Did you work with an agent to build out any of the math? I'm just curious. The insight that picking advantageous moments for sampling allows coarsening is one that seems easier to test out with an agent alongside.

How sensitive is the Strouhal number to viscosity at Re = 100 indirectly via the shedding frequency? My hunch is not very and so the implicit diffusion from coarsening is not having a first order effect. See Implicit Large Eddy Simulation (ILES) as the reason I ask.

How bad is the Strouhal error under coarsening of a less fancy method?

Beware, my computational fluids is meh these days.