The Space GPS Reflectometry Experiment On

the UK Disaster Monitoring Constellation

Satellite
Martin Unwin, Scott Gleason, Michael Brennan, Surrey Satellite Technology Ltd

BIOGRAPHY INTRODUCTION

Martin Unwin heads the GPS team at Surrey Satellite In 1993 the concept of using GPS and other signals
Technology, responsible for spaceborne GPS and GNSS reflected off the Earth’s surface as a means of remote
receiver design and operation. He holds a BSc from sensing the environment was proposed by the European
Lancaster University and a PhD from the University of Space Agency [1]. Potentially reflected GNSS signals
Surrey. could be used to accurately determine sea height from
space as a low cost alternative to satellite altimetry. As
Scott Gleason received his BS in Electrical Engineering the ideas and the theory were subsequently developed,
from the State University of New York at Buffalo and an other potential GNSS-derived measurements emerged,
MS in Engineering from Stanford University. He is such as sea state, wind speed, ice sensing and soil
currently a member of the GPS team at SSTL, where his moisture sensing. Reflectometry experiments from a
duties include software development and GPS variety of platforms followed, including from aeroplanes,
applications research. He is presently researching a PhD balloons and the Space Shuttle [2] [3] [4], and most
thesis in the area of GNSS Reflectometry at the recently from the SAC-C GPS receiver [5]. Nevertheless,
University of Surrey. it is recognised that there is still a shortage of data from
LEO altitudes to validate the current models.
Michael Brennan is a GPS engineer at SSTL, and was the
work package manager for the GPS receivers on UK- SSTL has an interest in the potential of GNSS
DMC and NigeriaSat. He holds a B.Eng from the reflectometry as an enabling technology for low cost
University of Greenwich in Electrical and Electronic ocean sensing constellations such as that proposed in the
Engineering. GANDER concept [6]. As a GNSS reflectometry
receiver is a passive instrument, the payload could fit on a
very small satellite, perhaps even a sub-10 kg
ABSTRACT nanosatellite, significantly lowering the cost of any
constellation.
A spaceborne GNSS reflectometry experiment is to be
undertaken by Surrey Satellite Technology Limited with SSTL has expertise in GPS receiver design, and SSTL’s
support from the British National Space Centre (BNSC) SGR has been flown on numerous LEO missions [7,8,10].
to investigate the reception of reflected GPS signals off When an opportunity arose on a forthcoming SSTL
the ocean for oceanography applications. The GPS satellite, it was a logical step to upgrade the capability of
receiver on the UK-DMC satellite has been modified to the SGR from just a navigation tool into a scientific
accommodate a downward (nadir) pointing high gain instrument to demonstrate this new technique.
antenna, to send sampled data to a solid-state data
recorder and to process reflected signals in real time.
This instrument will be used to search for signals THE UK-DMC SATELLITE
reflected off the ocean and when detected, characterise
their relationship with the sea state. If successful, this The UK-DMC satellite is one of a small constellation of
experiment could prove an enabling step towards a 700km altitude polar-orbiting satellites intended to image
constellation of low cost small satellites that could make disaster areas and provide images promptly to relief
real-time measurements of the sea state around the globe. agencies. Alsat-1 was the first DMC satellite to be
launched, and has been in orbit since November 2002 [9].
Three more, including UK-DMC, are scheduled for
launch in September 2003. A number of secondary
experiments are carried on UK-DMC, including an is expected to be used for assisting the search for reflected
Internet router, a miniature propulsion system and the signals, rejection of direct signals from nadir antenna, as
GPS reflectometry experiment. well as for orbit determination

