Anne Karin Magnusson

On estimating extremes in an evolving wave field

Abstract Observations of waves as a time series from a fixed or moored sensor are shown to underestimate the extreme waves occurring in the vicinity. The underestimate arises because of the difference in phase and envelope propagation of surface water waves so that the highest crest, for example, is recorded only when the crest coincides with the envelope maximum at the location of the wave recorder. In addition, the dispersive nature of water waves can lead to the coalescence of groups of waves as longer waves catch up to shorter slower waves, so that the group shape changes markedly as the waves propagate. Moored accelerometer buoys introduce another error that exacerbates the underestimate of the highest crest—the quasi-Lagrangian motion of the buoy leads to flattening of the crests and sharpening of the troughs so that apparent mean water level is raised and thus the height of crests above it is reduced. An analysis approach to correct these underestimates is outlined and tested with observed data against the predictions of extreme values based on narrow-banded random wave theory.

Projected changes in significant wave height toward the end of the 21st century: Northeast Atlantic

Wind field ensembles from six CMIP5 models force wave model time slices of the northeast Atlantic over the last three decades of the 20th and the 21st centuries. The future wave climate is investigated by considering the RCP4.5 and RCP8.5 emission scenarios. The CMIP5 model selection is based on their ability to reconstruct the present (1971–2000) extratropical cyclone activity, but increased spatial resolution has also been emphasized. In total, the study comprises 35 wave model integrations, each about 30 years long, in total more than 1000 years. Here annual statistics of significant wave height are analyzed, including mean parameters and upper percentiles. There is general agreement among all models considered that the mean significant wave height is expected to decrease by the end of the 21st century. This signal is statistically significant also for higher percentiles, but less evident for annual maxima. The RCP8.5 scenario yields the strongest reduction in wave height. The exception to this is the north western part of the Norwegian Sea and the Barents Sea, where receding ice cover gives longer fetch and higher waves. The upper percentiles are reduced less than the mean wave height, suggesting that the future wave climate has higher variance than the historical period.