Résumé
What is the impact of the segmentation on the propagation and therefore size of major earthquakes in continental domains such as the Himalayan arc? During the last decades, several studies suggested crustal structures vary laterally along the Himalayan arc. However, the lack of paleoseismological data in key areas, such as Bhutan and the major NE-SW Yadong cross structure, hampered progress in that line of research. Here, we address this issue using paleoseismological investigations and Bayesian stratigraphic modeling from two trenches in central Bhutan. Our trench results reveal that Bhutan has been struck by at least two great earthquakes that ruptured the Main Frontal Thrust in 1713 and in medieval times. The total cumulative vertical slip reaches at least 10 m. Combined to previously published works carried out in Central Nepal, Sikkim and Assam our study supports the potential occurrence of either i) a giant earthquake between AD 1107 and AD 1141 with a surface rupture extending from central Nepal to Assam or ii) a sequence of great earthquakes along the Himalayan arc between AD 1025 and AD 1547. Following several studies on the relation between segmentation and rupture area of great subduction earthquakes, we propose to interpret our results in terms of segmentation of the Himalayan arc. Indeed, the Yadong cross structure acts as a barrier to the propagation of the rupture for great earthquakes (e.g. the AD 1713 earthquake). Conversely, giant earthquakes such as the medieval event could cross this structure and rupture more than one segment. In the case of a unique medieval earthquake along the Himalayan arc, the 800-km-length-rupture can produce an event with a magnitude in the range of 8.7-9.1.