Severe thunderstorm over KwaZulu-Natal

By Lee-ann Clark and Kevin Rae (South African Weather Service)

Southern Africa is notorious for its high frequency of thunderstorms, particularly during the (Southern Hemisphere) summer season. A host of severe weather phenomena often accompany such thundershowers, including a few tornadoes.

An F4 tornadic event occurred at Mount Ayliff (Transkei, south of KwaZulu-Natal) on 18 January 1999. For the case study in question, fairly widespread thundershowers developed over the eastern highland regions of South Africa during the afternoon of 3 January 2005, triggered partly by high diurnal surface heating and partly by low-level moisture convergence over the high-lying topography of the Drakensberg mountains, forming the western border of the KwaZulu-Natal province (see geographic map source: SAWS). Under the influence of fresh, predominantly westerly winds aloft, the deep convective cells then migrated seawards over the interior of the province. One can also infer this from the thin cirrus anvil 'blowoff' (identifiable on the RGB and VIS image/s) to the east and south-east of some of the cumulonimbus cells.

A great number of severe storm reports flooded in, following the passage of the storms through the region. Severe aspects relating to the thunderstorms included many reports of heavy rain/flash flooding, hail as well as wind damage resulting from strong surface wind gusts. Greytown (in the central part of the province) reported 500 people left homeless when house roofs were blown off. Many businesses and shops experienced similar damage. Hail (of unknown size) was reported from widespread localities such as Pietermaritzburg, Cedara as well as at Durban International Airport (on the coast). Heavy rain was reported at Margate (south coast) as well as at Nottingham Road. Hilton, near Pietermaritzburg, experienced approximately 70mm of rain in an hour.

Eight people died after being struck by lightning in two separate incidents. Rivers in the Mhlahlani region of Ulundi burst their banks, while in the Pietermaritzburg area, two young boys were swept away by sudden flash-flooding of the river they were attempting to cross. One boy was rescued but his 11-year-old friend was overcome by the fast-moving torrent and drowned. Storm water drains in the surrounding suburbs overflowed, while an entire house was destroyed by a mudslide, trapping a youth inside.He was later rescued and taken to hospital.

Later that same evening, renewed thunderstorm development, much further west (closer to the surface dry line), resulted in scattered thundershower activity moving through Gauteng province (where Pretoria and Johannesburg are located). Very frequent and intense lightning activity was experienced, resulting in many power failures. Strong surface winds and gustiness were also experienced in the suburbs, toppling trees and accompanied by grape-sized hail.

Of note, however, is the relatively high temperature/dewpoint (T/Td) values being reported, with early afternoon (insolation maximum) surface temperatures of between 28 to 31°C and dewpoints of 16 to 19°C. Wet bulb values ranged between 24 to 27°C. These values (particularly the moisture variables) are extremely high, even for Southern African conditions. No well-defined 'tongue' of wet bulb values could be identified on this scale.

Surface inflow in the Greytown area was likely to be light north-easterly to northerly, underlying brisk (30–40 knots) west-south-westerly winds in the layer between 600hPa and 300hPa. Visually assessing the hodograph of the midnight (00:00 UTC) Durban balloon sounding, one would favour storm structure to most likely be of a 'multicellular' nature, rather than 'supercellular' (i.e. a hodograph profile showing little curvature with altitude, dominated by relatively fresh winds aloft; as opposed to a 'supercellular' profile in which speed AND directional shearing would be in evidence, with an anticlockwise (Southern Hemisphere) directional shift, given increasing altitude). The relatively linear hodograph at 00:00 UTC is also reflected in an 'almost negligable' 0–3km storm-relative helicity (S–RH) value.

The upper air soundings released at Durban, representative of 00:00 UTC and 12:00 UTC respectively (see graphs below) both exhibit pronounced free lift or 'bouyancy'. Both soundings suggesting in excess of 2000 Joules/kg of CAPE. It warrants noting that both soundings also suggest fairly generous mid to upper level instability, with a (parcel) Lifted Index of -5.8 and -5.7 for the 00:00 UTC and 12:00 UTC soundings respectively.

