We summarize a theoretical analysis of the velocity resolution of an ultra-wideband (UWB), along-track interferometric SAR (ATI-SAR) when used to measure ocean surface currents. In particular, we investigate the question of whether spreading the total system power,PT, over multiple channels spanning an ultra-wide RF band,Btot, can improve the surface current velocity resolution. We analyze the resolution achievable when a frequency modulated continuous wave (FMCW) system is channelized to effectively operate aspindependent ATI-SARs, each with a bandwidth ofBtot/p, and an average power ofPT/p, but with different center frequencies. Our results indicate that for moderate and higher wind speeds and for the low altitude, drone-based configurations we envision for such a system, spreading the energy across multiple RF sub-bands does indeed produce finer velocity resolution than if the energy is concentrated into a single channel. While a single channel produces a higher signal-to-noise ratio (SNR) than the individual sub-band channels due to higher average power and pulse compression gain, this higher SNR does not always decrease the interferometric phase noise and current speed uncertainty. In contrast, the independent samples in thepsub-bands can be used to reduce the phase noise through a weighted average, resulting in lower overall uncertainty. However, as the altitude and thus the range increases, the phase noise shifts into a regime where it is more sensitive to SNR. The performance of the single-channel system is then superior over a progressively wider span of wind speeds.

Doppler scattering simulations from sea-like surfaces are sensitive to non-linear hydrodynamic effects in wave propagation. Two commonly considered approaches to introduce such non-linearities – the Creamer transform and the Choppy wave model – are examined in this paper. It has been observed that the two models yield simulated Doppler spectra with conspicuously different centroids. Based on the modulation transfer function theory for the Doppler moments of sea backscatter, it seems plausible that such a behavior could result from model differences in the hydrodynamic modulation of smaller, scatter-producing ripples by larger waves. We perform numerical experiments to measure the corresponding hydrodynamic modulation transfer functions (HMTF). An established approach uses a single-harmonic large-scale surface. A correlation-based technique is also proposed that allows HMTFs to be estimated in Monte Carlo simulations of full-spectrum ocean-like surface profiles. Both methodologies show the Creamer HMTF values to be significantly larger than their Choppy-wave counterparts. Estimates made with ocean-like surfaces are also found higher than those based on the solitary-wave approach, particularly for the Creamer model. Application of the first-order Romeiser-Thompson theory with the estimated HMTF values produces Doppler shifts that are in general agreement with those observed in exact scattering simulations using either surface representation. The results of the study reveal considerable differences between the hydrodynamic modulation effects of the two surface models and demonstrate that those disparities are important contributors to the observed deviations in Doppler centroids. More work is required to conclusively determine a preference of one model over the other.

Doppler spectra obtained from simulations of ocean-like surface bistatic scattering are shown under the first order small slope approximation of rough surface scattering. Spectral properties are examined as a function of wind speed, polarization, and observing geometry in order to assess the potential for the bistatic remote sensing of various sea surface properties.

The spectral description of the wind-roughened ocean surface usually involves an omnidirectional spectral model and an azimuthal spreading function that describes the fraction of the waves of a certain length propagating at a given angle with respect to the wind direction. A number of such functions have been introduced and refined over the years. We use the 2nd-order Small Slope Approximation (SSA2) to calculate and compare the backscattered normalized radar cross sections resulting from several of them. In the process, the same Hwang omnidirectional spectrum is used. The SSA2 allows evaluating both co-polarized and cross-polarized backscattering cross sections. This study focuses on L and Ku radar bands and wind speeds between 5 and 9 m/s and covers a wide range of incidence angles. The results are compared with geophysical model functions where available.

The accuracy of sea scattering simulations depends on both the robustness of the chosen electromagnetic model and the adequate description of the ocean surface itself. The latter consideration is particularly relevant for Doppler simulations that are sensitive to hydrodynamic (non-linear) effects in surface propagation. There are two commonly used approaches to introduce such non-linearities – the Creamer transform and the Choppy wave model. We compare these two models and their impact on the Doppler characteristics of the computed backscattered signals using Monte Carlo simulations in the 2-D space. The electromagnetic part of the problem is solved exactly using a boundary integral equation technique. Significant differences in Doppler centroids corresponding to the two models are observed for intermediate to large incidence angles, with the Creamer transform producing larger shifts. The widths of the spectra are more consistent, with the Creamer model leading to somewhat larger values. Possible reasons for the discrepancies are discussed.

