1. A discussion on Subduction Zones: Links between Structure and Seismogenesis Aron Meltzner TO Brownbag Discussion Group 29 November 2005 Aron Meltzner TO Brownbag Discussion Group 29 November 2005 Image of southwestern Japan. Figure from Sugiyama (1994) overlying image from Google Earth ©MDA EarthSat.

2. Background coseismic slip in great subduction earthquakes is usually nonuniform and contains asperities: regions of higher slip or moment release coseismic slip in great subduction earthquakes is usually nonuniform and contains asperities: regions of higher slip or moment release debate lingers as to whether asperities represent permanent geologic structures that control the rupture process or rather represent the filling of seismic gaps debate lingers as to whether asperities represent permanent geologic structures that control the rupture process or rather represent the filling of seismic gaps there is evidence to support both arguments there is evidence to support both arguments

3. Background in some cases, successive large earthquakes within a given fault segment have not ruptured the same source region (e.g., the 1957–1986–1996 sequence along the Andreanof segment of the Aleutian arc) ; stress transfer and other dynamic processes may be responsible for the sequence in some cases, successive large earthquakes within a given fault segment have not ruptured the same source region (e.g., the 1957–1986–1996 sequence along the Andreanof segment of the Aleutian arc) ; stress transfer and other dynamic processes may be responsible for the sequence some asperities appear to persist from one seismic cycle to the next: GPS data from Japan and southern Alaska indicate that centers of slip in previous great earthquakes are now preferentially accumulating strain some asperities appear to persist from one seismic cycle to the next: GPS data from Japan and southern Alaska indicate that centers of slip in previous great earthquakes are now preferentially accumulating strain

4. Background in great earthquakes, rupture may extend over 1000+ km in length and the entire width of the seismogenic zone, producing huge asperities in great earthquakes, rupture may extend over 1000+ km in length and the entire width of the seismogenic zone, producing huge asperities structures potentially correlative with these asperities should be recognizable in the large-scale architecture of the forearc structures potentially correlative with these asperities should be recognizable in the large-scale architecture of the forearc along the Nankai Trough, some locked patches and high-slip regions appear to correlate with offshore basins; history suggests that the source regions have persisted over many seismic cycles along the Nankai Trough, some locked patches and high-slip regions appear to correlate with offshore basins; history suggests that the source regions have persisted over many seismic cycles

5. Wells et al. (2003) 1944 1946

6. Basic Structural Features inverted L -shaped high or composite anticline fringing a forearc basin inverted L -shaped high or composite anticline fringing a forearc basin reverse faults along unit boundary showing relative uplift of left (west) block reverse faults along unit boundary showing relative uplift of left (west) block nose structure protruding landwards along unit boundary nose structure protruding landwards along unit boundary folds with axes perpendicular to trench on the right (east) side of the unit boundary folds with axes perpendicular to trench on the right (east) side of the unit boundary dextral and sinistral megakink folds on the right and left sides of the unit boundary, respectively dextral and sinistral megakink folds on the right and left sides of the unit boundary, respectively

7. Coseismic Deformation at the Earth’s Surface for a N-dipping right-oblique thrust 1. uplift above the S and E margins of the fault plane 2. subsidence above the NW corner of the fault plane 3. E–W shortening along E margin of the fault plane 4. both ( 1 ) and ( 3 ) decrease northward ( 1 ) correlates with inverted L -shaped structural high ( 2 ) correlates with forearc basin ( 3 ) correlates with N–S-trending reverse faults and folds northward decrease of coseismic uplift agrees with northward tilting of marine terraces at promontories northward decrease of E–W shortening consistent with nose structure and megakink folds

9. Background the forearc basins may simply indicate strong forearc crust at depth, with passive basin fill trapped behind the growing accretionary prism [Byrne et al., 1988] the forearc basins may simply indicate strong forearc crust at depth, with passive basin fill trapped behind the growing accretionary prism [Byrne et al., 1988] alternatively, the downward motion of the lower plate during earthquakes may have created the depression overlying the rupture zone [Mogi, 1969] alternatively, the downward motion of the lower plate during earthquakes may have created the depression overlying the rupture zone [Mogi, 1969] Sugiyama [1994] considered the basins in part to be the product of cumulative interseismic subsidence not recovered during earthquakes Sugiyama [1994] considered the basins in part to be the product of cumulative interseismic subsidence not recovered during earthquakes

10. Wells et al. (2003)

11. Background structural segmentation results from along-strike changes in plate geometry and strain partitioning in the upper plate structural segmentation results from along-strike changes in plate geometry and strain partitioning in the upper plate common causes include: common causes include: incoming fracture zones incoming fracture zones oblique subduction and margin parallel deformation oblique subduction and margin parallel deformation inherited transverse structures inherited transverse structures ridge subduction and seamount “tunneling” ridge subduction and seamount “tunneling”

12. Kodaira et al. (2002)

13. Briggs et al. (submitted) Sumatra: Uplift & Subsidence Sumatra: Uplift & Subsidence

15. Sumatra: Bathymetry & March 2005 Coseismic Slip Briggs et al. (submitted); slip model by Y. Hsu after Sieh & Natawidjaja (2000)

16. Wells et al. (2003)

19. From Wells et al. (2003): The lack of correlation between forearc lows and slip in the great Alaska earthquake is in large part due to the anomalous outer arc gravity high. The high may be caused by the unusually flat-lying, dense lower plate, a local doubling of oceanic crustal thickness, and resultant uplift of the upper plate due to the collision of the Yakutat terrane.

