CICEET Progress Report for the period 3/15/05 Through 9/15/05

Project Title: Advanced Laser Fluorescence (ALF) Technology for Estuarine and Coastal Environmental Biomonitoring
Principal Investigator(s): Alexander Chekalyuk, Kenneth Moore, David White, and Dwayne Porter
Project Start Date: 9/15/03

Figures

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Objective for the Reported Period
To conduct extensive laboratory and field tests of the prototype ALF system on participating NERR sites

Tasks to meet objectives

• To conduct initial ALF tests at the South Slough (OR) NERR
• To conduct a shipboard ALF survey in the Chesapeake Bay and York River and laboratory sample measurements at the Chesapeake Bay (CB-VA) NERR
• To conduct additional field ALF tests in Pacific coastal zone of San Diego
• To initiate thorough analyses and evaluation of acquired field data to meet the overall project objective

Progress on Tasks nad Preliminary Data
ALF Instrument Development
Development of the compact, laptop-footprint prototype ALF system (Figure 1) was finished in March-April 2005 and the system was configured for its operational field tests. The ALF instrument includes blue and green diode lasers, a hyperspectral CCD sensor with a fiber light collector, a PDP sensor of variable fluorescence with a PMT detector and a waveform digitizer, a flow-through sample unit with a miniature pump, a USB GPS navigation module, and a laptop PC with a USB controller. The ALF software with a user-friendly interface (Figure 2) was developed under the LabView software environment (National Instruments, Inc.). The system can operate on AC power or internal rechargeable battery; in the latter case providing up to 6 hours of continuous high-speed underway measurements in a waterproof Pelican case (Figure 1, Figure 5) at a rate ~600 samples per hour (tested in the York River up to 33 miles/hour, see Figure 5). It can also be used for discrete sample analyses and long-term flow-through monitoring at stationary settings (tested at the Pier of Scripps Institution of Oceanography (SIO), see upper-right photos in Figure). Figure 2 displays a typical screen capture, which present hyperspectral emission signatures stimulated with blue and green lasers (upper panels) and PDP fluorescence induction measurements (lower right panel).

ALF Field Deployments
The ALF-1 instrument was successfully tested during 2 estuarine field campaigns at the Sough Slough (OR) and Chesapeake Bay (VA) NERR sites in April and July 2005. The initial field measurements with the ALF instrument at the South Slough NERR in close collaboration with the NERR research coordinator Dr. Steve Rumrill, who provided excellent support. The tests included a laboratory study of ALF fluorescence signatures for water samples collected in a range of bio-environmental conditions on the NERR site (Figure 1, upper left). The ALF capability for continuous flow-through measurements onboard a small motorboat was also successfully tested (see results of the ALF mapping in Figure 3, thicker and brighter lines represent increased magnitudes of the ALF variables). A high-resolution data set, which provides information about phytoplankton biodiversity, physiological status, pigment biomass, water turbidity and concentration of dissolved organic matter, was acquired in about 2 hours of the underway motorboat ALF monitoring. Figure 4 displays a GPS measurement map and detailed along-transect spatial distributions. Numbered arrows marks along horizontal axis of the plot here and below correspond to the marked locations on the respective map. Supporting data (not presented here) include YSI measurements of standard water quality parameters (STD, dissolved oxygen, and water turbidity) at selected site locations; more detailed information was provided by HPLC and optical microscopic analysis of collected water samples.

A more extensive ALF survey was conducted in the Chesapeake Bay and York River (Figures 5-8), including the CB-VA NERR site, in July 2005 in close collaboration with the project CoPI and the CB-VA Research Coordinator, Dr. Ken Moore (Virginia Institute of Marine Science (VIMS)). Results of the ALF motorboat mapping presented in Figure 6A show significant meso- and small-scale spatial variability in the ALF bio-optical characteristics. Figure 6B present spatial distributions of the key water quality variables recorded with the supporting flow-through DataFlow system, which operated along with the ALF instrument on board a motorboat. The acquired data sets provided for quite comprehensive characterization of bio-environmental situations in the surveyed area, which would not be possible with ‘conventional’ monitoring techniques. In particular, both CDOM and water turbidity gradually increased up the river. The former showed significant spatial correlation with pigment distributions in the local algal bloom found between the transect points 9 and 10. The magnitudes of variable fluorescence, Fv/Fm, varied in the 0.2-0.3 range over most of the transect, thus indicating poor phytoplankton photo-physiological status, which was later confirmed by the laboratory measurements of the dark-adapted water samples (for example, see Figure 7D and discussion below). The elevated Fv/Fm magnitudes were found in the upper York River (points 9 and 10 in Figure 6A, B), where extremely high and spatially variable Chl-a concentration, reaching 100 mg/l (panel A), indicated a patchy, but very intense small-scale algal bloom. Despite the general correlation between the Chl-a and phycoerythrin (PE) spatial distributions, PE fluorescence varied in a smaller, 5-fold range.

