SEISMIC ENERGY-RELEASE IN MEXICAN SUBDUCTION ZONE EARTHQUAKES

We estimated radiated seismic energy (E(s)) from 18 shallow, thrust Me xican subduction zone earthquakes (4 X 10(22) less than or equal to M( 0) less than or equal to 1.1 X 10(28) dyne-cm; 11 less than or equal t o H less than or equal to 37 km) using digital accelerograms from the Guerrero Accelerograph Array and neglecting stations with large site e ffects. E(s) is computed by integrating squared velocity spectra, afte r applying geometrical spreading and Q corrections. We discarded epice ntral recordings for the largest Michoacan event. We find that log (E( s)/M(0)) = -4.152 +/- 0.275, which gives a median value of E(s)/M(0) a nd apparent stress (sigma(a)) of 7.1 x 10(-5) and 24 bars, respectivel y. The median E(s)/M(0) value is in accordance with Gutenberg and Rich ter's (G-R) (1956) formula for E(s), in which E(s)/M(0) = 5 x 10(-5) i s implicit. Worldwide E(s)/M(0) data, where E(s) is computed from loca l records, mostly fall between 5 x 10(-5) and 5 x 10(-4) for events wi th M(0) greater than or equal to 10(22) dyne-cm. On the other hand, E( S)/M(0) values, generally, lie between 5 x 10(-6) and 5 x 10(-5), if E (s) is estimated from teleseismic records. Especially anomalous are th e E(s)/M(0) data from Kikuchi and Fukao (1988) for large and great ear thquakes, which fall near 5 x 10(-6). Thus, while E(s) from local data suggests that the G-R relation seldom overestimates seismic energy re lease, the teleseismic data point to the contrary. The cause of this d iscrepancy may lie in the difficulty of resolving incoherent radiation from the fault and inappropriate choice of t in the analysis of tele seismic data.