Structural interpretation of the great earthquakes of the last millennium in the central Himalaya

A major question about the Himalaya remains open: does a great earthquake (like the Mw similar to 8.1 1934 earthquake) release all the strain stored by the Tibet-India convergence during the preceding interseismic period and only that strain, or can it also release a background store of energy that remained unreleased through one or more earlier earthquakes and so potentially engender giant events or a relatively random sequence of events? To consider this question, the history of the great earthquakes of the last millennium is investigated here by combining data provided by the historical archives of Kathmandu, trenches through surface ruptures, isoseismal damage mapping, seismites, and the instrumental record. In the Kathmandu basin, the location of the epicenter of the 1934 earthquake was determined from the arrival of high-energy P-waves that caused sedimentary dikes and ground fractures perpendicular to the epicenter azimuth. The epicenter of the Mw similar to 7.61833 earthquake can therefore be determined analogously from dike orientation, and its location to the NE of Kathmandu indicates an overlap with the Mw similar to 8.1 1934 rupture. The 1934 earthquake released strain not released by the 1833 earthquake. Comparison of the historical records of earthquakes in Kathmandu with C-14 ages from paleo-seismic trenches along the Himalayan front suggests that: (I) the 1344 Kathmandu event ruptured the surface as far away as Kumaon and was therefore a giant Mw >= 8.6 earthquake; and (2) the 1255 event that destroyed Kathmandu is attested by surface ruptures in central and western Nepal and by seismites in soft sediment as far away as Kumaon. Geometric and rheologic controls for the different types of ruptures during the medium (Mw similar to 7), great (Mw >= 8), and giant (Mw > 8.4) earthquakes are illustrated in structural cross-sections. It is found that the epicenters of great Himalayan earthquakes are located on the basal thrust farther north or close to the locked zone, which is defined from geodetic measurements of regional deformation during the interseismic period; this suggests that great earthquakes initiate in a wide transition zone between exclusively brittle and exclusively creeping regimes, the extent of which depends on the dip of the Main Himalayan Thrust. The succession of the great earthquakes during the last millennium has released all the 20-m millennial Himalayan convergence; even in the central seismic gap which has been locked since 1505, the millennial seismic release rate is close to the convergence rate. Nonetheless, no evidence of a succession of characteristic earthquakes has been found: the similar to 1100, 1833, and 1934 earthquakes in the eastern Himalaya are characterized neither by constant displacement nor by constant recurrence. Furthermore, some great earthquakes do not release all the strain elastically stored by the Himalayan and Tibetan upper crust: after the 1255 event, there was still the potential for a slip of several meters for the Mw similar to 8.1 1505 event. This suggests a rather random release of seismic energy; great earthquakes could occur anytime and in any part of the central Himalaya. Furthermore, a future giant earthquake of Mw >= 8.6 cannot be excluded. (C) 2013 Elsevier B.V. All rights reserved.