Using 12 years of frequent high precision profiles collected on a wave-dominated sandy oceanic beach to 8-m depth, the characteristics and interpretation of depth of closure — or the seaward limit of significant profile change — is critically examined. This includes evaluation of the predictive capability of d 1 — the seaward boundary of the littoral zone — as originally defined by Hallermeier (1981) [Hallermeier R.J., 1981. A profile zonation for seasonal sand beaches from wave climate. Coastal Eng. 4, 253–277.]. Depth of closure during major erosional events is usually produced by the seaward limit of offshore bar movement. Following the original recommendation of Hallermeier (1981), d 1 based on 12 h exceeded wave height and a reference depth of mean low water provides a robust limit to the observations using a 6-cm change criterion. Empirically, the observed depth of closure is 69% of d 1, although the scatter is large. This scatter is partly controlled by pre-event profile shape, most particularly bar configuration. Depth of closure under accretional conditions can also be measured, but it is time-scale-dependent as accretion is a slow, steady process, d 1 underpredicts closure for accretional situations. Time-interval (e.g., annual) depth of closure represents the integral effect of erosional and accretional events. An important observation of the data is that depth of closure increases with time scale. However, the full population of time-interval observations is not resolved at Duck due to the measurement limit (8-m depth). As time interval increases from 1 year to 8 years, less cases close. Most non-closing time-interval cases coincide with the periods influenced by the most energetic wave events. However, time-interval closures are generally deeper than the biggest event closure in the period, showing that it represents more than the largest event. The frequency distribution of the data suggests that most, if not all, of these missing data simply represent closures deeper than 8-m depth. Up to a 4-year time interval, d 1, t appears to provide a reasonable limit to the quantified observations. It is an interesting result that d 1 (an event-based approach) might provide a limit to closure over periods up to 4 years as these time-interval closures are generally larger than those produced by single storms. Therefore, the general applicability of this result for more complete data sets, or on rapidly evolving (i.e., eroding or accreting) coasts is uncertain. The present data also suggest that on swell-dominated coastlines where accretional processes are dominant, d 1 may underpredict closure, suggesting an important limitation to this approach. But as sediment would be moving onshore, this may not be a practical problem. Therefore, within the limits of the data set, Hallermeier's (1981) approach is found to define robust estimates of depth of closure, particularly for individual erosional events. This useful result is expected to find widespread application in coastal geology and engineering.