Supergranulation

A second, hidden honeycomb on the Sun

Point a high-resolution telescope at the Sun's surface in white light and you see granulation: a shimmering carpet of bright cells about 1,000 km across, each a column of hot plasma boiling up, spreading out, cooling, and sinking in the dark lanes between. Each granule lasts only minutes. That is the convection everyone knows.

Now stop looking at brightness and start looking at motion. Measure the Doppler shift of a spectral line across the disk and a completely different, far larger pattern emerges — one you would never have guessed from the intensity image. The surface is tiled by enormous cells about 30,000 km wide, each one spanning a bit more than two Earths laid side by side, and each one persisting for about a day. This is supergranulation. It is not a brightness pattern at all; it is a flow pattern. Plasma wells up gently in the middle of each cell, streams horizontally outward across it at several hundred metres per second, and sinks at the boundary it shares with its neighbours.

The two patterns coexist on the same surface. Granulation is the small, fast convection you see in light; supergranulation is the large, slow convection you see in velocity. Understanding why the Sun convects on two such different scales — and whether there are scales in between — is one of the durable puzzles of solar physics.

How a flow you can't see was found

Supergranulation was first noticed by A. B. Hart in 1954, who saw quasi-periodic fluctuations in the line-of-sight velocity along a strip of the solar disk. The phenomenon was properly mapped and named in a landmark 1962 paper by Robert Leighton, Robert Noyes, and George Simon, using the spectroheliograph at Mount Wilson to build full Doppler maps of the surface.

Their trick was geometric. A supergranule's vertical motion is feeble — the upflow at the centre and the downflow at the rim are only about 20–30 m/s, producing an intensity contrast under a few percent that is nearly impossible to see. But the horizontal outflow is an order of magnitude larger, 300–500 m/s. At the centre of the solar disk that horizontal flow is perpendicular to your line of sight and invisible to the Doppler effect. Move toward the limb, though, and the horizontal flow tips into the line of sight: the near edge of each cell shows plasma rushing toward you (blueshift), the far edge shows plasma rushing away (redshift). The whole surface lights up in a Doppler map as a quilt of blue-and-red cell pairs. That projection is the entire reason supergranulation is observable, and it explains why the pattern essentially vanishes at disk centre.

The mechanism: convection, conveyor belt, and flux sweeping

A supergranule is, at heart, a convection cell — the same physics as a granule, just on a different scale. Hot plasma is buoyant, rises, reaches the surface, radiates away its heat, becomes denser, and sinks. What is distinctive is the geometry: the cell is very flat. Its horizontal extent (~30,000 km) dwarfs the modest vertical velocities, so the flow is overwhelmingly sideways. Think of it as a slow, wide pancake of overturning gas rather than the tall, narrow plume of a granule.

That sideways flow does something important. The photosphere is an excellent electrical conductor, so magnetic field is effectively "frozen in" to the moving plasma — the magnetic Reynolds number is enormous, and field lines are dragged wherever the gas goes. Weak field scattered across a cell's interior is therefore continuously swept by the 300–500 m/s outflow toward the converging downdrafts at the cell boundary. There the flux piles up, concentrates, and is compressed into intense kilogauss flux tubes. Over many cell lifetimes this builds the magnetic network: a web of concentrated field outlining the supergranule edges. Seen in the chromosphere — in the cores of the Ca II H and K lines, or in H-alpha — this same web glows as the bright chromospheric network, and it is the launch site for spicules and the footpoints of coronal loops. Supergranulation is, in this sense, the organising scaffold of the Sun's surface magnetism between active regions.

The numbers: three scales of solar convection

The cleanest way to grasp supergranulation is to put it beside its neighbours. The Sun convects on at least two firmly established scales — granulation and supergranulation — with mesogranulation proposed in between and "giant cells" predicted at the very largest scale.

Scale	Cell size	Lifetime	Dominant velocity	How it's seen
Granulation	~1,000 km	8–20 min	Vertical, ~1–2 km/s	Continuum intensity (bright cells, dark lanes)
Mesogranulation (proposed)	~5,000–10,000 km	~3 h	~60 m/s	Granule-tracking, local correlation
Supergranulation	~30,000 km	~1 day (24–48 h)	Horizontal, 300–500 m/s	Doppler maps off disk centre
Giant cells (predicted)	~100,000–200,000 km	weeks–months	~few m/s	Marginal helioseismic / long-baseline tracking

Two things stand out. First, the size jump from granule to supergranule is about a factor of 30, and the lifetime jump is roughly a factor of 100 — these are not the same convection slightly rescaled. Second, supergranulation is the only one of these whose vertical motion is negligible compared with its horizontal motion. That horizontal dominance is exactly why it stays hidden in brightness and only confesses itself in velocity.

