Lava-Cooling Operations During the 1973 Eruption of Eldfell Volcano,
Heimaey, Vestmannaeyjar, Iceland
U.S. Geological Survey Open-File Report 97-724

Title Page

Table of Contents

Editor's Introduction

Lava Cooling on Heimaey: Methods and Procedures

Appendix 1

Appendix 2

Thorbjörn Sigurgeirsson
Science Institute, University of Iceland
IS-107 Reykjavík, Iceland


When lava fields in Iceland are examined, it becomes evident that lava flows can be quite varied, and that the area covered by each is different. Sometimes lava will flow for a distance of 100 km, while at other times it does not extend 1 km beyond its source (fig. 4).

Figure 4. Branching lava flows at Heiðmörk, southeast of Reykjavík in southwestern Iceland

Figure 4. Click on figure for larger image with caption.

Among the factors that have an important effect upon the areal extent of lava flows are the viscosity of the lava mass, its total volumetric output, the speed of its formation, and the geomorphic character of the terrain. A steep slope permits the lava to flow farthest from its source, but experience has shown that lava can still flow very far on nearly flat ground, as, for example, in Flói [slope gradient of 1:500; the region between the Ölfusá and Þjórsá (rivers) in southwestern Iceland]. The lava outflow which occurs during a volcanic eruption becomes centered in a given place, and, as is well known, this unavoidably causes changes in the previous landscape. In some cases the lava comes to rest on the terrain as an extensive lava field; in other cases it piles up in the vicinity of the crater.

It is the viscosity of the lava that governs its ultimate movement. Viscosity depends upon the chemical composition of the lava and the temperature of the molten lava; viscosity increases quickly as the lava cools. The mobility of basalt is at a maximum at temperatures of 1,000 to 1,200ºC when it emerges, but when the lava's temperature has decreased to 800 ºC it firmly solidifies and does not flow any more. How far the lava reaches before it solidifies depends, therefore, upon the velocity of flow and the rate of cooling. [Editor´s note: Volatiles are also an important factor.]

In some cases, cooling begins before the lava erupts from the crater. This happens in submarine eruptions and in subglacial eruptions, where water has access to the eruptive fissure or crater vent. Contact between the water and the viscous magma is dominated by gas bubbles released from the lava that combine with steam and boiling water. Cooling could then be so rapid that no molten lava manages to flow out from the vent, but, rather, the lava is ejected as scoria or [finer-grained] ash [all airborne ejecta is collectively called tephra] from the eruptive vent of fissure. Examples of this are the eruption of Surtsey, where the sea had direct access to the eruption fissure and crater vent, the eruptions of Katla (a subglacial volcano under the eastern part of the Mýrdalsjökull [ice cap] in southern Iceland) and also the many palagonite (hyaloclastite) ridges that formed from volcanic eruptions during the Pleistocene [and Holocene Epochs, collectively known as the Quaternary Period, a span of 1.6 million years]. Some minor cooling could also take place when fragments of lava are thrown into the air and then fall back into a crater.

After lava has flowed from a crater [or fissure] and expels most of the gas, its agitation diminishes, and cooling slows down. Because of the solid crust that develops on the surface of lava flows, a powerful lava stream flowing forward into the sea can move for a distance along the seafloor just as on land. A slow lava stream, on the other hand, cools when it reaches the sea. A piling up takes place at the coast, and the molten lava is forced to flow on top of the pile-up. Below the surface of the sea, a lava-flow front has the slope angle of scree.

Back to Top

One sort of natural lava cooling is the so-called "pseudo" volcanic eruption that is known from places where lava flows into very shallow water, or over marshy ground (for instance, at Skútustaðagígar on the southern shore of Mývatn in northeastern Iceland, or the area around Kirkjubæjlig;jarklaustur in southern Iceland). In such cases it is presumed that water is trapped inside the lava or beneath it, and that steam escapes upward into the molten lava proper, where it is often expelled violently, creating groups of pseudocraters. [There cooling occurs in a similar way as craters that receive water from the upper surface.]

Another type of natural cooling was evident during the volcanic activity on Heimaey. At certain places that had previously been covered by the sea, vapor steamed continuously out of the lava. It was evident that in such areas the cooling of the lava impeded the flow, because the leading edge of the lava became indented, behind which the lava piled up due to the pressure from behind. Information regarding this phenomenon is recorded on vertical aerial photographs.

Efforts to reduce damage caused by the force of the lava flow have been made several times, notably in the form of the construction of dikes or ramparts. The protection measures at Heimaey are undoubtedly the most extensive that have ever been used in a volcanic eruption. The chief reliance was upon cooling by water. This method has been tried previously on a small scale in Hawaii in about 1960, where the spraying was done directly on the lava margin and was considered to have produced results [Editor's Note: Bolt and others, 1977; Blong, 1984; Macdonald and others, 1986]. Perhaps an experiment of this kind was also [said to be] undertaken at Mt. Etna (on Sicily, Italy) a few years ago.

