Skip navigation

RAMMB: Regional and Mesoscale Meteorology Branch logo CIRA: Cooperative Institute for Research in the Atmosphere logo NESDIS: NOAA Environmental Satellite, Data, and Information Service logo

Announcement: Please visit out new web application, SLIDER, for every pixel of real-time GOES-16 and Himawari-8 imagery.

MODIS-based Blowing Dust (Yellow/Pink)

Blue-light Absorption Technique

Fig. 1. A large dust front lofted by post-frontal northerly winds advances southward through the Mojave Desert. The dust enhancement (right) depicts significant areas of suspended dust as pink, land as dark green, and clouds/snow as cyan. The brightness of the enhancement is a proxy for algorithm confidence (i.e., brighter implies that more of the various detection tests produced strong signals based on the detection thresholds assumed). The GOES-R Advanced Baseline Imager (ABI) will be able to produce imagery of similar quality, but at dramatically higher temporal refresh (minutes, vs. the 1-2 times per day currently available from MODIS Terra and Aqua). As examples to follow will illustrate, the enhancement can also be presented as yellow dust against blue backgrounds, to accommodate users having difficulty with green/red color discrimination. Both versions are provided in the Proving Ground demonstrations.

1) Product Information:

- Who is developing and distributing this product?

The Cooperative Institute for Research in the Atmosphere (CIRA) in Fort Collins, Colorado and the Naval Research Laboratory (NRL) in Monterey, California are developing and distributing the Blowing Dust Enhancement product.

- Who is receiving this product, and how?

The Blowing Dust Enhancement products are sent to the National Weather Service (NWS) Regional Headquarters from which they are distributed to Weather Forecast Offices (WFOs) for display on their local AWIPS systems. Imagery updates are available approximately two times per day from the MODIS sensors on board Terra (~10:30 AM local time) and Aqua (~1:30 PM local time).

- What is the product size?

The size of Blowing Dust Enhancement product images is determined by the span and resolution of the AWIPS domain itself. Since current AWIPS system displays accommodate 1-byte per pixel, a good rule of thumb is that the size of the imagery (in bytes) corresponds roughly to the total number of pixels in a given AWIPS domain. For example, an AWIPS domain having dimensions of 1000 x 1000 pixels will require approximately 1 Megabyte (~106 bytes).

2) Product Description:

- Purpose of this product:

The purpose of the Blowing Dust Enhancement imagery product to provide a visually intuitive depiction that is useful to experts and non-experts alike. Instead of toggling between visible imagery, infrared imagery, and various spectral differences, the current product attempts to put several of these elements together and depict the consensus of where the tests have high confidence in the presence of dust in the scene.

- Why is this a GOES-R Proving Ground Product?

The Blowing Dust Enhancement product demonstrates the kind of imagery that will be possible in the GOES-R era. GOES-R will feature the Advanced Baseline Imager (ABI) sensor whish will be able to produce versions of the imagery shown here at much higher time resolution.

- How is this product created now?

Since we do not have all the required channels on present-day GOES satellites to do the Blowing Dust Enhancement, we can only demonstrate it now via the Moderate Resolution Imaging Spectroradiometer (MODIS) instruments which fly on the polar-orbiting NASA Terra and Aqua satellites. This limits the temporal frequency of coverage to about two passes (mid-morning and mid-afternoon) per day.

Depicting lofted dust in enhanced satellite imagery requires that we isolate its signal (in terms of measurable properties) in order to separate it in color-space from other components of the scene. The science literature describes many ways to do this. As the chemical properties of lofted dust are very diverse, it is often the case that one approach will work better than another for a given region. The current approach combines several of the more reliable discriminators which take advantage of the spectral bands that will be available to GOES-R. The principal considerations for the enhancement are:

