The Integrated Ocean Observing System High-Frequency Radar Network: Status and Local, Regional, and National Applications

Home , Radar

PAPER The Integrated Ocean Observing System High-Frequency Radar Network: Status and Local, Regional, and National Applications

AUTHORS ABSTRACT Jack Harlan A national high-frequency radar network has been created over the past 20 years NOAA IOOS® Program or so that provides hourly 2-D ocean surface current velocity fields in near real time Eric Terrill from a few kilometers offshore out to approximately 200 km. This preoperational net- Lisa Hazard work is made up of more than 100 radars from 30 different institutions. The Inte- Scripps Institution of Oceanography, grated Ocean Observing System efforts have supported the standards-based ingest Coastal Observing R&D Center and delivery of these velocity fields to a number of applications such as coastal search and rescue, oil spill response, water quality monitoring, and safe and efficient Carolyn Keen marine navigation. Thus, regardless of the operating institution or location of the Scripps Institution of Oceanography, radar systems, emergency response managers, and other users, can rely on a com- Institute of Geophysics and mon source and means of obtaining and using the data. Details of the history, the Planetary Physics physics, and the application of high-frequency radar are discussed with successes Donald Barrick of the integrated network highlighted. Chad Whelan CODAR Ocean Sensors, Ltd. wave height spectrum at the Bragg In the 1990s, the Office of Naval Stephan Howden wave number. Barrick was invited to Research and the National Science Stennis Space Center, University present his results at seminars in Foundation funds were used to acquire of Southern Mississippi Boulder, Colorado, as the National radars at several universities includ- Josh Kohut Oceanic and Atmospheric Administra- ing the Oregon State University, the Rutgers University tion (NOAA), and its Boulder labora- Rutgers University, the University of tories were being formed in 1970. A California-Santa Barbara, the Naval group was formed within NOAA’s Postgraduate School, the University of History and Technical new Environmental Research Labora- Rhode Island, and the University of Background for HF Radar tories to build a compact antenna sys- Connecticut. This was followed by a History tem to be used for coastal ocean surface surge in acquisition because of the Na- The present state of the U.S. na- current mapping. This was the Coastal tional Oceanographic Partnership Pro- tional high-frequency (HF) radar net- Ocean Dynamics Applications Radar gram, an NOAA/Office of Naval work has resulted from nearly 40 years (CODAR) program. After demon- Research/National Science Foundation of research and applications. HF radar strating its effectiveness, the NOAA/ program that funded coastal oceano- observations of the ocean surface truly National Ocean Service formed a Tran- graphic research at many of these same began with Crombie’s (1955) experi- sitional Engineering Program in 1978 universities. mental discovery of the mechanism to encourage development of a com- In 2002, California voters approved behind his puzzling analog sea-echo mercial version of CODAR. With funds that led to a program called spectral plots. Don Barrick (1968, only a small potential market, no exist- the Coastal Ocean Currents Moni- 1972) theoretically derived the model ing radar companies were interested in toring Program, which allowed for that indeed showed that this resonant commercializing CODAR so a small the investment of $21 million to scatter was in fact “Bragg scatter” and group left NOAA to start CODAR create a California network of HF related the echo strength to the ocean Ocean Sensors, Ltd. in the early 1980s. radars to measure ocean surface

