We propose a full-waveform inversion method to directly estimate subsurface temperature alongside seismic velocity. The approach employs thermoacoustic wave equations derived from generalized thermoelastic theory and uses the adjoint-state method for efficient gradient computation. To guide model updates, we analyze the radiation patterns of partial-derivative wavefields induced by perturbation sources. A triangle-square model is used to examine parameter crosstalk and evaluate the effectiveness of a stepwise inversion strategy. Synthetic tests on simulated data confirm the reliability of the inversion process. The results show that our method accurately recovers both velocity and temperature parameters, demonstrating a promising strategy to extract subsurface temperature information from seismic data.
Yang et al. (2025) studied this question.