The Case for Building a Current-Mapping Over-the-Horizon Radar
T. M. Georges
NOAA Environmental Technology Laboratory
325 Broadway
Boulder, CO 80303 USA
J. A. Harlan
NOAA Environmental Technology Laboratory/Science and Technology Corporation
325 Broadway
Boulder, CO 80303 USA
Abstract--Results of recent NOAA tests with the U.S. Air Force and U.S. Navy over-the-horizon (OTH) radars show that it is possible to map near-surface currents with 10-15-km resolution to ranges greater than 2,500 km. The technique is similar to that used by commercially available high-frequency current-mapping radars, except that range is greatly extended by bouncing the radar beam off the ionosphere. Current maps made with U.S. Navy OTH radars often show new detail never before seen, for example, the complex space-time structure of the Florida Current, as well as large currents driven ahead of an Atlantic hurricane. However, the military taskloads of these radars preclude their operational use for continuous, large-area ocean monitoring. For that purpose, a dedicated low-cost radar system would be required. A possible installation would cover the Gulf of Mexico, the Caribbean Sea, and the hurricane approaches to the U.S. East Coast and Gulf Coast. We examine the costs, benefits, problems, and potential customers of such a radar.
I. Introduction
Shore-based high-frequency radars that map ocean surface currents are now in common use, are commercially available, and their uses are described elsewhere in these proceedings. Depending on the frequency used, their range is limited to 50-100 km by the attenuation of electromagnetic surface waves over the ocean. NOAA's tests with military over-the-horizon (OTH) or skywave radars show that, with care, the range of current maps can be extended to more then 2,500 km by using ionospheric reflections. As with surface-wave radars, two OTH radars are required to map vector currents [1], but a single radar can map surface wind directions [2]. Many examples of unique current and wind-direction maps obtained in this way can be viewed on our website. Skywave current measurements have been validated by comparison with in-situ acoustic current profilers [3] and with altimetry-derived geostrophic currents [4].
Piggybacking on military OTH radars has permitted us to demonstrate long-range current mapping, but the present military taskload of these radar systems precludes their operational use for routine, large-area ocean monitoring. To fully exploit this remote-sensing technology, it is necessary to design and deploy a skywave radar system dedicated to ocean monitoring. By relaxing military radar specifications and removing the requirement for detecting and tracking manmade targets, it should be possible to build a radar whose ocean-monitoring capabilities are comparable to those of the Navy's Relocatable Over-the-Horizon Radar (ROTHR) but at a fraction of its cost.
In this paper, we outline the problems, costs and anticipated benefits of deploying a skywave radar for ocean monitoring. The benefits are illustrated with data we obtained sporadically between 1995 and the present, using the two ROTHR radars, one located near Norfolk, VA and the other located near Corpus Christi, TX [6].
II. Benefits
The current-mapping capabilities of a skywave radar are amply demonstrated by the many examples shown on our website. Space permits us to show here only two examples of current maps obtained with the U.S. Navy ROTHR: One is a current-vector map in the southern Florida Straits between Florida, Cuba and the Bahamas (Figure 1). The other example is the result of the first attempt to map the surface current field in the vicinity of a hurricane (Figure 2).
Our access to the ROTHR is presently limited to approximately 4 minutes of actual radar time per week, and our view is limited to only a small patch of ocean at a time, generally about 70,000 km2. Even so, virtually every measurement yields surprises and new views of ocean circulation features. Access to the radar's full 106 km2 coverage area (which includes the Gulf of Mexico, the Caribbean Sea, and part of the subtropical Atlantic Ocean) would provide an ongoing synoptic view of the ocean circulation in that region. While propagation limitations prevent mapping currents over the entire radar coverage at one time, composite images derived from runs days apart can help fill in gaps in the coverage. An illustration of this potential is two data sets we obtained in 1995, when we briefly had full access to the Texas ROTHR. These are shown on our website on the 1995 ROTHR data page.
