Geol Rundsch (1994) 83: 822-831 © Springer-Verlag 1994 N. Kukowski. R. von Huene • J. Matavieille S. E. Lallemand Sediment accretion against a buttress beneath the Peruvian continental margin at 12 ° S as simulated with sandbox modeling Received: 21 January 1994/ Accepted : 16 June 1994 Abstract Reflection seismic data from the Peruvian continental margin at 12 ° S clearly reveal an accretionary Introduction wedge and buttress. Sandbox experiments applying the physical concept of the Coulomb theory allow the The concept of a buttress or backstop has been firmly systematic investigation of the growth and deformation implanted in published work on convergent margins of such an accretionary structure. The style of deforma- (Byrne etal., 1988; 1993). It is considered to be the tion of the buttress and the internal structure of the wedge seaward edge of the island arc or continental margin older is observed in the sandbox models. The possibility of rock framework against which an accretionary prism underplating material beneath the buttress and the forms (von Huene and Scholl, 1993). Conceptually, this amount of tectonic erosion depend on the physical body is the anvil against which offscraped trench sedi- properties of the materials, mainly internal friction, ment is accreted and tectonically deformed. cohesion and basal friction. Boundary conditions such as A good example of an actual buttress is imaged in the height of the subduction gate and the thickness of seismic records across the Peruvian continental margin. incoming sand also constrain the style of growth of the Beneath the slope, a wedge-shaped body of crystalline model accretionary structure. rock with probable continental affinity has been imaged The configurations of two experiments were closely through careful processing of seismic records (Moore and scaled to reflection seismic depth sections across the Taylor 1988; von Huene and Miller 1988; yon Huene, in Peruvian margin. A deformable buttress constructed of press). Against this wedge-shaped buttress a prism of compacted rock powder is introduced to replicate the trench sediment about 10 km wide has been accreted. basement rock which allows deformation similar to that Deformation visible in the overlying sediments may give in the seismic data. With the sandbox models it is possible a hint that deformation also occurs in the narrow tapered to verify a proposed accretionary history derived from end of the buttress. seismic and borehole data. The models also help in The evolution and deformation of the accretionary understanding the mechanisms which control the amount body in front of a buttress can be forward modeled of accretion, subduction and underplating as a function applying a sandbox technique. However, in most former of physical properties, boundary conditions and the models the buttress was modeled as a hard, non-defor- duration of convergence. mable member and sand was accreted against a landward Key words Accretionary prisms • Convergent margins- or seaward dipping rigid stationary buttress (Byrne et al., Peru. Analogue modeling. Tectonics • Accretion - Tecto- 1988; Mulugeta 1988; Lallemand et al., 1992). To achieve nic erosion. Marine geophysics more realistic conditions, Byrne et al. (1993) used a defor- mable backstop made of wet sand. Considering the tectonic deformation in the buttress off Peru, we intro- duced a deformable buttress (compacted rock powder) in the sandbox modeling. Furthermore, we scaled the mo- N. Kukowski (~:~); R. von Huene dels after observations in seismic records across the GEOMAR, Forschungszentrum ffir Marine Geowissenschaften, Peruvian margin to replicate features observed in nature Wischhofstrasse 1--3, D-24148 Kiel, Germany, Fax: (49)-431- 72 02-293, e-mail: [email protected] and we allowed material to leave the system through a 'subduction window' to simulate that in nature a certain J. Malavieille - S. E. Lallemand Equipe de G6ophysique et Tectonique, Universit6 Montpellier II, amount of material is transported to greater depths. Case postale 060, Place E. Bataillon, F-34095 Montpellier Cedex In this paper we report the results of our sandbox 05, France modeling. A main objective of our models was to test 823 8°S IO°S 12°S 82°W 80°W 78°W 76°W 74°W Fig. 1. Map of the Peruvian continentalmargin showing available lines were pre-stack depth migrated at GEOMAR (von reflection seismic lines near 9°S and 12°S latitude and ODP drillholes Huene and Pecher, unpublished data). From early wide-angle seismic data (Hussong and Wipperman, 1981) it was known that seismic velocities of a kinematic restoration proposed from geophysical and 5 km s -1 and greater characterize upper plate rocks geological observations off Peru that include information landward of the first 10 km of the margin. In seismic from ODP drilling (yon Huene, Suess et al., 1988). By reflection records, coherent reflections were visible along observing the intermediate