Rimed Snow Crystals vs Graupel: How Supercooled Droplets Turn a Dendrite Into a Pellet

Most explanations of graupel stop at "snow pellets." That's why so many winter observers can't tell a rimed dendrite from true graupel when one lands on their sleeve. You see something white and lumpy, log it as graupel and move on. Often it's a rimed crystal with its six-armed frame still intact underneath, and the difference tells you a lot about what the cloud overhead was doing. I made that mistake on my own observation sheets for two full winters before I bought a decent loupe.

This article covers how riming actually works: how 10–40 micron droplets freeze on contact, why the first ones land on branch tips, where the 0.2 mg tipping point sits, and how density and fall speed change at each stage. It also shows how to classify what lands on your collection board without lab equipment.

How a Supercooled Droplet Freezes Onto a Snow Crystal

Clouds below 0 °C are often full of liquid water. Those droplets stay liquid because pure water needs something to freeze onto, like an ice nucleus or an ice surface. A falling snow crystal is that surface.

Rime accretion happens when supercooled droplets roughly 10 to 40 microns across hit a crystal and freeze almost instantly. The freezing is so fast that each droplet keeps its spherical shape. Under magnification you don't see a smooth ice coating. You see tiny frozen beads stuck to the crystal like condensation on a cold glass, except each bead is solid ice.

Three features of this process matter for identification:

The droplets keep their round shape. Small droplets lose their latent heat to the surrounding air fast enough to freeze in place before they can spread out.
Air gets trapped between the beads. Frozen spheres stack loosely, so rime is porous. That's why heavily rimed particles are still low density compared to solid ice.
The crystal underneath stays intact. Light riming adds mass but doesn't break the hexagonal frame. The original crystal is still there and still identifiable.

Why Small Crystals Barely Rime

Not every crystal that passes through a droplet-rich layer picks up rime. A droplet has to actually hit the crystal, and small droplets tend to follow the air as it flows around a falling object instead of hitting it. Small crystals fall slowly and push only a little air aside, so most droplets drift past them.

Lab and field work dating back to the late 1960s puts the onset of riming for plate-type crystals at roughly 300 microns across. Columns need to be about 50 microns wide before collection becomes efficient. Below those sizes you'll see clean crystals even inside a cloud loaded with supercooled water. I've caught tiny, perfectly clean plates mixed in with heavily rimed dendrites in the same snowfall. People assume that means two different cloud layers, but it's simply size-dependent collection.

Why the First Droplets Land on Branch Tips

Rime doesn't spread evenly. The first beads collect on the outer edges and branch tips of a dendrite or plate, because those parts stick farthest into the droplet stream. When I look at a lightly rimed stellar dendrite under 15x, the tips look frosted while the center hub is still glassy. If you see beading only at the edges, you're looking at early-stage riming, not graupel.

When Riming Outpaces Crystal Growth

A snow crystal grows in two competing ways:

Vapor deposition: water vapor deposits directly onto the ice lattice and builds the familiar hexagonal arms and facets.
Accretion: supercooled droplets hit the crystal and freeze as rime.

When droplets are scarce, deposition wins and you get the snowflakes from textbook photos. When the cloud holds a lot of liquid water, accretion takes over. Graupel is riming at its extreme, where droplets are colliding with the crystal faster than vapor can build new structure.

The working threshold I use is about 0.2 milligrams of rime per crystal. Past that point, the original hexagonal symmetry is completely buried. What's left is a low-density pellet, usually conical, with a density between 0.1 and 0.3 g/cm³. That's still far lighter than solid ice at about 0.92 g/cm³.

The process speeds itself up. Each droplet adds mass, so the particle falls faster. A faster particle sweeps through more cloud volume per second and collects more droplets, which adds more mass. Once a crystal gets past the heavily rimed stage in a droplet-rich layer, the jump to graupel happens quickly.

Why Graupel Often Ends Up Conical

A rimed particle that falls in a stable orientation collects droplets mostly on the face that leads into the airflow. The original crystal stays at the narrow tip, and the rime builds up into a widening cone behind it. That's conical graupel, the classic shape.

If the particle tumbles, rime builds on all sides and you get lump graupel, which is rounded and irregular. A plate that rimes evenly while still flat becomes hexagonal graupel, where you can still see a faint six-sided outline under a heavy rime shell. All three shapes show up together in convective snow showers, and the mix gives you a hint about how turbulent the air was.

Rime Accretion Stage and Physical Property Classification

This is the reference table I keep taped inside my observation notebook. The crystal visibility column is the one you'll use most in the field, because it's the only one you can judge by eye.

Accretion Stage Name	Droplet Density per mm²	Original Crystal Visibility	Particle Density (g/cm³)	Terminal Fall Velocity (m/s)
Unrimed Crystal	0 droplets	100% (fully visible)	0.01 to 0.05	0.3 to 0.8
Lightly Rimed	10 to 50 droplets	>80% visible	0.05 to 0.08	0.8 to 1.1
Heavily Rimed	200 to 500 droplets	<20% visible (outline obscured)	0.10 to 0.18	1.2 to 1.8
Graupel (Soft Hail)	Continuous droplet shell	0% (crystal completely enveloped)	0.20 to 0.35	1.8 to 3.0

Notice how fast terminal velocity climbs. A graupel pellet can fall up to ten times faster than the unrimed dendrite it started as. You can see this without any equipment. Unrimed dendrites drift and wander as they fall, while graupel comes down straight and fast and makes an audible tick on a car roof or jacket hood.

