A peak can hide
between samples.
Shrink the profile. Move the mesh or refine it. A taller true peak can still escape a coarse scan.
Reduced motion is enabled. Choose still frames with the controls.
What this microscope reveals
Invented divergence-free Gaussian swirl U(s)=(-s_y,s_x,0)e^(−|s|²). Its kinematic scaling is uτ(x)=τ^(−α)U(x/τ^β), with τ≥0.125. It has not been shown to satisfy a fluid momentum equation.
energy ratio=τ^(3β−2α)
The strip shows axial curl along x at y=z=0, normalized to its centre height. The eyepiece also rescales x by ℓ, so its shape stays fixed. Dials carry absolute normalized readings. Units are illustrative.
N intervals cover x∈[−3,3]. Offset shifts the grid, never the field. Dots are direct evaluations; the sampled positive peak is compared with the known centre. Negative wings are retained. This is one sampling error, not a certified PDE error bound.
Energy may fall while the centre grows. That does not prove fluid blowup, supply a smooth force or validate the episode’s research claims. There is no Euler/NS solver, neural network or interval-arithmetic certificate here. Playback stops at a finite cutoff.
Wang etal.: self-similar numerical searches ↗
Fefferman: equations and bounded energy ↗