Permafrost thermal amelioration techniques have been tested for 4 years near Schefferville (mean annual temperature −6 °C). The most important amelioration measure tested was the use of snow fences to increase snow depth and hence reduce winter heat losses, which affect the permafrost much more than changing the summer heat input conditions. However, when different summer treatments were tested, stripping the vegetation, darkening the ground and using thin transparent covers proved beneficial. Monitoring included deep ground temperature measurements (to 25 m), radiation instruments, lysimeters, and measurement of ground thermal properties. At 5 m depth, between 20 and 40 × 106 Jm−2 are gained and lost each year under natural conditions, whereas the amelioration gave a continuous gain of 20 × 106 Jm−2 per annum (nearly 2% of net radiation), increasing mean temperature by 2.5 °over 4 years. At the 10 m depth corresponding gains were 10 × 106 Jm−2 and 1.5°. The rocks (iron-rich) have high thermal conductivities, and slower amelioration results are likely in other areas. Limited plot size (7500 m2) has resulted in significant lateral heat loss. The overall test result is that the active layer has been greatly deepened but no thaw has yet been held through a full winter. The method would be useful in several mining applications, including the prevention of new permafrost.