The generic design of the DMC satellites includes both a In addition to raw IF sampling, the GPS receiver can be
dual antenna SSTL SGR -10 GPS receiver and a solid- reprogrammed in orbit with code to process the reflected
state data recorder, fortuitously carried in the same signals in real-time using its 24 parallel channels and a
module tray, and so the reflectometry instrument could be relatively powerful processor (ARM60B) to handle the
constructed from electronics mostly already present in the additional calculations.
satellite.
To accommodate this GPS reflectometry experiment, an
additional front-end was added to the UK-DMC GPS
INSTRUMENT OVERVIEW receiver, the sampled IF signals were connected to the
data recorder via an LVDS daughter-board and a new
A block diagram of the instrument is shown in Figure 1. nadir antenna was added.
Three antennas are used, two space-pointing (zenith) and
one high gain Earth-pointing (nadir) antenna. The signals A new antenna des ign was commissioned from European
from each antenna can be down-converted to IF Antennas Ltd for this experiment as it had fairly special
(intermediate frequency) by three GPS L1 front-ends. requirements. The antenna was to be GPS L1 LHCP, as
The sampled IF signals can either be processed in real- models indicate that most of the signal received at nadir
time using the 24 GPS C/A code channels, or directly would be LHCP. Furthermore, the antenna should be as
logged by a d ata recorder at IF. high gain as possible without interfering with other
antennas present on the satellite, and have a 10° offset
The IF is 2-bit sampled at 5.71 Mbps and the data from bore-site to allow for some yaw steering. European
recorder logs sampled signals from the nadir and from Antennas provided a three-patch array integrated with
one of the zenith antennas. The data recorder also logs the parasitic elements that neatly fitted in a gap between two
occurrence of the GPS receiver pulse per second signals other SSTL antennas, as shown in Figure 2. The gain
to help match the data collection time with GPS time. achieved was just under 12 dBic.
The data recorder has 128 MBytes of storage space, and
the data is stored in a compact bit-by-bit fashion, so 40
seconds of continuous sampled data can be recorded. As Reflectometry
the DMC satellites have been designed to download Antenna
images of up to 1 GByte in size, the 128 MBytes of
reflectometry data can be easily handled by the
communications system.
Direct
Antennas

LNA SGR RF
Front-Ends SGR 24 Channel PVT & Processed
1&3 Parallel Digital Data from RF1/4
Processing

ARM Processor
LNA SGR RF Command and control High Rate
Front-End Downlink
4

Sampled
data
Figure 2, Antenna Mounted to the Spacecraft
High Gain 5.71 MHz Solid State Data
Nadir Antenna 2 bit samp.
F/E 1 & 4
Recorder and& OBC
OBC The hardware was tested using GPS simulators and
Plus PPS SSTL’s software receiver tools. The satellite itself was
Figure 1, Space GPS Receiver Reflectometry subjected to standard environmental testing (see Figure
Configuration 3).

While sampling is taking place, the GPS receiver will be

positioning according to normal operations using signals
received by the zenith antenna. The information typically
available from a GPS receiver will be stored
(simultaneously with the raw sampled data): position,
velocity, time, satellites tracked, satellite positions,
pseudoranges, Doppler shift, phase, etc. This information
 Reflectometry
NigeriaSat-1 Antenna

UK-DMC

Figure 3, DMC Satellites in Thermal Vacuum

Chamber
Figure 4, Reflection Geometry

MODELLING THE EXPECTED SIGNALS

In the traditional link margin calculation, variables such
There is still a degree of uncertainty over the as the loss due to path delay and antenna patterns can be
calculated to within a manageable degree of accuracy.
characteristics of reflected signals received by a
Other variables involved in this scenario such as σ0 and
spaceborne platform, as the signals are expected to be
quite different to those received from airborne the actual area contributing to reflected power on the sea
experiments. Link margin calculations can be used to surface present greater difficulties under random sea
conditions. For example, Figure 5 and Figure 6 represent
help select the optimum satellite altitude and nadir
antenna gains for a spaceborne experiment. the approximate area on a 2 dimensional path of sea that
will contribute to the scattering process for wind of
Being an experiment of opportunity, the altitude of the 15 m/s and 5 m/s respectively. This general model uses
UK-DMC satellite is higher than ideal, and the gain of the the directional wave spectrum of Elfouhaily [20] and is
antenna that could be accommodated at short notice is not used to indicate relative power distribution under different
conditions and not absolute peak power. It is expected
as high as desired. The preliminary link calculations for
this experimental configuration indicates that detecting that lower winds will result in stronger signals despite the
fact that the effective area of the scattering will be
signals could be challenging at times but realistic under
smaller.
calm ocean conditions.

