Road Management & Engineering Journal
Road Management & Engineering Journal
January 1, 1998
TranSafety, Inc.
(360) 683-6276
Fax: (360) 335-6402

Snow Road Enhancement
by Deborah Diemand, Russ Alger, and Valeri Klokov1

(This report was reproduced from the Transportation Research Board's Transportation Research Record 1534: Soils, Geology, and Foundations--Geosynthetics: Cold Regions, Flexible Pavements, and Other Issues published by National Academy Press in Washington, D.C., 1996.)


Snow roads are used extensively in areas where seasonal access to remote areas would otherwise be difficult or impossible for wheeled vehicles. Forestry operations in Scandinavia and Canada, petroleum operations in Alaska and Canada, and almost all activities in Antarctica make extensive use of this technology. Many techniques of preparing snow roads and runways have been used and studied, but the most intractable problems remain unsolved: how to extend the service life of the road as the warm season approaches and how to bridge damaged or transitional sections. Other, less important problems include sinkage of parked vehicles, damage to heavily trafficked areas, damage caused by fluid spills and infiltration by saltwater, and use limited to vehicles with low tire pressures. Research addressing these problems was conducted, and the preliminary results are encouraging. A short test section of road was constructed with geocells. This material is designed for use with sand or gravel but, instead, the cells were filled with packed snow. The resulting surface was very hard, stable, and resistant to damage by repeated passes by wheeled traffic. Paving blocks were also prepared by converting snow directly to ice by using very high compaction pressures in a hydraulic press. The material was very strong and was resistant to the infiltration of fluids of all kinds. The application of these two techniques would greatly reduce most problems encountered in the use of snow roads and runways.

Compacted snow roads and runways have been studied and used for many years in polar areas. Many techniques have been used with varying degrees of success, but the aim of all these techniques is to plane and compact the existing snow cover and allow it to sinter to increase its strength. The density of the resultant material is generally in the range of 0.5 to 0.7 g/cm3 , and many months may be needed to reach maximum strength. Under ideal conditions a road surface prepared in this way is still limited to use by vehicles with relatively low ground pressures and is easily damaged by excessive speed and various environmental influences (warm temperatures, fluid infiltration, etc.).

It is often the case that a compacted snow pavement made by conventional means is adequate for its task over most of its length but becomes impassable over short transitional or damaged areas. This study has concentrated on developing methods to overcome these problems and has explored two methods that have shown considerable promise both in the initial preparation of the road and for the subsequent repair or reinforcement of an existing road: (a) the use of geocells, in which, using a heavy, cellular, plastic geosynthetic material designed for soil stabilization, a stable road section that could withstand heavy wheeled traffic was prepared, and (b) the use of compacted ice blocks, in which, by compacting snow at very high pressures to a density greater than 0.80 g/cm3, ice blocks suitable for use as paving blocks that would be both strong and impermeable to fluids were produced. Which of these two methods should be used depends largely on the materials and equipment available.


A conventional snow road is normally made by processing (e.g., harrowing) the natural snow cover to increase the density of the upper layer. This is followed by repeated planing and rolling, gradually increasing the weight of the roller. It is a lengthy personnel- and equipment-intensive process requiring frequent maintenance after it is complete. It is also easily damaged, especially by wheeled vehicles either stopping or accelerating too quickly. The problem is that under these circumstances the tires tend to sink into the relatively soft road surface, displacing the unconfined material sideways and producing ruts. This is a result of local weakness of the compact snow rather than the global weakness of the road surface. The problem is similar to that faced by wheeled vehicles in sand. Because geocells have been used with great success under sandy conditions, we believed that this material would have similar value in snow and undertook a field test in December 1991 in Houghton, Michigan.

Another advantage of using this method is that in areas with heavy snow accumulations, the geocells can be laid down quickly and filled relatively quickly and, once this is done, can be used immediately, at least for less demanding applications. The bearing strength will increase with time as the material sinters.

The geocell used was an expandable plastic web designed for soil confinement and stabilization. The expanded cells were about 20 cm (8 in.) in diameter. The geocells have been used successfully by the military in sandy regions, and, when filled with sand or gravel, the material produces a hard and durable surface that can be used easily by wheeled vehicles of all types. It is available in a number of different sizes and either in black or white. Figure 1 shows the geocells as delivered (collapsed) and also expanded before filling.

Geocell material in collapsed form, as delivered.
Geocell material after stretching prior to filling with snow.

