Anomalous dissipation is considered one of the defining features of turbulence. Energy dissipates at a rate independent of viscosity — turn the viscosity dial, and the dissipation rate stays the same. This has been observed experimentally, predicted theoretically, and enshrined as fundamental. Kankaria, Mukherjee, Murugan, Rosti, and Ray show that it depends on which triadic interactions you keep.
Triadic interactions are the elementary unit of nonlinearity in the Navier-Stokes equations — three Fourier modes coupling together to transfer energy. In full turbulence, all possible triads participate. The authors systematically remove subsets of triads and measure what happens. When certain classes of interactions are suppressed, anomalous dissipation vanishes. The energy dissipation becomes viscosity-dependent again. Normal behavior returns.
Reader note (2026-11-17): Checked against the full text of arXiv:2603.19180 (Kankaria, Mukherjee, Murugan, Rosti and Ray; v1), which I had earlier held as matching on a lighter read. The authors, the triadic framing and the headline match: the paper says anomalous dissipation is not a generic property of the Navier-Stokes equations and requires the full combinatorial richness of the triadic interactions. Four points differ. First, what is removed: the paper decimates Fourier modes, either fractally or homogeneously, which thins the set of active modes and with it the triads; the essay's certain classes of interactions suggests a targeted removal, whereas the paper describes removing even a statistically homogeneous subset. Second, what happens: the paper reports that the mean dissipation rate decays toward an almost vanishing value as the Reynolds number grows, with the flow approaching Kolmogorov scaling; I searched for the wording viscosity-dependent, depends on viscosity and independent of viscosity and found no hits, so the essay's statement that dissipation becomes viscosity-dependent again and normal behavior returns is the essay's wording, not the paper's. Third, the paper's evidence is direct numerical simulation, and it says the large-scale cascade and a power-law energy spectrum remain comparatively robust while the fine-scale structure is sensitive to the connectivity of the network. Fourth, the paper's conclusion is about anomalous dissipation, multiscaling and multifractality being not robust to reducing the network; the essay's sentence that turbulence is not a robust phenomenon is broader than that.
This means anomalous dissipation is not a generic consequence of the Navier-Stokes equations. It's a specific consequence of the full combinatorial richness of the nonlinear term. Simplify the nonlinearity — remove certain triads — and the defining pathology of turbulence disappears.
The implication is philosophical as much as physical. Anomalous dissipation was treated as an inevitable feature of any sufficiently complex fluid flow. It's actually a fragile feature that requires the complete interaction network. Turbulence is not a robust phenomenon — it's a contingent one, dependent on the full combinatorial structure of its nonlinearity. Remove the right pieces and the monster becomes tame.