Evolution of Ocean Island Volcanoes
Dennis Geist, Department of Geology and Geological Engineering, University of Idaho, Moscow, Idaho 83844
Wendy Bohrson, Department of Geology, Central Washington University, Ellensburg, Washington 98926-7418
Karen Harpp, Department of Geology, Colgate University, Hamilton, New York 13346
The GSA Penrose Conference "Evolution of Ocean Island Volcanoes" convened in the Galápagos Islands, Ecuador, June 412, 1998. The conference was cosponsored by the Charles Darwin Foundation and the International Association for Volcanology and Chemistry of the Earth's Interior. An international group of 65 scientists met in Darwin's islands to assess our current understanding of the evolution of ocean island volcanoes, to forge new interdisciplinary alliances, and to evaluate the course of future research.
Field Excursion and Keynote Presentations
The conference began with a five-day field trip to examine ocean islands in various stages of volcanic evolution. Several geologists who have had field experience in the Galápagos, as well as three Galápagos National Park naturalist guides, led the trip. Keynote talks were given between stops.
San Cristobal, the first stop, lies at the older, eastern end of the archipelago. It has experienced numerous Holocene eruptions and has erupted more diverse magma types than most ocean island volcanoes, ranging from mid-ocean ridge-like tholeiites to more typical alkaline basalts. Tectonically, the Galápagos hotspot is unusual because it lies adjacent to the Galápagos spreading center. Dennis Geist and Karen Harpp discussed the complexities that may have arisen from the conjunction of hotspot and ridge magmatism, including non-age-progressive volcanic activity, geographically widespread eruptions of similar age, a diversity of rock types atypical of ocean island volcanoes, and unusual spatial patterns displayed by isotope ratios and trace elements. Another stop, Espa-ola island, also lies at the eastern end of the archipelago. Unlike San Cristobal, however, activity on Espa-ola ceased ca. 3 Ma, and the island is strongly disrupted by normal faults.
Alexander McBirney described the history of geological exploration of the archipelago, focusing on the writings of Darwin. At the time of his voyage, Darwin had a keen interest in geology-petrology in particular. On the basis of astute field observations, he addressed issues that continue to be debated today, such as the role of crystal fractionation, the tectonic alignments of the islands, and hydrovolcanism. Mark Kurz reviewed the controversial interpretations of helium isotope data from the Galápagos and other ocean islands, especially the meaning of temporal changes in the helium isotopic ratios at certain volcanoes. After presenting several hypotheses for the unusually high abundance of primordial 3He in ocean island basalts, he suggested that these rocks provide modified samples of the lower part of the mantle. Some participants raised concerns about these interpretations in light of radiogenic isotopic evidence for ancient subducted lithosphere as a source of ocean island magmas.
Bob Reynolds led the trip to Sierra Negra, an active volcano on the westernmost island of Isabela. The volcano summit offered impressive views of the caldera and the remarkable sinuous ridge that may be the result of recent, shallow intrusion beneath the caldera floor. Wendy Bohrson summarized the different evolutionary styles of ocean island volcanoes the world over. Although there are broad unifying themes, there is also an extraordinary diversity in terms of their style of eruption, petrologic character, and mantle sources. Clearly, it is misleading to consider any single ocean island chain as "typical." Notably, although Hawaii is commonly considered the basis of comparison for all other ocean islands, volcanism there is probably best viewed as an extreme end-member-i.e., the most vigorous hotspot on Earth. Peter Mouginis-Mark presented a synthesis of planetary volcanism, including spectacular images of volcanoes from Venus and Mars. He proposed that Galápagos shields may be the best analogues for some extraterrestrial volcanoes.
At Alcedo, another of the active western shields, trip participants saw fallout deposits from the archipelago's largest siliceous eruption. Geist and McBirney debated whether these highly evolved rocks originated by crystal fractionation of a basaltic magma or by crustal anatexis. Dave Bercovici presented a keynote talk on important recent breakthroughs in fluid dynamic models of mantle plumes. Perhaps most important are conclusions about entrainment of mantle material during plume ascent. Previous modeling may have suggested entrainment because of unrealistic starting conditions in the models; in the more realistic case where plumes originate at a finite boundary layer, results indicate that they do not entrain surrounding material to significant extents. This is important because geochemists have called on entrainment to explain much of the diversity observed at hotspot volcanoes.
