The purpose of this study was to identify the currently available nondestructive evaluation technology that holds the greatest potential to detect moisture in flexible pavements and then apply the technology in multiple locations throughout Virginia. Ground-penetrating radar (GPR) was chosen for use in a field investigation because of its ability to measure large areal extents and reports of successful implementation by other researchers. This technology was used to determine the moisture content of the subgrade beneath five flexible pavement sections in Virginia. The GPR survey was conducted at normal driving speeds, and data were collected at a sampling rate of 1 scan per foot. For each site, three scans were collected in the travel lane (in the right wheel path, the center of the lane, and in the left wheel path). Existing passing lanes were also scanned. Initial data processing subdivided each pavement section into a three-layer system composed of the hot-mix asphalt layers, the aggregate base layers, and the subgrade. The processing also included calculating the dielectric constant of each layer. These raw data were used to conduct further analyses considering data from only the subgrade. The data were normalized to highlight those areas with the highest dielectric constants since it is known that moisture will have the greatest influence on the dielectric properties of the material. This study showed that GPR can identify areas of varying dielectric constant attributed to variations in the moisture content of the subgrade of various pavement sections. The use of the GPR offered a safe and rapid means for nondestructively surveying large areas of pavement as the survey was conducted at normal driving speeds. In addition, the use of a statistically based data normalization procedure allowed GPR to be used to assess qualitatively the moisture condition of the subgrade of flexible pavements. Two advantages of GPR testing are that it can be used to provide a continuous reading of subgrade moisture conditions (rather than a point location) and can be performed at highway speeds with no traffic control. A typical network-level study involving one subgrade bore per mile would cost $1,200 per location. The cost for the current study averaged $0.04 per data point. Assuming one data point every foot and three scans per lane, the resultant cost was approximately $680 per mile. VDOT maintains approximately 225 lane-miles of flexible pavement with subsurface drainage layers and does not routinely inspect the condition of the outlet pipes. In approximately 1 workday, a two-person crew could use a push camera to inspect approximately 5 lane-miles worth of outlet pipes for subsurface drainage layers or the GPR system could scan approximately 330 lane-miles (assuming operation at 55 mph for 6 hours). Thus, the entire system of flexible pavement with subsurface drainage could be inspected in approximately 2% of the time.