Bearing in mind the media reports of widespread surface wind damage, note the calculated WINDEX for both soundings. WINDEX is an estimation of (maximum) surface wind gusts in a dry downburst environment. The WINDEX for 00:00 UTC is 27 knots but rises to 43 knots, based on the afternoon sounding — entirely in keeping with observed events of surface wind damage and widespread gustiness. This, together with an absence of observed funnel vortices or 'twisting type' damage on the ground, further strengthens the likelihood of the wind damage having been of a (non-tornadic) 'dry microburst' variety. The reports of heavy rain and flash-flooding correlate well with a RAOB calculation of excessively high precipitable water loading, based on the 12:00 UTC sounding, namely 46mm. Generally, 25mm or more of precipitable water is considered to be a significant criterion when considering the likelihood of heavy rain.

Forecasters and severe weather enthusiasts will readily recognise all the classic signs of Charles Doswell's 'loaded gun' type sounding when viewing the overall shape or form of the 12:00 UTC Durban ascent. Extremely hot and humid at the surface (high temperature and dewpoint), with moisture extending up to (beyond) 60 hPa; then drying out rapidly aloft.

A conditionally unstable atmosphere, with moderate to high CAPE, indicative of extreme instability in the mid to upper regions of the troposphere, together with the presence of a low-level 'capping inversion'. In Doswell's analogy the capping inversion is the trigger of the gun. Once overcome or 'broken' (due to sufficient diurnal surface heating), the 'gun' is discharged; putting in motion the vigorous ascent of warm, moist unstable air; otherwise known as 'deep convection'.

A notable feature of the cumulonimbus (Cb) development over KwaZulu-Natal on the afternoon in question , is the occurrence of a clear, cloud-free zone over the KwaZulu-Natal Midlands, immediately westwards of the most active development. This feature is readily identifiable on both VIS and IR imagery (as a longitudinal corridor to the north-west of Durban).

In a weak to moderately speed-sheared environment, it is likely that storm qualities relating to 'bouyancy' may well have dominated storm development processes (with speed/directional shearing perhaps playing a lesser role on the day). Note that the surrounding area was becoming increasingly dominated by thick (anvil) cirrus outflow aloft, effectively cutting off heating (insolation) at the surface and maximizing reflectivity of the thick cirrus shield.

Typically, such Cb cells would become dominated by downdrafts, or at the very least, neutral buoyancy. By contrast, the very active thunderstorm cells (on the eastern side of the clear corridor) were still in full sunlight, with the summer Sun beginning to sink in the west. Re-radiated longwave radiation from the surface (within the cloud-free slot) was in all likelihood also sustaining the warm updraft towers. The steep, near-vertical structure of the upwind (western) side of the thunderstorm updraft towers can be inferred from the VIS imagery (even with some indications of 'backbuilding').

The animated RGB loop (see above) highlights the presence of 'overshooting tops', as the locally 'warmer' vigorous updrafts (small ice crystals dominating) were breaking through the dense cirrus/cirrostratus anvils (characterised by much larger ice crystals). The overshooting tops stand out as a more yellow against a background of anvil cirrus (predominantly orange/red in this particular RGB representation).

In terms of remote sensing (and especially Meteosat Second Generation (MSG) in this case), the ability for a weather forecaster to accurately identify overshooting tops, surface outflow boundaries and other similar (storm scale) features of thunderstorm development, will without doubt greatly enhance his or her ability to issue accurate (and timely) severe weather nowcasts.

Skew-T/Log-Pdiagram (RAOB) of Durban sounding, 3 January 2005 — CAPE in red, CIN in blue

Additional content

Map showing the geographic location of KwaZulu-Natal province (source: SAWS)
Channel 12 (HRV) Image (3 Jan 2005, 16:00 UTC)
Channel 12 (HRV) Animation (3 Jan 2005, 12:00–16:00 UTC)