the surface reflectance and the atmospheric transmission (related to the sought amount of absorbing material). There are two potential advantages to use an active source rather than the sun. One is the ability to measure everywhere at all season, when the passive method requires a high enough sun source. The other potential advantage, using the time resolved capabilities of the lidar, is to measure a vertical profile of the absorbing material. In practice, the high accuracy requirements for the monitoring of greenhouse gases, and the rather low atmospheric scattering at the wavelengths with significant absorption, does not allow such a vertical profile retrieval. Only the strong surface echo can be used for the differential absorption technique, and a total column can be retrieved with some weighting function. Nevertheless, the time-resolved capabilities of the lidar make it possible to differentiate the surface echo from the atmospheric contribution. The differential absorption technique can then be applied to the surface echo only, excluding the contribution of atmospheric scattering. This eliminates the largest error contributor to the solar spectroscopy method, which greatly affects passive instruments (Ehret et al., 2008).

The TOVS instrument has been onboard the NOAA operational meteorological platforms for almost 30 years. It includes a thermal infrared radiometer with 19 channels between 4 and 15 μm and a microwave radiometer. Some of the bands are affected by CO₂ absorption, while others are mostly affected by water vapour absorption. The TOVS mission has been highly successful and its observations have been used for operational weather predictions.

The TOVS observations have also been used for climate analysis and a large number of atmospheric variables have been monitored from the measurements. A group at the Laboratoire de météorologie dynamique (LMD), Palaiseau, France, has conducted pioneering work to estimate the CO₂ mixing ratio from a carefully selected set of TOVS channels (Chédin et al., 2003a) following a theoretical analysis that applied to both IASI and TOVS (Chédin et al., 2003b). Note that the weighting function of the TOVS products is rather high in the atmosphere. Most of the sensitivity is between 100 and 500 hPa. It was soon recognized that the retrieval of CO₂ is much easier in the tropics than in the mid- or high-latitude, both because of a less variable temperature profile and a thicker troposphere. The initial results shown in (Chédin et al., 2003a) compared favourably with high altitude measurements derived from commercial airliners and showed a concentration growth rate in agreement with that derived from surface observations.

An extensive evaluation was performed against both modelling results and available high altitude observations. The evaluation clearly demonstrated that “the large spatial variability of TOVS retrievals reflects substantial regional biases and noise which need to be reduced before remotely-sensed CO₂ from TOVS will help constrain our knowledge of the carbon cycle” (Peylin et al., 2007). The LMD group performed further research to identify the cause for the biases. It was eventually found that there is insufficient information in the TOVS channels to distinguish and correct the relative effects of ozone, water vapour and temperature profile. As a consequence, it became clear that the errors in the TOVS CO₂ estimates are too large to provide useful information (i.e., information that is not already in the 3D fields inferred from surface observations with models) on the CO₂ spatial distribution (Chevallier et al., 2007).

Nevertheless, there may be some interesting information in TOVS spectra relative to CO₂ fluxes. The monthly maps of night versus day XCO₂ differences show spatial and temporal patterns that are highly correlated with biomass burning activity (Chédin et al., 2008). This persistent diurnal cycle of XCO₂ above regions where biomass burning occurs is shown in Fig. 2. This observation led to the hypotheses that CO₂ from fires is rapidly uplifted to high altitude (i.e., above 5 km where TOVS has some sensitivity) during the afternoon and then transported away from the sources by strong high altitude winds. Because the biomass burning activity has a large amplitude diurnal cycle, this suite of processes may explain the night-day difference in the CO₂ maps. Therefore, it was hypothesized that the TOVS CO₂ night-day difference can be used as a proxy for biomass burning activity (Chédin et al., 2007). Across Africa, the TOVS CO₂ difference shows high spatial correlations with emissions independently developed from burned area (van der Werf et al., 2003). There is still debate in the scientific community on the process that generates the observed signal on the TOVS data. Indeed, the hypothesis that CO₂ from fires is rapidly lifted up to the higher troposphere is not supported by several observations, in particular the lidar profiles of aerosols that are emitted, together with CO₂, by the fires.

In summary, it is now clear that TOVS measurements cannot be used to infer sources and sinks of CO₂. It is still debated on whether the TOVS observations can provide quantitative information for carbon cycle studies, in particular for biomass burning activity.

