Time Series Measurements of Chlorophyll Fluorescence in the Oceanic Bottom Boundary Layer with a Multisensor Fiber-Optic Fluorometer

AUGUST 1997 D'SA ET AL. 889

EURICO J. D'SA,STEVEN E. LOHRENZ, AND VERNON L. ASPER Institute for Marine Sciences, University of Southern Mississippi, Stennis Space Center, Mississippi

ROY A. WALTERS Ocean Optics, Dunedin, Florida (Manuscript received 4 June 1996, in ®nal form 31 December 1996)

ABSTRACT An in situ multisensor ®ber-optic ¯uorometer (MFF) has been developed to acquire long-term chlorophyll ¯uorescence measurements in the oceanic bottom boundary layer to characterize the ®nescale pigment structure at vertical spatial scales comparable to physical measurements. The eight ¯uorescence sensors of the MFF are composed of dual optical ®bers of varying lengths (1.5±8 m), with the ®ber ends oriented at 30Њ to each other and enclosed by a small light baf¯e. Strobe excitation blue light is passed through one of each pair of optical ®bers and stimulated chlorophyll ¯uorescence is carried back to a photomultiplier. Two sets of four ¯uorescence sensors assigned to high- and low-sensitivity photomultiplier detectors enable chlorophyll a measurements in two ranges, 0±50 mg mϪ3 and0±200mgmϪ3, respectively. Aspects of the design of the ®ber-optic sensor are described that were intended to optimize detection of ¯uorescence signals and minimize interference by ambient light. The ®ber-optic sensor outputs were stable with minimal instrument drift during long-term ®eld operations, and measurements were not affected by turbidity and ambient light. A vertical array of ®ber-optic ¯uorescence sensors supported on a tripod has been deployed at coastal sites for up to seven weeks and chlorophyll ¯uorescence was obtained with suf®ciently high vertical spatial and temporal resolution.

1. Introduction ¯uorometers have been used to obtain information on Flow conditions and suspended sediment in the bot- the temporal variability of phytoplankton biomass tom boundary layer of coastal waters have been widely (Whitledge and Wirick 1983; Wirick 1994) and provide studied to determine sediment movements and transport a means for obtaining chlorophyll a ¯uorescence in con- (Trowbridge and Nowell 1994). However, little infor- junction with other bio-optical and physical measure- mation about particulate organic matter distribution and ments (Dickey et al. 1991; Marra 1994). Measurements transport in the near-bottom waters has been reported of the photosynthetic pigment ®eld in the bottom bound- (Thomsen et al. 1994), primarily due to dif®culties in ary layer at resolutions comparable to the ¯ow ®eld the measurement process. Chlorophyll and its degra- (Williams et al. 1987) have been constrained primarily dation products ¯uoresce and can be used to trace that by size and long-term operating capability of ¯uorom- fraction of particulate organic matter associated with eters. Requirements for such a ¯uorometer are an in- phytoplankton (Lorenzen 1966). Measurements have strument with multiple and small sensors to minimize been made using sensitive pro®ling ¯uorometers to de- ¯ow disturbance and enable their location to be close ®ne the small vertical spatial scales of variability of the to one another and to other sensors such as current me- chlorophyll ¯uorescence in the upper ocean (Derenbach ters, measurement capability at high spatial resolution, et al. 1979; Astheimer and Haardt 1984; Cowles et al. and long-term in situ operation. A ¯uorometer design 1989). Such observations have led to increasing aware- using optical ®bers met the above requirements since ness that biological distributions and rate processes in the ®bers have high bandwidth, transmit signals ef®- the upper ocean may be characterized by small length- ciently from remote locations, and do not cause any scale heterogeneity on the order of 10Ϫ2 m. Moored appreciable distortion of water ¯ow. This paper describes the development and performance evaluation of a multisensor ®ber-optic ¯uorometer that was designed to overcome the limitations of Corresponding author address: Dr. Eurico J. D'Sa, World Precision Instruments, Inc., International Trade Center, 175 Sarasota Center conventional ¯uorometers in obtaining estimates of the Blvd., Sarasota, FL 34240-9258. photosynthetic pigment ®eld in the bottom boundary E-mail: jerico@¯net.com layer. The principle of operation and description of the

᭧1997 American Meteorological Society

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC 890 JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY VOLUME 14

FIG. 2. Basic optical con®guration showing the excitation and detection optics of a single sensor of the ®ber-optic ¯uorometer.

