Sentinel-1 SAR System and Mission

Ramón Torres, Ignacio Navas-Traver, David Bibby, Svein Lokas, Paul Snoeij, Björn Rommen, Steve Osborne, Francisco
Ceba-Vega, Pierre Potin and Dirk Geudtner
European Space Agency, ESTEC and ESRIN
Noordwijk, The Netherlands and Frascati, Italy
[email protected]

Abstract— The paper provides an overview of the Copernicus observation of the marine environment, including oil spill
Sentinel-1 mission, which includes Sentinel-1A (S-1A) and detection and Arctic/Antarctic sea-ice monitoring, the
Sentinel-1B (S-1B) satellites, and the characteristics of its SAR surveillance of maritime transport zones (e.g. European and
system. Sentinel-1 uses pre-programmed SAR mode operations North Atlantic zones), as well as the mapping of land surfaces
to provide a high revisit frequency and systematic global SAR
including vegetation cover and mapping in support of crisis
image coverage. This is mainly based upon the operational use
of the novel TOPS (Terrain Observation with Progressive Scans situations, such as natural disasters and humanitarian aid [1].
in azimuth) SAR imaging mode. Sentinel-1A, the first component, was launched on 3 April
In particular, we present results of the SAR system performance
2014. It was commissioned in orbit during the following
analysis focusing on the instrument stability and the achieved
radiometric accuracy, as well as the Noise Equivalent Sigma
months leading to an In-Orbit Commissioning Review
Zero (NESZ). In addition, we discuss the cross- S-1A/S-1B SAR (IOCR) in September 2014 and entry in operation in October
Interferometry (InSAR) performance considering the effects of 2014. The identical Sentinel-1B was launched on 25 April
burst synchronization and SAR antenna pointing on the 2016, two years after Sentinel-1A, commissioned to an IOCR
achievable common Doppler bandwidth. Results of differential on 14 September 2016 and an immediate entry in operation,
cross interferograms are presented showing the coseismic to complete the constellation with Sentinel-1A.
surface displacement caused by the central Italy earthquake. Both Sentinel-1 models have been designed for a nominal
operational lifetime of 5 years, however with internal
Keywords—Sentinel-1; SAR performance, TOPS; InSAR
resources to extend the life for additional 5 more years.
Nevertheless, new Sentinel-1 models, i.e. Sentinel-1C and
I. INTRODUCTION Sentinel-1D, are already being developed to be ready in time
In the framework of the European Commission to replace the Sentinel-1A and Sentinel-1B satellites.
(EC)/European Space Agency (ESA) joint Copernicus Sentinel-1A (S-1A) and Sentinel-1B (S-1B) fly in a near
program, ESA is undertaking the development of a series of polar, sun-synchronized (dawn-dusk) orbit at 693 km altitude
six Sentinel missions with the objective to provide routinely and in the same orbital plane with 180 deg. phased positions.
Earth observation data for the implementation of operational The 12-day orbital repeat cycle of each satellite, which
Copernicus and national services. The EC Copernicus means 6 days for the entire Sentinel-1 constellation, along
services comprise operational mapping, monitoring and with the use of pre-programmed SAR mode operations,
forecasting activities for Land, Marine, Atmosphere, provides a high revisit frequency and systematic global SAR
Emergency, Security, and Climate change. image coverage.
In addition, the 6-day repeat orbit interval along with small
II. SENTINEL-1 MISSION orbital baselines enables cross-SAR interferometry (InSAR)
coherent change detection applications, such as surface
The Sentinel-1 Mission, as part of the Space Component of deformation mapping and cryosphere dynamics.
Copernicus, is based on a constellation of two satellites to
fulfil revisit and coverage requirements, providing robust III. SENTINEL-1 SAR SYSTEM
datasets for Copernicus Services.
The Sentinel-1 mission has been specifically designed to A. Sentinel-1 SAR Imaging Modes
acquire systematically and provide routinely data and The Sentinel-1 SAR instrument with its active phased
information products for Copernicus Ocean, Land and array antenna supports four exclusive imaging modes
Emergency services, as well as to national user services. providing different resolution and coverage: Interferometric
These services focus on operational applications such as the Wide Swath (IW), Extra Wide Swath (EW), Stripmap (SM),

