In the last section we examined the origin of the basaltic melts that form the Hawaiian and Emperor seamount chains. The alkalic magmas appear to rise and erupt directly from the mantle, whereas the tholeiitic magmas form long-lived, large magma chambers at shallow depth within the shield volcanoes. Here we'll examine some ideas as to why these large shallow magma chambers develop. As you will recall, the ascent of magma through the asthenosphere probably occurs as discrete rising diapirs, looking something like tadpoles swimming towards the Earth's surface. The viscous nature of the asthenosphere allows the diapirs to rise by displacing and pushing to the side the surrounding material. Forces necessary to push the melt through the mantle are generated in several ways. The first melts at 60-80 km (we know this from deep earthquakes beneath Kilauea) probably form as thin films around unmelted grains of olivine and pyroxene, much like the film of water around grains of wet sand. This melt is driven together by capillary or "wicking" motions to form larger pockets of melt. Upon melting there is also a 2-5% volume increase and a corresponding reduction in density (solid mantle is 3.2 times denser than water while melt is about 2.8 x water). The volume increase forces the melt through plastic cracks while the density decrease makes sure the direction it is headed is upwards. The situation changes dramatically, however, at the base of the lithosphere. Unlike the asthenosphere, the lithosphere behaves more like a brittle material, necessitating a completely different mode of transport.
The figure on the left is a cutaway showing the relationship between magma ascent in the asthenosphere and the internal plumbing of an Hawaiian volcano. This figure and the next were taken from an article in National Geographic. The large arrow shows the direction of plate motion over the hotspot. As magma rises, it is temporarily blocked as it encounters the base of the lithosphere. This phenomenon is known as underplating. From here, a brittle material transport mechanism probably takes over.
One proposed model for brittle transport in the lithosphere is shown on the right. The white, curved arrows near the bottom of the figure are meant to suggest rising asthenospheric material, partially melting as it becomes shallower because of the effect of decreasing pressure on the melting temperature of rock. As the lithosphere bends downward under the weight of the volcanoes being built above, the lower surface is put into extension. Presumably this creates tensional cracks that allow melt to rise some distance into the overlying plate. Note that the magma path as it is ascends through the lithosphere is shown as being discontinuous. It is unlikely that the conduits feeding shallow magma storage structure within the volcanoes are continuous open pathways as is often shown in simple diagrams. Instead, as proposed by Hill, it is likely that magma passes between a series of expanding and contracting "cracks" that act as temporary, small storage reservoirs during ascent. As magma rises from one such structure to the one above by buoyancy as before, the lower crack collapses as the one above expands. It has been suggested that it is this exchange of material between adjacent pods that is the source of long period caldera earthquakes that are often seen during time of high magma supply rate. (long period earthquakes are look more like squiggles than sharp tectonic earthquakes) Magma continues to rise through the lithosphere, but the density contrast is much less (2.9 to 2.8). Eventually magma rises until it is equal to or heavier than the surrounding rocks. This causes magma to pool as a magma chamber within the volcanic edifice, as the overlying lava flows have lots of bubble holes and are very light (lower density). Also depicted is magma accumulating near the base of the crust, although the existence of a storage reservoir at this depth has not been established.
A sketch from Hill's paper illustrating the brittle ascent mechanism is shown on the right (note the inset showing the location of the cross-section. Melt is passed from pod to pod along a series of connecting shear cracks or faults. The zone over which this occurs is drawn as a funnel narrowing towards the surface. Seismicity (earthquake locations) beneath Kilauea do not reveal this widening with depth. Also, since the installation of a dense seismic network at HVO, no clearly defined seismic structure that could be associated with the vertical plumbing beneath Mauna Loa. Undoubtedly the truth is rather more complicated than these simple models suggest.
As we mentioned above, magma can only ascend when it is lighter than the surrounding rocks. At this depth the magma pools forming a summit magma chamber, the top of which is roughly 3 km beneath the surface of Kilauea Volcano. We can see the rough outline of the magma chamber as a zone without earthquakes extending from about 3-6 km beneath the summit. Because the magma is a liquid, you cannot break it, hence there are no earthquakes generated from within the chamber. Earthquakes that outline the chamber are generated by breaking of the surrounding rocks from pressure increases and expansion of water to steam in cracks. Most of these earthquakes are seen around the top of the chamber. The width of the magma chamber is probably close to about 3 km, from the width of the summit caldera. The bottom of the magma chamber is harder to quantify, but the absence of small, brittle earthquakes suggests the bottom may be at depths of 6-8 km. Somewhere around this same depth we see an increase in seismic velocities to about 8 km/second (at 8 km depth), a reflection of very dense material directly beneath the magma chamber. This very dense material is only found beneath the summit at shallow depths and has a density about equivalent to the mantle. One might ask "What is this stuff?", a very good question. Since it is unlikely that the mantle squeezed up the cracks with the melt, it must be something of similar density, thus probably made of olivine, the main mantle mineral.
