Yellowstone Caldera Chronicles is a weekly column written by scientists and collaborators of the Yellowstone Volcano Observatory. This week's contribution is from R. Greg Vaughan, research scientist with the U.S. Geological Survey.

Sources/Usage: Public Domain. View Media Details

When interpreting thermal infrared satellite images, context is important—the season, time of day, vegetation, ground characteristics, and other factors all contribute to the information contained in these images.  The methods used to process the data are also important, just like the methods used to develop photographs on film ultimately control how the photo looks.  So, how do we process thermal infrared satellite images to allow us to better understand Yellowstone’s heat emissions? 

First, some background information. Every object that has a temperature emits energy (electromagnetic radiation) into its surroundings in all directions.  The characteristics of this emitted radiation are a function of the object's emissivity, which is a measure of how efficiently the energy is radiated, and temperature.  The thermal infrared radiation that we use to study Yellowstone has wavelengths ranging from about 7.5 to 13 microns, which is longer than the visible light radiation that we can see (0.4 to 0.7 microns).  

So, with thermal infrared imagery we are not directly measuring temperature.  Instead, we are measuring a flow of energy per unit time, but only at certain wavelengths, and only the energy coming from a certain area on the surface that is detected by the optics of the instrument—a quantity called spectral radiance.  Using a principle of physics called Planck’s Radiation Law, which describes the relation between an object’s temperature, emissivity, and emitted radiance at different wavelengths, we can use the spectral radiance information in a thermal infrared image to calculate a temperature for each pixel in an image. 

That might seem straightforward, but the natural world is pretty complicated—no materials emit energy with perfect efficiency at all wavelengths.  In reality, the emissivity of a material affects how much energy is radiated, and emissivity can vary with wavelength.  Most materials that cover Earth’s surface, like rocks, soils, water, and vegetation are not that reflective in the thermal infrared and have relatively high emissivities. 

Fortunately, there is a huge body of research that has determined the spectral emissivity values for a wide range of natural materials—essentially a spectral library!—and satellite datasets have been used to create a global map of spectral emissivity.  So, when collecting thermal infrared radiance data from Yellowstone, we have good estimates of surface emissivity that we can use to retrieve surface temperatures from the data.

This is where the image processing comes in to play.  Image processing is the process of performing mathematical operations on images to achieve a certain visual effect, like enhancing the contrast between parts of an image, or extracting some type of information, like pixel temperature values.  Satellite data have pixel values stored as simple numbers that don’t have an obvious meaning, but there are equations that can be used to convert these numbers to, in the case of thermal infrared data, spectral radiance values.  Once you have spectral radiance values for each pixel, and you know the wavelength of the image data, you can assume an emissivity and use Planck’s Radiation Law to calculate the temperature of each pixel in a thermal infrared image. 

Sources/Usage: Public Domain. View Media Details

With these temperature data, we can learn quite a lot about Yellowstone and other volcanic areas.  For example, heat output—either in terms of increased temperatures or increased areas showing slightly elevated temperatures—could be a sign of volcanic unrest, so determining the thermal output of a volcanic area over time can provide an indication of the potential for future eruption.  In the case of Yellowstone, it is possible to measure the thermal output of the entire system, as well as individual geyser basins, using satellite data—these are the data that helped YVO scientists discover the new thermal area near Tern Lake in 2019.  These changes can be small, but can still be seen from space thanks to advanced thermal infrared sensors and knowledge of a little bit of physics!