The Department of Defense (DoD) is currently involved in a number of unexploded ordnance (UXO) site remediation efforts where rapid transition of advanced technologies can save substantial sums of money and expedite the transfer of lands for reuse. One of the most prominent efforts is the ongoing cleanup of UXO at the bombing ranges on the island of Kaho'olawe in Hawaii. The significant magnetic anomalies from geologic sources and near-surface fragments make traditional magnetometer-based surveys impractical. Standard EM-61 metal detection surveys have also performed poorly due to the very high magnetic susceptibility response of basalt and basaltic soils. The focus of this project was to evaluate the Geonics EM-63 multigate time domain metal detector to quantify its detection, discrimination, cost, and production rates.

This project was designed to incorporate the lessons learned from previous UXO technology demonstrations and to extend the results of the Jefferson Proving Ground (JPG) Phase IV demonstrations. In evaluating the Geonics EM-63 multigate time domain metal detector, the project was divided into three main tasks. The first task was conducted at NAEVA's office in Charlottesville, Virginia and the Naval Research Laboratory (NRL) Test Field at the Army Research Laboratory's Blossom Point Research Facility in Maryland to establish an ordnance decay curve analysis database and to develop algorithms for discrimination. The second task, conducted at JPG in Madison, Indiana, was a government-sponsored blind test to determine the capabilities of the EM-63 and other electromagnetic induction (EMI) instruments under a realistic nonmagnetic clutter environment. Finally, the third task was conducted on the island of Kaho'olawe to determine detection and discrimination capabilities in a high magnetic background environment.

In areas of nonmagnetic background, the probability of detection was determined to be approximately 92 percent and discrimination of ordnance from clutter was approximately 75 percent. In areas of magnetic background, the probability of detection and discrimination of ordnance from clutter was approximately 61 percent. This is an improvement over conventional methods, but many factors influenced these results. For the Phase V demonstrations at JPG, a nonmagnetic background site, a number of objects were emplaced at depths below the detection threshold. The EM-63 is not as effective in detecting small items (e.g., 20 millimeter), and the site was not completely cleared before the demonstration began. These factors, in addition to site access difficulties, also influenced the Kaho'olawe results.

This controlled site demonstration was the first to employ realistic conditions, allowing side-by-side comparisons of discrimination performance, production rates, and costs. Acceptance of the discrimination technology (that is, not digging part of a prioritized geophysical target list) will ultimately require evaluation by regulatory agencies. Cost savings represented by this technology can be significant. If an "average" daily EM-63 survey produced 100 targets and the demonstrated technology was able to eliminate only 18 percent of those targets from excavation, the cost savings would approximate the entire cost of conducting the digital geophysical mapping and discrimination. If the demonstrated technology were able only to discriminate metal objects (both UXO and non-UXO) from magnetic rocks/soil, 32 percent of target excavations, more than 3,800 targets would have been eliminated with a savings of more than $760,000 (estimating $200 cost per dig). (Project Completed - 2003)