Penrose Conference Report some role. Penetrative deformational fab­ rics are present in most exhumed moun­ tain belts and provide clear evidence that Exhumation Processes: Normal Faulting, ductile flow is an important process that can either increase or decrease the rate of Ductile Flow, and Erosion exhumation, depending upon whether Conveners: the flow causes thinning or thickening in Mark T. Brandon, Department of Geology and Geophysics, Yale University, P.O. Box the vertical direction. Erosion is also an 208109, NewHaven, CT 06520-8109; [email protected] important exhumational process, as indi­ Uwe Ring, Institut fiir Geowissenschaften, Johannes Gutenberg-Universitiit, Postfach 3980, cated by the large volumes of sediment 55099 Mainz, Germany; [email protected] found adjacent to most contractional orogens. Within this context, we organized a Penrose Conference to examine all pro­ Over the past 25 years, there has One of the most difficult questions to cesses that contribute to exhumation of been a growing appreciation of the role answer in most orogenic belts, especially deep-seated rocks in ancient and modern that tectonic processes play in exhuming the older ones, concerns the relative con­ orogens. We started with three broad metamorphic rocks. This trend began with tributions of these different exhumation objectives: (1) to review and synthesize the discovery of highly attenuated crustal processes. Even so, the tectonics commu­ our knowledge about normal faulting, sections in the Basin and Range province nity has moved toward a casual consensus ductile flow, and erosion as exhumation and the recognition that attenuation was that normal faulting is the primary mech­ processes; (2) to examine the geologic caused by regional-scale horizontal exten­ anism for exhumation of deeply seated evidence needed to quantify the relative sion, as manifested by normal faulting. rocks. The most commonly cited evidence contributions of these different processes, This discovery caused many geoscientists is the presence of "younger-over-older" using information from metamorphic to rethink the role that horizontal exten­ relationships, where large faults, with low petrology, isotopic thermochronology, sion or, as it is more commonly called, to moderate dips, have placed younger structural and kinematic analysis, synoro­ tectonic extension, might play in other rocks on older rocks or lower grade rocks genic stratigraphy, geomorphology, and orogenic settings. Of particular interest is on higher grade rocks and, in the process, paleoelevation analysis; and (3) to exam­ emerging evidence that tectonic extension have cut out significant thicknesses of ine relevant geodynamic models and their might be responsible for exhuming meta­ stratigraphic or metamorphic section. But predictions for conditions that might trig­ morphic rocks within convergent orogens, recent papers have shown that this type of ger gravitational collapse. such as onland thrust belts (e.g., Hima­ evidence, by itself, is not diagnostic. Con­ layas, European Alps, Betic Cordillera of tractional faults can also cut out section if THE CONFERENCE southern Spain, Brooks Range of Alaska) the section was tilted back toward the hin­ Theconference was held October 9-13, and subduction-related convergent mar­ terland prior to faulting. A typical contrac­ 1996, at the Orthodox Academy of Greece, gins (e.g., Franciscan of California, Sanba­ tional fault would climb upward toward located near the town of Chania on the gawa of Japan, Hellenic-Aegean conver­ the foreland but would appear to cut · island of Crete in southern Greece. It gent margin of Greece, Hikurangi downward through the tilted section. included three days of presentations accretionary wedge of northeastern Top-to-the-foreland motion would pro­ and two days of field trips. There were 94 New Zealand). duce the appearance of normal offset participants, 42 from the United States, We use the term "exhumation" to where the fault cut through the tilted 15 from England, 13 from Germany, five refer, in a generic way, to all processes section. An additional problem is that from Australia, four each from Greece and that contribute to the unroofing of deeply the total offset and original dip of most Canada, three from Switzerland, and one seated rocks and their rise to the Earth's crustal-scale normal faults, especially those each from France, Israel, Japan, The surface. There are several exhumation pro­ involving deep crustal rocks, are typically Netherlands, New Zealand, Norway, cesses. Normal faulting and extensional poorly resolved. All of these factors make Poland, South Africa, and Spain. The par­ ductile flow operate within the Earth it difficult to quantify the relative contri­ ticipants included 16 Ph.D. students. The and are a direct manifestation of tectonic bution of normal faulting in exhuming conference had a relatively large number deformation, whereas erosion operates at deeply seated rocks. of women participants, 13 professionals the Earth's surface but can be indirectly Despite the current preference for and 9 students, marking a positive trend influenced by tectonically driven changes exhumation by normal faulting, it is clear in topography. that ductile flow and erosion must play Penrose Conference continued on p. 18