FIG. 1. Schematic diagram of the optics and electronics of the multisensor ®ber-optic ¯uorometer. The excitation optics uses four ducts the ¯uorescence emission light from the sampling xenon strobe lamps, blue interference ®lters, and eight optical ®bers volume through a combination of a red interference ®lter that carry the excitation light to the eight sensors. The detection optics (Rolyn Optics) with a peak transmission at 676 nm and consists of two photomultipliers (PMT), red ®lters, and eight optical a 665-nm high-pass ®lter (Rolyn Optics RG665) into a ®bers that conduct the ¯uorescence emission light from the sensors to the two photomultipliers. photomultiplier. The electronics of the MFF is controlled by a Tattle- tale microcomputer (Onset Computers model 7) inter- instrument is presented along with laboratory and ®eld faced with a 500-kHz analog-to-digital converter (Ocean experiments to evaluate the instrument. The ¯uorometer Optics, ADC500) and a 120-Mbyte hard disk for data has been successfully deployed at shelf sites in the Gulf storage. The software was written in C programming of Mexico and the United States east coast in conjunc- language that allowed control of instrument operations. tion with physical characterizations of the bottom These included the sampling interval and signal aver- boundary layer. aging of the strobe stimulated ¯uorescence, background signal subtraction (i.e., detector signal in absence of strobe excitation), data storage, switching to standby 2. The multisensor ®ber-optic ¯uorometer low-power mode between samplings, and communica- a. The general description of the MFF system tion with a personal computer. A 12-V battery pack consisting of alkaline batteries supplied power to the The MFF system used four excitation xenon strobe system. The four xenon strobe lamps in the MFF are lamps (model LS-128A; EG&G Electro-optics) and two triggered in sequence by the microcomputer and typi- photomultiplier detectors (Hamamatsu HC-125), there- cally have a total ¯ash duration of about 100 ␮s and a by providing a capability for independent ¯uorescence FWHM duration of only about 8.2 ␮s. As the peak measurements at eight locations using variable length ¯uorescence intensity corresponds to the peak in strobe dual ®ber-optic sensors (Fig. 1). Excitation light from intensity, the signals from the photomultiplier and the each strobe was conveyed through two excitation optical reference photodetector were digitized by a 12-bit an- ®bers to two sensors. The two excitation optical ®bers alog-to-digital converter at a high sampling frequency were epoxied into a single SMA ®ber-optic connector of 500 kHz for 200 ␮s (100 samples per strobe ¯ash) and then inserted into an SMA mating adapter on the and the signal stored on hard disk. A typical ¯uores- strobe housing. An additional ®ber inserted into the cence value obtained during a laboratory experiment or strobe housing conducted the strobe light to a photo- ®eld operation consisted of an averaged response from detector. Detection ®bers from the sensors conveyed 10 strobe ¯ashes that was recorded along with the time ¯uorescence emission light to the two photomultipliers stamp and the sensor number. An additional strobe ¯ash provided with a housing to hold a pair of red ®lters and was used for monitoring strobe intensity by the pho- mating adapters for the detection ®bers. The basic op- todetector. A serial port on the MFF enabled program- tical con®guration of the excitation and detection system ming and computer control for setting up of the MFF of a single ¯uorescence sensor of the MFF is shown in operating parameters and downloading of stored data Fig. 2. The broadband excitation light is focused through ®les for further processing. a lens and a blue interference ®lter (Omega Optical The pressure case was constructed of T-6061 alu- 455DF) with maximum transmission at 455 nm and minum and was 1 m long and 0.2 m in diameter. A bandwidth of 70 nm at FWHM (full width at one-half double O-ring end cap was ®tted with 16 optical pe- of maximum) into a 600-␮m core diameter optical ®ber netrators (Ocean Optics, Inc.) or feedthroughs to couple (3M FP-600-UHT). The blue excitation light is con- the internal±external optics of the ¯uorometer. The pres- ducted by the ®ber to the sensor sampling volume. The sure rated optical feedthroughs had SMA adapters at detection system consists of a detection ®ber that con- either end to couple to SMA connectors and a short

