BREAKING WAVE MEASUREMENTS WITH SAR DEPOLARIZED RETURNS Paul A. Hwang 1 , Biao Zhang 2 and William Perrie 2 1 Remote Sensing Division, Naval Research Laboratory, Washington DC, USA 2 Fisheries and Oceans Canada, Bedford Institute of Oceanography, Dartmouth, NS, Canada Email: paul.hwang@nrl.navy.mil ABSTRACT The wind generates a distribution of small slope waves and sporadic steep breaking events. Such double structure of the sea surface is expected to have a strong impact on the radar scattering from the ocean surface. The signature of the double structure is in the wind speed dependence of radar returns: linear for scattering from gentle waves and cubic for breaking contribution. The composite-surface Bragg resonance (CB) theory describes the former very well. Detection of the breaking contribution remains difficult. Here we show that the depolarized (de-pol) radar return exhibits the typical double structure, its wind speed dependence increases with wind speed from linear to cubic. The increased sensitivity of the de-pol returns in high winds is ideal for hurricane wind retrieval. The strong breaking connection offers an opportunity to measure wave breaking and the associated energy dissipation and area of foam coverage from space, their quantification is important in air- sea interaction and electromagnetic and electro-optical remote sensing. Index Terms— Depolarized radar cross section, Breaking wave, Hurricane wind 1. INTRODUCTION Active radar backscatters and passive microwave emissions from the ocean surface are our major sources of wind measurements in the ocean. Without spaceborne active and passive microwave sensors, marine weather forecasts and simulations of three quarters of the earth surface would have to rely on the limited number of instrumented buoys and weather ships distributed irregularly over the ocean. Clearly, active and passive microwave remote sensing has become an integral part of our weather monitoring systems. Phillips [18] addresses the radar scattering problem based on the structure of ocean surface waves, which are the roughness elements that scatter the radar waves on the ocean surface. He establishes an analytical framework of the double structure of sea surface scatterers. Applying the results of his detailed wave dynamic analysis [17] and radar scattering theories [22], he shows that the wind speed dependence of the normalized radar cross section (NRCS) is expected to be linear in the Bragg contribution and cubic in the breaking contribution. However, based on subsequent comparisons of the theory with several field observations, he concludes that the assembled data do not reveal the cubic wind speed dependence expected of the breaking contribution. The data set [7] he examined in most detail reveals a comprehensive linear dependence of NRCS on wind speed throughout the whole range of radar frequencies and incidence angles in the data set (0.428 to 8.91 GHz and 30° to 85°, respectively). In Section 2, we present results from analyzing the co-pol (σ 0VV and σ 0HH , respectively, vertical transmit vertical receive and horizontal transmit horizontal receive) and de- pol (σ 0VH , horizontal transmit vertical receive or vertical transmit horizontal receive) returns from RADARSAT-2 (R2). The co-pol data are in general agreement with the CB theory but de-pol data show significant departure in high winds. Of special interest is the wind speed dependence of de-pol returns, linear in mild to moderate winds and cubic in high winds, reflecting the characteristics of the double structure described by Phillips [18]. The significant connection between de-pol returns with breaking offers an opportunity to monitor wind wave breaking from space, and the enhanced sensitivity in high winds is especially valuable for hurricane wind retrieval (Section 3). A summary is given in Section 4. More detailed discussions have been presented in Hwang et al. [14][15]. 2. ANALYSIS OF RADARSAT-2 RETURNS Quad-polarization (quad-pol) and dual-polarization (dual-pol) measurements with de-pol radar returns from the ocean surface are now available from many satellites, among them the R2. Here we present analysis results of quad-pol (414 points) and dual-pol (372 points) backscatter data and collocated wind velocities from ocean buoys maintained by the National Data Buoy Center (NDBC). Figure 1 displays examples of the radar returns of quad-pol data set as a function of wind speed, U 10 . We also show the curves predicted by the CB theory [22] and empirical geophysical model function (GMF) CMOD5 [10][16] for