Basin-centered asperities in great subduction zone earthquakes - Implications for 
seismic hazards and tectonic erosion at the plate boundary

Ray Wells
USGS Menlo Park
 

Abstract: 
We have compiled published coseismic slip distributions for 29 of the largest 
megathrust earthquakes in order to determine how earthquake slip is related to the 
convergent margin structure revealed by satellite gravity data. Our motivation was 
to examine the landward limit of large slip for earthquake hazard studies, but our 
results may have implications for the long term evolution of subduction zones. In 
our study, which covered 7500 km of rupture length along the Circum-Pacific margin, 
the areas of highest slip or seismic moment release tend to be statistically focused 
beneath the offshore deep-sea terrace and its forearc basins. On average, 71% of an 
earthquake's seismic moment and 79% of its high-slip area occur within the free-air 
gravity low marking the deep-sea terrace. 57% of the high- slip area lies beneath 
the terrace's forearc basins, which comprise only 21% of the seismogenic zone by 
area. In SW Japan, slip in the 1923, 1944, 1946, and 1968 earthquakes was largely 
focused beneath five forearc basins. The steep gravity gradient marking the landward 
edge of the basins coincides with the 350°C isotherm on the plate boundary 
megathrust, commonly cited as the down-dip limit of the "locked zone." Basin-centered 
coseismic slip also occurred along the Aleutian, Mexico, Peru, and Chile subduction 
zones, but was ambiguous for the great 1952 Kamchatka and 1964 Alaska earthquakes. 
The deep-sea terrace and its basins may thus define the locus of long-term seismic 
moment release. The inferred source zone for Cascadia's 1700 AD earthquake in the 
northwest U.S. and adjacent Canada contains five large, basin-centered gravity lows 
that we suggest may indicate potential high-slip areas at depth. Seismic profiling 
and/or drilling on the terrace and outer shelf off NE Japan, Chile, Peru, and Alaska 
document kilometers of late Cenozoic subsidence above the high-slip areas. In most 
cases, the sustained subsidence had previously been interpreted to be the result of 
crustal thinning due to basal subduction erosion. The coincidence of basin-centered 
coseismic slip with evidence of sustained subsidence and subduction erosion suggests 
that the erosion may be occurring in the seismogenic zone. These deep-sea terraces 
and basins may be created in part by permanent interseismic subsidence not recovered 
during earthquakes. The subsidence history of these basins and their possible link to 
subduction erosion could be tested with additional seismic profiling and deep drilling.