P A P E R Waves Initiative within SEACOOS AUTHORS ABSTRACT 1 George Voulgaris Amongst other ocean state parameters, the development of a wave measurement pro- Marine Science Program, Department gram was supported as part of the Southeast U.S. Atlantic Coastal Ocean Observing System of Geological Sciences, University of (SEACOOS). The program focused on supporting nearshore wave measurements using South Carolina both cabled and autonomous systems but also examined the feasibility of using HF Radar Brian K. Haus systems for remote estimation of wave parameters. The nearshore stations have provided Rosenstiel School of Marine and Atmos- a significant database on directional wave climate for a number of nearshore locations in pheric Sciences, University of Miami the region that provide valuable information to coastal engineers and managers for sustain- able development along the coast of the southeastern United States. The ability of Wellen Paul Work high-frequency radar (WERA HF) to provide wave information was evaluated through a field School of Civil and Environmental experiment in SE Florida. The results were encouraging and placed some initial bounds on the Engineering, Georgia Institute of confidence to be associated with empirically derived wave height information. Coordination Technology, Savannah efforts for the development of a comprehensive waves program for the Southeast U.S. were Lynn K. Shay initiated and contributed to the development of the National Wave Observations plan. They Rosenstiel School of Marine and also led to the development of a new Regional Coastal Ocean Observing System (RCOOS) Atmospheric Sciences, University that includes developing systems in support of local weather forecast offices in their surf- of Miami zone and rip- forecasts.

Harvey E. Seim activities, and navigation, particularly near surface gravity waves as these are strongly Department of Marine Sciences, University ports and inlets. affected by local bathymetry and gradients of North Carolina at Chapel Hill Although not a standalone theme of in atmospheric forcing. Finally, goal (4) was Robert H. Weisberg the program initially, the recognition of to implement wave forecasting capabilities College of Marine Science, University the importance of surface wave parameters with high spatial resolution to resolve wave of South Florida led to SEACOOS leadership providing condition variability in the nearshore. resources for a wave measurement program In order to achieve the goals defined James R. Nelson with the vision of later proceeding to wave above, a number of pilot wave measure- Skidaway Institute of forecasting. This program was a key com- ment installations/studies were carried out ponent of a second phase of development that focused on: (i) continued operation for SEACOOS. The activities described of existing sites for the provision of wave 1. Introduction herein, termed the “Waves Initiative”, information (e.g., South Atlantic Bight urface waves are an important physi- consisted of planning and pilot studies Synoptic Offshore Observational Network cal mechanism influencing a number of undertaken in the initial phase of program [SABSOON] towers in Georgia); (ii) de- Soceanic processes ranging from mixing development (SEACOOS Implementation velopment of new sites in the nearshore at Plan, 2004). The specific goals of the sur- scientifically and socially relevant locations and CO2 exchanges between the ocean and the atmosphere, to sediment transport, face wave program within SEACOOS were (e.g., Springmaid and Folly Beach Piers, coastal erosion and coastline evolution. to: (1) create a directional wave data set SC, Savannah River Entrance Channel, In the context of regional ocean observ- (climatology) for several nearshore areas for GA), with a focus on directional wave ing, the latter and other coastal processes use as design criteria for sediment transport measurements; and (iii) adaptation and are of considerable societal and economic and coastal engineering studies; (2) provide evaluation of existing or developing tech- interest, especially given various scenarios wave data real-time to the public and other nology (i.e., WERA Radar) for remote of global climate and local relative sea level stakeholders for operational purposes (e.g., measurements of waves with high spatial changes (IPPC, 2007a, b). In addition, search and rescue) and decision-making; resolution. The first two activities were surface waves can strongly influence pub- (3) develop a database of nearshore direc- initiated as a proof of concept, while the lic safety through their impact on search tional wave data for the development and latter was developed and carried out later and rescue operations, recreational beach calibration of coastal (inner shelf and surf on in the project. zone) wave forecast systems. Such systems The Waves Initiative was partially suc- 1Email: [email protected], require high resolution wave fields that cessful in achieving objectives (1) and (2) Tel: + 803 777-2549 resolve the directional characteristics of with limitations in observational continuity

68 Marine Technology Society Journal being the primary obstacle. These breaks in current profiler (ADCP). Finally, a large (SABSOON) program initiated in 1998 data coverage are related to general opera- section is dedicated to describing research with funding from the National Oceano- tional difficulties explained in Nelson and and development activities designed to graphic Partnership Program (NOPP). The Weisberg (this volume) but also to imple- evaluate the use of HF Radar technology local water depth is 27 m and the sensor mentation- and site-specific limitations for wave measurements. is installed 6 m below mean sea level; it is as described below. Limited availability sampled continuously at 2Hz. The sensor of funds prohibited the development of is cabled and power supply and data trans- a sufficient number of systems to resolve 2. In Situ Observations mittal is integrated within the SABSOON the regional and sub-regional variations in The in situ wave monitoring systems infrastructure. Spectral analysis and linear nearshore waves. Goals (3) and (4) were not used in SEACOOS varied from the wave theory are used to compute surface advanced significantly within SEACOOS simplest, single transducer point wave height after compensation for pres- due to lack of financial support and effec- measurement of non-directional wave sure attenuation with depth (e.g., Bishop tive coordination between the regional and characteristics, to directional wave obser- and Donelan, 1987; Voulgaris et al., 1995). federal partners. The implementation of the vations using ADCPs with the Waves® Wave statistics are generated hourly and the National Program on Wave Observations option, and wave buoys. Based on the in- data are transmitted to the National Data (NOAA/IOOS, 2008), which was formed stallation and data transmission methods, Buoy Center (NDBC) for real-time display after the end of the SEACOOS project, the systems are categorized as cabled or (see Station SPAG1 at www.ndbc.noaa.gov) may provide a mechanism for improving autonomous systems. and archiving. these collaborations in future efforts. It is This station has provided an almost characteristic that within the SE region, 2.1 Cabled Systems continuous record of wave conditions for there were areas that had developed wave 2.1.1 Non-directional Wave Character- over 7 years, creating a long-term database forecasting capability (e.g., Jacksonville, istics (SABSOON Pressure Sensor) on wave conditions. These data have been FL, had in place a wave forecasting system One of the first attempts for wave meas- instrumental in assessing the role of waves developed by the Naval Research Labora- urements focused on utilizing existing sen- in benthic primary production in the mid- tory with funds through the National sors and infrastructure to provide real-time shelf sandy seabed of the south Atlantic Oceanic and Atmospheric Administra- measurements of surface waves. A Parosci- Bight (Jahnke et al., 2008). Subsurface tion/National Weather Service, [Welch, entific pressure transducer is installed on visible light (PAR) data collected as part of pers. comm.]) but had no in situ data to the Navy tower R2 (offshore Georgia, see the same measurement system reveal that ground-truth the system. On the other Figure 1) as part of the South Atlantic Bight prior to the existence of coastal observing hand, there were regions with operational Synoptic Offshore Observational Network systems, ship-based estimates of average at- wave measurement stations (e.g., in South Carolina and Georgia) but no correspond- Figure 1 ing forecasting capability. Locations of wave in situ sensor measurements operated by SEACOOS in the GA, SC area discussed in this paper. Although funding shortfalls did not permit the SEACOOS program to begin its planned second phase of development, a significant amount of experience was gained and the progress made is invalu- able. This contribution describes the efforts and progress in the area of surface waves; it describes the development of the in situ measuring stations operating in the SEA- COOS region (see Figure 1). The develop- ment and operation of two cabled nearshore directional wave installations are described first, with some results on nearshore wave climatology developed from those stations. This is followed with a description of a wave buoy observational activity that included an intercomparison of surface wave parameters measured by the buoy with those measured by an RD Instruments acoustic Doppler

