About the Lesson

This lesson introduces you to seven of the ten infrared imager bands on the GOES R-U ABI (Advanced Baseline Imager). By exploring the spectral characteristics of each band, you will gain a better understanding of band selection and what each band observes, and begin to appreciate some of the many potential applications.

Please note that information on the three water vapor bands not included here can be found in the IR Water Vapor Bands lesson.

On completing the lesson, you should be able to:

• Identify the various phenomena that can be detected in the seven ABI infrared bands covered in this lesson
• Identify sun glint in shortwave infrared imagery
• Identify which band is sensitive to hot spots and daytime aerosols
• Describe what is meant by "clean" vs. "dirty" infrared window bands
• Identify which band is most sensitive to stratospheric ozone
• Identify which bands are sensitive to emissions of volcanic sulfur dioxide
• Identify the bands useful for determining cloud properties

Please note that this lesson was originally published prior to the launch of GOES-16 and used imagery from the Himawari-8 geostationary satellite to illustrate GOES-R ABI capabilities. The Himawari-8 imager, the AHI (or Advanced Himawari Imager), has bands nearly identical to the ABI on board the GOES R to U series. In the interest of time, some of the original Himawari-8 imagery is still included where geographical context is not critical for understanding the concept.

Introduction

The earth and atmosphere emit energy primarily in the infrared at wavelengths greater than about 3 micrometers and peaking near 10 micrometers. The Advanced Baseline Imager's (ABI) infrared bands are in the 3 to 14 micrometer region of the infrared spectrum where much of this peak emission occurs.

 

Atmospheric gases absorb selectively in the longwave infrared, particularly water vapor between 5 and 8 micrometers, ozone between 9.2 and 9.8 micrometers, and carbon dioxide at wavelengths greater than 13 micrometers. Outside these major gas absorption regions, the clear atmosphere is highly transparent to longwave infrared radiation. Window regions are denoted "clean" when very little atmospheric absorption is present, and "dirty" when there is some absorption by gases, but not enough to prevent energy from the surface and clouds from reaching the satellite.

 

The amount of energy emitted depends on the temperature and emissivity (emission efficiency) of the earth's surface, clouds, aerosols and gases.

 

Atmospheric gases absorb and re-emit infrared energy at discrete wavelengths. The jagged yellow earth emitted radiance curve shown in the graphic, varies with wavelength due to absorption and emission by gases at different pressure levels. We'll explore this further as we step through the individual bands.

 

When superimposing the various infrared band locations, we see that some are located in infrared "window" regions where we see Earth's surface and clouds without a significant contribution from atmospheric gases. Other bands are sensitive to absorption by gases, with the degree of sensitivity determining the relative mix of atmosphere, clouds, and surface we'll be able to see. As a general rule, the greater the band's sensitivity to a gas, the higher the altitude the infrared signal originates from, and the less emission we'll see from the surface and lower atmospheric layers, including cloud cover.

 

Another convenient way to show the impact of selective absorption on the ABI's infrared bands is to plot total atmospheric transmittance as a function of wavelength. In the graph below, the gray jagged background curve is atmospheric transmittance shown for a representative atmosphere. When transmittance approaches one (or 100%), most Earth-emitted energy reaches space. And when transmittance dips to smaller values, that implies more absorption by gases and therefore less energy reaching space. Zero transmittance would be an atmosphere that's totally opaque and very cold from a satellite's vantage point. The blue shaded curves represent the instrument response functions for ABI's IR window bands and show the spectral interval covered by each band.

 

Now let's look at each of the seven infrared bands covered in this lesson beginning with the 3.9 "shortwave IR window" band.

Shortwave IR Window Band

The shortwave IR window band at 3.9 micrometers has demonstrated its usefulness for many applications on board both geostationary and polar-orbiting meteorological satellites since the late 1970s. Some of these applications include:

• An improved ability to identify and monitor the evolution of fog and low clouds
• Cloud phase determination
• Delineating between supercooled water and ice clouds
• Locating water clouds over snow during daytime
• Detecting hot spots such as fires and volcanoes
• Monitoring atmospheric aerosols and volcanic ash during daytime
• Sea-surface temperature measurements at night, and
• Estimating low-level cloud motion winds

The chart shown here summarizes three radiative transfer concepts that are key for exploiting the many capabilities of the shortwave IR band. We'll touch briefly on each concept in this section.

