Global evaluation of wave data reanalysis: Comparison of the ERA5 dataset to buoy observations

Wave reanalysis data are critical to multiple industries including commerce, fisheries, petroleum, offshore wind, the design and the construction of maritime structures. One of the most used datasets is ERA5 developed by the European Centre for Medium Range Forecast. The accuracy and reliability of this dataset have not been rigorously assessed on a global scale. We present the first global evaluation of ERA5 significant wave height against measurements from 444 buoys worldwide. We demonstrate the influence of critical control factors including distance to shore, water depth, regional characteristics and seasonality. Results showed that ERA5 underestimates significant wave height in extreme conditions. This underestimation could lead to underestimating the severity of predicted wave conditions when designing vessels or foundations for maritime structures. ERA5 accuracy increases farther from shore, but does not change significantly with increasing water depth. In the Northern Hemisphere, the RMSE and absolute bias between ERA5 and measured significant wave height are up to 0.25 m lower in summer than in winter. Seasonal patterns are less pronounced in the Southern Hemisphere. The results provide global confidence levels for the robust use of ERA5 data, and establish a methodology to test future datasets.