In the present trial sections with an expanded size of 2.43 m (8 ft) by 6.10 m (20 ft) were used. Use of both black and white geocells determined whether black material would absorb enough heat from sunlight to have any significant effect on the strength of the road. (It was found that the black material was about 2C warmer than the white material.)

Using geocells with two different thicknesses, 15.2 cm (6 in.) and 20.3 cm (8 in.), permitted testing whether the thickness of the geocell would have a noticeable effect on the road's bearing capacity. Second, the study team wanted to determine if the snow in the cells could be packed to a uniform density throughout the thickness of the section.

Figure 2 depicts a diagram of the 50-m (160-ft) prepared test section showing the four 12.2-m (40-ft) component sections, with half being the thinner material (black and white). This test section was laid out on a flat field with snow cover of about 30 cm (12 in.). A week before the test the snow cover was compacted and smoothed with a tracked vehicle followed by a drag. The snow temperature at the time of the test was about -1C.

Layout of test section depicting 5 sections each
of black and white geocells both 20 and 15 cm thick.

To stretch the geocell sections to fill them with snow, a tubular frame was used (Figure 3). After the cells were stretched , the first two rows of cells at each end were filled with enough snow to anchor the web section; the frame was then removed. In this way it was possible to lay out a number of expanded, empty sections before beginning to fill them, as shown in Figure 4.

Stretcher frame used to expand geocells before filling with snow.

Push-type snowblower used to fill geocell sections.

A large Blanchet snowblower was used to fill the geocells with snow by making successful passes along the edges of the test sections and out into the surrounding field with the blower and depositing the disaggregated snow onto the honeycomb sections. After sufficient snow had been blown onto the cells, small blowers were used on the road section itself to distribute it more evenly. Finally, the snow was pushed around on the honeycomb with shovels and scoops to try to fill all of the sections evenly, a difficult task because the snow was very wet and heavy. At the end of this process the geocells were completely filled and covered with snow. Colder, drier snow would probably be much easier to distribute and compact.

The last stage in the preparation of the road section was the compaction of the snow into the cells. This was accomplished with repeated passes of a 3/4-ton pickup truck such that the wheels passed over the entire surface at least twice. At the end of the compaction process the road surface appeared very hard and strong.

Unusually warm weather continued for the next few days and included about 7.5 mm of rain the day after the road section had been prepared. Three days later, when normal cold temperature conditions returned, the compacted wet snow and the subsequent rain froze; the resulting road section became extremely hard and strong. An M113 and a 5-ton water truck drove on the section, causing no damage. Attempts at quantifying the strength of the wet snow where not successful.

Although study participants were not able to obtain measurements of the density or strength of the compacted snow, the evidence of the performance test road suggests that the use of geocells could offer significant benefits to polar operations.

The conditions prevailing during this test were ideal, and the final product was suitable for use by heavy equipment after a very short time. Although such conditions cannot be relied on, the results justify more extensive field tests under more challenging conditions.


A serious problem in the use of compacted snow roads isvapor movement and infiltration of liquids, which may cause a loss of strength through recrystallization or partial melting. This occurs because at the relatively low snow density (usually less than 0.7 g/cm3) of conventionally prepared snow roads, circulation of air and liquids is possible. The density at which snow becomes ice, that is, when any pores are closed off and circulation is no longer possible, is thought to be 0.82 g/cm3. (1) Techniques involving the addition of heat or water, or both, have been used to produce a high- density road surface, but these tests have yielded uneven results and are often so fuel intensive as to make them practically infeasible for large-scale use. The study team therefore undertook the production of compact snow blocks with densities in the range of 0.8 to 0.9 g/cm3 using a hot pressing technique.

Hot pressing, in which particles of parent material near its melting point are compacted at high pressures, is often used in metal and ceramic manufacture because it reduces defect size and grain growth over the standard method of forming the material under pressure and the heat to complete the sintering. The end product has effectively completed the sintering process, that is, reach its maximum strength, before it has left the mold.

After preliminary testing (2) it was found that compaction pressures of about 7 MPa (1,000 lb/in.2 ) were necessary to achieve the desired density. A round form was used in the preliminary tests because it was the most efficient from the standpoint of producing blocks. However, from the standpoint of applying the blocks in their final role as a pavement, a square or rectangle shape would have been preferable. As a compromise between the two, a hexagonal shape was chosen as the optimum shape for the paving blocks; the optimum size was about 10 to 15 cm (4 to 6 in.) thick, with sides of about 23 cm (9 in.). The final product, shown in Figure 5, weighed about 22 kg, with densities from 0.80 to 0.89 g/cm3.

