Content uploaded by Jeffrey Dean Paduan
Author content
All content in this area was uploaded by Jeffrey Dean Paduan on Dec 30, 2014
Content may be subject to copyright.
INTRODUCTION TO HIGH-FREQUENCY RADAR: REALITY
AND MYTH
By Jeffrey D. Paduan (NPS) and Hans C. Graber (RSMAS)
he concept of using High Frequency (HF)
radio pulses to remotely probe the ocean surface
has been around for decades. In this note, and the
companion paper by Teague et al. (this issue), we
strive to introduce this technique to a broad
oceanographic audience. Teague et al. provide the
historical context plus an outline of different
system configurations, while we focus on the
measurements of primary interest to coastal
oceanographers, i.e., maps of near-surface currents,
wave heights, and wind direction. Another goal of
this note and, indeed, this entire issue is to present
a realistic assessment of the state-of-the-art in HF
radar techniques vis-á-vis coastal oceanography.
When evaluating any new measurement technique,
it is important to separate issues related to system
design from fundamental limitations of the
technique. The former are engineering short-
comings, which are subject to continuous
improvement. The latter are real limitations in the
use of the particular geophysical signal in the
presence of realistic noise. Most of the “myths”
about HF radar measurments, in our view, stem
from the confusion of these two issues.
One common misconception about HF radar
stems from the word “radar” itself. A more
descriptive name would be HF “radio,” as the HF
portion of the electromagnetic spectrum is within
the radio bands. Figure 1 shows a broad range of
the electromagnetic spectrum, including the
nomenclature commonly applied to different
portions of the spectrum. The HF band, with
frequencies of ~3-30 Mhz and wavelengths of ~10-
100 m, sits between the spectral bands used for
television and (AM) radio transmissions. More
commonly, the term radar is applied to instruments
operating in the microwave portion of the
spectrum, for which wavelengths are measured in
milimeters or centimeters.
Throughout oceanography, many different
instruments exploit many different portions of the
electromagnetic spectrum. Figure 2 illustrates
several of these remote sensing techniques used to
extract information about the ocean surface. The
figure is adapted from the review by Shearman
(1981) and it contrasts space-borne systems, such
as altimeters and scatterometers, which use
microwave frequencies with shore-based systems,
which use a range of frequencies depending on the
application. (Not shown are aircraft-borne
systems, which also operate in the microwave
band.) The figure also illustrates the different types
of transmission paths, including true line-of-sight
paths, “sky-wave” paths, which reflect off the
ionisphere, and “ground-wave” paths, which
exploit coupling of the radiowaves with the
conducting ocean water to achieve extended
ranges. For HF radars, instruments that operate
using sky-wave transmissions are often referred to
as over-the-horizon (or OTH) radars (e.g., Georges,
1980), although HF ground-wave radars, which are
the major focus of this issue, also achieve beyond-
the-horizon ranges.
Reflection (or backscatter) of electromagnetic
energy from the sea surface can be expected to
produce an energy spectrum at the receiver, even if
the energy source was single-frequency, because of
T
Jeffrey D. Paduan, Code OC/Pd, Naval Postgraduate
School, Monterey, CA 93943, USA; Hans C. Graber,
Rosenstiel School of Marine and Atmospheric Science,
University of Miami, Miami, FL 33149-1049, USA.
Fig. 1: Electromagnetic spectrum showing the HF band relative to other radio wave bands and the broader
spectrum.
the complicated shape and motion of the sea
surface. Interpreting these spectral returns for
various transmit frequencies is the key to extracting
information about the ocean. Many instruments
rely upon a resonant backscatter phenomenon
known as “Bragg scattering,” which results from
coherent reflection of the transmitted energy by
ocean surface waves whose wavelength is exactly
one half as long as that of the transmitted radar
waves. The inset in Fig. 2 attempts to illustrate this
process by showing how energy reflected at one
wave crest is precisely in phase with other energy
that traveled 1/2 wavelength down and 1/2
wavelength back to reflect from the next wave
crest. These coherent reflections result in a strong
peak in the backscatter spectrum. Scatterometers
exploit Bragg scattering from capillary waves (~1
cm) to obtain information about winds. HF radars,
on the other hand, exploit Bragg scattering from
surface gravity waves (~10 m) to obtain
information about currents (and winds).