A recent ship-based field campaign demonstrated the use of a maritime multi-channel reflectometer using Sirius-XM broadcast radio downlinks as Signals-of-Opportunity. This campaign involved the collection of a large data set of bistatic ocean surface scattering measurements, which coincided with ground truth data detailing local oceanographic parameters. The experiment was unique in that most of its measurements were taken well outside the plane of incidence, with illumination from an oblique angle. This study offers a novel opportunity to explore the fidelity of electromagnetic models (coupled with oceanographic models) to predict radar scattering in these unique geometries. Match-ups are shown between measured and modeled bistatic normalized radar cross sections and range profiles, with the intention of both 1) supporting shore- and pier-based reflectometry through bistatic surface scattering model validation and 2) demonstrating the utility of using reflectometry for oceanic remote sensing from a sea-faring vessel.

A method for determining the cutoff wavenumber used in the geometrical optics theory of scattering from a random rough surface is presented. The method is based on the “alpha stable distribution” approach described in previous studies, and yields a cutoff wavenumber that depends on the spectrum model applied, the angle of incidence, and the frequency of interest. Use of the cutoff wavenumber so determined is found to improve agreement between predictions of the geometrical optics and physical optics theories of near specular scattering from the sea surface in some situations.

This paper examines errors in commonly used sea backscatter characteristics, such as the normalized radar cross section (NRCS) and first moments of the Doppler spectrum, when calculated using the Two-Scale Model (TSM). The latter is used in its “deterministic” or “coherent” formulation that allows evaluation of the complex scattered field for any given surface profile. The average quantities of interest are compute in the two-dimensional space using large-scale Monte Carlo simulations at X band. These descriptors are benchmarked against their “exact” counterparts obtained from a Method of Moments (MoM) solution for the same surface realizations. The calculations cover incidence angles from 0 to 85° and scale separation lengths from 0.5 to 8 electromagnetic (e-m) wavelengths. The produced error maps in these two variables look very different for the NRCS and for the Dopple moments. The error landscape for the latter is rather complex, but a depression is discernable where the separation length is inversely proportional to the sine of the incidence angle. Picking this path, however, is not optimal for the NRCS and would lead to large errors at smaller incidence angles. For the angle-independent choice, various optimization approaches suggest the best separation value to be around one e-m wavelength for all considered quantities.

Simulation results are used to assist algorithm development for extracting the ocean surface significant wave height (SWH) from microwave scattering of satellite broadcast signals from the ocean surface. The analysis shows that the Doppler frequency spectrum width is a reliable parameter for extracting the SWH. The numerical experiments include both deep and shallow water conditions. The correlation between Doppler frequency spectrum width and SWH is robust and applicable to all wind speeds, inverse wave ages, and water depths tested in this study (5–21 m/s, 0.8–2.0, and 3 m–$\infty $, respectively).

This work presents data from a recent field campaign in which an S-band passive radar using noncooperative satellite-based illumination was used to measure bistatic radar scattering from the sea surface. Significant wave height (SWH) information from a nearby Waverider buoy was reported concurrent with passive radar measurements. These simultaneous data collections fostered a study of the sensitivity of passive radar amplitude and Doppler measurements to SWH. This work demonstrates that Doppler and amplitude measurements with passive radar using satellite-based illumination can provide a promising approach to coastal remote sensing of SWH and mean wave direction, particularly when the scene is illuminated by multiple transmitters with a large physical separation between them.

Numerical experiments are carried out to explore the retrieval of ocean surface significant wave height (SWH) with the Doppler spectrum width (DSW) of microwave bistatic scattering from the ocean surface. Variations of DSW with incidence and scattering angles, wave height, wind speed, wave age, local water depth, and swell condition are described. Physical interpretation is offered on the relationship between SWH and DSW through the surface wave orbital velocity. Among the various environmental factors, the swell exerts the largest influence on the resulting relationship between DSW and SWH. The analysis considers a stationary transmitter and receiver only in order to provide insight into the impact of wave motions on the Doppler signatures.

This study implements direct numerical simulations of in-plane bistatic scattering in 2-D space to mimic the setup of an actual wavetank experiment. The latter investigated feasibility of passive sensing of water surface conditions using commercial geostationary satellites as transmitters of opportunity. While range resolution afforded by the signal bandwidth and sensing geometry is very coarse (nearly a hundred meters at best), finer spatial sensitivity can still be obtained due to narrow pattern of the receiving antenna located close to surface. Variations of the signal power and short-term Doppler shifts with time – all observed in the field data – can be used as additional sources of remote sensing information. Therefore the numerical simulations took account of antenna patterns and locations, while using a representative measured wave spectrum. We demonstrate that undulations in power and spectrogram centroids are indeed present in the modelled data and can be related to the dominant wave period. Critically, the numerical experiment allows considering a broad range of scattering angles (receiver antenna orientations) to indicate when such wave period retrievals are feasible.

The study investigates the accuracy of the Doppler and normalized radar cross section (NRCS) predictions when using the coherent scattered field expressions of the Two-Scale Model (TSM). This is done through large-scale Monte Carlo simulations in the 2D space, followed by comparisons to the results of integral equation-based exact numerical calculations. Of particular interest is the choice of the scale separation parameter (intrinsic to the TSM) that minimizes the error for either the Doppler spectrum moments or the NRCS. The optimal separation length is found to be much smaller (about 1-1.5 electromagnetic wavelengths) than the conventionally accepted values. This quantity is seen decreasing with stronger wind.

In this paper, we report initial results of an experimental investigation into measurement of ocean surface currents with an ultra high frequency synthetic aperture radar (UHF SAR) that supports an along-track interferometric (ATI) mode. We first describe the unique system, the NRL Multiband SAR, used to collect the data. We then present initial results related to two different approaches to current measurement. In the first approach, the surface current is estimated using standard ATI-SAR processing. We show progress in the implementation of this method with an interferometric phase cut across the north wall of the Gulf Stream. In the second, a time-sequence of images of propagating wave-like patterns is first generated using sub-aperture processing of the wide-beam, strip-map data, and this sequence is then transformed into wavenumber-frequency space through a 3D FFT. The surface current is then estimated from the displacement of the wave energy relative to the theoretical surface wave dispersion relation in the absence of any underlying current. We illustrate our progress in this approach with a wavenumber-frequency analysis of a 41-second sub-aperture image sequence of propagating waves.

Doppler characteristics are intrinsic to scattering from ocean surface and can be used for sea state retrievals. In this study, bistatic reflections of time-harmonic S-band signal from evolving sea-like surface are simulated in 2-D space using an exact technique based on the first-principles boundary integral equation formulation. Large Monte Carlo ensembles allow obtaining well-averaged Doppler spectra. The behavior of basic spectral parameters (mean shift and width) at both VV and HH polarizations is analyzed across a broad range of scattering angles (the incident angle is kept at 45°). Changes in those parameters with the sea state (wind speed) are examined. The simulations are also repeated with the two approximate scattering models: the coherent Two-Scale Model and the second-order Small Slope Approximation, providing rigorous assessment of their ability to predict Doppler properties of sea surface scatter.

This paper investigates new approaches to estimating the motion of the dynamic ocean surface using a multi-channel synthetic aperture radar (MSAR) with M phase centers arranged in an along-track configuration. The objective of this paper is to determine the processing methods that produce the finest velocity resolution, an issue that arises due to the finite coherence time of radar backscatter produced by the sea surface. The investigation is carried out both theoretically, using synthesized data produced from a modeled MSAR covariance matrix, as well as experimentally, using images collected with a 16-channel system. Three processing methods are considered: linear regression along with a multi-baseline phase progression, estimation of the velocity centroid, and coherent averaging of the shortest baseline interferograms. Both the theoretical and experimental results indicate that simple averaging of the shortest-baseline interferograms often produces the best velocity precision.

The study simulates and analyses backscattering of a time-harmonic (continuous-wave) X-band signal from evolving ocean-like surfaces in the 2D space. For the same surface realization, the scattered field is evaluated using the direct numerical solution of a first-principles boundary integral equation, and with two approximate models: the coherent Two-Scale Model (CTSM) and the second-order Small Slope Approximation (SSA2). Large Monte Carlo ensembles are built up to evaluate well-averaged Doppler spectra. The investigation focuses on the two spectral parameters: centroid and width, with alternative definitions for the latter being discussed. A broad and finely sampled range of incidence angles is considered, providing a comprehensive and detailed view at the behavior of these quantities from nadir to low grazing. Using the same set of surfaces when calculating backscattered fields according to the exact and approximate methods provides a rigorous test for the accuracy of the chosen scattering models. The reported high-quality data from direct numerical simulations can serve as a benchmark for other scattering formulations.

This paper presents a new interferometric approach to forming ocean surface velocity images using a multi-channel synthetic aperture radar (SAR) with M equally-spaced phase centers arranged in an along-track configuration. The technique coherently combines the M-1 shortest-baseline interferograms to produce motion estimates with superior velocity resolution, as compared to those produced by both the Velocity SAR technique as well as by an alternative interferometric method that analyzes the phase progression across a sequence of M-1 interferograms with progressively longer baselines. The new single-baseline approach also provides the widest possible unambiguous Doppler range, owing to the use of the shortest available baseline. The technique is analyzed theoretically and illustrated experimentally using ocean imagery collected with a multi-channel SAR system that supports 16 along-track phase centers.