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC AUGUST 1997 D'SA ET AL. 891 length internal 600-␮m ®ber that allowed light to be coupled from ®bers mated on either end of the feedthrough. One set of eight optical penetrators was used to couple excitation blue light from the four strobes to the sensors, and another set of eight penetrators coupled the ¯uorescence emission light from the sensors to the two photomultipliers located inside the pressure casing. Index matching liquid (Norland Products, Inc.) was used to enhance the ef®ciency of coupling of signals between the ®bers and the optical feedthroughs. A four-pin un- derwater connector permitted serial communication between an external computer and the Tattletale microcomputer, thus enabling instrument control without FIG. 3. Inter®ber angle versus relative ¯uorescence for the 600- opening and closing the pressure case. The sensor ®ber ␮m-diameter (solid line) and 1000-␮m-diameter (dashed line) optical lengths varied from 1.5 to 8 m, allowing remote ¯uo- ®bers. rescence measurements to be made at varying distances from the instrument. surements with plastic-core ®bers indicated high levels of ®ber background emission (data not shown). Though b. The dual-®ber sensor the 1000-␮m-diameter ®ber gave slightly higher ¯uores- Key considerations for the ®ber-optic sensor design cence signal (Fig. 3), the 600-␮m core ®ber was selected were the geometrical con®guration of the ®bers, ®ber because of ¯exibility and low minimum bend radius. The type, and a light baf¯e to minimize ambient light in- ®ber sensor ends were separated by a minimal distance terference while simultaneously providing minimum of 0.1 mm or less. The 30Њ inter®ber angle between the impedance to water ¯ow. Selection of the optical ®bers excitation and detection ®bers, beside giving a maximal was based on cost, performance, and ease of use in the ¯uorescence response, also provided easy handling of the ocean environment. The light-detection ef®ciency of an ®ber-optic sensors during ®eld operations of the MFF. optical ®ber depends on the numerical aperture and the The silica-clad silica core ®bers had a kevlar-reinforced ®ber core diameter: the active sampling volume increas- polyvinyl chloride (PVC) jacket. At the sensor end, the es with larger values of both the parameters (Plaza et PVC jacket and kevlar were removed to optimize the al. 1986). However, large core optical ®bers are sus- alignments between the excitation and detection ®bers. ceptible to breakage and dif®cult to handle during ®eld operations in the oceanic environment. A trade-off is c. The light baf¯e therefore involved in the selection of the core size for the optical ®bers. The optical sensor for the ¯uorometer In situ chlorophyll ¯uorescence measurements can be is made up of a pair of polished ends of the excitation affected by interference from ambient red light, es- and emission ®bers that are oriented at an angle to each pecially during daytime operations (Hitchcock et al. other. The active sampling volume is a cone de®ned by 1989). Initial ®eld trials of the ®ber-optic ¯uorometer the numerical aperture, the ®ber core diameter, and the con®rmed the need for a light baf¯e to prevent ambient sensor geometry. The overlap between the excitation light from affecting chlorophyll ¯uorescence measure- and collection cones is the region of sampling volume ments (data not shown). Though the ¯uorescence emis- from which ¯uorescence is measured. sion signal due to chlorophyll in the sampling volume Experiments were conducted to select ®ber type and could be separated from the ambient background signal the best geometrical con®guration for the ®ber-optic ¯uo- by background subtraction or other synchronous detec- rescence sensor, with the parameters of interest being the tion techniques, the net effect of high background signal inter®ber angle, the distance between the bare polished due to ambient light containing photons in the red wave ®ber ends, and the optical ®ber-core diameter. The ¯uo- band was a reduction in the ¯uorescence dynamic range rescence response for different con®gurations was eval- measured by the instrument. This problem was ad- uated in a solution of coproporphyrin (Aldrich Chemical, dressed by construction and testing of a series of sensors Inc.), a substance with ¯uorescence properties similar to with different light baf¯e designs (D'Sa et al. 1994). chlorophyll. The use of this solution enabled ¯uorescence The selected design of the light baf¯e (Fig. 4) provided response of the optical ®bers to be evaluated without the for minimal ambient light interference and only mod- variability due to chemical degradation of chlorophyll. erate impedance to water ¯ow. The baf¯e was con- Fluorescence intensity was maximal for a 30Њ inter®ber structed out of PVC and composed of three cylindrical angle for the 600- and 1000-␮m-diameter silica-core op- blocks of 9 cm in diameter and 2.5 cm thick. The ®ber tical ®bers (Fig. 3). In comparison, experiments with the holder block had two holes oriented at a 30Њ angle. 200- and 400-␮m silica-core ®bers gave lower ¯uores- Excitation and emission ®bers with polished ends were cence response than the 600-␮m silica ®ber, while mea- secured in the holes with epoxy. A center block was an