The general geometry of a bistatic GPS reflection is

shown below in Figure 4 where Rx and Tx represent the
receiving spacecraft and the transmitting GPS satellite
respectively. The point on the earth’s surface where a
specular reflection occurs will satisfy Snell’s law, where
the incident and reflected angles are equal (θ I ).

There are several ways to calculate the power of the

received signal at Rx and many variables that can be
adjusted that greatly affect the overall result. It is
expected for bistatic or forward reflection that there will
be a combination of specular coherent reflections (using
geometrical optic principals) and a considerably larger in-
coherent “scattered” diffuse component (including a radar
scattering calculation, σ0 ). The current literature often
places an arbitrary cut-off between the two types as a
function of sea height and transmission wavelength [16].
This subject of the “mixing” of diffuse and coherent Figure 5, Approximate Distribution On Sea Surface
components could benefit from empirical measurements Of Scattering Power For 15 m/s Winds
to help validate and improve the existing models.
Concerning the sea surface plots: The yellow dot (in the northern hemisphere particularly in the Gulf of Mexico,
middle on the left) represents the receiving satellite (R x) the Mediterranean and between the islands off the shores
sub-satellite point, the red curve the 3 dB ellipse of the of South East Asia [17].
antenna on the UK-DMC experiment and the faint blue
circle the 1 C/A code chip iso-range ellipse. Models also indicate that under calm conditions, the
strongest signals will be returned from specular points
directly under the receiving satellite, i.e. when the GPS
satellite passes overhead. The 10° offset of the high gain
nadir antenna will slightly modify the direction from
which the strongest signals are to be received.

Once reflected signals have been detected, and

measurements derived, there will be a need for validation
of the data through independent determination of the sea
state. The sea state can be monitored either using existing
in-situ measurements, such as from buoys, or by
identifying cross-over points with other remote sensing
satellites, such as NASA-CNES Topex and Jason-1
altimetry satellites.

An operational planning tool was developed in software

that took into account these requirements to enable GPS
Figure 6, Approximate Distribution On Sea Surface reflectometry operations to be scheduled. Specular
Of Scattering Power For 5 m/s Winds reflection points on the Earth between user satellite and
GPS satellites can be predicted according to the following
specified constraints:
In spite of the valuable modelling work being undertaken
• Location on the Earth’s surface
in the oceanography and GPS community, there remains
considerable uncertainty as to the received signal levels of • Size of region of interest
reflected GPS signals at LEO and an important • Scheduling within specified time into future
experimental objective of this experiment is to provide an • Maximum off-pointing angles permitted
answer to some of these questions. • Antenna may be off-pointing from nadir by
specified offset
If signals are strong enough, the experiment may provide • Ground-track cross-over with additional
an opportunity to exploring and verifying models for specified satellite(s) if necessary
other potential non-oceanography applications, for
example surface moisture content monitoring [14] and ice Once suitable targets have been found, the program gives
sensing. the following results that can be used to control
operations of the satellite:
• Times, locations and details of predicted specular
OPERATIONS PLANNING AND TARGETING GPS reflections.
REFLECTIONS • Attitude manoeuvring information to steer nadir
antenna to specular point (an orbital
The UK DMC satellite is in a near-polar orbit, and so demonstration of this capability is described in
experimental operations could be scheduled anywhere [12]).
over the whole globe over the course of a few days.
However, as the data recorder can store only 40 seconds The software was designed to include a simple user
of sampled GPS data at one time, the data gathering must interface to assist operators to rapidly schedule data
be targeted carefully, according to some selection criteria. collections based on requests from users (see Figure 7).