Time-dependent properties of coherent range-resolved backscatter from a sea surface with simple floating targets are studied using two-dimensional (2-D) direct numerical simulations based on a first-principles boundary integral equation technique. The deterministic sea surface is generated according to a wind-driven wave spectrum. Target motion is modeled in a simplified way as that of a massless object heaved by orbital currents. Calculations can be performed at a broad range of grazing angles and at both HH and VV polarizations. The paper gives examples of range-time records of the X-band backscatter and then focuses on the Doppler characteristics of the simulated echoes. Differences between the responses of floating targets and sea surface are rather conspicuous, and appear to be caused primarily by dissimilarities in cross-section modulation behavior.

The concept of the “velocity synthetic aperture radar” (VSAR)-a multiaperture sensor capable of measuring radial velocities in the scene and utilizing this information to correct motion-induced imaging distortions inherent to SAR-was proposed two decades ago. Lately, with the emergence of truly multichannel systems featuring antenna arrays with dozens of elements, the approach has been enjoying a renewed interest. The viability and effectiveness of the algorithm were successfully demonstrated in a series of airborne field campaigns that involved imaging both man-made targets and natural maritime features. These experiments and the wealth of resulting data also underscored the need for comprehensive mathematical descriptions of expected target signatures in the collected “image stacks” and for further refinements of the VSAR imaging theory. This paper addresses both tasks by building upon the available mathematical results developed for the along-track interferometric SAR imagery of distributed evolving targets. The approach allows simultaneous accounting for all essential effects known to impact SAR imagery of a target or an extended feature: its azimuth velocity, radial velocity and acceleration, as well as finite coherence time. The emphasis is on obtaining closed-form expressions that could readily illustrate the structure and behavior of the VSAR stack spectrum of such a target and help gauge anticipated focusing improvement stemming from the VSAR image correction. In particular, it is rigorously shown that the VSAR algorithm is successful in situations when SAR defocusing arises predominantly from radial motion and short coherence times-the resulting resolution is generally no worse than that of the corresponding real-aperture radar. On the other hand, strong defocusing due to azimuth translation may be problematic to compensate within the VSAR approach framework.

This paper describes the Naval Research Laboratory Multichannel Synthetic Aperture Radar (NRL MSAR) and presents initial results from the first field deployment of this system. The NRL MSAR is an airborne test bed designed to investigate remote sensing and surveillance applications that exploit multiple along-track phase centers, particularly applications that require measurement of scene motion. The system operates at X-band and supports 32 along-track phase centers through the use of two transmit horns and 16 receive antennas. As illustrated in this paper, SAR images generated with these phase centers can be coherently combined to directly measure scene motion using the velocity SAR algorithm, and these measurements can then be used to correct the image distortion that the motion causes. In September 2014, this unique radar was deployed for the first time on an airborne platform. This paper presents a description of the system and results from its inaugural deployment, including the correction of distorted maritime signatures. These data were collected over an ocean inlet and contain a variety of moving backscatter sources, including automobiles, ships, shoaling ocean waves, and tidal currents.

This paper describes the Naval Research Laboratory Multi Aperture Synthetic Aperture Radar (NRL MSAR) and presents initial results from the inaugural field deployment of this system. The NRL MSAR is an airborne test bed designed to investigate remote sensing and surveillance applications that exploit multiple along-track phase centers, in particular, applications that require measurement of scene motion. This paper presents the results of initial coherent analyses to estimate scene and target motion and to correct the image distortions that these motions induce. These images were collected over an ocean inlet and contain a variety of moving backscatter sources, including automobiles, ships, shoaling ocean waves, and tidal currents.

Joint statistical properties of range-resolved sea backscatter at different polarizations are investigated using large Monte Carlo ensembles of numerically generated data. The simulations are based on the first-principles boundary integral equation technique in the two-dimensional (2-D) space and produce noise-free sets of backscatter corresponding to well-defined wind conditions. This study focuses on the joint probability density function (PDF) of vertically (VV) and horizontally (HH) polarized clutter intensities, investigating the behavior of the function's orientation and shape with variations in the incidence angle and wind speed. Comparisons with the results generated using coherent formulation of the popular Two-Scale Model reveal strengths and weaknesses of the latter in reproducing detailed statistical descriptors of sea clutter such as the joint VV-HH PDF.

Joint properties of high-resolution coherent sea surface backscatter at different polarizations are studied using two-dimensional (2-D) direct numerical simulations based on a first-principles boundary integral equation technique. The quantities considered include the joint probability density function for HH and VV intensities and the distribution for the phase difference between the two polarization channels. Various distribution moments and polarimetric parameters such as entropy and “alpha angle” are also examined, including their sensitivity to wind speed and noise. A broad span of incidence angles (ranging from 20 to 85 degrees) is considered, which helps elucidate the trends as backscattering geometry evolves towards low grazing regime.