978-1-4673-8823-8/17/$31.00 ©2017 IEEE 1582

and Wave (WV), see Fig. 1. All modes, except the WV mode The instrument internal calibration is based on internally
can be operated in dual polarization. generated chirp signals that are routed through almost the
complete transmit and receive signal paths plus dedicated
calibration paths comprised of very stable passive elements.
As the calibration signals experience the same gain and
phase changes as the nominal radar signal, calibration
measurements are used by the SAR processor on ground to
compensate for variations in gain, phase and internal delay.
The Sentinel-1B instrument stability measured by
calibration pulses during a 25 minutes datatake is shown in
Fig. 2. During this datatake, the temperature of the Transmit/
Receive modules (TRMs) increased by 20°C, the gain varied
by 0.3 dB and the phase changed by 20°.

Figure 1. Sentinel-1 SAR imaging modes

Both the IW and EW mode operate in the novel TOPS,

which utilize ScanSAR-type burst imaging with an additional
antenna beam steering in azimuth referred to as Terrain
Observation (Terrain Observation with Progressive Scans in Figure 2. Instrument stability during a long datatake as measured by
calibration pulses
azimuth) mode [2]. The Sentinel-1 IW and EW TOPS modes
provide a large swath width of 250 km and 400 km, Fig. 3 shows the long-term Sentinel-1B instrument stability
respectively with enhanced image performance as compared during a period of 2.5 months by combining measurements
to the conventional ScanSAR mode. In fact, the IW TOPS from different SAR modes and sub-swaths.
mode is the main mode of operations for the systematic data
acquisition over land and coastal areas. The EW mode is
mainly used for sea-ice mapping and oil spill monitoring.
In the IW and EW TOPS imaging modes, the SAR
antenna azimuth beam is steered from aft to the fore at a
constant rate. As a result and contrary to ScanSAR, all targets
on ground are observed by the complete azimuth antenna
pattern eliminating almost entirely the scalloping effect and
achieving constant azimuth ambiguities and signal-to-noise
ratio (SNR) along azimuth. Figure 3. Long term instrument stability measured by calibration pulses
Key parameters of the Sentinel-1 SAR imaging modes are
summarized in Table 1. The dedicated RF Characterization (RFC) mode is used
for the in-orbit monitoring of the health and stability of the
TABLE 1. SUMMARY OF SENTINEL-1 SAR IMAGING MODES
individual TRMs, whilst all modules are operating under
Mode Incidence Chirp Single-Look Ground realistic thermal conditions and nominal power loads.
Angle Bandwidth Res. (rg x az) [m] Variations in the TRMs can be used as input to the Sentinel-1
[deg] [MHz] Antenna Model to predict the resulting changes in the antenna
SM (S1-S5) 20-43 87.6-42.2 5x5 patterns. Fig.4 depicts the permanent failure of 2 TRMs,
IW (1-3) 30-42 56.5-42.8 5 x 20
which occurred during the S-1B Commissioning phase and
resulted in a negligible degradation of the overall system
EW (1-5) 20-44 22.2-10.4 20 x 40 performance.
WV (1&2) 23 & 36.5 74.5 & 48.2 5x5

B. Sentinel-1 SAR Instrument Performance

The Sentinel-1 SAR instrument utilizes an internal
calibration system that measures and monitors in-orbit short-
term variations of the instrument gain and phase resulting
from temperature variations during imaging, but also long-
term changes due to seasonal variations and ageing effects of
the hardware such as the degradation of thermal surfaces. Figure 4. Antenna health monitoring based on RFC measurements

978-1-4673-8823-8/17/$31.00 ©2017 IEEE 1583

 The Sentinel-1 radiometric accuracy was verified based Using Sentinel-1A/B InSAR scene and long data take
on Radar Cross Section (RCS) measurements in SAR images pairs, as well as multi-temporal stacks, we measured only
acquired over the ESA radar calibration transponders. Fig. 5 very small offsets and variation of the burst synchronization
depicts the RCS measurements using 15 S-1B dual-pol IW of 3ms, see Fig. 7.
mode images. The radiometric accuracy for all SAR modes
on both S-1A and S-1B was measured to be better than 0.5dB
and the overall calibration factor was adjusted by 1.5 dB.