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC 892 JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY VOLUME 14

FIG. 4. Light baf¯e design of the dual ®ber-optic ¯uorescence sensor with (a) three-dimensional view of the top block that has two optical ®bers epoxied to the block at 30Њ angle to each other, and (b) section view of baf¯e sensor assembly consisting of three blocks that are held by three bolts and separated by spacers. Arrows indicate ¯ow through the baf¯e. FIG. 5. Fluorescence response of the dual ®ber-optic sensors ref- erenced to the ¯uorescence of a ``standard'' Turner Designs ¯uorometer 10-005R for laboratory-grown phytoplankton culture Nanno- chloris performed in August 1995. Solid and hollow circles denote the experiment after a 7-week ®eld deployment (23 August±10 Oc- inverted V-shaped 2.5-cm-thick hollow cylinder. The tober 1995) off New Jersey. bottom block had two sharply inclined holes to allow settling particles and ¯uid to exit the baf¯e. The blocks were held together by bolts that were inserted in three reduced in size and made smaller than other miniature holes drilled at a 120Њ separation angle at the center of ¯uorometers. the V-grooves. The 1.25-cm spacers were placed on the bolts and between the blocks to separate them. The two 3. Performance evaluation of the multisensor inverted V-shaped 360Њ openings around the baf¯e al- ®ber-optic ¯uorometer lowed water to ¯ow through the baf¯e and the dual- ®ber sampling volume. A ¯ume experiment using a dye a. Stability and drift analysis (thymidine blue) indicated that at an approximate ¯ow In applications involving long-term unattended in- Ϫ1 speed of 15 cm s , the dye had a residence time of 1 strument operation, such as the bottom boundary layer Ϫ1 s, while at 25 cm s ¯ow speed, an approximate res- deployments of the multisensor ®ber-optic ¯uorometer, idence time of 0.5 s was recorded. These high ¯ow rates measurement stability is an important concern. The ¯uo- were only used to determine the water residence time rescence measured by the different sensors should re- within the sample volume and revealed the minimum ¯ect the chlorophyll concentration in the medium as impedance provided by the baf¯e to the water ¯owing determined by the calibration coef®cients (assuming through the sampling volume. Analysis of sensor ¯ow constant ¯uorescence yield). However, due to various characteristics and saturation excitation irradiance in- factors such as environmental stress on sensors and ®- dicated an increased phytoplankton ¯uorescence yield bers, ®ber degradation, biofouling, and strobe and elec- within the dual-®ber-optic sensor sampling volume tronic drift, the instrument response may vary. To mon- (D'Sa 1996). A laboratory experiment was conducted itor instrument stability, a number of laboratory cali- to test the effect of the baf¯e walls on the ¯uorescence brations were conducted on the eight-sensor ®ber-optic response of the dual ®ber sensor by enclosing the ®ber ¯uorometer using stirred laboratory-grown pure sensor in volumesÐ678, 381, 172 (baf¯e volume), and phytoplankton cultures and by referencing the ¯uores- 22 cm3Ðand monitoring its ¯uorescence response in a cence output of the baf¯ed ®ber-optic sensors to the diluted phytoplankton culture solution. A steady ¯uo- ¯uorescence of a ``standard'' Turner Designs ¯uorom- rescence signal of 581 counts was measured by the ¯uo- eter 10-005R, which was previously calibrated in ace- rescence sensor for the different baf¯e enclosures and tone-extracted chlorophyll. The long-term ¯uorescence indicated that the baf¯e, if necessary, could be further response of the dual ®ber-optic sensors (Fig. 5) were

