Dirty Snow Absorbs Two to Five Times More Sunlight
Snow looks white because it reflects most of the visible light that lands on it. How much sunlight a surface reflects, or bounces back, is called its albedo. NASA reports that fresh snow reflects 80 to 90 percent of the sunlight that falls on it, while dusty snow reflects only 50 to 60 percent and absorbs the rest[1].
Flip the numbers around and the difference gets bigger. Fresh snow absorbs about 10 to 20 percent of the sunlight. Dusty snow absorbs about 40 to 50 percent. So, under the same sunlight, dusty snow absorbs roughly two to five times the fraction that fresh snow does (20 to 40 percent at the near end, 10 to 50 percent at the far end). These are rounded ranges from one source, so think of them as a rough guide, not an exact rule. Snow that is not brand new also darkens as its grains grow and it collects dirt, so even "clean" old snow reflects less than a fresh snowfall.
Still, a gap of a few tens of percent sounds small next to a gap of weeks. Something must be multiplying it.
Why a Little Dust Keeps Making Snow Darker as It Melts
The multiplier is a loop. According to NASA, when dust lets the snowpack absorb more heat, the snow grains begin to cluster into bigger grains. As the grains grow, the snow absorbs more sunlight[1]. More absorbed light means more melting, and more melting means bigger grains and less reflection. A small difference at the start grows over time.
Not every kind of dirt is equally strong. In a classic scientific review, Stephen Warren reported that very small amounts of soil dust, measured in parts per million (a very tiny amount), can noticeably lower snow's albedo. An automated summary of the review adds that soot is about 50 times more effective than dust at the same weight[5]. That is a comparison per unit of weight. In real snow, dust is often present in much larger amounts than soot, so both the kind and the amount of dirt matter.
So a loop can turn a thin film of dirt into a big effect. What does that add up to on a real mountain?
How Much Earlier Dust Made Colorado's Mountain Snow Disappear
According to NSIDC's summary, researchers in the San Juan Mountains of southwestern Colorado found that a layer of dust on a snowpack can shorten the time snow covers the ground by 21 to 51 days. For comparison, the same summary gives about 5 to 18 days for an air temperature rise of 2 to 4 °C[2]. That result comes from this mountain range, so it should not be read as a number for every mountain.
Dust arrives in bursts, blown in from dry land far away. NASA reports that in the winter and spring of 2005 only four dust events blew over the San Juans, and 2006 had about twice as many. Models and observations together showed snow melting up to a month earlier in some places[3].
The effect does not stay on the mountain, because snowmelt becomes river water. Many people in the western United States depend on rivers such as the Colorado, which is fed largely by mountain snow. In a 2010 study, Thomas Painter and colleagues used a computer model to compare runoff with and without the extra dust that settled after the American West was settled by farmers and ranchers. They estimated that the peak of the Colorado River's runoff at Lees Ferry, Arizona, came on average about three weeks earlier, and that annual runoff fell by about 5 percent[4]. This is a model estimate, not a direct measurement. NASA adds that by 1890 airborne dust was about five to seven times greater than before settlement[1].
If dark specks make snow melt faster, can too much dust make snow melt more slowly?
Why a Thick Blanket of Dirt Can Slow Melting Instead
It can. Warren's review notes that very large amounts of dust on snow can actually inhibit melting by insulating it[5]. A thin dusting absorbs sunlight and passes the heat to the snow. A thick layer works more like a blanket: it can absorb the light at the top but keep that heat from reaching the snow underneath. The same idea explains why a pile of gravel on ice can leave a lump of ice standing after the surrounding ice has melted.
So the rule is not "darker always melts faster." It depends on how much dirt there is, and also on the weather and on how the snow is lying.
Dust Is Not the Only Thing That Darkens Snow
The U.S. National Snow and Ice Data Center points out that dark particles such as dust and soot absorb more sunlight and speed up snowmelt. It also notes that living things can color snow. Cold-loving algae with a red pigment make what is called watermelon snow, which looks pink or red[2]. What darkens snow depends on the place: desert dust in the Colorado mountains, soot near towns, algae in some other snow. The reasoning above applies to the dust-and-soot cases the sources measured. This article does not claim it for every dark snow in the world.
How to Compare Light Absorption Using Ice on White and Black Paper
You can see the main idea at home, with an adult nearby. Lay a sheet of white paper and a sheet of black paper side by side on a sunny table, such as one by a window. Put an ice cube of the same size on each, and check them every two or three minutes. Compare their sizes after ten minutes. If the ice on the black paper shrinks faster, you have seen darker surfaces absorbing more light.
Ice is not the same as snow, and your result will depend on how strong the sun is and how big the cubes are, so keep everything else the same. Another easy thing to try is to compare roadside snow with snow in a clean yard, and see which melts first. Do it from a safe place well away from traffic.
To read further, start with NASA's "Dusting the Virtues of Snow" and the NSIDC page on the science of snow. If you enjoy numbers, the abstract of Painter's 2010 study shows how scientists turn mountain snow into a river forecast.