GSA home

Log In | GSA Community | GSA Store | Join GSA | Donate | Contact Us

GSA home

| GSA Community | GSA Store | Donate | Contact Us

About GSA

Career Resources

Connected Community

Divisions &
Associated Societies

Education & Outreach

GSA Foundation

GSA International




Public Policy



Find Your Science at GSA

Field Forum

Assessing the State of our Knowledge of Continental Arc Volcanism: The Tatara–San Pedro Complex, 36°S, Andean Southern Volcanic Zone

4–12 February 2007 • Talca and Tatara–San Pedro, Chile

Claude Jaupart Institut de Physique du Globe de Paris, 4, Place Jussieu, 75252 Paris Cedex 05, France
Tom Sisson U.S. Geological Survey, 345 Middlefield Road, Menlo Park, California 94025-3561, USA
Jon Blundy Department of Earth Sciences, University of Bristol, Wills Memorial Building, Bristol BS8 1RK, UK
Richard Arculus Research School of Earth Sciences, Australian National University, Canberra ACT 0200, Australia


Estero Molino South
Estero Molino South

The Tatara–San Pedro Volcanic Complex, Chile, is a well-exposed continental arc volcanic center ranging from basalt to rhyolite in composition. Its exposure, compositional diversity, moderate accessibility, and setting within the geochemically and tectonically segmented Andean Arc have prompted numerous detailed studies of its geologic history and magma genesis. Consequently, it is among the best-studied continental arc volcanic centers in the world. Forty international participants from diverse scientific backgrounds gathered for this Field Forum to discuss the processes involved in the construction of such volcanoes and the origins of their magmas. The forum opened in the Talca municipal library with two days of presentations by invited speakers that provided a general background to the complex and fostered vigorous debate. Field work commenced with two days of helicopter-supported visits to all stratigraphic components of the complex, followed by two days on foot from a spectacularly beautiful lakeside camp on the complex’s west flank, where stratigraphic and petrologic relationships were examined and discussed in depth. A hike to the trailhead was followed by a concluding day in Talca during which key issues were discussed by separate working groups and then in plenary session.


The Tatara–San Pedro Field Forum identified four linked behavioral characteristics of long-lived (ca. 1 Ma) volcanic systems. First: The complex grew as a series of temporally distinct eruptive phases whose products overlap widely in location, composition, and appearance. Detailed mapping, intensive sampling, and abundant accurate and precise eruption-age measurements were essential for unraveling the volcanic history and specific petrogenetic events; broad-scale reconnaissance sampling of complex volcanoes almost certainly lead to highly misleading conclusions. Dozens of precise (better than ±10 ka) 40Ar/39Ar ages are necessary to determine the time-composition-volume relations of Pleistocene–Holocene volcanic centers like the Tatara–San Pedro Complex. Second is the scarcity of near-primary (i.e., high-Mg) magmas, despite an abundance of olivine-rich rocks, which are mostly the result of magma contamination. The third characteristic is the role of paleotopography and climate in controlling the eruptive distribution and subsequent preservation of the volcanic products. Most of the Tatara–San Pedro Complex formed during the Pleistocene, and it is unclear how much material has been removed by glacial ice, river incision, and/or sector collapse. Fourth: Local basement tectonics influenced the distribution of volcanism. Incision is sufficient around the complex to expose grabens, dike systems, and hydrothermally altered areas that mark the former locations of Pleistocene volcanoes removed by erosion.

Field Forum participants at Volcán San Pedro
Field Forum participants at Volcán San Pedro. Click on photo for larger image and participant names.

Participants discussed at length the specific and general implications of the large data set for major and trace elements, as well as mineral compositions and assemblages of Tatara–San Pedro Complex lavas. Compositions range from basalt to high-silica rhyolite, encompassing most of the range and types present in the Andean Southern Volcanic Zone, within which near-primary magmas are absent. Magmatic diversity at the complex shows that local investigations and interpretations of regional along-arc compositional trends are highly complementary; for example, isotopic, trace element, and major element indices are to varying degrees systematically variable along the Southern Volcanic Zone between 33°S and 41°S, and these variations are in part related to segmentation of the arc.

