COMPS, SEACOOS and Near-shore Waves Rick Cole
Neil Trenaman
University of South Florida, St. Petersburg, FL
RD Instruments, San Diego, CA
Robert Weisberg
Kevin Amundsen
University of South Florida, St. Petersburg, FL
Benthos, Inc., North Falmouth, MA
University of South Florida 140 Seventh Ave. South St. Petersburg, FL 33701 USA
[email protected] compared with other continental shelf regions, little is known about the west Florida shelf circulation. There is an immediate need to make more effective use of existing resources and new technologies to provide a more timely detection and prediction of the coastal environmental conditions from changes in weather and sea state to changes in habitat and living marine resources. USF’s Ocean Circulation Group and Ocean Monitoring and Prediction Group maintain the “Coastal Ocean Monitoring and Prediction System” (COMPS, Fig. 1), a real-time monitoring program along the WFS. Data from this array will lead to a better understanding of the responses of the inner shelf to forcing by waves, tides, winds, seasonal heating and cooling, fresh water inflows, and the interactions with the adjacent deep Gulf of Mexico.
The Ocean Circulation Group at the Abstract: University of South Florida (USF), College of Marine Science maintains a real-time monitoring program on the eastern Gulf of Mexico’s west Florida shelf as part of the Coastal Ocean Monitoring and Prediction System (COMPS) and the Southeast Atlantic Coastal Ocean Observing System (SEACOOS). Offshore is an array of surface buoys and bottom mounted instruments for surface meteorology, currents, temperature, and salinity from the Big Bend to the Dry Tortugas. Recognizing the need for wave measurements near-shore, testing began on both instrumentation and telemetry techniques for achieving real time waves capabilities. This paper documents a collaborative effort between USF, RD Instruments (RDI), and Benthos, Inc. in producing a system for real time directional wave data acquisition culminating in a successful 17-day deployment spanning a period of calm winds followed by a hurricane. Reported on is a system consisting of an RDI acoustic Doppler current profiler (ADCP™) with Waves Technology, and NEMO, RDI’s new Real Time Waves Processing Module, linked to the surface by Benthos, Inc. acoustic modems, and telemetered to USF by FreeWave radio for real time Web posting.
I. INTRODUCTION A large number of local state and federal agencies are responsible for different but overlapping aspects of the coastal ocean environment. This requires environmental observations, analysis, research, forecasts and management. Understanding the circulation on the west Florida shelf (WFS) in the eastern Gulf of Mexico is important for a variety of reasons. Knowledge of the currents and sea level variation has application to coastal erosion, recreational and commercial navigation, search and rescue operations, and the tracing of the movement of hazardous material spills. Interactions between the coastal waters and the offshore loop current in the Gulf of Mexico affects the distribution of biological and chemical properties that affect fisheries and red tides (Harmful Algal Blooms).
Fig. 1: The COMPS Domain As coastal monitoring projects continue to evolve new programs will come online. In May of 2000, the Chief of Naval Research, the Administrator of NOAA, and the President of the Consortium for Ocean Research and Education announced the formation of OCEAN.US,
A concern to all coastal residents and tourists is the storm surge response of sea level to severe weather events.
1
an organization dedicated to the formation of an integrated and sustainable ocean observation system. The Southeastern Atlantic Coastal Ocean Observing System (SEACOOS, Fig. 2) was formed and is to be one of the regional systems ringing the U.S. to form the coastal component of the Integrated Ocean Observing System (IOOS). A collaborative university partnership that collects, manages, and disseminates integrated regional ocean observations and information products for the coasts of North Carolina, South Carolina, Georgia, and Florida. The Office of Naval Research (ONR) has provided funding for this effort. The vision for this ocean observation system requires that observing systems scattered across the country cooperate to collect and disseminate data and data products to serve the critical and expanding needs of environmental protection, public health, industry, education, research, and recreation.
