Synthetic 3D Seismic Volumes and Fault Labels for Strike-Slip Flower Structures in Carbonate Reservoirs

This dataset contains synthetic three-dimensional seismic volumes and corresponding fault-label volumes generated for deep-learning-based fault detection in carbonate strike-slip fault systems with flower structures. The data were generated following the workflow described in the manuscript “Synthetic Seismic Modeling of Strike-Slip Fault Systems with Flower Structures in Carbonate Reservoirs”. The workflow integrates fractal-based stratigraphic modeling, strike-slip fault construction with flower-like geometries, carbonate-specific velocity assignment, wave-equation-based seismic forward modeling, least-squares reverse time migration (LSRTM), and dataset extraction. The generated models include structurally variable fault segments, including translational, transpressional, and transtensional styles, to represent complex carbonate fault systems. Each data volume is stored in NumPy `.npy` format and has a shape of 129 × 129 × 179. The files can be read directly in Python using: `numpy.load("filename.npy")` The dataset is intended for training, validation, and testing of seismic fault detection methods, especially data-driven approaches targeting complex carbonate reservoirs and strike-slip fault systems. The seismic volumes provide synthetic migrated seismic responses, and the associated label volumes provide the corresponding fault annotations.