122 Marine Technology Society Journal currents to ensure the monitoring of Data Buoy Center (http://hfradar. nications Conference in January 2012 coastal water quality. The acquisition ndbc.noaa.gov/) while data failover re- (Figure 1). began in 2005 with 40 CODAR dundancy is also provided at Rutgers radars eventually being integrated University. Data file management Physics of HF Radar with the then-existing 14 CODARs and distribution follow internation- Current Monitoring in California. ally accepted standards, for exam- Why HF radar? On a national scale, the Integrated ple, netCDF-CF file and metadata HF denotes that part of the electro- Ocean Observing System (IOOS ) formats and OpenGIS Web Coverage magnetic spectrum having frequencies Program has been facilitating the de-® Service Interface Standard® for inter- from 3 to 30 MHz, which is equivalent velopment of a national data manage- operable delivery of gridded data. to radio wavelengths of 10 to 100 m. ment and distribution system for all Nationally, an additional focus has HF radar has been shown to be the op- U.S. HF radars as well as radars operated been the effort to acquire primary timal method for coastal sea surface by the Canadian Coast Guard in Nova radio frequency licenses. To form an current mapping for a number of rea- Scotia. Presently, more than 100 HF ra- operational network, the radars need sons. First, the targets required to pro- dars and 30 institutions are part of the to operate at dedicated radio frequen- duce coherent sea echo using HF are network, and their data are delivered cies, which requires the approval of surface gravity waves, typically of sev- by IOOS national data servers. The the International Telecommunications eral to a few tens of meters wavelength, development server and data display Union as well as U.S. agencies. The which are well understood and nearly are provided by Scripps Institution of process to acquire those frequencies always present in the open ocean. Sec- Oceanography’s Coastal Observing Re- has been supported by NOAA IOOS ond, vertically polarized HF waves can search and Development Center (http:// for nearly 5 years, with the expec- propagate over conductive seawater cordc.ucsd.edu/projects/mapping/), tation that the final approvals will be via coupling to the mean spherical sea and its mirror is at the NOAA National given at the World Radiocommu- surface, producing measurement ranges

FIGURE 1

Montage of U.S. HF radar site locations. Green sites are sending data on schedule. Red sites have delayed data. (Color versions of ﬁgures available online at: http://www.ingentaconnect.com/content/mts/mtsj/2010/00000044/00000006.)

November/December 2010 Volume 44 Number 6 123 beyond line of sight, out to 200 km Doppler proportional to the relative apriorito properly remove the Bragg or more offshore. Third, Doppler sea speed of the ocean wave traveling di- wave phase speed: echo at HF, under most wave condi- rectly toward or away from the radar. sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi tions, has a well-defined signal from The transmit frequency of the radar gλo 2πD wave–current interactions that is easily determines the radar wavelength and, c ¼ tanh 2π λo distinguishable from wave–wave pro- hence, determines the length of ocean cesses. This allows for robust extraction waves from which the radar wave will of current velocities. It is primarily backscatter. Because attenuation in- Range and Velocity Determination these three features, along with the creases as frequency increases, the result All HF radar systems currently used spatial resolutions that are possible is that higher frequency radars (shorter for ocean measurements use some due to the frequency modulation dis- radar wavelength) have a shorter maxi- form of frequency-modulated contin- cussed below, which place the HF mum range. Approximately one third uous wave (FMCW) waveform for band in a unique status for coastal of the radars in the United States are range determination. FMCW has the current monitoring. in the 4- to 5-MHz band, which can benefitofmuchlowermaximum achieve 200 km or more, depending power requirements to achieve the on conditions. Another third operates same average power and, therefore, Physics of HF Sea Scattering in the 12- to 14-MHz band and can range performance as older time- The two environmental conditions achieve an approximately 90-km gated pulsed radars (Barrick, 1973). necessary for HF current mapping are range. Approximately one quarter For closely spaced or colocated trans- conductive surface water and the pres- of the radars operate in the 24- to mit and receive antennas, a pulsed ence of surface gravity waves of suffi- 27-MHz band and achieve ranges of and gated FMCW (or FMiCW, “i” = cient length and height. Conductivity approximately 45 km. At the higher interrupted) waveform is desirable of water is primarily determined by frequencies, it is possible to obtain whereby the transmit signal is cycled salinity, which is typically 32–37 PSU greater radio spectrum bandwidth on and off and radar echo received in in the open ocean. As salinity decreases, that in turn allows for higher range res- opposition over a period determined so does the strength of the sea echo and, olution. The resolutions vary from less by the system’s achievable range. therefore, range of measurement. Since than 1 km to approximately 6 km. Re- This is done to prevent saturation of freshwater is inherently 5,000 times less gardless of the operating frequency, the the electronics as well as the received conductive than seawater, HF signals do physics is the same. Assuming a sta- echo by the much stronger transmit not travel nearly as far (e.g. Fernandez tionary radar, the relative wave speed signal. For both cases, the fundamen- et al., 2000). It has been observed in is comprised of the phase speed of the tal range determination is the same bays and around river mouths that dur- Bragg wave plus any underlying cur- (Barrick, 1973). ing times of high freshwater discharge rent. For deep water, the phase speed A continuous linear frequency ranges can be significantly reduced forsurfacewavesiswellknownasa sweep (or chirp) over a fixed band- λ (e.g., Long et al., 2006). function of o: width and pulse repetition frequency The ocean surface, at any given mo- rffiffiffiffiffiffiffi is generated in the receiver and ampli- ment, contains a random structure of gλ fied for transmit. As scattered energy is c ¼ o crests and troughs, the slopes of which 2π received, it is delayed by the two-way scatter radar signals in all directions. travel time and shifted in Doppler on However, within the random surface, which can be subtracted leaving only the basis of the target velocity. When it is only the periodic structure of sur- the velocity of the current. This velocity mixed with the coherent linear sweep λ face waves whose wavelength, o,is is the projection of the actual current still generated inside the receiver, the precisely half the radar wavelength, λ, along the ray from the radar location time delay of the received echo results that will produce coherent backscatter. to the scattering area and is generally in a difference frequency train, which This is an analytic result known as referred to as a radial velocity.Inwater is digitized for range and Doppler pro- Bragg scattering. In the case of a stan- of shallow or intermediate depth, the cessing. By applying a fast Fourier dard backscatter (or monostatic) radar, water depth, D,mustbealsobe transform to the digitized signal, the the scattered energy will be shifted in known at each measurement location data can be sorted into discrete range