During the volcanic activity on Surtsey, a team was organized to test the effect of pumping water onto the lava flow, but the team was forced to turn back because of unfavorable weather. Sea-water pumping was later tried on a much smaller scale, but this test yielded no results that could be relied upon. The effect of sea-water cooling upon the lava flow at Surtsey was, nevertheless, evident, because the lava flowed for a long distance along the beach and because the surf cooled the lava front, thereby created a protective wall.

Indeed, the effect of water on molten lava has been noted before the last decade (1960's). The biography of the clergyman Jón Steingrímsson indicates that it was evident to him that water did retard the flow of lava. In his account of the Fire Mass held at Kirkjubæjarklaustur (central part of the south coast of Iceland) on the 5th Sunday after Trinity in 1783, he states inter alia: "... then God was invoked earnestly, and His judgment pronounced that the fire should not come the width of a foot farther than it has been before the divine service was celebrated, but it should pile up on top of itself in a mound. Thereby all neighboring waters came down upon it, and suffocated it most decisively." [Pastor Jón Steingrímsson was referring, of course, to the Skaftáreldahraun (or lava flows associated with the famous Laki eruption of 1783)2].

Back to Top

Nature of the Lava Cooling

Rock is a poor conductor of heat, so that a lava slab 10 m thick is known to be hot for years, if it is [solidly] intact and not fractured, even if its surface is continually cooled by water. Generally, however, lava is fractured; joints are formed during the abrupt cooling that has taken place.

Water has been found to be the most practical means to achieve lava cooling. Water absorbs heat from the lava; even more so, if it heats up to the boiling point and changes to steam. For most efficient cooling it is important that all of the water should be converted to steam and not flow off, down through, or away from the lava. The experience at Heimaey, where the molten lava was actually covered by a thick layer of tephra and scoria shows that near maximum efficiency is often attainable. Jets of water pouring on the same place in the lava flow at the rate of 100 l s-1 for several days or weeks, showed scarcely any diminution of steam generation. At first the steam generation was intense in the vicinity of the water jets. After a while, evaporation near the jets dropped, but internal movement of water-generated- steam extended in all directions, so that, eventually, one hectare (ha or 0.01 km2) of lava surface could be covered with a single stationary stream of water.

Under these circumstances, it is fairly easy to calculate approximately how large a lava mass will solidify due to water cooling. As the water heats up from 10ºC to a temperature of 100ºC, the steam absorbs 630 kilocalories for each kilogram of water that evaporates. Each kilogram of lava that solidifies and cools from 1,100ºC to 800ºC gives off about 190 kilocalories, and it is safe to estimate that lava that has solidified and cooled all the way down to 100ºC releases a total of about 380 kilocalories. This means that each kilogram of water cools 1.7 kilograms of lava, or that each cubic meter of water cools about 0.7 m3 of lava. This implies that a mass of lava cooling down to 100ºC as the result of contact with an ample source of water rapidly reaches this temperature on its surface. Furthermore, if water cannot enter [easily through] cracks and joints in the rock, cooling proceeds from the outer to the inner parts of the rock before the temperature has fallen below 100 ºC.

Back to Top

From the above considerations and figures, one can calculate, on the basis of a water stream flowing at the rate of 100 l s-1 over a 6,000 m2 area of molten lava, under average cooling conditions and conversion of all water to steam, that a solid layer of lava about 1 m thick will form in 24 hours. This cooling rate has been confirmed by measurements taken in drillholes bored into the lava after the water cooling ended [Valdimar Kr. Jónsson and Matthías Matthíasson, 1974]. In the boreholes, the cooling extended down to depths of 13 to 15 m, after cold [sea] water had been poured on the lava for about two weeks. A hole was bored in an area which had not been cooled by water, and molten lava was encountered directly below the surface.

Although water cooling is a comparatively slow process, it would still require a period of time a hundred times longer in order to cool lava to a sufficient depth, if we relied only on the thermal conductivity of the rock.

Therefore, for water cooling to be effective it is concluded that water must penetrate down into the lava through cracks and joints that form during the [initial] cooling period. This conclusion was confirmed when lava was removed from the terminus of a lava flow that had flowed as far as the fish-cannery plant near the harbor (see fig. 22). Fracture planes within the artificially cooled lava showed salt encrustation precipitated in all the joints where steam had been generated. Even adjacent to the lava margin the salt encrustation reached down through the lava, showing that the cooling water had penetrated the lava to the bottom. Wherever the cooling water makes contact, the lava also exhibits many fissures. The same could be seen in the lava at Kirkjubær (east of the new hospital) that experienced significant surface displacement during the cooling operations, resulting in the opening of large fissures well into the interior of the lava flows. Apparently water cooling proceeded quickly at depth when lava was in motion, because its motion opened cracks and fissures before cooling was complete.

Previous Page

Title Page

Back to Top

Next Page

[an error occurred while processing this directive]