Earthquake recurrence in NW and central Himalaya Hilmar Bungum a,⇑ , Conrad D. Lindholm a , Ambrish K. Mahajan b a NORSAR, P.O Box 53, N-2027 Kjeller, Norway b School of Earth and Environmental Sciences, Central University Himachal, Kangra H.P. 176206, Himachal Pradesh, India article info Article history: Received 3 October 2016 Received in revised form 24 January 2017 Accepted 26 January 2017 Available online 4 February 2017 Keywords: Himalaya Plate motions Seismicity Compressional tectonics Strain rates Seismic moment rates Earthquake deficit abstract In this study we compare Himalayan seismic moment release estimates derived from strain rate obser- vations with those derived from large historical earthquakes, and to this end we use a reassessed cata- logue of historical earthquakes from western and central Himalaya since the beginning of the 16th century. We have computed seismic moment rates within six contiguous segments along the Himalayan arc and compared these, using Kostrov’s formula, with moment rates computed from recent global strain rate estimates and regional studies. While the ratios between strain-based moment-rate estimates and those inferred from observed seismicity vary significantly between the segments, we find on the average consistently larger strain-based values by about a factor of two, based on seismicity from the last 515 years. The moment-rate ratio is, however, significantly reduced when shorter catalogues are used, to 1.28 for the last 215 years and to 1.05 for the last 115 years, which is an almost perfect match. The possible inclusion of afterslip in the model would further improve the 515-year match. This is indi- cating that a significant part of the difference, possibly most of it, is likely to be caused by incompleteness of the longer earthquake catalogue, possibly combined with underestimated magnitudes. The difference between geodetic and seismic estimates for the more complete part of the catalogue is smaller than pre- viously reported along the western Himalayan frontal thrust. In fact, the only region where a significant moment-rate difference is found in our study is in SE Himachal Pradesh. In terms of seismic hazard it is found that the moment rate reduction of about a factor of two, when going from 115 to 515 years, leads to a reduction in the 475-year PGA of about 26%. It is also found that using 50 years of USGS seismicity data between 1963 and 2012 leads to a 40% lower hazard as compared to using moment release for the last 115 years. Ó 2017 Elsevier Ltd. All rights reserved. 1. Introduction With a total length of about 2400 km, the highest mountains, the deepest lithospheric roots and the highest uplift rates within a continent, the Himalayan mountain chain is an undisputed end member on the scale of seismotectonic significance. Given the large-scale motions it may appear surprising that several earlier studies have predicted an earthquake deficit in this region, with a GPS-based moment release rate exceeding the observed rate. Such studies, albeit based on quite different methods and approaches and also with quite different deficit estimates, include Bilham and Ambraseys (2005), Meade (2010), Ader et al. (2012), Schiffman et al. (2013) and Stevens and Avouac (2015, 2016). Ader et al. (2012) find, using a subduction zone inversion tech- nique, that the Main Himalayan Thrust (MHT; Fig. 1) appears to be fully locked from its foothill surface expression to beneath the front about 100 km to the north, and that the moment deficit accu- mulates at a rate of 6.6  10 19 Nm/yr, which corresponds to a seis- mic moment deficit of about a factor of five. Stevens and Avouac (2015) also find that the fault is fully locked along its complete length and over about 100 km width (see also Li et al., 2016 and Jouanne et al., 2017), and that the moment deficit builds up at a rate of 15.1 ± 1  10 19 Nm/yr for the entire length, which is more than twice the rate of Ader et al. (2012). Based on a different approach and one not much different from what is used here, Bilham and Ambraseys (2005) report a deficit of about a factor of 3 to 4. All of these studies are predating the 25 April 2015 M W 7.8 Gorkha earthquake and the connected 12 May 2015 M W 7.3 earthquake, which ruptured locked portions of the MHT (e.g., Avouac et al., 2015). Here we start with a re-evaluation of the largest historical earthquakes along the western and central Himalayan arc, and pro- pose a segmentation based on combined tectonic and seismologi- cal criteria (Fig. 2). The observed seismic moment rates are first compared with estimates directly from the GPS velocity vectors, http://dx.doi.org/10.1016/j.jseaes.2017.01.034 1367-9120/Ó 2017 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. E-mail address: hilmar.bungum@norsar.no (H. Bungum). Journal of Asian Earth Sciences 138 (2017) 25–37 Contents lists available at ScienceDirect Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes