Distant storms, tsunamis, and earthquakes generate waves on floating ice shelves. Previous studies, however, have disagreed about whether the resulting wave-induced stresses may cause ice shelf rift propagation. Most ice shelf rifts show long periods of dormancy suggesting that they have low background stress concentrations and may therefore be susceptible to wave-induced stresses. Here, I quantify wave-induced stresses on the Ross Ice Shelf Nascent Rift and the Amery Ice Shelf Loose Tooth T2 Rift using passive seismology. I then relate these stresses to a fracture mechanical model of rift propagation that accounts for rift cohesive strength due to refrozen melange, ice inertia, and spatial heterogeneity in fracture toughness due to the presence of high toughness suture zones. I infer wave-induced stresses using the wave impedance tensor, a rank three tensor that relates seismically observable particle velocities to components of the stress tensor. I find that wave-induced stresses are an order of magnitude larger on the Ross Ice Shelf as compared to the Amery Ice Shelf. In the absence of additional rift strength, my model predicts that the Nascent Rift should have experienced extensive rift propagation. The observation that no such propagation occurred during this time therefore suggests that the Nascent Rift experiences cohesive strengthening from either refrozen melange or rift tip processes zone dynamics. This study illustrates one way in which passive seismology may illuminate glacier calving physics.