Around 11:00 UTC on 5 April an isolated thunderstorm developed rapidly southwest of Ouagadougou, the capital of Burkina Faso.
More information and detailed analysis of the feature can be found in the In Depth section.
by HansPeter Roesli and Henk Verschuur (EUMETSAT)
Over the Sahel the boundary layer was heavily laden with dust and smoke (see Dust RGB from 11:30 UTC, PNG, 844 KB, Aqua MODIS image from 13:25 UTC, PNG, 813 KB, picture 01, JPG, 472 KB, picture 02, JPG, 519 KB, source: H. Verschuur).
The Dust RGB indicates the presence of desert dust straddling the border of Mali with Burkina Faso and Niger that had been lingering there for several days. The MODIS image, on the other hand, shows a faint smoke veil that had been pushing up from south.
The outbreak of the thunderstorm was very typical for African thunderstorms, being at the intersection of two outflow boundaries, one of unknown origin and of wavy character arriving from the north, the other one generated by a convective cluster moving westwards over Benin-Togo-Ghana joining from the east (see IR10.8 image from 11:00 UTC (PNG, 306 KB), IR10.8 animation (MPG, 3 MB), VIS0.8 image from 11:00 UTC (PNG, 245 KB) and VIS0.8 animation (MPG, 1 MB)).
Later in the afternoon the Ouagadougou thunderstorm merged with the westward moving cluster. Besides the dynamic forcing there was also sufficient water vapour available, especially in the lowest 2–3 km. A sharp boundary separated the dry Saharan air from the humid air closer to the coast of the Gulf of Guinea (see cross section of specific humidity from 12:00 UTC ECMWF analysis, PNG, 22 KB).
The animated sequence of images shows the general westward movement and development of the convective complex (see Met-8 RGB Dust animation, MPG, 4 MB). It does not appear that the suspended dust changed very much during the day, i.e. most probably the dust sucked into the convective towers was readily replaced by freshly lifted dust. The display of the cloud top temperature gives an excellent insight into the multitude of convective cells (see Met-8 IR10.8 animation, MPG, 4 MB). Zooming in on the Ouagadougou thunderstorm (see Met-8 IR10.8 animation (zoom), MPG, 3 MB) reveals alternating so-called doughnut- and U-shaped (or V-shaped) temperature structures that point to the storm's particular severity.
Another indication of the severity is found when inspecting the freezing level of the cloud water (see Scatterplots of T-Reff, PNG, 1 MB, source: D. Rosenfeld). Region 2, the Ouagadougou thunderstorm, appears to freeze only very close to the -40 °C level, the level of spontaneous freezing. This can be verified by inspecting the T-Reff diagrammes (top temperature versus effective radius of the cloud droplets) shown in the same picture. The green and yellow curves (average and median radii of the droplets, respectively) cross the 15 micron line just below the -40 °C line.
At values warmer than the crossing points the droplet size increases very slowly (steep curves). Beyond the -40 °C the radii increase much faster, indicating frozen particles. Cloud glaciation so close to the spontaneous freezing point of water is very unusual for a convective cloud. This points to the presence of a huge amount of condensation nuclei (smoke particles!) in the boundary layer that competed for the available water vapour leading to the formation of many, but small (<15 micron) cloud droplets.
The dust particles (ice nuclei) would then not find much water vapour once they got activated below the freezing point. A narrow spectrum of small cloud droplets inhibits droplet growth and delays glaciation, i.e. glaciation and the release of latent heat occur at an elevated level invigorating the storm even more. In all the other regions (1 and 3–5) glaciation occurs at much warmer temperatures in the range of -10 °C to -20 °C.
More information on the microphysical analysis of RGB colours may be found in: Rosenfeld D. and I. M. Lensky, 1998: Satellite-based insights into precipitation formation processes in continental and maritime convective clouds. The Bulletin of the American Meteorological Society, 79, 2457–2476 (see PDF, 1 MB).
Unfortunately no rain gauge or radar data could be recovered up to now that would better document this case. In particular, it would be interesting to know whether hail was detected somewhere (hypothesis is 'no hail'). The thunderstorm's high winds and heavy rain were witnessed by the participants of the Sahel-2007 Conference discussing activities of rain stimulation in the Sahel (see group picture, JPG, 664 KB, picture of the rain, JPG, 574 KB, source: H. Verschuur).
Some more details of the storm, including an outflow boundary, may be gleaned from a picture at 250 m spatial resolution from an AQUA-MODIS overpass (see 13:25 UTC channel 1 image, JPG, 447 KB, source: M. Setvak) . It may also be compared to animated sequences from Meteosat-8 HRV and Natural Colour RGB images (see HRV animation, MPG, 3 MB, and Natural Colour RGB animation, MPG, 2 MB).
Meteosat-8 HRV Image
Meteosat-8 RGB Composite Images
RGB Composite (Dust RGB)
IR12.0–IR10.8, IR10.8–IR8.7, IR10.8
Full Resolution (PNG, 836 KB)
Animation (05:00–21:00 UTC, MPG, 4 MB)
RGB Composite & T-Reff scatterplots
VIS0.8, IR3.9r, IR10.8
Full Resolution (PNG, 1 MB)