Measuring Currents
The history of HF backscatter measurements is
better outlined by Teague et al. (this issue). We
point to the work of Crombie (1955) as the first to
identify strong sea echoes in the HF band with
resonant Bragg scattering. Bragg waves in the HF
band happen to be “short” surface gravity waves,
which can be assumed to be traveling as deep-
water waves, except in very shallow depths of a
few meters or less. This is important because it
allows information contained in the Doppler shift
of Bragg peaks to be used to estimate ocean
currents.
Figure 3 illustrates the Doppler technique for
ocean current determination from HF radar back-
scatter. It shows an actual spectrum from the
Ocean Surface Current Radar (OSCR) system. The
spectrum contains obvious Bragg peaks due to the
presence of Bragg waves traveling toward and
away from the receiver. The frequencies of these
peaks are offset from that of the transmitted energy
for two reasons: 1) the Bragg waves are moving
with the deep-water phase speed given by c =
sqrt(gλ/4π), where λ is the wavelength of the
transmitted energy and g is the gravitational
acceleration, and 2) the Bragg waves are moved by
the underlying ocean current. Because the
expected Doppler shift due to the Bragg waves is
known, any additional Doppler shift is attributed to
the current as shown in Fig. 3.
It is important to keep in mind the following
points about HF radar-derived currents: 1) a single
radar site is capable of detecting only the
component of flow traveling toward or away from
the site for a given look angle, 2) the effective
depth of the measurement depends on the depth of
influence of the Bragg waves and is quite shallow
(~1 m), 3) stable estimates require scattering from
hundreds of wave crests plus ensemble averaging
of the spectral returns, which sets the space-time
resolution of the instruments, 4) the precision is
limited by the frequency resolution of the Doppler
spectrum and is typically 2-5 cm s–1, and 5) the
accuracy is controlled by numerous factors, such as
signal-to-noise ratios, pointing errors, and
geometry.
Since a single radar station measures only the
component of flow along a radial beam emanating
from the site, “radial” currents from two or more
sites should be combined to form vector surface
current estimates. Figure 4 illustrates this principle
using radial data from two radar sites. It also
illustrates the “baseline problem” that occurs where
both radar sites measure the same (or nearly the
same) component of velocity, such as along the
baseline between the sites or at great distances
from both sites. Generally two radials must have
an angle greater than 30 degrees and less than 150
degrees to resolve the current vector. This
geometric sensitivity has been compared to the
familiar geometric dilution of precision, or GDOP,
in the Global Positioning System (Chapman and
Graber, this issue). If currents are assumed to be
constant over several radial bins, it is also possible
Fig. 2: Scematic representation of various remote sensing methods exploiting signals
backsattered from the sea surface (after Shearman, 1981). The inset illustrates the
resonant Bragg scattering process that occurs due to reflection from waves whose
wavelength is 1/2 as long as that of the incident energy.
to estimate velocities using a single radar site as
was done by Bjorkstedt and Roughgarden (this
issue), although the GDOP-related errors will be
relatively large in this case.
The current measurement by HF radars is close
to a “true” surface current measurement. Because
radar pulses scatter off ocean waves, the derived
currents represent an integral over a depth that is
proportional to the radar wavelength. Stewart and
Joy (1974) show this depth to be, approximately, d
= λ/8π. Since wavelength depends on the radar
frequency, it is feasible to use multi-frequency HF
radars to estimate vertical shear in the top two
meters of the ocean.
Present system and coverage capabilities of HF
radars are quite impressive. Measurements can be
made in range as short as 1 km and as long as 150
km from the shore at a resolution of about 1 to 3
km along a radial beam. Radio interference or high
sea states can limit the actual range at times as well
as the ground conditions in the vicinity of the
receive antennas. Wet and moist sandy soils
enhance the ground wave propagation, while dry
and rocky grounds reduce signal strengths. Typical
azimuthal resolutions are ~5˚. Near the coast, this
gives a measurement width of ~0.5 km; the width
is ~10.0 km at range cells 100 km offshore (Fig. 4).
