22 THENEWYORKER, MARCH 23, 2020
placid-appearing mass which no man
had seen before—or even suspected.”
He concluded, “There are apparently
random plastic flows and currents within
the snow cover whose causes and effects
were unknown, and still are.”
In 1805, the Irish hydrographer Sir
Francis Beaufort developed a scale for
measuring wind speed at sea by obser-
vation. Later, it was adapted for use on
land. In his book “Defining the Wind,”
from 2004, Scott Huler argues that the
descriptions accompanying the scale,
which were written anonymously, should
count as literature. At Beaufort 0, the
wind is “calm; smoke rises vertically.” At
Beaufort 3, a gentle breeze, one sees “leaves
and small twigs in constant motion.” At
Beaufort 5, a fresh breeze, “small trees in
leaf begin to sway; crested wavelets form
on inland waters.” The poetic descrip-
tions connect subjective impressions to
objective reality. A near-gale—a Beau-
fort 7—is defined by “whole trees in mo-
tion; inconvenience in walking against
wind.” See and feel those things, and you
know that the wind is between thirty-two
and thirty-eight miles per hour.
Atwater devised an analogous guide
to snow. His language is evocative, but
there’s less authority in the descriptions.
“Unstable damp snow is tacky,” he wrote.
“It slithers out from underfoot and rolls
away in balls or slips blanketwise.. ..
Well settled snow has good flotation
and makes a clean, sharp track.” Snow
is less forthcoming than the wind. Its
chaos hides beneath the surface.
O
ne crisp, bright morning in Febru-
ary, I walked along a brook just out-
side the center of Davos, toward the head-
quarters of the Swiss Institute for Snow
and Avalanche Research. In Davos, the
train from the valley potters up through
wooded hills, picking up locals in ski
boots; the S.L.F., as the institute is now
known, occupies a squat building a few
minutes from the train station. A small
exhibit in the lobby explains the history
of snow and avalanches in Switzerland.
In 1951, while Atwater was experi-
menting with explosives, Switzerland
experienced the worst avalanche season
in its recorded history. Ten feet of snow
fell in ten days. About a hundred peo-
ple were killed; villages that had sur-
vived avalanches for centuries were de-
stroyed. The S.L.F., which was founded
in 1942, suddenly became an institution
of national import.
Henning Löwe, the forty-six-year-old
head of the institute’s Cold Lab, wears
an earring in his right ear; before taking
up the study of snow, he received a Ph.D.
in theoretical condensed-matter physics.
Dressed in jeans, black Nikes, and a worn
fleece shirt, he led me inside the lab, where
computers sat beside refrigerated rooms
with three-inch-thick steel doors. The
lab’s goal, he explained, was to find out
what the wetness or heaviness or hoari-
ness of snow really meant, on the level
of its crystals. “We are connecting phys-
ical properties of snow to structure,” Löwe
said. He picked up a palm-size cube that
looked elaborately hollowed out, like a
plaster mold of a termite’s nest. A twenty-
millimetre-wide sample of snow had
been taken from the crown of an ava-
lanche—the pit that’s left when a slab
releases—scanned with X-rays, and then
3-D-printed in plastic, at high magnifi-
cation: the layer cake, under a micro-
scope. The weak, bottom layer was com-
posed of what looked like large popcorn
kernels. The top layer, which had settled,
was a tight tangle, like instant ramen.
“You start to shear this thing”—Löwe
made a chopping motion where the two
layers met—“it’s ninety-nine per cent
sure that this will break there.”
Snow science has come a long way
since Atwater’s experiments at Alta. The
basic process by which newly fallen snow
crystals sinter into a cohesive slab can
now be seen in slow motion: it resem-
bles the way ice cubes in an empty glass
fuse together. The process of recrystal-
lization—the re-separating of the cubes—
was more mysterious. Löwe opened a
closet, and pulled a cylinder from a shelf
marked “Snowbreeder 3.” The device al-
lows scientists to observe a snow sam-
ple while applying varying degrees of
heat and pressure. At his computer, Löwe
played a time-lapse video of “snow meta-
morphism” in the Snowbreeder. “In the
beginning, it’s typical snow, it’s round-
grained snow, the crystals are small,” he
said. Then heat was applied from below.
The lower crystals began evaporating
their moisture to the crystals above, which
used it to grow downward. “We see that,
here, a facet’s growing. There, a facet’s
growing,” he said, pointing. This was
hoar—the snow becoming spiky, brittle,
weak. “Seeing something is always the
beginning of understanding,” he said.
The scientific study of snow layers has
refined our understanding of avalanches.
In 2008, a study published in Science by
a group of Scottish and German mate-
rials researchers modelled how, when one
part of a heavy layer of snow collapses
onto a weak layer, it can produce a wave.
Their model explained a curious obser-
vation from the field: skiers occasionally
trigger deadly avalanches above or below
them, even when standing on flat slopes.
The weak layer, it turns out, behaves like
the coils in a mattress: apply force in one
place, and it spreads all over the bed. The
concept is now a cornerstone of avalanche-
safety education, where it is known sim-
ply as “remote triggering.”
Snow research also has applications
beyond avalanches. Spinning his keys
around a finger, Löwe led me through
the cold rooms. In one, a humidifier gen-
erated tiny clouds of perfect, lab-grown
powder; in another, snow from the Arc-
tic, Finland, and Iceland had been care-
fully preserved. Scientists are studying
how snow’s crystal structure determines
its color, or “albedo,” which, in turn,
affects its ability to act like a giant mir-
ror and mitigate global warming.
In an upstairs office with mountain
views, Perry Bartelt, a gray-haired re-
search engineer, works on Rapid Mass
Movement Simulation, or RAMMS—
software for simulating avalanches. The
week before, an avalanche in Turkey had
killed half a dozen people; dozens more
died during the rescue, when the moun-
tain avalanched a second time. Turkish
researchers had rushed data from both
slides to Bartelt. RAMMS calculated that
the first avalanche had hit the bottom
of the slope with five times the force
needed to knock down a building. Its
core had the density of wood.
Using a terrain map, RAMMS predicts
the path and the power of an avalanche.
Its central innovation is its ability to treat
an avalanche as a “granular shear flow,”
using statistics to average out the activ-
ity of millions of interacting grains. Imag-
ine a box of cereal, full of flakes and
marshmallows; now pour it out. Some
bits will fly straight, carried by their own
momentum. Others will catch on the
surface they’re sliding down. Many flakes
will shake against one another, breaking
up and settling below the intact marsh-
mallows. (In granular flows, small things
