IEEE Geoscience and Remote Sensing Magazine - June 2022 - 34

Tables 3 and 4 summarize the sensitivities derived from
the different experiments and model predictions carried
out during the project. The results are quite consistent, with
the airborne data exhibiting a higher sensitivity but referring
to drier conditions.
The impact of topography is clearly highlighted by the
work, both experimentally and from the models. There was
a consensus in the need to upgrade the simulation models
introducing the topography effect when predicting the signal
from satellites.
The spatial resolution can also be good (on the order
of kilometers) for very flat surfaces if the signal is mainly
coherent; otherwise, it goes down to about 25 km. However,
the measurements are sparse (they are not arranged
as in one image), and they are not characterized by exact
-25
-30
-35
-40
-45
-50
-55
-60
-65
1,500
2,500
3,500
4,500
5,500
500
19 19.5 20 20.5 21 21.5 22 22.5
Latitude (°)
TDS
SAVERS
EMPIRICAL II = 0-30° THEORETICAL
Surface Elevation
FIGURE 23. The TDS-1 peak reflectivity compared to the reflectivity
simulated with SAVERS for a bare soil. The error bars indicate
the reflectivity variations for surface small-scale roughness ranging
from 3.2 to 3.8 cm. The surface elevation is extracted from the
SRTM DEM.
-20
-25
-30
-35
-40
-45
-50
-55
-65
-60
-12.5 -11.5 -10.5 -9.5 -8.5 -7.5
Latitude (°)
TDS
SAVERS
SP Near Water
Tree Height
FIGURE 24. The SAVERS simulated and observed peak reflectivities
versus latitude. The error bars indicate the reflectivity variations for
biomass ranging from 80 to 120 t/ha. The tree height is estimated
from lidar data.
34
40
35
30
25
20
15
10
5
TDS-1 (VOD
theoretically
corrected)
CyGNSS
7 dB/1 m3m−3
3 dB/0.4 m3m−3
15 dB/1 m3m−3
6 dB/0.4 m3m−3
CyGNSS binned ~18 dB/1 m3m−3
(ii ~independent)
~7 dB/0.4 m3m−3
CyGNSS
VOD and v
empirically
corrected
GLORI
26 dB/1 m3m−3
(ii ~independent)
~10 dB/0.4 m3m−3
~18 ÷ 11 dB/0.4
m3m−3 (relatively dry
case)
6 ÷ 6.8 dB/0.4
m3m−3
and fixed repetition cycles. Thus, some aggregation in space
and time must be carried out to obtain a product having a
raster (i.e., a gridded) format. The GNSS-R technique is also
affected by a poor coverage and temporal resolution. As a
matter of fact, an SP for repeating passes is located in different
positions so that, for inhomogeneous targets, this can
generate a variable signal as a function of time. For these
reasons, a constellation of small satellites must be considered
to overcome the problem [39], [72].
The work has also addressed the need of ancillary data
for retrieving the target parameters to compensate for the
effects of other features affecting the signal, including the
small-scale surface roughness and the topography (this is,
however, a common situation in remote sensing). Dual
polarized GNSS-R observations have proved quite useful
from low platforms but are not yet tested from satellites,
except for the experiment exploiting the SMAP radar as a
GNSS receiver.
In addition to the influence of many parameters on the
signal and the fact that observations are not regularly repeated
on the same point in inhomogeneous areas, large
fluctuations (a sort of speckle) were observed, even when
TABLE 3. SUMMARY OF SENSITIVITY TO SM FROM THE
DIFFERENT EXPERIMENTS. INDICATED IN DB PER UNIT
OF VOLUMETRIC SOIL MOISTURE (IN m3/m3) AND AS
DYNAMIC RANGE OVER THE TYPICAL RANGE OF SM
(BETWEEN 0 AND 0.4 m3/m3).
TABLE 4. SUMMARY OF SENSITIVITY OF
SPACEBORNE REFLECTIVITY TO DIFFERENT
FOREST PARAMETERS DERIVED FROM
LOCAL AND GLOBAL SCALE ANALYSIS.
SENSITIVITY TO LOCAL
VOD (Np)
AGB (t/ha)
GLOBAL
-0.05 dB/tha−1
18 dB/[50-
400 t/ha]
Tree height (m) -0.7 dB/m
-0.01 dB/
tha−1
-0.23 dB/m
THEORETICAL
-16 dB/1 VOD -7 dB/1 VOD < -11 dB/1
VOD
26 dB/[50-
400 t/ha]
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE JUNE 2022
Peak Reflectivity (dB)
Peak Reflectivity (dB)
LIDAR Tree Height (m)
Surface Elevation From DEM (m)
IEEE Geoscience and Remote Sensing Magazine - June 2022

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - June 2022