PARTICIPANTS Torgeir B. Andersen Jean Crespi Herman Griitter Karen L. Kleinspehn Alan Roberts Stuart Thomson Vaios Avdis George H. Davis Larry Guenther Maarten Krabbendam David Rowley Robert J. Twiss Hans G. John F. Dewey Bradley Hacker Anthi Liati Leigh Royden 0. Vanderhaeghe Ave Lallemant Bill Dinklage Bernard Hallet Gordon S. Lister Elizabeth Schermer Jaume Verges Dov Avigad David Dinter Andreas Henk Lidia Lonergan Fried M. Schwerdtner Lisa A. Verts Gary Axen Theodor Doutsos Ralf Hetzel Neil Mancktelow Jane Selverstone Simon Wallis Suzanne Baldwin W. Gary Ernst Paul Hoffman Stanislaw Mazur Dianne Seward Nicole Wawrzenitz Samantha Barr Hilmar von Eynatten Daniel K. Holm Peter Molnar Eberhard Seidel Laura E. Webb Manfred Brix James Faulds Niels Hovius Patrick O'Brien Sarah Sherlock John Weber Fraukje Brouwer Paul Fitzgerald Simon Inger John S. Oldow Virginia B. Sisson John Wheeler Doug Burbank Marnie Forster Carl Jacobson John P. Platt Colin Stark Donna L. Whitney B. Clark Burchfiel David Foster Rebecca A. Jamieson Raymond A. Price Bernhard Stockhert Sean D. Willett Erin Campbell Wolfgang Franke Barbara E. John Meinert K. Rahn John Tarney Paul F. Williams Christian Chopin Wolfgang Frisch Christopher Johnson Tim Rawling Micheal P. Terry Brian Windley Peter Copeland Dieter Gebauer Sue Keay Steve Reddy Christian Teyssier Meg Coleman Allen F. Glazner Adiamantis Kilias Stephen J. Reynolds

GSA TODAY, May 1997 17 Penrose Conference continuedfrom p. 17 exhumation probably operates at very found in the Alps and the Norwegian slow rates. Other evidence is needed to Caledonides, and now known from in the reduction of the gender gap, at least distinguish between tectonic and ero­ numerous places around the world. UHP in the area of exhumation research. sional exhumation. metamorphic rocks are continental or Presentations were divided into six John Platt followed with a keynote oceanic crustal rocks that were metamor­ half-day sessions. Each session had about presentation on synexhumation defor­ phosed within the coesite-eclogite or dia­ two hours of oral presentations, about mation in mountain belts. He reviewed mond-eclogite fades. P-T-t data for these one hour of poster presentations, and geologic evidence for detecting tectonic rocks demonstrate that both oceanic and about one hour for a panel discussion. exhumation, the main clues being the continental crust can be subducted to This report highlights the oral presenta­ excision of metamorphic gradients, the depths >100 km and then returned to the tions and panel discussions. Those who shape of pressure-temperature-time (P-T-t) surface. are interested in further details will find paths, and the kinematics of ductile shear The keynote presentation by Chris­ a complete listing of the conference pro­ zones and brittle faults. tian Chopin provided a general review of gram, including the titles of all presenta­ Jane Selverstone challenged the use the UHP problem. Simon Wallis described tions, at the following Web site: http:// of P-T-t paths to distinguish exhumational some UHP rocks from China and then love.geology.yale.edu/-brandon/exhume. processes. The shape of a P-T-t path is typi­ explored the role that buoyancy might html. cally difficult to resolve, and the youngest play in returning these rocks to the sur­ part of the path, which is most diagnostic face. Paddy O'Brien discussed attempts FIELD TRIPS of process, is the most difficult to con­ to determine the P-T-t path for HP meta­ The field trips were led by Bernard strain. She emphasized her point by show­ morphic rocks in the Bohemian massif Stockhert (Ruhr-Universitat Bochum, ing two sets of published P-T-t paths, one of Germany. Germany) and Eberhard Seidel (Univer­ set attributed to erosion and the other to Dieter Gebauer reviewed his applica­ sitat zu Koln, Germany), and focused on tectonic exhumation. There was no obvi­ tion of SHRIMP U-Pb dating to determine geologic evidence for Miocene exhuma­ ous difference between the sets. The infer­ ages of igneous and metamorphic growth tion in a convergent-margin setting. ence is that steep decompression curves of zircons from HP and UHP metamorphic The island of Crete is.a forearc high that should not be considered diagnostic of rocks of the Alps. His work indicates an formed above the south-facing Hellenic extensional exhumation. Eocene