124 Marine Technology Society Journal bins at each sweep. Application of a FIGURE 3 second fast Fourier transform at each Representative HF radar Doppler spectrum. range bin over multiple sweeps pro- duces a Doppler spectrum at each range. A typical Doppler spectrum for a single receive antenna is shown in Fig- ure 2. The characteristic Bragg peaks from surface wave echoes are indicated with a positive Doppler shifted peak resulting from waves approaching the radar and negative Doppler shifted peak from waves retreating from the radar. Each peak is further spread because of the underlying current velocities present across the entire arc at the selected range. Also shown is the weaker second-order sea echo, which is a har- monic of the first order, whereby longer waves, not currents, modify the Doppler of the Bragg waves. Wave state information can be extracted from second-order termine the bearing to which each ve- wavelength apart. Phase differences spectra for certain wave conditions that locity can be attributed. In general, exist between signals received on the vary by radar frequency (Lipa, 1977) there are two classifications of bear- array elements that depend on the di- (Figure 3). ing determination commonly used rection of arrival. When the Doppler for HF radar: beam forming and direc- spectra of the individual array elements tion finding. are summed with the proper phase dif- FIGURE 2 Direction finding uses the phase ferences applied for a given bearing, a CODAR SeaSonde on San Clemente Island, CA. and amplitude differences between re- digital narrow beam is formed and a ceive antenna elements, known as the peak-picking algorithm used on the antenna response pattern. These dif- resultant spectrum. The digital beam ferences are applied to each Doppler width depends on the ratio of the bininthespectraoftheindividual wavelength divided by the array length elements to determine the most likely and on the bearing toward which the direction of arrival. Direction finding beam is steered (Skolnik, 1990). can be applied to compact directional antennas or to phased array antennas. Methods for Combining Radial It is most commonly used with the Current Vectors three colocated elements of the compact Although there are a variety of uses cross–loop/monopole configuration for radial vectors by themselves, most (e.g., Miller et al., 1985). Approxi- often the radial velocity vectors from mately 90% of the HF radars in the two or more sites must be combined Bearing Determination Methods United States use a direction finding to produce a 2-D map of the surface The final stage of processing radial method. current velocity. The problem inher- vectors is bearing determination. A Beam forming uses an array of re- ent in any combining method, how- single antenna can detect all of the cur- ceiving antenna elements, typically be- ever, is that each radar inherently rent velocities present at a given range, tween 8 and 16 in a linear alignment outputs radial vector data in a polar but more information is needed to de- and spaced about half of the radar grid centered on the radar location.