As the models indicate the reflected signals could be very The program uses NORAD-generated Two Line Element
weak under rough sea conditions, calm seas will initially (TLE) files that can be downloaded from the Internet and
be targeted. At any time, certain regions of the oceans are updated frequently [18]. The program uses an SGP4
have a good statistical probability of being calm due to propagator [19] to estimate the locations of the receiver
geography or seasonal effects, and so successful signal satellite and all the GPS space vehicles at commanded
detection is most likely over these regions. The satellite time steps from a synchronized epoch time. It can then
launch is scheduled for September 2003, during which calculate specular reflection point locations for all GPS
time the more traditionally calm areas of ocean are in the satellites in view. Cross-overs with other remote sensing
satellites can also be predicted by using their respective
TLEs. An offset in the boresight of the nadir antenna can SOFTWARE RECEIVER
be accommodated in the software, and satellite attitude
manoeuvres can be planned to slew the maximum gain Once the raw sampled data has been collected and
towards the anticipated specular point on the oceans downloaded, it is processed by a software GPS receiver
surface. that performs the usual down -conversion and correlation
processes required to detect signals. The great advantage
of post-processing over real-time tracking is that the data
can be processed over and over again, varying parameters
and optimising for known or predicted signals.

A software receiver and related tools were developed at

SSTL using a combination of C and Matlab
programming. The functioning of the software was
proven using reference data generated by a Novatel 8-bit
RF front-end supplied by Stanford University.

To recover any GPS signals from sampled data, the

incoming data must be:
• multiplied by a local oscillator with the correct
Doppler shift,
• multiplied by the correct GPS satellite C/A code
signal at the correct code phase
• integrated over 1 millisecond (or more)

The steps required for the software receiver to process

Figure 7, Software Prediction Utility Main Display and detect reflected signals are shown in Figure 9.
Showing a Near Intersection Between a DMC and
Jason-1 Satellite Near Antarctica
1) Extract bits to raw data
Sampled RF IF stream
An example of scheduling with a pre-defined target box (Nadir & Zenith) 2) Perform 1 ms correlations
PPS 3) Perform non-coherent
off the west coast of Africa is shown in Figure 8. The 128 MBytes integrations
satellite is travelling “down” the image with the red dot 40 seconds
“behind” it representing the location on the oceans surface
corresponding to the maximum gain of the nadir-facing
antenna. The white patches around the satellite represent 1) Detailed searches using
Specular Point FFT
predicted specular reflection points on the oceans surface. Predictions 2) Flexible plotting of results
UK DMC PVT 3) Specular point verification
GPS Satellite PVT 4) Specular parameter
Real time corrections calculations
5) Link margin analysis

1) Code domain map

2) Frequency domain map

Figure 9, Sampled Data Processing Flow Diagram

Even though there are only 40 seconds of data, there is

still a vast search space to find potentially very weak
signals in the data, with parameters that can vary such as
satellite codes, code phase, Doppler shift, rates of change
of code and Doppler. The PVT and PPS information
available helps in the prediction of the reflected signals,
and significantly reduces the search space. Nevertheless,
advanced search techniques involving FFTs have been
Figure 8, GPS to UK-DMC Specular Reflection Point developed that will find satellites with a more poorly
Prediction over the Eastern Atlantic Ocean predicted Doppler shift.
 from one second to the next allows for more flexibility in
The nadir antenna is expected to also detect relatively the software receiver signal search. The multiple traces in
strong direct signals that may “leak around” the UK DMC each colour are an effect of the sampling and of taking an
satellite structure (effectively via back-lobes of the inherent 3-D plot and representing it in 2 dimensions.
antenna). The PVT information makes the prediction and The sampling frequency of 5.71 MHz translates into a
detection of these signals in the raw data rather limited number of correlation points (or correlator
straightforward. If present in the raw data, these direct spacings in more common usage) per GPS C/A code chip.
signals gives us confidence that the collection process has In this regard the on-board calculations will be of critical
been successful. importance. We will be able to estimate accurately the
expected C/A code delays and Doppler offset frequencies
During the integration of the spacecraft, there were and feed these into the software receiver.
opportunities to test the GPS receiver and data recorder
operation. These allowed us to validate the software The decreasing in magnitude sinc functions of the
receiver as well as the hardware itself under various frequency plot represents different C/A code “slices”.
conditions before the launch of the UK-DMC satellite. Conversely, the code chips (delay) domain plot contains a
Shown below in Figure 10 and Figure 11 are the typical trace for numerous frequency slices and appears as a
correlation functions of a GPS signal in the time/space dense spike because as the correlation tails off over a
and frequency domains. A GPS simulator was configured relatively larger span of frequencies the effect is to
to produce signals as if picked up directly fro m a receiver gradually fill the correlation peak.
in low earth orbit. This was to reproduce as accurately as
possible the expected signal dynamics typically expected
from a LEO satellite.