Airborne along-track interferometric synthetic aperture radar (AT-InSAR) observations of the ocean surface often exhibit peculiar drops in coherence that can be quite compact (a few meters in size) and seem to be associated with breaking waves and whitecaps. Intriguingly, these signatures in the coherence map can be quite sharp and rather conspicuous, whereas the corresponding SAR image intensity in the same area may either suffer from apparent motion-related defocusing or lack any recognizable corresponding feature altogether. Such coherence drops are often accompanied by manifest perturbations in the interferometric phase. While there have been suggestions to use these details for breaker detection and even for breaker velocity retrieval, a basic understanding of why and how the patterns form needs to be developed first. To gain some insight in these phenomena, we consider AT-InSAR imaging of a simplified model feature that is a moving distributed target with a certain decorrelation time. The approach yields closed-form expressions that can be analyzed to identify possible mechanisms for forming such signatures. It appears that the intuitive explanation linking the drop in coherence to faster temporal decorrelation (something expected for a turbulent breaker) has its limitations when the feature is azimuthally small. Interference of the defocused and/or displaced image with the background may be the mechanism in the latter case.

Low-grazing wide-band HF radar echoes from evolving ocean-like surfaces in 2-D space are simulated employing a first-principles boundary integral equation technique. The synthesized signals achieve range resolution of about 15 m and correspond to observations with a hypothetical narrow-beam system having a 10-MHz effective bandwidth. Range-time plots of backscatter magnitude appear to contain a pattern consistent with the long surface waves. The double Fourier transform of the signal magnitude or power in range and time reveals strong harmonics located along the dispersion curve of the propagating gravity waves. The effect is discernable due to the high range resolution employed and is not predicted by the first-order Small Perturbation Method (SPM) approximation. It is explained by considering the next, second order of the SPM under certain simplifying assumptions such as perfect surface conductivity (with the Norton attenuation factor set to 1). The derived analytical expression is then used to investigate the feasibility of the surface profile retrieval from such time-varying range-resolved HF backscatter (gathered in the course of the “numerical experiment” through the exact scattering simulations). For the case of perfectly conducting sea-like surface developed under 10-m/s wind conditions, the results are very encouraging. The approach is sensitive to the surface harmonics with wavelengths far exceeding the Bragg resonance scales of the HF band. Extending this type of analysis to cases with realistic finite conductivity and/or rougher surface (assuming that the second-order SPM is still valid) will require proper accounting for ground-wave propagation effects in perturbation expressions. The study offers a glimpse into ocean observation possibilities that the wide-band HF systems can provide if and when they become a reality.

The dual-beam interferometer is an airborne instrument that combines two vertically polarized C-band along-track interferometric synthetic aperture radars (AT-InSARs) observing the surface below at different squints. The system was designed by the University of Massachusetts and saw several deployments in the early 2000s. An imagery of a small vessel with a rather pronounced wake pattern captured during one of such flights is the subject of this letter. Specifically, the interferometric phase in the turbulent wake exhibits a conspicuous banding structure that is still visible at distances more than 1 km behind the craft. The phase signatures from the fore and aft looks are combined to retrieve both longitudinal and lateral velocity components along cuts traversing the wake 400 and 750 m behind the boat. The results identify appreciable variations in the longitudinal velocity across the turbulent wake which are apparently consistent with the combined effect of the hull drag and the propeller backwash. A persistent pattern for the lateral component is also observed but is harder to interpret without the detailed knowledge of the vessel. The examples demonstrate the utility of AT-InSAR and, particularly, of a dual-beam AT-InSAR, for studies of centerline ship wakes. Readily available velocity signatures of a turbulent wake obtained with such systems can help with vessel classification tasks.

The technique for extracting wave period and wave direction from a navigation radar backscattering intensity is well developed but the determination of spectral density or wave height is hindered by the complex nature of the modulation transfer function. In contrast to backscattering intensity, Doppler velocity from coherent radar is the radial velocity of the scattering objects. Its oscillatory component is contributed by ocean waves. The spectral peak component of Doppler velocity is close to the peak wave period measured by a nearby buoy and the significant wave height can be accurately calculated. With radar range coverage on the order of ten dominant wavelengths, reliable assessment of peak wave period and significant wave height is achievable with radar data as short as one second. Wave direction can also be determined with a scanning system.

