Volcanic eruptions of Mount Karthala

By Anina Cirillo, Cecilie Wettre, Jochen Kerkmann and Gordon Bridge (EUMETSAT)

November 2005

The Karthala volcano forms most of the landmass of Grande Comore (also called Ngazidja), the main island of the Union of the Comoros. It is situated at 11.75 S and 43.38 E in the Indian Ocean and is one of the largest active volcanoes in the world. Over the last two hundred years, it has erupted every 11 years on average.

According to the Kathala National Observatory, the eruption on 24 November 2005 was probably phreatic in its initial phase. Phreatic means, caused by vaporisation of ground water in contact with a shallow magma body underneath the crater. In these cases water vapour forms immediately, with the resulting explosion breaking the surrounding rocks and throwing them into the caldera.

In the capital, Moroni, there had been ash falling from two o'clock (local time) in the morning on the 25th, with near zero visibility. At eight o'clock in the morning the sky was as dark as it would have been at seven o'clock in the evening. Ash deposits in the streets were as much as 10cm thick in some places.

Very significant amounts of dust and ash were released during this eruption, which were largely carried to the south-west area of the island. The eruption killed at least one child, infiltrated homes, shops and offices and contaminated domestic water supplies, and this during the height of the dry season. In fact, water supplies for about 120,000 residents, mainly from rural villages near the volcano, had been contaminated by the ash, which had also raised fears of an increased number of respiratory ailments.

The eruption which started around 18.00 UTC on 24 November entered the second phase (the magmatic phase), in the afternoon of the 25th, with the appearance of a red cloud above the caldera, caused by the formation of a lava lake. During the magmatic stage, the lava lake replaced the earlier water lake in the crater. Additionally, the head of the Meteorological Department of the international airport of Karthala reported that several international flights (Air Tanzania, Air Austral) and several local flights were cancelled during the period 26–27 November.

Even when air traffic resumed late on the 27th, a local flight coming to Moroni from Madagascar had to return to its departure airport because it encountered a large ash cloud on the way to Moroni. The Karthala Volcano Observatory reported that the eruption ended, after 14 days, on 8 December 2005.

As there has not been any significant cloud coverage in the area, it was possible to monitor the Karthala eruption using Metosat-8 images right from the beginning of the event. The eruption can be best observed in the so-called Ash RGB combination image, using the channel combination IR12.0–IR10.8, IR10.8–8.7 and IR10.8 (see upper left image and its recipe/interpretation).

This RGB exploits the fact that thin ash clouds tend to have a positive BTD between IR12.0 and IR10.8, while water/ice clouds (in particular thin clouds) have a negative BTD. Thus, thin ash clouds have a strong reddish colour, whilst water/ice clouds have less red, in particular thin ice clouds which appear very dark. In the case of the Karthala ash cloud, the BTD between IR12.0 and IR10.8 reached values of about +3K (in the early phase of the eruption). In other volcano cases even higher values (around +10K) have been observed.

The second important component of the above mentioned Ash RGB is the BTD between the IR10.8 and IR8.7 channels. This BTD helps to detect thin clouds (in particular ice clouds are more transparent at IR8.7 than at IR10.8), but it is not so useful for discriminating ash clouds from water/ice clouds.

However, as described in the case study Sulphur Plant Explosion over Northern Iraq, due to a (weak) SO₂ absorption band at around 8.7 microns, the BTD between the IR10.8 and IR8.7 channels allows one to detect (pure) volcanic SO₂ plumes. Detection depends, of course, on the purity of the SO₂ cloud (mixed SO₂/ash/ice clouds are difficult to detect in this combination), on the SO₂ concentration and on the temperature difference between the underlying surface and the SO₂ cloud.

In general, the larger the latter and the larger the SO₂ concentration, the stronger will be the BTD signal. It is not yet clear which is the minimum SO₂ concentration (e.g. expressed in Dobson Units) detectable with MSG. Extreme cases like the sulphur plant explosion over northern Iraq in June 2003, or the eruption of Nyiragongo, in July 2004, produced very strong signals in MSG imagery, but also the much weaker signal from the Karthala eruption could be easily monitored. In the animation of the Ash RGB, the SO₂ cloud is visible during the second phase of the eruption with a green colour.

