AbstractSeismic energy is a physical concept related to broadband information on the source radiation; this is different from seismic moment, which is related to low-frequency information. Improved estimates of seismic moment and radiated energy can improve the scientific framework for interpreting observed seismicity in terms of physical source processes. This study evaluates the utility of regional Lg waves, the dominant feature of seismograms for regional earthquakes, for the estimation of radiated seismic energy. The rms Lg, which is defined as the root mean square of Lg-wave ground velocity within a certain group velocity range, is proportional to the square root of Lg-wave-radiated energy flux. We find that rms Lg is a good quantity for estimating the energy release of earthquakes because a large fraction of the total seismic energy is radiated as Lg waves. We scale the network-averaged rms Lg value to a reference distance of 100 km for a set of 52 earthquakes in the northeastern United States, using an empirically-derived relation between rms Lg and epicentral distance. We apply the source parameters, estimated for these earthquakes, to Boatwright's deceleration source model and calculate the total seismic energy expected to have been radiated by each event, assuming it can be modeled by this source. We then obtain for earthquakes in the northeastern United States the scaling of the total radiated seismic energy of earthquakes with network-averaged rms Lg values. From synthetic seismograms, it is found that different velocity structures have an effect on the efficiency of Lg-wave radiation. Using measurements of rms Lg made from synthetics, we find that the vertical strike-slip source tends to radiate less Lg-wave energy than other source mechanisms. |