Figure 7. Burst synchronization results for S-1A/S-1B IW scene pair

(8 bursts) acquired over the Salar de Uyuni site, Bolivia (left) and a long
datatake (10 slices) acquired over China

The burst mis-synchronisation can be directly converted into

Figure 5. Sentinel-1 RCS measurements over ESA transponders a difference in Doppler centroid difference [5]. As a result,
(left: Sentinel-1A RCS, right: Sentinel-1B relative RCS)
the InSAR common Doppler bandwidth is larger than 95% of
the available total Doppler bandwidth (300 Hz), see Fig. 8.
The Noise Equivalent Sigma Zero (NESZ) was measured
using slant range profiles of cross-polarized (HV & VH) SAR
data acquired over the Pacific Doldrums. The Image data
consists mainly of receiver noise after correction of the
elevation antenna pattern. On average the Sentinel-1 NESZ is
better than -25dB. Fig. 6 depicts the NESZ measurements for
S-1B IW and EW HV mode.

Figure 8. S-1A/S-1B cross-burst synchronization (left) and resulting loss

of common Doppler bandwidth (right)
The azimuth antenna pointing is very stable for both S-1A
and S-1B due the total zero Doppler attitude steering of the
spacecraft causing only a very small difference in Doppler
centroid of 20Hz when forming S-1A and S-1B,
interferograms, respectively. However, during the
Figure 6. Sentinel-1B IW and EW HV mode NESZ measurements vs. Commissioning of S-1B, we measured an initial antenna
theoretical (design) values pointing offset (i.e. yaw and pitch) that is equivalent to a
mean Doppler centroid difference of 200Hz, see Fig. 9, due to
IV. SENTINEL-1 SAR INTERFEROMETRY a relative star tracker misalignment. However, due to the
PERFORMANCE TOPS antenna beam steering the resulting effective Doppler
centroid difference is smaller than the mean Doppler centroid
The generation of high-quality Sentinel-1A &B TOPS IW
difference. As a result, the common Doppler bandwidth of S-
mode cross-interferograms and coherence maps requires an
1A/S-1B cross-interferograms was temporarily reduced to
accurate synchronization of the azimuth scanning patterns
65%. However, at the end of the S-1B Commissioning phase,
(i.e. bursts) and a very stable azimuth antenna pointing to
the S-1B antenna mis-pointing was corrected to achieve a
achieve the maximum common Doppler bandwidth. (i.e.
Doppler centroid difference of 20Hz, which is equivalent to
azimuth spectral alignment).
more than 95% of the common Doppler bandwidth.
300
C. Burst Synchronization and Aziumuth Antenna Pointing 200
S1A
S1B
Doppler centroid [Hz]

To achieve highly accurate burst synchronization, i.e. the 100

instrument burst data acquisition from repeat-pass orbits starts 0

at the same time over the same location on the ground, the S-1
-100
mission exploits the novel concept of position-tag
commanding using the orbit position angle and its on-board
-200

conversion into sensing time along with an orbital point grid

-300
Jan 2015 May 2015 Sep 2015 Jan 2016 May 2016 Sep 2016

[4]. The calculation of the burst synchronization is based upon

Figure 9. Mean Doppler centroid difference for S-1A and S-1B IW image
the use of the orbital state vectors and the annotated azimuth stack acquired over Chicago, USA
start time.