20. Sumatra: Bathymetry & December 2004 Coseismic Slip after Sieh & Natawidjaja (2000) Subarya et al. (submitted); slip model by R. McCaffrey Subarya et al. (submitted); slip model by M. Chlieh

21. Points for Discussion from Wells et al. (2003) the commonly observed trenchward decrease in seismic slip could reflect landward bias in geodetic observations, but tsunami and seismic inversions give similar results and suggest the pattern of deeper, basin-centered asperities is real the commonly observed trenchward decrease in seismic slip could reflect landward bias in geodetic observations, but tsunami and seismic inversions give similar results and suggest the pattern of deeper, basin-centered asperities is real although a few tsunami inversions document coseismic slip beneath the prism, in general, the highest slip region tends to be deeper beneath the forearc basins and upper slope although a few tsunami inversions document coseismic slip beneath the prism, in general, the highest slip region tends to be deeper beneath the forearc basins and upper slope how does the Nias-Simeulue earthquake fit it? how does the Nias-Simeulue earthquake fit it?

22. Points for Discussion from Wells et al. (2003) some transverse, intrabasin highs apparently overlie areas of lower slip in great earthquakes some transverse, intrabasin highs apparently overlie areas of lower slip in great earthquakes 1952 and 1968 earthquakes off Hokkaido ruptured on either side of the gravity high at Cape Erimo, but little or no slip occurred beneath the Cape in either earthquake 1952 and 1968 earthquakes off Hokkaido ruptured on either side of the gravity high at Cape Erimo, but little or no slip occurred beneath the Cape in either earthquake the Cape is either a potential asperity storing up great slip, or it is relatively weak, possibly because of trapped heat or fluids beneath the thicker crust the Cape is either a potential asperity storing up great slip, or it is relatively weak, possibly because of trapped heat or fluids beneath the thicker crust GPS inversion at Cape Erimo is consistent with less than full locking of the underlying plate boundary GPS inversion at Cape Erimo is consistent with less than full locking of the underlying plate boundary

23. Points for Discussion from Wells et al. (2003) a weaker fault beneath the Cape could allow some aseismic slip there, causing earthquakes to nucleate along the Cape Erimo gradient and rupture primarily under the adjacent basins a weaker fault beneath the Cape could allow some aseismic slip there, causing earthquakes to nucleate along the Cape Erimo gradient and rupture primarily under the adjacent basins analogies include: analogies include: the Kii Peninsula, bounding the 1944 & 1946 earthquakes along the Nankai Trough the Kii Peninsula, bounding the 1944 & 1946 earthquakes along the Nankai Trough the Shumagin Islands high off Alaska the Shumagin Islands high off Alaska the Batu Islands and the “Simeulue Saddle” ?? the Batu Islands and the “Simeulue Saddle” ??

24. after Sieh & Natawidjaja (2000)

25. Points for Discussion from Wells et al. (2003) not all forearc highs are areas of lower seismic slip not all forearc highs are areas of lower seismic slip in the 1952 Kamchatka earthquake, one of the three asperities was centered on a prominent transverse high in the 1952 Kamchatka earthquake, one of the three asperities was centered on a prominent transverse high in the 1946 Nankaido event, significant slip occurred beneath the southwestern Kii Peninsula in the 1946 Nankaido event, significant slip occurred beneath the southwestern Kii Peninsula along the Nankai margin, the transverse anticlinal highs accommodate significant oblique slip on splay faults and may be important components of coseismic deformation along the Nankai margin, the transverse anticlinal highs accommodate significant oblique slip on splay faults and may be important components of coseismic deformation

26. Points for Discussion from Wells et al. (2003) the tendency for coseismic slip to be focused beneath the terrace and its basins, rather than beneath the intervening highs, could reflect along-strike variations in the temperature, fluid pressures, and stresses on the megathrust caused by variations in overlying crustal thickness and density the tendency for coseismic slip to be focused beneath the terrace and its basins, rather than beneath the intervening highs, could reflect along-strike variations in the temperature, fluid pressures, and stresses on the megathrust caused by variations in overlying crustal thickness and density large-scale segmentation of the source zone would then result from oblique convergence, subducting fracture zones, or some other second-order process large-scale segmentation of the source zone would then result from oblique convergence, subducting fracture zones, or some other second-order process the question remains whether there is a relationship between forearc basin evolution and seismogenesis the question remains whether there is a relationship between forearc basin evolution and seismogenesis

27. The End

28. 1944 1946

29. Wells et al. (2003) 1944 1946