This study presents a novel spectral inversion method, specifically designed to account for inhomogeneous attenuation structures; it thus can lead to bias‐reduced estimation of both source and site effects. The generalized inversion technique (GIT) often assumes uniform attenuation structures, which can be invalid in regions with evident structural variations. To overcome these limitations, the proposed method divided the study region into equally spaced cells. The Q‐value was evaluated as a cell‐specific value. By formulating a constrained inverse problem that incorporated both spatial and frequency smoothness, high resolution and stability were achieved in Q‐value estimations. The proposed method was validated using synthetic data tests, confirming that the assumed true values could be accurately reproduced. Then, the proposed method was applied to the aftershock region of the 2008 Iwate–Miyagi Nairiku earthquake. Results revealed low Q‐values in the western region, associated with active volcanoes, whereas higher Q‐values in the eastern region were consistent with previous tomographic studies. Notably, site amplification varied up to three times compared with that estimated using GIT, and the estimated stress drop was approximately half on average. The results generally agreed with those obtained using the block inversion technique (BIT), another method that considers inhomogeneous Q‐values, albeit some discrepancies were observed in certain cells. These differences likely stemmed from variations in the assumed constraint conditions. These findings indicated that the proposed method could estimate reasonable solutions while accounting for inhomogeneous Q‐values in a finer, equally spaced mesh. These findings also highlighted the importance of considering inhomogeneous attenuation structures in spectral inversion analyses. The proposed method has the potential to enhance strong ground motion predictions and contribute to improved seismic hazard assessments by providing attenuation estimations and bias‐reduced site and source estimation.

You do not have access to this content, please speak to your institutional administrator if you feel you should have access.