Statistical analysis of global ocean significant wave heights from satellite altimetry over the past 2 decades

Abstract. The analysis of global ocean surface waves and of long-term changes requires accurate time series of waves over several decades. Such time series have previously only been available from model reanalyses or from in situ observations. Now, altimetry provides a long series of observations of significant wave heights (SWHs) in the global ocean. The aim of this study is to analyse the climatology of significant wave heights and extreme significant wave heights derived from remote sensing in the global ocean and their long-term trends from 2002 to 2020 using different statistical approaches as the mean, the 95th percentile, and the 100-year return level of SWH. The mean SWH and the 95th percentile of SWH are calculated for two seasons: January, February, and March as well as July, August, and September and for each year. A trend is then estimated using linear regression for each cell in the overall grid. The 100-year return levels are determined by fitting a generalized Pareto distribution to all exceedances over a high threshold. The trend in 100-year return level is estimated using the transformed stationary approach, which, to our knowledge, is used for the first time to draw a global map based on altimetry. Predominantly large positive trends over 2002–2020 for both SWH and extreme SWH are mostly found in the Southern Hemisphere, including the South Atlantic, the Southern Ocean, and the southern Indian Ocean, which is consistent with previous studies. In the North Atlantic, SWH has increased poleward of 45° N, corroborating what was concluded in the Fifth IPCC Assessment Report; however, SWH has also largely decreased equatorward of 45° N in wintertime. The 100-year return levels of SWH have significantly increased in the North Atlantic and in the eastern tropical Pacific, where the cyclone tracks are located. Finally, in this study we find trends of SWH and 95th percentile of SWH over 2002–2020 to be much higher than those indicated in the literature for the period 1985–2018.