I. NEAR-SHORE WAVES The offshore component COMPS utilizes acoustic Doppler current profilers (ADCP) for measuring ocean velocity profiles. These are employed either surface mounted on buoy bridles in a downward looking mode or bottom mounted in trawl resistant racks along with other sensors. Profiling ranges in these west Florida shelf (WFS) applications have varied between the 10 m isobath and the shelf slope. Data acquisition may be accomplished using a variety of methods, including internal recording, telemetry by radio, satellite, or cables, and most recently by acoustic telemetry. All methods have advantages and disadvantages, and in certain applications it is first necessary to transfer data from the bottom to the surface before broadcasting to a home site. Cabling is straight forward, but subject to damage by trawling and anchoring. Acoustic modems alleviate this cabling issue, but are limited to the amount of data that can be transferred in a given time. Thus while the acoustic transmission of ADCP velocity and other oceanographic data is well developed, the transmission of sufficient information for full wave directional spectra from an ADCP has been difficult. The OCG, in collaboration with RD Instruments (San Diego, CA) and Benthos, Inc. (North Falmouth, MA) set out to test an integrated telemetry system for the purpose of real time acquisition and telemetry of velocity profile and directional wave data. Using an RDI Workhorse ADCP with Waves Technology, linked to a home station at USF by a combination of Benthos, Inc. Telesonar acoustic modems and FreeWave radios, we initiated in-water testing in February of 2002. After solving a series of problems over the course of three deployments, including the recognition that the waves data need to be preprocessed prior to acoustic transmission, we arrived at a successful systems design in August 2004.
Fig. 2: The SEACOOS Domain Realizing that surface wave measurements are important issues for coastal erosion and navigational hazards, SEACOOS has initiated a program for near-shore wave measurements and is in the process of developing a wave forecasting program. Nowcasting and forecasting of near-shore wave conditions will close the loop between offshore components and sites along the coastline within the IOOS. Numerical modeling will play an important roll in the open ocean, providing wave and swell information on large scale wave forecasting. Federal agencies (NOAANDBC) have experience in maintaining offshore sites providing data that have input on larger scale domain numerical models. A plan is now being structured to set procedures and develop technologies for making waveforecasting systems available throughout the entire southeastern US. Discussed here is the OCG’s first attempt at achieving “real-time” wave measurements locally along the near-shore environment of the WFS.
II. SYSTEMS COMPONENTS A. RDI Workhorse 600kHz ADCP with patented Waves technology: The RDI Waves ADCP uses velocity information from the three to five bins nearest the surface along with pressure and echo ranging from the surface to compute the wave height and directional spectra. Three independent calculations of wave height derive from orbital velocity, pressure and surface location. Directional spectra derive from a Maximum Likelihood Inversion of the individual beam data. Waves sampling bursts consisted of 2400 pings at 2 Hz. Current sampling bursts consisted of 60 pings over 10 min.
2
F. Mooring Systems, Inc., fiberglass ADCP rack (trawl resistant bottom mount, TRBM):
B. RDI NEMO Real Time Waves Processing Module: Designed specifically for RDI Waves users, NEMO is a self-contained devise, linked directly to the Workhorse Waves-ADCP, that processes real-time current and waves data at the source to provide condensed data packets suitable for transmission to the surface or shore via an acoustic modem or a hard wire link. Output hourly were one 20 min wave burst with wave height, period, and direction along with water level and six velocity profiles.
This lightweight, small boat deployable mount was used to secure all sub-sea instrumentation. G. Figure 1 provides a systems schematic.
C. Benthos, Inc., Telesonar Acoustic Modems: An ATM-885 remote modem with an AT-408 omni directional transducer serves as the underwater component of the acoustic communication link (interfaced with the Waves-ADCP). This system acoustically transmits (or receives) data from a local ATM-880 surface modem (housed in a NEMA weather resistant enclosure with connectors for one remote transducer, an RS-232 serial port for PC interface, and external DC power) using an AT-408 omni directional dunking transducer located near the surface. Both modems require 21-volts of input power and operate at the middle frequency (MF) range of 16-21 kHz.
