Assessing the Impact of TEC Fluctuations on
ALOS-PALSAR Images
Luca Spogli
1,2
, Elvira Musicò
1
, Claudio Cesaroni
1
, John Peter Merryman Boncori
1
,
Giorgiana De Franceschi
1
and Roberto Seu
3
1
Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, Rome, Italy
2
SpacEarth Technology, Rome, Italy
3
Electronics and Telecommunications, Sapienza University of Rome, Via Eudossiana 18, 00184 Rome, Italy
luca.spogli@ingv.it, elvira.musico@ingv.it, claudio.cesaroni@ingv.it, john.merryman@ingv.it,
giorgiana.defranceschi@ingv.it, roberto.seu@uniroma1.it
Keywords: Ionosphere, SAR, InSAR, Total Electron Content, GNSS, Mid-latitude Ionosphere.
Abstract: Trans-ionospheric waves experience delay proportional to the Total Electron Content (TEC), being the
number of free electrons present along a satellite-receiver ray path. TEC is a highly variable quantity,
influenced by different helio-geophysical parameters, such as solar activity, season, time of the day, etc.
Such large variability may lead to TEC spatial and temporal fluctuations over different scales, affecting the
quality of the Synthetic Aperture Radar (SAR) signals and, in turn, limiting further developments of
interferometric techniques, such as InSAR (Interferometric SAR). In the specific, the need of catching
qualitative and quantitative correspondences between TEC fluctuations and InSAR image streaks is a key
point to drive the development of future mitigation techniques to improve the quality of the SAR imaging.
In this paper calibrated TEC values, derived from the RINEX data provided by the RING (Rete Integrata
Nazionale GPS) network of GPS receivers, are analysed to assess the ionosphere conditions during the
ALOS (Advanced Land Observing Satellite) PALSAR (Phased Array type L-band Synthetic Aperture
Radar) passages over central Italy.
1 INTRODUCTION
The InSAR is among the most used techniques for
geophysical applications, because it is able to
provide high-resolution imaging, independently on
all tropospheric weather conditions (Bürgmann et
al., 2000). SAR signals emitted by ALOS-PALSAR
in L-band frequency range cross the bulk of the
ionosphere, because they are emitted at about 700km
altitude, i.e. in the topside ionosphere. By
consequence, such signals experience changes in
their phase and group velocity induced by the
ionosphere. Changes are proportional to the
distribution of free electrons encountered during
their travel from SAR satellite to ground and back.
The InSAR technique extracts the position and/or
the displacement of a surface by calculating the
phase difference between two SAR images acquired
in different days, but at the same local time and on
the same geometrical area (Ferretti et al., 2007).
Thus, to ensure the performance of the InSAR
imaging and to assess the ionospheric impact on it,
the status of the ionosphere in both passages must be
taken into account, as it is known to be responsible
of the appearance of streaks on SAR co-registration
images. In the specific, the shift of the co-
registration is proportional to the ionospheric TEC
gradients. (Chen & Zebker, 2014).
TEC can be obtained by means of GNSS dual
frequency receivers located at ground. GNSS slant
TEC (STEC) is defined as the total number of free
electrons within a cylinder with a cross section of 1
m
2
and height equal to the signal ray path, from
satellite to ground (in the GNSS case). STEC is then
a measure of the integrated electron density Ne
according to the formula:
15
Spogli L., MusicÚ E., Cesaroni C., Boncori J., Franceschi G. and Seu R.
Assessing the Impact of TEC Fluctuations on ALOS-PALSAR Images.
DOI: 10.5220/0006226700150021
In Proceedings of the Fifth International Conference on Telecommunications and Remote Sensing (ICTRS 2016), pages 15-21
ISBN: 978-989-758-200-4
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
𝑆𝑇𝐸𝐶 =𝑁𝑒
(
𝑠
)
𝑑𝑠,
(1).
where s is the ray path of the satellite-receiver link.
Being GNSS satellites located at an altitude of about
20200 km, TEC measured with suitable ground
receivers provides information also above the ALOS
altitudes. However, the contribution to TEC of the
ionosphere above 700 km is of few %’s (Kelley,
2009).
TEC is a highly variable quantity influenced by
different helio-geophysical parameters such as the
solar activity, the season and the time of the day.