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC AUGUST 1997 D'SA ET AL. 893

and the different sensitivities of the detectors. Pure phytoplankton cultures were used in the calibration of the MFF sensors using the dilution method. The chlorophyll concentration of the culture was determined ¯uoro- metrically (Parsons et al. 1984) using a Turner Designs ¯uorometer. There exists a high degree of correlation between the chlorophyll concentrations and in vivo ¯uorescence (r2 Ն 0.9 for most sensors). The measurement range of chlorophyll concentration for the sensors assigned to the high- and low-sensitivity photomultipliers are 0±50 mg mϪ3 and 0±200 mg mϪ3, with detection limits of 0.1 mg mϪ3 and 0.5 mg mϪ3, respectively. The lower ¯uorescence detection limits (equals 2͙ 2 times the standard deviation of the mean of the baseline signal) in ¯uorescence counts for the eight sensors of the MFF are shown in Table 1 and were obtained from seven measurements taken in arti®cial seawater during various calibration runs. We evaluated the possibility that the ¯uorescence signal was affected by drift in ¯uorometer components such as the strobe and electronics. The strobe intensity and photomultiplier background signal were recorded during a 7-week deployment of the MFF in the bottom bound-

FIG. 6. Chlorophyll a ¯uorescence calibration of the multisensor ary layer on the United States east coast and are shown ®ber-optic ¯uorometer conducted in laboratory-grown phytoplankton in Fig. 7. Analysis of peak strobe signal intensities and culture of Nannochloris using the dilution method in July 1994 before their background revealed stable operation of all four a 6-week ®eld deployment near Cape Hatteras. The chlorophyll con- strobes. The excitation light intensities for the four centration of the phytoplankton culture was determined ¯uoromet- rically with a Turner Designs ¯uorometer using the extraction method. strobes were about the same, but the relative strengths measured by the photodetector varied due to the posi- tioning of the optical ®bers carrying the strobe light in monitored prior to and following a 7-week ®eld de- relation to the photodetector and is the reason for the ployment in the United States east coast by comparing different average peak intensities recorded (Fig. 7a). the ¯uorescence response of the MFF sensors in labo- Monitoring of the strobe intensity allowed for the scal- ratory-grown phytoplankton culture with the ¯uores- ing of the ¯uorescence signal in case of drift in the cence response of the same culture on the Turner De- xenon strobe lamp during long-term deployments. Anal- signs ¯uorometer and indicated no degradation in sensor ysis of the time series photomultiplier background signal performance. Sensor s7 shows higher ¯uorescence re- (i.e., the photomultiplier signal in absence of the strobe sponse after the deployment and could have been due excitation light) during the deployment also indicates to adjustments made to the sensor and better light cou- minimal effects of detector and electronic drift in the pling between the optical feedthrough and the sensor signal (Fig. 7b). Background signal for the two pho- ®bers. Performance degradation in some sensors oc- tomultipliers during the whole deployment period varied curred mainly due to biofouling. To minimize biofouling by less than 10 counts out of 4096 and includes inter- effects, the antifouling agent TBT (tributyl tin) was ap- sampling variations (at 1-h interval) and drift (Table 2). plied to the inside of the baf¯e and around the ®ber- These measurements also indicate that the baf¯e was optic sensors. The antifouling agent was not applied to very ef®cient at minimizing any contributions of diel the ®ber faces as it would have affected the optical ambient light variations to the ¯uorescence signal. properties of the sensor. Changes in instrument temperature (due to the MFF Sensors have different responsitivities and calibration being exposed to different water masses during the slopes (Fig. 6) due to the characteristics of each sensor 7-week ®eld deployment) also did not appear to cause

TABLE 1. Lower ¯uorescence detection limit for the eight ®ber-optic sensors used in the MFF. Values in ¯uorescence counts for the mean, standard deviation, and detection limit are shown. Fluorescence counts Sensor 1 Sensor 2 Sensor 3 Sensor 4 Sensor 5 Sensor 6 Sensor 7 Sensor 8 Mean 51.0 29.85 25.85 30.14 41.71 93.85 40.0 44.85 Standard deviation 38.18 17.45 11.11 24.97 24.4 45.05 38.09 35.41 Detection limit 107.9 49.3 31.3 70.6 69.0 107.7 127.4 100.1