The focus at Santiago volcano, in the central, thus middle-aged, part of the archipelago, was on surge-deposited tephra and shallow-submarine geology. Mike Garcia presented an update of recent work on the Hawaiian Islands, including evidence for crustal assimilation by basaltic magmas, petrologic processes in the submarine (Loihi) stage, and time-series data from the ongoing Puu Oo eruption. Bill White gave an overview of isotopic constraints on ocean island volcanism, the main point being that there is an emerging consensus among isotope geochemists that ancient subducted lithosphere is reactivated to form ascending plumes, eventually constituting the principal source of most ocean island magmas.
On day five, the participants disembarked on Santa Cruz island, the location of the Charles Darwin Research Station, our hosts for the remainder of the conference. The scientific program was designed to take a top-down look at ocean island geology, starting with volcanic processes and ending in the mantle.
Don Swanson, in a keynote talk on hazards generated by ocean island volcanoes, emphasized that work in Hawaii has shown that such volcanoes do not just emit quiescent flows of basalt. If groundwater and magma interact at shallow levels, enormous explosions can occur, and the giant landslides that are promoted by rift zone processes, although infrequent, present extreme potential for destruction both directly and as a result of the giant tsunamis that they may cause. Several posters addressing case studies of hazards at various ocean island volcanoes underscored that because of increasing population pressure, more and more neighborhoods are being built in locations that are potentially in the paths of destructive lava flows. Scott Rowland reviewed the variations in morphology of ocean island volcanoes, noting the complexities introduced by the presence of vents, lava flows, pyroclastic flows, intrusive events, and the consequences of erosion. One of the most important distinctions in the morphology of different ocean island volcanoes is the presence (e.g., Hawaii) or lack (e.g., Galápagos) of rift zones. Two features that can be important in the development of rift zones are the presence of sediments (which may permit rift zone development) and whether new volcanoes grow on preexisting ones. Additional talks and posters characterized the submarine stages of volcanism and emphasized the importance of stresses that develop during ascent and intrusion of dikes and in magma reservoirs.
Howard Snell, director of conservation for the Charles Darwin Research Station, spoke on the geologic controls of biodiversity and the evolution of new species in the Galápagos. He highlighted controlling factors such as distances between emerged islands, sea-level changes, and ages of different islands, presenting quantitative links between these factors and speciation. One aspect that may be overlooked by geologists is the importance of very small islands, which may serve as ecological staging areas.
Lithospheric Controls on Ocean Island Magmatism
Doug Toomey discussed seismic evidence for magma chambers and partial melting, emphasizing the extraordinary results from the East Pacific Rise MELT experiment. Most important is that the partial melt zone is much wider and deeper than many models have suggested. Unfortunately, comparable surveys have not been accomplished yet in the ocean island environment. One of the principal problems is that in many places, notably Hawaii, the lithosphere is too thick to image the melt zone with existing technology. Clearly, such a survey in an archipelago like the Galápagos could revolutionize our understanding of melting and melt transport in the ocean island setting.
Participants discussed the origin of chemical diversity among ocean island volcanoes, specifically the importance of lithosphere-magma interaction. In contrast to traditional assumptions, recent evidence suggests that many ocean island magmas are measurably affected by lithospheric assimilation. This topic is of more than provincial petrologic importance because of the overwhelming role ocean island basalts have played in modern interpretations of mantle dynamics. It was suggested that although assimilation may be important in some settings, it is recognizable, and the isotopists may not be altogether misinterpreting mantle signatures. L. Kenny Rubin discussed the use of actinide-series disequilibria and noted that, in general, the lack of reliable eruption ages from this technique is linked to complexities in processes of melting and differentiation, the low abundances of the necessary elements in many minerals, and analytical difficulties. Several posters highlighted measurements of magmatic water, halogens, and oxygen isotopic ratios, exploring how magmas interact with hydrothermal systems and the extent to which they interact. Others dealt with the controls of regional tectonic stresses on the volcanic and petrologic development of ocean island volcanoes and the relationship between alkaline and tholeiitic magmas from several archipelagos. The scale of chemical diversity was a recurring theme, and the findings illustrate that compositional variation can occur on scales as small as melt inclusions and as large as volcanic edifices.
Ocean-island volcanoes have played the primary role in revealing large-scale dynamic features and long-term evolution of Earth's mantle. Two aspects of ocean islands provide essential evidence in this view: (1) the time-transgressive nature of hotspot tracks and their supposed fixity, and (2) the isotopic and trace element compositions of ocean-island basalts. Paul Wessel reviewed the "hotspotting" technique, whereby all volcanoes in a chain are backtracked to indicate the position of their source. He reached two important conclusions. First, many island chains either are not produced by fixed hotspots or, if they are, positions of the islands have been shifted by secondary processes. Second, there is evidence for a major reorganization of the Pacific and surrounding plates ca. 6 Ma. Alice Gripp presented the results of her latest update of a global inversion for absolute plate motion, which may cast some doubt on this reorganization event. One conclusion emerged from the discussions: Absolute motion of the plates, particularly for the past few million years, is not as well understood as many would hope.
