The geometry of the subducted Rivera and Cocos plates beneath the North American plate in southern Mexico was determined based on the accurately located hypocenters of local and te!eseismic earthquakes. The hypocenters of the teleseisms were relocated, and the focal depths of 21 events were constrained using a body wave inversion scheme. The suduction in southern Mexico may be approximated as a subhorizontal slab bounded at the edges by the steep subduction geometry of the Cocos plate beneath the Caribbean plate to the east and of the Rivera plate beneath North America to the west. The dip of the interplate contact geometry is constant to a depth of 30 kin, and lateral changes in the dip of the subducted plate are only observed once it is decoupled from the overriding plate. On the basis of the seismicity, the focal mechanisms, and the geometry of the downgoing slab, southern Mexico may be segmented into four regions ß (1) the Jalisco region to the west, where the Rivera plate subducts at a steep angle that resembles the geometry of the Cocos plate beneath the Caribbean plate in Central America; (2) the Michoacan region, where the dip angle of the Cocos plate decreases gradually toward the southeast, (3) the Guerrero-Oaxac.a region, bounded approximately by the onshore projection of the Orozco and O'Gorman fracture zones, where the subducted slab is almost subhorizontal and underplates the upper continental plate for about 250 kin, and (4) the southern Oaxaca and Chiapas region, in southeastern Mexico, where the dip of the subduction gradually increases to a steeper subduction in Central America. These drastic changes in dip do not appear to take place on tear faults, suggesting that smooth contortions accommodate these changes in geometry. The inferred 80 and 100 km depth contours of the subducted slab lie beneath the southern front of the Trans-Mexican Volcanic Belt, suggesting it is directly related to the subduction. Thus the observed nonparallelism with the Middle American Trench is apparently due to the changing geometry of the Rivera and Cocos plates beneath the North American plate in southern Mexico, and not to zones of weakness in the crust of the North American plate as some authors have suggested. 1969; Stoiber and Carr, 1973; Dean and Drake, 1978; Nixon, 1982; Beyis and lsacks, 1984; Burbach et aL, 1984; Ponce et al., 1992]. A steep subduction geometry in southeastern Mexico [Ponce et al., 1992] and in the Jalisco region of western Mexico [Pardo and Sudrez, 1993] bound a slab dipping apparently at a very shallow angle in central Mexico.On the basis of teleseismic data, Burbach et aI. [19134] concluded that the subducted Cocos plate can be divided into these three major regions dipping at different angles. Beyis and lsacks [1984], using a hypocentral trend surface analysis, find no evidence of tear faulting separating these segments of the subducted lithosphere •[nd" suggested that the transition 1Now at

No abstract

No abstract

Abstract. The configuration of the Pacific plate subducted beneath the Kamchatka peninsula and the stress distribution in the Kamchatka subduction zone (KSZ) were studied using the catalog of the Kamchatka regional seismic network, focal mechanism solutions estimated from P wave first motions, the formal inversion of long-period waveforms, and centroid moment tensor solutions. To the south of-55øN, the slab shows an approximately constant dip angle of -55 ø. To the north of-55øN, the dip of the slab becomes shallower reaching -35 ø. The maximum depth of seismicity, Din, varies from -500 km depth near 50øN to -300 km depth at -55øN. The volcanic front is almost linear along the main part of the KSZ whereas it is sharply shifted landward to the north of-55øN. The variation of Dm is apparently consistent with the standard empirical relation Dm --f (rp), where rp is the thermal parameter of the subducted slab.To the north of-55øN, the slab is offset toward the northwest, and it is sharply deformed in a narrow contorted zone which is -30 km wide (-56øN, -161 øE). To the north of this contortion, Dm decreases to -100 km. The landward shift of the northern part of the slab is reflected by a sharp deviation of the volcanic front to the northwest which follows the -90-160 km isodepth range of the subducted slab. The observed value of Dm in the northern segment significantly diverges from the global relation Dm --f (rp). We interpret this as an effective decrease of the thermal thickness of the subducted lithosphere.

A great earthquake (surface-wave magnitude, 7.8) occurred along the coast of central Chile on 3 March 1985, causing heavy damage to coastal towns. Intense foreshock activity near the epicenter of the main shock occurred for 11 days before the earthquake. The aftershocks of the 1985 earthquake define a rupture area of 170 by 110 square kilometers. The earthquake was forecast on the basis of the nearly constant repeat time (83 +/- 9 years) of great earthquakes in this region. An analysis of previous earthquakes suggests that the rupture lengths of great shocks in the region vary by a factor of about 3. The nearly constant repeat time and variable rupture lengths cannot be reconciled with time- or slip-predictable models of earthquake recurrence. The great earthquakes in the region seem to involve a variable rupture mode and yet, for unknown reasons, remain periodic. Historical data suggest that the region south of the 1985 rupture zone should now be considered a gap of high seismic potential that may rupture in a great earthquake in the next few tens of years.