Many apparent phenocrysts in Tatara–San Pedro Complex lavas, particularly of olivine, were shown through detailed chemical studies to be out of equilibrium with their host magma composition and/or with one another. This was documented thoroughly in the forum guidebook by chemical zoning profiles and textures. Considerable discussion ensued as to the nature of the process(es) that introduced these grains. Olivine and clinopyroxene are abundant as “phenocrysts” in Tatara–San Pedro Complex basaltic andesites and basalts, but are rare as constituents of common rocks of the upper continental crust, including those exposed near the Tatara–San Pedro Complex. Olivine and clinopyroxene are, however, phases that would grow as near-liquidus minerals from Tatara–San Pedro Complex basaltic and basaltic andesitic magmas. Some forum participants took this as evidence that the non-equilibrium olivine and clinopyroxene grains were entrained from within the magmatic plumbing system and grew from predecessor magmas generally similar to those that delivered the crystals to the surface. Such grains can be termed “antecrysts,” distinguishing them from true phenocrysts or from wholly foreign xenocrysts. Some fraction of the suspended crystals that are not in equilibrium with host lavas carry subtle, easily overlooked evidence that they are xenocrysts derived from solid lithologies. Establishing the genetic relationships among whole-rock chemistry, crystal populations, and mineral compositions is difficult but essential.

There also was a lot of discussion about contamination mechanisms. Xenocrysts can be picked up during ascent from conduit walls, or from reservoir walls during storage. One question that came up repeatedly concerned the controls on the physical mechanisms of entrainment and dispersion of xenocrysts/antecrysts and associated melts throughout a large magmatic system, and the criteria for distinguishing xenocrysts from phenocrysts from antecrysts. It was suggested that in any long-lived volcanic center where magma pathways and conduits are repeatedly exploited and cannibalized by multiple generations of magmatic pulses, entrainment of antecrysts is likely to be the norm rather than the exception.

Some approaches were identified for clarifying the architecture of a magmatic system. Diffusion modeling of chemical zoning in xenocrysts/antecrysts can be used to estimate their residence times and hence the ascent or storage times for their host magmas. Volatile-saturation pressures of melt inclusions can give depths of phenocryst growth. U-Th and U-Pb ages of zircons from plutonic xenoliths present in some Tatara–San Pedro Complex eruptives could distinguish if these originate from intrusive parts of the active magmatic system or are older and unrelated. Forum participants also endorsed the general need for geophysical studies to determine the locations, sizes, and shapes of the intrusive portions of active arc magmatic systems, as well as the distributions of melt and crystals. The Field Forum participants stressed the community’s need for integrated geologic-geophysical-petrologic studies of dormant volcanic systems that can complement results obtained by studying active eruptions.

Observations at the Tatara–San Pedro Complex brought into focus the magnitude and efficiency of erosional processes, including sector collapse. Erosion is a mixed blessing, as it allows access to exposures deep within a volcano but can lead to total loss of some eruptive units. Lots of debate ensued on the magnitude and timing of events and processes. Results from Tatara–San Pedro Complex and other well-studied arc volcanoes show that thorough three-dimensional geologic mapping, supported by radiometric dating and geochemistry, are required to determine the full history of a volcanic edifice. Emergent digital mapping techniques, coupled with remote sensing, are well-placed to tackle the difficulties of mapping in remote, rugged, and arid terranes, such as the Andes.


A Complete Data Set

Many important issues are difficult to address without comprehensively documenting the volcano in question, and the basic requirement for this task is information on the compositions, ages, and distributions of preserved eruptive products. Such results reveal the strongly episodic eruptive behavior of the Tatara–San Pedro Complex, also seen at some other arc volcanoes, and allow estimates of the volumes and masses of erupted materials, although the accuracy of these estimates diminishes with age and degree of erosion. The extent and ubiquity of open-system processes at the Tatara–San Pedro Complex also leads to questions about the underlying crust. The immediate vicinity of the Tatara–San Pedro Complex exposes folded Tertiary arc volcanic and volcaniclastic rocks intruded by granodiorite to leucogranite plutons. Networks of Pleistocene and Holocene dikes and larger intrusions of ultramafic, mafic, and intermediate compositions must underlie and feed the Tatara–San Pedro Complex itself, but these are poorly understood. Do Tatara–San Pedro Complex magmas interact widely with older, unrelated crust at shallow levels, or do they mainly entrain and assimilate various combinations of antecrysts and evolved interstitial liquids left by earlier magmas? The relative youth of nearby arc-generated crustal rocks precludes strong contrasts in Nd, Sr, and Pb isotopes that could otherwise distinguish the sources of contamination, although oxygen and osmium isotope measurements might be informative. Volatile contents of melt via glass inclusions and phase equilibria studies should be made systematically at volcanoes that are the objects of long-term exhaustive investigations in order to constrain magma storage depths and their evolution with time.


