Rupture de surface du séisme de Bihar-Népal 1934 : implications pour l'aléa sismique en Himalaya du Népal

Surface rupture of 1934 Bihar-Nepal Earthquake : implications for seismic hazard in Nepal Himalaya

• Auteur / Autrice : Soma Nath Sapkota
• Direction : Paul Tapponnier, Yann Klinger
• Type : Thèse de doctorat
• Discipline(s) : Géophysique interne
• Date : Soutenance en 2011
• Etablissement(s) : Institut de physique du globe (Paris ; 1921-....)

Mots clés

Mots clés contrôlés

Himalaya (Népal)

Géomorphologie

Paléosismologie

Népal

Morphotectonique

Résumé

Although the Main Himalayan Frontal Thrust (MHT/MFT), largest and fastest‐slipping continental megathrust, poses a major threat to the northern Indian sub‐continent, seismic hazard along it remains to be quantified. Based on historical descriptions of the two main 20th century earthquakes (1905, 1934), a consensus has emerged that neither produced surface ruptures, a view recently reinforced by paleo‐seismological investigations in which only faulting much older than 1900 was found. This leaves us with fundamental, unanswered questions (recurrence times, rupture lengths, geomorphic signature of large events), and the ominous perspective of even greater quakes with displacements in excess of 15m, potentially on par with M ≈ 9 oceanic subduction events. Knowing the precise geometry and history of earthquake ruptures along this very active fault is thus more than ever critical to assess seismic hazard in the area. To address such questions we have engaged high‐resolution geomorphic and paleo‐seismic studies of Main Frontal Thrust (MFT) in eastern Nepal. Long wavelength warping of river terraces show that late Pleistocene/Holocene deformation is well expressed across frontal folds above the thrust, which have been successfully used to determine a shortening rate on order of 2 cm/yr, but the surface trace of the MFT, where sharpest, remains the best location to document whether large earthquakes break the ground and to determine their sizes and recurrence times. Our survey of the area between the Mahara Khola and Arun/Sun Khosi valleys shows many clear outcrops with young, pristine scarps, challenging the consensus that no surface rupture took place in the 19th century. This area is also located at the centre of the region comprised entirely within the 1934 isoseismal VIII. In the valley of the Sir Khola, which crosses the northern branch of the MFT, we logged in detail a refreshed river‐cut face across the 26 m‐high cumulative thrust scarp. Newly dated charcoal samples collected in the gravel layers of an uplifted strath terrace offset by thrusts reaching the surface confirm the young age of the last event. Six distinct 14C calibrated dates indicate that the terrace was emplaced less than 250 years ago, in the 18th or early 19thcentury, and was subsequently offset by F1, with a vertical throw of ≈ 1. 5 m (≈ 3 m of slip). This same terrace is offset again by another fault at the base of the main scarp. Since no other large earthquake than the Bihar‐Nepal event was recorded locally in the 19th and 20th century, it must be concluded that the Sir Khola rivercut exposes the first unambiguous surface trace ever found of the 1934 earthquake. Similarly, a trench excavated 30 m east of the Natural river‐cut shows two events approximately 700 years apart. We thus interpret the great 1934 earthquake to be a repeat of the 1255 AD event that destroyed Kathmandu. Other charcoal ages in lower fluvial units of the footwall rapidly jump to 3000 years BP, and to older ages just below (up to 7000 yrs), nearly at the same level as that of the present river. This implies that stratigraphic section is missing, but also that there has been no significant longterm incision by the river just south of the thrust. In the Charnath Khola area, the dating of different uplifted terraces on the MFT hanging‐wall, and of one on the footwall, supports our findings at the Sirkhola and Ratu Nadi. Here, the 14‐16 m high Tintale terrace, whose age is only 1‐1. 25 kyr, appears to have been uplifted by at least two events (possibly three) with 4‐5m of vertical coseismic throw in each event since its abandonment. 14C dating of the lowest hanging‐wall terrace (≈ 4 m‐high) in Tintale creek, which was abandoned less than three centuries ago supports the existence of the 1934 rupture. Geophysical surveys including three shallow seismic profiles and Electrical Resistivity Tomographic (ERT) sections, ≈ 1. 5 km‐long each, in the Charnath, Sirkhola and Ratu Valleys, and one Ground Penetrating Radar (GPR) profile at Thapatol near Bardibas, add invaluable information at various depths, complementing our morpho‐tectonic interpretation of the area. In particular, the seismic profiles shot across the MFT image well the shallow part of the thrust‐plane down to ≈ 400mepth. To our knowledge, at least in Nepal, this is the first effort of this kind to study the MFT at such detailed scale in combination with paleo‐seismological trenching.