Climate change signal and uncertainty inCMIP5‐based projections of global ocean surface wave heights

This study uses the analysis of variance approaches to quantify the climate change signal and uncertainty in multimodel ensembles of statistical simulations of significant wave height ( H s ), which are based on the CMIP5 historical, RCP4.5 and RCP8.5 forcing scenario simulations of sea level pressure. Here the signal of climate change refers to the temporal variations caused by the prescribed forcing. “Significant” means “significantly different from zero at 5% level.” In a four‐model ensemble of H s simulations, the common signal—the signal that is simulated in all the four models—is found to strengthen over time. For the historical followed by RCP8.5 scenario, the common signal in annual mean H s is found to be significant in 16.6% and 82.2% of the area by year 2005 and 2099, respectively. The global average of the variance proportion of the common signal increases from 0.75% in year 2005 to 12.0% by year 2099. The signal is strongest in the eastern tropical Pacific (ETP), featuring significant increases in both the annual mean and maximum of H s in this region. The climate model uncertainty (i.e., intermodel variability) is significant nearly globally; its magnitude is comparable to or greater than that of the common signal in most areas, except in the ETP where the signal is much larger. In a 20‐model ensemble of H s simulations for the period 2006–2099, the model uncertainty is found to be significant globally; it is about 10 times as large as the variability between the RCP4.5 and RCP8.5 scenarios.