So, as Bill Nye would say "Consider the Following": The melt that accumulates in the summit area of the volcano is a long way from home, where it was very very warm when it melted (perhaps 1400-1600 degrees Celsius). Well the magma chamber is very near the cold (for lava anyway) surface, so one would expect it to cool down before erupting. And in fact, the normal eruption temperature of Hawaiian lava is about 1150-1175 degrees C. Turns out you can use the amount of Mg (magnesium) in the glass (frozen lava) of flows to determine the temperature. These temperatures work out to about 7.5-8 % by weight of MgO in the melt. Well, recently, Dave Clague found some bits of lava glass on the deep ocean floor around Kilauea that erupted from Kilauea. These lava bits had MgO contents of around 14% MgO in the glass, indicating they formed at about 1500 degrees C. That's hot! This lava probably represents our best guess as to what the melted mantle is like. Being a clever sort of guy, Dave used the difference in MgO from the high MgO lava that we think came directly from the mantle and the low MgO mantle that erupts to form the volcano. He found out that about 15-20% of the original lava had crystallized as olivine in the summit magma chamber prior to eruption. Well this is a lot of olivine when you consider the size of these volcanoes (remember they are the biggest mountains on the planet). Olivine weighs about 3.3 x water, so it sinks like a proverbial rock through the 2.8 density magma. Having no where to go but down, the best explanation for all the dense stuff beneath the summit is that it is olivine. Now this is pretty cool, because it means that the lava doesn't actually even see the lighter 2.9 ocean crust on the way up, but is surrounded by very dense material all the way to the base of the magma chamber. This gives the magma a strong driving pressure all the way into the chamber. This is important, because if the magma just rose to the level where it weighed the same as the surrounding rocks, there would be no force left over to "inflate" or overfill the magma chamber and cause an eruption.
As magma continues to move into the summit area, the magma chamber expands as pressure increases. Obviously such expansion cannot continue indefinitely. Eventually the summit magma pressure becomes sufficient to overcome the strength of the surrounding rocks, and a dike begins to form. A dike is a crack filled with molten rock propagating through the volcanic edifice. Sometime these cracks move upwards and a summit fissure eruption occurs. Other times (slightly more frequently) the dike moves out horizontally, moving downrift along the axis of one of two radial rift zones running from the summit area. Eventually such dikes may intersect the surface of the volcano resulting in a rift zone fissure eruption as discussed in the next lesson.
Summit inflation and deflation is shown on the tiltmeter record on the right. The data was taken from an instrument near HVO which is oriented such that rising or inflation of the caldera produces positive (uphill) slope on the graph. The period shown runs from 1955 through 1994 and shows a slow inflation of the summit reaching a peak in about 1975, just after the eruption of Mauna Ulu shield ended. The big deflation (drop) in 1975 is related to the M 7.5 Kalapana earthquake. During this event, the entire southcoast of Kilauea slid seaward, causing the summit elevation to drop. The next big drop in late 1977 was due to the Puu Kiai eruption in 1977 that sent lava into the outskirts of Kalapana. Lava draining from the summit caused this deflation. Since 1975 the summit has been slowly deflating until now it is somewhat less than it was in the beginning. Perhaps one would like to conclude from this that we might expect the current eruption to slowly wane. However, the slow deflation seems more connected to a gradual seaward sliding of the volcano and lowering of the overall elevation rather than to depressurization of the magma chamber. This slow inflation and deflation is only one of the phenomena shown here. Superimposed on the general trend are numerous smaller changes, each associated with some kind of eruptive or intrusive event. For example, the deep valley on the left side of the plot shows the removal of summit magma during the 1960 eruption near the town of Kapoho near the Eastern tip of the Big Island. This eruption was interesting because it erupted for several days before the summit began to deflate, suggesting that there is a deep underground pathway between the summit and the lower rift that is filled with magma (i.e. it is a narrow extension of the summit chamber).
Summit fissure eruptions resulting from a vertically propagating dike are often associated with an initially positive tilt then negative change while rift zone fissure eruptions are almost always recorded as a sudden decrease in tilt reflecting a decrease in summit magma pressure. The increase in tilt during summit eruptions is related to the forcible intrusion of magma into shallow cracks above the chamber causing the summit area to lift up.
At even smaller scale other important features can be seen as is shown in the inset. Here very small changes in summit pressure associated with high fountaining near Pu`u `O`o produce a sawtooth effect. The positive ramps show the slow recovery of the summit following a previous episode of fountaining followed by a steep decrease as the next episode begins. These periods of high fountains, discussed in more detail later, occurred roughly once every month and lasted generally about a day or less. Notice that the summit pressure recovers approximately to its previous recent maximum before the beginning of the next high fountaining event.
The figure on the right is a very generalized sketch of the interior of Kilauea. The brittle feeder cracks are shown beneath the summit magma chamber. Note the relationship between the extent of the chamber and the location of the caldera bounding faults. The magma chamber expands as fresh melt arrives from below and deflates as melt is injected into dikes or erupted onto the surface of the ground. If the volcano is sliding seaward, a lot of cracks can be created underground for magma to be forced into as dikes and not erupt. The location of magma within the rift zones is shown toward the right with an eruption conduit suggesting the configuration for the current eruption at Pu'u `O`o. The overall supply rate for Kilauea has been determined to be about 0.1 cubic kilometers of lava per year as calculated by measuring the volume of erupted material over the last 20-30 years. Interestingly enough, this coincides nicely with Shaw's calculations of lava supply for the southernmost islands using the volume of the islands.
Extra Reading Clague, D. A., 1987, Hawaiian xenolith populations, magma supply rates, and development of magma chambers: Bulletin of Volcanology, Volume 49, p. 577-587
If you have comments or suggestions, email me at email@example.com