Dense bacterial suspensions generate turbulence at zero Reynolds number — flows that look chaotic, form vortices, and display power-law energy spectra despite having no inertia whatsoever. The standard explanation treats the “activity” — the self-propulsion strength of the bacteria — as a uniform background parameter. Turn it up and chaos appears. Turn it down and it stops. The activity is the dial; the turbulence is the response.
Sahoo, Perlekar, and Pandit ask a deceptively simple question: what if the dial isn't uniform?
In real bacterial suspensions, activity varies in space and time. Nutrients aren't evenly distributed. Bacteria cluster, deplete resources locally, die in patches. The activity field is heterogeneous. More importantly, it gets advected — carried along by the very flow it generates. The bacteria swim, their swimming creates flow, the flow rearranges the bacteria, and the rearrangement changes the activity field. The parameter that drives the chaos is itself reshaped by the chaos.
What happens is confinement. The turbulence doesn't fill the entire fluid. It retreats to regions where activity is high, and those regions develop sharp boundaries — activity fronts that separate turbulent zones from quiescent ones. The fronts aren't static; they morph, merge, and split as the flow advects the activity field. The interfacial morphology is complex and dynamic, but the crucial observation is that the turbulence is localized. The chaos has edges.
The deeper consequence: universality in active turbulence is local and transient, not global and steady-state. The energy spectra that look universal in a homogeneous system become position-dependent in a heterogeneous one. Different regions of the same fluid can exhibit different spectral regimes simultaneously, with transitions occurring at the activity fronts. The system doesn't have a single phase — it has coexisting phases separated by moving boundaries.
This inversion is what makes the result interesting. In classical turbulence, the fluid properties are fixed and the flow is variable. In active turbulence with advected activity, the flow properties themselves are variable — determined by a field that the flow rearranges. The system isn't just nonlinear. It's self-modifying: the output reshapes the input at every step. The appropriate question isn't “what spectrum does this system produce?” but “where, and for how long?”