**4.3.1 Sea surface data collection**

The first data collection was carried out on December 2010, from Sardinia Eastern coast at 157 m above the sea surface, near Cala Gonone. This region is characterized by high cliffs like those shown in Fig. 10.

Fig. 10. The Sardinia Eastern Coast near Cala Gonone (©Google Maps)

GNSS Signals: A Powerful Source for Atmosphere and Earth's Surface Monitoring 193

For satellite 16 (Fig. 12 (right)) two different echoes are visible. The first echo is coming from the coast (range closest to the observer and lower intensity due to scattering from the terrain); while the second is characterized by a greater delay, a different Doppler and successive returns lasting about 2 chips. In this case the correlation peak expires after 24 samples. This is a typical example of the capability to extract informations also from the un-

Thus, our receiving system is able to track coherent and un-coherent reflections and to contemporary distinguish between echoes with different delays, Doppler shifts and

During the second data collection of May 2011, an experiment performed flying over an area placed in the Piedmont region (north west part of Italy), the receiving system was placed on board a small aircraft in order to track reflections from rice fields. Since rice fields are flooded during this month, they are a perfect scenario to study reflection phenomena.

Like in the previous experiment, the geometry of reflections was analyzed and all the satellites with elevation lower than 33° were discarded, since below this elevation the specular reflections did not enter inside the -3 dB beam-width of the LHCP nadir looking

The signal to noise ratio detected from the reflected signal was normalized respect to the correspondent direct signal; moreover, we compute all the specular reflection points visible

On board the aircraft, a video camera was placed to see which fields were really flooded during the acquisition. The panoramic view extracted from the video was superimposed on ©Google Maps, together with the specular reflection points (Fig. 13). After the superposition, we have noticed a good agreement between the fields' state and the received power (see Fig. 13 and 14). The minimum received power correspondent to a low

Fig. 13. Specular reflection points tracks for satellite 8 and 26 over Piedmont rice fields with

Furthermore, we compare the signals of two different satellites with similar elevation but different azimuth; we notice a high correlation between the two specular reflection point tracks both from the qualitative (Fig. 13, Fig. 14) and the quantitative (Fig. 15) point of view. The quantitative comparison is performed considering the reflected power coming from the same longitude, considering a bean of 0.01°. Further investigations on this behavior are

relative normalized signal to noise ratio. See Fig. 13 for the red rectangle zoom.

and to each point we associate the relative normalized signal to noise ratio.

normalized signal to noise ratio is clearly associated to not flooded fields.

coherent part of the signal.

**4.3.2 Rice fields data collection** 

intensity.

antenna.

under development.

Considering the satellites in view, we pointed our antenna towards 170° of azimuth respect to geographical North. We successfully track reflected signals coming from GPS satellite 16 and GPS satellite 30. The geometry of acquisition was determined computing the iso-delay lines with ½ C/A code chip step and the specular reflection points, shown in Fig. 11 (in red for the 16th and in green for the 30th) together with the antenna footprint (depicted in light blue). All the points have been superimposed on ©Google static Maps and georeferenced in UTM. For both satellites the relative delay-doppler maps were computed over 1 s of noncoherent integration time and normalized from 0 to 1. Results are shown in Fig. 12. For satellite 30 (Fig. 12 (left)), the map is characterized by a very low noise, since the expected scattered signal is almost coherent and limited to one iso-range area, with no successive returns with delay greater than 1 chip (8 samples rising to the maximum, 8 samples going down to the noise floor).

Fig. 11. Specular reflection point and iso-delay lines superimposed in UTM on ©Google static Maps (satellite 16 in red, satellite 30 in green)

Fig. 12. Delay-Doppler maps for satellite 16 (right) and satellite 30 (left)

For satellite 16 (Fig. 12 (right)) two different echoes are visible. The first echo is coming from the coast (range closest to the observer and lower intensity due to scattering from the terrain); while the second is characterized by a greater delay, a different Doppler and successive returns lasting about 2 chips. In this case the correlation peak expires after 24 samples. This is a typical example of the capability to extract informations also from the uncoherent part of the signal.

Thus, our receiving system is able to track coherent and un-coherent reflections and to contemporary distinguish between echoes with different delays, Doppler shifts and intensity.
