Intermittent Turbulence, Fast Flavor Conversion, and Observable Supernova Probes

Fast flavor conversion (FFC) in core-collapse supernovae is usually analyzed in homogeneous backgrounds or with smooth stochastic turbulence closures. We construct an exact linear benchmark in which the matter-noise memory kernel is instead generated by a finite She--Leveque log-Poisson cascade. Projecting the marginal FFC channel onto this kernel gives a causal Volterra equation whose non-Markovian memory closes into a finite local system. The resulting Laplace-space resolvent is rational, with one pole pair for each cascade level, so the dispersion relation, characteristic polynomial, and time-domain solution can be checked analytically. We then connect this benchmark to the realization-level toy model and gain-region heating proxy used in the supplementary derivation. For the updated intermittent choice $\delta\rho/\langle\rho\rangle=0.4$, $\bar\lambda/\mu=1$, and hence $\kappa_0=0.16$, the representative $N=2$, $r=2$ cascade gives $\sigma_{\rm int}^2=1.124$ and an intermittent conversion fraction $1-P_{\rm base}\simeq0.455$. The older weaker normalization $\kappa_0=0.05$ gives $1-P_{\rm base}\simeq0.324$. The corresponding Mori-like heating ratios are $Q_{\rm int}/Q_{\rm hom}=1.060$ and $1.041$, whereas the Wang/Fornax-like ratios are $0.855$ and $0.899$. Thus intermittency mainly controls the conversion fraction, while the neutrino spectral hierarchy controls the sign of the heating correction.