National Oceanic and Atmospheric Administration
Environmental Technology Laboratory

Bistatic GPS Signal Scattering from an Ocean Surface: Theoretical Modeling and Wind Speed Retrieval from Aircraft Measurements

V. U. Zavorotny

Cooperative Institute for Research in Environmental Sciences,
University of Colorado/NOAA, Environmental Technology Laboratory
325 Broadway, Boulder, CO 80303-3337
email: Valery.Zavorotny@noaa.gov
Phone:303-497-6616; Fax: 303-497-6181

This work was supported by NASA Headquarters under the Physical Oceanography Program RTOP 622-47-55

Workshop on Meteorological and Oceanographic Applications of GNSS Surface Reflections: From Modeling to User Requirements July 6th 1999, De Bilt, The Netherlands


Acknowledgments
The presentation is based on results obtained by:

  • P. Axelrad
  • G. Born
  • J. Garrison
  • S. Katzberg
  • A. Komjathy
  • A. Voronovich
  • V. Zavorotny


Outline of talk


Introduction

  • GPS gave birth to many new applications
  • The bistatic geometry, circular polarization and special temporal structure of GPS signals provide unique opportunities for remote sensing
  • Currently the feasibility of an airborne GPS delay- mapping scatterometer is shown
  • Ocean-surface state and wind speed retrieval has been demonstrated using this scatterometer
  • A satellite GPS scatterometer is under the consideration
  • A concept of the high-resolution, fully-polarized, delay-Doppler-mapping, bistatic GPS radar is proposed
  • Other applications: wave height, wind vector, wetland, salinity, ice state determination


Cox&Munk photographs of the sun's glitter (to give an idea how surface reflections form a glistening zone)



Geometry of multistatic GPS radar system


By employing code-range measurements, a receiver can distinguish particular annulus zones on the ocean surface illuminated by GPS signals.


Modeling of GPS bistatic surface scattering

In the GPS receiver the signal at time t0 is cross-correlated with a replica of PRN code a taken at a different time t0 -

(1)

The average power density as a function of a time delay :

(2)

Eq.(2) is a bistatic-radar equation with the function D2 2 |S|2as a resulting radar footprint. An annulus zone appears due to the code-correlation process:

(3)

Function |S|2 defines a Doppler zone. The scattering cross-section 0 in the geometrical optics leads to quasi-specular scattering:

(4)

W is the PDF of surface slopes. The width of 0 in terms of defines a glistening zone
For isotropic Gaussian slopes:

(5)


Size of the Doppler and annulus zones as a function of the receiver amplitude for 90 degree elevation angle


Click for full-size image.


Doppler shift along longitudinal coordinate for different elevation angles.


Click for full-size image.


Relative position of the Doppler zone function and annuli at different time delays for 90 degree elevation angle.


Click for full-size image.


Relative position of the Doppler zone function and annuli at different time delays for 40 degree elevation angle.


Click for full-size image.


Model for ocean surface statistics

  • Statistics of the slopes is assumed to be Gaussian, with anisotropic surface slopes having wind-dependent variances and correlations.
  • These parameters can be derived from a solely wavevector spectrum of full surface elevations by integrating it over wave numbers smaller than a dividing parameter of the two-scale model


  • Two spectral models for sea roughness were used in the case of well-developed seas: (a) J. R. Apel, JGR, 99, 16,269, (1994); (b)T. Elfouhaily, et al. JGR, 102, 15,781, (1997).


Comparison between mean-square slope (mss) values obtained from Apel (solid) and Elfouhaily (dashed curve) spectra.


Click for full-size image.


Waveforms for upwind and downwind for the Elfouhaily spectrum
(alt = 5 km, elev = 45 deg)


Click for full-size image.


Waveforms for various receiver altitudes and wind speeds
at =45° (a) h = 1 km; (b) h = 10 km; (c) h = 300 km.


Click for full-size image.


Waveforms with and without taking into account Doppler spreading for h = 5 km, =45°, vrec=0.17km/s.


Click for full-size image.


Waveforms obtained at different integration times Ti
(h=225 km, =45°, vrec=7.9km, U=8m/s, c=0.1 µs)


Click for full-size image.


Aircraft experiments with the delay-mapping GPS receiver

  • For the receiver, a commercially available GPS-development kit from GEC Plessey (now MITEL) including a 2021 correlator chip fed by two 2010 RF front ends were used
  • Two low-gain L-band antennas were used: a zenith-oriented RHCP antenna, and a nadir-oriented LHCP antenna
  • Aircraft remote sensing experiments started since 1997
  • November 1997: NASA Stennis Space Center (NASA Super King Air B-200 aircraft; Gulf of Mexico)
  • May 1998: Wallops Island (C-130 aircraft; Gulf Stream )
  • August 1998: Wallops Island (Virginia Space Grant Consortium Balloon Experiment)


LHCP antenna on bottom of fuselage of Langley B-200 aircraft.


Delay-mapping receiver hardware in the Langley B-200 aircraft.


The delay-mapping GPS receiver


The shape of the waveform versus code delay is a function of ocean surface roughness and hence wind speed.


The experimental waveforms for PRN21 agrees well with the theoretical model for wind speeds 8-10 m/s.


Scheme illustrating the formation of footprints.


Sensitivity to the wind direction: Signal as a function of time delay and Doppler offset
(H=5km, U=8 m/s, wind dir. = 0 deg. Ti=10ms)


Click for full-size image.


Sensitivity to the wind direction: Signal as a function of time delay and Doppler offset
(H=5 km, U=8m/s, wind dir. = 90 deg., Ti=10 ms)


Click for full-size image.


Sensitivity to the wind direction: Function versus Doppler offset for six wind directions
(H=5km, U=8m/s, teta=90°, Ti = 10ms)


Click for full-size image.


Fully-polarized delay-Doppler mapping using GPS signals

  • The current delay-mapping technique has several limitations: the footprint is an elliptical zone, a spatial resolution is low, a relatively low SNR, only a left-hand circularly polarized antenna for surface scattered signal
  • Adding a second down-looking right-hand circularly polarized antenna makes the system fully-polarized
  • Larger coherent integration times and scanning of Doppler compensation frequency make the system similar to the synthetic aperture radar
  • The intersection of hyperbolic equi-Doppler zones and elliptical equi- range zones form a small footprint on the mapped surface
  • A higher spatial resolution can be achieved using the P-code sequence of the GPS signal, or using a modified signal processing


Conclusions

  • It has been demonstrated in aircraft experiments that GPS scattered signals reflected from the ocean surface can be used as a remote-sensing tool, particularly for wind-speed retrieval
  • Theoretically modeled GPS surface scattering agrees well with experimental data
  • Potentials of this technique can be enhanced by employing fully- polarized delay-Doppler mapping receiver
  • Surface-reflected GPS signals can be used in various other remote-sensing applications: wind vector, wave height, salinity, ice studies, delineation of wetlands
  • Satellite-borne instruments will give a unique opportunity to use GPS as a global scale remote-sensing tool to infer various Earth surface parameters


NOAA | ETL | Radar | Bistatic GPS Signal Scattering from an Ocean Surface: Theoretical Modeling and Wind Speed Retrieval from Aircraft Measurements: http://www.etl.noaa.gov/~vzavoronty