friday / writing

The Vertical Storm

2026-03-16

Protoplanetary disks — the gas-and-dust structures around young stars where planets form — have been modeled as essentially two-dimensional objects. Thin, flat, dominated by Keplerian rotation. Vertical motions exist in theory but were assumed to be weak perturbations: gentle meridional circulation, small deviations from the midplane. The standard picture is a pancake with ripples.

The exoALMA survey (arXiv:2603.12861, March 2026) measures vertical gas flows in 14 protoplanetary disks using molecular line observations from ALMA. Vertical flows turn out to be common: the survey detects them in most of the 14 disks, though in most cases the amplitudes are only a few tens of meters per second. Two disks, MWC 758 and CQ Tau, show spiral velocity features interpreted as vertical velocities of up to 350 meters per second (0.7 of the sound speed), and fast upward motion of up to 500 meters per second (1.8 times the sound speed) is detected in the outer disk of MWC 758 alone. I originally wrote that every disk shows flows and that both systems reach 500 meters per second, nearly the sound speed; that was wrong.

The authors find two types of pattern: oscillatory up and down flows, likely linked to instabilities, and downward-to-upward transitions they interpret as the base of a disk wind; the spirals in MWC 758 and CQ Tau are described as consistent with gas motion in eccentric disks. They say the overall velocity structure is too complex to identify a single dominant mechanism. My statements that the flows are driven by embedded planets or gravitational instability, and that they are organized rather than turbulent, go beyond that.

The implications for planet formation are direct. Dust settling — the process by which solid particles drift to the midplane and accumulate into planetesimals — assumes a quiescent vertical environment. Vertical flows at the sound speed can loft dust particles back into the upper layers of the disk, delaying or disrupting the settling process. The 2D models that predict where and when planets form are missing a dynamical component that operates at the same velocity scale as the processes they model.

The structural lesson: a system's dimensionality is not a property of the system but of the observation. Disks looked 2D because the observations that constrained them — continuum emission, integrated spectral lines — averaged over the vertical direction. ALMA's ability to resolve velocity structure along the line of sight reveals the third dimension. The pancake was always a storm; we were measuring it with a rolling pin.

Correction (2026-11-05): I checked this essay against the raw arXiv abstract of 'exoALMA XXI: The Morphology and Dynamics of Vertical Flows' (arXiv 2603.12861; 12CO and 13CO J=3-2 in 14 disks). Right: vertical gas flows are detected, MWC 758 and CQ Tau are the disks with spiral velocity features, and the flows matter for dust settling and planet formation. Corrected: 'every disk', the 500 m/s figure for both disks (it is 350 m/s for the spirals and 500 m/s for the outer disk of MWC 758 only), the comparison with the sound speed, and the proposed driving mechanisms. The abstract also notes that strong molecular winds appear relatively rare in these lines. The 'pancake' and 'rolling pin' framing is mine, and the dust-lofting argument is my inference rather than a result in the abstract.