The AIRS instrument is the first high-spectral-resolution infrared sounder developed by NASA in support of operational weather forecasting by the National Oceanic and Atmospheric Administration (NOAA). It was launched onboard the AQUA platform on May 4, 2002.

The heart of the instrument is a cooled (155 K) array grating spectrometer operating over the range of 3.7–15.4 μm at a spectral resolution (λ/Δλ) of 1200. The instrument concept requires no moving parts for spectral encoding and provides 2378 spectral samples, all measured simultaneously in time and space. AIRS views the ground through a cross track rotary scan mirror which provides ± 49.5° ground coverage along with views to on board spectral and radiometric calibration sources every 2.67 s scan cycle. The AIRS IR spatial resolution is 13.5 km from the nominal 705.3 km orbit.

AIRS provides daily near-global coverage both day and night. The primary objective of AIRS is meteorology with the retrieval of temperature and water vapour profiles. The goal for temperature profiles is a vertical resolution of 1 km and an accuracy of 1 K RMS. Water vapour profiles are expected with a resolution of 3 to 5%. These numbers apply to clear scene or significantly above the cloud layer.

In addition to using the AIRS data to generate relatively well established “core products”, i.e., temperature and moisture vertical profiles, surface temperature, and total column ozone, members of the AIRS science team are also working on a number of products where feasibility, achievable accuracy, precision, spatial and temporal coverage are not fully established. Products under development include carbon dioxide total column and methane concentration.

In the past year there have been several studies to assess whether AIRS measurements could be used to estimate the concentration of carbon dioxide (Chédin et al., 2003a; Engelen and Stephens, 2004). As said above, the primary variable that controls the radiance measured by the satellite is the temperature. Atmospheric CO₂ concentration does have an effect, but this effect is relatively small when accounting for the typical variability of both temperature and CO₂ concentration. Nevertheless, radiative transfer simulations studies indicate that some useful CO₂ signature may be obtained by combining the measurement of channels that peak at similar levels in the atmosphere, but sensitive to the concentration of different gases (such as CO₂ and oxygen). Because oxygen is a well-mixed gas, the relative concentration of CO₂ can be estimated. The CO₂ signal is still small with regards to the radiometer noise level so that considerable averaging may be needed to derive a useful concentration at the better than 1% level (Engelen and Stephens, 2004). Moreover, the radiative transfer simulations clearly indicate that the method is sensitive to concentrations above about 700 hPa. There is no sensitivity to the atmospheric boundary layer.

Strow and Hannon (2008) have confirmed that the increase of CO₂ mixing ratio in the atmosphere has a measurable effect on the globally averaged AIRS temperatures. Initial attempts to retrieve CO₂ monthly products from AIRS data followed the same approach as with TOVS, i.e., a neural network applied on carefully selected channels (Crevoisier et al., 2004). The retrievals were limited to the tropics for the same reason as given for TOVS above. Although the global CO₂ distributions appeared reasonable, it was rapidly realized that the retrieved spatial features were not solely due to CO₂ and that very significant biases affected the retrievals.

A group at ECMWF assimilated the AIRS radiance in the operational weather forecast system to retrieve the CO₂ mixing ratio of the upper troposphere (Engelen et al., 2009). The initial results were limited to the tropical zone, and the monthly mean random error was estimated to be 1%. The first comparison against results from several models (constrained by surface observations of CO₂) show model-observed differences that are larger than model–model differences, together with very intriguing feature in the observations (Tiwari et al., 2006).

Another group attempted another method to retrieve CO₂ from AIRS (Chahine et al., 2005). It was shown that the retrieved CO₂ seasonal cycle was in agreement with the result of aircraft observation, although with significant noise. The same group produced global maps of retrieved CO₂ that differed significantly from those derived from best knowledge of surface flux and atmospheric transport models (Chahine et al., 2008) (Fig. 3). These differences were attributed to error in the atmospheric transport simulation. At present, consensus in the community is rather that biases remain in the retrieval, in particular due to undetected clouds.

There is a large choice in the selection of the AIRS channels that may be used to estimate the CO₂ weighted column. All choices that have been proposed so far lead to a sensitivity limited to the upper troposphere. However, Strow and Hannon (2008) suggested a combination of channels that led to a sensitivity lower in the atmosphere, with a limitation to oceanic sampling for a lower variability of surface emissivity. The analysis was performed over large ocean basins or zonal means and demonstrated an accuracy close to 1 ppm at this scale. Spatial averaging is needed to get a sufficient number of clear observations and to statistically decrease the noise in the retrievals.

