On this page, you can find descriptions of my research, past and present, and, if you are a prospective grad student, you can get information on directions for possible thesis topics that fit in with my overall research goals.
Predicting Garnet-Melt Partitioning
The spreadsheet file for carrying out the predictive modeling of garnet-melt trace element partitioning published in Contributions to Mineralogy and Petrology in 2007 is available here.
Publications
For a full list of my publications, including links to PDF files of many of them, please visit my Institute of Meteoritics page.
Other Planetary Petrology Sites
Here follow links to some well-established groups involved in petrology and planetary science, some of which I have had an association with at some point; and to some useful sources of geological and petrological information.
Have a look at my former institution, Macquarie University's Department of Earth and Planetary Sciences (formerly the School of Earth Sciences) in Sydney, Australia.
Check out the Department of Geological Sciences at the University of Oregon's homepage; this is where I got my doctorate under the supervision of Dana Johnston.
Here's the page for the University of Bristol's Department of Geology, and for CETSEI, the Centre for the Experimental and Theoretical Study of the Earth's Interior.
The Lunar and Planetary Laboratory at the University of Arizona.
The School of GeoSciences department at the University of Edinburgh. Their claim as the "birthplace of geosciences" is no exaggeration.
The Australian National University's Research School of Earth Sciences.
Caltech's GPS Division Home Page.
Andrea Koziol has put together an excellent resource, Mineralogy and Petrology Research on the Web.
The MELTS Home Page. This is the link to Mark Ghiorso's site for his model of magmatic differentiation based on free-energy minimization.
Lamont-Doherty Earth Observatory Home Page
The Planetary Geology Group at Arizona State University.
The Woods Hole Oceanographic Institution WWW Server.
The Carnegie Institution is home to two groups important both for historical reasons and for the work they do currently: the Geophysical Laboratory, birthplace of modern petrology and at which no geophysics is studied, and the Department of Terrestrial Magnetism, at which nobody studies magnetism. Not that there's anything wrong with that. I'm in the Institute of Meteoritics, and I don't know jack about meteorites!
MTU Volcanoes Page. Lots of good volcano images.
The University of North Dakota's Volcanoes of the World page. A good place to take kids of all ages.
Research in Planetary Petrology
My main research areas are in Planetary Petrology. Planetary Petrology means, in my case, an emphasis on experimental investigations of processes of formation of terrestrial and extraterrestrial basaltic magmas and their relationships with their mantle source regions. Basalts are melts of the upper mantles of these bodies, and as such are probes of the interiors, bringing with them (typically subtle) clues about their source regions. This is hugely important, because we can't take field trips to the mantle and hence cannot sample those regions directly... Basalt is by far the most abundant rock type on Earth: the entire ocean floor, which covers three-quarters of Earth's surface, is made of basalt. Similarly, both Mars and the Moon are dominated by basaltic rocks. It therefore follows that understanding how a planetary interior works means learning how to decipher what basalts have to tell us.
Areas of emphasis also include study of terrestrial mantle xenoliths (rare pieces of Earth's upper mantle brought to the surface by some basaltic lava flows), the role of H2O and CO2-rich fluids, and mineral-melt trace element partitioning. Areas of major concentration in my earlier work included formation of subduction- and rift-related magmas, very low-degree silicate melts as mantle metasomatic agents, and the solution behavior of noble gases in magmatic systems. Summaries of these research avenues appear below.
Experimental Petrology
The main tools I use in my work are those of experimental petrology, particularly at elevated pressures using solid-media apparatus such as multi-anvil and piston-cylinder presses. These devices can mimic conditions that prevail in the mantles of Earth, Moon, and Mars where planetary basalts are generated. We can control conditions of temperature, pressure, volatile content, oxygen fugacity, and others in order to constrain how these materials form and evolve. My primary responsibility within the Institute of Meteoritics is in managing the High Pressure Experimental Laboratory, and you can find more details of the lab at that link.
In late 2000, I started the ExPet Yahoo email group for experimental petrologists, intended as a successor to the first incarnation of such a list organized by Henry Shaw back in the early 90s. Take this link to join the group if you're interested.