age for metamorphism which subduction zone. The island has spectacu­ Clark Burchfiel reviewed the develop­ would require exhumation rates of -10 to lar exposures of high-pressure-low-tem­ ment of the South Tibetan detachment 30 mm/yr for some of the Pennine nappe. perature (HP) metamorphic rocks that system, a north-dipping system of faults A major problem, however, is that, as of formed at depths of -35 km and were and ductile shear zones with top-north yet, there is no direct tie between exhumed rapidly, starting in the Miocene. offsets that crops out for more than 700 km Gebauer's zircon ages and the formational Geodetic data and the presence of young along the southern flank of the Himalayan ages of the diagnostic HP and UHP meta­ normal faults indicate that the surface of crest. The detachment system has proba­ morphic assemblages. the island is being extended in a strike­ bly been active since 22 Ma and appears The panel, chaired by Gary Ernst, normal direction, presumably due to to have evolved diachronously along its emphasized that UHP rocks are typically southward rollback of the subducting slab. length. There is local evidence in the found as large, internally coherent slices Extensional deformation on Crete is com­ Mount Everest area that the detachment at least several kilometers wide. The pres­ monly viewed as a southern manifestation system caused rapid exhumation, but at ervation of UHP assemblages suggests dry of the active extensional province that present it is not known if this is a general metamorphic conditions during exhuma­ underlies the Aegean Sea to the north. feature of the entire system. tion. Tectonic extrusion was reviewed as But an important distinction is that Crete Steve Reddy reviewed the pitfalls in one way to account for the exhumation lies in the forearc above the Hellenic slab, recognizing extensional faults. He showed of these unusual rocks. whereas the Aegean metamorphic core an example from the Pennine zone of the General comments from the floor complexes have formed in the arc and Swiss Alps where metamorphic section highlighted several additional issues. Are back-arc regions. was thinned by a doubly vergent system UHP rocks metamorphosed above the At present, Crete has very rugged of extensional faults that moved top­ Moho within a very thick crustal root or topography, with local relief >2 km northwest and top-southeast. below the Moho as slices subducted into (highest point is Psiloritis at 2456 m). The panel, chaired by Ray Price, the mantle? The paucity of mantle rocks The rugged landscape suggests rapid ero­ focused on the correct use of kinematic in association with UHP rocks favors the sion rates, but the relative contribution indicators and on the distinctions between crustal-root option. But the mantle option of erosion to total exhumation remains strain and displacement and between crys­ is favored by the observation that crustal poorly resolved. This problem was a com­ tallization ages and cooling ages. There rocks are not strong enough to support the mon theme throughout the conference. was also some discussion about the diffi­ exceptional high topography that should culties in using time-temperature data to form above a region underlain by very SESSION 1: estimate a time-depth path. One panelist thick (>100 km) continental crust. A thick LOCAL EXPRESSION OF TECTONIC noted that the present northward dip crustal root would be further inhibited by EXHUMATION: STRUCTURE, of the South Tibetan detachment system the higher radiogenic heat production METAMORPHISM, AND could be caused by isostatic uplift associ­ typical of continental rocks, which should THERMOCHRONOLOGY ated with deep erosion of the high Him­ lead to thermally induced softening Mark Brandon opened the conference alayas. This possibility highlighted the and/or melting. For the modem Earth, with a short introduction. He made the general difficulties in resolving the origi­ continental crust is thought to be no point that exhumation rates alone are not nal dip of exhumational structures. thicker than about 60 to 70 km. But what diagnostic of process. A common assump­ defines the Moho in areas where the crust tion is that tectonic exhumation is fast SESSION 2: has been metamorphosed to UHP assem­ and erosion is slow. But erosion has been FORMATION AND EXHUMATION blages? A study of the seismic-velocity clocked at rates up to -15 mm/yr1 (e.g., OF UDP METAMORPHIC ROCKS properties of UHP rocks is needed to Southern Alps of New Zealand). Con­ The most intriguing examples of understand the relationship of the seismi­ versely, there are settings , such as long­ deeply exhumed rocks are the ultra-high­ cally determined Moho to the petrologi­ lived continental rifts, where tectonic pressure (UHP) metamorphic rocks, first cally defined base of the crust.