This must be taken into account when considering
the ionospheric conditions of the days to which the
images obtained by InSAR techniques refer. Many
researches about ionospheric effects on SAR have
been carried out, most of which are based on
numerical simulations, use of global ionospheric
models like WBMOD (Meyer & Agram, 2015) or
large-scale interpolated ionospheric maps that do not
have the spatial resolution needed for InSAR
(Hanssen, 2001). Currently, only few studies
investigated the relationship between ionospheric
variability and SAR/InSAR imaging by means of
independent measurements (like GNSS) and, in
general, a final word has not been told about the
ionospheric influence on real SAR images (Zhu el al
2016). For this reason, high-resolution determination
of TEC and of its spatial-temporal fluctuations is
needed. In this work, we show the preliminary
results obtained by comparing ALOS-PALSAR
images of central Italy and TEC measurements
obtained by means of the Rete Integrata Nazionale
GPS (RING, http://ring.gm.ingv.it/) network of
GNSS receivers, managed by the Istituto Nazionale
di Geofisica e Vulcanologia (Italy).
The paper is organized as follows: section 2
presents the ALOS-PALSAR and GNSS data and
how they have been treated to obtain the results,
discussed in section 3. Then, conclusions are
provided in section 4.
2 DATA AND METHODS
2.1 ALOS-PALSAR
InSAR is an imaging technique that evaluates the
pixel-to-pixel phase difference between two SAR
images, acquired over the same area, to produce an
interferogram. In this work, two SAR images (a
“master” and a “slave”) are acquired with a single
receiver in two different epochs with almost the
same incidence angle and on the same area, i.e. in
the so called “repeat-pass interferometry mode.
Since the orbital positons of the two SAR passages
are slightly different, it is necessary to coregistrate
the master and slave images, with the accuracy of
sub-pixel, to calculate the phase difference of two
corresponding pixels and, then, obtaining high
quality InSAR images. In the present work, we
concentrate on the results of the co-registration,
without looking at the final InSAR product
(interpherometric phase), with the aim to minimize
the tropospheric error affecting phase measurements
(Hassen, 2001) and to focus on the ionospheric
contribution only.
Figure 1: ALOS ascending tracks over central Italy
We consider the azimuth shifts obtained by the
coregistration of three images acquired by ALOS-
PALSAR over Italy. The observed area is that
covered by the ground track 638 of ALOS, which is
located over central Italy. Shifts are estimated by
using the Multiple Aperture InSAR (MAI) method.
The MAI method, based on split-beam InSAR
processing, is able to extract along-track
displacements from InSAR data more efficiently
than the pixel amplitude correlation (Bechor &
Zebker 2006).
Fifth International Conference on Telecommunications and Remote Sensing
16
2.2 GNSS
As mentioned before, GNSS signals passing through
the ionosphere experience transmission delay
proportional to TEC. Starting from RINEX data,
dual frequencies GNSS receivers are able to
calculate TEC along the slant path, according to the
following formula:





 
 
 
where 1 (1575.42 ) and 2 (1227.60 MHz) are
the two frequencies of the signal transmitted, 1 and
2 are the corresponding pseudoranges and ε
represents the biases induced by the receiver,
satellite, multipath etc. In order to minimize the
biases, it is necessary calibrate the STEC (Cesaroni
et al., 2015a).
To such scope, the GOPI calibration software
(http://seemala.blogspot.it/2011/04/rinex-gps-tec-
program-version-22.html) has been used. The GOPI
software provides TEC values projected to the
vertical, under the assumption that the ionosphere
can be represented by a single, thin, ionized layer
located at 350 km (Mannucci et al., 1998) that is
suitable for a quiet mid-latitude ionosphere.
Hereafter, we refer to calibrated and verticalized
TEC as TECcal. The projection to the vertical allows
having TEC values not dependent on the position of
the GNSS receivers at ground. Since ALOS
ascending passages of T638 over the area of interest
occurred during night, it is reasonable to assume that
a mid-latitude ionosphere in the nighttime is
“frozen”, i.e. not meaningfully changing, during a
time interval of 5 minutes around the two passages
of ALOS.
To find the portion of the ionosphere crossed by
the SAR signal, some geometrical consideration are
here reported and sketched in figure 2.
The distance D between the point P at ground
and the projection of the point P_i (which is the
intersection of the SAR signal and the ionosphere) at
ground is approximately given by H_i*tan (α)
(Reuveni et al 2015), where the look angle, α, is
related to the incidence angle, θ, through the
following formula:




; (3)
In which Re is the Earth’s radius and H_i is the
altitude of the thin layer approximated ionosphere.