Martian basalts and the evolution of the martian mantle
There are only a couple dozen martian meteorites, yet as a community we have been able, from them, to unravel some profound insights into how that planet formed and evolved over time. Our primary key to this understanding is the study of basaltic meteorites from Mars, known originally as shergottites. These samples reveal that Mars was probably entirely molten very early in its history, and that relatively soon after the martian core formed, the remaining molten silicate portion of Mars crystallized to form distinct geochemical reservoirs that remained unmixed until just a few hundred million years ago. The martian basalts display arrays of geochemical features that can be understood as mixtures of two different reservoirs, and it is likely that these formed during the crystallization of the initial martian magma ocean. Our work on this topic has been highlighted in a couple of recent articles in the highly readable Planetary Science Research Discoveries series: The Multifarious Martian Mantle and A Primordial and Complicated Ocean of Magma on Mars. We have attempted to combine insights from high-pressure experimentation with numerical modeling to understand how these reservoirs formed and were involved in the generation of martian basalts. Experiments with this aim have included melting of possible martian mantle compositions over a range of temperatures and pressures in an attempt to constrain the source regions for martian basalts. Most recently, phase relations of the most primitive martian basalt recovered to date, the Yamato 980459 shergottite, have been determined under volatile-bearing conditions. These latter experiments show that the anhydrous phase relations are shifted to significantly lower temperature with 2 wt% water added, but that this water content is not sufficient to stabilize a hydrous phase near the liquidus of this composition.
In addition to this hydrous work, new experiments testing our numerical models of martian magma ocean crystallization will be undertaken. Our models have important predictions for the nature of the source regions for martian basalts, and these predictions require experimental testing in our continuing efforts to understand the martian interior. Papers arising from the work summarized here have been published during the early to mid 2000's in Geochimica et Cosmochimica Acta, Physics of the Earth and Planetary Interiors, and Earth and Planetary Science Letters.
Studies of lunar basalts
Despite the visits to the lunar surface during the Apollo program and the hundreds of kilograms of samples returned, there is still a great deal to be learned about the Moon. The influence of the Moon on our own Earth is very profound, from its role in maintaining the angular momentum of Earth (so that our axis of rotation doesn't roll over chaotically, preserving our climate) to its influence on oceanic tides, its role in timekeeping, and its appearance in human mythology. Over the centuries, a variety of ideas about its formation have come about, but the most recent and widely accepted is that a large, Mars-sized body collided with the proto-Earth very early in Earth's history, jettisoning copious material (both from the proto-Earth and the impactor itself) that eventually formed the Moon. This cataclysmic event probably resulted in total melting of both Earth and the Moon, and working out the consequences of these melting episodes is crucial in understanding their histories. From the point of view of the evolution of the Moon, important unanswered questions include the composition and internal structure of the Moon, particularly the lunar upper mantle; the extent of the lunar magma ocean; whether there may exist a deep, possibly garnet-bearing, remnant of undifferentiated lunar mantle and whether some lunar picritic glasses originated in such a horizon; and the origin of KREEP-rich, magnesian samples.
Work in our group on these questions has dealt with the determination of near-liquidus phase relations for both very high- and very low-titanium picritic glass bead compositions thought to be near-primary, and on determinations of trace element partitioning between these liquids and garnet at high pressure. See my Publications page for references on these studies. Hopefully proposed new work will be funded that will attempt to test experimentally some predictions of lunar magma ocean crystallization by mimicking that process in the laboratory. Despite the widespread acceptance of the Giant Impact / Lunar Magma Ocean hypothesis, it has never been directly tested experimentally, and doing so should help resolve some of the unanswered questions listed above. Work on this topic resulted in mid-2000s papers in Geochimica et Cosmochimica Acta and American Mineralogist.
Trace element partitioning
Knowing how elements partition between coexisting mineral and melt phases is critical in modeling a wide range of processes taking place within planetary interiors, from formation of metallic cores to crystallization of planet-scale magma oceans to the generation of magmatic liquids during partial melting and their subsequent differentiation prior to and during eventual eruption. Thus a great deal of effort has been expended by many researchers over the years to measure element partitioning and gain understanding of the factors governing the process. These factors can include mineral and melt composition, temperature, pressure, water content, oxygen fugacity, and more. Many elements are useful for elucidating interesting geochemical processes, such as siderophile, large-ion lithophile, high field-strength, and rare-earth elements.
