GNSS-R, an Innovative Remote Sensing Tool for the Mekong Delta, Jamila Beckheinrich, Brest 2013 Slide 1 GNSS Reflectometry, an Innovative Remote Sensing Tool for the Mekong Delta J. Beckheinrich 1, S. Schön 2, G. Beyerle 1, M. Semmling 1, H. Apel 1, J. Wickert 1 1.GFZ, German Research Center for Geosciences, Potsdam, Germany 2.Leibniz University Hannover, Institute of Geodesy, Hannover, Germany Space Reflecto 2013 Brest, France

WP within the WISDOM project Experimental Setup Data Analysis Latest Results Conclusion and Outlook Slide 2

WP within the WISDOM project Experimental Setup Data Analysis Latest Results Conclusion and Outlook Slide 3

Slide 4 WP within the WISDOM Project: Severe changes could be observed in the Mekong Delta: extreme flood events important to monitor coastal area with dense population Test the possibility of using low elevation GNSS- Reflectometry for flood monitoring of the Mekong Delta Develop and implement algorithms to process GPS signals into water levels

WP within the WISDOM project Experimental Setup Data Analysis Latest Results Conclusion and Outlook Slide 5

Experimental setup: Slide 6 Gauge Measurement location 6 km 150 m

Experimental setup: Two time series with two different antenna heights above the reflecting surface: Terrace with ~10 m Roof with ~20 m Slide 7

Experimental setup: Two antennas (Antcom): RHCP oriented to the zenith (direct signal) LHCP tilted to the reflecting surface (reflected signal) Slide 8 L1/L2 GPS RHCP RHCP/ LHCP L1/L2 GPS Direct signal Reflected signal

Experimental setup: Slide 9 a: GORS b: Control unit c: Laptop d: External Hard Disk

WP within the WISDOM project Experimental Setup Data Analysis Latest Results Conclusion and Outlook Slide 10

Antenna height ~ 10 m~ 20 m Recording time ~84 h~60 h Reflection events ~194 h~ 191 Continuous phase observations > 2 min. ~26 %~ 29 % Continuous phase observations > 10 min ~22 % Slide 11 Data Analysis:

Slide 12 Data Analysis: Roof (~20 m) Hotel River Terrace (~10 m) Incoherent Observations Coherent Observations

Slide 13 Data Analysis:

Slide 14 Data Analysis: River Receiver Reflection track

Empirical Mode Decomposition (EMD) Method: 1998 Huang [1] 2008 and 2013 Hirrle [2] Motivation: Fourier Transform assume stationarity and linearity Wavelet Transform assume linearity EMD is a fully data-driven signal analysis No need of an a priori base Slide 15 Data Analysis:

Slide 16 Intrinsic Mode Functions (IMF) Data Analysis: IMF Trend Input signal

Slide 17 Data Analysis: Input Signal

Slide 18 Data Analysis: Maxima and minima

Slide 19 Data Analysis: Cubic spline interpolation

Slide 20 Data Analysis: Mean

Slide 21 Data Analysis: Resulting Signal

Slide 22 Intrinsic Mode Functions (IMF) Data Analysis: IMF Trend Input signal

Slide 23 Data Analysis: Trend

Slide 24 Data Analysis: Horizontal Reflector h = 13 m Vertical Reflector h = 1 m

Slide 25 Data Analysis: IMF 1 vs Horizontal Reflection IMF 2 vs Vertical Reflection Completness

Slide 26 Data Analysis: PRN 12, Vietnam, Can Tho City, 28th February 2012 Frequency

Slide 27 Frequency Data Analysis: PRN 12, Vietnam, Can Tho City, 28th February 2012

Slide 28 First results: EMD Observations

WP within the WISDOM project Experimental Setup Data Analysis Latest Results Conclusion and Outlook Slide 29

Slide 30 Latest Results: EMD Beckheinrich J. et al, GNSS Reflectometry: Innovative Flood Monitoring at the Mekong Delta, submitted JSTARS august 2013

Slide 31 Latest Results: Vietnam, Can Tho City, 29th February 2012 With EMD: std (1σ) = 0.05 m Without EMD: std (1σ) = 0.12 m

Slide 32 Latest Results: Vietnam, Can Tho City, 29 th February 2012

WP within the WISDOM project Experimental Setup Data Analysis Latest Results Conclusion and Outlook Slide 33

Although the measurement were made under challenging conditions we could reach a correlation of 0.84 with the data recorded from a gauge instrument GNSS-R could contribute to monitor water level changes of the Mekong Delta Improvement of stochastic and functional model Determination of sigma reflected signals (Experiment) Include Phase Wind up effect Amplitude Better understanding of the EMD parameters and there impact on the IMF decomposition Slide 34 Conclusion and Outlook:

[1]Huang N.E. et al., The Empirical Mode Decomposition and the Hilbert Spectrum for non linear and non-stationary time series analysis, Proc. R. Soc. London A, (1998), pp [2]Hirrle A. Et al., Estimation of Multipath Parameters using Hilbert Huang Transform, ION GNSS, Nashville, Tennessee, USA, 2012 [3]Beckheinrich J. et al., GNSS Reflectometry: Innovative Flood Monitoring at the Mekong Delta, submitted at JSTARS, August 2013 Slide 35 References:

Thank you for your attention Slide 36

Slide 37 First Fresnel zone, 10 m height antenna

First Fresnel zone, 20 m antenna height Slide 38