GMES Sentinel-1 Mission, System and Products

Paul Snoeij Evert Attema,Berthyl Duesmann,

Malcolm Davidson, Nicolas Floury, Betlem Rosich, Bjorn Rommen,

Guido Levrini

Overview

• Sentinel-1 System• C-SAR instrument• Data Compression• Conclusions

Sentinel-1 System• Space Segment

– A constellation of two satellites carrying a C-Band Synthetic Aperture Radar payload

– Nominal lifetime in orbit of 7 years / satellite (consumables for 12) – Global coverage – Near-Polar Sun-Synchronous dusk-dawn orbit @ 693km– Exact repeat period 12 days (for each satellite)– The second satellite in the same orbit but with a different Mean Anomaly

• Ground Segment– Mission operations for a system of satellites over a period of 20 years– S-Band station (Kiruna proposed), with a back-up for S/C contingencies– Downlink currently assumes three X-Band receiving stations

Sentinel-1 Mission Plan• Sentinel-1 has one main operational mode

(Interferometric Wide Swath mode) that:– satisfies most currently known service requirements– avoids conflicts and preserves revisit performance– provides robustness and reliability of service– simplifies mission planning & decreases operational costs– satisfies also tomorrow’s requests by building up a consistent

long-term archive• However

– Mutually exclusive modes are provided for continuity reasons (w.r.t. ERS & Envisat) and for accommodation of emerging user requirements

– Two mutually exclusive dual polarisation modes are provided

Observation Geometry

36.50

100 Km

200 Km

1

2

3

4

5

80 Km

Mode Access Angle

Single Look Resolution

Swath Width Polarisation

Strip Map 20-45 deg. Range 5 m Azimuth 5 m

> 80 km HH+HV or VV+VH

Interferometric Wide Swath

> 25 deg. Range 5 m Azimuth 20 m

> 250 km HH+HV or VV+VH

Extra Wide Swath

> 20 deg. Range 20 m Azimuth 40 m

> 400 km HH+HV or VV+VH

Wave mode 23 deg.and36.5 deg.

Range 20 m Azimuth 5 m

> 20 x 20 km Vignettes at 100 km intervals

HH or VV

For All ModesRadiometric accuracy (3 σ) 1 dBNoise Equivalent Sigma Zero -22 dBPoint Target Ambiguity Ratio -25 dBDistributed Target Ambiguity Ratio -22 dB

Sentinel-1 C-SAR Block

Diagram

Sentinel-1 T/R module single polarisation

Data compression• Raw SAR data compression is used to reduce the

instrument data rate to fit the available data downlink.• Application of a BAQ like raw data compressor adds

quantisation noise to the raw data which degrades the data SNR.

• A BAQ that applies a constant bit rate independent of radar signal power is not optimal. It produces a degradation of the SNR which varies with the power of the detected radar signal.

• FDBAQ controls the bit rate in a flexible way as a function of the input radar signal power.

The FDBAQ principle• FDBAQ (Flexible Dynamic Block Adaptive Quantization) exploits a Variable Bit Rate (VBR) scheme.• The number of quantization bits is selected according to a local estimate of clutter to noise ratio CNR.• FDBAQ principle:

FDBAQ quantization bits

estimation

Configuration parameters

InputblockSignal

Outputquantized

Signal

Input Signal Quantization

This approach leads to a Time variant bit rate

Bit Rate estimation Criteria• Non uniform quantization improves CQNR considering ECBAQ:

R = number of quantization bits

• FDBAQ: R selected applying the following criteria:

Thermal Noise + Quantization Noise ≤ fth = -22 [dB]

Mean radarBackscatter

NESZ(beam/orbitDependent)

R Estimation R[bit/sample]

1

N

R quantization levels

[ ] ][ 95.556.1 dBRNCCQNR dB +−==

( ) ( )595.0

156.0)(log)()(log)(2

1022

10 +−−+=

rfrrrR zthzc σσσ

Sentinel-1 Data Compression

-25

-22

-19

-16

-13

-10

-25 -20 -15 -10 -5 0 5 10Backscatter Coefficient (dB)

Tota

l Noi

se (d

B)

FDBAQ ECBAQ

Sentinel-1 Phase Noise

0

2

4

6

8

10

12

14

-25 -20 -15 -10 -5 0 5 10Backscatter Coefficient (dB)

Phas

e N

oise

(deg

)

FDBAQ ECBAQ

Bit Rate estimation using C-Band MosaicC-band GM Envisat Mosaic considered for FDBAQ performances evaluation

Mosaic absolutely calibrated

Gives an estimation of the mean radar backscattering coefficient

Mosaic doesn’t consider the reflectivity dependency with the incidence angle

C-Band Reflectivity models appliedNOTICE: ice-covered regions high reflectivity

Higher bit rate expected using

FDBAQ

Total Noise Power: theoretical performances

IW1 IW2 IW3

• High values of Total Noise Power are expected over ice-covered regions.

IW1 IW2 IW3

• Chosen test sites:

Sentinel-1 IW simulation

North America – (USA) (Land)

Greenland (Ice)

Atlantic Ocean (Sea)

• 3 different Sentinel-1 IW simulations (TopSAR acquisition mode).

• A typical Sentinel-1 orbit has been used.

• Requirement on mean Bit Rate:

Descending Orbit

Mosaic

Acquisition length ( for each Strip):

400 [s]

• 8 different Strips (Topsar IW) have been considered:• Reflectivity Model is applied: C-Band Ice Model over ice-covered areas.

C-Band Sea Model over sea areas

Sentinel-1 IW: Greenland Simulation

12345678

• Discrete quantization levels not applied to estimated rate R.

Greenland Simulation: quantization bits

Greenland Simulation: Noise Power Estimation

-25 -22-23.5

STRIP 4

-25

-22

Estimated Mean Bit Rate Greenland Discrete

quantization levels NOT applied

Quantization levels applied

[3 3.4 3.8 4.2 5]

Mean Bitrate(along all beams and strips):

245.39 [Mbit/s]

Mean Bitrate(along all beams and strips):

245.08 [Mbit/s]

Conclusions• Sentinel-1 System introduced• Dual channel receiver for simultaneously

receiving co- and cross-polar signals• A technique for the evaluation of FDBAQ

quantizer performances has been presented• FDBAQ performances have been evaluated

A figure of merit (Total Noise Power ) has been evaluatedFDBAQ reduces mean bit rate by 20%

Thank you for

your attention