The objectives of this investigation were to determine the feasibility of deriving vertical velocity patterns, the magnitudes of those patterns, and the three-dimensional wind field in severe convective storms. A technique called Tracking Radar Echoes by Correlation (TREC), also called the NEXRAD Transverse Wind algorithm, and data from a single radar were used to determine the internal storm motions. The radar data were composed of Nature Severe Storms Laboratory (NSSL) Doppler data collected on 19 June 1980. TREC uses correlation analysis to determine motion within and around a given storm. TREC was used to determine the u and v wind components. These components were gridded using a standard objective analysis technique. It has been shown that in weak reflectivity regions random pattern matching can occur, therefore in areas of reflectivity less than 10 dBZ TREC-derived data were viewed with caution. Important results of this study are: 1) there was good agreement in both areal extent and magnitude between the TREC-derived vertical velocities and the dual-Doppler velocities; 2) for time increments of 5 min or less between radar volume scans, the horizontal wind field derived by TREC agreed with the dual-Doppler wind field during the multicell and transition phases of the storm; 3) the effects of the generator level motions on the TREC derived wind field were evident only in the supercell phase and; 4) vertical cross sections indicated regions of updrafts, storm-relative inflows especially in regions of updrafts, updrafts extending through a bounded weak echo region, and an apparent gust front. (rh).