Sea of Okhotsk Mw7.8 z = 633 km July 5, 2008 02:12:05 53.9 North 152.9 East
Movies 1-7 use 450 US Array BHZ stations. Movie 8 uses 175 European BHZ stations.
1) linear stacking
2) cube root stacking
3) linear stacking, color scale renormalized each frame
4) cube root stacking, color scale renormalized each frame
5) cube root stacking, color scale renormalized each frame, pP is reference phase
6) linear stacking, color scale renormalized each frame, sP is reference phase (filtered 0.4-2.0 Hz)
7) linear stacking, color scale renormalized each frame, US Array, 640 seconds long
8) linear stacking, color scale renormalized each frame, Europe, 640 seconds long
All movies are shown in mapview at z=633km and data are filtered bw 0.8-2.0 Hz. Until a nearby aftershock occurs for me to use as a reference event, its tough to make good interpretations.
Summary for those of you interested in backprojection applications:
-Most of the energy was released in 2 bursts separated by ~20 seconds
-The second burst is always in the direction of the receiving array, suggesting that it occured at a different depth. However, in both cases (US Array+Europe) it appears to closer to straight north than the direction of the swimming artifact, which go in the direction of the array.
-pP has virtually the same moveout at P, which is why pP and P movies look like time shifted versions of each other. pP does not have significant energy.
-sP has a moveout relative to P of about 4sec/degree at these distances and for this source depth (this is a lot)
-sP only shows up at lower frequencies and its image, which should be at the hypocenter, appears to be displaced by a few hundred km in the direction of the array. Might be useful to help constrain depth?
-Combining backprojection images from different arrays is a good idea for shallow events, but isnt so straight forward for deep events where slabs, discontinuities, depth phases, and most importantly, larger relative travel time differences with depth play a significant role.
-There are lots of things which are clearly not random noise. Forward modelling with large aperture array stacking methods (backprojection/migration/beam forming/slowness stacking, theyre all basically the same) along with some 2 or 3d (or point scattering) synthetic seismograms may be useful.
From the plots below (left is linear yaxis, right is logrithmic) one can see that:
-pP should be showing up ~110 seconds after P (at time=210 sec) for both US Array and Europe, but it cant be seen. Could be nodal?
-Both US Array and Europe image something energetic at 260 sec (160 sec after P). sP trails P by at least 200 sec, so its not that. If you look at Time=260sec in the movies 7 and 8, youll see that the peaks at those times are not co-located, but are in the direction of the receiving arrays. I think its too fortuitous to be noise, especially since US Array data are 45-60 deg, and European data are 70-80 deg away. Its probably an aftershock or a scattered phase at a different depth (which could explain the discrepency in location).
I'll ask Loki to give me an aftershock in this region... this would really help.