Exceptional Saharan dust transport over the Atlantic Ocean

Filter by


EUMETSAT Users Twitter

RSS Feed

RSS Icon Image Library

In June 2020 large amounts of Saharan dust, declared by the media as an "historic Saharan dust plume", travelled on easterly trade winds all the way to the Caribbean and south-eastern parts of the continental US.

Date & Time
5 June 2020 00:00 UTC–26 June 21:00 UTC
Meteosat-11, Suomi-NPP, GOES-16, CALIPSO
Dust RGB, True Colour RGB, Thermal Infrared Channels

By Natasa Strelec Mahovic and Jose Prieto (EUMETSAT)

Saharan dust transport is a phenomenon that occurs every year, mostly during summer months, often reaching Central and South America.

Being rich in elements such as nitrogen, iron and phosphorus, Saharan dust brings important nutrients to the plant-life of the Amazon's rainforest and acts as a fertiliser for the production of phytoplankton.

However, although the occurrence is not unusual, in June 2020 the amount of the dust transported and the density of the dust cloud was exceptional, causing very poor visibility and major dust deposits in many places in the Caribbean.

Synoptic analysis

Because of its long duration and exceptional intensity, the phenomenon could be followed from meteorological satellites throughout the month of June (Figures 1 and 2).

Figure 1: Suomi NPP True colour RGB, 05 June–26 June, showing dust being transported from Sahara to the Caribbean.

Figure 2: Meteosat-11 Dust RGB, 13 June 00:00 UTC–26 June 21:00 UTC

The daily sequence of Dust RGB composites at 12:00 UTC (Figure 3) shows that the largest plumes actually generated from soil. Some occurred in the east parts of Africa, and travelled through the southern Sahara to reach the Atlantic, for example one such plume was generated on the 9-10 June.

Figure 3
Figure 3: Main dust outbursts in the course of June over the Sahara

The Bodélé depression, often a factor in Saharan dust transportation, was not a major source of dust this month, even if its natural wind channel contributed to a large accumulation around 5 June, which eventually reached the Caribbean.

Although there were several episodes of dust being lifted from Sahara at beginning of June, the presence of a low pressure centre and the circulation in the lower layers of the atmosphere kept the dust 'floating' over West Africa, with only a small amount reaching the Atlantic (Figure 4).

Figure 4
Figure 4: Meteosat-11 Dust RGB, with MSLP and 10 m wind overlaid, 6 June 15:00 UTC

Since the dust cloud was confined to the lowest 4 km of the troposphere, as can be seen from the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) instrument onboard the CALIPSO satellite (Figure 5), it was transported by the winds in the lower atmospheric layers.

Figure 5
Figure 5: Transect of the dust over the Dominican Republic, CALIOP 23 June. Source: NASA

The strongest transport started mid-June with strengthening of the Azores High and the associated easterly winds in the lowest 3 km.

The animation (Figure 6) shows the Azores High located half way between Europe and America, with the largest pressure gradients persisting over 15N–20N latitude for more than two weeks, causing strong 850 hPa winds with speeds more than 20 m/s. This synoptic situation was very favourable for the intensive transport of dust. Also note how the maximum of the wind moves westwards, transporting the large amounts of dust far west.

Figure 6: Meteosat-11 Dust RGB, with MSLP and 850 hPa wind overlaid, 14–26 June


Meteosat/GOES comparison

Thermal window channels estimate dry particulate matter (PM) concentrations and their heights. However, the viewing angle plays a significant part in the quality of the aerosol optical depth (AOD)— satellites detect dust better on the boundary of the field of view than near the centre. The reason is the longer interaction of the radiation with dust particles, through a more slant path in the atmosphere.

The average number of particles touched by a straight line through the plume is called the aerosol optical depth (AOD) of the plume. This number is bigger when the line traverses the plume diagonally (obliquely), so it is bigger for the most distant satellite (Figure 7). Multi-satellite imagery can mislead the detection of dust and typically underestimate its concentration over the Atlantic.