Eruptions are by nature transient events, and volcanic activity exhibits different types of episodic behavior. Some results at the Tatara–San Pedro Volcanic Complex are consistent with a link between eruptive periods and crustal unloading during deglaciation, but the strength of this inference was debated vigorously. A few forum participants questioned whether the number and precision of radiometric age measurements and the extent of outcrop preservation are adequate to draw this conclusion. There was much additional discussion of spatial episodicity, such that magmas are channeled into discrete eruptive centers. Estimates of a volcano’s chemical budget would be improved if the dimensions of the system’s deep “capture zone” could be determined, for example, by geophysical techniques.

Liquid Lines of Descent

The Tatara–San Pedro Complex is characterized by a significant diversity of parental magma compositions. Associated evolved magmas that erupted close in time to these parental compositions define generally coherent trends, but these trends differ from one eruptive episode to another. For example, some eruptive groups classify as arc tholeiitic series and others as calc-alkaline based on FeO*/MgO versus SiO2 systematics. The compositional trends result from various combinations of fractional crystallization, magma mixing, and assimilation, wherein both earlier Tatara–San Pedro Complex intrusions and older crust may be implicated. Despite the wealth of geochemical information for the Tatara–San Pedro Complex, experiments have not been undertaken to determine liquid lines of descent for potential parental magma compositions. Participants expressed concerns about how pristine parental compositions, unperturbed by open-system inputs, can be identified and how the characterization of extraneous crystals (xenocrysts and antecrysts) can contribute to an understanding of such processes. Conversely, determining liquidus phase relations on rocks whose crystals are of uncertain parentage is an exercise of dubious value. The recognition that many volcanic rocks are mixtures of melts and entrained crystals has considerable implications for the ways in which we use experiments to constrain phase relations and liquid lines of descent.


Forum participants identified the following areas of future study: (1) architecture of a volcanic system, (2) the mantle signal, (3) episodic behavior, (4) phase equilibria and volatiles, and (5) natural hazards.

Architecture of a Volcanic System

The geochemistry of Tatara–San Pedro Complex lavas suggests that the volcano lies above mafic-ultramafic bodies with complex internal structures, and this inference is supported by an abundance of mafic to ultramafic xenoliths in Tatara–San Pedro Complex lavas. Some Tatara–San Pedro Complex magmas carry evidence for interaction with such bodies in the form of antecrysts or xenocrysts. Lava compositions reflect interaction and assimilation of a variety of rock types, implying that the architecture of a volcanic system must be reconstructed over a large depth range.

Depth of storage is a key factor for volatile evolution and mineral precipitation. The small density differences between magma and many crustal rocks imply that there may be multiple storage zones, which may be associated with neutral buoyancy horizons. The lack of volatile analyses of melt inclusions from Tatara–San Pedro Complex lavas and pyroclasts represents an important limitation in this regard. The relationship between plutonic and volcanic rocks would be best studied through exceedingly rare exposures that permit documentation of the connection between plutons, conduits, and the overlying volcanic system. The well-exposed Tertiary intrusives in the vicinity of Tatara–San Pedro Complex may provide useful clues to the configuration of the intrusive counterparts of the Pleistocene volcanic edifice.

The Mantle Signal

Using erupted lavas to determine the composition of primary melts generated in the subduction zone mantle wedge is extraordinarily difficult in a long-lived continental arc and a mature volcanic system. Eruptions of near-primary, unequivocally mantle-derived magmas are relatively scarce even in island arcs, and they are extremely rare in continental arcs. This is well established by ~1000 analyses from the Tatara–San Pedro Complex, where assimilation of various materials, some derived from the magmatic system and some not, is common. Thus, instead of back-calculating a primary liquid from the erupted lavas, it is probably better to study small mafic cones that are likely to erupt mafic end-members. Such cones flank many long-lived arc volcanoes, but they are remarkably rare in the vicinity of the Tatara–San Pedro Complex.

