2003 Progress Report: Coastal Wetland Indicators

EPA Grant Number: R828677C003
Subproject: this is subproject number 003 , established and managed by the Center Director under grant R828677
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).

Center: EAGLES - Atlantic Coast Environmental Indicators Consortium
Center Director: Paerl, Hans
Title: Coastal Wetland Indicators
Investigators: Morris, James T. , Herrick, Gabe , Hopkinson, Charles S , Marshall, Helen , Novakowski, Karyn I. , Rodriguez, Diana , Torres, Raymond , Valentine, Vinton
Current Investigators: Morris, James T. , Gallegos, Charles L. , Herrick, Gabe , Hopkinson, Charles S , Marshall, Helen , Montane, Juana M. , Novakowski, Karyn I. , Rodriguez, Diana , Torres, Raymond , Valentine, Vinton
Institution: University of South Carolina at Columbia
Current Institution: Marine Biological Laboratory
EPA Project Officer: Hiscock, Michael
Project Period: March 1, 2001 through February 28, 2005
Project Period Covered by this Report: March 1, 2002 through February 28, 2003
RFA: Environmental Indicators in the Estuarine Environment Research Program (2000) RFA Text |  Recipients Lists
Research Category: Ecological Indicators/Assessment/Restoration , Water , Ecosystems

Objective:

The objectives of this research project are to: (1) develop a suite of indicators of the condition of coastal wetlands that are based on physical and biological criteria with emphasis on higher plant-based pigment indicators and (2) link these indicators to remote sensing capabilities.

Progress Summary:

The Atlantic Coast Environmental Indicators Consortium (ACE INC), Coastal Wetland Indicators component has been active since August 2001. In the initial months of the project, most of the activity was organizational. Two postdoctoral fellows have been hired: Helen Marshall, an expert on plant pigments and bio-optical modeling with a Ph.D. from the University of Wales, and Vinton Valentine with a Ph.D. from the University of Delaware. Dr. Marshall is working on a major objective of our project—to develop a suite of indicators of the condition of coastal wetlands that are based on biophysical criteria. Dr. Valentine brings to the project expertise in image analysis and geographic information systems. The structure of the program, its elements, and principal scientists remains as diagrammed (see Figure 1).

Pigment Indicators Progress

Plant pigments and the spectrum of reflected light from plant leaves are variables that make attractive indicators as they clearly have application to remote sensing. Results of scanning with a spectroradiometer showed that P-treated plants (Spartina alterniflora) had significantly higher reflectance in the NIR, irrespective of N treatment, in a spectral region that is largely determined by cell structure. The observed differences in spectral reflectance should be great enough to be detectable by remote sensors and could provide a means of monitoring nutritional status.

Schematic of the Structure, Elements, and Principal Scientists of Coastal Wetland Indicators

Figure 1. Schematic of the Structure, Elements, and Principal Scientists of Coastal Wetland Indicators

We also have made progress in relating the density of chlorophyll in the plant canopy to the reflectance data in remote imagery, and we have successfully trained a neural network (NN) to classify remote imagery. With great success, the NN was able to map chlorophyll density as well as plant community distributions and major landscape features. This is an important step, because chlorophyll concentration provides information about a plant’s condition. The concentration of chlorophyll in plant tissues varies with phenology and nutrition. Because photosynthetic rate and chlorophyll a (Chl-a) concentration are directly related (Bokari 1983), Chl-a is actually a more sensitive indicator of the condition of higher plants than biomass and should be investigated as an index of stress. Accessory pigments, measured by HPLC, provide even more information about the condition of plants. Further, because Chl-a is highly absorbent of radiation in the range of Landsat Thematic Mapper spectral band 3 (630-690 nm) and reflective in spectral band 4 (760-900 nm), it should be feasible to use remote sensing techniques to monitor the condition of vegetation and the density of pigments in the plant canopy (see Figure 2). At North Inlet previous attempts to remotely sense pigments have met with some success (see Figure 4) using spatially precise ADAR data, but our experience has shown that hyperspectral data is needed to make significant advances in the remote sensing of plant pigments. We have great success, however, in training NN to interpret remote data, and we expect that significant progress will be made using NN to interpret our new hyperspectral data. Continued development of the NN also has allowed very favorable comparisons with the currently popular CHEMTAX model (for assessing phytoplankton populations from HPLC analysis of water samples). The NN can already reproduce the results of CHEMTAX, and further development is addressing current problems with CHEMTAX (e.g. variability of pigment ratios in the phytoplankton).

