It starts with a soft, almost soothing, hiss, but it is not a comforting sound. The hiss -- the sound of wind and snow grains sliding along the ground--is the prelude to a mighty symphony of noise, swirling snow, and danger. The snow symphony spawns whiteout (Fig. 1) and gridlock. When your work is studying snow, a blizzard is both a fascinating event and an occupational hazard.

Let’s start with the basics: blizzards come in two types: 1) those that are the result of a combination of wind and new falling snow, and 2) those that are the result of winds picking up snow that is already on the ground (ground blizzard). From the swirling insides, these two look identical, but the second type needs stronger winds to get it started because snow on the ground is often bonded and the wind has to break those bonds before the grains start moving. The rule of thumb is that the ground wind needs to be about 6 m s-1 before that happens. Oddly, once the second type is underway, it takes less wind to keep it going because the surface is slicker than when new snow is falling.

What differentiates a blizzard from a mere windstorm is the drifting snow. There are three mechanisms that move the snow: creep, saltation, and suspension. As the wind increases in speed and violence, each of these mechanisms in turn adds to the cacophony of the blizzard. Creep is just another name for rolling and sliding, and it is best seen in action out on a frozen lake or on hard packed snow (Fig. 2). The wind starts in motion sinuous tendrils of grains. These slither downwind like wispy snakes, the snow grains sliding along the slick surface.

More often, the surface is slightly rough, and soon a few of the grains will hit a bump and bound into the air. Down close to the surface, the wind speed follows a logarithmic profile. It might be blowing at 3 m s-1 at 1 cm height, but 6 m s-1 at 10 cm height. The wind’s capacity to move a snow grain is at least a cubic power law, so this increase in wind speed magnifies the wind’s carrying capacity by a factor of 8 or more. Now those grains that have popped up 10 cm are in the stronger winds, so they really go flying! Moreover, as these grains descend (because gravity is pulling them back down) they impact other grains, ejecting them upward, adding to the flying mass in a cascade that increases the flux of grains many-fold. The whole process is called saltation, and it is thought to be the mechanism that moves the most snow during a blizzard (for wind less than 15 m s-1), although it is not the mechanism that makes blizzards the most dangerous.

The individual saltating grains follow beautiful parabolic trajectories that rise 5 to 15 cm above the snow surface. These trajectories have fascinated scientists for more than 80 years. The cascading collision process, in particular, has been the subject of intense scrutiny because it not only sustains, but also increases the flux of grains with time and distance downwind. Daiji Kobayashi in Japan was the first to successfully photograph these beautiful trajectories (Figs. 3a and 3b). Personally, I accidently learned about a novel way to observe saltation during a winter climbing trip in the Wrangell Mountains of Alaska in 1988. A big windstorm came in and pinned us in a snow cave on the lee side of a ridge for 6 days. The cave was in the downhill side of a crevasse, so when I popped up out of the crevasse to check the weather (which was frequently), my head was right at the surface of the snow, where I was blasted by saltating grains. When I could keep my eyes open, I could see the collisions, the bounding, the subsequent ejection of grains, and the wonderful parabolic trajectories, all up close and personal. Right up until my eyelids froze shut.

Strangely enough, much of what we know about saltating snow grains comes from the deserts of North Africa. Brigadier Ralph Alger Bagnold (1896-1990) (Fig. 4), a British soldier, explorer, and scientist, became fascinated with deserts while serving in Egypt. Using Model T Fords, he explored the Libyan Desert in a series of bold expeditions in the 1920s and 1930s. This fascination led Bagnold to conduct a classic series of wind tunnel experiments in the late-1930s and the publication of The Physics of Blown Sand and Desert Dunes in 1941. The method of travel he pioneered is still used today with expeditions and tourists visiting Libya, and virtually every modern study of blowing snow and sand can be traced back to his work.

Bagnold was the first to recognize a fundamental (and surprising) piece of physics that allows blowing snow and sand to produce a bewildering range of features. These features vary from ripple marks to barchans to dunes and cornices (Fig. 5). The physics works like this: for both sand and snow there is a threshold surface friction velocity necessary to start the grains moving. Researchers usually call this u*critical. Think of it as the speed at which the air needs to be moving in order for it to be able to pluck out grains.

Clearly, this critical speed is related to the size and weight of the grains as well as the wind speed. But the surface friction velocity is also an inverse function of the surface roughness. As the surface roughness goes up friction velocity goes down. Saltating grains in the air stream make the effective surface roughness go up… a lot. They rob the wind of its momentum, so the wind speed has to go down (Fig. 6).

The practical implication of this mechanism is that no sooner has the wind increased enough to start grains saltating, when the wind is slowed by those very grains. This drops the wind below threshold, and whatever grains had been entrained in the wind now fall out and are deposited rather than eroded. As soon as this happens, the wind speed goes back up (because the surface roughness increase due to saltation has been removed) and it starts eroding once again. In a sense, the saltation process has built into it an inherent pulse, with erosion and deposition alternating downwind at scales determined by whether the wind is rising, falling, or staying steady, and how much snow is available for entrainment.

The last mechanism that moves snow in a blizzard is suspension (Figs. 7a, 7b, and 7c). During both creep and saltation the snow grains collide and break, producing smaller grain fragments that are also lifted by the wind. These particles, however, have much lower settling velocities than the saltating grains, so instead of descending back to the snow surface, they can be lofted higher and higher into the air. Collectively they are called the suspended load, and they are responsible for the poor visibility associated with blizzards. I have often found myself in a blizzard where looking horizontally I can barely make out my companions 10 meters away, but looking straight up I can see the sun shining.

Sadly, I have also known of several fatalities when aircraft have tried to land in a blizzard when the suspended load was making it impossible to see the ground. The suspended load is constantly settling out of the blizzard and being recharged at the same time. The suspended particles are easily observed in the lee of a building, cutbank or snow fence. There, where one is protected from the wind (mercifully), a steady rain of small snow particles descend out of the sky and covers everything (Fig. 7c). Because of their tiny size (and sharp points) these particles tend to be very clingy. In no time, everything is covered in this flour-like snow, which goes by the name spindrift.



An ongoing area of our research and others is focused on determining “Just how much snow can a blizzard move?” The answer to this question affects how we design building complexes and our estimates of snow removal costs in places where there are blizzards. The consensus is that that amount increases with the wind speed to the third or even forth power. In other words, if snow starts moving at 6 m s-1, by the time it has hit 18 m s-1 the flux has increased 30-fold. Translating these numbers into more tangible terms, a 24-hour blizzard with winds in excess of 10 m s-1 and an ample supply of snow can move 20 m3 of snow per lineal meter perpendicular to the wind.

Where I work near Barrow, Alaska we get about six of these blizzards each winter, producing about 130 m3 m-1 of snow moved during the season. We check this theoretical value by measuring how much snow has been trapped in the lee of a large snow fence (Fig. 8), and they agree pretty well. That particular fence protects a housing subdivision on Cakeater Road. Without the fence, which is over 2 kilometers long, 17 million kg of snow might otherwise end up amongst the houses and would have to be dug out using front-end loaders, plows and shovels.

Estimates of the costs associated with cleaning up after blizzards are difficult to make, and probably wildly inaccurate, but almost certainly would exceed hundreds of millions of dollars each year in the U.S. Similarly, with the exception of people trapped in cars or crashes resulting from the low visibility associated with blowing snow, the actual injury and death toll due blizzards is hard to pin down. Nevertheless, a few simple physical process combining snow and wind have a profound effect on our lives.