978-1-4673-8823-8/17/$31.00 ©2017 IEEE 1584

D. Orbital Tube and InSAR Baseline and routine operational activities. The Sentinel-1 mission is
Both Sentinel-1 satellites are kept within a tube around a operated via a pre-defined observation plan. All data
reference orbit. This is implemented via a ground-track dead- acquisitions are systematically downlinked and processed to
band of ±120m, which in turn is equivalent to an orbital tube generate a pre-defined list of core products within specific
with a radius of less than 100 rms. This orbit maintenance timeliness.
strategy leads to small perpendicular S-1A/S-1B cross-InSAR
G. Observation Scernario
baselines of ± 100m, see Fig. 10.
200
With both Sentinel-1 satellites in orbit, most of the global
S1A
S1B
landmasses are mapped on average every 12 days at least.
100 The European land and coastal waters are systematically
Perp. baseline [m]

mapped within the 6-day constellation repeat cycle in both

0
ascending and descending passes, using dual-pol (VV+VH).
-100
Tectonic/volcanic areas outside Europe are covered,
alternating between ascending and descending passes, each
-200
Jan 2015 May 2015 Sep 2015 Jan 2016 May 2016 Sep 2016
within a 24-day repeat-pass interval.
Areas for sea-ice and iceberg monitoring are observed
Figure 10. Perpendicular cross-InSAR baseline for S-1A/S-1B IW data stack almost daily. Also, specific observation campaigns are
acquired over Beijing, China performed for the Greenland and Antarctica ice-sheets.
The current observation plan is summarized in Fig. 13,
E. Sentinel-1A/ Sentinel-1B Cross-InSAR and also available at the following website:
The cross-InSAR capability of the Sentinel-1 sentinels.copernicus.eu/web/sentinel/missions/sentinel-
Constellation to map wide-area, small-scale surface 1/observation-scenario
deformation has been successfully demonstrated by the
generation of differential cross-interferograms of the area
affected by the M 6.2 central Italy earthquake that occurred
on Aug. 24th, 2016 using S-1B and S-1A IW mode image data
acquired on Aug. 21st and Aug. 27th, 2016, respectively. The
high quality (i.e. no phase jumps between subsequent bursts
and neighboring sub-swaths) of this coseismic deformation
map, see Fig. 11, demonstrates the excellent compatibility
(i.e. synchronization of burst image timing) and stability of
the radar instruments on both satellites, as well as the
accurate orbit control of both spacecraft.
Figure 13. Sentinel-1 Constellation Observation Scenario

ACKNOWLEDGMENT
The authors would like to thank the Sentinel-1 Project and Mission
Teams, as well as industry, especially Thales Alenia Space and
AIRBUS for their on-going support.

REFERENCES
[1] Torres,, et al. “GMES Sentinel-1 Mission”, Special Issue of Journal of
Remote Sensing of Environment “The Sentinel Missions – New
Opportunities for Science”, Vol. 120, pp. 9-24, May 2012.
[2] F. De Zan and Andrea Monti Guarnieri, “TOPSAR: Terrain
Observation by Progressive Scans”, IEEE Transactions on Geoscience
and Remote Sensing, Vol. 44, No. 9, September 2006, pp 2352-2360.
Figure 12. S-1A/S-1B IW mode cross differential interferogram of Italy [3] P. Prats et al., “Processing of Sliding Spotlight and TOPS SAR Data
earthquake. The effective InSAR baseline is 28.1m, the burst mis- Using Baseband Azimuth Scaling”, TGARS, Vol. 48, no. 2, pp. 770-
synchronization is 3.12 ms and the mean Doppler centroid frequency is 110 780, February 2010.
Hz (S1B) and 54 Hz (S1A), respectively. [4] D. Geudtner et al. “Sentinel-1 SAR Interferometry Performance
Verification”, Proc. EUSAR 2016, Hamburg, Germany.
V. SENTINEL-1 MISSION OPERATIONS CONCEPT [5] Yague-Martinez, N., Prats, P., Gonzalez, F.R., Brcic, R., Shau, R.,
Eineder, M., Geudtner, D. and Bamler, R., “Interferometric Processing
F. Operations Concept of Sentinel-1 TOPS Data”, IEEE Trans. Geoscience and Remote
The main objective of the Sentinel-1 operations strategy is Sensing, Vol. 54, No. 4, pp. 2220-2234, 2016.
to ensure a reliable provision of data to users with systematic

978-1-4673-8823-8/17/$31.00 ©2017 IEEE 1585