Metastable olivine wedge beneath northeast China and its applications

When the Pacific slab subducted into the mantle transition zone, there might exist a metastable olivine wedge (MOW) inside the slab due to the phase transition. Lots of researchers have adopted such various methods to detect the characteristics of this MOW as the forward modeling of travel times, shear wave amplitude patterns, teleseismic P wave coda, receiver function imaging, thermodynamic simulation and so on. Almost all results could be more or less affected by the source, the receiver and/or the velocity model passed through by the seismic rays. In this study, we have used 21 deep earthquakes, greater than 400 km and locating beneath northeast China, to study the velocity within the MOW. For more precisions, we have done further modifications in two ways based on our previous studies. (1) Double-difference location method is used to relocate all events with an error of 1-2 km with the data recorded by stations both at northeast China and at Japan. All relocated events locate in a zone about 30 km away from the upper boundary of Pacific slab. (2) Double residual travel times, generated by an event-pair at a common station at only Japan, are used to constrain the velocity anomaly rather than the residuals themselves. As a result, we have found that an ultra-lower velocity zone (ULVZ), averagely -7% relative to the iasp91 model, exists within the subducted Pacific slab around the deep earthquakes, which might be represented as the metastable olivine wedge. Because of the lower-velocity corresponding to the lower-density, the MOW would provide upward buoyancy forces which might prevent the slab from free subduction into the mantle transition zone. This feed-back mechanism of MOW to the slab is called ';parachute-effect', which is characterized by other researchers. In addition, the existence of the ULVZ or the MOW in the slab may supply a possible mechanism for triggering deep earthquakes, called ';phase transformation faulting', which was already proposed few decades ago but now is proved further by our result. Therefore, our study is not only instructive for realizing the dynamic process of slab subduction, but also helpful for analyzing the mechanism of deep earthquakes.