Core Collapse Supernova Gravitational Wave Sourcing and Characterization based on Three-Dimensional Models

We present for the first time an analysis of high-frequency gravitational wave (GW) emission from proto-neutron stars (PNS) in core collapse supernovae (CCSN) that combines spatial decomposition and modal decomposition to both source and characterize the emission using three-dimensional CCSN simulations. We analyze simulations initiated from 15 and 25 solar mass progenitors with Solar- and zero-metallicity respectively. We decompose the GW strains into five spatial regions and find that strains are initially largest in the PNS surface layers from accretion and later largest from the Ledoux convective and convective overshoot regions of the PNS. We compute the fractional GW luminosity as a function of enclosed radius and observe that most of the luminosity moves from the PNS surface to deep within the PNS at later times. Using a self-consistent perturbative analysis, we investigate the evolution of the oscillation modes of the PNS. We find that the frequency of the evolving high-frequency component of the GW signal is well matched to the eigenfrequency evolution of the ${}^2g_2$-, ${}^2g_1$-, and ${}^2f$-modes over time. We show that the ${}^2g$-modes emit most of their power in GW initially from the PNS surface region, but within a few 100 ms after bounce, it is the convective overshoot region of the PNS that emits the most GW power for the ${}^2g_1$-mode. Eventually, the ${}^2f$-mode is the dominant mode producing GWs, and they are emitted primarily from the convective overshoot region. Thus, we show that, while the GW emission is global, it is possible to source the dominant contributions to it. We find that the source of the high-frequency GW emission from the PNS in CCSN is more complex than assessed by other methods, as well as time dependent, first emitted by ${}^2g$-modes driven by accretion onto the PNS and later emitted by the ${}^2f$-mode driven by sustained Ledoux convection.