Eruption Processes

Vent Location and Historical Flows

Here is the a digital elevation map of the south half of the Big Island. Mokoweoweo, the summit crater of Mauna Loa volcano, can be seen half way down on the left side of the figure with its two prominent rift zones. The summit of Kilauea and its East Rift Zone occupy the bottom right quarter of the image. Even though currently the summits of both of the volcanoes are occupied by deep calderas, it is clear that the highest parts of the volcanoes must represent the region of the most frequent and highest volume eruptions. Clearly the calderas must be transient features that get filled and then form again throughout the life of the volcanoes.

Next to the calderas, the most prominent features on the volcanoes are the ridges which extend to the east and the south of the summits of both Mauna Loa and Kilauea. The simple fact that these ridges exist suggest that they are important sites for additional eruptions, otherwise the volcanoes would have a symmetrical circular form. The decreasing elevation along the ridges suggests a decreasing amount of lava output away from the summit. When we look at these ridges they are also defined by vent structures and large crack systems. Linear regions of vents are referred to as "rift zones" by geologists, and the ridges are known as the Southwest Rift Zone and the East Rift Zone of Kilauea and the Southwest and Northeast Rift Zones of Mauna Loa.

This map shows the historical flows from Mauna Loa, Kilauea, and Hualalai during the past two hundred years. Several things should be immediately obvious. First is that most of the recent flows covering the area of the map originate along one of the major rift zones. They result from dikes that propogate along the rift zone before intersecting the surface and erupting as a "curtain of fire" or "fissure eruption". The exceptions are the radial eruption arising just north of Mauna Loa's summit, and the flows on Hualalai, which actually are on a rift zone of that volcano. Note how the flows from Mauna Loa's Northeast Rift Zone are topographically focused toward Hilo. The other point worth making is that although Kilauea is more volcanically active than Mauna Loa, judging from the areal extent of flows during the past two hundred years, the opposite would seem to be the case. Actually, since about 1955 Kilauea has been producing lava at a rate roughly twice that of Mauna Loa. However, during the preceding 150 years the reverse was true. Remember, conditions on these volcanoes are always changing. In geology we have a concept known as uniformitarianism, which holds that the processes active today were the same as those in the remote past. While this is very true over large time scales, the geological situation on an active volcano changes so fast that it is important not to assume that current activity is typical or even indicative of modal behavior even in the recent past.

This map shows a close-up of Kilauea's East Rift zone activity through about 1980. The current eruption, which began in early 1983 is not shown. The large lava field flowing from the rift zone near the summit is the Mauna Ulu flow field which was built from the sustained eruption of Mauna Ulu shield from 1969 through much of 1974. The Mauna Ulu eruption may not appear much bigger than the other rift zone eruptions, but in fact it produced about 50-100 times more lava than the others. Many of these flows were emplaced on top of each other, giving the appearance of a relatively normal sized eruption. The current eruption from Pu`u `O`o has produced about 3 times as much lava as the Mauna Ulu eruption (1.5 km3 versus 0.5 km3). These large rift eruptions represent a significant change in behavior on Kilauea during historic times. From 1950 to 1969, eruptive activity consisted mostly of short-lived summit eruptions followed within a few months by an eruption on the rift. Before that, from 1925 to 1950 there was very little activity anywhere on Kilauea. Lava may have been intruding underground, but there was little seismic data collected back then, making it impossible to track magma movements underground.