 

The shortwave IR band is located in the shortwave window region of Earth's emitted spectrum. Radiation in this region is not significantly absorbed by Earth's atmosphere. There is also much less earth-emitted energy in this wavelength region than at longer wavelengths.

 

 

The spectral region covered by the shortwave IR band also includes reflected solar radiation. In the zoomed in plot, we see that there is an overlap in the region covered by the shortwave IR band. The addition of reflected solar radiation from clouds and land produces a noticeable warming effect when viewing daytime shortwave IR imagery.

 

Large differences in the scattering (reflection) properties of water and ice particles in the shortwave infrared allow the shortwave IR band to discriminate cloud phase and detect snow cover during daytime.

The graph below shows that reflection at 3.9 micrometers by a cloud, is a function of both cloud phase and droplet size. Droplet radii in water clouds range between 5 and 20 micrometers, while ice particles in cirrus clouds are often 50 micrometers or larger. Therefore, water clouds are much more reflective (warmer) than ice clouds at 3.9 micrometers during daytime.

 

During nighttime, the emissivity (or emission efficiency) of water clouds at 3.9 micrometers is less than at longer infrared wavelengths. This results in colder brightness temperatures for water clouds in the shortwave IR band than in the longwave IR bands. For relatively thin cirrus however, the reverse is true.

Notice in the image comparison that thin cirrus appear warmer while lower level liquid water cloud tops appear cooler.

 

The shortwave and longwave IR bands have long been used in combination, typically by employing a difference technique to produce a nighttime image product that delineates ice and water clouds.

 

Detection of fires and hot spots is another special capability of the shortwave IR band. A band near 3.9 micrometers is much more sensitive in terms of the energy detected by the satellite when the temperature of an object increases, than a band centered near 11 micrometers. This effect is noticeable when an image pixel is partially filled by a hot spot or fire.

A plot for brightness temperatures for subpixel fires is shown in the left panel. The 3.9 micrometer brightness temperature is measurably warmer than the 10.7 micrometer temperature, especially for small values of N (hot spot fractional coverage).

 

A similar warming effect can be seen for subpixel cloud cover where cloud top temperatures are typically colder than the surrounding surface. Comparing the 3.9 micrometer band with either the 10.3 or 11.2 micrometer bands for a scene with small cumulus cloud elements for example, we would expect the 3.9 micrometer image to appear slightly warmer.

The animation below shows a very active brushfire as it appears in color enhanced 3.9 micrometer band imagery beginning in the overnight hours. Thicker portions of the smoke plume are also evident given the band's sensitivity to larger aerosol particles.

As the loop enters daytime, solar energy reflects off the surrounding land surfaces and hence we see warmer brightness temperatures. Some of the warming (in brightness temperatures) is also due to infrared emission from a warming surface. Despite the additional solar input, the fire's hotspots remain visible. Some of this is due to the band's increased saturation temperature on AHI and ABI. The newer 3.9 micrometer band saturates at 410 K, which is about 70 K warmer than with legacy GOES. This allows for improved identification of fires, distinguishing a fire from nearby hot reflective surfaces for example, and for better sub-pixel fire characterization such as a fire's aerial coverage and burn temperature.

 

As in the visible and near-IR, sun glint must be considered when viewing shortwave IR imagery over water areas. When water surfaces are smooth, large reflections occur when the sun and satellite are properly aligned.

Two different image enhancements are shown in the loops below. The loop on the left uses brighter shades to indicate warmer brightness temperatures and darker shades to indicate colder brightness temperatures. The loop on the right uses a proposed AWIPS IR color enhancement.

Notice the bright relatively warm region moves from east to west as the sun moves. We also see patches that are relatively warm and cold within the larger glint pattern. Some patches change from colder to warmer as the sun's reflection approaches, and then back again. These are indicative of relatively calm waters surrounded by a rougher sea surface.