The BTD between the IR10.8 and IR8.7 channels reached values of around +5.5K, lower than in the Nyiragongo case (+7K), but still high enough to fall outside the normal range for cloud-free areas (between 0 and +4K, depending on the humidity and the surface emissivities).

The SO₂ retrievals from SCIAMACHY (see link under 'See also') from 26 November at 07:10 UTC also clearly show the SO₂ cloud with a relatively low maximum concentration of around 3 DU, although this result is to be considered more qualitatively as no fine tuning has been performed. SO₂ retrievals have also been performed using MODIS and AIRS data, giving a higher maximum SO₂ concentration of 8.3 DU (see links under 'See also'). More details about the Karthala eruption are available in the Bulletin of the Global Volcanism Network.

As can be seen in the images below, the IR3.9 channel is also very useful for the detection of volcanic ash, especially during day-time, because of the relatively high reflectivity of volcanic ash in this channel (as compared to ice clouds). It also allows one to detect the heat signal from the eruption (hot spot). However, during night-time this channel is not as useful for ash detection because (thin) ash and (thin) ice clouds give more or less the same signal. Therefore, the IR3.9 channel has not been included in the Ash RGB product, although it is very useful for ash cloud detection duriing day-time (and may be also for SO₂ detection during night-time).

An example of a RGB product using the IR3.9 channel is shown in the list under 'See also'. It shows a strong difference in colour between the volcanic ash cloud (green) and the ice-dominated cloud (red). The latter cloud probably represents the steam cloud, which was pushed high into the atmosphere at the beginning of the eruption when the initial phreatic eruption produced a huge mass of steam as the lake that had formed in the caldera evaporated. This volcanic steam/ice cloud is partly mixed with ash particles and SO₂, and it would be interesting to study the effect of the ash aerosols on the microphysical properties of this cloud.

Similarly, it would be worthwhile to study the effect of the ash aerosol cloud on the (deep) convective clouds over Madagascar. Looking at the animation of the Ash RGB (in particular, when animating at high speed), it appears as if the ash cloud 'suppresses' the convection in northern Madagascar while, further south, in the area not influenced by the ash cloud, convective activity remains strong. This would indicate some suppression mechanism by volcanic ash cloud on convective clouds over northern Madagascar, possibly a rapid collapse of convection caused by enhanced rain-out. However, this hypothesis needs detailed study of the microphysical properties of the clouds over Madagascar.

Finally, the height of the volcanic ash cloud can be determined using trigonometry and an early-morning visible image. The HRV image from 03:00 UTC (see link under 'See also') shows a distinct shadow of the volcanic cloud on the ocean surface. If L represents the length of this shadow and Alpha the angle of the Sun above the horizon, then the height H of the cloud can be determined from: H = tan(Alpha)*L (neglecting curvature of the Earth's surface). From L being about 111 km at 03:00 UTC and Alpha 5.3 degrees, it follows that the height is 10.45km. Probably, this value does not correspond to the maximum of height of the ash cloud during the eruption, but it is in good agreement with the maximum height estimate of the US Air Force Weather Agency (ash to FL 380 = 38,000 feet, approximately 11.6km, see MODIS image under 'See also'). Unfortunately, due to the strong sun glint in the area of interest at this time of the year, it was not possible to do further height estimates at later stages (the shadow was simply not visible).

May 2006

A new eruption of Karthala started on the evening of Sunday 28 May 2006 at 18:05 local time, and could be observed in Meteosat-8 images.

Fresh magma formed a lava lake inside Chungu Chahalé crater in the caldera, lighting up the sky, and sending scores of frightened residents onto the streets. The NIR1.6 and IR3.9 channels of Meteosat-8 clearly show the presence of the lava.

No explosive activity was reported at the beginning of the eruption, only a thick plume of gas and steam above the crater, extending 60km to the northwest, over the Mozambique channel. A large SO₂ plume can be monitored with Meteosat-8 RGB imagery, using the Ash RGB composite image.

The 2,361m Mount Karthala and its forested slopes form most of the land mass of Grande Comore, the main island in the Comoros chain which lies 300km off east Africa. The Karthala Volcano is notoriously active, having erupted more than 20 times since the 19th century.

The last eruption at Karthala occurred in November 2005 (the second eruption that year), when the volcano belched out huge amounts of ash, polluting water supplies and leaving 250,000 people without clean drinking water.