The two correlation peaks visible are due to the same

signal being processed at two successive time intervals.
A single second separates the black and blue traces and
the movement of the peaks agrees with that is expected in
a low earth orbit (the peaks are 9 chips apart, indicating a
change in range of about 2.7 kilometers).

Figure 11, Correlation magnitude vs Frequency (Hz)

(“Doppler domain”) of Test Signal, One Second
Between Black and Blue Traces.
The shapes of both the delay and Doppler plots are
expected to look quite different when dealing with
reflected signals, and the characteristics will depend on
the sea state. Key measurements will be derived from
these plots, specifically the maximum peak power, and
the slopes of the raising and trailing edges of the
correlations in both delay and Doppler domains. Using
the appropriate algorithms, it is hoped that these
Figure 10, Correlation magnitude vs C/A Code Chips measurements can then be inverted to recover sea state
(“delay domain”) of Test Signal, One Second Between information.
Black and Blue Traces.
The sharp shapes of the correlation peaks in Figure 10 are
DETECTING WEAK SIGNALS
those of direct GPS signals (as generated by the
simulator). If reflected signals are detected, it is predicted
that the shape will elongate considerably with respect to In some terrestrial applications, very weak signals can be
recovered by coherent correlation with long integration
the direct signal as was observed in [4].
times. Models and measurements connected with
The Doppler map of the test signal (Figure 11) also moves spaceborne reflectometry, however, indicate that no gain
with respect to time, the relatively slow shift in frequency will be achieved by correlating for longer periods than
1 millisecond, as the satellite velocity and ocean surface
features prevent longer coherency [4]. Instead, the These routines will be validated using the reflected data
processing gain of weak signals can be increased by non- against the known sea state. It is anticipated that such
coherent integrations, accumulating millisecond inversion routines will initially be a mixture of theory and
correlations over periods of several seconds. This empirical modelling, although subject to later refinement.
additional gain will help to recover weaker signals, for
example over rough seas, or perhaps improve the SNR of The reflected data and reference sea-state measurements
stronger signals. will also be useful in planning for future more
sophisticated oceanography and altimetry missions.
During the testing of the flight hardware prior to assembly Researchers can use the data for extending detailed
in the satellite, some tests were performed to evaluate reflectometry models to LEO e.g. [21], analysis of
how much gain could be obtained using non-coherent specular and diffuse reflections, the mean slope variance,
integration techniques. It was found that correlation determine how surface roughness affects altimetry offsets,
peaks could still be detected when the signals were and perhaps investigate the dependence of the signals on
7-8 dBs below the normal acquisition threshold of the less direct parameters such as salinity, etc.
GPS receiver (down to –170dBW) [15].