The Dual-Beam Interferometer is an airborne instrument that combines two along-track interferometric synthetic aperture radars observing the surface below at different squints. This configuration allows retrieving vector velocities of surface flows in a single aircraft pass. The system was designed by the University of Massachusetts and saw several deployments in the early 2000s. An imagery of a boat with a rather pronounced wake system captured during one of these flights is the subject of this paper. The velocity of the vessel is estimated based on the “train off the tracks” displacement in one of the looks. This estimate, which will be affected by uncertainty in target position due to smearing, is then used to remove unknown phase biases in the interferometric channels. Retrieved velocity variations are examined along cuts traversing the wake, with the focus on its narrow “turbulent” part. We find that the reconstructed velocities 150 m behind the boat are intuitively satisfying, but we are unable to fully account for the cross-wake component of velocity at 400 m behind the vessel.

Synthetic aperture radar (SAR) single-aperture systems are designed to image fixed scenes. Dual-antenna along-track interferometric (ATI) SAR systems are designed to detect moving targets and estimate their motion parameters. Although SAR systems are not designed to characterize moving targets, for localized targets, such as vehicles or ships, this problem has been addressed in the literature with some success. Distributed moving targets are hard to characterize, even for ATI systems. Estimating surface water speeds is an ideal example of this since the water returns are generated from weak returns randomly distributed in both time and space. The question asked here is can one measure surface water speeds with a single-phase center SAR system? The answer presented is a qualified yes, with the aid of an upper bound on the water speed, but not with the same accuracy as an ATI system. This upper bound and the collection geometry provide design criteria for the filtering of the phase information. Time-frequency (TF) methods provide another speed estimate as well as a rough profile of the speed across the water channel. A robust nonparametric TF method was developed and applied to estimate the speed. Comparisons between the ATI estimate and single-phase estimates are made using data from an X-band ATI-SAR system.

A numerical study of the retrieval of sea surface height profiles from low grazing angle radar observations is described. The study is based on a numerical method for electromagnetic scattering from 1-D rough sea profiles, combined with the ldquoimproved linear representationrdquo of Creamer for simulating weakly nonlinear sea surface hydrodynamics. Numerical computations are performed for frequencies from 2975 to 3025 MHz so that simulated radar pulse returns are achieved. The geometry utilized models a radar with an antenna height of 14 m, observing the sea surface at ranges from 520 to 1720 m. The low grazing angles of this configuration produce significant shadowing of the sea surface, and standard analytical theories of sea scattering are not directly applicable. Three approaches for retrieving sea height profile information are compared. The first method uses a statistical relationship between the surface height and the computed radar cross sections versus range (an incoherent measurement). A second method uses the phase difference between scattering measurements in two vertically separated antennas (ldquovertical interferometry) in the retrieval. The final technique retrieves height profiles from variations in the apparent Doppler frequency (coherent measurements) versus range and requires that time-stepped simulations be performed. The relative advantages and disadvantages of each of the three approaches are examined and discussed.

The anomaly of radar sea spikes, defined here as the non-Bragg scattering events with backscattering cross section of horizontal polarization exceeding that of vertical polarization, has been associated with steep wave features possibly going through wave breaking process. This property is employed for remote detection of breaking waves. The results are used to quantify the effect of wave breaking on radar returns. Large increase due to breaking is found in the Doppler velocities of both polarizations (about 50% faster with breaking). The effect of breaking on the backscattering cross section is much stronger for the horizontal polarization (with 15 to 20 dB enhancement) and relatively small in the vertical polarization (on the order of 0.5 dB fluctuations). The presence of swell reduces the impact of breaking waves on radar return in comparison to the scattering from wind seas.

Joint Time-Frequency Analysis (JTFA) can estimate the speed of moving targets in Synthetic Aperture Radar (SAR) images. These targets are usually solid targets. JTFA estimation of water surface speeds is more difficult since the returns are generated by Bragg scattering epochs that are randomly distributed in the radar's footprint - in both time and space. The spectral characterization of these returns depends strongly on the geometry of the collection, which also affects the techniques needed to estimate the speeds. A novel phase differencing technique is presented to measure the frequency offset caused by the range motion of the surface water. JTFA is applied after extensive filtering to extract the motion induced component of the phase signal.

Radar backscatter data are collected using a coherent, dual-polarized radar from a fixed tower in the ocean. Statistical analysis is performed to investigate the 1D and 2D probability density functions of backscatter intensity, Doppler frequency, coherence function and polarization ratio. The fraction of sea spike coverage generally increases with wind speed but the trend of increase is modified by the intensity of background swell condition. Parameterizations of sea spike coverage combining both wind and wave factors show some apparent advantage than parameterizations with wind or wave factors alone.