November/December 2010 Volume 44 Number 6 125 Mapping multiple sets of radial vec- There are presently 30 institutions Ltd.) standard-range and long-range tors from displaced polar grids onto a that contribute their data to the na- HF radar systems operating on the Cartesian coordinate system results tional HF radar network data manage- eastern coast of the United States. in variations in data density, signal ment system, which is funded by The USCG assessed the improvement strength, and geometric dilution of IOOS but relies on the voluntary ad- from HF radar data in their SAR plan- statistical accuracy (Chapman et al., herence to data file format standards ning process (Ullman et al, 2003). This 1997) across the field of coverage. by the HF radar operators. Users from study showed better comparison when A number of combining methods these institutions also routinely volun- CODAR-derived currents were com- have been developed including but teer their time for workshops such as pared against available NOAA tidal not limited to simple interpolation the Radar Operators Working Group current predictions. Along with these with vector addition, least squares (http://www.rowg.org), information key comparisons, an equally important methods on vectors falling inside a de- collection efforts such as the gap product was developed, the Short- fined averaging circle, and objective analyses, and standards compiled for Term Predictive System (STPS), which mapping. Recently, efforts have been the creation of the National Surface provides a 24-h forecast of surface made both in applying modal analysis Current Mapping Plan (http://www. currents based on the statistics of the to multiple radial data sets (Lekien ioos.gov/hfradar) and advisory panels previous 30 days of CODAR surface et al., 2004) as well as assimilating ra- such as the National HF Radar Tech- current data. Following these evalua- dial velocity data directly into models nical Steering Team to help make the tion studies, available in situ data without performing a separate ra- transition to an operational national were used to evaluate and define ap- dial combining step (Shulman et al., HF radar network. propriate parameters for inclusion in 2007). the USCG search planning tool. In May 2009, the current velocities from National Applications the Mid-Atlantic long-range CODAR On a national scale, there are two network and long-range STPS fore- IOOS HF Radar: An main applications presently underway: casts were included in the operational Exemplary Partnership (1) the U.S. Coast Guard (USCG) USCG SAR Optimal Planning System. In 1999, a number of HF radar Search and Rescue (SAR) operations For SAR cases in the Mid-Atlantic, researchers gathered informally in and (2) the NOAA oil spill response planners now have access to these data Oregon as a side meeting to a National operations. These applications use and forecasts within their operational Oceanographic Partnership Program ocean surface current data to track planning tool. awardees workshop. The clear benefits and predict the flow of the uppermost Because SAR is a national mission to everyone from having meetings layer of the ocean, and IOOS within encompassing all U.S. coastal waters, specifically designed to exchange in- NOAA is providing resources to bring the IOOS Program in NOAA is extend- formation and research about HF new capabilities to both of them. ing these Mid-Atlantic data products to radar gave birth to the Radiowave all coastal areas where HF radars are Oceanography Workshop (http:// USCG SAR Optimal located. This is a partnership with the radiowaveoceanography.org/) series Planning System USCG, the Scripps Institution of of meetings starting in 2001, which Beginning in 2000, the USCG Re- Oceanography, the University of Con- have continued annually ever since. search and Development Center began necticut, the Rutgers University, and Although completely self-funded, a multiyear investigation into the utility the Applied Sciences Associates that these meetings have been successful of real-time HF radar surface–current will extend the STPS and also provide in annually bringing together HF measurements for search and rescue a gap-filled current velocity field using radar experts at a dedicated forum in (SAR). In collaboration with the Uni- optimal interpolation (e.g., Kim et al., which to share state-of-the-art knowl- versity of Connecticut, the Univer- 2008) as input to the STPS. These edge. This series of workshops illus- sity of Rhode Island, and the Rutgers groups provide expertise from a spec- trates the level of cooperation and University, these drifter-verified tests trum of topics that are needed to pro- commitment within the HF radar were based around the CODAR vide a real-time end-to-end product, community. SeaSonde (CODAR Ocean Sensors, including data handling from the

