A dedicated C-band radar is deployed on AMT26 to validate Sentinel-1 Synthetic Aperture Radar (SAR) imager products to investigate sea surface properties.
The radar sends a microwave signal, with the same wavelength (around 6cm) as used by the Synthetic Aperture Radar on-board Sentinel-1A. If the sea surface is perfectly flat, it will act like a mirror: all the microwave energy will bounce on the surface, go forward, and will never make the trip back to the radar. However, if the sea surface is wrinkled, the microwave energy will be scattered in all directions, and some will return to the radar. Therefore, the radar measures the intensity of the signal that is actually reflected back to the radar.
Anything that wrinkles the sea surface will leave a mark on the image – this can be swell of the ocean, wind, oceanic currents or internal waves, or many other parameters. Therefore a major challenge of this technique is to understand what the sea surface actually looks like and what is received by the satellite. This is why on the AMT cruise there is a camera alongside the radar to take pictures of the sea surface which is synchronized to the radar data stream. This capacity of the sea surface to backscatter the microwave signal is called the Normalized Returned Cross Section (NRCS).
Below are some plots of data from the radar. The radar “pings” 50 times per second, and at each ping (horizontal axis) the intensity of the backscattered signal is plotted as a function of distance (vertical axis). The strength of the signal is is measured in dBm (decibel-milliwatts) and ranges from weak (blue) to strong (red) in the colour scale bar. If you look carefully, you can sometimes see tilted 'features' in the time/distance axis: which means that something is reflecting microwaves effectively, and is seen over many successive pulses, moving towards or away from the ship. It can also clearly be seen that the contrast in the images decreases progressively as you move away from the ship: which means that the signal from distant ranges is too faint for the radar to measure.
As well as different wavelengths of microwaves, there are also different orientations of the electric field in the wave with respect to the propagation direction. The E-field is always perpendicular to the propagation direction, but it can point in any direction in the plane perpendicular. So if you look towards the sea surface from the radar, the microwaves that are produced could have their E-field looking straight up and down (12/6 o'clock), or E-field pointing in the 3-9 o'clock direction. When the e-field is in the 12/6 o'clock direction, the wave is called "vertically polarized" (V). When the e-field is in the 3/9 o'clock direction, it is called "horizontally polarized" (H). But not all the light that hits the sea surface with a vertical polarization will come back with a vertical polarization. A large part of it does, that's the VV signal and some of it gets converted to H polarization. That's the VH signal. Then some of the H-polarization transmitted signal comes back as H also HH, and some of it comes back as V. That's the HV signal.
The synthetic aperture radar onboard Sentinel-1A and Sentinel-1B can measure all or some of these signals at the same time. We tried to do the same. Hence the four plots. Which look similar, but are not entirely. Some things do show up better in VV than in HH, or in the cross-polarized signals that in the co-polarized ones.
We can already use the backscatter intensity and images to investigate on the sea surface properties that are actually sensed from the Sentinel 1 satellites. But over the next two years we also intend to produce a radar that will be sufficiently stable to be able to check on the calibration of the NRCS data that we receive from the Sentinel-1 satellites, in a similar way as the ISAR instrument is used to check on the calibration of the SLTSR and the infrared radiometers on-board.
Louis Marie, IFREMER