Fig.1: Wireless Waves Configuration: Bottom Mount, Surface Mount and Repeater Station on the Beach III. DEPLOYMENTS From the initial attempt in 2002 to the present successful application a total of four test deployments were made. The first was off the seawall at the USF campus, the second near Benthos in Buzzards Bay, and the third and fourth offshore in St. Pete Beach at Pass-A-Grille channel. In all cases the out of water telemetry links were by FreeWave RF modems with a PC interface to collect data in real-time via the RDI Waves processing software Waves Monitor (WavesMon). WavesMon is designed for real time data collection and processing of the wave data gathered by an ADCP, allowing for the evaluation of waves and currents at a glance. The program also creates waves records (.wvs) that can be displayed by the WavesView software that searches time series for interesting wave energy events. In three deployments velocity data were received and processed via WavesMon, but the wave bursts information were problematic. This led to the development of the NEMO for preprocessing waves data in situ at the ADCP location and our successful fourth deployment.
D. FreeWave Wireless Data Transceiver: Once at the surface a line-of-site, FreeWave radio based system is used to telemeter these data back to USF. A multipoint network consisting of the offshore master broadcasts to a slave receiver (or multiple receivers), either directly or through a repeater station. The received messages are entered through a serial port to the host computer. This application used a master, a repeater and one slave receiver, as follows: 1) Master station located offshore on United States Coast Guard aid to navigation Pass-A-Grille Channel Light 2 (LLNR 1400), St. Pete Beach, FL, 27º 40’ 36”N, 082º 46’ 00”W. 2) Repeater station located Restaurant, St. Pete Beach, FL
at
the
Hurricane
3) Slave station located at the CMS, USF, St. Petersburg, FL
The Pass-A-Grille site, located approximately one and a quarter miles west of the St. Pete Beach, consisted of two acoustically linked components (Fig. 1): 1) an in situ measurement rack situated on the approximate 7 m isobath some 110 m from 2) a fixed channel marker platform on which we mounted the FreeWave radio master. With the assistance of Neil Trenaman and Egil Rassmussen of RDI and Adam Lipper from Benthos all instrumentation was readied in the OCG lab and deployed from the R/V SUBCHASER. Figure 2 shows the instrumentation mounted on the bottom rack.
E. PC Interface at USF: A laptop PC was configured to receive hourly data, which was then routed to the COMPS webpage for internal viewing over the 17-day deployment interval.
3
modem will ask the distant modem to re-send the subpacket, which contained the error. Using these techniques all data being sent from the distant modem are able to make the journey through the water unchanged and ready for delivery to a standard computer “serial port” to the end user. The receiving modem calculates a checksum for each data sub-packet and if there is an exact match to the transmitted checksum, chances are extremely good that the data has arrived error-free. If not, the modem has the ability to analyze the probability of different outcomes and will make one or more “guesses,” often arriving at the correct solution and “recovering” the lost data without the need for it being re-sent. V. INSTRUMENT SET-UP The standard waves setup calls for 20-minute, five bin data sets. Required are three bins with the extra two used for redundancy. A 20-minute sample interval provides sufficient record length for wind wave burst analysis. At 2 Hz this results in 2400 samples per burst. Increasing sampling rate and duration increases power consumption. Along with the waves bursts velocity profiles were sampled every 10-minutes with 60 pings. Table 1 summarizes the setup that was used.
Fig. 2: R. Cole, N. Trenaman and J. Law Prepare the TRBM for Deployment IV. ACOUSTICS Many factors must be taken into consideration before deploying an acoustic telemetry system. Operating depth, required communication range, multipath, ambient noise and environmental conditions are all critical. Multipath, the most restricting affect both on baud rate and reliability is the result of sea-surface and seafloor reflections, reflections from objects near the receiving modem, thermal gradients and water turbulence. The wave-modems deployments were of prototype designs and corrected on a “trial and error” basis. Subsequent to the initial three attempts RDI redesigned the waves data processing and data transfer, while Benthos incorporated software and hardware modem modifications. The modems start with the user’s binary digital data, which can be representative of any type of information that can be encoded digitally. For example, any data that can be passed through the Internet can be handled by the modems. Whereas typical computers use copper wires or radio signals as the data transmission medium, acoustic modems use water, and this greatly reduces the effective data transfer rates and distances.