Questions that are worthy of investigation are many, including: Does the mantle signature change with time due to mantle heterogeneity and/or successive episodes of melt production and percolation? What are the fluid contributions from the subducted oceanic crust, subducted sediments, and the hydrated mantle wedge and how do these vary temporally and spatially?

Episodic Behavior

The apparent episodicity of eruptions and eruptive cycles at the Tatara–San Pedro Complex might be a sampling artifact related to extensive glacial erosion and insufficient age measurements, but episodic behavior has been documented at a number of other well-dated arc volcanoes, so the causes probably lie elsewhere. Many discussions dealt with a wider perspective on eruptive behavior. It was felt that determining time variations of eruption rate was the best method to study episodicity, as opposed to just tallying the number and ages of eruptions. To accomplish this requires geologic mapping to provide 3-D reconstructions of erupted units and extensive supporting chemical and age information. Episodicity can be attributed to regional causes, such as tectonic forcing of magma production rate or glacial maxima and minima, and to local ones, such as edifice growth and sector collapse affecting the eruption rate. One key question is whether the volcano remains perched on an instability threshold, such that small perturbations to the underlying magma system may trigger an eruption.

Phase Equilibria and Volatiles

The pre-eruptive volatile contents of Tatara–San Pedro Complex magmas are not well known and deserve specific study. Tephras provide the optimal materials for melt inclusion volatile studies, but tephra deposits are rare at the Tatara–San Pedro Complex mainly due to its effusive eruptive style and to the lack of tephra preservation during Pleistocene glacial conditions. Some discussions addressed sulfur- and ore-forming potential with the observation that the olivine and clinopyroxene antecryst suite bore some similarity to Alaskan-Urals type zoned mafic-ultramafic complexes that host platinum group element deposits. It was felt that one should try to determine a sulfur budget for the volcano and also that accessory phases, including magmatic sulfides, deserve studies of their own, because they provide information on trace element contents.

Natural Hazards

Any study of a volcano contributes to a better understanding of volcanic hazards. The long history at the Tatara–San Pedro Complex could be used to test petrologic monitoring methods that are currently developed elsewhere for more restricted eruption sequences. Key factors in hazard assessments are estimating the style and total volume of an eruption and how these vary from one eruption to the next. Some avenues for investigation are to use chemical and petrological data to infer magmatic recharge and discharge rates and to link these to episodicity and eruptive volumes. Much of the Pleistocene activity of the Tatara–San Pedro Complex consisted of effusive eruptions of mafic to intermediate lavas that extended to no more than about 15 km from the present summit location. The population is low in the immediate vicinity, and similar eruptions in the future would not cause major losses of life or property. The Tatara–San Pedro Complex produced two voluminous debris avalanches in the Holocene that were shed from its steep south face and that contain appreciable hydrothermally altered materials. A dominantly andesitic cone, Volcán San Pedro, grew atop the collapse scar and constructed a new summit to the edifice. Silicic eruptions tend to be more explosive than mafic ones, so it would be helpful to gain a better understanding of why and how the Tatara–San Pedro Complex switched several times from mafic to more evolved eruptive products. The copious geochemical data on the Pleistocene edifice could make such investigations more successful than elsewhere. Collapse-prone hydrothermally altered rocks are exposed to the east and northeast of the new summit and would be amenable to study due to their only moderately difficult access and the well-known volcanic history of the area.


Physics of Magmatic Processes

Discussions on almost every topic during the Field Forum focused at one stage or another on an examination of the physical processes responsible for what is observed. Many participants felt that inadequate progress has been achieved in the quantification of the physics of volcanological and petrological processes and that more effort is required. Over 25 years have passed since the publication of the pioneering textbook The Physics of Magmatic Processes (Hargrave, 1980, Princeton University Press); a fresh examination of the problems addressed in that book is overdue.