The ALF measurements of the PE/Chl-a ratio indicated significant structural changes in phytoplankton assemblages in the surveyed area, which was later confirmed by the HPLC analyses. ALF hyperspectral measurements of pigment emission (for example, see Figure 6C; curve numbers correspond the marked locations in panel D; the CDOM and water Raman components were removed using spectral deconvolution ­ see below) indicated significant spectral changes in pigment fluorescence bands associated with the observed taxonomic variability. In particular, in the Chesapeake Bay area adjacent to the York River mouth, where cyanobacteria were abundant, the ALF pigment measurements indicated respectively elevated concentration of phycocyanin (PC, Figure 6C; curve numbers correspond the marked locations on the map in Figure 6D) and dominance of the PE spectral form with maximum at 579 nm (see Figure 6D and Figure 7A). In the York River, where cryptomonads were the dominant PBP-containing algal group, the PE maximum was respectively shifted towards 590 nm (for comparison, see our laboratory spectral measurements with pure cultures in Figure 7B). This shift showed high correlation (Rsq = 0.88, see Figure 7C) with the HPLC-measured concentration ratio of zeaxanthin and alloxanthin, the pigment biomarkers of cyanobacteria and cryptophytes, respectively. This observation opens up a pathway to development of a new method for real-time in vivo quantitative ALF assessments of these two ecologically important algal groups. A smaller, but noticeable spectral shift in the Chl-a fluorescence peak (see Figure 6C), which seems to spatially correlate with the HPLC-revealed change from diatom to dinoflagellate dominance along the transect, was also observed.

The original field deployment plan was extended to include ALF testing in significantly distinct coastal marine environment. In June 2005 the ALF prototype was deployed in the Pacific coastal zone of San Diego for high-resolution measurements of small-scale spatial variability in a bloom of the non-toxic ‘red tide’ dinoflagellate L. polyedrum, an organism that blooms regularly off the Southern California coast. This study was conducted in close collaboration with Dr. B. Gregg Mitchell and his research team at Scripps Institute of Oceanography (SIO, UCSD), who made it possible by providing their excellent support, laboratory and other facilities. Figure 9 presents spatial distributions of Chl-a concentration, CDOM and variable fluorescence acquired during the small skiff survey near the SIO pier. Panels A, B and C indicate that Chl-a concentration exhibited extremely high spatial variability with 10-20 fold concentration gradients over sub-100 meter scale, with unusually high concentrations of 40-50 mg/m----3 in dense algal patches. Measurements of variable fluorescence, Fv/Fm, suggested generally depressed physiological status of phytoplankton (0.2-0.36), likely caused by nutrient limitation at the late stage of bloom development. The strong spatial structure in Fv/Fm (see panel B) indicates that the community had small-scale forcing of its physiology, information that could not be attained with any alternative sampling regimen. The CDOM peaks also were spatially correlated with the Chl-a peaks, suggesting their biological origin. This initial test provided valuable information to evaluate ALF performance in different coastal marine environments and respectively adjust its measurement and analytical protocols. It also showed the ALF potential as a useful tool for detection and monitoring dynamics of phytoplankton coastal blooms, which may be extended in the future towards harmful algal bloom (HAB) applications.