From about 1800, when the first records were made, to 1924 most of the activity of Kilauea was confined to Kilauea caldera. The main activity was concentrated around the site of Halemaumau, though at times the entire floor of the caldera was engulfed by lava. It must have been a spectacular sight. In 1924, lava moved underground to the Kapoho region (detected as a large swarm of earthquakes in the area) and probably out to sea on the submarine part of the East Rift Zone. The draining of this lava caused several major collapses of Halemaumau and the accompanying strong explosions blew out the rocks that cover the area around Halemaumau. Prior to 1790 (the exact time remains a mystery) 2 shield volcanoes covered the summit area of Kilauea and sent lava both eastward and southwestward for at least 5 centuries. There was no caldera at the summit during this activity. Exactly when the caldera formed is unknown, but it predates the explosive summit eruption of 1790 that killed a number of Hawaiian warriors crossing the summit region. In the summit area, particularly west of Halemaumau, you can see the explosive deposits draping the caldera wall. This indicates that the caldera wall already existed when the explosions occurred. Whether the caldera collapse precede this event by days, weeks, months, years, or decades remains one the great mysteries of Kilauea.

Kilauea Summit Activity

The summit of Kilauea Volcano is an interesting and dynamic place. Although many of the changes are happening almost instantaneously by geological reckoning of time, it is still slow from a human perspective. The summit slowly rises and falls controlled by differences in supply rates and extraction of material to feed eruptions and intrusions. Slowly over time layer after layer builds up, adding to the height of the volcano. This building is punctuated by abrupt collapses of the summit - forming calderas such as the one present today. This pattern of slow, protracted growth with intermittent collapses is a bit like three steps forward, two steps back. The long term growth rate for Hawaiian volcanoes taking both of these phenomena into account is roughly 1 cm/yr. The layers covering the summit are visible in the northern summit caldera walls shown here on the right.

This cliff formed during the summit collapse preceding the phreatic summit eruption in 1790. Note the gentle slope which is typical in the near-summit region of basaltic shield volcanoes. Also note the matching slope of Kilauea Volcano in the background. Compare this to the post-shield alkalic cones clustered on the summit of Mauna Kea in the background.

It is thought that summit collapse and large phreatic ash eruptions are closely related. It is hard to be certain of this, since it has only happened once historically. The theory is that the summit magma chamber is rapidly drained or deflated by a downrift eruption, possibly on the seafloor, or massive intrusion and it is no longer able to support its own "lid". As hot, water-saturated rocks fall into the magma chamber, high pressure steam is created that leads to an ash column towering 10 to 20 km above the summit. At present much of the summit is covered by a thick ash layer (the Keanakakoi Ash). However, considering the infrequency of ash layers in the caldera walls (only one is found at the base of the caldera and has been dated at 2200 years old), such eruptions must be somewhat rare, possibly occurring every few thousand years or so.

The current long-lived rift eruption represents a very different form of behavior than that of the recent past. From 1790 through 1924 the summit was occupied by a swirling lake of molten rock, as shown in this early painting by C.H. Hitchcock. The center of this activity has long been centered near the current location of Halema'uma'u, which is itself at the summit of a low, broad shield slowly building on the floor of the larger caldera. This can be seen by taking the short hike from the base of the cliff beneath Volcano House across the crater to the parking lot on the opposite side near the Southwest Rift Zone. Eventually this shield will grow large enough to overflow the summit region, as was apparently the case about 300-500 years preceding the ~1790 summit collapse. During this earlier period, lava flowed from a center near the current Kilauea Iki through the massive Thurston Lava Tube, covering the area currently known as Hawaiian Paradise Park (where your teacher lives). Another shield was present on the west side of the summit about where the Hawaiian Volcano Observatory sits today. This shield spewed lava to the southwest. It seems the current caldera is a relatively recent occupant of the summit, as are the chain of craters. So for the past 500-700 years, the summit of Kilauea has been the focal point of volcanic activity rather than the rift zones.

The scene was much the same when painted by Titian Ramsey Peale in 1842. Presumably this larger lava lake filling much of the current caldera had persisted since the caldera formed in the late eighteenth century. It seems odd to think of this kind of activity as typical, because it has been virtually unknown since 1924. As we shall see in later lessons, the current activity on the East Rift Zone is also very unusual for historic times. The drastic changes in style of activity lends credibility to ideas first presented by Robin Holcomb, that Kilauea undergoes distinct cycles of activity. It is possible that the current rift activity represents the reopening of the deep pathways beneath the East Rift zone that could lead to another catastrophic draining and caldera forming event in the future. Probably before that we might see a sequence of submarine eruptions as well. The focus of activity might also return to the summit when the entire system is pressurized, as was the case just before the 1790 eruption.

