The Earth is covered by a rigid shell, called lithosphere. What makes the lithosphere rigid, and where its strength resides, remains a subject of active research. Food analogies play a big role in the discussion of the lithospheric strength , with models like "jelly sandwich" and "crème brûlée" being accepted research terms. While the former argues for a weak, ductily deforming lower crust between relatively stronger upper crust and the upper mantle, the latter posits a uniformly strong crust above a weaker upper mantle. One of the more compelling pieces of evidence that investigators marshal in promoting any particular model is the depth to which seismic activity extends inside the lithosphere, the argument being that earthquakes are necessarily caused by brittle failure and hence are incompatible with ductile deformation.

Figure 1 Locations (green stars) and focal mechanisms (beach-ball symbols) of earthquakes in northwestern Tibet, from this study (#1-#5) and from previous publications (#6-#8 Chen and Yang [2004], #9 Fan and Ni [1989]). White circles are other well-located earthquakes [Huang et al., 2009]. The size of beach-ball symbols is proportional to earthquake size, the largest being Mw=6.0. Red lines show two major strike-slip faults in northwestern Tibet. Grey lines locate the profiles of hypocenter projection in Figure 2.

Figure 2. Hypocenter projection (green crosses) for earthquakes shown in Figure 1. Zero lateral distance corresponds to the trace of the Altyn-Tagh fault. All three projections (A, B, C) are stacked onto one profile to show the relative positions of earthquakes, indicated by green crosses, with respect to the fault trace.  The red dash-lines show estimates of crustal thickness from Wittlinger et al. , EPSL 2004.

Using data collected on the northern edge of the Tibetan plateau, we identified earthquakes at depths between 40 and 60 km, in the lowermost part of the crust. Other studies showed that earthquakes in this region also take place deeper, in the upper mantle. Notably, patterns of seismic wave radiation from earthquakes we studied imply extension and shearing at the source. Deep earthquakes routinely observed at sites of lithospheric plate convergence typically display evidence of compression. Thus we associate earthquakes in northwestern Tibet with brittle deformation within the body of the Eurasia lithosphere, both at the whole crust and in the uppermost part of the mantle.

We investigated the region of an ongoing collision between the Indian and Eurasian tectonic plates that results in widespread deformation of the continental lithosphere. Over the past decade numerous regional studies were conducted between the Himalaya and the Tien Shan mountains, each illuminating a small part of the area. We combined the data from a number of portable and permanent networks to construct an ~1800 km long profile of the lithospheric properties across three very different tectonic domains: the Tibetan plateau, the Tarim basin, and the Tien Shan mountains.  We used records of distant earthquakes to construct receiver function gathers for each station. The uniformity of processing ensures that our results are comparable along the transect. We examined receiver function gathers at each site, and ranked their quality on the basis of the number of records, noise levels, and directional stability of the wavefield. We selected 32 sites with high-quality data. For these we constructed average receiver function traces using data in 60°-85° distance range, and used them as a guide to the lithospheric layering beneath the region. On most receiver functions we have constructed the most prominent feature is a positive phase likely associated with the crust-mantle transition. The timing of this phase varies significantly over the length of the profile. Beneath the Tibetan plateau delay times ~7-8 s are seen close to the Himalayas, and nearly 10 s delays are found further north. Delays of 6 to 8 s are seen beneath sites in the Tarim basin and the Tien Shan mountains, and a nearly 10 s delay are seen at the border between them. In addition to the pulse associated with the crust-mantle transition we see other locally-consistent features, for example a negative phase with delay values between 3 and 5 s beneath much of the Tibetan plateau.

Crustal structure between the Himalayas and the Kazakh Shield

Deep earthquakes beneath NW Tibet suggest the crust is brittle throughout

The figure above shows a summary of results to date. A map of the transect is shown in the lower panel, with crosses placed every 50 km. Circles with names show locations of seismic stations, red lines show major faults. In the upper panel we show topography; second panel shows crustal thickness obtained from literature (black line) compared with estimates of crustal thickness based on the timing of the Ps phase in the receiver functions (green boxes). The third panel shows the receiver function transect. Blue (red) pulses correspond to positive (negative) P-SV conversions, and signify an increase (decrease) of seismic impedance with depth. We associate the positive pulses at 6-10 s delay with the crust-mantle converted phase (Ps). The timing of this phase at individual beams is plotted as a green line.