Measuring Winds and Waves
Although the focus of this special issue, and
many of the experiments using ground-wave HF
radar systems, is on surface currents, it is also
possible to extract information about surface waves
and winds from HF backscatter spectra. Wave
techniques are discussed by Wyatt (this issue) and
by Heron and Graber (this issue), while the method
for extracting wind direction is discussed by
Fernandez et al. (this issue). Very crudely, wave
information is obtained by fitting a model of
surface wave backscatter to the observed second-
order portion of the spectrum (Fig. 3), which is due
to reflections from waves at all frequencies and not
just the resonant Bragg waves. Wind direction, on
the other hand, is related to the ratio of the strength
of the advancing and receeding Bragg peaks.
System Configurations
While the basic scattering principle is the same
for all existing HF radars, distinct differences are
found in the antenna configurations that transmit
and receive the electromagnetic signals. The
compact antenna system utilized by the Coastal
Ocean Dynamics Applications Radar (CODAR)
consists of crossed loops and a whip for receiving
and a whip for transmitting radio pulses (Barrick et
al., 1977). This antenna system is small and lends
itself for deployment in highly populated and rocky
coastal areas (e.g., cover photos). Radars of this
type have been in use in Germany (Essen et al.,
1981) and the Monterey Bay area (Paduan and
Rosenfeld, 1996; Paduan and Cook, this issue).
The omnidirectional characteristic of the cross-loop
whip combination makes it possible to scan wider
ocean sectors (e.g., Fig. 4), but this requires
software-intensive, direction-finding techniques to
determine angle for a given range cell (Lipa and
Barrick, 1983; Barrick and Lipa, this issue).
In contrast, linear phased-array antennas consist
of numerous (typically 8 to 16) elements separated
by one ocean wave length and aligned normal to
the principal receive direction (e.g., cover photos).
These radars, such as the University of Miami’s
OSCR system, are positioned at the seaward edge
of a beach or cliff and require open space up to 100
m in length. The radio pulses are transmitted from
a separate antenna array, which is a four-element
Yagi array in the case of OSCR. Azimuthal
resolution (direction) is obtained from well-
established beam forming techniques. Other radars
utilizing phased arrays are found in Germany
(Gurgel 1997), Japan (Hisaki, 1996), Australia
(Heron et al., 1985), France (Forget et al., 1981),
Canada (Howell and Walsh, 1993) and United
Kingdom (Wyatt, 1986; Prandle, 1991).
It is misleading to attempt to describe one HF
radar configuration that will be optimum for all
situations. Direction-finding (DF) and phased-
array systems each have their advantages and
disadvantages. For example, DF systems like
CODAR were developed to be able to deploy the
Fig. 3: Sample backscatter spectrum showing prominent Bragg peaks due to
waves advancing toward and receding from the receiver. The smaller
Dopper shift,
Δ
f, is due to ocean currents that, in this example, are moving
away from the receiver.
antennas on a small coastal outcrop, or even on a
building, where a long secure stretch of beach or
cliff may not be available. In addition, the angular
coverage from DF techniques is much greater than
the, at most, 60˚ sector that is available using
phased-array pointing techniques.
At the same time, phased-array systems have
important advantages over DF systems. Because
the “beam” can be steered to a particular look
direction, it is possible to collect backscatter
spectra from a single patch of ocean (e.g., Fig. 3)
and, thereby, infer surface wave characteristics
from the second-order portions of the spectra. (DF
systems, by contrast, collect spectra based on ocean
backscatter over an entire range cell, which
obscures the wave information.) The determination
of wind direction is also more straight forward
when using individual spectra from phased-array
systems.
Conclusions
The purpose of this special issue on HF radars
is to describe in simple terms how the radars work
and demonstrate the usefulness and capabilities of
such instrument technology for todays problems in
ocean research. The following short feature
articles present a wide variety of applications that
are important in physical and biological
oceanography in the coastal zone. Beyond their
utility to the scientist, these measurements are also
of great interest to both military and civilian coastal
engineers, public safety officers, and planners who
must maintain navigational seaways, mitigate
ocean pollution, conduct search and rescue
operations, and attempt to balance the health of
coastal habitats, public access, and private property
rights.