Fall 2008 Volume 42, Number 3 69 tenuation of PAR were biased toward calm- in the establishment of two nearshore wave ADCP unit to the station at the end of the er, lower attenuation conditions (Jahnke et and current monitoring stations along the Pier. The armored cable was selected for al., 2008). The data revealed that significant coast of South Carolina. The two sites durability and for its ability to self-bury wave heights greater than 2.5 m, common were adjacent to fishing piers (Springmaid on sandy substrates. The ADCP itself was in fall and early winter, greatly reduce and Folly Beach Piers in Myrtle Beach and installed on a trawl-resistant bottom mount benthic PAR flux. This result provides an Folly Island, respectively). The existence of manufactured by Mooring Systems, Inc. example of how data from regional coastal pier structures, extending approximately The cable was secured to the underwater observing systems can also be used by the 1100 feet into the nearshore, provided mount and the pilings of the pier respec- scientific community to relate results from ideal locations for the development of such tively using armored cable grips to provide short-term, process-based experiments to stations (see Figure 6 in Nelson and Weis- relief from tension . longer-term issues with potentially more berg, this volume). The guiding principle A suite of software programs was writ- general applicability. for the development of these stations was ten using Perl scripting language to auto- the need for the collection of directional mate file management at the pier end and 2.1.2 Nearshore Directional Waves wave conditions and currents at locations to automatically transmit data using FTP to and Current Measurements outside the breaker zone suitable for coastal a server at the University of South Carolina The development of the RDI Waves® engineering applications and for forecast- in Columbia, SC, for further processing module for measuring waves and currents ing of nearshore wave conditions. A base prior to disseminating via an Internet with a single upward-looking acoustic station was established at the end of each web page. Significant effort was placed Doppler current profiler (Strong et al., pier where a PC-104 computer is used in presenting the complicated nature of 2000) presented a very attractive methodol- to control data collection and process- directional wave information collected by ogy for the simultaneous collection of mean ing of the ADCP wave data using the the system. The adopted approach focused flow and directional surface wave informa- WAVESMON® software provided by the on displaying all the spectral information tion. Although ADCPs had been routinely instrument manufacturer. Internet access available in a variety of formats allowing used for measurements of mean currents in was established at each station through the user to extract the amount of informa- real-time operations, the development of either the deployment of a fiber optic line tion suitable for his/her purposes. The wave measuring capabilities resulted in a (Springmaid Pier) or the use of an Ethernet data displayed on the web include tables significant increase in the amount of data wireless bridge (Folly Beach) that ensured showing representative wave height (Hs), collected per unit time, making wireless remote control of the systems as well as peak and mean wave periods (Tp and Tm, underwater data transmission very diffi- real–time data transmission. respectively) and direction of propagation cult with the available technology. Cabled An underwater, seven-conductor, of the peak period wave. Visual displays of operation of these systems provided an armored cable (Rochester Stock Type 7-H- the full directional spectrum, the distribu- alternative that was adopted by SEACOOS 422A) was used to connect the underwater tion of wave energy by frequency and the

Figure 2

Example of visualization of directional wave (a) and current (b) information from the cabled ADCP-based nearshore stations operated by the University of South Carolina. (a) Left: non-directional wave energy spectrum (top) and directional distribution of total wave energy (bottom). Right: Tabular form of wave and current statistics (top) and directional wave energy distribution with the current speed superimposed on it (yellow vector). (b) Three-dimensional vector plot of currents as well as their relation to the coastlines. These images are automatically created from the ADCP data after it has been processed by the Wavesmon® software, and updated hourly.