Figure 7
Figure 7: Concept for the larger aerosol optical depth and colder infrared signal from the oblique satellite

On the comparison of GOES and Meteosat views of the dust plume for several locations on its path (Figure 8), taken on 23 June at 15:10 UTC, each satellite enhances the dust plume closer to the other satellite. Notice that both satellites look at precisely the same dust concentrations, just from the different viewing angles.

Figure 8
Figure 8: GOES-16 (above) and Meteosat-11 (below) tell different stories on dust on 23 June at 15:10 UTC. In the GOES image, large amounts of dust over the Caribbean go almost unnoticed to GOES (weak magenta hue), around the satellite vertical. Meteosat-11 SEVIRI Dust RGB shows large amounts of dust over the Caribbean on a very slanted view, while dust near Africa is less outstanding.
Figure 9
Figure 9: Over Cape Verde in the eastern Atlantic (location B Fig 8) scatter plots for the split window (GOES left, Meteosat right; difference 10 µm–12 µm in the vertical axis, channel around 10 µm in the horizontal)

Details of satellite measurements for three locations, marked in Figure 8, A (21N–8E) on land in Africa, B (17N–23E) over the ocean near Cape Verde, and C (18N–70E) in the Caribbean, can be analysed with the help of scatter plots.

Channels on ABI are spectrally different from channels in SEVIRI close in wavelength. The readings differ by a fraction of a degree due to spectral properties, most of the time. But differences due to the viewing angle are bigger. That is why it is assumed that 10.3 µm on ABI is equivalent of 10.8 µm on SEVIRI, and 12.3 µm on ABI equivalent of 12.0 µm.

Figure 10
Figure 10: Over the Caribbean (location C), same as Figure 9. See text for interpretation.

The graphs for location B over the Cape Verde (Figure 9) show the dust distribution of brightness temperatures (BT, also called glow temperatures). On the shorter wavelength channel at 10.x µm, GOES BTs are about 5 K colder than on Meteosat, 287 K versus 292 K. A longer path through the dust reduces the signal (see Figure 7). Also the difference with the longer wavelength gets deeper, -1.5 K versus -0.6 K, part of which is a bias due to different weighting functions for ABI and SEVIRI channels, and part due to a higher AOD for GOES. The upper dust contribution for thick plumes is colder in the diagonal than along the vertical.

Figure 11
Figure 11: Over continental Africa (location A), same as Figure 9. See text for interpretation.

For the location C in the Caribbean region (Figure 10), the behaviour is similar, with an AOD advantage for Meteosat, which reaches -2.5 K in the difference (vertical axis) compared with just 0.5 K for GOES almost in the vertical, due to the high humidity above the sea favouring 10.3µm channel. Also the 10.x µm values show 15 K difference (29 5 K versus 281 K), which points to a large concentration of dust, hard to notice in GOES imagery.

Finally, for the African soil (Figure 11), the graph again shows a large bias for the shorter wavelength of 15 K. A larger area is covered in the GOES graph (left), so different layers of dust can be identified as the three 'legs' in that graph, at 277, 285 and 293 K. These are BT values, an approximate average of the dust temperatures and ground temperatures.

The Meteosat-11 Dust RGB animation from 6–22 June (Figure 11) shows the various African dust sources and, at the end, the dust lingering over the Caribbean, on the limb of the field of view.

Figure 11: Meteosat-11 Dust RGB, 06 June 00:00 UTC–22 June 00:00 UTC

The GOES-16 animation for 16–27 June (Figure 12) focuses on the Atlantic-Caribbean scene, with the dust signal becoming faint (in contrast to Meteosat).

Figure 12: GOES-16 Dust RGB, 16 June 00:00 UTC–27 June 12:00 UTC
We use essential cookies to ensure that we give you the best experience on our website. To analyse website traffic we also use third-party performance cookies. If you are ok with the use of essential as well as non-essential cookies, please select Accept & Continue. Instructions on how to prevent the use of non-essential cookies are available under our Terms Of Use, or simply select Decline Cookies.