Knowledge of the spatial distribution of subsurface fractures is critical for improving the production of geothermal and oil/gas. The double-beam (DB) method using acoustic waves is capable of characterizing the irregularly distributed fractures, including multiple coexisting fracture sets. We have extended the DB method to elastic waves, in particular for P-to-S waves scattered by fractures, using 3C seismic data. This elastic double-beam (EDB) method can add additional information to cross-validate the DB images of P-P waves and increase our confidence in the final fracture characterization results. To test the EDB method, we use a 3D layered reservoir model with multiple nonorthogonal coexisting fracture sets, which captures a wide range of geologic scenarios. Based on this model, we model 3C data using an elastic full-wave finite-difference method. For each subsurface target, the EDB method outputs DB images of P-P, P-Sin, and P-Santi waves, where Sin and Santi represent the in-plane (polarized perpendicular to the fracture plane strike) and antiplane (polarized parallel to the fracture plane strike) scattered vector S-waves, respectively. Numerical results indicate that EDB is a reliable tool to detect fracture distribution using the self-verification feature. In other words, P-P, P-Sin, and P-Santi should all give the same fracture parameters for truly existing fractures. Using EDB, the existence and orientation of fractures are more reliably estimated than fracture compliance for irregular fracture sets. EDB is not sensitive to random noise in data up to high noise levels. EDB can work accurately when velocity models have mild deviations from the true models. The EDB amplitude is large where fractures are dense or the compliance value is large, or both; these can be important in the interpretation of the fluid transport properties of a reservoir.