Kilauea Summit Phreatic Eruptions

In 1924 we got a small taste of some of the violent summit activity of which Kilauea is capable. This image shows the eruptive column that formed above the summit shortly after the summit magma lake drained, not to be seen again for 17 years. Although there were many earthquakes reported near Cape Kumakahi, the Easternmost point on the Big Island, no lava was reported on land or in shallow water off shore. Still, it is reasonable to speculate that lava was emplaced in large quantities in the lower rift zone, and that this emplacement then produced the seismic activity. While this event was much less energetic than the 1790 collapse, one observer was killed when he ventured too close. This kind of eruption is termed phreatic, if only ash from previous flows is involved, and phreatomagmatic if juvenile or fresh lava is extruded as well. Nearly all of the ash in 1924 was non-juvenile.

The sketch on the right shows one model that seems to explain the eruptive phenomena in 1924. Since Kilauea's summit is itself only about 1 km above sea level, this puts the top of the magma chamber in Kilauea roughly 2 km beneath sea level, so that the surrounding rocks may be saturated with water. At the time of the eruption, the summit lava lake was full as shown in the first panel. As lava drained away, saturated rock exposed during withdrawal was subjected to a dramatic and drastic decrease in pressure. As is well known, water under pressure boils at much higher temperatures. This is the secret to the functioning of a pressure cooker since under pressure the water boils at a higher temperature and thus the food cooks more quickly. As the pressure is released on saturated rocks (that may have been as hot as 700 C) pore fluids rapidly boiled and expanded 100-1000 times. This rapid expansion of steam fragmented the rocks to a fine ash and ejected it in a buoyantly rising cloud called an "eruption column". The eruption manifested itself as a serious of massive explosions, like a canon firing. With each blast, ash and rocks the size of small cars were thrown from Halema'uma'u. These eruptions went on periodically for several weeks in May 1924, giving us a hint as to the duration of the main caldera collapse around 1790.

Summit/Rift Connection

As pressure builds in the summit magma chamber, a thin blade of molten rock can be intruded into the summit area or the upper part of the rift zone. Since the mechanisms are similar, we'll only discuss the emplacement of dikes in the rift zone here. This melt-filled, propagating crack is known as a "dike". Generally they are about 3 km in vertical extent and about 3 to 10 m thick. They travel down-rift at a rate of a few km/hr. Eventually such a dike intersects the free surface somewhere along the axis of the rift zone, resulting in a "curtain of fire" several km long.

Rift Zone Intrusions

Here is a simplified sketch of the internal plumbing system of Kilauea's East Rift Zone. This figure outlines the classic view of magma behaviour at Kilauea. Once the summit is pressurized, dikes are thought to move down the rift zones by fracturing rocks until they reach the surface and form fissure eruptions. The emplacement of these dikes leaves a distinctive (and large) seismic trail.

Whatever the basic cause or character of shallow rift zone dikes, they do not go silently into the night. Basically dikes are a propagating, magma-filled crack. In the rift zone they are about 2 or 3 km in vertical extent, and the propagating crack widens to about 2 or 3 meters. As it propagates, the rocks in front of the moving crack tip are pried apart. This requires breaking rock under tension, and the process of breaking rock underground is what an earthquake is all about. The sound of rock breaking is transmitted through the rock of the edifice, and eventually recorded on any of several nearby seismographs (shown here). From the time it takes the sounds to propagate, each quake can be located to within about 100 meters, and the position of the leading crack tip associated with the dike can be tracked. The seismogram on the right shows part of the dike propagation associated with the current eruption just before the initial outbreak in early 1983. The small events towards the top of the record are the events we have been discussing. The large amplitude signal about half way down is known as tremor and occurs when magma is moving rapidly through the edifice. Tremor is somewhat analogous to the sound you hear from the water pipes in your house when someone is running water.

