Letters to Nature -- Feb 1, 1996 (vol. 379, pp. 434-436)
Large-scale mapping of ocean-surface currents
with dual over-the-horizon radars
T. M Georges1, J. A. Harlan2, and R. A. Lematta3
1. Environmental Technology Laboratory, National Oceanic and Atmospheric Administration, Boulder, Colorado 80303 USA
2. Cooperative Institute for Research in Environmental Sciences, NOAA/University of Colorado, Boulder, Colorado 80303 USA
3. Research Development Test and Evaluation Division, Naval Command Control and Ocean Surveillance Center, San Diego, California 92152 USA
Detailed information about near-surface ocean currents is needed for effective fisheries management, pollution mitigation, search and rescue, and climate studies, but the present generation of measurement techniques provides only limited spatial and temporal resolution or coverage1,2. In near-coastal environments, pairs of shore-based high-frequency radars have been used to map surface currents over an area of a few hundred square kilometres3,4. The potential for mapping open-ocean current fields has been demonstrated using military high-frequency radars that can be used to 'see' over the horizon for thousands of kilometres by reflecting signals off the ionosphere. But using one radar, only one current component can be mapped by this method5. Here we report the mapping of surface current vectors obtained from simultaneously employing two such radar systems with overlapping coverage. We obtain a current map in the Florida Straits, about 1,500 km from the radars, covering two 70,000 km2 areas at a resolution of 10 km and 0.1 m s-1. As it employs only about 2% of the radars' potential coverage, the test shows the potential of this technique for mapping the more energetic features of ocean circulation - such as boundary currents and mesoscale eddy systems - over vast ocean areas.
Decametric radio waves interact with the ocean primarily by Bragg-resonant scattering from surface gravity waves. For the special case of radio waves scattered back toward their source, the Bragg-resonant ocean waves are aligned perpendicular to the line-of-sight direction and have a wavelength equal to one-half the radio wavelength. (Ten-meter-long ocean waves backscatter 15-MHz radio waves.) Radially traveling gravity waves Doppler-shift the backscattered energy by an amount proportional to their phase velocity. The spectrum of a sea echo thus consists of two sharp Doppler-shifted lines: a positive-shifted line from the Bragg-resonant waves traveling toward the radio source and a negative-shifted line from those receding from the source. If, in addition, surface currents transport the gravity waves, an extra Doppler shift is imposed that is proportional to the radial component of the current. These principles form the basis of HF (high-frequency or 3-to-30-MHz) current-mapping radars, which are now commercially available.3,4 A single radar maps the radial current component; two radars are required to map surface-current vectors. Relying on line-of-sight and surface-wave illumination of the sea surface, a pair of such shore-based radars can effectively map surface currents, for example in bays and estuaries, to ranges of about 70 km.
Because HF radio waves reflected from the ionosphere can reach great ranges (up to 3500 km in one "hop"), it was at first thought that the range of current-mapping radars could be usefully extended in this manner to cover millions instead of hundreds of square kilometers of open ocean. Two obstacles stand in the way, however: The first is the need for very large aperture antenna arrays. Beams of 1 or less in azimuth are required to resolve the most interesting ocean current features at ranges of 1000 km and more. For a frequency of 15 MHz, for example, a 1 beam would require a kilometer-long array. The second obstacle is the distortion the sea echo suffers after two reflections from the ionosphere (one outgoing and the other on the return path). Ionospheric reflections spread and shift the sea-echo spectrum, making it difficult to measure exact Doppler shifts and to distinguish Doppler shifts caused by the currents from those caused by ionospheric motions.
Large-aperture HF over-the-horizon radars developed and deployed as part of military air-defense systems would satisfy the requirement for narrow radar beams, if access to these systems were allowed. Since 1989, the U.S. Air Force and the U.S. Navy have permitted NOAA to conduct limited ocean-monitoring tests with their OTH radar systems, some of which have been deactivated with the end of the Cold War. The details of these radars, as they pertain to ocean monitoring, are unclassified.6
We deal with the second obstacle, ionospheric distortion, by recognizing that the space and time scales of ionospheric variability are generally different from those of ocean currents. Much of the Doppler "noise" superimposed on the sea echoes is caused by ionospheric motions with time scales of minutes to hours, whereas most open-ocean currents vary on time scales of days or more. This allows us to use spatial and temporal filters to separate the two effects on the sea echo. For example, current features that persist from day to day are presumed not to be ionospheric artifacts. Ionospheric changes on a diurnal time scale exhibit a roughly predictable climatology and impose biases that are nearly uniform over large ocean areas. Such biases can be removed by high-pass spatial filtering, in which a time-dependent areal-average current is subtracted from each measurement. In addition, we take advantage, whenever possible, of the stable reflections provided by the daytime ionospheric E and F1 layers as well as zero-Doppler references provided by nearby land echoes. When such stable echoes are not available, current measurements may have to be averaged over several days to obtain suitable accuarcy.
Our first current-mapping tests of OTH radar used only a single radar and thus mapped only radial currents.5 Soon after the U.S. Navy deployed the second of its two Relocatable Over-the-Horizon Radar (ROTHR) systems in Texas, we set up both radars to illuminate the same patch of ocean in the Florida Straits, about 1500 km from the Texas and Virginia radars. A frequency of 14.5 MHz was selected using real-time sweep-frequency diagnostic soundings over both oblique radar paths. A 24.6 s Doppler frequency transform was used, giving a nominal radial-current resolution of about 0.42 m s-1. Using parabolic interpolation to locate the spectral peaks more accurately, we can achieve a current resolution exceeding 0.1 m s-1. Greater Doppler resolution is possible using longer coherent integration, but the limit imposed by ionospheric shifting and smearing is not yet known.
