Gravity measurements were made with a very sensitive gravimeter in permafrost terrain containing massive ground ice and other segregated ice. Measurements were first taken along a line over undisturbed terrain where a road cut was to be made; a second gravity profile parallel to the first profile but laterally displaced from it by about 36 m was subsequently made along the edge of the roadbed after road construction. Data from pre-construction borings and a profile of subsurface soil and ice conditions, synthesized from information obtained during cutting, were used for ground-truth information and compared with the gravity measurements. The horizontal dimensions and locations of the deposits of ground ice embedded in the soil layer correlated reasonably well with the dimensions and locations of the lows in the gravity profile. However, the second profile, taken along the roadbed, also showed significant variation even after the usual types of gravity corrections were applied, suggesting that there are significant horizontal variations in the density of the topmost layers of the underlying bedrock (schist) through which the cut was made.
The density contrast of the undisturbed ice-rich soil as a function of distance along the first pro-file was estimated assuming the contrast was produced by infinitely long, transverse, rectangular blocks of given dimensions but unknown density. A set of equations dependent (to a first approximation) only on the unknown block densities was constructed from the corrected gravity data and solved by the Gauss-Seidel method. The maximum contrast for one block was found to be about 0.4 Mg m3 which gives a volumetric ice content of about 80% for the block, if the mean den-sity for all the blocks is taken to be 1.45 Mgg m3
A third gravity profile was made over an artificially-constructed ice mass with dimensions of 34 × 0.69 × 3.2 m buried at a depth of 1.2 m. This profile did not show conclusively the presence of the ice mass, partly because the anomaly it produces is close to the nominal limit of detection of the gravimeter.
It is concluded that large massive ground ice can be detected by means of its gravitational field using sensitive commercially-available gravimeters in conjunction with some ground-truth data. However, the application of such gravimeters to routine pre-construction investigations and terrain reconnaissance for ground ice is limited by their sensitivity and by the requirement for a stable measuring platform. At present, the gravity method and possibly impulse radar are the only non-contacting remote methods for obtaining an estimate of the excess ice in permafrost.