The IASI instrument has been launched by ESA on the Metop platform. It provides high spectral resolution in the thermal infrared, similar to that of AIRS, and also shares with this instrument the main objective of estimating temperature and water vapor profiles for weather forecast purposes. Similarly to AIRS, it may also be used for climate purposes and for the estimate of atmospheric compositions. There is no reason to believe, however, that IASI can do much better than AIRS. To our knowledge, there has not been any successful attempt for the estimate of carbon dioxide or methane from IASI measurements.

In summary, one shall recall that the AIRS instrument was launched more than 7 years ago with a focus on temperature retrievals. There have been several attempts at estimating CO₂ from its measurements by groups in Europe and the US. Although it has been demonstrated that there is a CO₂ signal in the data, the errors in the retrieval are still too large to provide novel information on surface CO₂ fluxes (Chevallier et al., 2005). This disappointing result derives from the relatively high weighting function of the retrievals. In the upper troposphere, the CO₂ signal is dominated by the seasonal cycle and excursions around this cycle are limited to one or a few ppm. To provide useful information, the accuracy requirements are then even stronger than for an instrument that is sensitive to the low atmospheric levels. Such requirements have not yet been met and it is not clear whether they can be met. In addition, the retrievals appear to be affected by regional biases, which have a devastating effect on the estimates of surface fluxes. It is likely that retrievals of CO₂ from thermal infrared instruments such as AIRS or IASI will mostly be used to identify errors in the transport models rather than be used to constrain sources and sinks.

The SCIAMACHY instrument was launched onboard the Envisat spacecraft on 28th February 2002. Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY) is a spectrometer designed to measure sunlight, transmitted, reflected and scattered by the earth atmosphere or surface in the ultraviolet, visible and near infrared wavelength region (240–2380 nm) at moderate spectral resolution (0.2–1.5 nm). The absorption, reflection and scattering characteristics of the atmosphere are determined by measuring the extraterrestrial solar irradiance and the upwelling radiance observed in different viewing geometries.

SCIAMACHY has several modes of operation: Nadir, Limb, and occultation. In Nadir mode the atmospheric volume under the spacecraft and across-track is observed. Each scan covers an area on the ground of up to 960 km across track with a maximum resolution of 26 km × 15 km, albeit with resolution of typically 30 × 60 km for the channels sensitive to greenhouse gases. In Limb mode the instrument looks at the edge of the atmosphere. One novel characteristics of SCIAMACHY is the possibility to observe the same atmospheric volume first in limb and then, after about 7 min, in nadir geometry. In practice, however, no result concerning CO₂ and exploiting either the dual-view or the occultation technique has been published. On the other hand, there are many ongoing studies and published results that use the “nadir” mode of SCIAMACHY to estimate atmospheric concentration of CO₂ and CH₄. These are discussed below.

Due to its near-infrared nadir observation capability, SCIAMACHY is the first satellite instrument that is sensitive to CO₂ in the boundary layer where most variation occurs (Buchwitz et al., 2005). Therefore, SCIAMACHY CO₂ retrievals play a pioneering role and required the development of dedicated retrieval algorithms. The retrieval schemes adjust the measured radiance around 1.6 μm to the result of radiative transfer simulations, accounting for a number of atmospheric and surface parameters. The CO₂ mixing ratio that is retrieved is an averaged over the full column, with a weighting function that is close to 1 throughout the vertical profile. Note however that the measured spectra are sensitive, not so much to the CO₂ mixing ratio, but rather to the CO₂ column (in terms of mass or number of molecules). To derive a mixing ratio, it is therefore necessary to normalize the column using either the simultaneously measured O₂ column (using the A band at 765 nm) or a knowledge of the surface pressure derived from numerical weather analysis.

First attempts at deriving the CO₂ column from SCIAMACHY showed the possibility to observe the hemispheric gradient of atmospheric CO₂ (Buchwitz et al., 2005), although not with the expected accuracy. These first analyses rapidly showed the strong impact of atmospheric aerosols (Aben et al., 2007; Houweling et al., 2005). Several groups have taken up the challenge however, and the quality of SCIAMACHY retrievals has been significantly improved (Barkley et al., 2006a). This was due in part to a better understanding of the radiometer systematic measurement errors (Buchwitz et al., 2006) and to the use of accurate temperature profiles in the inversion process.

