The Harmony scientific workbench part 2

Abstract

This article introduces the Harmony scientific workbench (SWB), a tool for modeling the multistatic polarimetric ocean observations of the European Space Agency (ESA) Earth Explorer 10 Harmony mission. This work focuses on the retrieval of the mission’s radar-based level-2 (L2) geophysical data products, namely stress-equivalent wind at 10 m height (U10s) and total surface current (TSC) vectors. The retrieval uses simulated level-1 (L1) radar data, which is documented in a companion paper [1]. Different inversions of the simulated L1 observations are tested. We demonstrate that the assumptions made in traditional scatterometric retrieval approaches do not always hold at the high resolutions (O(1 km)) at which Harmony will operate. In particular, wave-Doppler suffers from representation noise, which leads to incorrect estimates of surface currents. The incorporation of the mean cross-section (MACS) and the spectral cutoff parameters improve the estimates of surface current substantially. In addition, iterative estimations with increasing resolution help to reduce the noise in the geophysical retrievals. Although it is possible to get close to 0.3 m/s root mean square error (RMSE) on the wind speed, and 0.1 m/s on the ocean currents at the scale of 1 km×1 km, there are limits to geophysical model function (GMF)-type retrievals, as they are not able to fully account for ocean surface characteristics such as the skewed and wavelength-dependent directional behaviour of wave systems. Despite these limitations, our results indicate that Harmony will provide the world’s first accurate high-resolution ocean wind and current observations based on SAR data.