New Zealand rip forecast matches 53 per cent of lifeguard ratings
A peer-reviewed evaluation of New Zealand’s rip-current forecast compared five warning levels with lifeguard observations. Exact agreement reached 53 per cent; allowing a forecast one level more cautious raised the figure to about 81 per cent. The algorithm combines waves, tides and beach shape. Historical rescue data from 58 beaches helped distinguish crowded swimming conditions from greater danger for each person entering the water.
Science··Night
Five warning levels were compared with lifeguards
Across 12 New Zealand beaches, 568 lifeguard assessments provided a comparison for the rip-current algorithm during summer 2024–2025. The hazard ratings matched exactly in 53 per cent of cases. Counting forecasts one level higher than the lifeguards’ ratings raised conservative agreement to about 81 per cent. The peer-reviewed evaluation, published on 16 November, assesses a five-level warning system for currents that can carry swimmers offshore.[1]
Rescue records were weighed against water use
Lifeguards supplied roughly 9,000 rescue records and water-user counts from 58 beaches during 2001–2022. Including the number of people exposed helped distinguish a busy day from conditions associated with greater danger per swimmer. Warning conditions encompassed 98 per cent of past incidents under the algorithm. That describes incident capture, without establishing a low false-alarm rate. Missing rescue times and estimated swimmer counts contributed uncertainty.[1]
Waves, tides and beach shape drive the forecast
The calculation combines breaking-wave height, tide level and whether sandbars are separated by channels. At beaches with that shape, average or larger waves together with lower tides were particularly important. It uses simple thresholds rather than directly simulating current speed. Internal use has begun within the organization, which is preparing to offer the algorithm to the public. Warning comprehension and actual safety outcomes require separate evaluation as the forecast moves toward wider use.[1]