Validation and intercomparison of ocean wave spectra inversion schemes using ASAR wave mode data

Abstract

This paper presents a validation and intercomparison of the non-linear Partition Rescaling and Shift Algorithm (PARSA) for deriving full two-dimensional ocean wave spectra from ENIVSAT Advanced Synthetic Aperture Radar (ASAR) Wave Mode (WM) data. ASAR WM data acquired globally are used for the validation exercise by comparing the retrieved significant wave height (SWH), zero up-crossing wave period (Tm02), swell SWH (H12, for waves with a period longer than 12 s), mean wave frequency and mean wave direction to in situ buoy measurements and results from the European Centre for Medium-Range Weather Forecast (ECMWF) reanalysis wave model and the Deutscher Wetterdienst (DWD) forecast wave model. An intercomparison of results from the PARSA algorithm with those from the quasi-linear retrieval algorithm adopted by the European Space Agency (ESA) to generate the ASAR WM level-2 product (ASA_WVW_2P, termed WVW hereafter) is also presented in this paper. In addition to the intercomparison with the existing ASAR WM WVW product, sea state parameters (SWH and Tm02), integrated from the PARSA spectra are also compared to the results derived using the empirical algorithm CWAVE_ENV. Validation results indicate that the PARSA inversion can yield full two-dimensional ocean wave spectra. The retrieved SWH corresponds well to buoy measurements with a scatter index of 21%, as demonstrated by 1247 collocated data pairs. By comparing SWH to the ECMWF reanalysis wave model and the DWD forecast wave model, better agreement is achieved, with scatter indices of 9% and 16%, respectively. In addition to comparing conventional integral wave parameters normally used to assess the quality of inverted spectra, a comparison of individual PARSA spectra chosen in different sea state with the nearest numerical wave model spectra and the WVW spectra is performed to illustrate two-dimensional spectral differences.