stages during convergence in boundaries above and below a high velocity wedge, but the sandbox, we gained many ideas about the possible incoherent events characterized its interior. Rocks reco- tectonic history, the material paths and the parameters vered from outcrops at the edge of the shelf and along the controling the evolution of the tectonic structure of lower slope indicated that the area of incoherent reflec- convergent margins. Observing the processes taking place tions could represent equivalents of the schists and in the sandbox, we saw that the ratio of material supplied gneisses that are exposed in the coastal mountains. After by the moving plate to material removed from the system drilling leg 112, these rocks were observed from a submer- through the subduction window may be one of the most sible in outcrops at the foot of the margin in Chiclayo critical parameters controlling the accretionary evolu- Canyon (Sosson et al., 1992). During leg 112 a shallow tion. water Eocene sandstone which is known to rest unconfor- mably on the metamorphic rock cropping out at the edge of the shelf was recovered from two sites on the lower slope near 12°S. The unconformity is seen in all multi- channel seismic records across the continental shelf and is Tectonic history of the Peruvian margin inferred to extend onshore. This regional unconformity, now up to 5 km below sea level, was a subaerial erosion Four seismic lines (1017, 1018, CDP-1, HIG-14) covering surface in pre-Middle Eocene time (von Huene, Suess about 30 km along the Peru margin near 12 ° S show the et al., 1988) because the shallow water continental origin accretionary wedge and the buttress (Fig. 1). Line CDP-1 of the sandstone on the unconformity makes it unlikely from the Nazca Plate Project and Shell 1017 were that the underlying basement is an Eocene deep water processed in the time domain to a pre-stack migration in accretionary wedge. Minor deformation affects the thin preparation for ODP drilling on leg 112 (yon Huene and tapered part and only extensional deformation of the Miller, 1988). HIG-14, which was also a part of the basement is observed to cut the basement from the middle drilling transect (Moore and Taylor, 1988), was post- slope landward (Fig. 2). These observations indicate that stack time migrated. The Shell line 1018 is also very close the wedge-shaped buttress is a relatively rigid body to the main drilling sites. Because of this proximity and probably composed of metamorphic rock similar to that because the Shell data are of fairly high quality, both Shell in the coastal massif (Kulm et al., 1988). 824 $W NE CDP o o o° ~ 8 ~ o~ 3 o° l I I I I I l I 4 ~ .... I .. 4 Is] [s] ! ..~. -~.~, 8 a - MIOCENE - PLIOCENE 4 - -4 6- -6 8- -8 10- -10 KM b Fig, 2. a Time-migratedstack ofthenear-trenchpart of CDP-1 and One aim of our study was to test this kinematic b interpreted drawing showing the accretinary prism and buttress interpretation physically with sandbox experiments. In the northern lines (CDP-1, Shell 1018, HIG 14), the Sandbox experiments buttress forms a wedge tapering about 20 °. The upper surface is usually bounded by a continuous reflection with Model concept strong amplitude (Fig. 2), which correlates with the Modeling the evolution of an accretionary wedge with the Eocene shallow water sandstone at ODP site 688. Its sandbox technique is based on the assumption that the lower boundary is less distinct because of the greater frontal part of a subduction zone, i.e. the accretionary depth and complex ray path; however, underplated wedge, the underplated sediment and the frontal part of sediment is consistently imaged there. Subducting sedi- the buttress, behaves as a Coulomb material (Davis et al., ment beneath the inferred metamorphic rocks is observed 1983; Dahlen, 1984). Thus the scale-independent Cou- many tens of kilometers landward along the plate bound- lomb theory can be applied. Dahlen (1984) derives an ary. exact solution for the critical taper a uniform wedge for The evolution of the accretionary wedge began when the case of a non-cohesive material. The mechanical state accretion was initiated after the Nazca Ridge trailing of the accretionary wedge is, for a given rock density, flank had been subducted (von Huene and Miller, 1988) a function of only four parameters, the internal and basal (Fig. 3). Several studies show that the area, where the pore pressures and friction angles, but not of time or Nazca Ridge subducts, migrated southward with time temperature. However, natural rocks are cohesive mate- (Cande 1985; yon Huene and Lallemand, 1990). From rials. Dahlen etal. (1984) showed that the Coulomb plate tectonic reconstruction (Nur and Ben Avraham, theory can be extended to cohesive materials with cohe- 1981) the time when ridge trailing flank subducted at the sion, Co, as the fifth of the parameters which characterize trench axis near the position of 1018 is estimated at 3 Ma.