DiffraPy: An open-source Python software for seismic diffraction imaging

Abstract

Seismic diffraction events, often overlooked as noise in conventional seismic imaging methods, contain valuable high-resolution information about subsurface structures and heterogeneities. Furthermore, the diffractive component of the wavefield is masked by the stronger reflection events. Although the literature presents innovative methods for separating diffracted events from the total wavefield, there are currently limited options of open-source alternatives in seismic diffraction processing and imaging. In this study, we introduce DiffraPy, an open-source Python package specifically designed for performing both conventional and diffraction-oriented migrations. The main workflow involves implementing an anti-stationary phase filter in the Kirchhoff migration for enhancing energy outside the Fresnel zone. Additionally, we propose using the semblance matrix obtained from the dip field calculation as an additional weight in conventional migration. The code provides functionality for forward modeling seismic data, allowing for new testing and experimentation. We demonstrate the general workflow of the code using a toy model and evaluate its capabilities on a synthetic salt model. In both cases, the program successfully suppressed the majority of reflected energy in the diffraction-oriented image. The final image highlights the responses of small lenses, discontinuities, and faults, providing valuable insights into the subsurface for interpretation. Despite the method sensitivity to variations in the correct velocity model, we demonstrate that our code is capable of imaging diffractors even when utilizing incorrect migration velocities of up to 5%. Our wavenumber analysis reveals the diffraction image preservation of smaller wavelengths, indicating higher resolution compared to conventional products. Future implementations may include extension to 3D datasets and improve the computational efficiency of our code. We expect our code to offer the geosciences community the first freely accessible and well-documented Python alternative for testing, reproducing, and exploring diffraction imaging examples.