Reported extreme wave heights off Ireland are artifacts

I read with interest the recent PNAS paper on Irish and New Zealand boulder transport (1), and agree with its conclusion that the Irish boulders clearly arise from storm wave action. However, there are errors in the analysis that must be addressed: in particular, descriptions of waves with significant heights H_s = 26.5 m and H_s = 22.34 m at the Wave Rider Belmullet Berth B site off the Irish coast. These waves would indeed be large, with significant heights greater than have ever been recorded at any location (2, 3). However, closer examination of the data suggests that these particular extreme values might best be described as artifacts. Fig. 1 shows results from the more complete dataset (freely available at https://www.marine.ie/Home/home) around the times of both extreme wave values: January 15, 2015 (Fig. 1A) and January 26–27, 2014 (Fig. 1B). Both extreme H_s values are large departures from the longer time series and are 2.7 and 1.7 times the immediately adjacent values. This is troubling, but not entirely conclusive. Additional reported data for H_max (generally taken as the largest single crest–trough wave height in a record) and H_1/3 (which is usually synonymous with H_s; it is not clear why separate values are given) tell a different story: H_1/3 values are very close to H_s values, except at the times of anomalously large wave heights. On January 27, 2014, the value of H_s = 22.34 m recorded at 1:02 AM, is 2.09 times greater than the H_1/3 = 10.71 m value, recorded just 2 min earlier. On January 15, 2015, the ratio is 2.57 at 15-min separation. These values are essentially impossible and suggest hardware or software issues.

Fig. 1.

Recorded H_s (black circle), H_1/3 (red square), and H_max (blue diamond) at Belmullet Berth B wave buoy on (A) January 15, 2015 and (B) January 26–27, 2014.

Recorded H_max values around the times of the H_s anomalies also show strong departures from statistically expected values: H_max/H_1/3 ratios of 2.97, 3.71, and 3.99 were recorded on January 15, 2015, and a corresponding ratio of 3.23 is found for January 27, 2015. Rayleigh distribution statistics and decades of wave-measurement experience (4, 5) suggest that the probability of these ratios for maximum wave heights is vanishingly small. The only remaining conclusion is that the wave records for this buoy contain intermittent errors for reasons that are not yet known and that the extreme published H_s values on these two dates, and potentially on others, are instead artifacts that should be discarded. For the authors’ (1) analysis, this will likely decrease significantly the size of the largest waves available to transport boulders.