Reconstruction of directional wind wave spectra from visual ship-based observations

For the first time, directional wave spectra are reconstructed from visually observed wind wave data over the North Atlantic Ocean. For this purpose, a novel analytical approach for calculating the directional spreading function from visual ship-based observations is proposed. Natively and inherently separated estimates of wind sea and swell heights, periods, and directions of propagation provide independent directional spreading functions for wave systems in a given area and time. Shape parameter and the mean angle are evaluated from spatially and temporally averaged sine and cosine projections of wave directions. The calculated directional spreading functions combined with frequency spectra of wind sea and swell allow for two-dimensional spectra reconstruction from visible wave elements on different temporal and spatial scales. The new approach was applied to visual wave observations in the North Atlantic for the period of 1970-2023. Visual wave observations were taken from the ICOADS (International Comprehensive Ocean-Atmosphere Data Set), consolidating all available observations by Voluntary Observing Ships (VOS). Directional wave spectra were reconstructed in two spatial grids: 10°x10° and 1°x1° with a temporal resolution varied from one day to climatological month. The results were intercompared to the directional spectra derived from the long-term model hindcasts of wind waves performed with WWIII spectral wave model for both individual daily spectra and regional spectral climatology of wind waves. We also analyzed interdecadal changes of directional spectra in both VOS-based and model products and their ability to explain observed and modeled changes in wind wave heights and directions. Thus, the ability to derive directional wave spectra from visual observations adds a new value to the conventional analysis of wind sea and swell systems in terms of heights and periods.