Georges Bank GLOBEC Shipboard ADCP Data Collection and Analysis

INVESTIGATORS:

Julio Candela
Clark Lab
WHOI
Woods Hole, MA 02543
e-mail: julio@mar.whoi.edu

Charles Flagg
Bldg 318
Brookhaven National Laboratory
Upton, NY 11973
e-mail: flagg@bnl.gov

GRANT PERIOD: October 1, 1993 to September 30, 1997

STATEMENT OF OBJECTIVES:

The GLOBal ocean ECosystem dynamics (GLOBEC) program seeks "to develop a fundamental understanding of the mechanisms that determine how the abundances of key marine animal populations vary in space and time" with the assumption that "the physical environment is a major contributor to the patterns of abundance and production of marine animals" (Peterson and Powell, 1991). GLOBEC focuses on planktonic forms of marine animals because these "animals are the most tightly coupled to the physics of the fluid medium" and thus would be most affected by changes in the physical environment as a result of modifications due to global warming. The first GLOBEC study is taking place in the northwest Atlantic Ocean over the Georges Bank. This location was chosen because the Bank is on a faunal boundary between polar and subtropical species and thus would be particularly sensitive to global fluctuations and because there already have been shifts in the faunal balance on the Bank. The total program, expected to last seven or eight years, began in early 1993 with the first of five field years beginning in 1994.

This component of the program deals with the collection and analysis of ship-board acoustic Doppler current profiler (ADCP) data taken from the ships that will transit Georges Bank during the mooring, broad-scale, and process surveys to the area. Julio Candela of the Woods Hole Oceanographic Institution (WHOI) and Charles Flagg of Brookhaven National Laboratory (BNL) jointly are responsible for the collection, processing, and analyzing the ADCP data collected on the cruises.

STATEMENT OF WORK:

There are three types of cruises to Georges Bank during the project. First, there are extensive broad-scale biological and hydrographic survey cruises covering the entire bank. So far, there have been nine of these cruises between May 1994 and July 1995, each lasting eleven to twelve days. A second series of cruises, the eleven process cruises began in January, 1995 and will continue through August, 1995. These cruises, whose goals are tied to specific regions or processes, cover a more restricted area than do the broad-scale cruises. The third cruise type is dedicated to the deployment, maintenance and retrieval of moored instrumentation. During all cruises to the area, ADCP data is and has been routinely collected. The ADCP data will be used to describe the broad-scale current and planktonic distributions and their spatial, annual, and seasonal variability. The results will be central to understanding the physical and biological interactions on the Bank which are at the heart of the GLOBEC goals.

In order to perform this task, the shipboard data acquisition as well as the shore based data processing, archiving, synthesis with other data sets, and analysis must be addressed. So far there have been four ships used in the Georges Bank GLOBEC program, the RV Albatross IV, RV Endeavor, RV Seward Johnson, and the RV Iselin. The ADCPs type and arrangement differ on each of the ships; the Seward Johnson had three different ADCPs all by itself. Some of the ADCPs are the older "narrow-band" variety, while some are the newer "broad-band" ADCPs. This diversity of equipment has meant that we were unable to use the same data collection and processing methodology throughout as we had initially planned. For those ships equipped with the narrow-band ADCPs we were able to use, for the most part, RD Instruments' venerable ADCP DAS, version 2.48 (Endeavor and the last cruise on the Seward Johnson). On some of the cruises with narrow-band ADCPs (Iselin and Seward Johnson) and for all those where the broad-band ADCPs were used (Iselin, Seward Johnson, and Albatross), data collection was controlled by RD Instrument's TRANSECT program.

There have been and will be too many cruises to the area for us to man all of them. Instead, arrangements were made with the ships' technicians and the scientists participating in each cruise to run the ADCPs and to monitor their performance. In addition, Julio Candela participated in the RV/Albatross IV cruise 9404 when the newly installed 300kHz BB-ADCP was fully operational for the first time. To alleviate the technicians' work-load and to assure a uniformity of ADCP data, the initial intent was to make use of a program developed at BNL which provides the ability to collect ADCP data without operator supervision using predetermined protocols tailored to different regions of the Bank. This program, however, is only compatible with the older ADCP DAS 2.48 which meant that those vessels equipped with broad-band ADCPs or those which were no longer able to run the ADCP DAS with its older architecture, needed to use TRANSECT. The automatic data collection system, referred to as the AUTOADCP programs, provides the ability to change the configuration of the ADCP and its data collection methodology based upon the ship's location. The program also provides a watchdog timer in case the system stalls as well as automatic logging of ADCP status and backup of the data files. These capabilities make the program ideal for the GLOBEC program because it frees us from having to be on every cruise. Using this program we have been able to setup the system before the ships leave the dock and collect the data upon the ships' return.

Processing of the ADCP data is being done at both WHOI and BNL. Julio Candela at WHOI has primary responsibility for the ADCP current data while Charles Flagg at BNL will be responsible for the planktonic biomass estimation. Aside from this division, all the TRANSECT data files are being initially processed at WHOI while all the PINGDATA files produced by the ADCP DAS are being initially processed at BNL.

Data Processing and Analysis
Data collected through the TRANSECT program mainly consist of processed (or averaged) binary files. In some cruises, when the hard disk capacity of the PC's computer running the ADCP permits, raw data binary files have also been stored along with the corresponding raw navigation data files. The collection of raw data files permits a finer sampling of the ADCP data compared to the usual half-minute ensemble averages used in the processed data files, in the event that future research of specific processes requires it. However to this point, the processing of TRANSECT files has only dealt with the averaged data files. After a binary/ascii translation of the averaged data files using a PC computer running the TRANSECT program, the ascii files are reformatted to be readable on a MATLAB environment and processed on a SUN 41 workstation. A series of MATLAB routines have been written at Woods Hole to do standard quality checks on the data, re-averaging it to suitable time and space intervals and performing a detiding of the current field using the methods described by Candela et al. (1992).

The binary pingdata files (collected with the DAS program) from each cruise are being processed on a Silicon Graphics Indy workstation using the ship-board ADCP processing programs developed by the ADCP group at the University of Hawaii. These programs consist of two parts, a set of processing programs and a data base system called CODAS (Common Oceanographic Data Access System) written in C, and a series of MATLAB analysis routines for post-processing and plotting the ADCP velocity and backscatter data. The CODAS routines were augmented to handle the backscatter intensity data in keeping with the methods developed by Flagg and Smith (1989) and Flagg, et al. (1994).

The ADCP data from the pingdata files were loaded into separate CODAS data bases for each cruise. The data were then edited based upon the percent good, a bottom search criterion, the error velocity, the variance of the vertical velocity, and the second vertical difference in the vertical and horizontal velocity components. Collectively, these parameters flag interference from the bottom, from instrumentation suspended from the ship, from large aggregations of fish or scatterers and other, possibly anomalous, behavior. Each velocity and backscatter profile was examined visually and in light of the flagging parameters. Calibration of the ADPC's orientation and velocity scaling was accomplished using data collected during each cruise from relatively straight, constant speed portions of the cruise. The calibration runs from the cruises yielded similar values. Once the ADCP profiles were edited and the velocities scaled and rotated, the navigation data were combined with reference layer velocities to produce smoothed versions of the ship's track and absolute reference layer velocities. The smoothed reference layer velocities and positions were then loaded into the data base so that the velocity data in the CODAS data base is ready to be extracted for plotting and analysis.

The currents on Georges Bank are dominated by semidiurnal tides with characteristic magnitudes of more than 50 cm/s. Therefore, detiding current data on the Bank is unavoidable to reveal the subinertial current field responsible for the net transport of suspended organisms. The 10 to 12 days of data characteristically available from each of the broad scale cruises, permits resolving the main semidurnal current field centered at the M2 tidal frequency. Distinguishing more of semidiurnal constituents as well as some of the diurnal constituents will require sampling the current field for a longer period of time. For example, resolving the main semidiurnal constituents M2 and S2 from a time series at a fixed point in space requires at least 15 days of observations. For the case of data like the shipboard ADCP in which the sampling is scattered over time and space it is not simple to determined exactly how much data is required to confidently resolve the M2 and S2 current fields over the Bank. However, it is reasonable to expect that this resolution will be reached when there are enough current observations to imply more than 15 days of data over most of the Bank. The detiding techniques being used (Candela et al., 1992) permit the combination of data from several cruises as well as information from moored instruments and/or currents deduced from drifters and should enable the resolution of the principal tidal constituent current field structure over the Bank. Although only individual cruises have been analyzed so far, one of the objectives of this work is to performed a comprehensive analysis from all the information available to the GLOBEC program and provide reliable procedures for detiding current observation taken anywhere on the Bank.

Backscatter intensity data, in terms of digital counts from each of the ADCP's four transducers, were loaded into the CODAS data base directly from the pingdata files along with the velocity data. Calibrating the intensities from each of the beams and combining them to produce a single calibrated intensity profile for each ensemble is based upon an expanded version of the method outlined in RD Instruments' technical bulletin ADCP-90-04 issued in December 1994. The processing produces a single calibrated and range corrected backscatter intensity profile for each ensemble. The calibrated intensities will be compared with the zooplankton net tow results when that data is available to develop an empirical relation between backscatter intensity and zooplankton biomass.

SUMMARY OF KEY FINDINGS:

At this stage in the data collection process significant findings about the currents and zooplankton distributions have yet to emerge. The effort to date has been to assure that the data is being collected correctly and that progress is made on processing the voluminous ADCP data set (approximately 60 MBytes of raw data to date). As of this writing, 25 GLOBEC cruises to Georges Bank have taken place, 22 where ADCP data were collected, with five more cruises scheduled. The ADCP data from 20 cruises are currently in hand, eight cruises worth at WHOI and the rest at BNL.

As the data have been processed and preliminary figures produced, the results will be made available through the World Wide Web either directly to WHOI and Brookhaven or through the GLOBEC home page.

References

Candela, J., R.C. Beardsley and R. Limeburner, 1992. Separation of tidal and subinertial currents in ship-mounted acoustic Doppler current profiler observations. J. of Geophys. Res., 97(C1), 769-788.

Flagg, C.N. and S.L. Smith, 1989. On the use of the RDI acoustic Doppler current profiler to measure zooplankton abundance. Deep Sea Res. 36(3): 455-474.

Flagg, C.N., C.D. Wirick, and S.L. Smith, 1994. The interaction of phytoplankton, zooplankton, and currents from 15 months of continuous data in the Mid-Atlantic Bight. Deep Sea Research II, 41, 411-436.