A major advance in understanding mineral-melt paritioning in silicate systems was made in the early 1990s by Jon Blundy and Bernie Wood. They applied lattice-strain theory to extract meaningful physical information from variations in partitioning D values as a function of the ionic radius of the trace elements. These parameters include the maximum D value for an element of a given charge entering a particular crystallographic site; the bulk modulus of that site; and the ideal radius of a trace element partitioning into that site without straining the lattice. Jon Blundy was awarded the F. W. Clarke Medal of the Geochemical Society for his work on this topic. Lattice-strain models of element partitioning have led to the ability to predict partitioning for important elements in phases such as clinopyroxene, plagioclase, and garnet.
I have made heavy use of the Blundy & Wood methodology in studying garnet-melt trace element partitioning in lunar and martian bulk compositions. These systems are more Fe-rich than is typically the case for terrestrial magmas. With my colleague Wim van Westrenen, a Blundy Ph.D. who extended lattice-strain theory to garnet-melt partitioning, I have been extending that predictive ability beyond what Wim was able to obtain for terrestrial magmas into compositions appropriate for extraterrestrial magmagenesis. Our new predictive models work very well at virtually any relevant temperatures and pressures under anhydrous conditions. Additional work is now needed to extend this predictive ability to systems containing water, where our new model fails badly. Work on element partitioning has produced recent publications in Physics of the Earth and Planetary Interiors, Geochimica et Cosmochimica Acta, and Contributions to Mineralogy and Petrology.
Readers of our 2007 papers in Contributions to Mineralogy and Petrology presenting our recent updated predictive model for anhydrous garnet melt partitioning can obtain the spreadsheet for performing the calculations by following this link.
Generation of basalts in subduction zones
Subduction zones, where one tectonic plate dives beneath another in a convergent-margin setting, play a key role in the formation of continental crust on Earth. They are sites of very active and explosive volcanism, where some of human history's greatest natural disasters have taken place (the eruption of Vesuvius, which buried Pompeii and Herculaeneum in AD 44, is a well-known example). Volcanic chains formed from subduction include remote locales such as the South Sandwich Islands, in the far south Atlantic ocean, but more commonly make up heavily populated island groups such as the Caribbean Antilles, Japan, and the Phillippines. In these regions, one oceanic plate is subducting beneath another. Where an oceanic plate is moving beneath a continental one, volcanic mountain ranges such as the Andes, the Cascades, the central volcanoes in Italy, and the Aleutian Islands are the result. The Pacific Ring of Fire is so named because a large proportion of the Pacific Ocean is bounded by subduction-related chains of volcanoes including the Andes, Cascades, Aleutians, Kamchatka, Japan, the Phillippines, and Indonesia.
Because of their locations in favorable climates and the propensity of eruptive products from these volcanoes to generate rich soils for agriculture, large human populations are concentrated in many areas formed by subduction-zone volcanism. But many volcanic eruptions in these areas are highly explosive, caused largely by the much higher volatile contents and viscosities of these magmas compared to places like Hawaii, where eruptions are much less violent. Thus a great deal of work is done trying to understand the dynamics of these eruptions to mitigate the natural hazard that they pose.
My work on subduction-zone magmatism, which formed a big part of my doctoral work, centered on trying to elucidate the processes by which the dominant rock type in arc volcanic terranes, high-alumina basalt, is formed. Back then, a strong debate was underway about the relative importance of the subducting oceanic slab and the overlying mantle wedge as source regions for subduction-related basalts. Some argued that the presence of water-rich sediments as part of the package of subducted materials allowed the downgoing plate to melt extensively, while others argued that fluids liberated from these sediments served mostly to lower the melting point of the overlying mantle peridotite. This debate has mostly subsided, with the "slab melting" idea losing favor except for the formation of less-common, distinctive compositions such as boninites. The canonical view is now that volatile-rich material liberated from the subducting slab promotes flux-melting of the wedge, and that these magmas rise and pond near the base of the lithosphere where differentation to produce a range of common arc rock types takes place.
