The basic mechanics of ice roads have been known for decades. As a laden truck moves over ice, it creates a shallow depression all around it \ufffd a sort of bowl in the ice, several inches deep and many yards across. The greater the speed, the deeper the depression. Above a critical velocity that varies with local conditions, a truck can damage the roadbed so severely that the next vehicle to come along will break through the ice. For this reason, the top speed on the Tibbitt to Contwoyto route is usually about 22 mph. In some stretches, as on Waite Lake, the maximum is just a few miles per hour.

As the concavity moves along beneath the truck, the displaced water forms pressure waves under the surface. When these waves reach the lakeshore, they bounce back, and the resulting interference with incoming waves can push the ice up and create a blowout. Whenever possible, engineers build ice roads at an angle to nearby shorelines to prevent blowouts under the roadbed itself.

Every day, a crew drives along the road, generating a profile of ice thickness with ground-penetrating radar. They drag an antenna, which sends 400-MHz radio waves down into the lake. A computer calculates the ice thickness based on the time it takes the signal to return from the ice-water interface. For even greater accuracy in calculating how much weight the road can bear, the data is calibrated against measurements taken from boreholes that workers drill with power augers.

One of the men driving the road with us today is Sam Proskin, a geotechnical engineer whose firm has been hired to help better understand what's going on beneath the ice. Proskin is trying to develop multidimensional maps of problem spots by using more sophisticated radar. This, he hopes, will help solve a vexing mystery: how water flow, sandbars and other subsurface features affect the underside of the ice. Such an understanding could help engineers re-route the road to minimize problems like the one here on Waite Lake.

"The physical properties of ice are well understood under laboratory conditions," Proskin says, "but when you extrapolate it into the scale of kilometers of road, it's a little more difficult to understand. It changes, it flows, it gets brittle. The ice is like a living thing."

For the roadbuilders, that means working with phenomena for which there is limited scientific understanding. For example, driving at a speed and weight that is appropriate for an ice sheet makes the ice stronger. Truckers say that they're "driving down the frost," but there's no evidence that such a physical process takes place: It's just another mystery of the ice. "Ice is an engineering challenge," Proskin says. "It's a solid near its melting point. If you heated steel to near its melting point and put loads on it, it would act strangely, too."