Grantee Research Project Results
2001 Progress Report: CISNet: Molecular to Landscape-Scale Monitoring of Estuarine Eutrophication
EPA Grant Number: R826944Title: CISNet: Molecular to Landscape-Scale Monitoring of Estuarine Eutrophication
Investigators: Morris, James T. , Noble, Peter , Lewitus, Alan , Porter, Dwayne , Jensen, John , Fletcher, Madilyn
Institution: University of South Carolina at Columbia
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 1998 through September 30, 2001 (Extended to September 30, 2002)
Project Period Covered by this Report: October 1, 2000 through September 30, 2001
Project Amount: $579,650
RFA: Ecological Effects of Environmental Stressors Using Coastal Intensive Sites (1998) RFA Text | Recipients Lists
Research Category: Aquatic Ecosystems
Objective:
This research addresses three hypotheses: (1) the composition and abundance of bacterial and phytoplankton communities will differ among estuaries as functions of nutrient availability; (2) bacterial and phytoplankton communities form associations that vary in complexity (species diversity) as a function of nutrient availability; and (3) at a landscape-scale, remote sensing of the concentration of chlorophyll in emergent wetland vegetation will provide a quantitative index of wetland condition and will demonstrate differences in nutrient loading among estuaries.
Progress Summary:
The first major component of this project focuses on hypotheses 1 and 2, and concerns the influences of the nutrient status of estuaries on the composition of bacterial and phytoplankton communities. The field work is concentrated in the North Inlet (NI) estuary, SC (a nutrient impoverished estuary), and the ACE Basin estuary, SC (a highly fertile estuary). Samples for bacterial community composition have been collected in tandem with phytoplankton samples. Bacterial samples are being analyzed using genetic profiling techniques (PCR/DGGE). Bacterial samples are being processed. Phytoplankton community composition is being assessed by analyzing the types of photo-pigments that are present in a water sample. These data now are being analyzed. Significant progress was made on five tasks identified with this part of the project.
Task 1: Develop the High Performance Liquid Chromatography (HPLC) pigment protocol to optimize separation of pigments of known chemotaxonomic importance.
The protocol being used makes use of HPLC and is based on a method used by Van Heukelem and Thomas (2001), which has been found to provide good pigment separation and sensitivity.
Task 2: Complete phase 1 of the Chemtax calibration process, which included high performance liquid chromatography (HPLC) analysis of pigments harvested from monocultures of 12 taxonomically diverse species of estuarine phytoplankton grown at different physiological states.
A new matrix factorization program, "CHEMTAX," has been used to relate HPLC pigment data to phytoplankton composition in the open ocean. We are in the process of applying this program to estuarine communities by calibrating it against phytoplankton monocultures isolated from estuarine environments. Thus far, HPLC chromatograms have been analyzed from 12 taxonomically diverse species acclimated to growth at saturating or limiting light, and harvested from exponential or stationary growth phase. The combined applications of CHEMTAX and neural computing to HPLC pigment profiles will lead to a major advance in current capabilities in determining phytoplankton community composition.
Task 3: Conduct weekly monitoring of physical, chemical, and biological variables at 2 sites within North Inlet estuary that vary in patterns of nutrient quantity and quality.
Water samples have been collected weekly from two sites in the NI estuary. One site is closer to the wetland forest, and therefore is influenced by runoff and groundwater to a greater extent than the other. The correlative relationships between pigments and physicochemical properties now are being assessed.
During the 1999 monitoring effort, a red tide was documented at one of the sampling sites. This was caused by a dinoflagellate species (tentatively identified as Peridinium sp.), which comprised greater than 95 percent of the total phytoplankton biomass, reaching greater 100,000 cell ml-1. Characterizations of the red tide thus far include spatial distribution (daily, areal, diel, depth, tidal stage) in relation to environment (nutrients, salinity, temperature, water flow, microzooplankton and phytoplankton composition); identification; pigment composition; and physiological and life cycle factors. Although the understanding of the physiological ecology of this dinoflagellate will be improved after the completion of several of the above ongoing analyses, early results are providing important clues. For example, dissolved inorganic carbon (DIC) and dissolved organic carbon (DOC) varied inversely during the red tide, but the patterns are counterintuitive to those expected during a phytoplankton bloom, where DIC should be taken up and DOC produced. The patterns infer that DOC was assimilated during the bloom. Alternative explanations can be eliminated when the results of bioassays are known.
