N. Pierdicca1, L. Guerriero2 , R. Giusto1, M. Brogioni3, A. Egido4, N. Floury5

1 DIET - Sapienza Univ. of Rome, Rome, Italy2 DISP - University of Tor Vergata, Rome, Italy3 CNR/IFAC, Sesto Fiorentino. Italy4 Starlab, Barcelona, Spain5 ESA/ESTEC, Noordwyik, The Netherland

GNSS Reflection from Bare and GNSS Reflection from Bare and Vegetated Soils: Experimental Vegetated Soils: Experimental Validation of an End-to-end Validation of an End-to-end SimulatorSimulator

2

ContentContent

Introduction: the Leimon project Simulator description Simulator validation during Leimon Conclusions

3

Land MOnitoring with Navigation signals Evaluating the potential of GNSS signals for remote

sensing soil moisture and vegetation biomass, through a ground based experimental campaign

LEiMON ProjectLEiMON Project

Developing a simulator to theoretically explain experimental data and predict the capability of a GNSS-R aerospace systems

The SAM dual pol GNSS-R instrument (by Starlab)

LEiMON set up

4

The LEIMON experimentThe LEIMON experimentFDR and TDR soil moisture probes

Meteo station

Soil roughness profilometer

Bare (different roughness) and vegetated fields

Plant height and water content

5

Time serie overviewTime serie overview

West field with developed plants

6

|Y |2 Processed signal power at the receiver vs. delay and frequency f.PT The transmitted power of the GPS satellite.GT , GR The antenna gains of the transmitting and the receiving instrument.RR, RT The distance from target on the surface to receiving and transmitting

antennas.Ti The coherent integration time used in signal processing. Bistatic scattering coefficient2 The GPS correlation (triangle) function S2 The attenuation sinc function due to Doppler misalignment dA Differential area within scattering surface area A (the glistening zone).

The mean power of received signal vs. delay and frequency f is modeled by integral Bistatic Radar Equation which includes time delay domain response ’ and Doppler domain response S2 (f’-f ) of the system (Zavorotny and Voronovich, 2000).

dARR

ffSGGPTfY

TR

RTTi 022

22

3

222

)'()'(

)4()ˆ,ˆ(

The Bistatic Radar EquationThe Bistatic Radar Equation

7

The BRE integrationThe BRE integration

In order to solve the integral we have to introduce the equations relating all those variables to the coordinates over the surface.

dARR

ffSGGPTfY

TR

RTTi 022

22

3

222

)'()'(

)4()ˆ,ˆ(

function of point scattering delay

function of point looking direction wrt boresight

function of incidence direction function of point

wrt RX and TX position

function of point bistatic angle (zenith and azimuth)

function of point scattering Doppler

Integrand variables are the coordinates defined on the surface

8

Local vs global framesLocal vs global frames

Elliptic Earth approximation x’y’z’ local frame centered in the specular point z’ axes along the geodetic vertical x’z’ plane coincident with the bistatic scattering plane

z’

x’

y’ TX

RX

SP=[xs,ys,zs]ECEFF

Earth ellipsoide

is =i

Earth surface

x

z

y

Given RX and TX trajectory/velocity & epoch, integral is computed in the local frame, where soil/vegetation parms are defined

9

° electromagnetic modelling° electromagnetic modelling

Indefinite mean surface plane with roughness at wavelength scale. Bistatic scattering of locally incident plane waves by AIEM

Homogeneous vegetation cover. Attenuation and multiple scattering by a discrete medium (Tor Vergata model)

INCOHERENT

component

COHERENT component

Scattering of spherical wave by Kirchoff approxi. (Eom & Fung, 1988) attenuated by vegetation

10

Polarization synthesis for real antennasPolarization synthesis for real antennas

hhhv

vhvvBSAtBSAr

tinc

rrec

rt ss

ss

P

PrSpSp 4

4 22

Nominal polarization unit vector are pRHCP=(0-j0)/√2 and pLHCP=(0+j0)/√2

Real antenna polariz. unit vector

SCS of “real” antennas and channel cross talk

x

y

z P=r,,

2

00

cos1

cos

jRHCP θp

2 orthogonal electric dipoles /2 phase shifted

2

2

1

2

1

21

cos

cos1

14

jjLR S

11

The signal simulator structureThe signal simulator structure

Geometric module:• GPS TX orbit• SP position

E.M. bare soil bistatic ° model

Simulator core

E.M. vegetation bistatic ° model

12

AIEM vs scattering direction AIEM vs scattering direction ss,,ss

i=31°mv=20 % z=1.5 cml=5 cmHRX=10 kmHPBW=120°

Incoherent component RR & LR more spread

Coherent component RR & LR focused around specular s=31°

Azi

mu

th

Azi

mu

th

Zenith Zenith

13

DDM output exampleDDM output exampleDDM’s (delay on the horizontal axes, frequency on the vertical axes) for incoherent (top) and coherent (bottom) component at RHCP (left) and LHCP (right)

