This report presents results of analyses of coordinated radar-aircraft data acquired form the 1983 experiment conducted at Hanscom AFB, Massachusetts. The objective of the experiment is to assess and validate the current NEXRAD algorithms for estimating aircraft turbulence from volume-scanned Doppler weather observations. Estimates of the turbulence severity index epsilon to the 1/3 power (a quantity used by NEXRAD) computed from radar and aircraft data are presented as a time series along each aircraft track. The radar point estimates of turbulence were averaged horizontally and vertically to yield layered Cartesian maps such as are intended for use by real time ATC controllers and pilots. The derived gust velocity (Ude), also used to indicate the intensity of aircraft encountered turbulence, was computed so that comparisons could be made of the turbulence intensity scales inferred from values of epsilon and U sub de. These quantitative comparisons indicate that for the turbulence generally encountered during the flights, both radar and aircraft estimates of epsilon to the 1/3 power significantly overstate the severity of turbulence as reported by the aircraft pilot. The data analysis also shows that radar-based estimates of epsilon to the 1/3 power, often significantly exceeded aircraft based estimates of epsilon to the 1/3 power. In contrast, the quantity Ude underestimates the aircraft reported turbulence intensity on all the flights. The uncertainty as to operationally useful thresholds for radar epsilon to the 1/3 power, aircraft epsilon to the 1/3 power and Ude is discussed as is the use of spectrum width as a turbulence indicator. It should be noted that the turbulence detection flights used in the study were conducted at ranges such that the radar resolution cell cross range extent was typically 1.5 to 3 km. With such resolution cell size extents, the hypothesis of spatially homogeneous turbulence may not hold and/or the assumed relationship of radar measured spectrum width to kinetic dissipation rate may not be fully accurate.