The global average concentration derived from SCIAMACHY is in reasonable agreement with the same quantity constrained by surface observations (Buchwitz et al., 2007). Near-global monthly maps of CO₂ column concentration have been derived from SCIAMACHY and have been compared to AIRS results (Barkley et al., 2006a), modelling results (Barkley et al., 2006b; Schneising et al., 2008) or column measurements provided by a sun-looking FTIR (Barkley et al., 2006b; Buchwitz et al., 2006; Schneising et al., 2008). In general, there is a good correlation between the satellite measurements and the validation products. There is also a rather large scatter of up to 10 ppm. Besides, the seasonal cycle of the satellite product in the Northern Hemisphere appears significantly larger than what can be expected from the numerous surface observations and a few airborne profiles (Barkley et al., 2006b; Buchwitz et al., 2007). There is currently a controversy in the carbon cycle community on whether the unexpected CO₂ features monitored by SCIAMACHY are real or the result of a bias in the data processing. At present, the data appears too different from the modelling outputs to be used in an inversion scheme for the retrieval of CO₂ surface fluxes. Nevertheless, SCIAMACHY was seen as a great tool to validate the OCO/GOSAT concept and to prepare the processing and analysis of these missions (Bosch et al., 2006).

The challenge of methane retrievals is somewhat easier as the spatial and temporal gradients generated by surface sources are, relative to the background concentration, larger than for CO₂. Similar techniques as those described above have been applied and maps of column CH₄ estimates have been produced, an example of which is shown in Fig. 4. The initial results have shown tropical concentrations much larger than expected (Frankenberg et al., 2005) which was tentatively explained by a yet unidentified source of methane produced by plants (Houweling et al., 2006). It was soon found, however, that the large tropical CH₄ excess was due to inaccuracies in the spectroscopic database used to interpret the measurements in terms of concentrations (Frankenberg et al., 2008). Despite the corrections, the CH₄ concentration maps are still rather different than those expected from the current knowledge of methane sources, chemistry and transport modelling (Meirink et al., 2008). It is still not clear whether this results from retrieval errors or from misunderstanding in the cycle of methane that could be resolved using remotely sensed data. Comparison with atmospheric transport simulations has shown biases in the retrievals, but also the potential of the product to identify large sources (Bergamaschi et al., 2007).

Note also that SCIAMACHY, even with continuous improvements in the data processing, will always have important limitations for the monitoring of CO₂ or CH₄ from space. One is due to the very large size of the field of view, so that most observations are cloud contaminated. Another arises from its spectral resolution that is insufficient to identify individual absorption lines. Finally, the viewing geometry is cross-track. This is well suited over land surfaces and actually provides better coverage than pure nadir instruments do. On the other hand, this geometry is rarely favourable over the oceans so that the monitoring of CO₂ is, for the most part, limited to land areas. These limitations (FOV size, spectral resolution, viewing geometry) are corrected for the forthcoming CO₂ dedicated missions.

In conclusion, the SCIAMACHY instrument allows a proof of concept for the monitoring of CO₂ and CH₄ from space using the differential absorption technique. The quality of the retrievals has been very much improved since the first versions of the processing algorithms and some features of the CO₂ and CH₄ spatial and temporal distributions can be measured. There is still no consensus on the value of SCIAMACHY data for an improved knowledge of CO₂ and CH₄ atmospheric concentrations, or to better constrain their surface fluxes. The instrument was not optimized for the monitoring of these gases and it is reasonable to expect significantly better results with the forthcoming missions that are dedicated to the monitoring of atmospheric concentration from space.

The GOSAT mission is developed as a cooperation between JAXA, the Japanese space agency and NIES, the National Institute for Environmental Studies. It is designed for the accurate measurement of CO₂ and CH₄ columns using the solar spectroscopy technique. GOSAT also carries a thermal infrared sensor, but the main products of the mission are expected from the Fourier Transform Spectrometer that operates in the solar spectrum. The satellite was launched in January 2009 and the first spectra acquired soon after. It is therefore the first satellite mission to be dedicated to greenhouse gases monitoring. At the time of writing, a few spectra and preliminary retrievals have been released. The spectra demonstrate that the instrument is functioning as anticipated. However, they are not yet properly calibrated. The CO₂ or CH₄ partial column are retrieved from the depth of their absorption lines in spectral bands around 1.6 and 2 μm while an oxygen absorption band at 0.76 μm is also used for reference and normalization: Since oxygen is a well-mixed gas in the atmosphere, the depth of the oxygen absorption lines in the spectra provides a reference indicative of the sunlight atmospheric path.

