The ability to safely and efficiently locate, identify, and remove buried objects on practice and test ranges is critical to the U.S. Department of Defense’s (DoD) mission and its goals of safe operation, sustainability, and environmental stewardship. Although several robust, advanced sensor technologies have been developed for discriminating buried objects on ranges as targets of interest (TOIs), widespread acceptance of the technologies requires that they be demonstrated on live sites, where the impact of operating and data acquisition and analysis parameters can be fully evaluated. The potential benefit of the technology is to reduce the number of subsurface investigations that are required to remove hazardous munitions and explosives of concern (MEC) in areas where subsurface clearance is required.

The overall objective for both sites used in the project (Fort Ord and RSA) was to demonstrate and validate Advanced Classification (AC) techniques in challenging site conditions. At Fort Ord, the primary challenge was to detect and classify munitions in areas with high subsurface metallic clutter. The primary challenge at RSA was to collect high-quality advanced sensor data in highly wooded areas that could be used for classification. Each site also had additional unique objectives that were determined in concurrence with CB&I, the U.S. Army Corps of Engineers (USACE), and project stakeholders during the project planning phase. The objective at both sites was also the knowledge transfer from research organizations (Black Tusk Geophysics [BTG], Acorn Science & Innovation, Inc. [Acorn], Geometrics, and U.S. Naval Research Laboratory [NRL]) to contracting companies such as CB&I.

Both the MM and TEMTADS 2x2 are time domain electromagnetic (TDEM) sensors, making use of three-dimensional receiver coils, with seven and four receiver cubes, respectively. In the larger MM sensor, transverse excitation of subsurface metallic objects is accomplished explicitly with three orthogonal transmitter coils. In the smaller TEMTADS sensor, transverse excitation was accomplished implicitly via the four vertical transmitter coils at offsets from the sensor center.

In general, the primary objectives for Fort Ord were met. CB&I successfully collected high-quality dynamic and cued data using the MM and were able to meet the performance objectives for data quality. All large munitions were correctly classified as TOI. However, one of the large munitions items recovered (155 millimeter [mm] projectile) was classified as a smaller munition as defined for this project. The objective for correctly classifying the smaller munitions was met but the final dig list for the smaller munitions included a high percentage of incorrectly classified non-TOI items.

CB&I successfully collected both dynamic and cued data within heavily wooded areas encountered at Redstone. Four of the blind quality control (QC) seeds were not classified as high confidence TOIs as per the performance objective but were included in the final dig list as “can’t decide” digs. Of the 26 additional TOI items recovered, 19 of these were classified as high confidence TOIs, while the remaining 7 were classified as “can’t decide” digs.

The most significant implementation issues during both data collection at Fort Ord and RSA were associated with the initial condition of the MM and TEMTADS units when received by CB&I, likely due to the age of both sensor systems during 2014 and 2015 when field work was undertaken. The MM unit required multiple repairs for transmitter boards and wiring. The initial TEMTADS unit from NRL was malfunctioning, and the electronics unit was swapped with a unit recently demobilized by another contractor. During field work at RSA, rain was common, and instructions from NRL staff were to avoid any rain, which created significant delays in initial testing and data collection. At Fort Ord and RSA, complications with data processing software related to the maturity of the software and the experience level of CB&I analysts, which resulted in portions of the data being processed and reviewed after data collection was completed. As a result, immediately administering corrective measures based on QC metrics was not possible.