126 Marine Technology Society Journal FIGURE 4 along the San Francisco waterfront. This closely matched visual reports Screenshots from USCG SAROPS. Left: search area without using HF radar data. Right: search area reduced by 2/3 when HF radar data used. Both after 96 h. of oil on the shorelines of Alcatraz, Angel Island, and San Francisco and on a map produced by the NOAA OR&R. Once the oil moved into the Gulf of the Farallones, the HF radar data accurately predicted that the oil would not beach there. As HF radar capabilities are integrated into California oil spill response, spills like the Cosco Busan’s (which occurred in dense fog) can be more effectively tracked, with mitigation efforts unimpeded by lack of visual data. radarsitetomultipledistributedna- radars, spanning more than 160 km The earlier Safe Seas exercise and tional servers, intermediate products of coastline and having 1- to 2-km res- use of HF radar data during the Cosco (STPS and optimal interpolation por- olution, provided continuous coverage Busan spill allowed OR&R to make a tions), and finally to the USCG Envi- during the 5-day exercise. This pre- seamless transition to utilizing Gulf of ronmental Data Server (Figure 4). paredness exercise provided a founda- Mexico HF Radar data soon after the tion for the use of HF radar data by Deepwater Horizon platform in the Oil Spill Response the NOAA OR&R spill response tra- northern Gulf of Mexico exploded Although the main impetus for cre- jectory modeling team (Figure 5). and sank in April of 2010. As of this ating the NOAA CODAR system in When the container vessel Cosco writing in August 2010, the HF radar the 1970s was for oil spill response, it Busan collided with the base of the data are still being used daily. Partners was not until 2006 HF radar was used Bay Bridge in San Francisco Bay in from the University of Southern Mis- by official government spill respond- November of 2007, spilling more sissippi and the University of South ers. In August of 2006, the National than 53,000 gallons of fuel oil, man- Florida have monitored their radar sys- Ocean Service and the USCG led an agers used surface current maps from tems constantly to ensure that they are interagency field exercise, Safe Seas HF radar data to monitor the spill tra- operating while the Deepwater Hori- 2006, in the San Francisco Bay area jectory, predicting movement as far zon spill continued and that the data to enhance the preparedness of oil north as Angel Island and westward were delivered to the IOOS national spill responders. As part of that exercise, the IOOS program collaborated with the NOS Office of Response FIGURE 5 and Restoration (OR&R) to create Schematic of data flow for new HF radar SAROPS project. hourly gap-filled maps of HF radar- derived surface currents. The IOOS partners at the San Francisco State University and the Naval Postgraduate School created new data handling soft- ware and implemented a real-time open-boundary modal analysis suite of algorithms (Kaplan and Lekien, 2007). These nowcasts were then for- matted into files that were readily ingested by the General NOAA Oil Modeling Environment. Eleven HF

November/December 2010 Volume 44 Number 6 127 FIGURE 6

HF radar currents for June 4, 2010, overlaid with oil coverage in the Deepwater Horizon spill area in the northern Gulf of Mexico, courtesy Rutgers University Coastal Ocean Observation Lab.

HF radar data servers at Scripps In- Regional Applications mation on ocean conditions that likely stitution of Oceanography and the Tracking Impacts on influence the survival rate of young NOAA National Data Buoy Center. Marine Populations salmon when they first enter the These Gulf of Mexico sites have been Ocean conditions change from year ocean. As smolts exit estuaries like the particularly valuable since they cover a to year and the ongoing measurements Russian River in early spring, strong good portion of the continental shelf of surface currents made by HF radar northerly winds and southward-moving in the Mississippi Bight, which is just are a crucial backbone for ocean obser- currents can carry weakly swimming to the north and northeast of the site vations along the coast. Unlike buoys small fish south to the predator-rich where the Deepwater Horizon was lo- and ships, which collect information Gulf of Farallones in some years or cated (Figure 6). at single points and times, HF radar alongshore to the north in others. Pre- Similar to USCG SAR, the op- provides full, archived mapping, day liminary evidence suggests that surface timally interpolated current velocity and night, of our coastal waters to flows in the months leading up to the fields, mentioned earlier, will also pro- 150 km offshore. Long-term monitor- spring emigration period may be im- vide a product that can be ingested into ing of surface currents is used to track portant for the survival of salmon smolts the NOAA oil spill response team’s impacts on marine populations. Off and returns to the Russian River sys- General NOAA Oil Modeling Envi- Bodega Bay, California researchers are tem years later (W.J. Sydeman/Farallon ronment model for application wher- using HF radar-derived surface current Institute and J.L. Largier/Bodega ever HF radars operate. data to obtain seasonal to annual infor- Marine Laboratory, unpublished data).