Task 4: Apply neural network (NN) and principal component (PCA) analyses to the monitoring data to determine relationships between environmental properties and phytoplankton pigment composition.
Phytoplankton communities from the two study sites (ACE and NI estuaries) were sampled along a salinity gradient to determine the effects of nutrient loads on their spatial and temporal distribution. The sites differ in size of their watersheds and their nutrient loads. Phytoplankton community composition was assessed using HPLC, multivariate statistics, and NN analyses. NNs were trained to relate HPLC inputs to environmental outputs (e.g., light attenuation coefficient, salinity, dissolved organic carbon). Sensitivity analysis was used to determine which of the 17 HPLC pigments contributed to training the NN. The Shannon diversity index, principle component, and cluster analyses revealed that ACE phytoplankton, receiving higher amounts of nitrogen and phosphate, had a lower pigment diversity in the spring and were compositionally different from NI phytoplankton, particularly during the summer time bloom (July to October). Higher nutrient loads in ACE estuary yielded high concentrations of chlorophyll a, alloxanthin, peridinin, chlorophyll c1, chlorophyll c2, and diadinoxanthin while lower nutrient loads in NI estuary yielded high concentrations of chlorophyll b, prasinoxanthin, and violaxanthin. The trained NNs predicted light attenuation (R2 equals 0.92), salinity (R2 equals 0.82), organic and total suspended solids (R2 equals 0.82, R2 equals 0.85, respectively), and temperature (R2 equals 0.80) with high accuracies. However, the NN predicted other environmental parameters poorly (dissolve organic carbon, R2 equals 0.63; ammonia, R2 equals 0.60). Sensitivity analysis revealed that: (1) the concentration of 19' butanoyloxyfucoxanthin increased, and the concentration of chlorophyll c2 decreased, with increasing light attenuation values; (2) the concentration of alloxanthin, total carotene, and peridinin increased with decreasing salinity; (3) the concentration of lutein and 19' butanoyloxyfucoxanthin increased with organic and total suspend solid concentrations; and (4) the concentration of 19' butanoyloxyfucoxanthin, prasinoxanthin, 19' hexanoyloxyfucoxanthin, and total carotene decreased, and fucoxanthin, violaxanthin, diadinoxanthin, zeaxanthin and lutein increased with increasing temperature. The results suggest that high nutrient loads affect the composition of phytoplankton communities of southeastern tidally dominated estuaries, and that sensitivity analysis of trained NNs revealed the association of specific pigments to environmental parameters.
Task 5: Sample several sites in NI and the ACE Basin (estuaries that differ in fertility) during the phytoplankton bloom period (summer) to determine the relationship between phytoplankton and nutrient quantity and quality.
See Task 4 for conclusions about differences between sites. Further, some differences in the HPLC pigment profiles between the ACE and NI estuaries are noteworthy. For example, zeaxanthin (a carotenoid found in cyanobacteria) and alloxanthin (a marker carotenoid for cryptophytes), which generally made up a similar proportion of the biomass in both estuaries in July, contributed a lower percentage of pigment biomass in ACE basin than in North Inlet in September, while the contribution of fucoxanthin, a carotenoid considered a marker for diatoms (this must be confirmed by microscopy), relatively was greater in ACE Basin in September. These patterns are consistent with the dogma that diatoms generally respond favorably to high nutrient (especially NO3) inputs, while marine cyanobacteria (Synechococcus spp.) are adapted to oligotrophic conditions.
Task 6: Obtain and classify remotely sensed imagery of the study area (to coincide with field sampling) during the spring and late summer.
The original proposal called for the acquisition of Thematic Mapper remotely sensed imagery of the North Inlet-Winyah Bay study area during periods corresponding to peak biomass and chlorophyll concentrations. Following preliminary data analysis of biophysical characteristics of several estuaries in SC, the decision was made to expand the study area also to include portions of the ACE Basin estuary located south of Charleston, SC. High spatial resolution (70cm), multispectral ADAR images of the ACE Basin, SC study area were taken on September 23, and of the NI, SC study area on November 4, 1999, and in both areas during October 2000. The ADAR System acquires digital photographs in four configurable spectral bands between 400 nm to 1,000 nm. Spectrotech, Inc. acquired three meter, 37-band multispectral imagery of the study area during September 2001. Although sacrificing spatial resolution, increased spectral resolution should aid in the development of more robust regression models. These data will be obtained and analyzed during 2002.
Task 7: Collect spectral response data and plant, biophysical data for use in developing relationships between chlorophyll and spectral reflectance of Spartina alterniflora and productivity models.
It was hypothesized that mathematical relationships could be developed between in situ biophysical data (e.g., standing chlorophyll biomass, CO2 exchange) and the brightness values associated with the remotely sensed imagery. In situ radiometric and biophysical (chlorophyll, biomass) data were obtained at approximately 50 locations within each of the Spartina alterniflora-dominated study areas. Canopy radiance data, leaf area index (LAI), chlorophyll and CO2 exchange data were collected in the ACE Basin study area during July, August, and September. Data were collected in the NI study area during July and November. Canopy radiance data were collected using a 350 nm to 1,050 nm field spectroradiometer. The geographical (x,y) location of each sample was recorded using a Global Positioning System (GPS). These biophysical data will be correlated with canopy radiance data collected using the field spectrometer and with remotely sensed data. Spectrophotometer scans of plant leaves from experimentally fertilized plots have demonstrated that significant differences exist between treatments in reflected light.
Task 8: Correlate biophysical and remote sensing data.
A long-term objective of this project is to develop linkages between remotely sensed images of coastal wetlands, fertility or nutrient status, and the chlorophyll biomass of the emergent wetland plant community.
Regression models have been tested using the ADAR remote imagery and canopy chlorophyll datasets. When a successful model is developed, it will be possible to produce "greenness" maps depicting the functional health and productivity of the estuarine wetland vegetation. This will provide baseline data for long-term monitoring of coastal estuarine wetlands, and the techniques being developing should be applicable to other coastal regions of the United States and around the world having similar phenological characteristics. Simple regression models using the ADAR datasets that were obtained do not yield satisfactory results. Better results are expected when the hyperspectral data are analyzed from a flight that occurred over NI in the fall of 2001 (data expected in mid-2002). Testing of the utility NN analysis has begun as a means of classifying remote imagery, and the preliminary results indicate that NN analysis is a far more powerful approach than conventional regression models.
Task 9: Maintain databases and develop and maintain metadata.
All biophysical field data collected by the remote sensing group has been entered into Excel spreadsheets, Quality Assurance/Quality Control performed, and backups generated. Spectral response data have been downloaded, reviewed, and backups generated. GPS data collected at each sampling location has been entered into our Geographic Information Systems (GIS) and backed up on CD.
Future Activities:
With respect to the first component of this project, nutrient controls on phytoplankton and bacterial community composition, the strategy during the final year is to summarize our results in publications. A major product will be a paper by Noble et al. that relates environmental variables to HPLC pigment profiles and taxonomic affiliations. That paper is being revised now.
The landscape-scale, remote sensing component of the project will focus on analyzing the biophysical and remotely sensed data and preparing publications. A new hyperspectral dataset will be available in fall 2001, to analyze, and the use of NN analysis will be explored to classify the remote images.
Journal Articles on this Report : 8 Displayed | Download in RIS Format
Other project views: | All 63 publications | 15 publications in selected types | All 15 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
DeLorenzo ME, Lewitus AJ, Scott GI, Ross PE. Use of metabolic inhibitors to characterize ecological interactions in an estuarine microbial food web. Microbial Ecology 2001;42(3):317-327. |
R826944 (2000) R826944 (2001) R826944 (Final) |
Exit Exit |
|
Jensen JR, Schill SR. Bidirectional reflectance distribution function (BRDF) characteristics of smooth cordgrass (Spartina alterniflora) obtained using a Sandmeier field goniometer. GeoCarto International 2000;15(2):23-30. |
R826944 (2001) |
Exit Exit Exit |
|
Lewitus AJ, Koepfler ET, Pigg RJ. Use of dissolved organic nitrogen by a salt marsh phytoplankton bloom community. Archiv fur Hydrobiologie Special Issues, Advances in Limnology 2000;55:441-456. |
R826944 (1999) R826944 (2000) R826944 (2001) R826944 (Final) |
Exit |
|
Lewitus AJ, Hayes KC, Willis BM, Burkholder JM, Glasgow HB, Holland AF, Maier PP, Rublee PA, Magnien R. Low abundance of the dinoflagellates, Pfiesteria piscicida, P. shumwayae, and Cryptoperidiniopsis spp., in South Carolina tidal creeks and open estuaries. Estuaries 2002;25(4):586-597. |
R826944 (2000) R826944 (2001) R826944 (Final) R827084 (Final) |
Exit |
|
Morris JT, Sundareshwar PV, Nietch CT, Kjerfve B, Cahoon DR. Responses of coastal wetlands to rising sea level. Ecology 2002;83(10):2869-2877. |
R826944 (2000) R826944 (2001) R826944 (Final) R828677 (2001) R828677 (Final) R828677C003 (2003) R828677C003 (Final) |
Exit Exit Exit |
|
Noble PA, Almeida JS, Lovell CR. Application of neural computing methods for interpreting phospholipid fatty acid profiles of natural microbial communities. Applied and Environmental Microbiology 2000;66(2):694-699. |
R826944 (1999) R826944 (2000) R826944 (2001) R826944 (Final) R824776 (Final) |
Exit Exit Exit |
|
Van Heukelem L, Thomas CS. Computer-assisted high-performance liquid chromatography method development with applications to the isolation and analysis of phytoplankton pigments. Journal of Chromatography A 2001;910(1):31-49. |
R826944 (2000) R826944 (2001) R826944 (Final) |
Exit Exit |
|
Wetz MS, Lewitus AJ, Koepfler ET, Hayes KC. Impact of the eastern oyster Crassostrea virginica on microbial community structure in a salt marsh estuary. Aquatic Microbial Ecology 2002;28(1):87-97. |
R826944 (2000) R826944 (2001) R826944 (Final) |
Exit Exit |
Supplemental Keywords:
estuarine eutrophication, remote sensing, productivity, nutrients, phytoplankton community, HPLC pigments, bacteria, gas exchange, canopy chlorophyll, neural net, RFA, Scientific Discipline, Water, Ecosystem Protection/Environmental Exposure & Risk, Water & Watershed, Nutrients, Ecology, Ecosystem/Assessment/Indicators, Ecosystem Protection, Chemistry, Ecological Effects - Environmental Exposure & Risk, Monitoring/Modeling, Environmental Monitoring, Ecological Risk Assessment, Biology, Watersheds, anthropogenic stress, aquatic ecosystem, coastal ecosystem, eutrophication, hydrological stability, nutrient supply, nutrient transport, remote sensing, ecological exposure, monitoring, scaling, chlorophyl, CISNet, wetland vegetation, estuaries, bacteria monitoring, coastal zone, remote sensing data, landscape-scale monitoring, esturarine eutrophication, phytoplankton nutrient, CISNet Program, phytoplankton dynamics, landscape monitoring, analytical chemistry, molecular monitoring, aquatic ecosystems, nutrient cycling, water quality, ecosystem health, stress responses, estuarine food webRelevant Websites:
http://zebra.biol.sc.edu/ Exit
Progress 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.