An ADAR Image, Classified to Show Chlorophyll Density in a Salt Marsh at North Inlet

Figure 2. An ADAR Image, Classified to Show Chlorophyll Density in a Salt Marsh at North Inlet.

Dr. Marshall also is looking at xanthophyll pigments, their role and reflectance. These pigments react rapidly to a variety of environmental stressors in both microalgae and higher plants. The xanthophyll cycle directly affects the rate of carbon fixation and occurs in reaction to environmental stress. This cycle causes changes in the proportion of different wavelengths of light reflected by the photosynthetic apparatus and as such can be used as a remote sensing indicator to assess nutrient availability and the rate of primary production in a variety of ecosystems. Previously, experiments had been performed at the Plum Island Estuary in Massachusetts and North Inlet in South Carolina. Reflectance was analyzed around 530 nm wavelength and the change from a ‘no feature’ was calculated and named delta reflectance. Delta reflectance correlated linearly with the epoxidation state of the xanthophyll cycle at both single leaf and plant canopy levels. In this way the environmental stress that salt marsh plants are under can be remotely sensed using the reflectance as an indicator. Work in the past year has focused on obtaining hyperspectral data from airborne AISA sensors, while sampling biophysical variables on the ground during the flyovers. Hyperspectral flight areas included North Inlet, Murrell’s Inlet, ACE Basin, and Grand Bay, and all ground sampling was conducted within a day of the flyover. Addition of a PAM fluorimeter and an infra-red temperature probe to the sampling equipment has allowed rapid photosynthesis and leaf temperature measurements to be taken to complement the suite of plant measurements already sampled. Sample species also has been widened to include common salt marsh species such as Juncus species and Typha species. Photosynthetic production and reflectance has continued to be found to be highly related to environmental stress, and final development of the pigment indicator is to produce an algorithm that is suitable for inclusion into a model that uses the NIR reflectance to separate out species (by morphological characteristics), and then calculates stress from those species between 530-550 nm reflectance.

To obtain a wide range of conditions, a ‘stress database’ has been constructed to isolate different stress factors. Data sets include large scale transects across the marsh, small-scale marsh features (see Figure 3), and laboratory experiments. This should allow the indicator not only to predict stress but also to provide information on the type of stressor.

Examples of Marsh Habitats Selected for Stress Studies

Figure 3. Examples of Marsh Habitats Selected for Stress Studies

Morphology of the marsh platform is being taken into account, and very good results have been found when matching the elevation of the marsh surface to plant stress and productivity. There appears to be an ‘optimal’ height above sea-level for marsh plan growth; however, changes in this height lead to changes in morphology as well as physiology, and so careful interpretation of reflectance data is required. Inclusion of PAM fluorescence measurements has helped in this and has added greatly to data interpretation.

Several conditions/controls also have been specified for effective use of the indicator. An important example is the need to standardize the time of day at which the indicator is optimal. Experiments have shown that light is an important factor in the pigment content of the plants, greatly affecting their reflectance (see Figure 4). To remove the effects of light, a standardized sampling protocol has been developed with regards to ‘time of day’ and ‘light intensity’ to remove possible artifacts.

Examples of Marsh Habitats Selected for Stress Studies

Figure 4. Diurnal Delta Reflectance Changes in S. alterniflora

Neural Network Analysis of Phytoplankton Pigments

We have made significant progress developing NN models as an alternative approach to the traditional CHEMTAX approach of analyzing pigment data. CHEMTAX is slow, it is unable to incorporate ancillary data, it assumes invariant pigment ratios, and it requires proprietary software (Matlab). The approach we have taken is to develop synthetic pigment data based on empirical knowledge of pigment ratios. The synthetic data incorporate the expected range of variability in pigment ratios. We have written software that uses a pigment matrix as a starting template, and then applies different degrees of variability to the ratios to simulate a collection of samples. Samples are generated to train the NN, a separate set is generated to use as a validation set, and a third set is used as the test set. We analyze the test using the trained NN as well as CHEMTAX, and the comparison shows that the NN generally outperforms Chemtax. Our results also demonstrate that pigment ratios are highly successful as indicators of most algal taxa, but not others, regardless of the method.