Quantified figures worth carrying around

A few concrete numbers anchor the phenomenon, all referenced to the standard solar radius R☉ = 696,000 km:

• Cell count. If a typical cell is ~30,000 km across, its area is roughly π(15,000)² ≈ 7 × 10⁸ km². The visible hemisphere has area 2πR☉² ≈ 3 × 10¹² km², so there are of order a few thousand supergranules across the disk and around 10,000 over the whole Sun at any instant. (The same surface holds millions of granules.)
• Crossing time. A parcel travelling 15,000 km from cell centre to edge at 400 m/s takes 15,000,000 m ÷ 400 m/s ≈ 37,500 s ≈ 10 hours — comfortably within the ~1-day cell lifetime, so a cell turns over roughly once or twice before it dissolves.
• Network field. Flux concentrated at the boundaries reaches kilogauss strengths (≈ 1,000–1,500 G) in flux tubes a few hundred km wide, against a quiet-Sun mean field of only a few gauss — a thousand-fold local enhancement driven purely by flow convergence.
• Super-rotation. The supergranule pattern rotates a few percent faster than the surface plasma it sits in, hinting that its roots reach into faster-rotating layers below.
• Wave dispersion. Helioseismic analysis finds the pattern drifts and oscillates with periods of order a week and phase speeds of tens of m/s — a genuine dynamical signature, not a frozen tiling.

The depth problem

For ordinary granulation the link between size and depth is intuitive: granules are about as wide as the few-hundred-km photospheric layer is thick. Apply the same reasoning to supergranules and their 30,000 km width implies driving at a depth of order 10,000–20,000 km — close to the layer where singly-ionised helium grabs its second electron (the He II ionisation zone), a natural candidate for a convective driving scale.

The trouble is that the observations disagree about how deep the flow actually reaches. Local helioseismology — using the travel times of sound waves between surface points (time–distance analysis) or the splitting of ridges in localised power spectra (ring-diagram analysis) — generally finds the supergranular flow concentrated in the upper few thousand kilometres, shallower than the size argument predicts. Some inversions even hint at a return flow at depth. There is no settled consensus: the depth of supergranulation, and indeed whether it is driven from above (by radiative cooling and granule-scale dynamics percolating upward in scale) or below (by buoyancy at an ionisation layer), remains an open research question that SDO/HMI helioseismology and high-cadence Doppler instruments are still chipping at.

Where supergranulation shows up

• The chromospheric network. The single most visible consequence. In Ca II K, H-alpha, and UV images from instruments like SDO/AIA and IRIS, the bright network you see framing the quiet Sun is supergranule boundaries lit up by the field that the outflow has herded there.
• Spicules and coronal heating. The concentrated network field at cell edges is the launch pad for spicules — jets of chromospheric plasma — and the footpoints of quiet-Sun coronal loops. The energy that escapes up these field lines feeds the still-mysterious heating of the corona to over a million kelvin.
• The solar dynamo and flux dispersal. Supergranular flows act as a random-walk diffusion engine for magnetic flux across the surface. This turbulent diffusion, combined with differential rotation and meridional flow, is built into surface-flux-transport models that reproduce the Sun's 11-year magnetic cycle and the migration of flux toward the poles.
• Helioseismic noise and signal. Supergranular flows imprint a measurable signature on the travel times of p-mode acoustic waves, making them both a target of local helioseismology and a source of correlated noise that other measurements must model out.
• Other stars. Sun-like stars must have their own supergranulation, and it is now invoked to explain part of the low-frequency "flicker" in the brightness of cool stars measured by Kepler and TESS — a granulation-and-supergranulation background that even limits the precision of exoplanet radial-velocity searches.

Common misconceptions and edge cases

• "Supergranules are just big granules." No. They differ in lifetime by ~100×, in size by ~30×, and qualitatively in flow geometry — granules are vertical-flow-dominated and seen in brightness, supergranules are horizontal-flow-dominated and seen in velocity. They may even be driven by different physics at different depths.
• "You can see supergranules in a sunspot photo." Not in intensity. The intensity contrast is only a couple of percent, far below granulation contrast. What you can see is their magnetic imprint — the chromospheric network — and their Doppler imprint away from disk centre.
• "The cells are static tiles." They evolve continuously, with cells appearing, merging and fragmenting, and the whole pattern carries a wave-like drift (Gizon, Duvall & Schou 2003) and super-rotates relative to the plasma.
• "Supergranulation is the largest convection scale." Probably not. Numerical models predict even larger "giant cells" reaching deep into the convection zone, though detecting them observationally is at the edge of what is possible. Mesogranulation, conversely, may or may not be a distinct scale at all.
• "The vertical flow doesn't matter." It is small (~20–30 m/s) but it is what makes a supergranule a convection cell rather than a horizontal vortex, and the gentle downflow at the boundary is precisely what concentrates the magnetic network. Small does not mean dynamically irrelevant.