Comparison of rip current hazard likelihood forecasts with observed rip current speeds

Abstract

Although rip currents are a major hazard for beachgoers, the relationship between the danger to swimmers and the physical properties of rip current circulation is not well understood. Here, the relationship between statistical model estimates of hazardous rip current likelihood and in situ velocity observations is assessed. The statistical model is part of a forecasting system that is being made operational by the National Weather Service to predict rip current hazard likelihood as a function of wave conditions and water level. The temporal variability of rip current speeds (offshore-directed currents) observed on an energetic sandy beach is correlated with the hindcasted hazard likelihood for a wide range of conditions. High likelihoods and rip current speeds occurred for low water levels, nearly shore-normal wave angles, and moderate or larger wave heights. The relationship between modeled hazard likelihood and the frequency with which rip current speeds exceeded a threshold was assessed for a range of threshold speeds. The frequency of occurrence of high (threshold exceeding) rip current speeds is consistent with the modeled probability of hazard, with a maximum Brier skill score of 0.65 for a threshold speed of 0.23 m s−1, and skill scores greater than 0.60 for threshold speeds between 0.15 and 0.30 m s−1. The results suggest that rip current speed may be an effective proxy for hazard level and that speeds greater than ~0.2 m s−1 may be hazardous to swimmers.