18 GSA TODAY, May 1997 What is the role of buoyancy in This type of model emphasizes the fact the primary factor in triggering the land­ returning metamorphosed crustal rocks that exhumation is not necessarily slides; climate has played a subsidiary role. to the surface? Chopin indicated that the restricted to vertical motions and may The panel, chaired by Paul Hoffman, UHP granitoid rocks found in the Pennine also involve the interaction between considered a range of topics. What factors nappes of the Alps have densities of 2840 erosional and tectonic processes. are most important for determining when to 3080 kg/m-3. These rocks would be The panel, chaired by Karen Klein­ a tectonically active landscape reaches buoyant with respect to the mantle but spehn, highlighted several additional steady state? In this context, steady state not with respect to average continental issues. To what degree are erosion rates means that the rates of rock uplift are bal­ crust. The buoyancy argument is further controlled by climate? When using ther­ anced by equal rates of erosion. Is the time challenged by the discovery of ultramafic mochronometry, how can one account for to steady state typically short (<1-2 m.y.)? rocks with UHP assemblages (e.g., meta­ the effects of topography on isotherms In that case, the geomorphic system morphic diamonds in the Beni Bousera and the rotation of isochronal surfaces would closely track tectonic uplift. Or is and Ronda peridotites of Morocco and after isotopic closure? How much of the the time to steady state typically long (>10 southern Spain; >120 kmassemblages exhumation history of a UHP terrane is to 20 m.y.)? If so, the geomorphic system from the Alpe Arami garnet peridotites preserved in its present structural and would be constantly out of phase with tec­ in the Alps). stratigraphic setting? Do shallow-level tonic uplift. This question is critical for structures bear any relations to the pro­ resolving the cause of the increased flux of SESSION 3: cesses associated with the early phases of terrigenous sediment to the world's oceans RATES AND PATTERNS OF LONG­ exhumation when the rocks were still at during the Quaternary, whether due to cli­ TERM EROSIONAL EXHUMATION great depths? mate change or to increased rates of tec­ Most currently cited erosion rates tonic convergence. are estimated from short-term (10-50 yr) SESSION 4: One panelist presented a comparison studies of sediment yield from modem PROCESSES AND CONTROLS OF of the exhumational history of the Tauem river drainages. A nagging question is LONG-TERMEROSIONAL window, which is the most deeply ex­ whether or not these rates are represen­ EXHUMATION humed part of the eastern Alps, with the tative of long-term erosion rates. This Geomorphology continues to provide record of erosional exhumation preserved problem is particularly important for the fundamental information about erosional in the Alpine foreland and offshore basins Quaternary because global climate has processes, but much of this research beneath the Mediterranean and the North been fluctuating on a period of -100,000 remains at an early stage of development. Sea. An integrated sediment budget yr due to the glacial cycle. In this context, In particular, there are many unanswered showed that orogenic sedimentation rates long-term erosion rates can be usefully questions about how short-term local­ were low when the Tauem window was defined as the average rate of erosion for scale erosional processes relate to the long­ being exhumed. The inference was that a period of time >-0.1 m.y., to ensure that term erosional behavior of a whole moun­ tectonic processes probably dominated the effects of the glacial cycle are averaged tain range. Nonetheless, this research is during deep exhumation of this part of out. essential for resolving the interplay be­ the Alps. In his keynote presentation, Doug tween tectonics, topography, global cli­ Another panelist showed that fast Burbank reviewed various methods for mate change, and regional-scale erosion. erosion can have a profound influence on estimating long-term erosion rates, such Bernard Hallet gave a keynote the shape of metamorphic isograds and, in as the reconstruction of eroded geologic presentation in which he compared some cases, can result in inverted isograds. features, isopach measurements of synoro­ modem rates of erosion in glaciated and genic sediments, thermochronologic dat­ nonglaciated active mountain belts. He SESSION 5: ing of exhumed bedrock or detrital grains noted that nonglaciated mountainous THE ULTIMATE CAUSE FOR in synorogenic sediments, or an inventory drainages in Asia and New Zealand have TECTONIC EXHUMATION: of erosional processes operating within a modem erosion rates of about 1.5 to ACTIVE RIFTING VS. PASSIVE modem drainage. Burbank cited geologic 8 mm/yr, which overlaps with exhuma­ GRAVITATIONAL COLLAPSE evidence indicating that the Salt Range of tion rates estimated for UHP rocks. He Tectonic exhumation has been recog­ Pakistan was eroded at long-term rates of then went on to show that warm-based nized as a common factor in continental 5 to 10 mm/yr. alpine glaciers in southern Alaska