The instrument has the capability to point on specific targets, but the normal mode of operation is cross-track scanning with up to 9 samples perpendicular to the satellite subtrack. However, as the number of cross-track samples increases, the signal-to-noise decreases so that it may be necessary to limit the number of cross-track observations. In addition, there is a limitation over the oceans as water is dark at the wavelengths of interest, except in the glint direction. Only a fraction of GOSAT observations are acquired in the favourable “sunglint” viewing geometry. This geometry constrain will strongly limit the density of valid observations over the oceans.

At the time of writing, there is no product that can be used for an estimate of their accuracy. One has therefore to rely on theoretical estimates. The signal-to-noise estimates vary as a function of solar angle and surface albedo. They lead to errors that vary between 0.5 and 1% depending on the scene. The usefulness of GOSAT products to better constrain sources of CO₂ and CH₄ will depend on the error statistics. For CO₂, there is a large difference in the potential impact of the product whether a 0.5% (≈ 1.5 ppm) or a 1% (≈ 3 ppm) accuracy is reached. In addition, the results will very much depend on the regional biases: A fully random error is acceptable as spatial and temporal averaging decreases its amplitude, while a regional bias of a fraction of a % may have a large impact on the flux distribution estimated. It will probably take several years to properly quantify the accuracy of GOSAT product and assess their usefulness for the carbon cycle community.

In the framework of the NASA ESSP program, the OCO mission was selected in July 2002. The launch occurred on 24 February 2009 but the rocket's payload fairing failed to separate properly, which prevented the satellite to reach orbit. The OCO instrument aimed at measuring carbon dioxide columns with the differential absorption technique, i.e., a technique similar to that of SCIAMACHY or GOSAT (Crisp et al., 2004; Connor et al., 2008). The stated accuracy objective was 1 ppm averaged concentration and numerical simulations indicated that an even better accuracy could be achieved in favourable conditions (Bosch et al., 2006; Connor et al., 2008). The instrument was designed to make very high spectral resolution measurements close to 1.6 μm and 2 μm, two spectral bands that show a few narrow CO₂ absorption lines, and that are free from other gases absorption, and at 0.76 μm that corresponds to an oxygen absorption band. The latter is used for correction of surface pressure and atmospheric scattering. The viewing geometry mode of OCO was very innovative with three modes of operation: Nadir, glint and target. In the nadir mode, the instrument looks straight down. In the glint mode, the instrument follows the glint direction. This is necessary over the ocean to get a sizeable reflectance from the surface. Finally, the instrument may aim at specific targets. The later mode is designed for the calibration and validation of the CO₂ products based on coincident observation with ground-based FTIR. In “nadir” mode, the measurement track is one-dimensional so that it seldom samples a ground truth point. The agile capability makes it possible to view a calibration site (target) almost everyday (weather permitting). The ground calibration and validation can partly be achieved through the measurement of a spectra derived from a Fourier Transform Interferometer (FTIR). At present, there is a global network of 10 ground based FTIR spectrometers (www.tccon.caltech.edu). These instruments were shown to accurately measure column CO₂ and CH₄ concentrations, when compared with aircraft vertical profiles on a campaign basis (Dufour et al., 2004; Washenfelder et al., 2006).

At the time or writing, there are discussions at NASA whether a copy of the original OCO mission should be built rapidly or whether it should focus on the next generation of CO₂ monitoring instruments, based on the active technique.

Two active missions have been proposed for the monitoring of CO₂ from space. The A-SCOPE mission was considered by the European Space Agency in the Earth Explorer context, while Active Sensing of CO₂ Emissions over Nights, Days, and Seasons (ASCENDS) is currently considered by NASA. The A-SCOPE concept was evaluated as too demanding for a launch in 2016 so that it did not pass the next round of selection that took place in January 2009. On the other hand, an airborne version of the ASCENDS concept has been built. The flights that took place confirm the potential concept and the theoretical considerations that ASCENDS meets the science requirement of 0.5% precision measurement of CO₂ mixing ratio along a 100 km horizontal segment. Nevertheless, it is clear that the active mission is much more demanding than a passive one. Besides, the accuracy of the CO₂ column estimates is not necessarily better than that of the passive missions using the differential absorption technique.