70 Marine Technology Society Journal directional distribution of the wave energy geol.sc.edu/gvoulgar/ww.html), the wave Despite these difficulties, the two for both the sea and the waves are also statistics are transmitted to NDBC for systems have been successful in collect- shown (see Figure 2a). The direction of display and archival. ing nearshore directional waves, enabling the local coastline is superimposed on the The Springmaid Pier station was the determination of wave climatologies directional spectrum display, so that the installed in December 2004 while the (see Figures 3 and 4) describing the local user can infer information on the direction Folly Beach Pier station installation was wave field and its alongshore direction. and strength of the longshore current that completed in February 2005. Both systems The climatology analysis reveals that 2 might be developed within the surf-zone. have been operational since that time and although wave power (defined as Hs ∙T) The three-dimensional structure of the the real-time data recovery has been 46% is at times incident on the site from both nearshore currents is also shown in a dia- and 60% of the total time for the Folly alongshore directions (in relation to the gram by itself (see Figure 2b). In addition Beach and Springmaid stations, respec- normal to the local coastline), most of it to local web displays (i.e., http://www. tively. Their respective mean water depths comes from the northeast, contributing are 6 and 4.5 m, reflecting the differences to a net southwestward directed longshore Figure 3 in beach morphology at the two locations. drift. It is characteristic (Figure 3), that Nearshore wave climatology for the two cabled The most significant problem encountered the magnitude and asymmetry is greater ADCP stations (Folly Beach and Springmaid Pier) with the cabled systems has been damage for Folly Beach, an area with significant along the coast of South Carolina. Under the same by lightning strikes. During a period of 4 erosion problems that has been the subject wind regime, the site at Folly Beach receives sig- years, 4 computer systems were replaced of numerous beach nourishment activities. nificantly more energy than the site on Springmaid and the ADCP boards were seriously dam- The wave period-wave height joint distri- Pier. At both sites the wave energy is directed south- ward, indicating a southward dominated longshore aged 3 times, requiring lengthy repairs by bution (Figure 4) reveals that the waves sediment transport. Data included cover the period the manufacturer. During these instances, approaching the Folly Beach area include from deployment (December 2004 and February lack of spare ADCP systems resulted in both local sea waves with periods of 4-6 s 2005, for Springmaid and Folly, respectively) to significant delays (>2 months) in bringing and swell (8-11 s period). On the contrary, December 2007. the system back online. In addition, a beach the waves in the nearshore at Myrtle Beach nourishment project at Folly Beach that was are more sea waves generated locally within carried out in 2006 led to the unexpected the embayment. accumulation of fine material (silt and clay) at the 4-5 m depth contour that resulted in 2.2 Autonomous Systems the burial of the ADCP transducers three There are nearshore sites where the use times, disrupting data collection. One of cabled systems is not feasible (e.g., heavy incident was reported where the ADCP shrimping activity, dredging operations etc) connection was damaged by trawling and stand-alone systems are preferred. In activity, despite the fact that the system such locations, the utilization of surface was deployed in an area where shrimping buoy and/or a modification of an ADCP activity is prohibited by state law. for use without cable are of merit and such Figure 4

Joint distribution of significant wave height and peak period at (a) Folly Beach, and (b) Springmaid Pier along the coast of South Carolina (for locations see Figure 1).

Fall 2008 Volume 42, Number 3 71 activities are briefly described below. In oth- agreement in spectra-derived parameters water acoustic modems (Cole and Weis- er areas where depth increases rapidly with reported by the two systems, with some berg, 2008). The NEMO® development distance offshore (e.g., Southeast Florida), significant differences at the upper and was driven by Coastal Observing initiatives bottom-mounted instruments may not be lower frequency measurement limits attrib- such as SEACOOS, but its utilization useful because of their inherent limitations uted to lower signal-to-noise ratios at these and performance for operational systems for observing higher frequency waves. frequencies. The wave buoy consistently depends on the quality and capabilities of reported greater wave energy at frequen- the underwater modem communications 2.2.1 Surface Buoy cies below 0.05 Hz, leading to larger mean systems used. A Tri-Axys wave buoy (manufactured and peak period estimates than reported by Axys Technologies, Canada) was used at by the ADCP. It was confirmed that the a site near the seaward end of the Savannah directional resolving power of the ADCP 3. Remote Measurements River Entrance Channel (see Figure 1) in was greater than that of the buoy (both da- of Waves Using High Georgia from 2004-2007 (Work, 2008). tasets were evaluated using the Maximum Frequency Radar Systems The buoy was deployed in water depth of Entropy Method) but this is inherent to Given the problems associated with 14 m (tidal range 2.1 m) and programmed the wave measurement methods employed installation and maintenance of in situ wave to report hourly estimates of directional by the two systems (i.e., six independent measurement systems, the development of surface wave energy spectra and related time series used by the buoy vs. twelve time empirical methods for extracting basic wave parameters via onboard Iridium telemetry series utilized by the ADCP wave array). It parameters from high frequency (HF) radar (with Inmarsat-D+ telemetry as backup). was noted that both systems gave similar backscattered Doppler spectra is highly Raw data were also logged onboard for mean and peak wave direction estimates, attractive. Such an approach can provide download during servicing. The ease of with the ADCP-derived wave energy being larger spatial coverage than is feasible using deployment for establishment of a real-time often more concentrated around the peak point measurement systems. Furthermore, data reporting station was a major factor in direction. This analysis confirmed that its remote, land-based operation signifi- the choice of this system. It required little wave parameters from the two different cantly reduces the operational difficulties development compared with the ADCP systems within the SEACOOS domain related to data collection in the marine cabled systems (see section 2.1.2), and no are comparable, although the ADCP has environment. However, it is imperative need for a coastal shore station since the the advantage of providing simultaneous that prior to adoption of these methods telemetry can reach any standard telephone estimates of ocean currents. However, in for operational purposes, an evaluation for modem with equal ease. The buoy provided areas where no ADCP can be deployed, reliability and accuracy is undertaken. real-time data for 62% of the time, com- water depth is too great, or no facilities are Empirically-based methods have been parable to the performance of the ADCP available for a coastal instrument station, tested and validated for the phased-array cable systems. As with the ADCP systems, the use of a surface buoy appears to provide Ocean Surface Current Radar (OSCR) HF lack of redundant hardware was the main comparable wave results. radars by Graber and Heron (1997). Their factor controlling the length of gaps in the approach was extended and extensively data set. The failures that did occur were 2.2.2 Stand-Alone, Bottom-Mounted tested by Ramos (2005) and Haus et al. mainly attributed to mooring failures and Systems (NEMO® – Acoustic Telemetry) (2006), where RMS differences in Hs of occasional malfunction following impacts Attempting to transmit all of the raw 0.21-0.50 m were found in comparisons of the buoy by passing vessels. data required for routing determination of with multiple in situ observations. Al- In addition to the operational collec- directional surface wave energy spectra will though larger differences were expected tion of wave data, the buoy was used in a overwhelm most telemetry systems. The in regions of high spatial variability of the 2.5-month intercomparison experiment two obvious to this problem are wave field because of the spatial smooth- with a 1200 kHz ADCP, similar to that to use a cabled telemetry system to increase ing inherent in the radar observations, used in the cabled systems described above. bandwidth or to process data in situ and the observed differences were of the same Hourly buoy observations from this period then transmit only the resulting processed order as those typically found between in were compared to simultaneous meas- parameters. During the project period, situ observations (Graber et al., 2000). The urements from the ADCP as described RD Instruments released the NEMO® same methods were applied to the Wellen in Work (2008). The data comparison as an add-on system that takes the latter Radar (WERA) systems used here to included both directional and non-di- approach; it carries out wave processing study the growth of surface waves over the rectional surface wave energy spectra and calculations underwater, thus reducing the Florida Current (Haus, 2007). These stud- bulk wave parameters (height, period, amount of data transmitted by three orders ies demonstrated the utility of empirical and direction). The results (see Work, of magnitude. This, in principle, allows the approaches for making spatially distributed 2008) indicated that there is a very close transmission of wave statistics using under- wave height measurements.