There are still some outstanding issues that have not been resolved, however. An important one is that primitive arc basalts do not appear to be derivable from presumably mantle-derived, parental magmas. These basalts plot squarely in the stability field of plagioclase, rather than olivine, at relevant pressures, and indeed do not show olivine saturation until well over 100 degrees beneath the liquidus. If these basalts are melts of the mantle, why do they not display olivine saturation? In some cases, plagioclase accumulation in a convecting, crystallizing magma can explain this discrepancy, but this process works for a limited subset of arc basalts. In my doctoral work, I sought to determine whether liquids representing these arc high-alumina basalts could be a product of peritectic reaction between olivine and precursor mafic parents. Although I did identify such a process, the liquids produced did not match the main target composition. In later (still unpublished) work, I sought to determine whether the presence of water in the system could suppress plagioclase crystallization in these melts, and thereby show conditions under which they could be consistent with separation from an olivine-bearing source. I found that such was not the case at 2 wt. % H2O, and even a few reconiassance experiments at 5 wt. % still failed to displace plagioclase from the liquidus. I really need to finish that project and get it published...! My earlier work was summarized in early-90s papers in Contributions to Mineralogy and Petrology and the Journal of Volcanology and Geothermal Research.
Mantle xenoliths and metasomatism
Metasomatism, in general, means simply for something to have undergone change of some kind. Mantle metasomatism is a process that has been often invoked to account for certain otherwise difficult-to-explain (usually trace-element) compositional features. The basic idea is that some kind of metasomatic agent -- a fluid, perhaps with a lot of stuff dissolved in it; or a carbonate melt; or a silicate melt; or some combination of these. Mantle metasomatism comes in two basic flavors. Patent metasomatism results in new mineral phases forming as a consequence of interaction with whatever agent was acting; this is clear to see by the presence, for example, of hydrous minerals, veins of quenched liquid or fluid phases, etc. Cryptic metasomatism, on the other hand, is manifest not by any obvious mineralogical change but only as compositional signatures for which "ordinary" petrological processes do not seem to account. Many studies of both mantle-derived basalts and of mantle xenoliths (pieces of mantle wallrock transported to the surface by volcanic eruptions) have cited one or both of these processes to explain such features.
It's been pet peeve of mine for a long time that falling back on cryptic metasomatism was an easy way out; one need simply invoke the involvement of some agent that is rich in this-and-that and poor in those-and-the-others -- whatever was needed to match the features one could not explain. In many such studies, there is no evidence for the existence of such a material except the inability to explain those elemental signatures. (I have no such issues with patent metasomatism, where the evidence is plain to see.) Near the end of my graduate work, I began studying a set of xenoliths from a new locality in southernmost Washington, the Simcoe volcano, which sits in the middle of a huge expanse of basalt-rich territory that totally lacks mantle xenoliths, except for these. In so doing I came across small veins, patches, and blobs of glass having a very strange composition: extremely rich in Si, Al, and the alkalies but poor in Ca, Mg, and Fe. In fact, they were very similar to alkali feldspar in composition despite being unquestionably quenched liquids.
To my surprise, it turned out that very many xenoliths contain bits of glass having similar compositions. Many of these could be explained as products of interaction of the host lava with the xenolith minerals, or of mineralogical breakdowns within the xenoliths occurring during ascent and emplacement. But once those were accounted for, a class of glasses remained, having the most extreme contents of Si, Al, and alkalies coupled with deficiencies in Mg, Fe, and Ca. It occurred to me that this material, being present in xenoliths from around the world and in every tectonic setting, may be a kind of crypic metasomatic agent. So I set out to use experiments to constrain the lithologies from which such material might be derived, fully expecting to find some strange assemblage of unusual minerals near the liquidus.
To my amazement, these liquids turned out to be in equilibrium with an upper-mantle mineral assemblage of olivine + orthopyroxene or olivine + orthopyroxene + clinopyroxene. Thus they could serve as metasomatic agents, because they would not be compelled to react with mantle minerals-- they are in equilibrium with them. Work by other experimentalists (dominantly the CalTech group) also suggested that very low-degree melts of the mantle could have some of these unusual compositional features and provided thermodynamic reasons for why this might be. However, others (mostly in Australia) objected to this interpretation. However these liquids ultimately form, though, it seems clear that they could be good candidates for a type of agent of cryptic metasomatism. Papers arising from this work were published in the mid-90s in the Journal of Geology, Journal of Petrology, and Earth and Planetary Science Letters.