Ocean waves are known to be both random in time and nonlinear. Surface elevation time histories measured in the Gulf of Mexico during Hurricane Camille in 1969 are re-analyzed. The average shapes of large crests and deep troughs in time are shown to be close to symmetric around the instant when the maximum (or minimum) occurs, with only slight evidence of asymmetry from wave breaking in the time histories. There is considerable vertical asymmetry with higher and sharper crests and smaller and more rounded troughs. Overall, the analysis supports the use of a focused wave group based on the scaled autocorrelation function (NewWave) as proposed by Lindgren and Boccotti, with sum harmonic corrections. There is a very small second order difference setup for both large crests and troughs, consistent with considerable directional spreading in the hurricane sea-state. This spreading is likely to be larger than that usually assumed for nontropical winter storms. The spectral tail is shown to have a decay rate proportional to –4.5 power law midway between the classical JONSWAP (Phillips) –5 form and the –4 slope proposed by Battjes et al. (1987, “A Reanalysis of the Spectra Observed in JONSWAP,” J. Phys. Oceanogr., 17(8), pp. 1288–1295) as a correction to JONSWAP.

Estimation of the long‐term behavior of wave climate is crucial for harnessing wave energy in a cost‐effective way. Previous studies have linked wave heights to the north‐south atmospheric pressure anomalies in the North Atlantic, suggesting that the wave climate fluctuates as a response to changes in zonal circulation in the atmosphere. We identify changes in wave power in the North‐East Atlantic that are strongly correlated to the dominant pressure anomalies, the North Atlantic Oscillation (NAO), and other modes. We present a reconstructed wave power climate for 1665–2005, using a combination of known and proxy indices for the NAO and other modes. Our reconstruction shows high interannual and multidecadal variability, which makes wave energy prediction challenging. This variability should be considered in any long‐term reliability analysis for wave energy devices and in power scheme economics.