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC 894 JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY VOLUME 14

that were not related to chlorophyll ¯uorescence and approaches to minimize their contribution. The optical ®lter backscatter experiment was conducted to identify and minimize the coupling or leakage of the excitation light from the excitation to the detection system of the ¯uorometer. Transmission characteristics of broadband blue ®lters analyzed on a spectrophotometer showed the presence of a small bandpass in the red that could affect ¯uorescence measurements due to the leakage of red light from the excitation to the detection system. The ®ber-optic ¯uorometer was op- erated with the excitation and detection ®bers facing each other in a 180Њ opposing orientation, and the ®lter combination with the best optical isolation between the excitation and detection system of the ¯uorometer was determined from this experiment. Different ®lter con- ®gurations were evaluated for the detection and the excitation system. A blue interference ®lter on the excitation optics and a combination 665-nm high-pass ®lter and a red interference ®lter were selected for the detection optics. The high-pass ®lter was located between the interference ®lter and the detector, with the interference ®lter having its re¯ecting side facing the sensor ®bers. To evaluate the effect of different concentrations of FIG. 7. Time series plot of (a) four-strobe excitation peak intensity non¯uorescing particulate matter (PM), the response of signals and (b) of the two photomultiplier background signal of the the ®ber-optic sensors in waters of different diatoma- multisensor ®ber-optic ¯uorometer measured during a 7-week de- ceous earth concentrations was determined. Results of ployment in the United States east coast continental shelf bottom boundary layer. the experiment show that high turbidity levels do not affect the ¯uorescence signal of the MFF. The ¯uorescence output of a low- and high-sensitivity ®ber-optic any signi®cant drift in electronics. Table 2 summarizes sensor (Fig. 8a) in waters containing zero chlorophyll the overall performance of the four strobes and the two concentrations indicate no consistent trend with increasing PM concentrations of up to 1500 mg LϪ1. The ¯uo- PMTs used in the ¯uorometer. The standard deviations rescence values obtained in this experiment may not of the peak intensities of the four strobes are quite small correspond to those of the calibration (Fig. 6) since the in relation to the mean peak signal. MFF performance was optimized before the ®eld deployment. The transmittance of the same water samples b. Scattering and turbidity analysis containing PM was measured on a 1-cm pathlength spectrophotometer (Fig. 8b) and con®rms that back- The problem of optical backscatter arises because of scattering of excitation light does not contribute signif- two main factors: 1) the blue ®lter on the excitation icantly to observed signals. Effects of increasing PM optics and the red ®lter on the detection optics do not concentrations added to the medium containing phy- perfectly isolate photons outside their cutoff wave bands toplankton indicates no effect on the ¯uorescence signal and the possibility of strobe excitation light leakage into (Fig. 8c). the detection system exists for the dual-®ber-optic sensor and 2) high turbidity in the water column due to high concentration of particulate matter may affect ¯uo- 4. Measurement of time series chlorophyll rescence measurements, as well as enhance non¯uo- distribution in the bottom boundary layer rescence light backscattering. Experiments were con- The MFF was attached to a 5-m-tall tripod along with ducted to evaluate the magnitude of extraneous signals benthic acoustic stress sensors (BASS) (Williams et al.

TABLE 2. Statistical properties of the four-strobe peak intensities and the two photomultiplier background signals in the absence of strobe excitation light during a 7-week in situ operation of the multisensor ®ber-optic ¯uorometer. Strobe-1 Strobe-2 Strobe-3 Strobe-4 PMT-1 PMT-2 Mean 1202.7 923.9 2402.4 670.5 47.9 14.6 Standard deviation 19.6 14.8 31.0 11.57 1.61 1.82

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC AUGUST 1997 D'SA ET AL. 895

FIG. 9. Time series of photosynthetic pigment estimates at four levels in the bottom boundary layer at a 20-m site on the inner continental shelf near Cape Hattteras and off Duck, North Carolina (36Њ11.932ЈN, 75Њ42.25ЈW), measured from yearday 204 to 234 (23 July to 22 August 1994).