A. What would it measure?
Skywave radars illuminate the sea surface in a sequence of patches that typically measure 150 km in the azimuth dimension and 200 km in the range dimension and each of which requires separate propagation management. Each patch is divided into several hundred cells that determine the radar's spatial resolution. From each cell (typically 6 x 15 km), the radar obtains a spectrum of the sea echo from which information about surface winds, waves, and currents is extracted [7].
Skywave radars complement active satellite sensors by providing temporal continuity within a fixed geographic area [5] and by filling data gaps between orbital tracks. In addition, HF radar's view of the sea is not obscured by clouds and rain. For the radar frequencies normally used, surface currents represent an average over the top meter of the ocean. With a coherent dwell time of 25 s and incoherent averaging over about 10 dwells, currents can be measured with an average accuracy of about 10-15 cm s-1. Increasing the dwell time does not appreciably increase this accuracy because ionospheric motions impose a limit on the accuracy to which the frequency of spectral peaks can be estimated. Current biases and geographic location errors caused by ionospheric motions are estimated and removed using zero-Doppler echoes from nearby landforms.
Just as valuable as the quantitative current measurements in each cell are the radar-derived patterns of spatial current variability. Unlike point sensors, they provide qualitative information about current structures analogous to that provided by satellite imaging systems.
In addition to mapping surface current, such a radar could also map sea state [8] and ocean-surface winds [2].
B. Who are the Customers?
Accurate maps of ocean surface currents, both in coastal zones and in the open ocean, are required by Navy Fleet operations, by Coast Guard search-and-rescue activities, by fisheries managers, by offshore drillers, by shipping interests, by coastal ecosystem managers, by oil-spill response teams, by global-climate-change researchers, and by the general public. Maps of the surface currents driven ahead of hurricanes could help improve understanding and forecasts of the storm surge that endangers coastal residents [6].
We have previously examined in detail the costs and benefits of NOAA taking over operation of the Air Force OTH-B radar for environmental monitoring. Further discussion of various applications and potential users of OTH radar can be found in [9], as well as on our website.
III. Problems
Although the physical principles underlying skywave current measurements are identical to those of surface-wave HF radars [10], there are many practical differences. Illuminating the sea over long ionospheric paths poses additional processing obstacles and added system complexity.
A. Large Receiving Array
The main component of a skywave radar system is a large-aperture steerable receiving antenna array. The need for a large aperture antenna array is not easily compromised. Beams of 1 or less are required to resolve the most interesting current structures at a range of 1,000 km or more. Larger beams would smear out spatial detail. At a range of 1000 km, a 1 beam would produce a radar cell about 18 km wide and would require a 1-km array aperture for a typical radar frequency of 15 MHz. Furthermore, a narrow azimuth beam is required to reduce azimuthal multipath caused by ionospheric inhomogeneity [11]. The array should be steerable over ±45 in azimuth.
B. Multiple Frequencies
A skywave radar must be able to operate anywhere between about 5 and 28 MHz to adapt to prevailing ionospheric conditions. Generally, only a narrow range of frequencies can provide stable propagation to a given patch of ocean at a given time. Finding the best frequency for a given path and time requires real-time 'propagation management,' that is, sampling of the ionosphere's reflecting properties at many frequencies. In addition, a spectrum monitoring system is required for avoiding interference to and from other users of the crowded HF spectrum.
C. Ionospheric Distortions
A major problem with a skywave radar compared with a surface-wave radar is dealing with the contaminations the radar signal suffers during two reflections from the ionosphere. Contamination often occurs, even if propagation management is performed correctly. Fortunately, a number of strategies have been developed over the years to cope with these contaminations. Some attempt to avoid the conditions known to cause them, whereas others correct for or remove distorted data after it is collected [12, 13]. For current mapping in particular, we have found it desirable to wait for stable ionospheric conditions, such as typically occur near midday, and to use lower ionospheric layers, (E and F1) whenever possible [12].
IV. Costs
The greatest challenge in designing a skywave radar for ocean monitoring is making the technology affordable to civilian users. Modern military over-the-horizon radars cost upwards of $100M--an order of magnitude more expensive than even a consortium of civilian customers would likely consider. Fortunately, many of the specifications and requirements of military OTH radars are superfluous to an ocean-monitoring mission. The cost and complexity of a oceanographic skywave radar would therefore lie somewhere between those of a military OTH radar (~$100M) and commercial surface-wave current-mapping radars ($200-500K).