This style of dike formation is well documented for fissure eruptions in the upper half of the rift zone. However, in the lower part of the rift zone, eruptions are known to begin in the Kapoho area and eventually begin to draw magma from the summit. These eruptions produce strong earthquakes as the local fracture opens up as a fissure vent by breaking rocks near the surface. Within a few days lava begins to drain from the summit magma chamber to the eruption zone, but it does so without leaving a seismic trail. From this we infer that there is a deeper pathway already occupied with some magma under the rift zones. Otherwise, lava from the summit would have to force its way along fracture systems (creating a lot of earthquakes) to get to the lower East Rift Zone. This deep rift zone conduit probably varies in extent, but at times must reach all the way to the bottom of Kilauea (5000 m or 18000 ft below sea level) allowing large volumes of lava to drain on the seafloor. This leaves the summit area unsupported and presumably results in collapse and formation of the summit caldera.

It seems likely that much of the core of the rift zone remains molten between intrusions and eruptive events. Lots of the lava emplaced in the rift zone never erupts, instead it forms shallow underground magma chambers that can stay molten for decades. Some of these magma chambers can be quite large. Occasionally this rift zone stored magma may move further down rift or be erupted to the surface. In this case, as with the summit when there is a significant reduction in pressure, the overlying rocks are left unsupported and can "stope" into the voided chamber, leaving a pit crater on the surface. The image on the right shows Napau crater, the last of the significant pit craters down rift. The current eruption began here, and the most recent episode began here as well. The two smaller depressions are also pit craters that are somewhat off the axis of the rift zone, showing that the rift zone is not a narrow feature but has a width approaching about 1 km. We do not know for sure if the pit craters that form the "Chain of Craters" all collapsed in the same event that produced the summit caldera, but it seems pretty likely that a lot of them did.

As dikes propagate up or down rift, eventually they may intersect the ground surface. The image on the right shows a dike revealed in the walls of a pit crater that did actually erupt. The rift zone is comprised of thousands of such feeder dikes, although most have been deeply buried by more recent extrusive activity.

Rift Zone Fissure Eruptions

Fissure eruptions, whether at the summit or along the rift zone appear to be the most numerous of the kinds of eruptions that occur on Hawaiian Volcanoes. That is not to say that they are the most important volumetrically, because this is almost certainly not the case. The sustained eruptions on the rift and at the summit as well as eruptions confined to the caldera itself would seem to be more significant in this regard. However, almost all Hawaiian eruptions begin life as fissure eruptions. This makes sense if you think of the fact that as the lava rises to the surface it is cracking the volcano, so it should be no surprise if the initial eruption comes from a crack in the ground.

Fissure eruptions, called "curtains of fire" are certainly some of the more spectacular eruptions and people arrive from all over the world to revel in their grandeur. The eruption begins (from eye witness accounts) as a fine hair-line crack in the tephra covering much of the rift zone. This occurs as the dike arrives to within a few tens or hundreds of meters of the surface. Soon gas and steam begins to drift from the crack which becomes longer and widens slowly until eventually small clots of lava spit from the opening. Then lava begins to flow from the crack, vesiculate, and form a wall of erupting lava (shown in the figure) with individual fountains reaching 30 to as much as 100 m high. The flows that form from such effusive activity are generally a'a because of the loss of volatiles (gas) and heat as the lava is ripped into small blobs and tossed into the air. It seems strange that a fissure eruption should begin so passively -- one might have expected something a bit more grandiose. Perhaps this is true, because there have been very few observations of the dike actually reaching the surface and erupting. Most observations begin when the eruption has been in progress for several hours. Consequently, it is difficult to know whether or not this description is typical. Most likely it is!