pigment concentrations encountered in the bottom boundary layer. Time series of estimated pigment concentrations were quite variable throughout the dataset and show a number of episodes with high pigment concentrations (Ͼ 100 mg mϪ3) that remained for periods of time varying from 0.5 days to approximately 2 days during the ®rst half of the deployment period (day 204 to day 218). These were recorded at all four levels in FIG. 8. (a) Response of the ®ber-optic sensors (numbers 2 and 6) the bottom boundary layer. For the duration of deploy- of the MSFF to different concentrations of particulate matter (PM) ment two distinct distribution patterns were discernible. measured using diatomaceous earth. (b) Percentage transmission of the same samples containing diatomaceous earth measured using a One was a high activity period at all levels in the bottom 1-cm-pathlength spectrophotometer. (c) Response of the low- and boundary layer prior to day 218, followed by a lower high-sensitivity ®ber-optic ¯uorescence sensors (number 1 and 5) activity at depths closer to the sea¯oor and some high to different concentrations of PM added to a medium containing pigment levels at the uppermost levels. Following day phytoplankton culture with a concentration of 118 mg mϪ3 of chlo- 218 and the occurrence of a frontal storm, a large signal rophyll a. was recorded by the upper two sensors that persisted for a number of days but was not detected by the lower 1987). The remote ®ber-optic sensors of the multisensor two sensors. In a subsequent paper (in preparation), we ¯uorometer were oriented facing the seabed and placed examine the relationship of observed variability to the at different heights above the bottom along one leg of physics and meteorology of the site. the tripod at depths corresponding to the BASS sensors. The MFF was programmed to acquire chlorophyll ¯uo- 5. Summary and conclusions rescence data from the remote ®ber sensors at 40-min intervals. Each ¯uorescence value was composed of an The multisensor ®ber-optic ¯uorometer provides a averaged response from 10 strobe ¯ashes for each sensor unique capability to obtain ¯uorescence time series at and was stored on the hard disk along with ancillary different depths simultaneously. Thus far, vertical res- data. Time series estimates of near-bottom pigment con- olution in moored applications of in vivo ¯uorometers centrations (Fig. 9) were obtained at a 20-m site near has been achieved through deployment of multiple in- Duck, North Carolina, at four discrete depths within 5 struments. The two photomultipliers and four strobes in m from bottom [at 4.44, 2.55, 1.2, and 0.24 m above the MFF allows for up to eight sensors on one instru- bottom (mab)] from 23 July (yearday 204) to 22 August ment. Addition of xenon strobe lamps or a photomul- 1994 (yearday 234). Pigment estimates were obtained tiplier in the MFF would accomodate more ¯uorescence from a predeployment calibration conducted with lab- sensors. The ®ber-optic system expands the effective- oratory-grown pure phytoplankton culture (green al- ness of individual strobe and detector packages by use gaeÐNannochloris) based on the assumption of a con- of multiple optical ®bers to allow increased vertical res- stant ¯uorescence yield. Chlorophyll measurements at olution, long-term unattended operations, and the design the four depths shown correspond to the high dynamic of small-sized ¯uorescence sensors allowing ¯exibility range MFF sensors that did not saturate at the high to place the ®ber-optic sensors in close proximity to one

