The arXiv preprint (v1, May 2026) from Barbani, Gaspari and collaborators uses GPU-accelerated 3D hydrodynamical runs of a stratified galaxy-group halo to show that chaotic cold accretion (CCA) remains super-Bondi across two subsonic turbulence regimes, yet the meso-scale (0.1–1 kpc) imprints distinct kinematic and variability signatures. Strong stirring sustains fragmented, multiphase filaments that boost inflow coherence, while weaker turbulence produces smoother cascades; both nevertheless deliver statistically indistinguishable central accretion rates. This decouples macro-scale weather from micro-scale feeding, a result that existing cosmological simulations—which rarely resolve below ~kpc—systematically miss. The power spectra of accretion rate follow broken power laws (pink noise flattening to red-noise tails), matching parsec-scale collisional damping observed in X-ray binaries and AGN light curves. Complementary diagnostics—the C-ratio near unity flagging soft X-ray condensation sites and k-plots revealing broader, overlapping velocity fields in stormy runs—highlight how meso-scale kinematics bridge halo rain to SMBH growth. Prior CCA studies (Gaspari et al. 2013, MNRAS; 2017, MNRAS) established the condensation mechanism but lacked the dynamic-range resolution needed to quantify this scale-dependent kinematic imprint. The present work therefore supplies the missing link for sub-grid prescriptions in large-volume simulations, implying that black-hole maintenance-mode feedback efficiency may be more universal than halo weather suggests. Limitations include idealised driven turbulence (no self-consistent cosmic rays or magnetic fields) and the absence of full radiative transfer; results remain simulation-based until confronted with XRISM or Athena X-IFU kinematics.