The ROTHR achieves its nominal 0.5 angular resolution with 2.58-km-long linear phased receiving arrays consisting of 372 twin-monopole elements. Range resolution is achieved by transmitting a 25-kHz FM-CW waveform. A radar resolution cell on the ocean surface is therefore about 6 km in range by about 15 km in azimuth, for the frequency and range used for this test. Only sea echoes that pass a built-in quality test are used for current measurements. The quality index is derived from the sharpness of the sea-echo spectrum, which can be degraded by ionospheric motions and multipath, or by interference from other users of the HF radio spectrum.
The first test used E-layer propagation near midday on 30 May 1995, and data collection from both radars required about 80 min. A second test took place 15 days later, using frequencies of 11 and 16 MHz, and covered the eastern part of the Straits. Radial-current measurements were interpolated to a common grid and combined within the overlapping region to produce the current vectors shown in Fig. 1 for the two test days. Some license has been taken in plotting the two measurements, made 15 days apart, on the same map, but only minor changes in these currents are expected in that time.
The map reveals the core of the Florida Current, which becomes the Gulf Stream off the U.S. East Coast. A maximum radar-derived current speed of 2.0 m s-1 is recorded just off the east coast of Florida, a value consistent with in-situ measurements in this region.7 The northward (counter-clockwise) flow at the left extremity of the map appears to be the eastern part of a semi-permanent feature sometimes called the Tortugas Gyre, centered at 84.5W, 25N in a concurrent TOPEX/ Poseidon sea-surface topography image. The southward flow that "peels off" the southeastern edge of the Florida Current at 80.5W appears to be diverting to the south of the Cay Sal Bank, centered at 80W, 23.8N. The very small current values (with unresolved directions) on the shelf to the north of the Florida Keys are consistent with the shallow waters (and associated bottom friction) there. More frequent measurements would be required to remove any embedded tidal currents.
No concurrent in-situ measurements were available for direct comparison with these radar measurements, underscoring the problem the radar technique addresses. However, day-to-day consistency of single-radar current maps in the same region from the preceding and following days lends credibility to these features, none of which has been consistently and quantitatively observed using other techniques. Subsequent tests have validated the OTH current measurements by comparing them with a shipborne acoustic-Doppler current profiler traversing the Gulf of Mexico Loop Current. This result is being prepared for publication.
These two examples of high-resolution maps of ocean-surface currents illustrate the potential of this OTH radar pair for mapping the surface circulation of the Intra-Americas Sea (IAS). Little is known about the circulation and its variability in and around the straits and channels of the West Indies, about the variable contributions to Florida Current transport by its subsidiary channels, and about the dynamics of the Gulf of Mexico Loop Current and its eddy-shedding process. Observations with enhanced space and time resolution of these and other only vaguely understood features could improve IAS circulation models (which are inadequately supported by observations) and lead to better understanding of how internal IAS dynamics modulates the western Atlantic boundary current and its poleward heat transport.
Since the specifications of OTH radars that apply to environmental monitoring are unclassified, other existing OTH radars that cover different ocean areas could be used for similar circulation studies. Our tests have shown that military radar resources can be time-shared so as not to interfere with the radars' primary mission. Since the strength of sea echoes observed with military OTH radars (which typically radiate 100 kW) is often 60 dB or more above atmospheric noise, consideration could even be given to designing low-power, low-cost OTH radars specifically for ocean monitoring.
1. _____Proceedings of the IEEE Fifth Working Conference on Current Measurement, Institute of Electrical and Electronics Engineers, Piscataway, NJ (1995).
2. Stewart, R. H., Satellite Oceanography, Univ. of Calif. Press, Berkeley, CA, 278-285 (1985).
3. Barrick, D. E., Evans, M. W. & Weber, B. Science 198, 138-144 (1977).
4. Hammond, T. M. et al., Continental Shelf Res. 7, 411-431 (1987).
5 Georges, T. M. & Harlan, J. A. Eos 76(15), 146 (1995).
6. Georges, T. M. & Harlan, J. A. IEEE Antennas and Propagation Magazine 36(4), 14-24 (1994).
7. Lee, T. N. et al., J. Geophys. Res. 100(C5), 8607-8620 (1995)
Figure 1. Result of the first attempt to map ocean surface currents using two over-the-horizon (OTH) radars. The two U.S. Navy ROTHR radars, one in Virginia and the other in Texas, reflected 14.5-MHz radio waves off the ionosphere to illuminate this 70,000 km2 ocean region in the southern Straits of Florida. The arrows show the direction of surface flow, and their lengths are proportional to the magnitude of the surface current (averaged over 2-m depth) at the arrow tip. A maximum current of 1.7 m s-1 is present at the entrance to the straits, and the southward flow of about 1.5 m s-1 at the left edge of the region coincides with the known location of the Gulf of Mexico Loop Current. The area illuminated for this test is about 1% of the ocean area covered by the radars, which includes the entire Caribbean Sea.