Tracks of cool, wet surfaces from precipitation

By Danny Rosenfeld (HUJI), Estelle de Coning, Evert Scholtz, Luis Fernandes, Louis van Hemert (South African Weather Service) and Jochen Kerkmann (EUMETSAT)

On 7 September 2005, shortly after 16:00 hours (local time), a severe thunderstorm with strong wind, hail and a tornado passed through the south-eastern Free State in South Africa. Eyewitness reports in 'Die Volksblad' from 8 September 2005 indicated that people saw the funnel cloud extending beneath the thunderstorm and it was accompanied by airplane-like noises. Hail as large as chicken eggs caused a lot of damage to farm buildings and vehicles. According to the farmers in the area 25mm of rain was measured at Verkeerdevlei and 20mm at Carnavon.

The convective storms can be seen in the Meteosat-8 images below from 7 September 2005. Note in particular the yellow to orange storm at 13:00 UTC, where the colour indicates the presence of small ice particles at the cloud top with reflectances of up to 11% in the IR3.9 channel.

Surface conditions on this day depicted a surface trough over southern Botswana, through the eastern parts of the northern Cape into a low pressure system over the eastern Cape interior. The distinction between the different air masses (dry v moist) was evident in the dry line (i.e. area where the dew point temperature gradient is extremely tight) over the southeastern interior. A dew point difference of 15K existed between Vryburg and Kuruman, and 11K between Bloemfontein and Kimberley. Moisture in the middle layers were in excess of 60% over the central and eastern Free State and vertical motion was enhanced by a jet stream of more than 100kt at 200hPa.

An interesting feature which could be seen well on the MSG imagery the day after the storms (see images from 8 September 2005), is tracks of wet and cool surfaces in the area of Bloemfontain, exactly where the convective storms passed through the previous day. They appear as stationary white streaks over the Free State in MSG Channel 07 (IR8.7, upper left) and as darker blue streaks on the RGB composite VIS0.8, IR3.9r, IR10.8 (bottom left). These streaks can also be seen on the previous day if one looks carefully at the animation of the RGB composite (see link below the upper right image).

Due to evaporative cooling, the precipitation tracks are about 8–10K cooler than the surrounding areas. The images clearly show where the convective storms started over the western Free State and eventually dissipated over the south-eastern Free State. Each rain shaft started at the head of a trail, expanded, merged with neighboring rain cells, expended more while weakening and eventually dissipating. The net effect is cooling the surface.

It should be noted that surface moisture effects are more evident at this time of the year since it is the end of the winter time and the surface is fairly dry and less vegetated.

Finally, the RGB VIS0.8, IR3.9r, IR10.8 and the brightness temperature difference image IR3.9–IR10.8 also show a number of fires (hot spots) further to the east, in areas where it probably had not rained. More information about the retrieval of land surface temperature and surface moisture can be found in the following publications:

• Thermal parameters extracted from diurnal cycles of land surface temperature (F.-M. Goettsche and F.-S. Olesen)
• Soil moisture and evapotranspiration Land SAF products (Keller et al.)
• Meteosat Second Generation data for assessment of surface moisture status (Sandholt et al.)

Summarising the above, the following four observations prove that the white streaks in the IR images are the tracks of the wet, cool surfaces and not thin cirrus clouds, dust storms or something else:
1. The tracks are stationary, because they reflect surface and not atmospheric properties.
2. The tracks appear with the diurnal heating, and vanish near sunrise and sunset.
3. The Dust RGB shows neither thin clouds nor dust in the area of interest.
4. During the afternoon of the previous day significant convective rain was observed over the area.

Additional content

Channel 01 (VIS0.6) (8 Sep 2005, 12:00 UTC)
Difference IR3.9– IR10.8 (8 Sep 2005, 12:00 UTC)
RGB IR12.0–-IR10.8, IR10.8–-IR8.7, IR10.8 (Dust RGB) (8 Sep 2005, 12:00 UTC)