128 Marine Technology Society Journal Reversing the collapse of the California California coast is experiencing an in- Within the Northwest Association salmon fishery requires an understand- creasing frequency and toxicity of of Networked Ocean Observing Sys- ing of the migratory paths of young harmful algal blooms (HABs), exacting tems region, the PacificNorthwest salmon as well as knowledge of the serious economic, human, and marine Harmful Algal Bloom Bulletin has movement of nearshore surface currents wildlife costs. Surface current mapping been developed by the NOAA and and upwelling events that comprise their has proven to be an essential tool for the University of Washington to pro- ocean going habitat. managers and scientists to assess and re- vide a comprehensive early warning in- Coastal surface currents can also spond to HABs and will be instrumen- formation system for Washington provide important input to establishing tal in developing the ability to predict coast razor clam toxicity and amnesic and evaluating marine protected areas these events. Like all food chain com- shellfish poisoning events. The bulletin (MPAs); it provides the only multiyear ponents, HABs are part of a larger builds upon the Olympic Region HAB data with enough spatial coverage to marine ecosystem driven by the phys- monitoring program and Ecology and assess how larvae of marine populations ics of winds, waves, and currents. Oceanography of Harmful Algal are dispersed from the location where HF radar has become a core technol- Blooms in the Pacific Northwest re- they originate to where they settle and ogy for understanding these ecosystem search by automating the aggregation grow to maturity. HF radar data from a processes. ofdataintoasinglelocationona regional network in California have A California statewide Harmful Web-based information dashboard. demonstrated the connectivity between Algal Bloom Monitoring and Alert Pro- Amongthearrayofchemicalandbio- central California marine protected gram that was initiated by the NOAA, logical information included are cur- areas (MPAs) by back-projecting tra- the California Ocean Science Trust, rents from HF radars that operate jectories from 10 MPAs more than a and the Southern California Coastal within Northwest Association of Net- 40-day period. Clarifying this connec- Water Research Project is supported worked Ocean Observing Systems tivity is an important step toward under- through the Ocean Observing Regional (Trainer and Hickey, 2010). standing the movement of invertebrate Associations. Weekly bottle samples The goal of assimilating HF radar- and fish larvae (Zelenke et al, 2009) measure chlorophyll, nutrients, domoic derived currents into numerical circu- (Figure 7). acid, and harmful algal species. Data are lation models has for a number of years HF radar data are also being used to posted to the Web and distributed via remained a priority within the modeling identify and track large eddy features the California Harmful Algal Bloom and HF radar research communities. (tens of kilometers wide) off Cape Monitoring and Alert Program Listserv. Generally, these models are developed Mendocino, Point Arena, and in the When HABs are detected, opportunis- for areas that scale to approximately Santa Barbara Channel. These eddies tic sampling from additional shore sites, that of an IOOS regional coastal play a critical role in connecting or dis- HF radar-derived surface currents, glid- ocean observing system. A number rupting marine populations that live ers, and boats determines their extent of successful modeling projects are along the coast of California. The and severity. described in the National Surface Current Mapping Plan (http://ioos. FIGURE 7 gov/hfradar), and a recent American Geophysical Union Meeting of the Color map: location of waters 40 days ago (red), 30 days ago (yellow), 20 days ago (green), 10 days Americas 2010 (Foz do Iguaçu, ago (cyan), and 5 days ago (blue) before reaching the labeled MPA (magenta). Connectivity maps on the basis of measured surface currents show what waters are influencing MPAs and the potential Brazil, program available here) held extent of surface water larval transport. a session on Application of HF Radar Networks to Ocean Forecasts.In addition, as part of the recently es- tablished National HF Radar Tech- nical Steering Team, the IOOS HF radar community is presently under- taking a comprehensive review of themanymodelingeffortsthatuse HF radar data throughout the globe.

