Fisher-Rao Geometry of Gluon Saturation Fronts

Abstract

How nonlinear gluon saturation erases the memory of initial conditions is a central question in high-energy QCD. We develop the Fisher-Rao information geometry of saturation fronts in the Balitsky-Kovchegov (BK) equation and its Fisher-KPP diffusive limit. Using the normalized coordinate-space front density $p = \partial N/\partial\ln r^2$, we isolate the transverse metric $g^\perp_{θθ}$ that quantifies shape distinguishability once longitudinal propagation along the flow is factored out. Nonlinear saturation arrests diffusive broadening and freezes the Fisher speed. Transverse deformations exhibit a universal $t^{-4}$ relaxation in the Fisher-KPP asymptotic analysis, with full BK evolution providing numerical evidence for the corresponding $τ^{-4}$ behavior. We trace this steep power law to a relaxation mechanism: deforming the initial profile is equivalent, to leading order, to a spatial translation combined with an effective time delay of the universal relaxation flow. Extracting the translation and the delay from neighboring runs alone predicts the ratio of $g^\perp_{θθ}$ to its shift-control value within $1\%$ for Fisher-KPP and for BK up to $τ= 15$. For Fisher-KPP, the asymptotic front expansion yields the $t^{-4}$ law conditionally on stated analytic hypotheses, with a universal amplitude $C_*$ that agrees with numerical extrapolations to sub-percent accuracy. For BK, the asymptotic translation information exceeds its diffusive approximation by a factor of 3.5, which we attribute to the non-diffusive Levin-Tuchin saturation region. These results provide an information-geometric characterization of the approach to geometric scaling, consistent with the persistent sensitivity of pre-asymptotic HERA fits to initial shape parameters, and motivate systematic studies of parameter sensitivity in EIC kinematics.