Wave hindcasts are often treated as fixed: run once, extract statistics, move on. But metocean design depends on how well the extremes are being represented in your data, and that can change over time as new storms enter the record.
We ran a simple test using our 45-year wave hindcast for Auckland (New Zealand) and Callao (Peru). The question was straightforward: how does the 99th percentile of Hs (P99) change as we keep extending the time series with new monthly data? Each bar in the figures shows the cumulative shift in P99 relative to a baseline of 1979 to 2024. As new months are added, we see significant changes occurring.
At Auckland, P99 rose to +2.7 mm in late 2024, then fell to −8.8 mm by August 2025 as calmer months dominated the tail of the distribution.
At Callao, P99 jumped +8.1 mm in early 2024, then dropped to −3.0 mm by mid-2025, reflecting a strong seasonal Pacific swell variability.
Using the same methodology and statistics, very different trends are uncovered. They may seem small, even millimetre-level shifts in high percentiles can propagate into design statistics, freeboard and crest elevation decisions, downtime thresholds, and safety margins – especially when embedded in joint extremes or return-period fits. That’s why we update our hindcasts monthly, to allow the tail statistics to evolve as the record grows.
When was your metocean dataset last updated? Want to see how your extremes evolve as new data is added? Our monthly-updated hindcast workflows are available here (link in the first comment)
#WaveHindcast #CoastalEngineering