Then, to highlight the difference (if any) between
the ionospheric features in correspondence with the
two ALOS passages, maps of TECcal values have
been considered. To obtain a fine representation of
the ionospheric features as derived from regional
TEC maps, it is necessary to have a very dense
GNSS network. To the scope, the RING geodetic
network has been selected, being composed by more
than 100 receiver spread all over the Italian territory
(Figure 3). TECcal maps boundaries have been
defined to correspond to the SAR images
boundaries, following the geometry described in
Figure 2.
Figure 2: Ionosphere crossed by SAR signal. P_i is the
intersection between the SAR signal and the ionosphere, D
is the distance between the point P at ground and the
projection of the point P_i on the ground, α and θ are
respectively the look angle and the incidence angle, H_i is
the altitude of the thin layer ionosphere.
Figure 3: Location of the GPS stations of the RING
network
Assessing the Impact of TEC Fluctuations on
ALOS-PALSAR Images
17
In such maps, TECcal are calculated at each IPP
(Ionospheric Pierce Point) for every satellite in view
by each receiver over a 5-minute interval (frozen
ionosphere) centred in the time of the track. TECcal
are then interpolated with the natural neighbour
method (Sibson, 1981). Such interpolation technique
has been selected according to Cesaroni (2015b) and
to Foster and Evans (2008), in which the authors
demonstrate that the natural neighbour interpolation
is the best choice when regional TEC maps are
considered. To avoid border effect due to the
interpolation, maps have been calculated all over
Italy and, then, restricted to fit the SAR image
boundaries, as mentioned before. To evaluate the
TECcal differences between the two passages, the
values of the maps that correspond to the master and
slave days are subtracted and a map of difference
(ΔTECcal) is produced.
3 RESULTS AND DISCUSSION
The variations of the ionosphere between the
different passes of the SAR cause an azimuth shift
between the positions of the pixels of the master and
slave image. This effect, also known as “azimuth
streaks”, influences the optimal coregistration of the
interferometric pair (Gray et al., 2000).
In this work, we concentrate on 3 ALOS-
PALSAR images of central Italy, whose
master/slave dates and parameters are summarized
in Table 1. For each coregistred image, azimuth
shifts have been determined and reported in figures
4 to 6.
Table 1: Images and relative parameters used in this work,
where θm and θs are respectively the incidence angle of
the master and the slave image, Bt stands for the temporal
baseline and Bp stands for perpendicular baseline.
Image #
1
2
3
Master
01/07/2007
01/07/2007
16/08/2007
θm (°)
38.7191
38.7191
38.7395
Slave
16/08/2007
01/10/2007
01/10/2007
θs (°)
38.7395
38.7223
38.7223
Bt(days)
46
92
46
Bp (m)
279.9401
539.2134
259.2523
In all three cases, the shift varies from -3 m to 3
m. By comparing such figures, a strong similarity
between image #1 (fig. 4) and #2 (fig. 6) can be
noticed. In particular, a similar pattern of the shift
are present, but with opposite sign. The sign shift is
because the day 16 August 2007 is used as slave in
figure 4, while in figure 6 it is used as master.
Figure 4: Shift obtained by coregistration of the master
(01/07/2007) and slave (16/08/2007) images
Figure 5: Shift obtained by coregistration of the master
(01/07/2007) and slave (01/10/2007) images
Fifth International Conference on Telecommunications and Remote Sensing
18
By looking closely to the red strike located at
around 41.8°N in figure 4 (or the corresponding blue
strike in figure 6), it is possible to note that it divides
into 3 smaller sub-strikes.
Figure 6: Shift obtained by coregistration of the master
(16/08/2007) and slave (01/10/2007) images
In Figure 5, the streaks are broader and less
structured than those in Figures 4 and 6. In the case
of Figure 5, 4 large streaks can be identified:
a positive shift around 43°N,
a negative shift between 41.7°N and
42.8°N,
a positive shift between 41.5°N and 41.7°N,
a negative shift between 40.7°N and
41.5°N.
Since the SAR image of the 16 August is used
both in Figures 4 and 6 but not in Figure 5, the
ionosphere in 16 august 2007 seems to be the
principal responsible for the streaks appearance.
