Potentially large post-1505 AD earthquakes in western Nepal revealed by a lake sediment record

According to paleoseismological studies, the last earthquake that ruptured the Main Frontal Thrust in western Nepal occurred in 1505 AD. No evidence of large earthquakes has been documented since, giving rise to the concept of a seismic gap in the central Himalaya. Here, we report on a new record of earthquake-triggered turbidites from Lake Rara, western Nepal. Our lake-sediment record contains eight possibly moderate-to-large earthquake-triggered turbidites during the last 800 years, three of which overlap in age with previously reported Mw ≥ 7 events in western Nepal. Shaking intensity modelling, together with instrumental records, suggests that near-field earthquakes (≤15 km) should have a minimum Mw  5.6, and regional earthquakes (≤80 km) a Mw > ~ 6.5, to trigger turbidites. We present a likely scenario that western Nepal may be as seismically active as central Nepal; however, more data are needed to revaluate the seismic risk in the central Himalaya.

The modelled earthquakes are indicated by a star. Turbidites associated with known historical/instrumental earthquakes are highlighted and the epicentral distance from Lake Rara is indicated.

Supplementary Discussion
The turbidites were identified using a series of sedimentological and geochemical criteria that included grain size, magnetic susceptibility, Ti concentrations, bulk organic geochemistry and radio-density. From the mean grain-size profiles, the turbidites are identified by their coarse base and a fining-upward sequence of fine sand to very fine silt, in sharp contact with the underlying background mud. Neither the turbidites nor the background sediment contained diatoms. The magnetic susceptibility, the Ti XRF profiles and the C/N ratio exhibit similar patterns as their behaviour is directly related to grain size ( Fig. 2; Supplementary Fig. 3). The spikes in magnetic susceptibility are interpreted as recording an input of para-to ferro-magnetic minerals, which are concentrated in the coarse and dense fraction of the sediment. Likewise, it has been shown that Ti concentrations reflect silt content in sediments 62, 63 . The turbidites are most visible on the 3D CT images of the cores, which highlight their dense bases due to their high contents of fine sand to silt, contrasting with the lower radio-densities of the muddy background.
Turbidites within lake sediments can be triggered by various factors such as floods, spontaneous slope failures, or earthquakes. In the main text, we argue why floods or slope failures are unlikely triggering mechanisms for the Lake Rara turbidites.
In the case of Lake Rara the "synchronicity criterion " 23, 64 cannot be directly applied as the lake is composed of a single basin, which means that the two coring sites are not entirely independent. Our best argument for attributing an earthquake origin to Lake Rara turbidites is a temporal correlation with known historical events 22, 23, 65, 66 . We have been able to relate turbidite T7 (1399-1570 AD; demonstrated that the occurrence of turbidites in Lake Rara is not due to random slope-failure processes but related to seismic activity implying ruptures on both the MFT/MHT and the WNFS. As stated in the main text, the geomorphologic and hydrographic context renders the potential of flood-triggered turbidites unlikely. The hydrographic system of Lake Rara contains 37 small streams flowing along gentle (<30°) and densely forested slopes ( Supplementary Fig. 1). These The low hydrodynamic activity in the catchment of Lake Rara is also suggested by the very low inferred sedimentation rates (0.3-0.5 mm/yr at site RA14-SC05 and 0.2-0.3 mm/yr at site RA14-SC06). These are less than half of the sedimentation rates generally observed in lakes used in paleoseismic research (~1 mm/yr or higher 22,27 ), resulting in higher slope stability and therefore rendering spontaneous slope failures unlikely. Finally, it is worth noting that sedimentation rates in Lake Rara have not significantly changed during the last millennium ( Supplementary Fig. 4a). This observation implies that land-use activities have had little effect on soil erosion and therefore did not affect the rate of sediment remobilisation and the earthquake recording sensitivity of the lake.
Of the eight turbidites recorded in core RA14-SC05 (site A; Fig. 1; Supplementary Figs. 1 and 2), only turbidite T7 (1399-1570 AD), which is attributed to the great 1505 AD earthquake, is also expressed in the sediments of the shallower site 2 (core RA14-SC06; Supplementary Fig. 2). The age of turbidite TA (1285-1606 AD) is statistically indistinguishable from that of turbidite T7 (1399-1570 AD), as demonstrated by the similarity of the Probability Density Functions (PDF) of both turbidites, which both peak around 1500 AD ( Supplementary Fig. 5). This synchronicity strongly suggests that turbidites T7 and TA represent the same event, and were likely generated by the same slope failure, which triggered a turbidity current affecting the entire basin of Lake