First, it is likely that transmitter power can be reduced substantially. The ROTHR transmitter produces 200 kW continuous power, but the sea echo Bragg lines are typically 30 dB above atmospheric noise and often much stronger. A 10 kW transmitter would reduce the cost of transmitting hardware significantly.
Because of the low waveform repetition rate required for sea echo sampling, it should be possible to use interrupted FM modulation and to collocate the transmit and receive equipment, thus saving the cost of two sites. (Normally OTH transmit and receive sites are separated by about 100 km.) It may also be possible to use a fixed or minimally steerable transmit antenna beam.
Substantial savings result from reducing the sensitivity required to track the weak echoes from manmade targets. Many hardware components, such as receivers, transmitters, sweep-frequency backscatter sounders, and spectrum monitors, are available commercially. Much of the signal processing can be done with ordinary PCs, rather than the specialized hardware required just a decade ago. It may be possible to adapt some of the signal-processing software from existing military OTH systems.
A major cost driver is the receiving antenna system, which cannot be made much more compact than its military counterparts without degrading the radar's spatial resolution to an unacceptable level. It may be possible to reduce the cost of the receiving array and associated receivers somewhat by using thinning techniques and compact superdirective elements, while maintaining a nominal 1-km effective aperture. Although the final cost of such a radar can be only crudely estimated at this point, it is clear that eliminating unnecessary functions and relaxing military specifications, along with innovative design and the use of off-the-shelf components could bring the cost of a useful oceanographic skywave radar to affordable levels. It would be interesting to see if one could be built for roughly $10M. Cost per radar could be further reduced by sharing development costs with multiple users and producing a common design for use in different geographic locations.
Such a radar could be deployed in phases, beginning with a prototype demonstrator system and augmenting it with improved resolution, processing, and diagnostics as funding permits. Although two radars are required to map vector surface currents, a single radar can provide substantial information, if there is sufficient a priori knowledge to estimate at least the sign of the transverse current component. The examples of radial current maps on our website show new details of the structure of the Florida Current and the Yucatan Current.
In addition to the greater cost to build a skywave system, operating it would also be more expensive. The need for hands-on, real-time assessment of the ionospheric propagation conditions is an added cost of a skywave ocean-monitoring radar, compared with a surface-wave system.
It is useful to keep in mind that ocean monitoring systems in general are expensive. For comparison, a typical new oceanographic research vessel costs about $50M, and an oceanographic satellite costs about $300M. Deploying and maintaining the 70-element TOGA-TAO array of buoys in the Pacific costs about $8M per year.
V. Conclusions
Our tests with military OTH radars have produced results well out of proportion to the minuscule radar resources used. A dedicated radar offers vastly increased ocean coverage in space and time and is the main argument for performing a rigorous design study and cost estimate. A possible configuration is shown in Figure 3 that would cover the Gulf of Mexico, the Caribbean Sea, and the hurricane approaches to the U.S. East Coast and Gulf Coast. The capabilities of such a system and the reliability of its data are already reasonably well known from smaller scale tests. Once the costs are more accurately known, potential users can decide whether they are justified in terms of the kinds of services we have demonstrated.
In general, observing systems such as this would be expected to yield their ultimate benefit to society as parts of an integrated regional ocean observing system. Such a system would combine data from multiple complementary sensors with ocean-circulation models and make usable products readily available to customers.
Acknowledgment
We thank the Office of the Secretary of Defense and the U.S. Navy for granting access to their over-the-horizon radars at no cost to NOAA.
References
[1] T. M. Georges, J. A. Harlan and R. A. Lematta, "Large-scale mapping of ocean surface currents with dual over-the-horizon radars," Nature, vol. 379, Feb. 1, pp. 434-436, 1996.
[2] G. S. Young, J. A. Harlan and T. M Georges, "Application of over-the-horizon radar observations to synoptic and mesoanalysis over the Atlantic," Weather and Forecasting, vol. 12, no. 1, pp. 44-55, 1997.