The advantages of HF radar as a non-invasive
measurement tool that can acquire vector surface
current, wave, and wind information should be
obvious. However, while the concept of this
technology is old, its acceptance in science,
government, and industry has been slow. Today
there is no reason not to proceed to develop better
hardware and software components while,
simultaneously, exploiting what existing systems
can tell us about the ocean. By analogy, the
Acoustic Doppler Current Meter (ADCP) was, a
few years ago, considered experimental and
mysterious by many in the oceanographic
community, while now its use is common. We are
confident that the use of HF radars will also
become commonplace, and that, as a result, a new
level of understanding of the coastal ocean will be
possible.
Acknowledgments
The editors of this special issue on High
Frequency Radar Remote Sensing gratefully
acknowledge the continued support of the Office of
Naval Research through grants N00014-91-J-1775
(HIRES), 92-J-1807 (REINAS), 94-1-1016
(DUCK94), 95-3-0022 (MRY BAY), 96-1-1065
(COPE), and 97-1-0348 (SHOALING WAVES).
Beyond these specific programs, we thank D.
Trizna, T. Kinder, and S. Sandgathe at ONR and
C.L. Vincent at USAE/ONR for their long-range
visions of HF radar’s potential.
References
Barrick, D.E., M.W. Evans, and B.L. Weber, 1977:
Ocean surface currents mapped by radar,
Science, 198, 138-144.
Crombie, D.D., 1955: Doppler spectrum of sea echo at
13.56 Mc/s, Nature, 175, 681-682.
Essen, H.H., E. Mittelstaedt, and F. Schirmer, 1981: On
nearshore surface current measurements by
means of radar. Dt. Hydrogr. Z., 1-14.
Forget, P., P. Broche, J.C. De Maistre and A. Fontanel,
1981: Sea state frequency features observed by
ground wave HF Doppler radar. Radio Sci. 16(5),
917-925.
Georges, T.M., 1980: Progress toward a practical
skywave sea-state radar. IEEE Tran s. Ant.
Propag., AP-28, 751-761.
K.-W. Gurgel, K.-W. and G. Antonischki, 1997:
Measurement of Surface Current Fields with
High Spatial Resolution by the HF Radar
WERA. Proc., IEEE/IGARSS'97, in press.
Heron, M.L., P.E. Dexter, and B.T. McGann, 1985:
Parameters of the air-sea interface by high-
frequency ground-wave HF Doppler radar. Aust.
J. Mar. Freshw. Res., 36, 655-670.
Hisaki, Y., 1996: Nonlinear inversion of the integral
equation to estimate ocean wave spectra form HF
radar, Radio Science, 31, 25-39.
Fig. 4: Sample radial current coverages for a phased-array radar (site 1)
covering a 60˚ swath, and a direction-finding radar (site 2), which in
pricinple can cover up to 360˚. At overlapping ocean bins (e.g., dashed
circle) a vector current estimate can be made, providing the angular
separation between the radial currents is large enough.
Howell, R. and J. Walsh, 1993: Measurement of ocean
wave spectra using narrow beam HF radar, IEEE
Journal of Oceanic Engineering, 18, 296-305.
Lipa, B.J., and D.E. Barrick, 1983: Least-squares method
for the extraction of surface currents from
CODAR Crossed-loop data: Application at
ARSLOE, IEEE J. Oceanic Engin., 8, 226-253.
Paduan, J.D., and L.K. Rosenfeld, 1996: Remotely
sensed surface currents in Monterey Bay from
shore-based HF radar (CODAR). J. Geophys.
Res., 101, 20669-20686.
Prandle, D., 1991: A new view of near-shore dynamics
based on observations from HF radar. Prog.
Oceanog. 27, 403-438.
Petruncio, E.T., L.K. Rosenfeld, and J.D. Paduan, 1997:
Observations of the internal tide in Monterey
Submarine Canyon. J. Phys. Oceanogr.,
Submitted.
Shearman, E.D.R., 1981: Remote sensing of ocean
waves, currents and surface winds by dekametric
radar. Remote sensing in meteorology,
oceanography and hydrology, A.P. Cracknell,
ed., Ellis Horwood, London.
Stewart R.H., and J.W. Joy, 1974: HF radar measurement
of surface current. Deep-Sea Res., 21, 1039-
1049.
Wyatt, L.R., 1986: The measurement of the ocean wave
directional spectrum from HF radar Doppler
spectra. Radio Sci., 12, 473-485.
