Abstract
The goal of this work is to develop and evaluate techniques for the retrieval of wind and internal wave (IW) information from marine X-band radar data. While ocean wind measurements are crucial for the transfer of energy and momentum across the air-sea interface, IWs play an important role in tidal energy transport. Marine radars work by transmitting microwave energy from a rotating antenna that also measures the backscatter. The radar backscatter from the sea surface is controlled by the wind-generated small ripple waves through the Bragg-scattering mechanism. Surface winds are thus the dominant factor for generating the radar backscatter. The varying surface current fields associated with IWs interact with the ripples, generating rough convergent and smooth divergent zones. Radars are capable of imaging such IW-induced surface signatures as bands of enhanced and weakened backscatter. The advantage of radar-based wind information is that it is obtained from a large area around the instrument. Marine radar wind data are therefore much less likely to be affected by platform-induced air flow distortions than in-situ measurements. Previous investigators have already demonstrated marine radar's suitability as a wind sensor [31, 30], however, these works have been limited to fixed-platform data. Here, the focus lies on shipborne marine radar data. Such data present the challenge that the existing wind streak-based approach for retrieving wind directions cannot be directly applied. This is because the wind streak signal may become obscured by the horizontal ship motion, since wind streaks become visible only after averaging over a sequence of radar images. In addition, moving platforms face a greater variability of conditions, which may further complicate a radar-based wind retrieval. Grazing incidence HH-polarized (horizontal transmit and receive) X-band radar data exhibit a single intensity peak in upwind direction. To retrieve the wind direction, this work proposes a least-squares fit technique that identifies the upwind peak in the range-averaged backscatter dependency on the antenna look direction. This technique requires no motion correction and is therefore well-suited for shipborne data. In addition, it functions well even if sections of the radar field of view are masked. An empirical model function is derived to retrieve the wind speed from the mean radar backscatter intensity. Data from the U.S. Office of Naval Research (ONR) Impact of Typhoons on the Ocean in the Pacific (ITOP) experiment are used for a comparison between radar-based wind estimates and anemometer measurements. The two data sets show good agreement. In addition, this work proposes a technique that uses geolocated marine radar data to extract wind streak information through a localized Radon transform. To compare streak- and upwind peak-based wind direction retrieval techniques, fixed and moving platform marine radar data from the ONR-sponsored High Resolution Air-Sea Interaction (Hi-Res) experiment are used. Wind directions obtained using the upwind peak method show a better agreement with the reference data than those obtained from the wind streaks. The difference between fixed and moving platform for the wind streak approach indicates that the image geolocation affects the wind retrieval negatively. Standard deviations as low as 6.0° and 0.42 m/s for the comparison between radar-based and reference wind data show that marine radars can yield highly reliable wind estimates. Regarding IWs, a new fully automated tool to retrieve IW signatures from marine radar image sequences is developed and applied to data collected during ONR's Non-Linear Internal Wave Initiative / Shallow Water '06 experiment (NLIWI/SW06). Marine radars have the advantage over satellite systems that their high temporal resolution enables the study of the IW evolution. The proposed technique employs our knowledge about the wind dependency of the radar backscatter to correct for the image ramp, i.e. the return signal's dependency on range and antenna look direction. The ramp-corrected radar images are then geolocated and averaged, which greatly enhances the IW signal. By determining the IW group velocity and correcting for it before the radar images are averaged, the IW signal is further enhanced. Such pre-processing enables a reliable retrieval of IW surface signatures by clustering local peaks and troughs, and tracking those clusters through time. This work also includes a detailed analysis of data collected during the tracking of a particularly energetic IW. The radar-derived time series of IW speed, direction, and soliton maps yield unique information about the IW's spatio-temporal evolution, including evidence of wave-wave interactions. In addition, it is demonstrated that marine radar data can be used to retrieve information about the interior ocean dynamics associated with the IW. The IW-induced backscatter modulation is correlated with the measured surface current gradients and IW amplitudes. Alternatively, results are shown where IW amplitudes were derived from the distances between positive and negative radar backscatter peaks. This approach was first introduced by Xue et al. [132] and is based on an extended Korteweg-de-Vries (eKdV) equation. This approach has the advantage that it is much less dependent on the prevailing wind conditions. To summarize, the marine radar backscatter dependency on wind is analyzed, and new wind retrieval techniques from shipborne radar data are proposed. The gained knowledge on the backscatter's wind dependency is then applied to marine radar data containing IW surface signatures. This work proposes a new methodology for retrieving these signatures and uses the resulting IW soliton maps to derive information about the IW-associated interior ocean dynamics.