72 Marine Technology Society Journal 3.1. SEACOOS HF Radar Wave Figure 5 Table 1

Measurements Wave coverage area of WERA as deployed for SEA- WERA system characteristics as deployed over The WERA HF radars used for these COOS based on a 50 km range limitation and a 120º the Southeast Florida Shelf during the mini-waves studies were deployed as a real-time, angular swath from each station. Crandon station experiment. (CDN) and north Key Largo station (NKL) shown. operational component of the Southeast Operating Frequency 16.045 MHz Atlantic Coastal Ocean Observing System Solid arrows show the direction of the normal to the receive array (boresight) of each radar station. Transmitted Peak Power 30 Watts (SEACOOS; see Shay et al., this volume). Shaded region denotes area where the two stations Bragg Wavelength 9.35 m The WERAs are phased-array systems that overlap to provide directional spectral observations. Measurement Depth ~0.8 m use a Frequency Modulated Continuous Box denotes location of in situ measurements and is Operational Range for Currents 80-120 km Wave (FMCW) transmission to interrogate shown in more detail in Figure 7. Solid circle shows the ocean surface (Gurgel et al., 1999). This location of Fowey Rocks CMAN station. Operational Range for Waves 40-60 km installation consists of two transmit/receive Range Cell Resolution 1.2 km stations, each with a linear sixteen-element, Integration Time 5 min. phased-array receiver and a rectangular Azimuthal Resolution (3 Db down) 2° four-element transmitter (Table 1). The Radial Doppler Velocity Resolution 2 cm s-1 two stations were separated by a distance of ~50 km from each other (Figure 5) with a twenty-minute repetition cycle. In order one station being in Miami-Dade County’s to compare with the hourly in situ observa-

Crandon Park (CDN) and the other in a tions, significant wave heights H( s) derived State of Florida Botanical Preserve on the from the five-minute spectra were averaged northern end of Key Largo (NKL). This and sub-sampled hourly. separation provided a large area for current mapping, with the region of consistent 3.2 In Situ Measurements Made current vector retrievals extending well during Mini-Waves Experiment out over the Florida Straits, but limited the Three RD Instruments ADCPs (one region for which directional spectra could 600KHz and two 1200KHz), one Sontek be measured using two-site methods to a ADP (1500 KHz), one Nortek AWAC relatively small area (Wyatt et al., 2005). and two Tri-Axys buoys were all deployed However, the empirical method for wave simultaneously near the shelf break, within height observations does not necessarily require overlap between two stations, con- Figure 6 sequently a larger area was available over a)Typical echo-Doppler spectrum (solid line) as observed by WERA deployed in Southeast Florida. dB scale which Hs could be computed (Figure 5). Echo-Doppler spectra (Figure 6) have normalized by peak of backscattered spectrum. Bragg peaks for 16.045 MHz are shown as straight vertical lines. Weighting function as derived by Barrick (1977) shown as a gray solid line. b). Spectrum normalized been archived at each radar station at by weighting function. Second-order regions used for wave heights shown in gray boxes. Only values in twenty-minute intervals since June 2004. more energetic half-space were used in wave calculations. The number of retrieved spectra varied with SNR, with, on average, 1000-2000 independent wave height and current observations being extracted from each data set. To provide the required in situ calibration of the WERA-derived wave heights and to validate directional spec- tra measurements, a multi-institutional (University of Miami, University of South Carolina, Georgia Institute of Technology, Savannah Campus) joint experiment was conducted along the Southeast Florida shelf from March-May, 2005. At the time of the study, WERA actively transmitted and received signals for 5 minutes (1,024 samples) successively from each site, with

Fall 2008 Volume 42, Number 3 73 the radar domain (Figure 7), from Year ed in section 2.2.1 (also see Work, 2008), To determine the correct scaling for Days (YD) 75 to 145 as part of the Mini- while the ADP provided directional wave the WERA observations of wave height Waves experiment in 2005. Deployment data based on the PUV method (Sobey the radar results were compared with the locations were controlled by the local and Hughes, 1999). Overall, the ADCP wave heights from the WADP. It should bathymetry and currents, and instrument uses data equivalent of a twelve-sensor be noted that ADP was chosen as the depth rating, confining the measurements array from which directional wave char- initial calibration instrument, because it to water depths less than 20 m. Hourly acteristics are derived, making the method was deployed a week earlier than the other wave statistics were estimated by the in situ more suitable than the PUV approach for instruments and captured data during a sensors based on twenty-minute burst sam- measurement of short period waves, and storm, providing a reasonable range of data pling. Intercomparisons between the RDI provides better directional resolving power, for the calibrations. ADCP and the Tri-Axys were carried out although divergence of the acoustic beams The resulting best linear fits between and the results are similar to those present- in the upper part of the water column still the CDN and NKL WERA datasets and imposes an upper limit on the observable the ADP, had slopes of 1.46 and 1.66 frequencies. Wave data from the Tri-Axys respectively (see Figure 8), with the radar- Figure 7 buoys (see Work, 2008) were confined to derived wave heights being systematically Expanded view of in situ measurement locations frequencies 0.03 to 0.64 Hz due to noise smaller than those from the in situ measure- (boxed region in Figure 6) for Mini-Waves Cal-Val limitations, although this is not a severe ments. The empirical parameters (see Ra- experiment. WADP – 1500 KHz Sontek ADP with constraint at the site. mos, 2005) were then adjusted to provide waves package, WADCP – 1200 KHz RDI waves the best fit between the WERA derived ADCP. AWAC – Nortek current meter. TAB-N/S – Tri-Axys directional wave buoys, North and South. 3.3 Calibration of the HF Radar for results and the in situ observations of Hs. Depth contours shown with depths in meters. ADP Wave Estimates Using the Empirical The individual time-series of wave and ADCP were current-only systems. Approach height from each WERA station (1682 and Wave heights derived from the radar 1836 cells for CDN and NKL, respectively) observations from cells located within 50 were then compared to the ADP-derived km of each transmit/receive station were wave height values obtained during the initially computed using an empirical ap- calibration period (YD 78-100). The linear proach developed by Ramos (2005). This correlation coefficients from both CDN method is based on the ratio of the 2nd order and NKL were highest within ± 45º of to 1st order scattering energy as derived the radar boresight (normal to the receive by Barrick (1977), with the processing array). The correlations for locations close optimized for use with the WERA system. to the boresight were higher than for cells The proportionality between the energy located closer to the ADP (Figure 9). The ratio and surface wave height must be poor correlations at large angles from the empirically determined for each particular boresight may result from contamination radar system. of the results by sidelobe contributions

Figure 8

Calibration series for wave height from WERA stations CDN (a) and NKL (b) (best correlated) with Hs from WADP.