Erik Hauri reviewed isotopic evidence for lateral compositional variability in the Hawaiian plume. Osmium isotopic evidence indicates that the variability is not related to lithospheric interactions. Despite geodynamic results indicating that entrainment of lower mantle is insignificant, isotopic evidence suggests the contrary. Deuterium measurements on melt inclusions presented at the conference may eventually constrain the time scales of mantle mixing and residence times. In an overview of helium isotopic data for many ocean islands, Dave Graham attributed helium isotopic variations to a spectrum of processes, the most significant being plume flux, velocity of the overlying plate, and proximity to a plate margin. Al Hofmann, who was one of the first to suggest that mantle plume compositions include a significant contribution from ancient subducted lithosphere, argued that the elevated strontium concentrations found in some ocean island basalts may be attributable to ancient gabbro in mantle plumes-trace element evidence from Hawaii indicates that gabbro-derived eclogites may constitute an important plume component.
Ocean Island Volcanic Processes
In a synopsis of recent research on a variety of ocean island provinces, Dave Clague presented an overview of the evolution of Hawaiian volcanism. MacDonald's pioneering model of the evolution of the world's largest volcanoes has been revised as a result of petrologic, volcanological, geochemical, and geophysical studies of the submarine stage at Loihi, as well as greater understanding of the causes of the evolution of Hawaiian volcanism. For example, evidence reviewed by several participants, including Clague, Hauri, and Rhodes, points to a concentrically zoned plume. Volcanoes appear to be dragged from edge to center to edge of the plume, thus producing notable variations in geochemical compositions. Significant questions remain; for example, the mechanism by which the plume becomes zoned is not fully understood, and there is controversy over the internal structure of the volcanoes, especially the nature of the submarine deposits. The upcoming Hawaiian Scientific Drilling Project, retrieving core from the deep interior of Mauna Kea, may enhance our understanding of petrologic and geochemical changes with time.
Dominique Weis's research on Kerguelen and its hotspot track and Kaj Hoernle's description of ocean island volcanoes of the eastern Atlantic showcased the extensive amount of data collected on ocean islands in the Indian and Atlantic oceans. Kerguelen is unique because it records 115 m.y. of hotspot volcanism. Kerguelen's lavas have remarkably consistent (and unique) radiogenic isotopic ratios, indicating that the plume source has been supplying chemically homogeneous magma continuously over this time span. Secular changes in magma production rates and some trace element ratios indicate decreasing extents of partial melting with time. Volcanoes of the Canary, Selvagen, and Madeira archipelagos are unusual because they lie on a very slow-moving plate and near a continental margin. Despite these complexities, the age and isotopic ratios of these volcanoes appear to be consistent with a mantle rich in ancient subducted sediments (the so-called HIMU component). In a demonstration formed by Anita Grunder, David Graham, and Wendy Bohrson, Hofmann demonstrated the latest high-technology method of mixing chemically distinct mantle "components."
The Galápagos is one of the best locations to study the interaction between mantle plumes and mid-ocean ridges; the Galápagos spreading center produces "excess" magma and has anomalous geochemical features, and the lavas have many characteristics in common with typical mid-ocean ridge basalt. Godfrey Fitton compared plume-ridge interaction at Iceland and the Galápagos via examination of excess niobium in the two provinces. He proposed that whereas the Galápagos exhibits a strong contribution from the mantle region that produces mid-ocean ridge basalt, the vast majority of Icelandic basalts show little evidence of MORB components. Derek Bostok emphasized the fluid nature of the niobium/imodium ratio. Dave Christie considered the ridge's perspective, discussing geochemical and tectonic features of the Galápagos spreading center and focusing on the influences on the ridge segments that propagate away from the hotspot. The data appear to support flow of plume material in both directions down the axis of the ridge. Moreover, the isotopic zoning observed in the Galápagos islands is partly preserved during this flow. Posters in this session were directed to detailed geophysical and geochemical studies of Iceland, the Reykjanes ridge, St. Paul-Amsterdam, and Easter Island. Dave Naar showed that the transforms in the Galápagos region are relatively young; thus, there is not likely to be a large difference in lithospheric thickness across the archipelago, as previously assumed. Instead, the pseudo-faults of propagating rifts may channel the flow of plume material to a ridge.