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC 896 JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY VOLUME 14 another and other instruments (e.g., current meters), grant through a subcontract (MAB-95-6) to the Mid- with minimal interference to the current regime. The Atlantic Bight project. lower limit of spatial sampling is approximately 10 cm and is set by the closest distance the baf¯es can be REFERENCES placed next to each other. The lower temporal sampling Astheimer, H., and H. Haardt, 1984: Small-scale patchiness of the limit is 1.2 s and is set by the strobe ¯ash frequency chlorophyll ¯uorescence in the sea: Aspects of instrumentation, (10 Hz), the number of speci®ed strobe ¯ashes, and the data processing, and interpretation. Mar. Ecol. Prog. Ser., 15, 233±245. number of operating sensors. Cowles, T. J., J. N. Moum, R. A. Desiderio, and S. M. Angel, 1989: In examining the response of the ®nal prototype ¯uo- In situ monitoring of ocean chlorophyll via laser-induced ¯uorometer both in the laboratory and at sea, we achieved rescence backscattering through an optical ®ber. Appl. Opt., 28, the goals required for remote, unattended operation in 595±600. Derenbach, J. B., H. Astheimer, H. P. Hansen, and H. Leach, 1979: an oceanic environment where corroboration of results Vertical microscale distribution of phytoplankton in relation to is dif®cult. Performance evaluation of the ®ber-optic the thermocline. Mar. Ecol. Prog. Ser., 1, 187±193. ¯uorometer consisted of evaluation of the optics and Dickey, T., J. Marra, T. Granata, C. Langdon, M. Hamilton, J. Wiggert, D. Siegel, and A. Bratkovich, 1991: Concurrent high resolution electronics of the instrument, stability and drift analysis, bio-optical and physical time series observations in the Sargasso effects of scattering and high turbidity on the ¯uores- Sea during the spring of 1987. J. Geophys. Res., 96, 8643±8663. cence signal, and intercomparison with a commercial D'Sa, E. J., 1996: Pigment dynamics in a coastal bottom boundary ¯uorometer. Over a 7-week ®eld deployment of the in- layer and its relation to the physical regime: Measurements using an in situ ®ber-optic ¯uorometer. Ph.D. dissertation, University strument there was minimal drift in the strobe excitation of Southern Mississippi, 185 pp. [Available from UMI, 300 N. signal and in the electronics. Multiple calibrations of Zeeb Rd., Ann Arbor, MI 48106.] the ®ber-optic ¯uorescence sensors indicate measure- , S. E. Lohrenz, V. L. Asper, R. A. Walters, M. J. Morris, and C. Rathbun, 1994: A multi-sensor in situ ®ber optic ¯uorometer. ments to be generally stable; however, biofouling effects Soc. Photo-Opt. Instrum. Eng., 2258, 33±43. were noticeable during long-term operations. The ¯uo- Hitchcock, G. L., E. J. Lessard, D. Dorson, J. Fontaine, and T. Rossby, rescence signal measured by the MFF was found not to 1989: The IFF: The isopycnal ¯oat ¯uorometer. J. Atmos. Oce- be affected by high levels of suspended particulate mat- anic Technol., 6, 19±25. Lorenzen, C. J., 1966: A method for the continuous measurement of ter in the sampling medium as indicated by laboratory in vivo chlorophyll concentration. Deep-Sea Res., 13, 223±227. experiments. Marra, J., 1994: Capabilities and merits of long-term bio-optical Field measurements with the ®ber-optic ¯uorometer moorings. Ocean Optics, R. W. Spinrad, K. L. Carder, and M. J. Perry, Eds., Oxford University Press, 189±201. at coastal sites has enabled the investigation of the long- Parsons, T. R., Y. Maita, and C. M. Lalli, 1984: A Manual of Chemical term vertical pigment structure of near-bottom waters and Biological Methods for Seawater Analysis. Pergamon Press, in conjunction with physical variables. 173 pp. Plaza, P., N. Q. Dao, M. Jouan, H. Fevrier, and H. Saisse, 1986: Simulation et optimisation des capteurs a ®bres optiques adja- Acknowledgments. We wish to thank the various peo- centes. Appl. Opt., 25, 3448±3454. Thomsen, L., G. Graf, V. Martens, and E. Steen, 1994: An instrument ple who contributed to the project; special thanks to for sampling water from the benthic boundary layer. Cont. Shelf Catherine Rathbun and Diane Arwood, who assisted Res., 14, 871±882. with the instrument development and ®eld tests. Mike Trowbridge, J. H., and A. R. M. Nowell, 1994: An introduction to the sediment transport events on shelves and slopes (STRESS) Morris of Ocean Optics provided valuable support dur- program. Cont. Shelf Res., 14, 1057±1061. ing the instrument development. Sandy Williams of Whitledge, T. E., and C. D. Wirick, 1983: Observations of chlorophyll WHOI contributed helpful suggestions on the mechan- off Long Island from a moored in situ ¯uorometer. Deep-Sea ical design and was responsible for the deployment and Res., 30, 297±309. Williams, A. J., J. S. Tochko, R. L. Koehler, W. D. Grant, T. F. Gross, recovery of the instrument that was set up on the BASS and C. V. R. Dunn, 1987: Measurement of turbulence in the tripod. Jeff Kinder provided assistance during the in- oceanic bottom boundary layer with an acoustic current meter strument deployment on R/V Cape Hatteras. array. J. Atmos. Oceanic Technol., 4, 312±327. Wirick, C., 1994: Exchange of phytoplankton across the continental Support was received on DOE Grants DE-FG02- shelf-slope boundary of the Middle Atlantic Bight during spring 92ER61663 and DE-FG05-95ER62071, and a NOAA 1988. Deep-Sea Res., 41, 391±410.

Unauthenticated | Downloaded 09/25/21 04:35 PM UTC