Cloud-top Phase Band

Located at the edge of the longwave IR window region, the ABI's 8.4 micrometer "cloud-top phase" band is new to the GOES satellites.

While similar to longwave IR window bands in the 10 to 12 micrometer region, the 8.4 micrometer band has an additional capability to help determine cloud-top microphysical properties during both day and nighttime. The sample cloud type product image shown below combines information from the AHI's 8.4, 11.2, and 12.4 micrometer bands and will also be available from GOES-R.

Combining the 8.4 and 11.2 micrometer bands allows for a more consistent day-night discrimination between water and ice clouds when compared to the more "traditional" fog/low cloud product which is limited to nighttime use due to daytime solar contamination.

Similar to the 3.9 micrometer band at night, the emissivity of dense water cloud tops at 8.4 micrometers is less than at 11.2 micrometers, resulting in cooler cloud tops at 8.4 than at 11.2 micrometers. In contrast, ice clouds tend to have much smaller temperature differences and even of the opposite sign for thinner cirrus.

Drag the slider to compare water vs. ice cloud-top brightness temperatures for the 8.4 and 11.2 micrometer images.

 

The ABI's 8.4 micrometer band also is lies in a "dirty" window region where water vapor absorption is stronger than in the 10.3 and 11.2 micrometer window bands. This results in brightness temperatures for surface features that are slightly cooler in the presence of a moist low-level atmosphere.

For desert scenes, despite a typically dryer lower atmosphere, cooler 8.4 micrometer brightness temperatures relative to the 10.3 and 11.2 bands, are common as well. This is the result of a surface emissivity (or emission effiency) for a desert surface that's typically around 90%, compared to darker soil types and vegetation cover where emissivities are closer to 100%.

The scene below shows the cooling effect in the 8.4 micrometer band compared to the 10.3 clean IR window band. Drier areas of the Intermountain West are compared to the more temperate and moist regions of the eastern U.S.

 

The 8.4 micrometer band, when combined with the 11.2 and 12.3 micrometer bands, is also useful for detecting volcanic ash clouds containing aerosols and sulfur dioxide (SO₂), especially for large concentrations of SO₂.

The atmospheric transmission plot for SO₂ shows two distinct absorption bands in the infrared. Notice that SO₂ absorption in the 8.4 micrometer band becomes pronounced only as SO₂ concentrations become quite large.

Ozone Band

The ABI's 9.6 micrometer "ozone" band is located in a strong ozone absorption region within Earth's infrared emitted spectrum.

Despite spatial variations, 90% of Earth's ozone resides in the stratosphere and with large concentrations just above the tropopause. This means that most of the ozone signal in this band originates from layers near and above the tropopause.

The ozone band is also slightly sensitive to water vapor, such that the combined absorption effects of ozone and water vapor will result in imagery that generally appears cooler than in the IR window bands. Cloud tops near or extending above the tropopause may however appear warmer in the ozone band due to ozone emissions from the lower stratosphere where temperature increases with height.

Drag the slider to compare a 9.6 micrometer ozone band image with the corresponding 10.3 micrometer clean IR window band.

 

The ozone band's brightness temperatures are also sensitive to changes in satellite viewing (or zenith) angle. As the satellite scans farther away from nadir, the point directly beneath the satellite above the equator, the scan angle increases and energy is detected from higher in the atmosphere.

In the weighting function animation below, notice that the stratospheric ozone contribution increases as the viewing angle approaches the limb near 70 degrees, and the brightness temperature 'T sub b' decreases. We can also see the ozone band's water vapor absorption as a contribution in the lower troposphere. Assuming a moist atmosphere, the height of that contribution also increases slightly with increasing scan angle, resulting in an additional cooling effect.

Looking at the ozone band image for a moment, we see that brightness temperatures cool with increasing distance away from the satellite's subpoint above the equator. The effect is more pronounced toward the poles since stratospheric ozone concentrations are greater there than along the equator. This cooling phenomenon, known as "limb darkening", is particularly pronounced for scan angles greater than 60 degrees. Limb darkening affects all of the IR bands since they all experience at least some degree of absorption by gases, but is most noticeable in the ozone and carbon dioxide bands where absorption is strongest.

