The distance over which planktonic larvae are dispersed and the variability within that dispersal distance are important for understanding gene flow and persistence of species with planktonic larvae in the coastal ocean. The breadth of spatial and temporal scales that can be important to dispersal in shelf seas makes direct observations of larvae, and thus these statistics difficult - instead, we often use numerical simulations of circulation to estimate the statistics of larval dispersal. However, meroplanktonic life histories are most common in coastal regions where drifter-based estimates of circulation are sparsely distributed, making validation of these numerical simulations quite difficult. We use a novel technique to validate climatological mean and standard deviation of dispersal distance at a global scale by drawing on the tens of thousands of sparsely distributed drifter observations on the shelf. Numerical dispersal estimates were made using Lagrangian particle trajectories calculated with circulation fields from a 1/12° global physical model and were validated against drifter data from the Global Drifter Program (GDP). The median dispersal distance of a climatological ensemble of numerical drifters released from a single location were found to match GDP drifter estimates quite well (with a mean deviation of 0.2%), while model estimates of dispersal were shown to underestimate the diffusivity of GDP drifters by 30 to 50%. Our results indicate that while global numerical estimates of dispersal statistics provide a close approximation of median dispersal distance in the coastal ocean, these numerical simulations underestimate the overall variation present in the coastal ocean.