The Continental Connection
Mantle plumes are also thought to produce continental volcanoes, such as those of continental flood basalt provinces and the Yellowstone-Snake River Plain. Although the chemistry of magmas produced by continental hotspots is complicated by the thicker and compositionally variable lithosphere, the ease of access and continuous record provide some advantages.
Gene Humphreys presented results from a tomographic study of the eastern Snake River Plain that has provided a remarkable image of the crust and mantle in the region. The axis of the plain is underlain by seismically slow upper mantle, which grades laterally to fast mantle under the parabolic wake of the hotspot. This spatial pattern is attributable to the presence of a partial melt zone that extends to 150 km and is surrounded by a zone of melt-depleted upper mantle. These results seem to be inconsistent with a deeply-rooted plume that pancakes against the lithosphere, as had been described by Bercovici earlier. Instead, self-perpetuating convective overturn caused by melt buoyancy may drive Yellowstone magmatism. Mike McCurry reviewed the petrology and geochemistry of the Snake River Plain, and also concluded that simple mantle-plume models may be inappropriate. Posters explored the connection between continental flood basalts, oceanic plateaus, and ocean island volcanoes. Although the case for the island-plateau connection appears strong, the link between ocean islands and continental flood basalts is much less clear: Are the chemical and eruptive differences due to the thicker continental lithosphere, or are the associations tectonically unrelated?
Participants reached several conclusions at this conference.
1. Although there are broad similarities in some characteristics among ocean island chains, all ocean island chains are unique in some ways. This is likely the result of the complex interaction among factors such as age, thickness, composition of lithosphere, proximity to a mid-ocean ridge, rate of plate motion, magma chamber dynamics, characteristics of the associated plume (if present), and different magma sources. Broad-scale similarities of particular characteristics may lead to the development of a spectrum of models for hotspot evolution and volcanic processes.
2. The assumed plume origin of many "hotspots" has been questioned on the basis of observations such as lack of age-progressive volcanism and simultaneous eruption of volcanoes that are geographically widespread. This underscores the question of the "origin" of plumes. If "strong" plumes come from the core-mantle boundary, as has been suggested, do "weaker" plumes come from a shallower boundary layer? Are other hypotheses, such as lithospheric cracking, tenable as explanations for some ocean island magmatism? Resolution of these questions requires integration of geochemistry and geophysics. In particular, as technology improves, the highly successful MELT experiment and the tomographic images from Yellowstone may provide templates for understanding crustal and mantle structure beneath ocean islands.
3. Evidence for chemical diversity of ocean island magmas abounds, but its origin remains incompletely explained. It is crucial that we improve our understanding of the physical nature of the chemically distinct mantle "components." While there seems to be a consensus that a major component of ocean island magmas is recycled oceanic lithosphere, the origin of much of the isotopic and trace element diversity remains unresolved. In addition, there is no consensus regarding the contribution lithospheric mantle makes to geochemical signatures of some ocean island volcanoes.
4. While there is evidence that some hotspot chains are characterized by coherent spatial and temporal variations in chemistry, the origin of such patterns is controversial. The apparent success of the Hawaiian zoned-plume model, where entrainment is called upon to generate chemical variations within the plume itself, is tempered by an apparent lack of support in geodynamical models. Yet, helium isotope signatures suggest some contribution from a relatively undegassed region, probably located in the lower mantle. Resolution of this controversy will require interdisciplinary studies that address both geochemical and geophysical constraints on plume-mantle interaction.
5. It is likely that volcanic hazards associated with ocean islands have been underestimated. As demonstrated at this conference, it is of paramount importance to acknowledge the potential consequences of giant landslides and associated tsunamis as well as recognize the possibility of highly explosive eruptions on volcanoes previously assumed to be quiescent. Identification and mitigation of the hazards being generated by increasing population density are essential. Progress in this arena relies on detailed mapping and dating of ocean island volcanoes, greater communication between geoscientists and public officials, and increased effort on the part of geoscientists to educate the public about volcanoes.
We thank Galápagos Travel, whose staff coordinated flights and arranged for field trip boats; the Charles Darwin Research Station staff, especially Ximena Naranjo, Rob Bengsten-Smith, and Heidi Snell; and Jack Nelson, who arranged accommodations at Hotel Galápagos.
The Geological Society of America and a grant from the National Science Foundation assisted in funding the student and postdoc participants. The International Association of Volcanology and Chemistry of the Earth's Interior (IAVCEI) also provided travel funds.
PENROSE CONFERENCE PARTICIPANTS
| Jamie Allan
| Alice Gripp
| Loren Kroenke
| Ian Ridley
L. Kenny Rubin