Dirty Snow

Vincent Giovannone

Introduction

It has been reported that snow containing soot will melt more rapidly under sunlight than pristine snow with little or no soot (see here, here and here). Through observations of roadside snow, which contains many impurities, including crystals of road salt, it is evident that sunlight melts dirty snow more rapidly than clean snow. An experiment was designed to test the melting rate of measured amounts of snow that had been seeded with a known amount of soot in comparison with pristine snow.

Figure 1. A layer of mineral and carbon dust coated a layer of snow (top) over clean snow (center and bottom) near Sunspot Observatory in New Mexico. (See here.) Forrest M. Mims III.

Methods

Snow samples for this study had no obvious contamination and were collected from an empty field far from roads, houses and trees. The samples ranged in age from a few hours to a few days. All samples were pristine, freshly fallen snow or snow from under the surface crust. Small party pie tins were used for sample collection. Each tin weighed about 1.50g on average based on measurements with a 100 x 0.02 gram digital scale. The pie tins were used to scoop snow samples. A sterile measuring scoop was used to add or remove snow in a pair of samples until each sample had an identical weight of between 50 to 60 grams.

After sample collection and weighing, sets of matched tins were labeled A and B. A measured amount of soot and soil (1.0 to 2.5 g) was added to Tin A, while Tin B was the control sample. The soot and soil was sprinkled onto the snow sample in Tin A and the samples were weighed again before being exposed to sunlight. The samples can be formed into snowballs or left flat in the tins.

It is important to note that even apparently pristine snow may already contain soot picked up as it fell from the sky or that later fell on the fallen snow. These particles, if present, are difficult to see or microscopic in size. Filters can be placed beneath snow samples to collect some of these particles. The filters should be carefully weighed before the experiment and placed over the party tin base. The filters I used for this experiment were of a medium grade paper and about 10 cm (4”) in diameter, a little larger than the tins. The filters were shaped into a shallow cone and placed in the tins. The samples were formed into snowballs and placed over the filters. The snow was allowed to melt, and the filters were weighed after they dried. The weight difference indicates the mass of the soot in the snow sample. The filters can be examined under a microscope for further study.

Samples were placed side by side in sunlight on a sloped window sill. This allowed the liquid water to pool and gradually spill out. Any particles were left on the filters. Filters were placed under both snow samples or just sample Tin B.

The start time was recorded the moment the samples were exposed to sunlight, and every ten minutes thereafter the weight of the samples was taken and recorded. The experiment was continued until both samples were completely melted and the experiment was ended.

After the filters were removed from the tins, any remaining water in the tins was poured through their respective filter. The filters were then placed on wax paper and allowed to to dry. The tins were cleaned with distilled water and left to dry.

Results

Table 1 shows the results obtained with a typical pair of samples. Tin A has been seeded with soot and Tin B is pristine snow. After 60 minutes, the sample with soot weighed only half as much as the pristine sample. The data in the table are plotted in the chart shown in Fig. 2. Note that the melt rate for both sample is is approximately linear.

Table 1. The melting rate of two snow samples in direct sunlight. The snow in Tin A has been sprinkled with soot and the snow in Tin B is pristine.

Figure 2. The data in Table 1 are plotted in this chart. Note that the melt rate for both sample is is approximately linear.

Conclusion

According to NASA scientist Jim Hansen (see here), "Black carbon reduces the amount of energy reflected by snow back into space, thus heating the snow surface more than if there were no black carbon." The data seems to support this statement. In all trials of this experiment the snow samples with added soot melted more rapidly than the pristine samples. In the trial presented in Table 1, after the first ten minutes there was a weight change of -0.24 g for Tin A (with soot) and -0.26 g for Tin B (without soot). After 20 minutes Tin A had lost 7.88 g and Tin B 4.52 g. After 30 minutes Tin A had lost 20.12 g and Tin B 14.42 g. The data show that a little time was required for the soot to become warm due to its increased absorption of solar energy and to transfer the warmth to the surrounding snow. At 60 minutes, Tin A had lost 52.66 g of liquid and Tin B had lost 39.48 g and still contained some unmelted snow when the experiment was ended. The approximate melting rate for Tin A was 0.88 g/min.and for Tin B was 0.66 g/min.

In all of the experimental trials, liquid water would spill from Tin A before Tin B. In the trial shown in Table 1, Tin A began to spill water after 13 minutes and Tin B 15 minutes after exposure to sunlight. The spill times were different for most of the trials, probably due to the difference in the snow (which ranged in age from a few hours to a few days) and the amount of soot. In one trial the snow already had a high water content, and Tin A began spilling at 4 minutes and Tin B at 7 minutes.

Earlier experiments involved the use of filters to trap particles in pristine snow. The filters were weighed before use. Three experiments were conducted, and material having a combined weight of 0.08 g was recovered. Examination through a 100-power microscope revealed that some of the particles were soot. A further analysis with a digital microscope could be conducted, but this goes beyond the scope of this report. Experiments were also conducted in an effort to measure the surface temperature of the snow in both Tin A and B. The bulb of a standard liquid thermometer was placed at the surface of the samples, but the results were inconclusive. Perhaps an electronic thermometer could be used to monitor the temperature as the sample melts. How much a role the air temperature plays, if any, is unknown