Particulate air pollution is a serious health concern, especially the particles small enough to remain in the air as it reaches the deepest, smallest areas in the lungs. These particles can cause health problems by themselves, and can act as "delivery devices" for adsorbed and far more dangerous chemicals from combustion (e.g. fossil-fuel burning engines, natural and human-induced fires) and other industrial processes.

The parameter PM2.5 is meant as a measure of these particles. The 2.5 corresponds to 2.5 microns, but it's not a hard cut-off, and actual particle size is less important than aerodynamic diameter. See https://diamondenv.wordpress.com/2010/12/10/particulate-pollution-pm10-and-pm2-5/

Local PM2.5 measurements are usually done using scatterometry; an airflow is directed past a laser beam and tiny pulses from photodetectors due to scattered light transients are analyzed and then counted. Size distribution is roughly inferred from the pulse height distribution.

Question: However, does NASA really measure ground-level PM2.5 concentrations from space, or anything even close to it? There are aerosols throughout the atmosphere, how could a satellite observation distinguish soot near ground level from small water droplets in the upper atmosphere?

below: From this 2014 NASAEarth tweet.

NASA PM2.5 map from a tweet

update: "This just in" at Earthobservatory.nasa.gov: Just Another Day on Aerosol Earth. Interesting article with links. Image has been downsized to 1 MB, (click for larger view)

caption: "A composite image using NASA satellite imagery shows the different kinds of aerosols in the earth's atmosphere."

Just Another Day on Aerosol Earth


1 Answer 1


You can detect such small particles by their absorption of infrared light. The quantity measured is called Aerosol Optical Depth (AOD). By observing the spectral distribution of light received by an satellite one can distinguish between various contributing factors: surface temperature (derived from the total distribution of infrared), cloud ceiling height, size of water droplets and dust or the amount of photosynthesis happening.

This measurement can be done in several ways: A calibrated light source is used, and the amount of light reaching a distant sensor is measured. The sensor could also be close to the light source and measure how much light is reflected (scattered, to be more precise) by the medium. The third option, used for satellite measurements is using the Earth and atmosphere itself as light source - this is simple as any object emits heat radiation, which at "normal" temperatures happens to be mostly in the infrared range.

One of these systems is MODIS aboard the TERRA and AQUA satellites. It scans the light coming from Earth in 36 spectral bands. Naturally, there is no direct measurement of particles in the atmosphere, but the combined data from all these channels provides a comprehensive image of all the components in the atmosphere down to the surface (in cloudless conditions) that can be fitted by simulations fitted to models based on laboratory- and Earth-based measurements of these factors.

As your question implies, you can't measure dust particles in the lower atmosphere under a cloud cover. Images containing clouds have to be carefully removed from the dataset during analysis. A study of the effects of clouds and vapor on the measurement is found in Relationship between Aerosol Optical Depth and Particulate Matter over Singapore: Effects of Aerosol Vertical Distributions

If you have access to it, there is also a closed-access paper giving a summary on these studies "Toward the next generation of air quality monitoring: Particulate Matter" by Engel-Cox et al.

  • $\begingroup$ Optical absorption is usually measured between a light source and a light detector. How is atmospheric absorption measured by a satellite? $\endgroup$
    – uhoh
    Aug 25, 2018 at 11:00
  • 1
    $\begingroup$ Any object at a "reasonable" temperature is a good light source in the infrared range between 1 and 10 µm due to thermal radiation. As the exact wavelength distribution of any thermal radiator is well known, one can calculate the absorption in any of the bands measured. $\endgroup$
    – asdfex
    Aug 25, 2018 at 11:06
  • $\begingroup$ Great! Thanks for the edit! Dalhousie University has an accessible copy: fizz.phys.dal.ca/~atmos/publications/Cox.pdf $\endgroup$
    – uhoh
    Aug 26, 2018 at 4:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.