ALF Algorithm Development
New automatic spectral deconvolution algorithms have been developed to account for spectral complexity of coastal and estuarine aquatic environments, where spectral bands of the seawater constituents are overlapped. For example, Figure 8A presents results of processing fluorescence emission spectrum (FEM, blue line) measured during the ALF motorboat survey in the Chesapeake Bay with green laser excitation. The deconvolution procedure generates a best fit of the measured spectrum by a weighted summation of spectral bands representing major water constituents typical for coastal and estuarine environments, i.e. water Raman scattering (dark red, 650 nm peak), CDOM (green, 570 nm peak), Chl-a (black, 685 nm peak) and PE (navy, 590 nm peak) fluorescence. The sum of the weighted constituents (red line) agrees excellently with the raw hyperspectral emission spectrum (blue line). A similar automated deconvolution procedure is routinely applied to the FEM spectral signatures measured with the blue ALF laser. Clearly, the multiple overlapped components in the emission spectrum can affect the accuracy of retrieval of individual components if the signal is not treated properly as the sum of the individual spectra. Despite reporting similar observations in the 80-90s, most of the fluorometers utilized for concentration measurements do not posses adequate resolution of the emission spectrum and, therefore, ignore the overlap, which results in reduced accuracy of their estimates. Our analysis, based on these extensive field tests in diverse environments, suggests that spectral deconvolution of hyperspectral seawater emission measurements is critical for accurate concentration assessment of seawater constituents. Additional significant improvement in the accuracy of the concentration assessments can be achieved by normalization of spectrally-deconvolved fluorescence intensities to the intensity of synchronously measured water Raman scattering, an ‘internal standard’ that allows to account for signal variability caused by absorption and scattering in the sampling volume and during light propagation in the media. For example, Figure 8B present a combined correlation plot (Rsq = 0.92) between the Chl-a laser assessments and HPLC pigment measurements in the South Slough (OR, Figure 4) and Chesapeake Bay/York River (Figure 6A). The remarkable match in the slope of the correlation line for measurements conducted in totally different bio-optical environments confirmed the accuracy of the ALF deconvolution/normalization algorithm.

The deconvolution/normalization protocol also provides for improved assessments of two other important components or coastal aquatic environments, CDOM and PBP pigments. For example, our earlier CDOM retrievals showed reasonably high correlation (Rsq = 0.79) between independently measured dissolved organic carbon (DOC) and Raman-normalized CDOM fluorescence, which is not achievable by conventional fluorometry and suggests that ALF technique may be a useful tool for the ocean carbon studies. The reported here ALF field deployments have resulted in another remarkable finding presented in Figure 8C. Two spectrally distinct CDOM chromophores were resolved in hyperspectral CDOM signatures; the relative intensity of the 493 nm chromophore was found significantly higher in the San Diego coastal zone (SDS FEM in Figure 8C) compared to estuarine waters of the York River (Chesapeake Bay area, CB FEM) and South Slough (OR, SS FEM). While it is known that different CDOM chromophores have different spectral excitation and emission matrices, there is no system to our knowledge that can resolve these in an automated flow through mode, and also use this spectral information in proper corrections for other critical constituents including Chl-a and PBPs.

The correlation analysis has revealed that laser measurements can provide additional useful information, not considered in the original ALF proposal In particular, ALF assessment of water turbidity was included in the ALF measurement protocol as a ‘bonus’ after analysis of our ALF field measurements at the CB-VA NERR conducted in June and September 2004. Rsq as high as 0.99 was observed between the Sept. 2004 data for the laser elastic scattering and water turbidity. Our June 2005 measurements in the Chesapeake Bay and York River have confirmed the ALF capacity for accurate assessment of water turbidity in a broad range of coastal and estuarine environments. For example, Figure 8D displays a correlation plot (Rsq = 0.98) between the DataFlow boat measurements of water turbidity at the stations, which are marked on the map in Figure 6D, and the ALF parameter, K_RAM, retrieved from the ALF FEM_532 spectra.

The feasibility to determine PE concentration from in vivo fluorescence measurements needs a special discussion. Earlier studies have revealed relatively low and variable correlation between PE concentration and PE fluorescence. We speculate that the observed low correlation may be attributed to spectral complexity of the seawater emission. For example, even with a green excitation typically used for PE fluorescence measurements, CDOM fluorescence near the PE peak (565-590 nm) may be comparable or exceed the actual PE fluorescence (see Figure 8A). An important element of our future research is to explore the validity of our speculation, which may result in accurate in vivo concentration measurements of PE and other PBP pigments, phycocyanin (PC) and allophycocyanin (APC) with the ALF technique. The recent ALF deployments in the Chesapeake Bay have already shown feasibility of the PC detection and assessments of its concentration from the ALF hyperspectral measurements of seawater emission with 532 nm laser excitation (for example, see curves 1, 2 in Figure 6C). As far as we know, this is the first observation of PC fluorescence in natural algal populations, which was made possible due to utilization of the ALF spectral deconvolution algorithms. Further development of the ALF deconvolution technique will allow us to discriminate the PBP and Chl-a pigment fluorescence bands and assess PBP pigment concentrations for improved characterization of phytoplankton populations in natural aquatic environments.

An important methodological result of our initial ALF field deployments is that it is feasible to retrieve accurate quantitative assessments of phytoplankton physiological status using sampling from the surface exposed to bright ambient light. For example, we present in Figure 7D a comparison of PDP Chl-a fluorescence induction measured in the dark flow-through sample chamber of the ALF prototype instrument during motorboat underway monitoring in the York River at station 6 (see a map in Figure 6D) at noon under bright sun (green curve) and in the dark-adapted seawater sample collected at the station (red curve). The intensity of the saturating PDP flash was the same. Although the time scale of the induction rise of underway measurements was longer (~600 vs. 60 ms), indicating adaptation of photosynthetic apparatus to the excessive ambient light to reduce the rate of photosynthesis, the Fv/Fm magnitude calculated from the degree of rise was comparable with that one measured in the dark-adapted sample. Such consistency has been confirmed for the data sets acquired in April-July 2005 in the Pacific coastal zone of San Diego, in the South Slough Estuary (OR) and in the Chesapeake Bay/York River area. Though these observations need biophysical interpretation, which we plan to work on, in practice they open up a pathway for assessments of phytoplankton photo-physiological status from continuous real-time flow-through underway measurements without traditionally assumed long-term (~30 min) dark adaptation. Furthermore, these ALF field deployments indicate that the rate of ongoing PSII photochemistry may be assessed from the ALF PDP and PAR measurements, which may provide a pathway towards estimations of primary productivity from the underway ALF fluorescence data.

Project Objectives for Next Reporting Period

Objectives
The project objective for the next reporting period will be to refine the ALF instrument configuration, measurement protocols, data processing and analytical algorithms based on the initial instrument field tests and deployments. A final project report will describe project activities, analyze research results, and provide recommendations how the Advanced Laser Fluorescence technology can be operationally utilized for bio-monitoring estuarine and coastal areas. This will complete the initial stage of ALF development funded by CICEET in 2003.

Anticipated Success in Meeting Project Objectives
We do not anticipate significant problems with meeting the project objectives in the scheduled project period.

Overall Project Timeline Update
The ALF project has been granted the requested no-cost extension Sept. 1 through Dec. 31, 2005.

Project Needs and Potential Difficulties
(1) Comparison of data acquired in the Chesapeake Bay (VA), North Inlet ­ Winyah Bay (SC), South Slough (OR) NERRs and in the Pacific coastal zone of San Diego has shown significantly different patterns in pigment and CDOM abundance, phytoplankton composition, and their photo-physiological status. The entire NERR network of 26 sites represents very diverse aquatic environments. Therefore, more extensive ALF field tests in a broad range of bio-environmental conditions are needed to adjust measurement protocols and analytical algorithms with regard to diverse regional, local and seasonal conditions.

(2) The ALF technology development is eventually targeted for operational utilization within SWMP biomonitoring by the NERR personnel. Though the ALF prototype instrument and software are designed to be user-friendly and do not require special professional skills for their operation, the ALF technology per se is relatively new and may need an ‘educated end user’ for its successful commercialization and future flawless incorporation in the SWMP biomonitoring program.

We consider resubmitting the follow-on CICEET proposal “Advanced Laser Fluorescence (ALF) Biomonitoring in the Middle-Atlantic NERR Sector and Great Lakes”, which addresses the above issues and was not selected for funding in 2005. The proposed research is the next important and logical step towards successful ALF operational utilization. At the following stage of the ALF development, we propose to conduct a series of seasonal and annual tests and evaluations of the ALF technology in the Middle-Atlantic NERR Sector. Field deployments will be also conducted in the Old Woman Creek (OWC, Lake Erie, OH) NERR for additional ALF evaluation in the Great Lakes-type freshwater aquatic environment.

Dissemination
The ALF research results will be presented at the 18th Biennial Conference of the Estuarine Research Federation (October 16-21, 2005, Norfolk, VA) as an invited talk during the Chesapeake Bay Colloquium and as an interactive poster at the SYM-04 Symposium “Observing the Coastal and Ocean Environment: Developments in Sensor Technology and the Use of Long-Term Data Sets for Operational Ecology”. The project results will be included in the planned presentation at the AGU/ASLO Ocean Sciences meeting in Honolulu, Feb. 2006. Two manuscripts are currently in preparation for publication in peer review scientific journals.

Expenditures
The expenditures were in the range anticipated for the work accomplished to date.