To understand the energy release process that operates at the end of the main shock rupture and start of the aftershock activity, we propose an inversion method that uses continuous high-frequency seismogram envelopes of the main shock and early aftershocks (i.e., events that occur at short times after the main shock). In our approach, the aftershock sequence is regarded as a continuous energy release process, rather than a discrete time series of events. To correct for the contribution of coda wave energy excited by multiple scattering, we use the theoretical envelope synthesized on the basis of the radiative transfer theory as a Green's function. The site amplification factors are corrected considering the conservation of energy flux and using the coda normalization method. The inverted temporal energy release rate for the 2008 M W 6.9 Iwate-Miyagi Nairiku earthquake, Japan, decays following t
À1.1, at the lapse time t of 40-900 s after the main shock origin time. This exponent of the decay rate is similar to the p value of the modified Omori law. The amount of estimated energy release is consistent with that calculated from the magnitude listed in the aftershock catalog. Although the uncertainty is large, the location of large energy release at the lapse times of 40-900 s approximately overlaps to that of the aftershocks, which surrounds the large energy release area during the main shock faulting. The maxima of the energy release rate normalized by the average decay rate distributes following a power law, similar to the Gutenberg-Richter law.