How to Identify Each Stage on a Collection Board

Unrimed: sharp edges, clear facets, and light passes through the arms. Branches look like glass.
Lightly rimed: scattered white beads, mostly at the tips. You can still count all six arms easily.
Heavily rimed: the arms look thick and crusty, and you can only make out the crystal's outline by tilting it under the loupe. Most people log these as graupel by mistake.
Graupel: opaque, white and round or cone-shaped, with no internal structure visible. Pressing it with a fingernail crushes it into rime fragments.

Common Graupel Explanations That Are Wrong

Here's where most popular explanations go wrong:

"Graupel is just snow pellets." The name describes what it looks like, not how it forms. It doesn't explain why you'll see graupel in one snow shower and clean dendrites an hour later.
"Graupel is the same as sleet." It isn't. Ice pellets (what US forecasters call sleet) are raindrops that refroze on the way down. They're clear or translucent, hard, and bounce like tiny marbles. Graupel is white, opaque and soft.
"Rimed snow means wet snow." Rime is frozen droplets, and rimed snow can be completely dry. Wet snow comes from melting, which is a different process.
"Graupel is soft hail, so it's basically hail." Small hail usually starts as graupel that then collected a layer of clear glaze ice. Once that glaze forms, it's a different particle.

Particle	Origin	Appearance	Squeeze Test
Heavily rimed crystal	Crystal plus rime, frame intact	White, outline faintly visible	Crumbles, arms detectable
Graupel	Crystal fully buried in rime	Opaque white, conical or lumpy	Crushes easily
Ice pellet	Refrozen raindrop	Clear or translucent, spherical	Hard, doesn't crush
Small hail	Graupel with glaze coating	Partly translucent shell	Hard outer layer

What Rime Tells You About the Cloud Overhead

Any rime on a crystal means the cloud contained supercooled liquid water. That alone tells you more than "it snowed."

Here's what I've learned to read from it:

Convective snow showers (cold air over relatively warm ground or open water) produce graupel reliably because updrafts keep supplying liquid water. One March afternoon under a lake-effect band, I watched the snow go from crisp stellar dendrites to lightly rimed crystals to bouncing cone graupel in about twenty minutes as a stronger cell moved overhead.
Secondary ice production. Between about −3 and −8 °C, a graupel particle that collects droplets throws off tiny ice splinters, a process known as the Hallett–Mossop process. Those splinters grow into new crystals, which is one reason graupel showers can switch suddenly to heavy, small-crystal snow.
Charge separation. Collisions between graupel and small ice crystals in the mixed-phase zone are a major way thunderstorm clouds build up electric charge. When you hear thundersnow, graupel is usually falling too.
Accumulation depth. Rimed snow settles into a denser layer than unrimed dendrites. If your snow board shows less depth than the forecast predicted while the precipitation total looks about right, heavy riming is often the reason.

How to Log Rime Stages Without Lab Equipment

My kit cost less than a pair of gloves:

A piece of black foam board or black velvet glued to cardboard
A 10x–20x jeweler's loupe, or a clip-on macro lens for your phone
A small notebook, or a notes app with a timestamp

My routine:

Pre-chill the board outside for ten minutes so crystals don't melt when they land.
Catch for 10–20 seconds, then pull the board under cover. Longer catches pile crystals on top of each other.
Hold your breath while viewing. Your breath carries enough moisture and heat to soften rime beads in seconds.
Judge outline visibility first and assign a stage from the table above.
Record the Magono–Lee code if you want to compare with formal datasets. R1 is rimed crystals, R2 is densely rimed, R3 is graupel-like snow, and R4 is graupel (hexagonal, lump or conical).
Note the time and air temperature. Changes in riming over a single event are more useful than one snapshot.

One habit that fixed most of my misclassifications: before writing "graupel," tilt the particle under the loupe and look for a straight edge or branch angle. If you find one, it's heavily rimed snow, not graupel.

FAQ

Can a rimed crystal still be classified by its original type?

Yes, as long as some of the frame is still visible. Formal schemes like Magono–Lee have separate rimed categories for exactly this, such as a rimed stellar dendrite versus a densely rimed one. Once the outline disappears completely, it moves into the graupel category.

Why does graupel bounce if it's so low in density?

Because rime is rigid even though it's porous. The frozen beads lock into a stiff structure, so the pellet rebounds from hard surfaces instead of sticking or flattening the way wet snow does.

Can riming happen when the air at the ground is above freezing?

Riming happens aloft, where the crystal and droplets are both below 0 °C. Surface temperature only affects what happens after the particle forms. Graupel often falls through a shallow layer just above freezing and still reaches the ground intact because it falls so quickly.

Is the 0.2 mg threshold something I can measure at home?

No. Use crystal outline visibility as your practical substitute. Weighing individual particles takes lab equipment. The visibility column in the table tracks the same transition well enough for observation logs.

Why do I see graupel mostly in spring and fall instead of midwinter?

Those seasons have more convective snow showers. Strong sun on bare ground or relatively warm water under cold air creates updrafts that keep clouds supplied with supercooled droplets. Steady midwinter snowfall from broad, stable cloud layers usually has less liquid water to rime with.

Does heavy riming change how snow behaves once it's on the ground?

Yes. Rimed snow and graupel pack differently from unrimed dendrites. Graupel pellets roll and collect in low spots, and heavily rimed snow settles into a denser layer. Logging the riming stage helps explain differences in depth and texture between events.

What to Do on Your Next Snow Day

Set up your board before the next forecast snow shower, not during it. Make a quick visual copy of the classification table for your phone's lock screen. During the event, take a catch every fifteen minutes and log only the stage and temperature. After three or four events, compare your logs against the radar loops. You'll start to see which kinds of showers produce graupel where you live, and you'll be able to tell from the precipitation alone what the cloud was doing.