74 Marine Technology Society Journal Figure 9 There was a significant effect of shal- low structure on the local wave field Linear correlation coefficient for Hs extracted from single site observations using method of Ramos (2005) for each individual WERA cell and WADP Hs observations over calibration period YD 79-100, 2005. Gray scale as indicated by localized low correlations indicates correlation coefficient. White ■ marks the position of the WADP within each radar measurement (see Figure 9). These are shown over shal- domain. Left: CDN. White ▲ is location of Fowey Rocks CMAN station. Right: NKL White ▲ is the location low shelf regions near Fowey Rocks Light, of Pacific Reef, white  marks Triumph Reef, White  marks Carysfort Light. Carysfort Light, Triumph Reef, and Pacific Reef. Fowey Rocks Light and Carysfort Light have large structures that could con- tribute to significant zero-Doppler returns, however Pacific Reef and Triumph Reef have no such large structure and the origin of the low correlation is likely due to local depth-limitations on the wave field. For the 16 MHz frequency used in this experiment, the Bragg scattering wavelength is 9.35 m (Table 1), implying that the system re- sponds better to surface waves of that same wavelength. While these waves will begin to deviate from deepwater propagation to the Doppler spectra or by the limited be the effect of the relatively large angle be- speeds and begin to exhibit non-linearities capacity of the approach to resolve waves tween wave propagation and radar direction in water depths less than 5 m, the simple propagating at large angles to the radar’s (Wyatt, 2002). In the present configuration empirical technique used here (Figure 6) central radial. The beam-forming of a the large angles from the radar are also as- does not necessarily require invoking linear phased-array system such as the WERA is sociated with large angles to the offshore theory. Further comparisons with in situ weighted to suppress sidelobe returns, but direction from which the greatest wave en- observations in very shallow water will its effectiveness reduces with increasing ergy is incident. This argument fails for the be required to determine if the weighting angle. Although this is not typically a prob- CDN site, where the angle between the wave function or scaling coefficients require lem for current measurements (Haus et al., propagation and the radar is small. A third adjustment to extract Hs in depth-limited 2004), the results indicate that it might be possibility might be the cumulative effect of wave conditions. significant for wave measurements. both sources of error with different relative The NKL correlations suggest that this influences for each site. This is not clear at 3.4 Validation of the HF Radar might not be a problem for this site, as they present and requires further investigation. In The empirically derived, calibrated are relatively high to the Southeast despite either case, limiting the wave observations WERA wave heights were compared to the large off-boresight angles. Alternatively, to narrower beams than those employed for data collected by the Tri-Axys buoy (TAB- the poor correlations at large angles might current observations is required. N) moored at the 19 m depth contour, 800 m offshore of the bottom-mounted Figure 10 current meters (located at the 9 m con- Validation time series for wave height estimates from (i) the Tri-axys buoy TAB-N (black), (ii) WERA observa- tour; Figure 7). During the validation tions from NKL (blue), and (iii) WERA observations from CDN (red). period (YD 100 -145) of the experiment there were persistent light winds and low wave heights, with the maximum hourly

Hs recorded at buoy TAB-N being 1.35 m. The time series revealed a qualitatively good agreement between the radar and the buoys (Figure 10). The scatter diagrams indicate an underestimation of the radar- derived wave heights relative to those reported by the TAB-N buoy (Figure 11a). It was also found that the ADP-derived wave estimates (which were used for the calibration of the radar) are lower than those reported by the buoy, especially

Fall 2008 Volume 42, Number 3 75 Figure 11 a) Scatter diagrams of WERA-derived wave heights along boresight, both sites vs. TAB-N (black squares = CDN, red circles = NKL), y = x line also shown. b) Average of data from both sites vs. TAB-N, RMS difference 0.2 m. Best-fit (--), y = x (-) and adjusted best fit by Sontek ADP vs. TAB-N regression from 1st part of experiment.

during peak wave activity (not shown 3.5 Directional Spectral first 22% of the record which is close to the here). Correcting the radar measurements Observations from the Radar 0.05 m mean difference between the two to adjust for this difference between the Systems buoys (Figure 12a). Periods of wave height calibration (ADP) and validation (buoy) Two weeks of the HF radar data were agreement coincided with times where instruments eliminated much of the offset processed using algorithms developed by the highest wave energy was found near (Figure 11b). Seaview Sensing (www.seaviewsensing. 0.2 Hz, although the direction estimates The correlations between wave heights com) to derive directional spectra over the differed by ~30º (see Figure 12b and c). computed for CDN and NKL and those region having sufficient overlap between The direction of the wave energy at high observed at all the in situ sensors were sig- the two stations (Figure 5). The WERA frequencies was coincident with the wind nificantly lower than for the comparisons derived spectra were compared with spec- direction as expected. between in situ observations (see Table 2). tral estimates from the Tri-Axys buoys This initial agreement was encouraging, Averaging the estimates from the two radar over the period YD 95-101 (Figure 12). but there were large errors (of the order of sites significantly improved the correlation 100%) for the last half of the observation Comparison of Hs derived from integrating between the radar and the buoys and acous- the directional spectra revealed that there period (Figure 12a). The reason for this tic sensors. This suggests that the variability was initially good agreement between the pronounced degradation in the quality of may be in part due to directional bias in the WERA and the buoys. The mean differ- the observations could not be conclusively radar observation relative to the wind direc- ence in significant wave height between determined. At this juncture it should be tion as suggested by Wyatt (2002). WERA and TAB-N was 0.06 m over the noted that these discrepancies might be caused by the limited length (5 min) of data Table 2 used for the HF radar spectra, compared Linear correlations for Hs for all wave measurement platforms during the mini-waves experiment. Radar to the 20 min datasets used for the in situ observations refer to the best correlated of the radar cells. sensors. It is hypothesized that increasing TAB/PW TAB/GV SNTK RDI/N RDI/S WERA/ WERA/ WERA/ the sampling period will suppress noise that NKL CDN Avg/ may be contributing to the radar overesti- TAB/PW 1 mates, but the appropriate duration should TAB/GV 0.97 1 be evaluated experimentally. SNTK 0.93 0.93 1 In addition to the limited sampling period, another problem inherent in HF RDI/N 0.95 0.94 0.93 1 radar observations is that they are sensitive RDI/S 0.95 0.95 0.92 0.95 1 to radio frequency interference (RFI). For WERA/NKL 0.80 0.81 0.83 0.80 0.80 1 the WERA surface current observations WERA/CDN 0.71 0.70 0.72 0.71 0.70 0.59 1 this can limit sampling range, but rarely WERA/Avg 0.85 0.85 0.86 0.85 0.84 0.91 0.87 1 affects data quality because of the high SNR