The ability of high-resolution imaging systems to resolve small-scale structure on the ocean surface suggests the possibility of using characteristics of Langmuir circulation to map the surface mixed-layer depth and near-surface current. We illustrate this using synthetic aperture radar and infrared imagery that are collected across the edge of the Gulf Stream (GS), which reveals surfactant streaks, or ldquowindrows,rdquo that are induced by Langmuir circulation. Based on changes in the windrow spacing and orientation, the mixed layer is estimated to deepen from 7 to 12 m across the edge of the GS and the current to increase from about 1 to 2 m/s. These spatial changes compare reasonably well with independent data, suggesting that the approach is plausible. It may also be possible to extract additional environmental information from the windrows.

Statistical properties of the X-band sea clutter are studied using 2-D direct numerical simulations. Surfaces are modeled as realizations of a Gaussian random process with the Pierson-Moskowitz or Elfouhaily spectrum. The Creamer transform is further applied to account for the lowest-order surface nonlinearities. Backscattered field at a given frequency is found using the first-principles boundary integral equation (BIE) technique. Calculations are repeated at a number of frequencies, which allows synthesizing the surface response to a pulse as short as 2.2 ns (the corresponding spatial resolution is 0.33 m). Large-scale Monte Carlo trials are used to evaluate the correlation properties and to obtain the probability distributions for the vertically- and horizontally-polarized clutter. This paper concentrates on the incident angle of 85deg (5deg grazing), with a few results for moderate 60deg incidence also reported for comparison. The effects of variations in wind speed (sea state) and radar resolution on the clutter statistics are investigated. An L-band example (with proportionally longer pulse) helps explore the role of a different electromagnetic (e/m) wavelength. The simulation technique also allows for the isolation and examination of the impacts of certain e/m and hydrodynamic approximations, including the replacement of rigorous solution to the BIE by a simpler analytical scattering model. The amplitude statistics of the simulated backscatter are compared to the Weibull and K distributions that are often used to describe surface clutter

A dual-beam interferometric synthetic aperture radar provides two velocity components of the ocean current remotely from a single fight pass. Combining two flight passes, all three orthogonal components of the surface velocity can be retrieved. An experiment was conducted near the Gulf Stream (GS) boundary. A sharp change of the surface velocity of about 1 m/s over a 500 m distance was measured. The wave condition is dominated by a 14-s swell system and low wind velocity. The wave variance inside GS is about twice the wave variance outside the GS in the present data. The large difference in the wave variance is considerably higher than that can be expected from hydrodynamic modulation. An ocean current system with strong current shears such as the Gulf Stream is a wave guide and can trap waves with the right combination of wavelengths and propagation directions. Numerical calculations indicate that the wave properties of the data set may satisfy the conditions for wave trapping by the Gulf Stream.

Retrieval of one dimensional ocean surface profile information is studied through the use of S and X band radar backscattering data at low grazing incidence. The required backscatter data is obtained through the use of numerical electromagnetic scattering codes. The simulations compute backscatter at multiple frequencies in each time step, so that range-resolved radar cross sections are obtained after an FFT operation. Profile retrievals are obtained through a simple "tilted Bragg" assumption, which allows local incidence angles on the surface to be directly determined from RCS data. These local incidence angles are then translated into surface slopes, and further integrated to construct the estimated profile. Initial results from this process are reported in this paper; further studies are in progress to quantify the overall accuracy achieved as a function of various surface and sensor parameters.

The dual-beam interferometer consists of two interferometric synthetic aperture radars (InSARs), one squinted at 20/spl deg/ forward of broadside, and the other 20/spl deg/ aft, to allow measurement of vector surface velocity with only a single aircraft pass. Estimates of surface velocity vectors in the coastal region during high tidal flow are presented. The data were gathered over the barrier islands west of Fort Myers, Florida, as part of a March 2004 deployment. Whereas no detailed bathymetry data were available, high-quality aerial photography appears to be a useful tool in inferring bottom topography and possible current obstructions. The retrieved velocity field clearly follows the expected outflow pattern. While comparisons with tidal current magnitudes predicted by the U.S. National Ocean Service do reveal discrepancies of up to 0.5 m/s, these differences are most likely due to the contribution of ocean surface waves to the overall InSAR velocity measurement. Velocity retrievals for the same area based on the data from different tracks show good consistency. The results constitute the first demonstration of vector retrieval of the surface velocity field with a single-pass InSAR system and confirm the robustness of the dual-beam interferometry principle.

The Dual Beam Interferometer (DBI) developed by University of Massachusetts (UMass) and an infrared camera operated by the Naval Research Laboratory (NRL) were jointly deployed during a March 2004 flight campaign off the east coast of Florida. Data collected produced simultaneous observations of features over the Gulf Stream's western boundary by both instruments. Use of instruments with different imaging capabilities enabled extraction of wind direction as well as current velocity measurement.