Several examples arose during the Field Forum. Assimilation and mixing are ubiquitous subvolcanic processes that may occur at various locations in the magmatic plumbing system and over a range of time scales. One should be able to assess the controls on the processes involved and their time scales (i.e., how long does it take to assimilate a specific amount of material). Configurations of volcanic plumbing systems remain elusive. Models that are currently entertained belong to three categories: direct magma ascent without storage, a large reservoir connected to the surface by a simple conduit system, and complex time-dependent plumbing systems with transient storage zones. Each of these scenarios is geologically possible, but greater efforts should be made to develop quantitative models that constrain the conditions under which such diverse plumbing systems develop. Chemical and petrological studies of volcanic products eventually lead to mass balance estimates for the magmatic/volcanic system, although this is not usually attempted on a transcrustal scale. An obvious question concerns the lateral extent of the system at depth. An eruptive center captures magmas coming from a wider area, which affects a comparison between the results of the petrological/chemical mass balance and the local crustal structure. However, very little is known about the physical extraction of melts from the sorts of mushy source regions and deep crustal reservoirs that are likely to lie at depth beneath many arc volcanoes. The presence of volatiles and silicic melts makes the problem radically different from its better-studied mid-ocean ridge counterpart.

The Ideal Volcano

Forum participants discussed the attributes of an ideal volcanic system that might serve as a case study for the comprehensive understanding of volcano behavior and hazards for human populations and cities. It was agreed that an ideal volcano should be active so that eruptive regimes are known or reconstructable from recent and well-preserved deposits, and because active centers produce a wealth of seismic, geothermal, volcanic gas, geodetic, and other monitoring signals that help to characterize the system. An obvious goal is to obtain a precise and complete history of eruptive and chemical evolution. Tephra layers are a high-fidelity record of eruptive activity, but they are fragile and erode rapidly. A long-lived lake, bay, or submarine area in the immediate vicinity would offer the opportunity to sample complete stratigraphic sequences at high temporal resolution. Exposures of deep sections, from sector collapse events for example, would allow access to parts of the volcano’s roots. Other desirable features: peripheral cinder cones, xenolith-bearing lavas, high isotopic contrast with the crust in order to facilitate geochemical interpretations, and frequent eruptions to allow for U-Th disequilibrium studies. In addition, the volcanic system must be large, the surrounding topography subdued to facilitate geophysical surveys, and the tectonic setting well-known. Another class of attributes deals with practical issues, such as logistical and political accessibility.

No volcano can boast every desirable attribute, but Forum participants identified some volcanoes that meet most requirements: Taupo, New Zealand; Sakurajima in Kagoshima Bay; Avachinsky in Kamchatka; Colima in Mexico; and Santorini in the Aegean Sea, among others. It was proposed that a workshop be convened that would include members of the geophysical community to explore possible joint geologic-petrologic-geophysical studies of an active arc volcanic center with the aim of developing better quantitative images and models of the integrated magmatic system. Such study would ultimately enhance the ability of geoscientists to correctly interpret monitored signals and so better anticipate and mitigate volcanic hazards.


We thank GSA for sponsoring the Field Forum. Cosponsoring institutions: the U.S. National Science Foundation, Immanuel Friedländer Foundation (ETH, Switzerland), SERNAGEOMIN (Chile), and the Swiss National Science Foundation. Mike Dungan instigated this forum, but he could not have managed without essential contributions from co-organizers Daniel Sellés, Carolina Rodríguez, Rebecca Lange, Ren Thompson, Fidel Costa, José Antonio Naranjo, and John Pallister. The execution of this logistically complex meeting was impeccable, the weather superb, the ésprit congenial, and the scientific discussions outstanding. A measure of the effort involved is a lavishly and comprehensively illustrated 112-page Tatara–San Pedro Complex Field Forum Guidebook, which has established a daunting benchmark for future meetings of this type. Mike Dungan was also prominent in the organization of the fourth International Association of Volcanology and Chemistry of the Earth’s Interior State of the Arc (IAVCEI SOTA) meeting held in southern Chile just prior to the Field Forum; all participants applaud Mike’s dedication to our scientific endeavors through his efforts to bring the IAVCEI SOTA meeting and GSA Field Forum to successful conclusions.