Spaceborne incoherent

airbone incoherent

14

Peak power output examplePeak power output example

mv=20 % z=1.5 cml=5 cmHRX=10 kmVRX=180 m/sHeadRX=45°HPBW=120°

Coherent and incoherent RR & LR peak power as the satellite moves along the orbit

15

The incoherent signal fluctuates (speckle) with long correlation time (15 sec) so only incoherent summation was feasible

Different soil/vegetation conditions investigated

Leimon dataLeimon data In the Leimon experiment the instrument records

temporal series of the direct and reflected signals (I&Q of the correlator output peak)

The mean reflected power (both LR and RR pol) normalized to the mean direct power (at RR pol.) is studied vs. the incidence angle

reflected

direct

16

August 26th SMC=10%

• East Z=0.6 cm

• West Z=1cm

April 8th SMC=30%

• East field Z=3cm

• West field Z=1.75cm

Simulator reproduces quite well LR signal versus at incidence angles ≤45°.

Validation: angular trendValidation: angular trend

17

August 26th SMC=10%

• East Z=0.6 cm

April 8th SMC=30%

• East Z=3cm

Theoretical simulations show that incoherent component strongly contributes to total signal when soil is rough.

Coherent vs. incoherent: soilCoherent vs. incoherent: soil

coherent

total

18

General overestimation of the DOWN/UP ratio but good correlation

Cross talk and/or different down and up antenna gains could explain the bias

RMSE=1.6 dB Bias=-1.4 dB

Bare soil overall comparison LRBare soil overall comparison LREast Side, West Side

Bare Soils 10%<SMC<30%

0.6<z<3cm

19

RMSE=7.25 dBBias=6.1 dB

The model underestimates RR polarization, especially at smaller angles

The RR measurements seem to saturate at -18 dB A deficiency of the mode could occur

Bare soil overall comparison RRBare soil overall comparison RREast Side, West Side

2020

RR underestimationRR underestimation

RR August 26th SMC=10% Z=0.6cm

At large incidence angle, the RR coherent component is the dominant one

Theoretical incoherent RR scattering (rough case) may be underestimated

RR April 8th SMC=30% Z=3cm

coherent

total

21

Vegetation overall comparison

Fair correspondence between model and vegetation data (except for largest angle =55°)

June 28, July 10,28

22<SMC<18%PWC=1.2, 3.6, 6.7 kg/m2

22

Sensitivity to SMCSensitivity to SMCBlack: Leimon data (May) and regression line Blue/Red: Theoretical simulations, two roughnesses

Model predicts 2.5 dB for a 10% SMC difference Leimon shows 2.8 dB

23

Sensitivity to vegetation Sensitivity to vegetation Magenta: Leimon data Green: Theoretical simulations vs. PWC

Model sensitivity reproduces the measured oneSensitivity to vegetation is quite low: about 2 dB for

the whole PWC range

24

Coherent vs Incoherent: vegetationCoherent vs Incoherent: vegetation

At 35°, the coherent component is attenuated by about 1 dB each 1kg/m2, which would predict a large sensitivity to PWC

The Simulator predicts a quite large incoherent component which explains the saturation effect with PWC in the data.

The model reproduces the measured trend thanks to the consideration of both component

25

ConclusionsConclusions A simulator has been developed which provides DDM’s

and waveforms of a GNSS-R system looking at land It takes as input arbitrary receiver position/velocity, in view

GPS satellite PRN code, surface properties (soil moisture, roughness, vegetation parameters).

It singles out coherent and incoherent signal components. A fair agreement has been found at LR polarization and

angles <45° (the antenna beamwidth) Incoherent component may be high in a ground based

configuration, and antenna properties must be considered Sensitivity to SMC is significant and well reproduced Sensitivity to vegetation is reproduced and it is quite low

because of the coherent and incoherent combination. A few discrepancies could be due to measurement or

model problems (of course) and are under investigation

26