November/December 2010 Volume 44 Number 6 129 Local Applications HF radar data were used to track move- Since 2008, several floatable events Coastal Water Quality ment of the effluents based on real-time along the New Jersey coast have In southern California, HF radar- observations of ocean surface currents prompted investigations on possible derived surface currents has allowed from the HF radar network. “[A scien- sources and ultimate fate of debris managers to track the movement of tist] was able to rapidly provide daily thathaswasheduponlocalbeaches. planned and unplanned discharges in and cumulative modeling of effluent For example, in August 2008, medical our coastal waters, enabling more pre- trajectories that really demonstrated waste washed up on the shores near cise and timely management decisions. the immediate value of the existing pro- Avalon, New Jersey. The New Jersey An Orange County Environmental gram,” said Michael Kellogg of the San Department of Environmental Protec- Health Engineering Specialist, familiar Francisco Public Utilities Commission. tion asked the mid-Atlantic HF radar with the Tijuana River outflow issues, This information significantly improved network managers at Rutgers Univer- wrote that “this real-time surface cur- the decision-making and response capa- sity to provide information on the pos- rents monitoring system has allowed bilities of the utilities commission. The sible source. Using the location and the the San Diego County Environmental trajectories showed a weak onshore flow, time of the initial beach location of the Health Agency to predict when con- indicating that the discharge would not debris, Rutgers radar scientists were taminated water from the Tijuana move toward beaches by the time the able to trace back its probable location River will impact the southern beaches rupture could be repaired; this allowed several days before the washup. The of San Diego County.” In November responding agencies to better manage weak currents indicated that if the de- of 2006, the City of Los Angeles di- beach closures, offshore and onshore bris were put into the ocean within sev- verted the flow from Hyperion—its water quality monitoring, and outfall eral days of the initial siting, it had to oldest and largest wastewater treatment repairs (Figure 8). be a local source. Consistent with the plant—from an outfall 5 miles from the shoreline to a rarely used pipe 1 mile FIGURE 8 offshore to allow inspection of the Upper panel: shows the near real-time Hyperion Outfall plume trajectory color coded based on 5-mile pipe. The diversion lasted particle age (dark blue—0 days; red—3 days). The color coding is based on approximate life cycle 3 days, and approximately 800 million of bacteria. Lower panel: distance along the coastline from the Hyperion Outfall with Los Angeles gallons of secondary-treated wastewater County sampling locations red if there is plume potential. was released 1 mile off the coast of Santa Monica. A division manager for the City of Los Angeles, Bureau of Sanitation’sEnvironmentalMonitor- ing Division writes that the city’s monitoring effort greatly benefited from information provided through the HF radar system and that “the real- time current information provided through [the program] enabled us to adaptively modify our sampling grid to better track the discharge plume and to predict the dispersion of the plume.” In October of 2007, the end gate to the Southwest Ocean Outfall offshore Ocean Beach in San Francisco was lost; a buoyant mixture was released from thepipe6.5kmoffshoreandroseto the surface. At the request of the San Francisco Public Utilities Commission,