According to Dst (Figure 7) and Kp (not shown)
indices, 16 August 2007 can be considered quiet
from the geomagnetic point of view (red box of
middle panel). Also 1 July 2007 (red box, top panel
of Figure 7) and 1 October 2007 (red box, bottom
panel of Figure 7) can be considered quiet.
In Figures 8 to 10, the ΔTECcal maps
corresponding to the images #1 to #3, respectively,
are reported. All maps presents structures of
ΔTECcal, that are likely responsible for the presence
of streaks in the ALOS-PALSAR images.
Figure 7: Dst index of July, August and October 2007.
(http://wdc.kugi.kyoto-u.ac.jp/dst_final/). Days of the
images are highlighted in athe red boxes
We remind that, to highlight the correspondence
between azimuth streaks and gradients of ΔTECcal,
the ionosphere maps are fitted on SAR area,
according to the method described in Section 2.2. By
comparing ΔTECcal behaviour in the figures, we
can see that the ΔTECcal values show largely
different patterns among the three days, reinforcing
the idea on how variable can be the ionospheric
conditions in different days, even if at mid latitudes.
Figures 8-10 show patterns of ΔTECcal
characterized by large ΔTECcal gradients mainly in
the N-NW direction of all over the maps, and, in
particular, gradients of 1-2 TECu in the latitudinal
range between 42.2°N and 41.5°N, where the
strikes, highlighted in figures 4-6, are present. This
confirm how variation of few TECu’s and below
have a great impact on the SAR imaging.It is
important to note that ΔTECcal map in figure 8 is
obtained by differencing TEC data acquired in the
same season, while the maps in Figure 9 and 10 have
been obtained by differencing summer and fall data.
Considering that the geospace conditions can be
considered quiet for all the 3 days, the seasonal
variation seems to be a key point in understanding
the impact of the ionosphere on the streaks
appearance in the analysed case.
Assessing the Impact of TEC Fluctuations on
ALOS-PALSAR Images
19
Figure 8: ΔTECcal maps (01/07/2007-16/08/2007) The
color bar shows the differences of the ΔTECcal values in
TECu
Figure 9: ΔTECcal maps (01/07/2007-01/10/2007) The
color bar shows the differences of the ΔTECcal values in
TECu.
Figure 10: ΔTECcal maps (16/08/2007-01/10/2007) The
color bar shows the differences of the ΔTECcal values in
TECu
The spatial scale of the identified ΔTECcal
gradients has been reached thanks to the fine spatial
density of the GNSS receivers at ground. Such
resolution is not reachable with the standard, global,
ionospheric models commonly used.
4 CONCLUSIONS
TEC fluctuation can significantly affect InSAR
applications. To investigate these ionospheric effects
on ALOS-PALSAR images, we analysed 3 maps of
azimuth shift and 3 maps of TEC differences
(ΔTECcal) over central Italy.
The InSAR images present meaningful streaks,
with shift ranging from -3 m to 3 m.
The ΔTECcal maps were obtained by:
using dual frequency GPS data of the RING
network,
assuming that the ionosphere is a thin layer
located at an altitude of 350 km above the
Earth’s surface,
processing the GPS data with a calibration
technique to reduce the errors introduced by
satellite receiver biases
Fifth International Conference on Telecommunications and Remote Sensing
20
interpolating the data with natural neighbour
interpolation
In order to show the condition of the ionosphere
in terms of TEC fluctuation in correspondence with
the coregisistration of two L-band SAR images
(Fig. 4-6), the maps of ΔTECcal were projected on
the same area. Results indicate how variations below
1-2 TECu observed with ΔTECcal provides the
appearance of streaks. The map (Fig. 9) in which
the maximum value of ΔTECcal is larger
than those of the other 2 maps
the gradient are less and smaller than those
in Figures 8 and 10
corresponds to the case with less strikes.
Furthermore, the results show that the 16 August
2007 seems to be the principal responsible for the
streaks appearance.
These case study suggests how variations of few
TECUs can affect considerably ALOS-PALSAR
images. Further refinement of the work will aim at
compensating for the seasonal variation, aiming at
catching finer ionospheric effects and different
geospace conditions and geographical sectors will be
investigated. A comparison between the results with
the TEC mapping here shown and common
ionospheric models (like GIM, WBMOD) is
currently ongoing. Future refinement of the work
will aim at quantitatively assess the impact of TEC
fluctuation on L-band SAR, by evaluating the TEC
gradient along the azimuth direction. Effect of the
zonal and meridional ΔTECcal gradients will be also
investigated.