Lava often issues from the initial fissure for several hours, sometimes oscillating between vigorous effusion and more subdued flow. Frequently lava will pool near the fissure, drowning the fountains, while flows continue unabated. The image here shows a mature, well-developed "curtain of fire".

This close-up of lava surging out a fissure during a rift zone eruption reveals some of the physics driving the eruption. As magma rises toward the surface, the pressure confining the fluid decreases. As this happens, water and other volatile compounds dissolved in the melt become saturated. Water in excess of the amount that can be dissolved in the magma forms steam bubbles in the rising melt, expanding it in volume, and making it much less dense. Both of these effects act to increase the vigour of the eruptive process. The steam bubbles that are too small to rise out of the lava as it cools become the small rounded vesicles that are so common on extrusive lava picked up on the surface of Hawaiian Volcanoes. The process is very similar to the bubbles that form in a soft drink (or hard drink, for that matter) when the cap is popped. Here as well, there is more of some volatile compound dissolved in the liquid than is possible, so the excess forms bubbles that rise to the surface. In the case of soft drinks, the gas is carbon dioxide. In volcanoes, it is water, sulfur dioxide, and carbon dioxide in that order of importance. The basalt that rises to form the Hawaiian Islands is 1% water by weight. As pressure on the lava decreases, it increases in volume by a factor of several thousand, expanding the melted rock into a foamy material. Most of the gas that is exsolved in this manner escapes at the surface and becomes part of Earth's atmosphere. The water vapor then condenses and becomes the world's oceans. Consider that almost all of the water found in the sea is either condensed volcanic gases or melted comets!

Volcanologists (this one included) like to play (work) near eruptive vents. This one is shielding himself with a metal plate from intense heat and spatter being thrown into the air. Most volcanologists have a certain amount of bodily damage associated with trying to occupy the same space as a blob of melted rock. As scientists go, volcanologists and geologists in general are somewhat short-lived. Hopefully you won't let this get around, because our insurance premiums are already more than we can bear. In any case, all danger aside, being a volcanologist is basically a really fun thing to be!

High Fountaining

Eventually, though, the pressure at depth driving the eruption wanes, leaving a glowing crack reaching deep into the earth belching sulfurous fumes. The image on the right shows a crack that became inactive just after dusk. Hydrogen gas issuing from the crack can be seen burning in this image. Sometimes the crack from a fissure eruption remains open for some period of time while magma pressure at the summit continues to build. When this occurs, lava may again well from the fissure, forming a new set of flows. If the feeding conduit is narrow or localized, this can also lead to truly one of the most impressive things a basaltic shield volcano can do -- high fountaining. High fountaining does not generally occur at the beginning of a fissure eruption, but tends to develop later in the eruption process -- generally after a period of quiescence. As lava reoccupies a narrow conduit and overflow, it begins to expand rapidly as before, but this time there is a clear path to several kilometers deep. All of the magma stored in the conduit begins to expand all at once, propelling melted rock into the air as if from a rocket engine pointed skyward. The vent itself is generally quite small, not more than a few tens of meters, but the jet of molten rock can reach 300 meters high.

Here is another volcanologist -- perhaps the same one as before. High fountaining tends to be periodic, lasting for several hours and then subsiding for several days to several weeks before repeating the cycle. The process is very similar to the cyclic behavior of a geyser, and the reasons for it are much the same. Before an eruption lava can be seen rising in the conduit until it is just at the point of spilling over the rim. This is generally cooler degassed lava that drained back into the vent at the end of the previous fountaining event. Once this plug of degassed lava is removed, high fountaining begins almost immediately, much like popping the cork on a champagne bottle. The volcanologist in the image on the right is monitoring the rise of magma in the conduit for the Hawaiian Volcano Observatory (HVO). Because of the regularity attending the periodic episodes of high fountaining, they are relatively easy to predict with high accuracy. The periods between fountains seem to increase with distance from the summit ranging from a few days to more than 1 month. The same is true for geysers, where the water rises into a funnel shaped pool which slows the rise of the free surface and consequently depressurizes the whole column. In the case of geysering the water in the column boils where in with high fountaining it exsolves, but other than that the two phenomena are remarkably similar.

