Yes, interesting. I've never heard of spatial smoothing of a stress field and don't know why that would occur. I imagine it is explained in their paper. From what I have heard from seismologist, you do need deviatoric/differential stress (I get them mixed up) to propagate an earthquake rupture, so it makes sense that if this stress drops to be low along part of a fault a rupture may not be able to pass. A seismologist and I discussed something similar for the Compton-San Pedro Escarpment fault for the proposal we finished yesterday...related to whether creep north of downtown Los Angeles translates to loading at the bottom of a large locked part of a megathrust system that extends offshore. We have some preliminary evidence that the answer is "yes"
Chris