Earthquake focal mechanisms, moment tensors, and the consistency of seismic activity near plate boundaries

Catalogs of moment tensors for more than 8000 earthquakes provide a more objective and complete description of the earthquake source than do focal mechanisms derived from first motions, and therefore moment‐tensors provide a valuable resource for tectonic analysis. We here present background information about the properties of moment tensors and examples of moment tensor analysis. We also introduce a new statistic, the seismic consistency C s , which measures the similarity of earthquakes within a group. C s is 1.0 if earthquakes are all alike and 0.0 if they cancel one another. Triangle diagrams provide a practical method for defining the fraction of normal, strike‐slip, and thrust fault components for an earthquake and are a new graphical method for displaying source properties of groups of earthquakes. We apply these methods to the Harvard centroid moment tensor catalog to study the characteristics of shallow earthquakes (<50 km depth) within 200 km of typical ridge‐transform and subduction zone plate boundaries. In this way, we have reached four major conclusions. First, even along relatively simple plate boundaries, there is considerable variation in the type and orientation of earthquake mechanisms. Second, along individual plate boundaries, groups of thrust, normal, or strike‐slip earthquakes generally have C s equal to 0.8 or higher. Thus for many types of tectonic analyses it is unnecessary to add moment tensors to study seismic deformation; rather, it is sufficient to add scalar earthquake moments of the individual events. Third, moment tensors for some individual earthquakes are quite different from those produced by slip along a planar fault. However, summing up moment tensors shows that the deformation pattern produced by groups of earthquakes is generally more like fault slip than the pattern of typical earthquakes within the group. Fourth, by dividing the sum of moments by the velocity of plate motion and the length of the boundary, we calculate R mom , the efficiency of seismic moment produced along individual plate boundaries. For the 12.75 years of data available, normal fault earthquakes along spreading ridges produce moment less efficiently than strike‐slip earthquakes along transforms. These in turn produce moment less efficiently than thrust earthquakes along subduction zones. For ridge‐transform earthquakes, boundaries with fast plate velocities produce seismic moment less efficiently than do boundaries with slow velocities.