had ero­ orogenesis, but the cause remains poorly Peter Copeland discussed the inter­ sion rates of 5 to 80 mm/yr. This observa­ understood. Is it a result of far-field dis­ pretation of cooling ages for detrital grains tion suggests that increases in alpine placements, as postulated for the back-arc derived from the Himalayas. The large vol­ glaciation, which could be caused by stretching that formed the Aegean meta­ ume Qf sediment in the Himalayan fore­ global cooling, increased precipitation, morphic core complexes, or is it due to land and in the offshore Bengal and Indus and/or growth of mountainous topogra­ collapse of overthickened crust, as has fans provides an important record of ero­ phy, could play a major role in exhuma­ been suggested for the Himalayan-Tibetan sional exhumation for the Himalayas. Paul tion of metamorphic rocks and in limiting orogen? Fitzgerald reviewed fission-track data that the maximum height of mountains. In his keynote presentation, Gordon indicate long-term erosion rates of -2 Sean Willett reviewed the results Lister stressed the importance of far-field mm/yr for Denali (Mount McKinley) in of numerical models of orogenesis that displacements. For the Aegean, he argued southern Alaska. account for both deformation and erosion. that back-arc stretching was caused by David Rowley described the exhuma­ He emphasized that deformation and ero­ rollback of the subducting slab. He also tion history of UHP rocks in Dabie Shan, sion are strongly coupled, so that their rel­ emphasized the transient nature of meta­ China, where exhumation rates are esti­ ative effects on the orogenic system may morphism in extensional settings. mated to have been 3 to 11 mm/yr. He be difficult to separate. Leigh Royden showed how convec­ presented a variant of the tectonic extru­ Niels Hovius presented the results tive flow in thickened crust might account sion model, where UHP rocks moved of a regional-scale inventory of landslides for deformation of the Tibetan plateau. upward as a large tabular body, with the in the Southern Alps of New Zealand as She argued that the style of deformation relative motion resolved along a lower observed over the past several decades. in a convergent orogen is controlled by thrust fault and an overlying normal fault. He estimated that erosion due to land­ the average strength of the crust, parti- The front of the extruded body is eroded slides was occurring at rates ranging from back at a rate close to the extrusion rate. 5 to 18 mm/yr. Earthquakes seem to be Penrose Conference continuedon p. 20

GSA TODAY, May 1997 19 Penrose Conference continuedfrom p. 19 the instability in the real mantle remain mantle peridotites. The presence of these poorly understood. Molnar finished with rocks at the Earth's surface suggests that tioning of strain, and the degree of cou­ an overview of a current project that com­ accretionary wedges must have sufficient pling between the crust and mantle. bines paleobotany with meteorology to upward flow to overcome the negative John Dewey presented evidence that estimate paleoelevation, with the objective buoyancy. This inference implies that UHP rocks of the Western Gneiss region to resolve the uplift history of the Col­ buoyancy is not a significant rate-limiting in the Norwegian Caledonides were orado Plateau and other high-elevation factor in the exhumation of metamorphic exhumed tectonically during oblique areas. rocks. continent/continent collision. Exhuma­ Alan Glazner and John Tarney The conference highlighted some tion rates were estimated to have been reviewed how magmatism might influ­ major challenges in understanding the -10 mm/yr. He discounted erosion ence extensional deformation. Glazner origin of UHP rocks. One important issue exhumation because basinal strata that focused on buoyancy of magmas and their concerns the relationship of the petrologic overlie the Western Gneiss contain no level of emplacement in the crust. Tarney Moho to the seismic Moho in areas with HP or UHP detritus. Dewey also noted discussed how magmatism might trigger anomalously thick continental crust. that the Western Gneiss was subjected crustal extension. Another concerns the tectonic setting to extreme vertical shortening, locally as The panel, chaired by Neil Manck­ in which UHP metamorphic rocks are much as -75%, which indicates that duc­ telow, started with a discussion about the formed. For instance, where would one tile thinning was probably an important concept of gravitational potential energy expect to find these rocks forming today? factor in exhuming these rocks. for understanding conditions leading to If UHP metamorphism occurs while The panel discussion, chaired by gravitational collapse. This was followed crustal slices are in the mantle, then this George Davis, started with the comment by two questions. How does magmatism phenomenon may be unrelated to tec­ that the Kapuskasing uplift of Canada, affect deformation and heat transfer dur­ tonic processes that thickened the crust. which exposes deep-seated Precambrian ing extension? What is the relationship of crust, may provide a particularly well UHP metamorphism to magmatism? IN