The Harmony mission features two bistatic synthetic-aperture radar (SAR) companions of Sentinel-1. As with any multistatic system, frequency deviations among the oscillators of the receivers cause a phase error in the phase of the demodulated SAR signal. Given that interferometry will be used to retrieve geophysical parameters from Harmony’s radar instruments, an erroneous phase difference between the SAR signals of the two companions will bias the retrieval. The companions will use a global-navigation system (GNSS)-based method to synchronize the phase of the signals. The residual phase that remains after the synchronization is significant enough to make the retrieval of relative sea-surface height (RSSH) impossible. In this paper, we present Multisquint with Overlaps (MuSO), a data-driven algorithm to remove the synchronization residual. The algorithm uses the multisquint processing approach, together with the overlap regions of the TOPSAR acquisition mode, to estimate the derivative of the residual. After running the algorithm, simulations suggest that the error signal reduces from a standard deviation of 4° to 0.01°, allowing the retrieval of RSSH from Harmony data.

Assessing the performance of interferometers and processing interferograms require accurate knowledge of the temporal lag, sensitivity, and spectral shift. While these parameters are well-defined for conventional interferometric configurations, their definition becomes opaque for complex configurations, such as bistatic systems with formation-flying satellites. According to the principle of diffraction tomography, each instrument samples a distinct region of the scattering surface’s Fourier domain. Using this principle, we introduce a wavenumber-domain method for calculating the temporal lag, spectral shift, and sensitivity to height of synthetic-aperture radar (SAR) interferometers. The method calculates interferometric parameters by aligning the ground-projected wavenumber support of the SAR images forming the interferogram. Although the wavenumber-support method agrees with the conventional geometric formulations of the temporal lag and sensitivity in geometrically simple cases, the two methods diverge in more complex geometries. We show that when the two SAR satellites fly in a close-formation or have lines of sight that are squinted with respect to the zero-Doppler direction, then the geometric formulations are inadequate and the wavenumber-support method is needed to accurately estimate the interferometric parameters.

We present a method that uses the spectral wavenumber domain to estimate the temporal lag of synthetic-aperture radar (SAR) interferometers. The method uses the surface-projected region of support of a given resolution cell of each of the two SAR images that forms the interferogram. Despite the spectral method agreeing with the conventional geometric method for the monostatic interferometer case with an along-track separation, the two methods diverge in bistatic cases of formation flying interferometers. We show that the geometric approximation overestimates the temporal lag when the two SARs fly in a close-formation or have a squinted line of sight with respect to the zero-Doppler direction.

Estimating sea surface height using cross-track interferometry requires high sensitivity because the ocean surface signal is in the order of 10 cm. Additionally, the interferometer requires a temporal delay of a few milliseconds to ensure coherency of the moving ocean surface. We show that a squinted line of sight, in combination with a Helix satellite formation allows optimizing the effective perpendicular and along-track baselines to satisfy these conditions. This paper presents a model to estimate the performance of a formation-flying cross-track interferometer with a squinted line of sight. The tenth Earth Explorer, Harmony, which features two bistatic SAR companions, and a theoretical system with one monostatic and one bistatic SAR are used as case studies. The standard deviation of the height estimate is 1 cm to 10 cm between 29° to 41° and increases to 30 cm at the far range (46°) at a wind speed of 5ms-1. The power spectral density of the elevation shows that spatial scales of 47km can be resolved. The performance improves at higher wind speeds due to higher backscattering. At a wind speed of 15ms-1, wavelengths of 27km to 11km can be resolved, depending on the elevation spectrum. The performance over a 250km swath enables the instantaneous estimation of the surface elevation at the submesoscales for the first time.

The paper investigates the polarimetry of bistatic Synthetic Aperture Radar (SAR) acquisitions over rough surfaces, with focus on the rotation of the scattered wave orientation at the companion antenna axes and on the optimal linear polarization in transmission. This latter is defined as the polarization achieving the maximum radar cross section and will be herewith recalled as principal polarization. The paper outlines a geometrical framework for the interpretation and the estimation of the principal polarizations. It is shown that the theoretical formulation provides a good agreement with the second-order analytical approach in [1]. The paper finally postulates that a bistatic illumination in the traditional$\mathrm{H}$and V linear modes can be considered equivalent to a compact$\psi$-pol, i.e. with the transmittion in a linear polarization rotated by$\psi$. For long baselines, such those as those envisioned by the ESA Harmony EE10 candidate, and for steep incidence angles, an equivalent$\pi/4$-pol might be possible for rough surfaces.

Ocean topography using SAR interferometry requires coherent observations of the sea surface. To observe the surface coherently, the along-track baseline between observations of the same scene must be kept to a minimum. Minimising the along-track baseline while maintaining a cross-track baseline that allows good sensitivity to relative surface height is difficult to achieve in satellite missions. This paper shows how a squinted line of sight allows single-pass cross-track interferometry with a wide swath over oceans. The Harmony candidate mission will have a formation that uses such an acquisition geometry to coherently observe the oceans.