 

Clean Longwave IR Window Band

The ABI's 10.3 micrometer "clean longwave IR window" band is one of four ABI infrared window bands. It's located in a window region of the IR spectrum with little sensitivity to absorption by water vapor, and is also the least sensitive to water vapor of the four longwave IR window bands (at 8.4, 10.3, 11.2, and 12.3 micrometers). This suggests that the 10.3 micrometer band provides the "cleanest" view of both cloud tops and Earth's surface.

The weighting functions (or contribution functions) shown below are for the four IR window bands under clear-sky conditions, with the 10.3 micrometer band's curve highlighted in blue. The area under and to the left of each curve indicates the relative contributions to the outgoing energy from various layers of the atmosphere. We see that all four weighting functions peak at the surface. This reinforces the point that the satellite brightness temperature signal is dominated by the surface for cloud-free scenes in all four IR window bands.

At the 10.3 micrometer wavelength, most surfaces and cloud types have an emissivity close to 1 (that's 100% in terms of efficiency), with a notable exception being thin cirrus. Therefore, the brightness temperature sensed by the satellite is close to actual surface skin or cloud-top temperature for scenes other than thin cirrus. Actually, thin cirrus permits more energy to pass through it than in the other IR window bands, and will therefore appear warmer in the 10.3 micrometer band.

For most surface features, the 10.3 micrometer band is also slightly warmer than traditional IR window bands due to its relatively weak sensitivity to lower tropospheric moisture.

The comparison below highlights some of the more commonly seen differences between the 10.3 and more traditional 11.2 micrometer bands. Given the similarity between these two IR bands, the 10.3 micrometer clean window band can be used much like the 10.7 micrometer band on legacy GOES as well as the Suomi-NPP and JPSS polar-orbiting satellites.

Drag the slider to compare 10.3 and 11.2 micrometer band images over the Southeastern U.S. during early summer.

Please note: The two images have been enhanced with the Rainbow IR color table to better highlight subtle temperature differences between the two spectral bands.

For derived products, the 10.3 micrometer band's "clean window" quality helps with improving atmospheric moisture corrections in SST and other surface products, estimating moisture profiles, estimating cloud-top particle size, and characterizing surface properties.

Longwave IR Window Band

Familiar to most as the traditional longwave infrared band or channel, the ABI's 11.2 micrometer "longwave IR window" band is the third of four ABI IR window bands covered in this lesson. It's located in a window region of the IR spectrum with only minor sensitivity to absorption by water vapor. As with the other three window bands, the result is a brightness temperature signal dominated by the temperature signal from cloud tops and the surface.

At the 11.2 micrometer wavelength, similar to the 10.3 micrometer band, most surfaces and cloud types have an emissivity close to 1 (that's 100% in terms of efficiency). Therefore, the brightness temperature sensed by the satellite is close to actual surface skin or cloud-top temperature for scenes other than thin cirrus.

Less longwave IR energy escapes to space through optically thin (semi-transparent) cirrus in the 11.2 micrometer band compared to the 10.3 micrometer band. Therefore, cirrus that is not opaque will appear cooler in the 11.2 micrometer band. The 12.3 micrometer dirty window band allows even less longwave energy to transit a thin cloud, so cirrus clouds will appear even cooler in that band.

In the images below, we see that thinner cirrus appears progressively cooler as the wavelength increases from 10.3 to 11.2 and then to 12.3 micrometers. Surface features appear progressively cooler as well, something we'll explore more in a moment.

Drag the slider to compare the GOES-16 IR images showing hurricane Harvey.

 

The weighting functions (or contribution functions) shown below include the ABI's four IR window bands under clear-sky conditions, with the 11.2 micrometer curve colored brown. The area under and to the left of each curve indicates the relative contributions to the outgoing energy from various layers of the atmosphere.

