Planetary embryos migrate through their birth disk by exchanging angular momentum with the surrounding gas. In calm disks, thermal effects dominate — the embryo heats its surroundings asymmetrically, creating density lobes that torque the orbit inward or outward in predictable ways. Chametla, Moranchel-Basurto, and Sánchez-Salcedo show that magnetic turbulence completely destroys these thermal lobes.
The destruction is total, not partial. In turbulent regions, the density structures that drive thermal migration vanish. What remains is stochastic torque — random kicks from turbulent fluctuations that push the embryo unpredictably in either direction. Migration becomes a random walk rather than a directed process.
Note (2026-11-15; read against the paper's full text): the essay says turbulence completely destroys the thermal lobes, that migration becomes a random walk in turbulent regions, and that real disks are magnetized and turbulent. The paper's result is from three-dimensional magnetohydrodynamic simulations of two cases, a 0.33 Mars-mass embryo and a 1 Earth-mass core, in an inviscid disk with an initially toroidal magnetic field; it finds that even with a weak field and irrespective of the luminosity, turbulence (which takes 1.5 to 3 orbital periods to develop) completely disrupts the thermal lobes, and the torque becomes strongly oscillatory. It says this suggests stochastic migration for masses between about 0.03 and 1 Earth masses, as expected in turbulent disks, and that thermal torques become inefficient in turbulent regions. I found no statement that the migration is a random walk, or about final orbital architectures; the mass range and the simulated set-up are the scope of the result.
This matters because most migration models assume thermal effects operate everywhere in the disk. The idealized simulations that predict orderly planetary migration — the ones used to explain why certain orbital architectures form — implicitly assume a level of disk calm that may not exist. Real disks are magnetized. Real disks are turbulent. In those regions, the orderly migration that populates theoretical models simply doesn't happen.
The implication is unsettling for planet formation theory: if embryos migrate randomly in turbulent zones, the final orbital architecture depends on the turbulence history of the disk, not just the initial mass distribution. The outcome becomes contingent on noise.