Table 1. ADCP Configuration VI. THE NEMO WAVES MODULE Options for real-time telemetry of oceanographic data are often limited by the available bandwidth of the telemetry technology or the cost of the service and in some cases, both. Since real time ADCP Waves processing requires the wave data to be sampled continuously at 2Hz for 20 minutes if data are lost during this interval the processing is compromised. The NEMO allows the Workhorse ADCP Waves and currents data to be buffered over the sample interval and then quickly processed and condensed for outputting a real-time wave and current profiling data packet. The effect of the NEMO is to condense the data from an original 150Kbytes to a 150byte data string suitable for acoustic transmission in reportready format.
The original digital data is sent through the water in “packets” and “subpackets.” A modem discovering data packets addressed to itself will decode the data and will then assemble the incoming packets into a seamless data stream for use by the host computer or destination device. The ocean environment is often a noisy place and signal interference can occur from the effects of wind-driven waves, passing ships, and snapping shrimp, to name but a few. If the modem receives a large data packet with a small amount of missing data due to noise interference, the
4
VII. RESULTS
Significant Wave Height ( Hs) 2.25 2
Along with a successful test deployment of this real time system we were in the water during the passage of Hurricane Francis just to the north of the Pass-A-Grille site (Fig. 3). Both the survival of the system and the associated data added to the success of the test deployment. Figure 4 shows the time series for Hs, Tp, Dp and Water Level during and after the passage of Hurricane Francis.
1.75 1.5 1.25 1 0.75 0.5 0.25 0 1
24
47
70
93
116
139
162
185
2 08
23 1
254
277
300
3 23
34 6
369
Ho ur s
Pe ak Wave Per iod (Tp) 10 9 8 7 6 5
During the height of the storm data telemetry was lost for nearly twenty-four hours due to large sea state (2 m waves), bubbles, and suspended sediment. The data did return acoustically and in real-time after the storm passed. Figure 5 shows the acoustic intensity contour and profile plots during the peak of Francis. The distinct local maxima seen in the profile data is used to detect the air-sea interface. The RDI Waves ADCP Array uses this information to calculate the wave spectra for the surface detection technique. The bottom panel shows the acoustic intensity contour plot and profile plot during the peak of Hurricane Francis. The contour plot and the corresponding profile plot indicate that the acoustic intensity signal is fully saturated, note there is no discernable acoustic peak in the signal. The most probable causes for this condition are high levels of suspended sediment and/or entrained air bubbles throughout the water column.
4 3 2 1 0 1
24
47
70
93
116
139
162
185
208
231
254
277
300
323
346
369
254
277
300
32 3
3 46
369
277 28
301 313
Ho ur s
Peak Wave Dire ction (Dp) 36 0 33 0 30 0 270 24 0 210 18 0 150 12 0 90 60 30 0 1
24
47
70
93
116
139
162
18 5
2 08
231
Ho ur s
Water Level 6.6 6.4 6.2 6 5.8 5.6 5.4 5.2 5 1
13
25
37
49
61
73
85
97 109 12 1 133 145
157 169
181 193 20
217 22
241 25 2 6
32
33
3 4 361 37
Ho ur s
Fig. 4: Individual Breakout Plots of Hs, Tp, Dp, and WL
Fig. 5: Acoustic Intensity Contour and Profile Plots. VIII. DATA DISSEMINATION Once data was received from the field, hourly, it was archived in the USF-COMPS base station computer. These data were then processed using decoder scripts written in Perl programming language. The wave observations of significant wave height, peak period, direction, water level and time series graphics were made available via the Internet internally, not public, for testing purposes, (http://comps.marine.usf.edu/rdi). The graphics scripts were written in IDL (Interactive Data Language) programming language. Figure 6 shows the test page menu with hourly observations, station position, pictures, tables and plots with the past 24-hour and 5-day
Fig. 3: Hurricane Francis Crosses Florida, September 5th and 6th, 2004 on a WNW track
5
Hs in meters
observations and a plots link of data parameters: Hs in meters, Tp in seconds, Dp in degrees and WL in meters. The peak to the right on the wave height time series (top) is the arrival and departure of Hurricane Francis on September 5th and 6th, the second of four major hurricanes to hit the state of Florida in the summer of 2004. Once this system is installed permanently it will become an added product on the COMPS web page.