130 Marine Technology Society Journal guidance from the HF radar data, the ities to develop a 3-D wind resource Barrick, D.E. 1973. FM/CW Radar signals investigation determined that the map to support the offshore wind and digital processing. NOAA Technical source was in fact a dentist who energy community. The work will use Report ERL 283-WPL 26, July 1973. dropped the waste from a boat just available forecast models and a new de- Chapman, R.D., Shay, L.K., Graber, H.C., off the beaches of Avalon the day be- ployment of a radar subnetwork (four Edson, J.B., Karachintsev, A., Trump, C.L., fore. This result, along with other sites) along the southern New Jersey Ross, D.B. 1997. Intercomparison of HF events in the region, has highlighted coast. This is a 2-year grant that lever- radar and ship-based current measurements. the need to extend the regional cover- ages IOOS infrastructure and creates a J Geophys Res. 102:18,737-48. doi: age of the present HF radar network higher resolution HF radar coverage 10.1029/97JC00049. closer to the coast. These local en- area within the Mid-Atlantic Bight. Crombie, D.D. 1955. Doppler spectrum of hancements are being initiated in the sea echo at 13.56 mc/s. Nature. 175:681-2. Mid-Atlantic Bight with leveraged doi:10.1038/175681a0. state agency resources to build out Summary nested high-resolution HF radar sites HF radar as a tool for ocean surface Fernandez, D.M., Meadows, L.A., Vesecky, and assimilation of these data into current mapping has been in existence J.F., Teague, C.C., Paduan, J.D., Hansen, P. coastal models tuned to track particles for more than 30 years. It has proven 2000. Surface current measurements by HF along the coast. itself in a number of applications of na- radar in freshwater lakes. IEEE J Oceanic Eng. tional, regional, and local significance, 25(4):458-71. Marine Navigation especially during the last 10 years or so. Kaplan, D.M., Lekien, F. 2007. Spatial The physics of the measurement and interpolation and filtering of surface current HFradardataareacorecompo- the technology that delivers the mea- data based on open-boundary modal nentofasimplebutveryeffective sured ocean current velocities provides analysis. J Geophys Res. 112:C12007, near real time, customized, interactive a robust method for coastal monitoring doi:10.1029/2006JC003984. Website displaying environmental from nearshore to more than 200 km conditions at the entrance to the Kim, S.Y., Terrill, E.J., Cornuelle, B.D. 2008. offshore. Through an integrated net- Ports of Los Angeles and Long Beach Mapping surface currents from HF radar work of radars distributed throughout Harbor: http://www.sccoos.org/data/ radial velocity measurements using optimal U.S. coastal waters, data are delivered harbors/lalb. This Website could serve interpolation. J Geophys Res. 113:C10023, in near real time for use in a number doi:10.1029/2007JC004244. as a template for ports throughout the of applications that are critical to the United States. This application is Lekien, F., Coulliette, C., Bank, R., Marsden, health, safety, ecology, and economies discussed more fully in a companion J.E. 2004. Open-boundary modal analysis: of coastal areas. article by Thomas et al. in this issue. interpolation, extrapolation, and Integrating HF radar data with ex- filtering. J Geophys Res. 109:C12004. doi:10.1029/2004JC002323. isting conventional in situ sensors will Lead Author: also occur in an upcoming demonstra- Jack Harlan Lipa, B.J. 1977. Derivation of directional tion project in Mobile Bay, Alabama, NOAA IOOS Program, ocean-wave spectra by integral inversion of involving Mobile’s NOAA Physical Silver Spring, MD® the second-order radar echoes. Radio Sci. Oceanographic Real-Time System 12:425-34. doi:10.1029/RS012i003p00425. Email: [email protected] (PORTS ) and two CODAR systems, Long, R., Garfield, N., D. Barrick, D.E. operated® by the University of South- 2006. The effect of salinity on monitoring ern Mississippi. This project may References San Francisco Bay surface currents using provide a basis for consideration of Barrick, D.E. 1968. A review of scattering Surface Current Monitoring Instruments a Gulfport, Mississippi HF radar- from surfaces with different roughness scales. (SMCI). Presentation at the California PORTS® equivalent. Radio Sci. 3:865-68. and the World Ocean ’06 Conference, September 17-20, Long Beach, California. Barrick, D.E. 1972. First-order theory and Offshore Wind Energy analysis of MF/HF/VHF scatter from the sea. Miller, P.A., Lyons, R.S., Weber, B.L. 1985. Rutgers University has been funded IEEE Trans Antennas and Propag. AP-20(1): A compact direction-finding antenna for by the New Jersey Board of Public Util- 2-10. doi:10.1109/TAP.1972.1140123. HF remote sensing. IEEE Trans Geosci

November/December 2010 Volume 44 Number 6 131 Remote Sens. GE-23:18-24. doi:10.1109/ TGRS.1985.289496.

Shulman, I., Kindle, J., Martin, P., deRada, S., Doyle, J., Penta, B., Anderson, S., Chavez, F., Paduan, J., Ramp, S. 2007. Modeling of upwelling/relaxation events with the Navy Coastal Ocean Model. J Geophys Res. 112:C06023. doi:10.1029/2006JC003946.

Skolnik, M. 1990. Radar Handbook. 2nd ed. New York: McGraw-Hill, Inc.

Trainer, V., Hickey, B.M. 2010. A Forecasting Bulletin for Harmful Algal Blooms in the Paciﬁc Northwest, Eos Trans. AGU, 91(26), Ocean Sci. Meet. Suppl., Abstract IT33B-06.

Ullman, D., O’Donnell, J., Edwards, C., Fake, T., Morschauser, D., Sprague, M., Allen, A., LCDR B, Krenzien. 2003. Use of Coastal Ocean Dynamics Application Radar (CODAR) Technology in U.S. Coast Guard Search and Rescue Planning. US Coast Guard Report No. CG-D-09-03.

Zelenke, B., Moline, M.A., Crawford, G.B., Garﬁeld, N., Jones, B.H., Largier, J.L., Paduan, J.D., Ramp, S.R., Terrill, E.J., Washburn, L. 2009. Evaluating connectivity between marine protected areas using CODAR high-frequency radar. In: OCEANS 2009 MTS/IEEE Conference, Biloxi, MS.

132 Marine Technology Society Journal