While all four weighting functions peak at the surface, we see that the 11.2 micrometer curve is located to the right of the 10.3 micrometer band curve below about 700 millibars. This indicates a greater sensitivity to lower tropospheric water vapor than at 10.3 micrometers. The result is a brightness temperature for clear-sky scenes that is slightly cooler in the 11.2 vs. the 10.3 micrometer band, but still warmer compared to the 8.4 and 12.3 micrometer bands where water vapor sensitivity is stronger. The cooling effect in the 8.4, 11.2, and 12.3 micrometer bands increases as low-level moisture increases, which has implications for assessing low-level moisture and diagnosing air mass boundaries.

Drag the slider to compare the 8.4, 10.3, 11.2 and 12.3 micrometer weighting functions and images.

 

GOES-R improvements for the IR bands include a 2-km spatial resolution, 15-minute refresh rate over the full Earth disc, 5-minute refresh rate over the extended contiguous U.S., and an even more frequent 1-minute cadence for two movable areas. With these improvements, the 11.2 micrometer band gives meteorologists an unprecedented ability to monitor the evolution, upscale growth, and decay of convection via cloud-top temperatures in conjunction with radar and the GOES-R GLM's ground breaking lightning observations.

The 11.2 micrometer band is also a significant contributor to all three main GOES-R derived product areas including: baseline, future capability, and envisioned. Examples for baseline products include imagery, cloud properties, satellite derived winds, sea surface temperature, and total precipitable water. Examples of future capability products include overshooting tops and total ozone. And the final class of products are those envisioned as follow-on applications that go beyond conventional uses of satellite information to include decision aids for improved predictions of convection and severe weather.

Dirty Longwave IR Window Band

As with the other three longwave IR window bands, the ABI's 12.3 micrometer band is located in the broad longwave IR window region. Of this IR band grouping, the 12.3 micrometer band is the most sensitive to absorption by low-level atmospheric moisture, hence its naming as the "dirty" longwave IR window band.

Similar bands have been available on both GEO and LEO satellites for many years due to this band's many uses. In combination with a cleaner IR window band, the 12.3 micrometer band can be used to estimate low-level moisture, analyze air mass differences and boundaries, diagnose cloud properties, and both observe and quantify volcanic ash, airborne dust and sand.

As with the 10.3 and 11.2 micrometer bands, most surfaces and thick opaque clouds have an emissivity in the 12.3 micrometer band close to 1. Beyond emissivity however, there are subtle yet important differences related to the band's sensitivity to low-level moisture, and more complex absorption and scattering processes within water and ice clouds, dust, and volcanic ash. These optical properties make the 12.3 micrometer band particularly useful in combination with other IR bands for numerous applications as we'll explore a bit further.

One such combination of different IR bands is what's known as a "split window difference" that uses the 12.3 micrometer band and a cleaner IR window band such as the 11.2 micrometer band. The difference product is particularly good for highlighting a variety of atmospheric and cloud properties already mentioned.

Let's look at identification of thin cirrus for a moment, a long standing use of the 12.3 micrometer band. Ice cloud particles absorb more energy at 12.3 micrometers than in the 8.4, 10.3 and 11.2 micrometer bands. This reduces the brightness temperature of thinner cirrus compared to the other three IR window bands. In the image comparison below, we see that thinner cirrus appear progressively cooler as wavelength increases from 10.3 to 12.3 micrometer wavelengths. Surface features also appear progressively cooler, something we'll look at next.

Drag the slider to compare the GOES-16 IR images showing hurricane Harvey.

 

As mentioned earlier, the 12.3 micrometer band is the most sensitive to atmospheric moisture of the four IR window bands. Weighting functions (or contribution functions) for the ABI's four IR window bands are shown below for clear-sky conditions and for a relatively moist summertime mid-latitude atmosphere. The 12.3 micrometer band's curve is colored orange.

Focusing on the 12.3 micrometer weighting function, we see that the curve is located farthest to the right of the four IR window bands with much of contribution due to water vapor absorption occurring below about 700 millibars. This indicates a more significant sensitivity to lower tropospheric water vapor. The result is a brightness temperature for clear-sky scenes that is slightly cooler compared to the other three bands, assuming that no strong temperature inversions are present within this layer.

In the three weighting functions below, notice that for most scenes, brightness temperatures cool with increasing wavelength, where there is greater sensitivity to low-level moisture.