Observations of mixing across mare-highland contacts using high-resolution Clementine data allow a new assessment of the relative importance of vertical versus lateral mass transport on the Moon at least at mare-highland contact. We analyze mare-highland contacts in the Grimaldi, Orientale, and Fecunditatis basins and Tsiolkovsky crater through image-based nonlinear spectral mixture modeling of Clementine ultraviolet-visible (UV-VIS) light spectrometer multispectral data. The mare in these regions differ in ages, but they are characterized by having simple geological contacts. The symmetric distribution of mare and highland soil constituents across their geological contact indicates that lateral transport is more efficient than deep vertical transport during the formation of the observed mixing zones. Observations across the lunar limb with the Clementine UV-VIS filters show that there is a stray light component of scattered light across high-contrast albedo boundaries. However, the magnitude and spatial properties of scattered light do not significantly affect the nature of compositional boundaries determined here. Since repetitive meteorite bombardment governs lateral transport across the geological contacts and is a random process, a stochastic model is developed to describe this lateral transport. Analysis of the power decay law for crater ejecta thickness indicates that the high-velocity ejecta travels long distances and follows a -3 power decay law. This results in an infinite variance, which requires an anomalous diffusion model since a classical diffusion model would be invalid. Mathematical modeling supports this result where it is shown that classical diffusion produces a relatively poor fit to profiles of material abundance, while the anomalous diffusion model fits the profiles adequately. These results indicate that high-velocity ejecta dominates the formation of the observed compositional gradients, while the ejecta near the crater rim is relatively less important. On the basis of these ejecta distributions and the assumption of random impact cratering, we derive a relationship that can be used as an index for relative age dating of geological contacts as well as for investigating cratering rates.