[3] J. A. Harlan, T. M. Georges, and D. C. Biggs, "Comparison of over-the-horizon radar surface-current measurements in the Gulf of Mexico with simultaneous sea truth," Radio Sci., vol. 33, no. 4, pp. 1241-1247, 1998.
[4] J. A. Harlan, "Observations of the western Caribbean Sea circulation: A multisensor approach," unpublished manuscript, 1998.
[5] T. M. Georges, J. A. Harlan, T. N. Lee and R. R. Leben, "Observations of the Florida Current with two over-the-horizon radars," Radio Sci., vol. 33, no. 4, pp. 127-1239, 1998.
[6] J. A. Harlan and T. M. Georges, "Observations of hurricane Hortense with two over-the-horizon radars," Geophys. Res Lett., vol. 24, no. 24, pp. 3241-3244, 1997.
[7] D. E. Barrick, "Extraction of wave parameters from measured HF sea-echo Doppler spectra," Radio Sci., vol. 198, pp. 415-424, 1977.
[8] J. W. Maresca, Jr. and T. M. Georges, "Measuring rms wave height and the scalar ocean wave spectrum with HF skywave radar," J. Geophys Res. vol. 85, no. C5, pp. 2759-2771, 1980.
[9] T. M. Georges, "Costs and benefits of using the Air Force over-the-horizon radar system for environmental research and services," NOAA Tech. Memo. ERL ETL-254, 39 pp., 1995.
[10] R. Stewart and J. Joy, "HF radio measurement of surface currents," Deep Sea Res. vol. 21, pp. 1039-1049, 1974.
[11] T. M. Georges and J. W. Maresca, Jr., "The effects of space and time resolution on the quality of sea-echo Doppler spectra measured with HF skywave radar," Radio Sci., vol. 14, no. 3, pp. 455-469, 1979.
[12] T. M. Georges, J. A. Harlan, R. R. Leben, and R. A. Lematta, "A test of ocean surface current mapping with over-the-horizon radar," IEEE Trans. Geosci. Remote Sens., vol. 36, no. 1, pp. 101-110, 1998.
[13] S. J. Anderson and Y. I. Abramovich, "A unified approach to detection, classification, and correction of ionospheric distortion in HF sky wave radar systems," Radio Sci., vol. 33, no. 4, pp. 1055-1067, 1998.
Figure 1. Dual-ROTHR map of the complex pattern of surface currents between Florida, Cuba, and the Bahamas on 14 May 1997. A portion of the Florida Current (upper left) is diverted to the south of the Cay Sal Bank, and into a cyclonic eddy to the east of the bank. The radars that made this map are in Texas and Virginia.
Figure 2. ROTHR-derived surface-current vectors in the vicinity of hurricane Hortense. The hurricane symbol shows the storm location at the time of the radar image. The storm was traveling northward at the time. The strong currents to the right of the storm (up to 1.8 m s-1) would generate a storm surge upon landfall. The center of the current circulation lies 50 km south of the storm center, apparently lagging the storm by 3 h. Near-inertial turning of the current vectors, as well as a region of surface divergence, can also be seen in the storm's wake.
Figure 3. A hypothetical arrangement of two low-cost skywave radars positioned to monitor the Gulf of Mexico and the hurricane approaches to the U.S. East Coast and Gulf Coast.
TABLE 1
| Surface-Wave
(Codar, OSCR) |
Military Skywave
(ROTHR) |
Low-Cost Skywave | |
| Coverage
(Range) |
~103 km2
(50 km) |
~106 km2
(3,000 km) |
~106 km2
(3,000 km) |
| Size | compact / ~60 m | ~3 km | ~1 km? |
| Power | 50-500 W | 200 kW | 10-20 kW? |
| Operation | Continuous, unattended | Staff of ~100, propagation mgmt, target tracking | Propagation mgmt,
subject to "ionospheric weather" |
| Frequency | few, fixed | 5-28 MHz | 10-28 MHz |
| Cost | $200-500K | $100-200M | $5-10M? |