The Hammer site included a variety of different objects buried at different levels, including metal objects, wooden objects, earthen "hearths," and brick. Objects were covered with homogenous sand.

Simulated model of buried features at the Hammer test site.

View of features at Hammer test site that will be covered with sand.

See the analysis of object identification with different antennas and conditions at the Hammer test site:

400 MHz:

Dry conditions: Surveys proved to be moderately successful. The metal and larger objects showed up easily, but wood and less-compact earthern features were essentially invisible.

Wet conditions: Surveys showed a large number of extraneous reflections, which are likely pockets of water. This may have been due to the artificial saturation of the site with a sprinkler, which may have not allowed enough time for water to percolate down. The wood objects showed up well when saturated.

900 MHz:

This antenna was better at reflecting smaller objects in this homogenous soil. At times, however, reflection profiles appeared too "cluttered." This is because this higher frequency antenna can resolve much smaller objects, and therefore resoloves much more. This was especially true in wet conditions, when the antenna likely reflected every small pocket of water.

900 MHz in Dry conditions.

Spatial Statisical Analysis

Correlations were calculated to compare the objects and their shapes to the amplitude maps' objects and shapes. Two spatial correlations were made:

1) To determine where GPR reflections were consistent with the model of the object's location and shape.

2) To determine where the model showed a buried object in a known place and GPR showed it at the same place (clutter was discounted).

400 MHz wet; 400 MHz dry

900 MHz dry; 900 MHz wet

Both wet and dry conditions with the 400 MHz antenna showed fair correlations. Water retention in the shallow amplitude slice maps lowered this correlation. The correlation for wet wood is good. In dry conditions, reflections were "smeared" to the south (transects were collected in a north-south manner), which also lowered the correlation.

Both wet and dry conditions with the 900 MHz antenna proved to be far superior to the 400 MHz antenna. The higher resolution of the 900 MHz antenna resulted in higher correlations.

Conclusions:

The Hammer test data demonstrates the importance of water in resolving buried features. When the ground was dry, most buried objects were visible (except the wooden objects). The wood may not have contrasted enough with the surrounding sand to be visible in dry conditions. When the ground was saturated, many of the objects that contrasted with the sand when dry are often invisible. This is because the water in the system tends to "blend" and "homogenize" the matrix and the artifacts. In contrast, the wooden objects are quite visible in wet conditions, as they have acted as absorptive agents for the water (like large sponges), and therefore their very high water content produces distinct radar reflections along interfaces with the surrounding sand.

These results indicate that muliple GPR surveys may be essential when ground conditions change during different seasons or moisture regimes. Also, some materials reflect radar energy in very different ways in dry and wet conditions.