  1. Color: to some extent we can distinguish dust from meteorological clouds with our own eyes through color differences (mineral dust clouds often have earth tones, owing to their strong absorption of blue light). Since most meteorological clouds appear gray/white, the color information provides a useful discriminator. This is why true color imagery can also be useful for dust detection. A satellite sensor that has the ability to distinguish color (e.g., via multiple narrow-band channels in the visible part of the spectrum) can combine this information in an analogous way to color vision for dust detection. This approach works well over dark surface backgrounds, like water, but becomes problematic over bright surfaces like deserts (and particularly, when the dust is over a surface background that has similar coloration to the dust itself). In this case, we need to appeal to other discriminators, and these are described below.
  2. Temperature: When dust is lofted into the atmosphere, its temperature rapidly adjusts to whatever the air temperature of the environment is. The dust layer will have a radiative influence on its environmental temperature (via absorption of upwelling thermal radiation from below, and reflection of incoming solar radiation from above) but these effects occur on longer time scales and are of secondary importance here, so we ignore them. Since temperatures in the lower atmosphere generally decrease with height (especially during the daytime hours), the dust layer cools as it rises, and soon produces a nice "thermal contrast" against the warmer surface. A satellite sensor that is sensitive to heat (infrared radiation) can therefore assist us in detecting an elevated dust plume based on the temperature contrast it produces against the warm background surface. If we combine this temperature contrast information with the color distinction information, now we have a more robust way of distinguishing between meteorological clouds (which are gray/white and cool) and dust plumes (which are earth-tones and cool) over both water and land surfaces than either test can provide on its own.
  3. Spectral Differences in Transparency: The chemical and physical properties of mineral dust layers give rise to different optical properties (scattering and absorption behavior) depending on what part of the infrared spectrum we view them in. In the same way that these properties result in preferred blue-light absorption in the visible part of the spectrum, differences occur in the middle and thermal infrared parts of the spectrum which gives rise to dust plumes appearing "thinner" (more transparent) or "thicker" (more opaque). As a result, measurements of a mid-level dust plume in a part of the infrared spectrum that corresponds to the more absorbing behavior may appear cooler than measurements of that same dust plume in a part of the infrared spectrum where dust appears more transparent. In the case of more absorbing behavior, we’re seeing more of the cool emissions of the dust layer’s environmental temperature, while in the case of more transparent behavior we’re seeing warmer contributions from the underlying surface that are able to transmit through the dust. Computing a difference between two such measurements, chosen carefully, provides an effective way to identify dust. The two channels used here are the 11 and 12 micrometer thermal infrared bands, which provide a dust detection signal that is opposite in sign to most meteorological clouds.

While any one of the above discriminators alone may not be sufficient to fully distinguish elevated dust, we can often provide a reasonable isolation of dust from other components of the scene through combining the various tests. When all tests are satisfied simultaneously (i.e., an intersection) we can say with greater confidence that the pixel being examined contains dust. It is important to point out that in the current algorithm these tests are not strictly "yes/no" but are cast as confidence factors that range between values of 0 (low confidence) to 1 (high confidence). The extent to which the tests are satisfied (strength of the dust signal) is communicated in the enhanced imagery in terms of brightness of the enhanced dust features...with brighter regions being higher confidence.

3) Product Examples and Interpretation

Fig. 2.: Another example of dust (yellow enhancement) in the desert southwest portions of the United States, as observed by Aqua-MODIS on 5 April 2010 at 2100 Z. Here, mineral dust from the Painted Desert region of Arizona is seen lofted in the strong pre-frontal southwesterly winds of a baroclinic system. The dust later affected the air quality and visibility near Grand Junction, and deposited on parts of the snow covered San Juan Mountains to the northeast.

The Blowing Dust Enhancement product is designed to simplify the detection of any significant dust regions present in a potentially complex scene. We do this by gathering together all the ‘dust discriminators’ as outlined in Section 2 of this training module, and present it in a visual form that helps the dust stand out from other constituents of the scene. The end result is imagery that is ‘false color’ (i.e., it no longer looks ‘real’ in comparison to true color imagery) but provides a less ambiguous depiction of where dust resides.