76 Marine Technology Society Journal Figure 12

(a) Hourly (black) wave heights derived from WERA directional spectra, and Tri-Axys spectra at TAB-N (green dots) and TAB-S (blue dots). (b) TAB-N spectra on April 6, 2005 (YD 96) at 3:50 GMT. (c) WERA spectra at cell 2457 on April 6, 2005 at 3:45 GMT.

of the first-order Bragg scattering peaks. surface wave observations and forecasts gram (CDIP) collaboration was considered The wave measurements appear to be more among the research, coastal management exemplary as a productive collaboration sensitive to RFI as expected because of the and emergency management communi- that works very well for the state of Cali- lower SNR of the second-order returns ties, SEACOOS sponsored a one-day fornia. It was identified that some of the used for wave observations. This interfer- meeting that took place at the University ingredients that have made this a success are ence typically varies from day to night and of South Carolina, Columbia, SC (Janu- the support by the State of California and on longer time scales as well. ary 26, 2005) to discuss the issue of Wave by congressional representatives, as well as Following these observations, the sam- Measurements and Forecasting within the appreciation of the role of waves in control- pling strategy for the WERA data collection southeastern U.S. The meeting explored ling coastal erosion. The latter has been the was changed in 2006 to ten-minute blocks the needs of the various federal partners and result of long–term research on the West per station to provide improved spectral the possible contributions that a regional Coast. It should be emphasized that the noise suppression. Additionally, two refine- coastal ocean observing system can make engagement of users’ groups in California ments to the sampling strategy were imple- in this area. Federal affiliate representatives (including state government and private mented to suppress RFI. The first is that the from NOAA/NDBC (Dr. Teng) and the citizens) has been instrumental in the suc- system now operates in a listen-before-talk U.S. Army Corps of Engineers ERDC/ cess of the cooperative program. In con- mode, where the available bandwidth is CHL (Dr. Jensen) were present. The in- trast, there has not been such a coordinated scanned immediately before transmission terest of the National Weather Service in effort within the Southeast to promote the to determine the frequency range with wave prediction was also recognized at the need for wave measurements. the lowest RFI and the frequency sweep time, but no representative was present During this meeting the important confined to that region. A second step is at that meeting. Representatives from the role of NOAA/NDBC in providing wave to use an RFI suppression approach where University of South Florida, University of measurements and data portals was noted, the RFI is measured during each transmis- Miami, Skidaway Institute of Oceanogra- as well as the role of NOAA/NWS in the sion interval and then removed from the phy, University of South Carolina, Univer- area of wind and wave forecasting (through observed Doppler spectra. Additional in sity of North Carolina at Wilmington and the National Centers for Environmental situ measurements are required to test the University of North Carolina at Chapel Prediction, NCEP). Given the mandate efficacy of these changes, particularly in Hill were present. for the local NWS offices to provide the regard to the directional wave spectral ob- During the meeting all federal affiliates public with forecasts of surf and servations which are considerably more sen- present expressed support for a coordinated conditions, they also represent important sitive to short sampling windows and RFI wave measurement program. Presentations users of local and regional wave informa- than measurements of mean currents. described the USACE activities in the area tion. In the area of numerical modeling, it of wave measurements and their interest in was recognized that the state of California a coordinated wave measurement initia- has been the pioneer in wave forecasting, 4. Coordination Efforts tive for the East Coast in general and the especially in the area of swell waves. In In addition to the data collection and Southeast in particular, especially since southern California, the REF/DIF wave research and development efforts in the area the latter is an area frequently influenced transformation model (Kirby et al., 2002) of wave measurements, efforts were made by the development of tropical storms and is employed for predicting swell conditions, to coordinate a wave measuring program hurricanes. The USACE/NDBC/Univ. of whereas in Jacksonville, Florida, a local wave for the region. Given the wide interest in California Coastal Data Information Pro- forecasting system was developed by NRL