The Dual Beam Interferometer consists of two interferometric SARs, one squinted at 20 degrees forward of broadside, and the other 20 degrees aft, to allow estimation of vector surface velocity with only a single aircraft pass. It was developed by the University of Massachusetts and is operated in collaboration with the Naval Research Laboratory. The paper presents initial estimates of surface velocity vectors in the coastal region during high tidal flow. The data were gathered over the barrier islands west of Ft. Myers, Florida, as part of March 2004 deployment. While no attempt has been made at this stage to correct for Doppler contributions from surface waves, the retrieved velocity field clearly follows the expected outflow pattern, and velocity magnitudes agree well with in-situ data.

A dual-beam along-track interferometric synthetic aperture radar that is entirely self-contained within an aircraft pod has been developed by the University of Massachusetts to study sea surface processes in coastal regions. The radar operates at 5.3 GHz with a bandwidth of up to 25 MHz. System hardware is described. Initial test flights aboard the National Oceanic and Atmospheric Administration's WP-3D research aircraft were performed to evaluate system performance over land and water surfaces. Imagery were collected for fore and aft squinted beams, though no interferometric data were collected. Notable look-angle dependences are observed in the sea surface normalized radar cross section under very low wind conditions.

A dual-beam along-track interferometric synthetic aperture radar which is self contained within an aircraft pod has been developed to study coastal regions. System hardware is described. Initial test flights aboard the NOAA WP-3D research aircraft were performed to evaluate system performance over land and water surfaces. Notable look-angle dependencies are observed in the sea surface NRCS under very low wind conditions.

Results from a Monte Carlo simulation of backscattering from one-dimensional (1-D) time-evolving sea surface models are reported. A numerical electromagnetic method based on an accelerated forward-backward approach is used to calculate backscattered returns from impedance surface profiles at incidence angles of 0/spl deg/ (normal), 40/spl deg/, and 80/spl deg/. Surfaces are initialized as realizations of a Pierson-Moskowitz spectrum and then stepped in time through a numerical hydrodynamic method. Results from three distinct hydrodynamic methods are compared: a linear evolution, the "improved linear representation" of Creamer et al. (1989), and the "Watson-West" approach of West et al. (1987). Instabilities in the West model due to formation of steep wave features limit the study to L-band backscattering for wind speeds less than 2 m/s, so that the surfaces considered are only slightly rough on an electromagnetic scale. The small slope approximation for electromagnetic scattering is shown to provide reasonable predictions in this limit. Statistics of the resulting surface profiles and backscattered fields are compared for the three models and are found to be similar in most respects. Backscattered field Doppler spectra, however, show differences, with the West model apparently capturing more nonlinear interactions in the surface evolution.

Results are presented for an experiment utilizing a pastoral land scene with a variety of eight classes, imaged by the NRL dual band (X and L) polarimetric synthetic aperture radar (NUWSAR) at a spatial resolution of 1.2 m. Projection pursuit (PP) statistical analysis tools were applied to a set of simultaneous L-band and X-band fully polarized images (six independent channels) to demonstrate the utility of land classification at high spatial resolution from a light aircraft using SAR. The statistical confusion matrix was used as a quantitative optimization measure of classification. Samples of eight classes from a portion of the scene were used to define a training set, then PP tools were used for classification. It is clear that L-band and X-band fully polarized data view the classes in a significantly different manner, and each brings independent information to the analysis. These results are not meant to be exhaustive at this time but to demonstrate the utility of applying PP tools to multiband and polarization SAR data and to give an indication of the quality of classification one can achieve with moderately high spatial resolution SAR data using a light plane platform.

The literature has seen the development of robust and efficient numerical techniques for exact calculations of rough surface scattering. We discuss how such methods, typically formulated for time-independent surfaces, can be extended to calculate scattering from time-evolving ocean-like surfaces. Estimates are provided for the choice of parameters in such time-varying simulations. The method of ordered multiple interactions (MOMI) is used to calculate time-varying scattering from surfaces generated according to linear and nonlinear (Creamer) models for incidence angles ranging from normal to low grazing. We discuss the runtime considerations and demonstrate that combining the MOMI with a fast multipole method (FMM)-type acceleration technique makes large-scale time-varying Monte Carlo simulations possible. The average Doppler spectra of backscattered signals obtained from such simulations are compared for different incident angles, polarizations, and surface models. In particular, the simulations show a broadening of the Doppler spectra for nonlinear surfaces, especially at low grazing angles (LGA) and a separation of the vertical and horizontal polarization spectra at LGA for nonlinear surfaces. This spectral separation at LGA is not observed when the linear surfaces are used.