Fig. 3.: In this example, a view over northwestern Africa, plumes of Saharan dust are making their way toward the west Offshore a tropical cyclone is in its formation stage. Tracking dust in this region is of interest to Hurricane forecasters, owing to the possible impacts the dust (or the dry air mass associated with the dust) has on the formation and dissipation processes of these storms. While the GOES-East satellite will not provide coverage this far to the east, this same algorithm will be available to future Meteosat geostationary satellites. Similarly, the future MTSAT Himawari geostationary satellites will provide similar capabilities to the west of GOES-W coverage. Using the international geostationary resources, we will be able to track U.S.-bound dust-laden atmospheric layers from their origins in the Sahara or the Gobi deserts all the way across the Atlantic/Pacific oceans.

Fig. 4.: Example of blowing dust over northern Texas, associated with strong westerly/northwesterly winds in the cold sector of a baroclinic system. Also noted are white plumes which correspond to grassland fires that are not enhanced by the current algorithm. Researchers are examining to what extent the entrainment of dusty air into convection influences the development/behavior of severe weather.

What to look for: Regions of pink/orange against dark green backgrounds (or in the case of the yellow-dust variant, look for regions of yellow against dark blue backgrounds). Sometimes dust plumes will display sharp boundaries (as in Fig. 1, and most often along the leading edge of a dust front), but more often possess a diffuse appearance and may ‘fan out’ like cirrus clouds.

Fig. 5: Example of cold terrain effects. Here, mountainous regions in the southwestern U.S. appear red.

What to watch out for: Cloud shadows, coastlines (both ocean and lake), and cold terrain (especially mountain ranges and elevated plateaus during the winter months; see Fig. 5) occasionally appear ‘enhanced.’ If used regularly, analysts will quickly become adept at identifying and discarding these residual artifacts of the algorithm. Dry lake beds may also appear bright pink in this enhancement. These lake beds often serve as point-sources for lofted dust.

Other Considerations: While major dust outbreaks do occur typically a few times per year over the southwestern deserts of the United States, they are far less prolific than the expansive deserts of Africa and Eurasia. A more common variety of dust storm to the United States is associated with the cold-pool outflow of thunderstorm complexes. Called "Haboobs" in the Middle East, these storms form literal walls of dust on scales of several miles (compared to hundreds or even thousands of miles for major dust storms). These storms can reduce visibilities at the surface to near zero in a matter of minutes. Oftentimes this variety of dust storm is missed by a polar-orbiting sensor like MODIS due either to time sampling (catching the event in progress from one or two snap-shot observations) or their being obscured by overriding clouds associated with the parent thunderstorm(s) at the time of observation. With the advent of GOES-R ABI, detection of these storms will be improved significantly.

Note: The current version of the Blowing Dust Enhancement product requires information from sunlight reflection, and thus is only valid for daytime observations. A future version this product that uses additional infrared bands that will be available from GOES-R ABI will provide a 24-hr capability. A variant on this product is being developed for the MSG-SEVIRI sensor.

4) Advantages and Limitations

The main advantage of Blowing Dust Enhancement imagery over conventional single-channel imagery is the ability to make a rapid assessment of dust in the scene. Its advantage over other dust enhancements is the ability to do so while suppressing (in color/brightness) the non-dust components of the scene. Another advantage of imagery-based enhancements (vs. pure quantitative products such as a ‘dust mask’) is that imagery retains the meteorological context of the scene. Dust is often lofted as a result of characteristic wind flow patterns (e.g., in the cold sector of a baroclinic system, terrain-locked circulations, or even as the result of thunderstorm outflows as often occurs in the desert southwest of the United States) that can often be inferred from the meteorological cloud field.

The main limitation is the depiction of the product in AWIPS as an 8-bit color (256 colors) image. Since the Blowing Dust product is in fact a 24-bit red/green/blue composite, there will be some degradation in quality when attempting to represent a broad color palette (256^3 = almost 17 million) in a reduced number of color tones. This challenge will go away in AWIPS II, which is advertised to have the capability to display 24-bit imagery.