A recent upward looking sonar profile survey of a ~10-km by 10-km area beneath the Arctic ice serves as a basis for an exploratory analysis of the quantitative charactertics of sea ice draft. In a companion paper (Goff, this issue) a method was developed for estimating profile statistical parameters and their uncertainties. These include mean draft, rms variation, characteristic length, fractal dimension, and normalized skewness. Here this methodology is applied to the intersecting profile data set, yielding 160 separate estimates for each parameter. Although not completely two-dimensional, the data nevertheless allow an opportunity to investigate some aspects of topographic anisotropy. Our most significant observations are (1) a strong positive correlation exists between fractal dimension and charactertic length, while normalized skewness is weakly negatively correlated with both parameters; (2) high and low rms/mean ice are morphologically distinct in the rms versus characteristic length and fractal dimension versus characteristic length parameter spaces; (3) the overall correlation between rms and characteristic length can be explained entirely by the difference between high and low rms/mean type morphology; (4) anisotropy appear to exist on a local scale but is highly variable over the entire survey. These observations could be explained by a systematic variation in morphology with age, perhaps including a progressive superposition of deformation events, and/or by regionally variable anisotropy. ¿American Geophysical Union 1995