Single, hexagonal, compacted-snow block measuring 23 cm (9 in.) on a side with a
thickness of about 15 cm (4-6 in.); density ranged from 0.80 to 0.89 gm/cm3.

Natural snow was collected from an undisturbed drifted area to make the blocks. The snow was gathered in a large metal barrel and stirred using a 3/4-in. electric drill equipped with a heavy wire whisk from a commercial food processor. The density of the snow after mixing ranged from 0.42 to 0.45 g/cm3. The snow temperature at the time of the compaction ranged from -5C to -15C and seemed to have little effect on the properties of the final product.

The hexagonal form was devised by inserting six plastic spacers inside a section of a large, iron pipe so that the hexagonal form would fit snugly inside the pipe, with the spacers supporting the sides of the hexagonal block against any lateral expansion during compaction. A 200-ton upward-acting hydraulic press was used to make the blocks. The final load was calculated by measuring the pressure of the hydraulic fluid and multiplying this by the area of the cylinder in the press.

Samples of these blocks as well as those made in the preliminary tests showed that this compacted, high-density snow ice is very strong and fine grained. The research team measured the strength in compression and flexure of this manufactured product as well as those of naturally occurring snow ice and clear lake ice. The results are provided in Table 1.

Average Strength in Compression and Flexure of Compacted Snow Ice and Two Types of Naturally Occurring Ice at -10C

Strength (MPa)
Compressive Flexural
Compacted snow ice 9.7 2.2 4.4 0.6
Natural snow ice 3.3 1.0 3.0 0.9
Lake ice 1.7 0.4 2.1 0.4

Finally, a short road segment was constructed and paved with the hexagonal blocks. Because there were not sufficient blocks to produce a full-sized road, we laid down two strips, each two blocks wide, so that the wheels of the test vehicle (a Jeep) would ride on the blocks leaving the central strip undisturbed. The test area was not prepared in any way. The site chosen was a rough, grassy area with about 15 cm of undisturbed snow cover. We simply placed the blocks on the snow, ensuring that they fitted snugly with their neighbors. The test road is shown in Figure 6. We installed ramps at one end of the road so that the vehicle could ride up onto the blocks more easily.

Road section paved with compacted-snow blocks.

After the Jeep had driven the length of this test section 30 times, the blocks were examined for damage. None were broken. Some were slightly displaced laterally away from the tracks because of the soft base layer, whereas those at the off end of the section were displaced considerably. This could have been prevented by installing ramps at this location as well. In short, it seems clear that the blocks need a fairly solid base and a degree of lateral stabilization. Further work is planned to determine the optimum conditions for use of these blocks for paving applications. However it is clear that this technique has great potential in certain demanding paving applications in cold regions.


The use of snow roads, although widespread in polar areas, imposes certain limits on the type of machinery that can be used and on the duration of operations. Often such roads are used in wet areas that are impassable in the summer, so all land transportation occurs during winter after muskeg has frozen sufficiently to bear the weight of traffic. In extremely cold areas, such as Antarctica, the snow never melts, but the prevailing low temperatures themselves make preparation of snow roads difficult because of the low water content.

By using either of the techniques described here, it would be possible to

The increased productivity resulting from the application of these snow road enhancement methods should more than compensate for the cost of implementing them. Continued research on these and related techniques is planned; areas of interest include identifying more efficient techniques to fill the geocells, retrieving the geocell at the end of the season, emplacing the ice blocks, and improving traction of the ice surface.


The authors are grateful to Presto Products Company, Geosystems Division, of Appleton, Wisconsin, for providing the geocell used in the test described here. The authors also thank Ted Strieter and the many technicians at Keweenaw Research Center, Houghton, Michigan, who helped with this work.


1 D. Diemand, Cold Regions Research and Engineering Laboratory, 72 Lyme Road, Hanover, N.H. 03755. R. Alger, Keweenaw Research Center, Michigan Technological University, 1400 Townsend Drive, Houghton, Mich. 49931. V. Klokov, Arctic and Antarctic Research Institute, Bering Street, 38, 199397, St. Petersburg, Russia.

Back to Index       Top of Page