8/27/04
29
31
9/02/04
04
06
08
9/10/04
Fig. 7: WavesMon Data vs. Transmitted NEMO Data Showing a Good Over-Plot of Both Data Sets During the Peak of Hurricane Francis X. CONCLUSION The OCG-RDI-Benthos waves test demonstrates that waves data can be successfully transmitted acoustically through the water column and then to shore for post processing. Benthos modems integrate to the Waves ADCP and the NEMO very well. Not only was the system test a success but the data set proved beneficial in checking the three methods RDI uses to record waves. An example was provided on the effects on acoustics (ADCPs & Modems) when the instruments are exposed to highly turbid and aerated water conditions (such as occurred during the passage of Hurricane Francis). Along with the temporary acoustic telemetry failure due to saturation by bubbles and sediments the ability of the ADCP to utilize surface tracking as the primary method for calculating the wave height was compromised. The ADCP backscatter plot (Fig. 5) shows the water column “saturated” with suspended sediment, so that the intensity signal can no longer detect the surface. This strengthens the argument for the need of more than a single sampling technique to ascertain wave height when using ADCPs. The spectral plots showing RDI’s three independent sampling clearly shows that the surface tracking beam technique does not work during the storm conditions (note both vertical or slant beams would fail as the RSSI signal is fully saturated and the surface detection peak in the signal is not discernable – it’s a flat line). However, having both the orbital velocity and the pressure sensor options allows the RDI ADCP to continue to measure the wave field even under extreme conditions, whereas the surface tracking technique works well during calmer conditions.
Figure 6: COMPS Waves Internal Web Menu and Time Series Plot Menu of Hs, Tp, Dp and WL
IX. DATA COMPARISON Comparison of the real-time transmitted data from the NEMO and the data processed using Wavesmon after the recovery of the instrument is shown in Fig. 7. This plot shows the increase in Hs during the onset, peak and decay of Hurricane Francis. The data processed by the NEMO and the post-processed Wavesmon data are in agreement except for about a 12-hour period during the peak of the storm. During the peak of the storm, data was not received from the site as shown by the missing NEMO data. The cause for the missing data was attributed to a drop out of the acoustic modem signal due to a saturation of the water column by entrained air bubbles and also high-suspended sediments.
6
ACKNOWLEDGEMENTS Support was derived partially from the Office of Naval Research, Grant # N00014-02-1-0972 for the Southeast Atlantic Coastal Ocean Observing System (SEACOOS), administered by the UNC under Task order # 3-12110-10. The authors would like to thank the following for their participation, efforts and patience. USF College of Marine Science: Jason Law, Jeff Scudder, Vembu Subramanian, Jeff Donovan, Cliff Merz, Patrick Smith, Dennis Mayer, Dr. Mark Luther/OMPL, Lauren Wetzel, Sherryl Gilbert, Brian Donahue and R/V PRICE, Tom Peacock and R/V SUBCHASER, All CMS Divers who have participated in Waves-ADCP deployments and recoveries. RDI: Brandon Strong, Darryl Symonds, Egil Rasmussen, Harry Maxfield, Earl Childress and Tony Ellis. Benthos, Inc.: Dale Green, Ken Scussel, Doug McGowen, Adam Lipper, Celeste Harmon, Tom deGroot (previously with Benthos), Jack Crosby (previously with Benthos) and Rick Babicz (previously with Benthos). The Hurricane Restaurant, St. Pete Beach, FL: Bruno and Rick Faulkenstein for allowing the use of their rooftop for installation of the FreeWave radio repeater station.
7