This image shows the beginning of the high fountaining. As lava spills over the rims of the conduit the column begins to expand, pushing the lava into a dome fountain. Imagine holding a garden hose so that the water falls back onto itself. This stage lasts for a few tens of minutes to sometimes a few hours until the fountain becomes well-developed. The conduit shown here is roughly 10 meters across.

Typically once started fountaining lasts for several hours. There may be a very slow loss in height, although generally this effect it slight until shortly before the eruption ends. Occasionally fountaining will abruptly stop for a few minutes and then resume as though nothing happened. At the end of the eruption the activity stops rather abruptly. Volcanologists witnessing an eruption have reported looking away briefly and when they looked back the fountains had vanished. On other occasions the termination of an eruption is followed by a tremendous release of gas roaring from the conduit so fast that clots of partially congealed lava are ripped from the inside of the conduit and thrown violently into the air. This activity is very short lived, but impressive! Sometimes towards the end the fountain pulsates, oscillating between fountaining and gas getting, as if huge gas bubbles are being forced up the conduit. As can be seen, the details of fountaining are many and varied.

This night-time image of fountaining was taken with a long exposure, revealing the internal structure of the fountain. Lava is jetting from the right side of the image (East or downrift) and being blown uprift by trade winds. The result is that the constructional edifice tends to be offset from the vent. In the case of Pu'u o'o, the vent was quite low on the downrift side of the cone. Molten particles on the outside of the column fall back to the base and produced large lava flows that have lost a lot of their original gases.

When the pressure within the system is strong enough to push lava to the surface continuously a new eruptive pattern takes place. The trigger for this during the current eruption was actually the height of the vent on Pu`u `O`o. By the third year of the eruption the vent was 600 ft higher than it was when it started (the cone was 1000 ft tall!). The pressure that it took to force the magma out and begin each new high fountain eruption became greater and greater. Finally at the beginning of the 48th event, lava reopened part of the original fissure, now about 700 feet lower than Pu`u `O`o and lava began to pour out all the time. Instead of having to wait a month for sufficient pressure to begin forcing magma from the vent. lava is continuously welling up and overflowing the vent system. When this occurs the lava piles up near the vent and forms a small shield volcano, like a minature Kilauea or Mauna Loa. The shields even have a minature crater filled with a lava lake.

Lava flows produced during this type of sustained eruption are fed by a lower rate of lava and are more likely to form pahoehoe flows. Pahoehoe lavas build long tube systems capable of transporting large volumes of lava for great distances (up to tens of km). The lava vent moved back to Pu`u `O`o in 1992, but found it a quite different place. The narrow conduit had collapsed leaving a large crater that lava could fill up at a lower level. Cracks in the base of the cone allowed lava to leak out to the surface continuously and begin to form shield structures around the original cinder cone. Today the crater keeps getting bigger due to collapse of the cinder cone and a few of the shields are getting close in elevation to the lava lake. Though we've had a few low spatter cones and dome fountains form at the base of the cone, it is very unlikely that we will ever see high fountains again because the narrow "gun-barrel" style of conduit necessary was destroyed long ago.

Examination Questions

1. Describe the mechanism of caldera collapse and phreatic eruptions at Kilauea's summit.
2. Describe how dikes move from the summit area and are emplaced in the rift zone. What evidence from dike emplacement and fissure eruptions in the lower rift zone suggest that there is a much deeper magma filled conduit in the rift zone?
3. Describe the 3 principle changes in eruptive style and vent morphology in a long-lived vent system. Why do these changes occur?