Extremes and Trends in Wave Climate. A regional and global study
MetadataShow full item record
- Geophysical Institute 
Wind generated surface waves represent a critical factor for offshore constructions and coastal development, and are highly relevant in scientific questions related to climate and weather. The sea state can be described by several parameters, but significant wave height is particularly important, traditionally defined by the average height of the highest third of individual waves. This study targets two aspects describing wave climate: extremes and trends in significant wave height. The Norwegian Reanalysis (NORA10) - a combined high-resolution atmospheric downscaling and wave hindcast based on the ERA-40 reanalysis covering the Norwegian Sea, the North Sea, and the Barents Sea is presented and validated. The wind-wave data archive, spanning the period September 1957 to August 2002, shows a significant improvement from ERA-40 over the whole range of data, but particularly in upper percentiles. Given the performance of NORA10, it provides a baseline climatology of significant wave height, which other datasets may be compared against. An extended version of NORA10, covering 1958 to 2009, is utilized to obtain 100- year return value estimates of significant wave height. The analysis explores three different approaches, where the applied extreme value distribution is dictated by the chosen subset of the initial data. This is done by: (i) peaks-over-threshold; (ii) annual maxima; and (iii) the r largest order statistic within blocks of one year. These subsets should conform to the generalized Pareto distribution (i) and the generalized extreme value distribution (ii/iii), respectively. By assuming stationary conditions, the three models provide results mainly within ±5%. In areas where the discrepancy is larger, (iii) is found less satisfactory. Method (i) yields good conformity when the threshold is set high. Here, the 99.7-percentile is applied. Given the better use of data, (i) is preferred over (ii). Based on (i), the 100-year estimates are peaking around 22 m northwest of Scotland, around 14 m in the North Sea and above 16 m in the Norwegian Sea. The robustness of a return value estimates often depend on the amount of available data. As most conventional time series of significant wave height do not extend further than 50 years, 100-year return value estimates can only be obtained by extrapolating some fitted theoretical distribution. In this study, a method for estimating return values from aggregated ensemble forecasts is presented. The archived ensemble forecasts originate from the European Centre for Medium-Range Weather Forecasts (ECMWF) and consist of 51 members run twice daily. To ensure that the aggregates are independent representations of the model climate, only 10-day forecasts are retained. By assuming each forecast being representative of a 6-hour interval, collectively 1 year of ensemble forecasts are representative of more than 25 years of data. In two separate papers, aggregates of ensemble forecasts, equivalent to more than 220 years of data, are utilized to obtain 100-year return value estimates of significant wave height and wind speed. The datasets are carefully validated against in situ measurements and altimeter data, and found representative of the observed climate. Return value estimates are obtained by traditional extreme value models, but also obtained directly from order statistics as the dataset, on average, should contain more than 2 events exceeding the 100-year return value. It is shown that the estimates come close in equalizing corresponding estimates from NORA10, and yield significant improvements compared to ERA-40 and ERA-I. Reanalyses may be considered homogeneous in the sense that they run with the same model configuration. However, they are highly dependent on assimilation and may suffer from the ever growing observational system. Ideally, assimilation should only be a mean to reduce random errors, but if a model exhibits systematic errors, data assimilation may correct bias. Once a model is run in forecast mode, the effect of assimilation is usually lost, as the model relax towards it biased state. ERA-I is a coupled wind-wave reanalysis produced at ECMWF, spanning the period 1979 onwards. Besides the reanalysis, ERA-I is run as 10-day forecasts. Altimeter observations are the only kind of wave data assimilated, first introduced in August 1991. Here, trends based on different forecast ranges are compared and validated against observations. It is shown that trends in significant wave height from analysis are highly affected by the transition in August 1991, especially in the northeast Atlantic and the eastern tropical Pacific. Finally, the 48-hour forecast range is proposed as a better candidate to obtain realistic trends.
Has partsPaper I: Magnar Reistad, Øyvind Breivik, Hilde Haakenstad, Ole Johan Aarnes, Birgitte R. Furevik and Jean-Raymond Bidlot. A high-resolution hindcast of wind and waves for the North Sea, the Norwegian Sea and the Barents Sea. J. Geophys. Res., 116, C05019, 2011. The article is available at: http://hdl.handle.net/1956/9251.
Paper II: Ole Johan Aarnes, Øyvind Breivik and Magnar Reistad. Wave Extremes in the Northeast Atlantic. J. Climate, 25, 1529–1543, 2012. The article is available at: http://hdl.handle.net/1956/9253.
Paper III: Øyvind Breivik, Ole Johan Aarnes, Jean-Raymond Bidlot, Ana Carrasco and Øyvind Sætra. Wave Extremes in the Northeast Atlantic from Ensemble Forecasts. J. Climate, 26, 7525–7540, 2013. Full text not available in BORA due to publisher restrictions. The article is available at: http://dx.doi.org/10.1175/JCLI-D-12-00738.1.
Paper IV: Øyvind Breivik, Ole Johan Aarnes, Saleh Abdalla, Jean-Raymond Bidlot and Peter A.E.M. Janssen. Wind and wave extremes over the world oceans from very large ensembles. Geophys. Res. Lett., 41, 5122–5131, 2014. The article is available at: http://hdl.handle.net/1956/9254.
Paper V: Ole Johan Aarnes, Saleh Abdalla, Jean-Raymond Bidlot and Øyvind Breivik. Marine wind and wave height trends at different ERA-Interim forecast ranges. J. Climate, 28, 819–837, 2015. The article is available at: http://hdl.handle.net/1956/9255.