MEMORIAM studied example of gravitational collapse. Just prior to the conference, one of The panel then considered some more SUMMING UP the student participants, Steve Thornley, speculative questions. Would the exhuma­ The meeting ended with summaries was killed in an accident in the Himalayas. tional phenomena be better understood from the panel chairs. The general con­ Steve had been invited to give a presenta­ by dividing the tectonic setting on the sensus was that the conference succeeded tion at the conference. He had just fin­ basis of the "absolute" motion of the in providing a broad overview of the ished a Ph.D. under Dick Walcott at upper plate? For instance, did the upper exhumation problem. In particular, it is Victoria University, Wellington, where he plate retreat (Aegean), advance (Alps), or clear that in most active mountain belts, studied extensional deformation operating surrender (lithospheric delamination in erosion and tectonism are dynamically at the surface of the Hikurangi accre­ the Himalaya)? Is fault dip related to tec­ coupled to the point where it may be diffi­ tionary wedge, New Zealand. We were tonic process, with gravitational collapse cult to separate cause and effect. It appears saddened by the death of this promising producing gently dipping normal faults that erosion can operate at very fast rates, young scientist. and rifting resulting in moderately dip­ perhaps as high as 15 mm/yr, given suffi­ ping normal faults? cient precipitation, steep terrain, and/or ACKNOWLEDGMENTS extensive alpine glaciation. These rates We are thankful to Alexandros SESSION 6: could be sustained for long periods of Papaderos, general director of the Ortho­ INFLUENCE OF DEEP-SEATED PHE­ time, as long as uplift rates continued to dox Academy of Crete, and Evanglos NOMENA ON THE GEODYNAMIC match erosion rates and climate condi­ Kastrinakis, manager of the academy, EVOLUTION OF MOUNTAIN BELTS tions remained favorable for fast erosion. along with their staff, for making our stay Deep-seated processes have also been Alpine glaciation appears to be the most comfortable and our scientific proceeding proposed as important factors controlling aggressive agent of erosion, one that is efficient and productive. Conference coor­ exhumation in mountain belts. In par­ particularly sensitive to global climate. In dinator Lois Elms did an excellent job of ticular, there has been much emphasis this regard, the relatively high sediment handling the arrangements and logistics, on thermal softening of an orogenic root production rate of the Quaternary may be especially given the complexities associ­ and delamination of a thickened mantle the result of a cooler climate and more ated with an overseas venue. We are root. Thermal softening would reduce the extensive alpine glaciation. A problem, grateful to Bernard Stockhert and Eberhard strength of the crust, and mantle delami­ however, is that much of our current Seidel for suggesting the academy as a nation would increase the average topog­ understanding of erosion rates is based conference site and for leading the confer­ raphy of the orogen. Both phenomena on relatively short records. There is a seri­ ence field trips. We thank Gordon Lister could lead to gravitational collapse. ous need for better long-term estimates and Marnie Forster for organizing and The theme of Peter Molnar's keynote using the sediment inventories or thermo­ leading an informal week-long post­ presentation was that normal faulting is chronometry. conference trip to the Aegean metamor­ typically initiated by some major change Buoyancy remains an often-cited phic core complexes. They put together in the geodynamic system, such as a factor for return of HP and UHP rocks to an excellent field guide, "Inside the change in boundary conditions (e.g., plate the surface, but there are an increasing Aegean Metamorphic Core Complexes." motions), a change in the constitutive number of examples where relatively Copies are available from Lister (gordon@ behavior of the crust (e.g., thermal soften­ dense rocks have been exhumed. Geolo­ artemis.earth.monash.edu.au). We also ing), or a change in the position of rocks gists have generally accepted the possibil­ thank Suzanne Baldwin, Meg Coleman, relative to a heterogenous stress field. He ity that subduction is able to carry conti­ Darrel Cowan, Carol Evenchick, John then reviewed the current understanding nental crustal rocks to depths >100 km, Garver, and Susan Monsen for their of Rayleigh-Taylor instabilities in thick­ perhaps because the driving force involves comments on a preliminary draft of this ened lithosphere, which might explain the negative buoyancy of subducting report. The conference was supported in the conditions needed for mantle delami­ oceanic lithosphere. What remains con­ part by National Science Foundation grant nation. This type of instability has been ceptually difficult is how tectonic forces EAR-9628304 and the Geological Society investigated using analog models, but the can return rocks to the surface, especially of America.• characteristic time and length scales of relatively dense eclogite-facies rocks and

20 GSA TODAY, May 1997