GLOBAL PATTERNS OF RADIATED SEISMIC ENERGY AND APPARENT STRESS

Radiated energies from shallow earthquakes with magnitudes greater than or equal to 5.8 that occurred between 1986 and 1991 are used to examine global patterns of energy release and apparent stress. In contrast to traditional methods which have relied upon empirical formulas, these energies are computed through direct spectral analysis of broadband seismic waveforms. Velocity-squared spectra of body waves are integrated after they have been corrected for effects arising from depth phases, frequency-dependent attenuation, and focal mechanism. The least squares regression fit of energy E(S) to surface wave magnitude M(S) for a global set of 397 earthquakes yields log E(S) = 4.4 + 1.5M(S), which implies that the Gutenberg-Richter relationship overestimates the energies of earthquakes. The least squares fit between E(S) and seismic moment M(o) is given by the relationship E(S) = 1.6 X 10(-5)M(o), which yields 0.47 MPa as the average global value of apparent stress. However, the regression lines of both E(S)-M(S) and E(S)-M(o) yield poor empirical predictors for the actual energy radiated by any given earthquake; the scatter of data is more than an order of magnitude about each of the regression lines. On the other hand, global variations between E(S) and M(o), while large, are not random. When subsets of E(S)-M(o) are plotted as a function of seismic region and faulting type, the scatter in data is substantially reduced. The E(S)-M(o) fits for many seismic regions and tectonic environments are very distinctive, and a characteristic apparent stress tau(c) can be derived. The lowest apparent stresses (<1.5 MPa) are associated with thrust earthquakes at subduction zones. The highest apparent stresses (>3.0 MPa) are associated with strike-slip earthquakes that occur at oceanic ridge-ridge transforms and in intraplate environments seaward of island arcs. Intermediate values of apparent stress (1.5 < tau(a) < 3.0 MPa) are associated with strike-slip earthquakes at incipient or transitional plate boundaries. In general the dominant mode of failure for a tectonic environment is associated with the faulting type that has the lowest apparent stress. An energy magnitude M(E) can complement moment magnitude M(W) in describing the size of an earthquake. M(E), being derived from velocity power spectra, is a measure of seismic potential for damage. M(W), being derived from the low-frequency asymptote of displacement spectra, is more physically related to the final static displacement of an earthquake. When earthquake size is ranked by moment, a list of the largest events is dominated by earthquakes with thrust mechanisms. When earthquake size is ranked by energy, the list of the largest events is dominated by strike-slip earthquakes.