Fall 2008 Volume 42, Number 3 77 (based on the SWAN model; see Rogers et provide measurements in the open ocean University of South Carolina and USACE al., 2007 ) with funds from NWS (through (deep water waves) and the large-scale (Field Research Facility at Duck, NC) and the Coastal Storms Initiative). It was clear wave forecast. Regional associations could funded by NOAA/COTS in 2007. This during the meeting that currently there is contribute key local, nearshore measure- project has a strong focus on wave forecast- not a consensus on model use and protocol ments important to particular constituents ing for the Carolinas and includes elements and that these will need to be established. and also run high resolution models that of nearshore assessment for use by Furthermore the wave forecasting issue are integrated with the larger scale mod- the local NOAA/NWS WFOs. is constrained by the resolution required by els. The roles that regional associations various users of these products. Although could play in this system were identified no explicit recommendations came out of as follows: (i) Evaluations of existing wave 5. Concluding Remarks the meeting, the need for some organiza- measurement technology. (ii) Continua- The waves initiative within SEACOOS tion and further action was established. tion of ongoing measurement programs led to the establishment of directional However, the following points were made: and develop new programs in order to wave measurement stations at several new (1) Wave measurement and prediction start building climatologies for a variety of locations, evaluation of the employed constitute an important link between coastal areas that can be used later to evalu- systems, and the transfer of a number of offshore atmospheric and oceanographic ate model performance. (iii) Development these systems or data streams to new pro- conditions and the nearshore. Waves im- of a unifying data product capitalizing on grams. These data have been invaluable in pact a wide variety of coastal users ranging the California CDIP experience. (iv) Selec- providing nearshore wave climatologies from recreational users (e.g., surfers, beach tion of two or three areas for test beds in for the deployment sites. Furthermore, users), to the local municipality (with the southeastern U.S. that have different NOAA/NWS have been utilizing these interests and responsibilities relating to wind/wave forcing and differing bathym- data for guidance on nearshore forecast- coastal erosion, permitting for develop- etry to be used for extensive measurements ing activities. The data from a variety of ment, and beach safety) to the state and and wave model evaluation. From such wave stations are fed to NOAAPORT by federal levels (including interests such as an exercise, a model or a suite of models the individual partners and then this is navigation, fisheries, search and rescue). might emerge that are suitable for use in distributed to the NWS local WFOs via the (2) NOAA/NDBC has invaluable experi- routine operational forecasting mode. (v) Advanced Weather Interactive Processing ence in maintaining offshore sites for the Development of procedures, protocols and System (AWIPS). measurement of waves and it might be technologies required to make these wave Expertise was established in using the organization best suited for providing forecasting systems transferable to different ADCPs for wave measurements. Particular wave information that facilitates data as- areas throughout the Southeast with the success was the creation of data display similation and verification of large-scale ultimate goal (a ten-year plan) to have the results for ADCP directional wave meas- domain numerical models. (3) Nowcasting whole Southeast covered. urements. These were widely disseminated and forecasting of nearshore wave condi- Although these items were brought up to various partners within SEACOOS but tions require resolution that is dictated by for discussion in the July, 2005 SEACOOS also were shared with RCOOS within the the gradient in offshore wave and wind workshop (Voulgaris and Nelson, 2005), it region. Comparisons of ADCP- and buoy- patterns and by bathymetry and coastline was deemed as too large of an effort to be obtained wave parameters have shown that morphology. It is likely that a number undertaken by SEACOOS alone. Partner- the two systems are in very good agreement of high-resolution wave transformation ing with and leveraging funds from federal and different sensors can be integrated in a models would be needed for different areas. organizations was proposed as a strategy wave observation program. These models could be maintained and run by which SEACOOS might achieve the The infrastructure of stations estab- by regional associations, obtaining their above-mentioned goals. Nevertheless it is lished by SEACOOS has been utilized by boundary conditions from the larger scale the feeling of the authors that the coordina- other entities for data dissemination. An federal backbone modeling and measure- tion efforts and discussions initiated with example is the use of the Springmaid Pier ment activities. the 2005 SEACOOS waves meeting con- wave station infrastructure in South Caro- The meeting concluded that regional tributed to the development of the Integrat- lina for the collection and transmission of associations and the federal partners could ed Ocean Observing System Operational Dissolved data in the nearshore for greatly benefit by collaboratively working Wave Observation Plan (NOAA/IOOS, the South Carolina Department of Ocean toward development of a high resolution 2008) by NOAA/NDBC and USACE. In and Coastal Resources Management (see: wave forecasting system. A straw-man addition, following the SEACOOS experi- http://carocoops.org/longbay/hypoxia/in- proposal for such a partnership was out- ence, the Carolinas RCOOS initiative was dex.html). lined, based on spatial resolution criteria. developed with partners from the Univer- The research and development efforts in It was proposed that federal affiliates could sity of North Carolina at Wilmington, the the area of using radars for estimating wave

78 Marine Technology Society Journal conditions with a high spatial resolution has in the type of data provided by these types References established the region as the pioneer in these of measurements. Barrick, D.E. 1977. Extraction of wave activities. The SEACOOS-sponsored radar parameters from measured HF radar sea-echo experiment demonstrated that the WERA Doppler spectra, Radio Sci. 12(3):415-424. technology is promising for providing wave Acknowledgments height estimates and led to identification of Funding for SEACOOS and the activi- Bishop , C.T. and M.A. Donelan. 1987. areas that need further research. This activ- ties described in this contribution was pro- Measuring waves with pressure transducers. ity fostered collaboration between different vided by the Office of Naval Research un- Coast Eng. 11:309-328. institutions and has provided the basis for der award N00014-02-1-0972. The award Cole, R. and R. Weisberg, 2008. Coastal the development of a new waves program was managed by the University of North Ocean Observing Systems Going Wireless. Sea under the auspices of the Regional Coastal Carolina, Office of the President. The Technol. 47(4):10-13. Ocean Observing Associations. contribution of Stephanie Obley, Laura Although the initial experiment pro- Azevedo and Jeffery P. Morin in the devel- Graber, H.C, E.A. Terray, M.A. Donelan, vided some encouraging results, its short opment and maintenance of the nearshore W.M. Drennan, J.C. Van Leer and D.B. Pe- duration failed to capture a variety of wave wave stations has been invaluable. Diving ters. 2000. ASIS-a new air-sea interaction spar conditions that could enable us to answer support for the SC nearshore stations was buoy: design and performance at sea. J Atmos questions such as: (i) what is the minimum provided by the South Carolina Institute of Oceanic Technol. 17(5):708-720. (threshold) wave conditions that WERA Archaeology and Anthropology, Maritime Gurgel, K.-W., G. Antonischki, H.-H. Essen responds to; (ii) how does wave non-linear- Research Division led by Dr. C. Amer. The and T. Schlick, 1999. Wellen�������������������� Radar (WERA): ity affect WERA-derived wave estimates; management of Springmaid Beach Resort a new ground-wave HF radar for ocean remote (iv) what are appropriate algorithms for Hotel and Conference Center is greatly sensing. Coast Eng. 37:219-234. wave parameter quality assurance and acknowledged for providing Internet con- control that would enable WERA-derived nection and access to the Springmaid Pier. Haus, B.K. 2007. Surface current effects on wave characteristics to be easily integrated Charleston County Park and Recreation the fetch limited growth of wave energy. within the existing operational protocols of Commissions are thanked for permitting J Geophys Res. 112, C03003, doi: NOAA/NDBC. Furthermore, long-range us to use the Folly Beach Edwin S. Taylor 10.1029/2006JC003924. HF Radars installed on the middle-Atlantic Fishing Pier, SC. Haus, B.K., R. Ramos, H.C. Graber, L.K. Bight (Shay et al., this volume), jointly oper- The crew of the R/V Savannah and Shay and Z.R. Hallock. 2006. Remote Ob- ated by the Skidaway Institute of Oceanog- other personnel from the Skidaway Insti- servation of the Spatial Variability of Surface raphy and the University of South Carolina, tute of Oceanography, particularly Trent Waves Interacting with an Estuarine Outflow. that are currently undergoing upgrading to Moore and Michael Richter, were critical IEEE J Oceanic Eng. 31(4):835-849. provide offshore wave information require to the success of the wave buoy program that the performance of radar systems for in Georgia. Haus, B. K., J. D. Wang, J. Martinez-Pedraja wave measurements be quantified in terms The following people helped provide and N. Smith. 2004. Southeast Florida Shelf of radar frequency of operation. sites for the radars: Dr. Renate Skinner, Circulation and Volume Exchange, Observa- Several additional lessons were learned. Mr. Jim Duquesnel, and Mr. Eric Kiefer tions of km-scale variability. Estuar Coast Shelf The development of nearshore wave prod- from Florida DEP; Mr. Kevin Kirwan S. 59(2):277-294. ucts was limited to the assets and method- and Mr. Ernest Lynk from Miami-Dade Heron, M.L., P.E. Dexter and B.T. McGann. ologies selected by each partner. Although County Parks and Recreation. Jorge Mar- 1985. Parameters of the air-sea interface by some of these were partially dictated by tinez, Thomas Cook and Mei Wang were high-frequency ground-wave Doppler radar. existing infrastructure, this resulted in a critical to this research through their work Aust. J Mar Freshw Res. 36:655-670. lack of uniformity amongst the different with the WERA measurement group at products. Some of the uniformity was the University of Miami. Jerome Fiechter IPPC, 2007a. Climate Change 2007: The provided through the channeling of wave and Mike Rebozo assisted with the field Physical Science Basis. Contribution of Work- data through NDBC. measurements. Richard Curry of Biscayne ing Group I to the Fourth Assessment Report Data management of wave products National Park assisted with the current of the Intergovernmental Panel on Climate and dissemination was never integrated meter deployments. Change. Solomon. Cambridge Press, 1009 pp. as a full product. Partially because of Dr. Chung-Chu Teng and NOAA/ the complexity of the collected data and NDBC Stennis Space Center are ac- partially because of lack of funds, the data knowledged for contributing one of the processing and management was left to Tri-Axys buoys for the HF radar evalua- individual participants with the expertise tion experiment.