This increase in processing gain should increase the ON-BOARD SOFTWARE

chances of detecting signals over a more diverse range of
sea conditions and off-pointing angles, and thereby The ability to download 40 seconds of sampled IF data is
increase the overall capabilities of the experiment. a vital step, but once signals have been detected and
characterised, the GPS receiver on UK DMC has the
flexibility to be reprogrammed in orbit to process these
INVERSION AND VALIDATION reflected signals in real time. To accommodate the real-
time reflectometry function, significant upgrades have
When signals have been detected then they will need to be been integrated into the existing operational GPS receiver
related to the actual sea-state. Operations will be software.
scheduled such that comparisons of signals with
independent information about the ocean can be made: The software development consisted of:
• Prediction of specular points for targeting
• Wind and weather models are available globally, reflections
but are likely to be more dependable near • Selection and allocation of satellites
coastlines • Detection of reflected signals
• Buoys and other in-s itu measurement sources • Delay-Doppler mapping of reflected signals
will be gathered (to be used to determine sea state)
• Cross-overs with other satellites such as Topex,
Jason-1 and QuickSCAT will be sought to The low level tracking and mapping functions have been
enable coincident measurements divided into several strategies or modes. They are based
around the general concept of staggering the available
Expertise exists in combining measurements from such tracking channels on the GPS receiver into bins where the
sources and so it is anticipated that detailed and accurate signal “shape” can be mapped to determine the distortions
sea-state information can be obtained at locations that are caused by the reflection process at the sea surface. For
being observed by the reflectometry experiment. example, the 12 spare tracking channels can be allocated
to a single satellite at ½ chip steps around the predicted
As a first aim, signal strengths can be simply related to specular point delay to be able to determine the change in
sea state (and off-pointing angle) to help understand the correlation between finely set intervals. Combining all
operational range for the UK-DMC experiment in terms the channel measurements will allow us to reconstruct the
of off-pointing angle and sea roughness. This will changed signal shape resulting from its interaction with
directly assist in scheduling future measurements to the sea surface. The same principle can be applied using
obtain a suitable range of representative cases. frequency offsets to produce a correlation shape in the
Doppler domain. A similar concept has been proven to be
The signal strength is one of the many measurements that successful in operation from an airborne platform for
can be compared with the model predictions. Some detecting and inverting information from GPS reflected
discrepancies are to be expected, and the data will help in signals [2,13].
the model refinement process. For the UK-DMC
experiment it is intended that basic inversion routines are The on-board reflectometry mapping software has been
developed that can derive approximately the sea-state implemented in the GPS receiver and tested in its
from the delay-Doppler measurements in near real-time. operational modes [12]. An engineering model SSTL
SGR-20 GPS receiver was set up on a mountain above [2] M. Armatys, “Estimation of Sea Surface Winds Using
Barcelona, and successfully detected GPS reflections. Reflected GPS Signals”, PhD Thesis, University of
The software was also loaded onto an SGR-20 GPS Colorado, 2001.
[3] E. Cardellach, “Results and Validation of the Mediterranean
receiver in the orbiting UoSAT-12 satellite to prove
Balloon Experiment (MEDEX), GPSR from the
successful functioning. Stratosphere”, Third European Workshop on GNSS
Surface Reflections, June 15th 2001, Barcelona.
Experience from processing the raw sampled data on the [4] S. Lowe et al, “First Spaceborne Observation of an Earth-
ground will inevitably be valuable for improving the real- Reflected GPS Signal”, Jet Propulsion Laboratory,
