Sifnioti, D.E.; Dolphin, T.J., and Vincent, C.E., 2020. Performance of hindcast wave model data used in UK coastal waters. In: Malvárez, G. and Navas, F. (eds.), Global Coastal Issues of 2020. Journal of Coastal Research, Special Issue No. 95, pp. 1284–1290. Coconut Creek (Florida), ISSN 0749-0208.
Long-term wave data are a necessity for various marine applications, but wave observations can be temporally and spatially sparse, are often absent in the required locations and can be costly and often cannot be obtained in a short notice. The models used are the UK's Met Office's European 38-year wave hindcast “ReMap”, which has spatial resolution of 8 km and temporal step of 3 hours, and ECMWF's (European Centre for Medium-Range Weather Forecasts) “ERA5” reanalysis dataset from 1979 to present with a spatial resolution of 31 km and a temporal step of 1 hour. Both datasets include wave and wind parameters, thus providing valuable long-term data that can be used in coastal regions where wave observations are unavailable. In this paper we assess the errors in performance of the ReMap and ERA5 data by comparing these with wave observations made at two coastal sites in the southern North Sea. ReMap and ERA5 showed a strong correlation with measured waves for significant wave height (Hs) (r = 0.84 - 0.9), fair correlation for peak wave period (Tp) (r = 0.58 - 0.68) and good correlation for mean zero up-crossing wave period (Tz) (r = 0.73 – 0.79), with different error statistics per location and per dataset. Analysis of the 90th percentile shows that the ReMap and ERA5 overestimated Hs by approximately 0.5 m, and by up to 1.2 s for Tp and 0.6s for Tz. Monthly and annual means are also presented and discussed. Overall the performances of ReMap and ERA5 against the measured waves were similar. It is concluded that the ReMap and ERA5 data can be used with reasonable confidence in coastal regions where no direct measurements are available for ‘average’ conditions, but caution needs to be taken for extreme conditions as an over prediction for extremes is likely.