About the AIRS Carbon Dioxide Product

AIRS provides satellite retrieval of mid-tropospheric carbon dioxide both day and night under clear and cloudy conditions over ocean and land, without the use of a priori information from models. AIRS retrievals are based on cloud-cleared thermal infrared radiance spectra in the 15 micron band and associated Level 2 geophysical profiles of temperature, water vapor and ozone, and achieves a 2 ppm accuracy in the tropics and mid-latitudes. AIRS measures the concentration of carbon dioxide (ppmv) with peak sensitivity at the 400 hPa pressure level at nadir resolution of 90 km x 90 km. The AIRS broad swath makes it able to map the global distribution of carbon dioxide every day.

The high spectral resolution and stability of AIRS allows a measurement accuracy between 1.5 ppm and 2 ppm, making it ideal for mapping the distribution and transport of carbon dioxide levels in the free troposphere.


Version 5 Level 3 carbon dioxide data have 2° latitude x 2.5° longitude grid boxes (dimensions of those files are 91° latitude x 144° longitude). Please note the future Version 6 carbon dioxide Level 3 files will have the same 1°x1° gridding scheme as other AIRS Level 3 products. 

Version 6 Level 2 carbon dioxide data will have different spatial resolution than other AIRS Level 2 products. The spatial resolution of the AIRS carbon dioxide Level 2 products is ~ 90 km x 90 km, so files will have dimensions of 15 x 22 (see explanation below).

Dimensions of Level 2 carbon dioxide data files

Carbon dioxide is reported on a 2x2 cluster of AMSU retrievals, and a standard AIRS Level 2 granule is made up of 45 scan lines, each containing 30 AMSU field-of-views (FOVs).

Each AIRS Level 2 granule contains a theoretical maximum number of retrievals equal to 45 x 30 = 1350. Thus the carbon dioxide "granule" will be comprised of 22 scanlines, each containing 15 clusters for a theoretical maximum of 22 x 15 = 330 retrievals per cluster.

We always skip the 45th AIRS Level 2 scan line of AMSU FOVs because incorporating it would require that we add the complexity of processing more than one AIRS granule at a time, and the indices of the clusters wouid then creep along as we move from one granule to the next.  

Therefore we process scanlines 1 & 2 together, then scanlines 3 & 4, then 5 & 6, all the way to 43 & 44, then skip 45 and in the next granule we are back to processing scanlines 1 & 2 together.

Brief Q&A with AIRS carbon dioxide lead Dr. Edward Olsen

Q: I'd love to speak with someone about how carbon dioxide data are being used, and any challenges to interpreting it and any effort to improve its collection or analysis. I'm also interested in hearing how these measurements over time are informing climate modeling.

Dr. Edward Olsen, JPL
Dr. Edward Olsen, JPL
A: So much has been published, most of it highly technical.  A report published in 1979 from Woods Hole (Charney Report) succinctly lays out the issues that have been the subject of intense research and modeling in the intervening years. It is an old, but good introduction.  Modeling has improved in the intervening 35 years, and the sensitivity of climate to the doubling of carbon dioxide is 3 °C +/- 1.5 °C.  There is still a wide disparity of predictions by the models, but the physics of propagation of radiation through the atmosphere is well understood.  The major disparity lies in the modeling of the response of the Earth (i.e., change in cloud cover) as well as the coupling to the deep ocean.

The atmosphere is the medium by which carbon dioxide is transported over the globe. Carbon dioxide is inert in the atmosphere, so it is a very good trace gas for studying the atmospheric circulation. All sources and sinks are at the surface, albeit there is minor destruction of carbon dioxide by ultraviolet radiation (photolysis) high above the stratosphere.  Most anthropogenic carbon dioxide is emitted in the northern hemisphere, and only about half of it remains in the atmosphere (source of the approximately 2 ppm per year increase). The majority of that removed is believed to be absorbed in the cold southern hemisphere oceans.  Thus the processes by which carbon dioxide is transported in the troposphere are an ongoing research effort that AIRS carbon dioxide data are well-suited to facilitate. Mapping the net flux from the surface requires sensitivity nearer to the surface than AIRS provides, and that is the goal of the Orbiting Carbon Observatory-2.

Small changes in climate, like rainfall patterns, have historically had enormous impact on human civilization. We have benefitted from having a relatively stable climate since the last ice age. The climate has undergone major changes in the past, and we are attempting to understand the climate processes and what impact we may have on them.