SUOMI NPP - VIIRS corrected reflectance (true color), 31 March 2018. source: NASA WorldView


Saharan sub-Sahel dust storm of late March – early April 2018

by Jos de Laat & Piet Stammes (KNMI), 30 April 2018

During the last few days of March 2018, the sub-Sahel region over eastern Africa was hit by an intense dust storm. The event, which started on the 28th of March and lasted approximately one week, stretched from Nigeria to the Sudan, where it was particularly severe. The airport of Khartoum was shut down for several days and flights were rerouted.

However, the event provided an excellent opportunity to test some newly developed applications of satellite data which will be part of the EUNADICS-AV project.

Figure 1 shows the GOME-2 absorbing aerosol index (AAI) plotted in top of the FRESCO cloud height and cloud cover. The date of these measurements was 31 March 2018.

The dust storm shows up nicely in the AAI data, with the largest AAI values over the Sudan. The plots also show a correlation between the AAI and the FRESCO cloud cover fraction. Presumably this is because FRESCO characterizes these clouds as semi-transparent, and with the AAI indicative of the amount of dust particles, the denser the cloud, the larger the associated cloud cover.

The plots also show FRESCO cloud top height, which is indicative of the height of the dust particle layers. According to GOME-2, the dust is located between 1-2 km altitude (mid-layer value). Surprisingly enough, preliminary TROPOMI data (not shown), indicates that the dust layer is located between 0-1 km altitude (mid-layer value), about a kilometer below the heights measured by GOME-2. How is that possible?

The time of the day when both satellites measure differs by approximately 4 hours: GOME-2 observes around 9:30 in the morning (local time), TROPOMI observes around 13:30 (local time). Because the sun is not in zenith at 10:30 in the morning, the light travels a longer path through the atmosphere than at 13:30. That also means a longer path in the upper parts of the semi-transparent dust layer. As a consequences, GOME-2 “observes” a higher altitude of the dust compared to TROPOMI.

On the other hand, both satellites indicate that the dust is confined to a rather shallow altitude band with a width of approximately 1 km. This suggests that for dust layers with sufficient amounts of particles, the apparently cloud height provides an accurate estimate of the height of the dust layer. Additional research will have to show whether this is a realistic value.

One possibility to do this is by using measurements from the CALIPSO satellite, which measures the height profile of clouds a dust along narrow orbits around the earth. Figure 2 shows an example of a so-called “curtain plot” as measured by the CALIPSO satellite. The pink circles indicate where CALIPSO has identified dust layers about half a day after the GOME-2 measurements. The top of the dust layers is located between 3 km and up to 5 km further south, although the layer further south appears to be a water cloud rather than a dust layer. The 3 km top layer (~ 1.5 km mid-layer) nicely corresponds with GOME-2 mid-layer heights.

How accurate this visual correspondence really is will be investigated in detail within the EUNADICS project.


Figure 1. Satellite measurements of the FRESCO cloud top height (upper left plot) and absorbing aerosol index (AAI) values (FRESCO cloud cover > 33 %). The upper right plot shows a scatterplot of AAI vs FRESCO cloud cover, with the lower right plot showing the corresponding probability distribution per 0.5 AAI index value intervals. The lower right plot shows the probability distribution of the FRESCO cloud height, color coded by AAI values. Data shown for 31 March 2018.


Figure 2. Satellite measurements of cloud and dust particles as a function of height along the satellite track indicated in the insert for 31 March 2018. The pink circle in the insert indicates the approximate part of the orbit shown in the figure. The figure is obtained from the NASA CALIPSO website (browse images: