Grantee Research Project Results
2003 Progress Report: Behaving Drifters as Gymnodinium breve Mimics
EPA Grant Number: R829370Title: Behaving Drifters as Gymnodinium breve Mimics
Investigators: Kamykowski, Daniel , Wolcott, Thomas G. , Janowitz, Gerald S.
Institution: North Carolina State University
EPA Project Officer: Packard, Benjamin H
Project Period: November 19, 2001 through November 18, 2004 (Extended to May 18, 2006)
Project Period Covered by this Report: November 19, 2002 through November 18, 2003
Project Amount: $423,493
RFA: Ecology and Oceanography of Harmful Algal Blooms (2001) RFA Text | Recipients Lists
Research Category: Water Quality , Water , Aquatic Ecosystems
Objective:
Because Gymnodinium breve was renamed Karenia brevis, the G. breve Population Mimic (GBPM) is now called the K. brevis Population Mimic (KBPM). The objectives of this research project are to: (1) follow the trajectories of free-ranging, buoyancy-adjusting floats programmed to act as KBPM on the west Florida shelf; and (2) incorporate the results into evolving physical-biological models. Hypotheses that will be tested remain the same, and are listed below.
In Support of Improving Behavioral Programming
Hypothesis 1: K. brevis exhibits positive chemotaxis toward inorganic and organic nutrients.
Hypothesis 2: K. brevis adjusts its migration in response to the nutrient sources that yield a chemotactic response when they are available in only part of a laboratory water column.
In Support of Testing KBPM Utility
Hypothesis 3: KBPMs and CODE drifters placed in proximity to each other follow similar tracks.
Hypothesis 4: KBPMs that migrate vertically in the same water column as other KBPMs distributed and maintained at different water column depths integrate the horizontal flows encountered during the vertical migration, and track differently than the KBPM maintained at any given depth.
Hypothesis 5: The environmental exposure recorded by the migrating KBPMs will support better relative growth than that recorded by the stationary depth KBPM when entered into an existing model of K. brevis' physiological responses.
Hypothesis 6: KBPM, programmed to vertically migrate, tracks in situ K. brevis populations when dropped into natural aggregations of K. brevis.
In Support of Improving Previous Numerical Models
Hypothesis 7: Physical-biological models incorporating information obtained from KBPM better simulate natural events (population growth and physical aggregation) than those that do not.
Progress Summary:
Hypothesis 1
Several chemotaxis experiments were performed with 10 different K. brevis strains in association with a 10-week undergraduate intern, Jessica Ramsey, during the summer of 2003. Two strains exhibited statistically significant increases in abundance in capillaries filled with culture media versus unenriched seawater. Preliminary nutrient-specific chemotaxis trials did not yield any statistically significant results.
Hypothesis 2
The mesocosm incubator room was upgraded with a new compressor and control valves to improve reliability and efficiency. Two mesocosms each illuminated at approximately 1,000 µm quanta m-2 s-1 were maintained at constant temperature during an experiment comparing nitrate replete and deplete mesocosms. Sample analysis is underway.
Hypotheses 3 and 4
For the KBPM, we are in the process of adapting a recently released Microchip microcontroller (PIC18F8720). It will eliminate external program memory and associated glue logic, requiring only nonvolatile external serial EEPROMs (up to 128 Kb) for logged data. It is capable of switching from high speed (up to 20 MHz crystal) to a slower oscillator (32.768 KHz) when rapid calculations are not required, dropping its current demand from around 15 mA to greater than 100 µA. The PIC18F8720 will handle the complex calculations at high speed, but also will be able to continuously monitor environmental variables at low power, allowing it to "wake up" only when something has changed enough to require action. The net result should be a considerably extended battery life. It has ample output pins and a variety of built-in peripherals (comparators, 10-bit analog to digital converters, two serial ports, I2C port, etc.). Furthermore, its program resides in flash memory, and can be self-reprogrammed by a resident "bootloader" routine. New programs may be loaded in a few seconds through the serial port, as a simple menu option. The entire PC board (except temperature sensor, IR communications transceiver, and radio beacon module) will consist of the PIC, four serial EEPROMs, real-time clock, sensors for rotation, pressure and light, and drivers for buoyancy pump and conductivity cell.
A bootloader program has been adapted for this application and programmed into the microcontroller. Reprogramming and testing new programs now is routine, so we have reorganized the program framework for the new processor, and are proceeding to link the remaining circuit blocks, write/adapt software routines, and test them.
The "speed through water" sensor is a new device, the Austria Microsystems AS5020, that encodes the angular position of a magnet suspended above it with 5.6-degree resolution, and sends the angle to the microcontroller as serial data. We have determined that a 6-mm diameter rare-earth magnet suspended a few mm above the chip on a low-friction pivot will provide a field whose rotational position the device can encode, while acting as a compass needle that constantly points north.
Program routines to read/write nonvolatile serial EEPROMS and real-time clock have been written for a 19 g PIC-based datalogger that we have been releasing on migrating blue crabs; they will be easily portable to the new PIC in the mimic. We intend to use a binary-encoded clock that simply keeps track of cumulative seconds.
The salinity sensor also has been developed and brought through several steps of refinement in the blue crab dataloggers. In the mimic application, we will simply transfer what we have learned (graphite electrodes, temperature correction routines, and linearization), with the luxury of using longer, hence larger diameter, tubes for the conductivity cell. This should reduce the probability of artifacts caused by particles or fouling.
The new PIC processor provides 10-bit A/D conversion; this is ample resolution, given the linearity and hysteresis of micromachined silicon pressure transducers. The technology for reading and calibrating these transducers also has been worked out in the blue crab dataloggers, (with 8-bit resolution) and has proved reliable. With the option of a larger platform, we intend to use a temperature- and gain-compensated transducer and a digital potentiometer rather than a hardware pot for calibration; this will allow resetting the span of the depth transducer from the menu.
The remaining sensor is for light (photosynthetically active radiation). We will use the new TAOS TCS230. This sensor has a group of unfiltered photodiodes, as well as groups with red, green, and blue filters. The various photodiode groups are selectable under software control. The device emits a pulse train with frequency proportional to the light intensity seen by the selected diodes, making it very easy to interface with a microcontroller. A similar single-diode device was used on the first mimic design, and the routines for reading the new sensor will require only the addition of diode-switching commands. We have analyzed the standard curves for the various TCS230 diodes, and have calculated weighting factors that provide a fairly flat response more than 400-700 nm. We intend to conduct ambient light to the sensor (located on the PC board) using a light fiber (Fujitsu, from Advanced Lighting Systems, Inc.) that will originate in a 2 cm Teflon ball (which provides unbiased spherical response) 10 cm above the pressure housing. It will be protected by a 10 cm long stainless steel tube to where it passes through the plug at the top of the pressure housing. We await spectral transmission curves for the light fiber material from the supplier, although they are alleged to be "quite flat." These curves will influence our choice of correction filters to eliminate the high sensitivity of silicon photodiodes to infrared wavelengths more than 700 nm. Our current intention is to use a dielectric "hot mirror;" these have a very flat transmission spectrum between 400 and 700 nm with quite sharp cutoffs above and below the passband.
The position-fixing ability of the mimic will be conferred by a Garmin global positioning system (GPS) receiver engine, communicating with the microcontroller over its own serial port (the second port eliminates the need for multiplexing communication with the GPS and with a host computer). The "C" code for collecting position and time data from a Garmin module was written for a previous adaptation of the first-generation mimic boards, and will require only minor tweaking for the new microcontroller. It includes code to update the system clock with the satellite time signal so that each time the mimic rises to the surface for a position fix, its clock will be corrected to within microseconds of exact time. The most recent Garmin GPS modules have one-half of the weight and power consumption of previous models; they are about 1 cm thick and are the size of a large postage stamp.
The radio beacon that will aid recovery of mimics will be switched by the microcontroller. We programmed the PIC microcontroller to encode the last several digits of the most recent latitude and longitude into the radio beacon signal. This stratagem allows the recovery vessel to come within 20 m of the beacon even before it is sighted. The beacon transmitters may be purchased (AVM Instruments) or constructed in-house. Communication with a host computer for loading programs, modifying parameters, issuing commands, and uploading data will be conducted with an infrared transceiver relying on IrDA protocols. A development kit for such an interface is on hand, and will be brought into the design in its turn.
The basis of the pressure housing is a fire extinguisher cylinder; these are inexpensive, readily available, and of almost exactly the right characteristics for 100 m maximum depth. Epoxy rings will be cast in silicone molds and will be fastened to the aluminum cylinders with urethane marine adhesive/sealant. O-ring seals will ensure that there is no leakage between the mating rings or between the rings and cylinder wall. The epoxy rings will include components that must be outside the pressure housing: the temperature sensor (a Dallas 13-bit 1-wire precision sensor), the IrDA transceiver, fittings for the pressure transducer and buoyancy pump tubes, and the salinity sensor (conductivity cell). The valve at the neck of the extinguisher cylinder will be replaced with a screw-in epoxy plug, into which are cast connections, components or fittings for the light collector, GPS antenna, radio beacon, and 32 KHz tracking/telemetry pinger.
The variable ballast (buoyancy adjusting) system has been redesigned for easier manufacturing and greater reliability. The solution we have selected is to use neoprene ballast bags (economical—available in a variety of configurations) and ethylene glycol (antifreeze—noncorrosive and with some lubricity) as the ballast fluid. Program routines for adjusting buoyancy to remain neutral, or to move up/down at biologically realistic rates, exist for the previous mimic design. We anticipate that only subtle adjustments will be required for the new version of the mimic, because we expect that it will be lighter (hence of lower displacement), and because it will need to incorporate information from the "speed through water" sensor.
Hypothesis 5
The field test of the Acoustic Doppler Current Profiler (ADCP) and digital GPS unit were successful. The environmental measuring system assembled by SubChem Systems, Inc. (a SeaBird SBE 19 plus CTD, a LICOR LI-190SA quantum sensor, a SubChemPak analyzer (nitrate, ammonia, and phosphate), and a WETLabs ECO-BB2F [chlorophyll fluorescence and particle concentration via backscatter]) was upgraded. In the SubChemPak analyzer, ammonia has replaced iron, and the depth capability has increased from approximately 15 m to less than 100 m. Also, all instruments now log on a WETLabs DH-4, and are accessed with the WETLabs WAPS software. Initial sea trials are complete, and the instrument is useable. Delivery is expected before the end of 2003, and the graduate student, Geoff Sinclair, supported by this grant, visited Gary Kirkpatrick at Mote Marine Laboratory to gain hands-on experience with their SubChem unit.
Hypothesis 6
Laboratory experiments were performed in support of behavioral model improvements for the KBPM. Radial photosynthetron experiments with 10 strains isolated from geographically diverse areas in the Gulf of Mexico and the Florida Atlantic coast fell along a gradient of light capability. In different experiments, the order of the different strains along the light capability gradient changed, depending on culture density and growth rate. These photoresponse characteristics will be incorporated into the KBPM productivity program.
Hypothesis 7
The time-dependent, three-dimensional distribution of a population of K. brevis was explored through the use of an Eulerian model (Janowitz, et al., 2004). The model combines a previously developed physiologically based behavioral model of these dinoflagellates with a simple model for a three-dimensional wind driven flow field over a variable-depth continental shelf. The behavioral model is simplified from that used in previous applications, and sigma coordinates are utilized in the model. Model results indicate that even for the relatively weak wind-driven currents used in our simulation, a nonquantized population can develop into two spatially distinct quantized populations in a period as short as 1 day, where, for present purposes, a quantized population is one in which all cells are at the same stage of the cell cycle.
Future Activities:
Hypothesis 1. We will continue to: (1) develop reliable chemotaxis protocols for K. brevis; and (2) run trials with specific nutrients contained in the culture media that elicited a statistically significant result in two K. brevis strains.
Hypothesis 2. Experiments with nutrient-stratified mesocosms are planned for Year 3 of the project.
Hypotheses 3 and 4. Most of the components described above for the KBPM have been designed or built. By the end of December 2003, a large proportion of the programming and testing should be complete. As soon as each of the subsystems have been tested, we can purchase the requisite parts and go into "production mode," with prototype tests conducted soon in the spring semester. The "behavioral" algorithms will be the most mathematically complex portion of the programming, but will be the simplest part. With a high-level language such as "C," they are simply entered as formulas, and the compiler does the rest. Once the circuit blocks are working as intended, we will rapidly develop and test behavioral models.
Hypothesis 5. Now that the major support equipment has been purchased and upgraded, additional deployment opportunities will be taken to develop operational skills that will be routine when the KBPM deployments begin.
Hypothesis 6. Laboratory experiments will continue to examine biochemical pools and physiological/behavioral rates critical to the parameterization of the K. brevis biological model. Small-scale experiments will continue to examine how swimming path characteristics change at different water column locations over the course of a diel vertical migration.
Hypothesis 7. The biological part of the biophysical model will be upgraded with the same parameterization used in the KBPM. Comparison runs will determine how the full range of K. brevis physiological/behavioral capabilities determined in the laboratory influences the ability to predict bloom development based on standard physical forcing.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
Other project views: | All 39 publications | 11 publications in selected types | All 10 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Nagai T, Yamazaki H, Kamykowski D. A Lagrangian photoresponse model coupled with 2nd-order turbulence closure. Marine Ecology Progress Series 2003;265:17-30. |
R829370 (2003) R829370 (2004) R829370 (Final) |
Exit Exit |
Supplemental Keywords:
marine, ecology, environmental chemistry, physics, organism, measurement methods, modeling, Southeast, ecosystem protection, environmental exposure and risk, water, biology, ecological risk assessment, ecosystems, oceanography, algal blooms, Ecology and Oceanography of Harmful Algal Blooms, ECOHAB, Karenia brevis population mimics, KBPMs, K. brevis behavioral submodel, K. brevis red tides, K. brevis toxins, Gymnodinium breve population mimics, GBPMs, G. breve behavioral submodel, G. breve red tides, G. breve toxins., RFA, Scientific Discipline, Water, Ecosystem Protection/Environmental Exposure & Risk, Ecology, Oceanography, algal blooms, Ecological Risk Assessment, Ecology and Ecosystems, Biology, brevetoxins, Gymnodinium breve toxins, ECOHAB, G. breve Population Mimics (GBPMs), G. breve red tidesProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.