The Segmented Overriding Plate and Coupling at the South-Central Chilean Margin (36–42°S)

Abstract

t. Identifying the parts of subduction zones that are suscep- tible to great earthquakes is a challenge that warrants consider- able attention. In south-central Chile, where the 1960 Mw 9.5 Valdivia earthquake occurred, we have combined surface geology and grav- ity data into a three-dimensional density model that helps to iden- tify trench-parallel changes in fore-arc properties between 36 and 42° S. In light of suggestions that gravity data predict the seismo- genic behavior of subduction zones, we use the gravity data and geological observations to separate the fore-arc in this region into three segments. The northern Arauco-Lonquimay segment, where gravity ano- malies are strongly positive, should be characterized by low cou- pling. In contrast, the plate interface under the Valdivia-Liquiñe and Bahia Mansa-Osorno segments to the south, where anomalies are negative or near-zero, should be highly coupled. The inferred dif- ferences in coupling are consistent with the extent of rupture dur- ing the Valdivia earthquake, which initiated under the southern part of the Arauco-Lonquimay segment but propagated southwards through the zone of inferred high coupling under the Valdivia- Liquiñe and Bahia Mansa-Osorno segments. A three-dimensional gravity model of this region, constrained by surface geology and, in part, by independent seismic informa- tion, shows that one major control on the changing gravity-anomaly characteristics is the depth to the slab below the fore-arc. The model suggests that a north-to-south increase in the depth to the slab of about 5 km is possible. This increasing depth to the slab (i.e. increas- ing fore-arc thickness) can account for the inferred increase in cou- pling under the southern segments. A deeper slab would lead to greater shear stress and an increased coupling force at the plate interface. This increase is a result of: ( ) greater normal stress acting on the plate interface induced by a thicker fore-arc, ( ) slab buoy- ) a shallower onset of sediment consolidation that ancy effects related to plate age (which also decreases from north to south), and ( increases the rigidity of material at the plate interface, thereby in- creasing the width of the frictionally coupled (unstably sliding) part of the subduction interface. Differences in fore-arc rheology, re- flected in along-strike compositional differences and seismicity patterns, also have an important influence on coupling. Other parameters that are often invoked to explain coupling dif- ferences (e.g. changes in trench sediment, convergence rate and sea- floor texture) cannot explain differences in coupling here because these parameters do not change over the length of the trench exam- ined, or they cause an effect that is contrary to the inferences based on gravity anomalies. The inferred coupling differences are also con- sistent with observed seismicity. In the Arauco-Lonquimay segment, where we infer coupling to be low, prominent seismicity is evident, indicating that the fore-arc is releasing the strain built up during con- vergence. In the Valdivia-Liquiñe and Bahia Mansa-Osorno segments, where we infer coupling to be high, fore-arc seismicity is limited. This suggests that strain is accumulating as the result of a locked plate interface, although aseismic slip cannot be ruled out