ACKNOWLEDGEMENTS
ALOS PALSAR data were provided by the
European Space Agency through Category-1
Proposal 26350. Authors are grateful to Dr. Antonio
Avallone for his support with RING data.
REFERENCES
Bechor, N. B., H. A. Zebker, Measuring two-dimensional
movements using a single InSAR pair, Geophys. Res.
Lett., 33, L16311, 2006, DOI:10.1029/
2006GL026883.
Bürgmann, R., Rosen, P. A., Fielding, E. J., Synthetic
aperture radar interferometry to measure Earth's
surface topography and its deformation. Annual
review of earth and planetary sciences, 28(1), 169-
209, 2000, DOI:10.1146/annurev.earth.28.1.169.
Cesaroni, C., Spogli, L., Alfonsi, L., De Franceschi, G.,
Ciraolo, L., Monico, J. F. G., Scotto, C., Romano, V.,
Aquino, M., Bougard, B., L-band scintillations and
calibrated total electron content gradients over Brazil
during the last solar maximum. Journal of Space
Weather and Space Climate, 5, A36, 2015a,
DOI:10.1051/ swsc/2015038.
Cesaroni C., A multi instrumental approach to the study of
equatorial ionosphere over South-America, PhD
dissertation, 2015b, DOI:
10.6092/unibo/amsdottorato/6889.
Chen, A. C., Zebker, H. A., Reducing ionospheric effects
in InSAR data using accurate coregistration. IEEE
Transactions on Geoscience and Remote Sensing,
52(1), 60-70, 2014,
DOI:10.1109/TGRS.2012.2236098.
Ferretti, A., Monti Guarnieri, A., Prati, C., Rocca, F.,
Massonnet, D., InSAR Principles: Guidelines for SAR
Interferometry Processing and Interpretation. The
Nederlands, Noordwijk: ESA Publications. TM-19,
2007, ISBN: 92-9092-233-8.
Foster, M.P., A.N. Evans. An evaluation of interpolation
techniques for reconstructing ionospheric TEC maps.
IEEE Trans. Geosci. Remote Sens., 46 (7), 2153
2164, 2008, DOI: 10.1109/TGRS.2008.916642.
Gray, A. L., Mattar, K. E., Sofko, G., Influence of
ionospheric electron density fluctuations on satellite
radar interferometry,Geophys. Res. Lett., 27(10),
14511454, 2000, DOI: 10.1029/2000GL000016.
Hanssen, R.F., (2001) Radar Interferometry: Data
Interpretation and Error Analysis. Kluwer Academic
Publishers, Dordrecht, 328 pp. ISBN: 0-7923-6945-9.
Kelley, M. C., The Earth's Ionosphere: Plasma Physics
and Electrodynamics (2nd ed.). Academic Press, 2009,
ISBN: 9780120884254.
Mannucci, A.J., B.D. Wilson, D.N. Yuan, C.H. Ho, U.J.
Lindqwister, T.F. Runge. A global mapping technique
for GPS-derived ionospheric total electron content
measurements. Radio Sci., 33 (3), 565, 1998, DOI:
10.1029/97RS02707.
Meyer, F. J., Agram, P. S. Developing an Error Model for
Ionospheric Phase Distortions in L-Band SAR and
InSAR Data. In FRINGE’15: Advances in the Science
and Applications of SAR Interferometry and Sentinel-
1 InSAR Workshop, Frascati, Italy, 23-27 March
2015.
Reuveni, Y., Bock, Y., Tong, X., Moore, A. W.,
Calibrating interferometric synthetic aperture radar
(InSAR) images with regional GPS network
atmosphere models, Geophys. J. Int. 202, 21062119,
2015, DOI: 10.1093/gji/ggv253
Sibson, R., A brief description of natural neighbor
interpolation (Chapter 2). In V. Barnett. Interpreting
Multivariate Data. Chichester: John Wiley. pp. 2136,
1981.
Zhu, W., Ding, X. L., Jung, H. S., Zhang, Q., Zhang, B.
C., & Qu, W. (2016). Investigation of ionospheric
effects on SAR Interferometry (InSAR): A case study of
Hong Kong. Advances in Space Research, 58, 564
576 2016, DOI: 10.1016/j.asr.2016.05.004.
Assessing the Impact of TEC Fluctuations on
ALOS-PALSAR Images
21