Fall 2008 Volume 42, Number 3 79 IPPC, 2007b. Climate Change 2007: Impacts, Strong, B., B. Brumley, E.A. Terray and G.W. Adaptation and Vulnerability. Working Group Stone, 2000. The performance of ADCP-de- II contribution to the Fourth Assessment Re- rived directional wave spectra and compari- port of the IPCC Intergovernmental Panel on son with other independent measurements. Climate Change. Cambridge Press, 986 pp. OCEANS 2000 MTS/IEEE Conference and Exhibition. 2:1195-1203. Jahnke, R.A., J.R. Nelson, M.E. Richards, C.Y. Robertson, A.M.F. Rao and D.B. Voulgaris, G. and J.R. Nelson, 2005. Regional Jahnke. 2008. Benthic Primary Productiv- Directional Waves/Sediment Transport Appli- ity on the Georgia Mid-continental Shelf: cations The role of SEACOOS for the SE U.S. Benthic Flux Measurements and High-Reso- Summary White Paper from the SEACOOS lution, Continuous In Situ PAR Records. Spring Meeting (July 2005, Jacksonville, FL) J Geophys Res Oceans. 113, C08022, do: 3 pp. plus Appendices (Available at: http:// 10.1029/2008JC004745. seacoos.org/teams/sediment-waves/sediment- waves-white-paper.pdf) Kirby, J.T., R.A. Dalrymple and F. Shi. 2002. Combined Refraction/Diffraction Model Voulgaris, G., M.P. Wilkin and M.B. Collins, REF/DIF 1 Version 3.0, Documentation and 1995. The in situ passive acoustic measure- User’s Manual, Research Report NO. CACR- ment of shingle movement under waves and 02-02, Department of Civil and Environ- currents: instrument (TOSCA) development mental Engineering, University of Delaware, and preliminary results. Cont Shelf Res. Newark, DE, 166 pp. 15(10):1195-1211.

Nelson, J.R. and R. Weisberg, 2008. In situ Work, P.A., 2008. Nearshore directional wave observations and satellite remote sensing in measurements by surface following buoy and SEACOOS: Program development and les- acoustic Doppler current profiler. Ocean Eng. sons learned. MTS Journal, this issue. 35(8-9):727-737. NOAA/IOOS, 2008. An Integrated Ocean Wyatt, L.R. S.P. Thompson and R. R. Burton, Observing System Operational Wave Observa- 1999. Evaluation of high frequency radar wave tion Plan, March 2008, Funded by the NOAA measurement. Coast Eng. 37:259-282. IOOS Program and developed by NOAA/ NDBC and USACE, with support from ACT. Wyatt, L.R. 2002. An evaluation of wave March 2008. 27 pp plus Appendices. (Available parameters measured using a single HF radar at: http://doc.aoos.org/nfra/Wave%20Plan%2 system. Can J Remote Sens. 28(2):205-218. 0Report%20low%20res.pdf). Wyatt, L.R., G. Liakhovetski, H. Graber and Ramos, R.J., 2005 2-D analysis of wave energy B.K. Haus. 2005. Factors affecting the accu- evolution using wavelet transforms. PhD The- racy of Showex HF radar wave measurements. sis, University of Miami, December 2005. J Atmos Ocean Tech. 22:847-859.

Rogers. W.E., J.M. Kaihatu, L. Hsu, R.E. Jensen, J.D. Gykes and K.T. Holland. 2006. Forecasting and hindcasting waves with the SWAN model in the Southern California Bight. Coast Eng. 54(1):1-15. Shay, L.K., H. Seim, D. Savidge, R. Styles and R.H. Weisberg, 2008. High Frequency Radar Observing Systems in SEACOS, MTS Journal, this issue.

Sobey, J and S.A. Hughes, 1999. A locally nonlinear interpretation of PUV measure- ments. Coast Eng. 36(1):17-36.

80 Marine Technology Society Journal