time software to optimise the detection, tracking and California Institute of Technology, Pasadena, California,
mapping process under varying sea states and off-pointing USA. (draft report, Jan 2000)
angles. Ultimately it is hoped that a real-time inversion [5] T. Meehan et al., “Observations of GPS Sea Surface
algorithm can be implemented into the on-board software Reflections from the SAC-C Spacecraft”, Workshop on
GNSS Oceanography with GNSS-Reflections, July 2003,
that can extract sea-state information from the
Barcelona.
measurements. [6] A da Silva Curiel et al. “The GANDER constellation for
maritime disaster mitigation”. Acta Astronautica, Vol44,
Nos 7-12, pp685-692, 1999.
CONCLUSIONS [7] M. Unwin et al. “Preliminary Orbital Results from the SGR
Space GPS Receiver”. Proc ION GPS-99, Nashville TN,
This paper has described the GPS reflectometry Sept 1999 [8] S. Purivigraipong and M. Unwin.
experiment due to fly on the UK-DMC satellite “Determining the Attitude of a Minisatellite by GPS”. GPS
opportunity in September 2003. Existing GPS hardware World, June 2001.
[9] A. Oussedik, Bekhti M. et al “AlSat -1 Microsatellite:
on a remote sensing satellite has been modified to
Disaster Monitoring Constellation (DMC-1)”. Centre
accommodate the facility for receiving GPS reflections National des Techniques Spatiales, Algeria. IAA-01-
off the oceans. The experiment permits both logging of IAA.11.1.a.06. 2002.
sampled IF signals and on-board pro cessing of signals to [10] SSTL’s Space GPS Receiver Data Sheet
be accomplished. Model predictions indicate that signals http://www.sstl.co.uk/datasheets/Subsys_SGR1020_HQ.pd
will be received under calm ocean conditions, but f
processing gain may permit weaker signals to be detected. [12] S. Gleason and M.J. Unwin , “Development and Testing of
a Remote Sensing Instrument Using GNSS Reflectometry
The reflectometry instrument on the UK DMC satellite is Concepts”, IGARSS 2003, Toulouse France.
[13] A. Komjathy, V.U. Zavorotny, P. Axelrad, G. Born and J.
small enough to be flown on a very small dedicated
Garrison. “GPS Signal Scattering From Sea Surface:
satellite, and could form the basis of a constellation for a Comparison Between Experimental Data and Theoretical
commercial marine storm-warning system. Model”. Fifth International Conference on Marine and
Coastal Environments, San Diego, CA 1998.
This experiment will generate data that will be useful for [14] D. Masters, S. Katzberg and P. Axelrad. “Airborne GPS
refining existing models, and could help pave the way for Bistatic Radar Soil Moisture Measurements During
future more sophisticated dedicated GNSS reflectometry SMEX02”. IGARSS 2003, Toulouse France.
remote sensing missions. [15] S. Gleason, M. Unwin. “Ground Validation of the UK
Disaster Monitoring Constellation Satellite GNSS-R
Experiment”. GNSS Reflectometry Workshop 2003,
Barcelona, Spain. July 18-19, 2003.
ACKNOWLEDGMENTS [16] P. Beckmann and A. Spizzichino, “The Scattering of
Electromagnetic Waves From Rough Surfaces”. Artech
Thanks are due to the British National Space Centre for House 1987.
providing the funds to carry out this research and to [17] D. Carter. Month by month mean wave height statistics
everyone at SSTL and European Antennas who were maps over the entire globe for 2002. Internal SOS
essential in the realization of this experiment. Documents provided to SSTL.
[18] NORAD Two Line Element Sets Current Data. CelesTrak.
We would like to thank Dennis Akos of Stanford http://celestrak.com/NORAD/elements/
[19] Hoots F. R., R. L. Roehrich. “Models for propagation of
University for providing data used during the preliminary NORAD Element sets; Project Spacecraft Report No. 3”.
design of the software receiver. Aerospace Defence Command, United States Airforce.
December 1980.
[20] T. Elfouhaily, Chapron B., Katsaros K. and Vandemark D.
REFERENCES “A Unified Directional Spectrum for Long and Short Wind-
Driven Waves”. Journal of Geophysical Research, Vol.
[1] M. Martin-Neira, “A Passive Reflectometry and 102, No. C7, pp 15781-15796, July 15th, 1997.
Interferometry System (PARIS): Application to Ocean [21] V.U. Zavorotny, “Scattering of GPS signals from the ocean
Altimetry”, ESTEC, Noordwijk, The Netherlands., ESA with wind remote sensing application”. (draft submission
Journal 1993, Vol 17. JGN) November 1998.