Remote Sensing of Surface Currents and Waves by the HF Radar WERA
Presented at the EEO'97 conference in Southampton
Klaus-Werner Gurgel and
University of Hamburg, Institute of Oceanography,
Tropolwitzstrasse 7, D-22529 Hamburg, Germany.
Tel: +49-40-42838-5742, Fax: +49-40-42838-5713,
A full paper has been published in the conference
HF radar uses the process of backscattering of electromagnetic waves of 10 m to
50 m wavelength from the rough sea surface and can be used to measure surface
currents and ocean wave parameters. The technique
has been rapidly developed during the past 25 years.
In Germany, the work on HF radar started in 1980 by adopting NOAA's CODAR
system, introduced in 1977 by D. E. Barrick. Recent developments lead to
an improved design called WERA (WEllen RAdar).
This paper describes the different techniques used by CODAR and WERA,
their advantages for different applications and the design used by WERA.
The basic physics of backscattering of electromagnetic waves by ocean waves
have been discovered and described in 1955 by D. D. Crombie.
The first commercial application of this physics has been the
CODAR system, introduced in 1977 by D. E. Barrick at NOAA.
This system has initially been designed for mapping of surface current fields.
Based on CODAR, further developments have been
made in the UK (Marconi OSCR), the United States (Barrick's Sea Sonde)
and in Germany (University of Hamburg WERA).
From CODAR independent developments
of HF radars have been made in Canada (C-CORE Northern Radar),
the UK (University of Birmingham PISCES),
France (University of Toulon), Australia and Japan.
There also exist several skywave HF radar systems all over the world.
In Germany at the University of Hamburg, the work on HF radar started in 1980.
NOAA's CODAR (COastal raDAR) has been bought and was
modified until 1985 to reduce internal noise, increase the sensitivity
of the system and optimize the algorithms.
In 1990, a shipborne CODAR has been developed to
measure the circulation at the Arctic Front and the ice edge in the North
Atlantic. Several limitations of the CODAR have been identified: The 4-element
receiving antenna array did not give direct access to ocean wave measurements
and the range resolution technique of CODAR was limited to
range cell sizes of minimal 1.2 km. Finally, the design did not allow as much
flexibilty as required to adapt the system to new applications.
Based on this experience, WERA (WEllen RAdar)
has been designed: Up to 16 receiving antennas
can be set up very flexible, e.g. as a 16-element linear array,
a 12-element linear array plus the 4-element CODAR array, or even on a circle.
The configuration can easily be reflected in processing software.
The range resolution technique has been changed from using pulses to a
using a frequency chirp, which allows more flexibility in changing the
range resolution of the system. To increase the signal-to-noise ratio of
WERA, parallel processing of 16 independent receiver channels has been
included in the design.
The operating frequency can be between 27 MHz and 30 MHz, but
other frequencies may be possible.
Range resolutions of 1.2 km, 600 m and 300 m can be selected.
Different Techniques for Spatial Resolution
Spatial resolution of radar systems has to be achieved in range and azimuth.
Several different techniques have been developed for the HF radars mentioned,
which can be described by the following principles:
Range resolution can be done on two different ways:
For a selected range resolution (e.g. 1.2 km), a specific bandwidth of the
radar signal (125 kHz) is required. This bandwidth is independent of the
technique used (8 micro seconds pulsewidth or 125 kHz chirp-sweepwidth).
The advantage of range resolution by pulses
is the simple design, however a high peak power has to be gained by the
transmitter. Range gating in the receiver leads to high data sampling rates
to be processed (2 micro seconds at 300 m range resolution).
The advantage of range
resolution by a linear frequency chirp is the simple way to change the
range resolution and a low data rate to be processed in the receiver,
however the transmitter and
receiver must be designed for high dynamic range and linearity.
- The time delay between a transmitted pulse and the
received echo gives the range, the length of the pulse gives
the extent of a range cell. This basically is range resolution
in time domain and used by CODAR.
- The frequency offset between a transmitted linear
frequency chirp and its echo gives the range, the bandwith of
the frequency chirp gives the extent of a range cell. This
basically is range resolution in frequency domain and used by
Azimuthal resolution can basically be done by
The advantage of Direction Finding is the use of a relatively small receiving
antenna, which can more easily be installed at the shore.
Beam forming is essential to allow direct access to the second order
sea echoes, which carry the information on ocean waves,
however significantly increased space is required to set up the array.
As WERA stores raw data, a comparison of Direction Finding and Beam Forming
results calculated from the same data set is possible.
- Direction Finding: Looking for the direction of the echo by
measuring the phase differences between all antennas (CODAR, WERA).
- Beam Forming: Stear a beam to different directions by adding
up the echos received by all antennas after applying a
phase shift (WERA).
The Design of WERA
In an frequency chirp (FMCW) radar system,
signals from increasing ranges appear at increasing
frequency offset. Range resolution is done by fourier transforming each
sweep separately, which for a selected frequency offset gives the time series
at that range. It appears, that a windowing function is required to avoid
smearing of short range echoes to longer ranges by convolution with the
phase-incontinuity of the repetition of the sweep.
To achieve a high signal-to-noise ratio, which besides access to the second
order sea echoes is important for optimal results of the wave processing
the signals received by the 16 antennas of the WERA system are processed
in separate receiver channels instead of multiplexing them.
Amplitude and phase calibration of
all 16 receiver channels including all individual cable lengths is very
important for the Beam Forming and Direction Finding.
To simplify the design of the WERA system, a direct conversion receiver with
I- and Q-channels is used. Before A/D conversion of the received signal,
a 7-pole lowpass filter keeps off all signals above Nyquist frequency. To
achieve a high dynamic range, a 16 bit A/D converter has been selected. Data
are preprocessed inside a VME-bus computer system sweep by sweep and stored
on a Unix workstation for final processing.
The linear frequency chirp is generated by a fast DDS synthesizer. To keep the
phase noise of the transmitted signal and the local oscillator of the receivers
low, an oven controlled high stability crystal oscillator has been
selected. This helps to extend the dynamic range of the system when processing
strong signals from near by and weak signals from long distances at
the same time.
Compared to CODAR, the new design of WERA offeres increased flexibility
in spatial resolution and allows simultanous Beam Forming and Direction Finding
techniques, as required by the application. Within the EC project SCAWVEX,
the WERA system measures surface currents and wave hight directional spectra
simultanously, using the University of Hamburg current algorithm and the
University of Sheffield wave algorithm. This is a further step in research
on current-wave interaction.
This work has been supported by the European Commission, DG XII, within
the Mast-2 programme, project MAS2-CT94-0103, SCAWVEX
(Surface Current And Wave Variability EXperiment).
I wish to thank the other members of the HF-radar group, Georg Antonischki,
Heinz-Hermann Essen, Florian Schirmer, Thomas Schlick and our technician Monika
Hamann for supporting the measurement campaigns, developing algorithms
and processing data.
last update 16-Apr-1997