Origin of the differentiated and hybrid lavas of Kilauea Volcano, Hawaii

Abstract

Kilauea Volcano has erupted lava from its summit caldera and from two rift zones that extend from the summit towards the east and south-west. Lavas erupted from the summit of the volcano differ from each other principally in their content of olivine and define lines of ‘olivine control’ on magnesia variation diagrams. Lavas erupted on the rift zones may be similar in composition to the summit lavas or may be differentiated by processes that involve minerals other than olivine. All of the differentiated lavas have less than 6·8 per cent MgO and plot off the extension of olivine control lines for the summit lavas. Prehistoric vents (before A.D. 1750) from which differentiated lavas have been erupted are found on the east rift zone and in the western Koae fault zone adjacent to the south-west rift zone; historic vents for differentiated lavas are confined to the east rift zone. Twenty-one new analyses are presented for several of the east rift differentiates and for the newly discovered differentiates adjacent to the south-west rift zone. The differentiates have MgO as low as 3·9 per cent and SiO 2 as high as 56 per cent; both extremes are found in the prehistoric lavas adjacent to the south-west rift. Detailed petrochemical studies suggest the following conclusions: The chemical composition of magma erupted at Kilauea summit varies with the date of eruption. Lavas erupted before 1750, during the eighteenth and nineteenth centuries, and in the twentieth century form groups that can be distinguished chemically. On a lesser scale, each Kilauea summit eruption in the twentieth century has a chemistry that is distinctive with respect to the chemistry of every other summit eruption. During late prehistoric time pockets of differentiated magma were formed within the rift zones by separation of the liquid remaining after partial crystallization of bodies of summit magma. This process presumably is still going on within the east rift zone, but the more recently separated liquids have not yet been erupted to the surface. The relative time at which these differentiated magmas were produced can be estimated from calculations based on their chemical compositions, which show that the differentiates could lie on the liquid line of descent for Kilauea summit magma of prehistoric composition but not on any liquid line of descent for younger summit magmas. Lava from some eruptions, notably the early part of the 1955 eruption on the lower east rift, has the composition of the liquid fraction as it is generated within the rift. Lava compositions of other eruptions, including those of the later lavas of 1955, are best explained by mixing of magma supplied from a central reservoir beneath Kilauea summit with the differentiated liquid in the rift. Lava from each summit eruption is unique chemically, so it is possible to recognize its presence or absence as components of mixing in such mixed lavas. It appears that summit magma of composition characteristic of the 1952 and 1961 Halemaumau eruptions contributed to the composition of the mixed lavas produced in the latter part of the 1955 eruption. Summit magma of 1961 composition is alone sufficient to explain the composition of mixed lavas erupted in 1960 and 1961. In rift lavas erupted from 1962 to 1965, the composition of lava erupted in Halemaumau in 1967, in addition to the 1961 composition, is a component of mixing, and it is the dominant summit component in the composition of the two 1965 eruptions. The proportion of summit magma to differentiated magma needed to explain the composition of lavas erupted on the upper east rift increases from 1961 to 1965; this increase indicates that the differentiated magma was being diluted and used up by repeated flooding of this part of the rift zone by magma supplied from the central reservoir. The fact that components of ‘summit composition’ appear in rift eruptions before they appear undiluted in Halemaumau suggests that the central reservoir is vertically zoned. Rift eruptions are fed from lower levels where younger magma is available, and summit eruptions are fed from the relatively older magma above. The chemical distinction between lava of successive summit eruptions implies that significant convective mixing of magma does not take place throughout the central reservoir. The unique and uniform composition of lava of each successive summit eruption also suggests that summit eruptions end when all of the magma of one composition has been erupted. The magma erupted from the upper levels of the reservoir during one cycle is continually replaced from below by younger magma of different composition. In order for eruption to be renewed in Halemaumau, new magma from the mantle must be held in storage at intermediate levels before it attains an ‘eruptive state’. The hypothesis presented in 2–4 above permits qualitative predictions concerning future lava compositions. The composition of the next lava to be erupted in Halemaumau is expected to be distinct from that of the 1967 eruption, and this composition will presumably be identified in rift eruptions occurring between 1967 and the time of its appearance in Halemaumau. Differentiates of prehistoric age also were apparently formed in the same way as those of historic age, but the mixing cannot be described quantitatively because of poor control on the stratigraphy and the compositions of erupted lavas. One lava in the Koae group, that from Yellow Cone, appears to be a mixture of a picritic magma (12 per cent MgO) with a differentiated liquid with less than 2·5 per cent MgO and nearly 60 per cent SiO 2 .