Tidal Drainage Network Mapping – Plum Island

Our research on marsh stability in northern marshes focused on (1) mapping and classification of marsh drainage networks; (2) calculation of geomorphometric and fractal dimension measures; and (3) examination of marsh sedimentation and pond metabolism.

We continued our collaboration with Dr. Thomas Millette at Mount Holyoke College to digitize drainage network features (channels, mosquito ditches, ponds, and pannes) of the Plum Island tidal marshes. We involved students by integrating the project into coursework and by leveraging work-study opportunities. We used 4 km x 4 km color orthophoto images, at 0.5 m resolution, from the on-line Massachusetts Geographic Information System (MassGIS) as the image base for the mapping effort. As of February 2004, digitization of channels, mosquito ditches, and ponds for most of the twelve tiles covering the area are complete. We are performing quality assurance, including final field checks, on the data completed to date, prior to their inclusion in the geomorphometric analyses.

We were successful in the identification, location, and assessment of historical aerial photography encompassing a number of years. In particular, we discovered the existence of photography from 1938, which will provide a detailed mapping source to support change analysis over a 60-year interval.

We were successful in obtaining supplemental funding for digital elevation model (DEM) generation of the marsh platform using either photogrammetry or lidar. We are in the process of contacting vendors to discuss capabilities and requirements and to obtain quotes. The data will be useful in developing basins boundaries and in investigating geomorphometric indicators that rely on relief measures, such as local relief and hypsometric interval.

Geomorphometric Indicators – Plum Island

To begin our investigation of geomorphometric indicators for tidal marsh condition, we have selected three basins within the Plum Island Estuary that appear to represent different development states. We established basin boundaries from a cost distance surface using terrain analysis modules in a geographic information system (GIS). Figure 5 displays and describes the basins and their networks.

Study Basins and Networks

Figure 5. Study Basins and Networks

After converting the channel and mosquito ditch polygon data to a centerline network, we calculated the following geomorphometric measures: drainage density (Dd), constant of channel maintenance (C), and length of overland flow (lo). We selected these metrics to study because they do not require stream ordering, which frequently fails in tidal drainage systems. The data in Table 1 agree with our impressions of the basin states.

Table 1. Geomorphometric Measures for Three Tidal Creek Basins in Plum Island Estuary.

Geomorphometric Measures for Three Tidal Creek Basins in Plum Island Estuary

The basin perceived to be undergoing active degradation (Shad) has the highest drainage density, the lowest constant of channel maintenance, and the lowest length of overland flow. The basin with the most stable marshes (Upper Parker) has the lowest drainage density, the highest constant of channel maintenance, and the highest length of overland flow. The Club Head basin represents a state between the above extremes and its measures correspond with this characterization. Given these results, the above indicators, in concert with other measures to be investigated (for example, number and mean length of headwater reaches, density and distribution of ponds and pannes), appear to provide useful information regarding tidal marsh condition.

Fractal Dimension

Fractal dimension of the tidal drainage network is another measure that we are investigating. Fractal dimension gives us an idea of the scaling behavior of objects. In the context of drainage networks, calculating this metric reveals at least two different scaling ranges, indicating multi-fractal behavior. The values associated with the ranges describe characteristics of the individual channels and of the entire network. The metric also may provide a more objective measure of channel sinuosity as well as an indication of the scales, and possibly types, of processes acting on the networks.

We calculated the fractal dimensions for the three basin centerline networks using the box counting technique. Figure 6 and Table 2 present our results. The patterns are similar to ones found for terrestrial drainage networks. The fractal dimension found at high resolutions, Ds, applies to the individual channels and may reveal information about channel sinuosity. The fractal dimension found at low resolutions, Db, describes information about the branching of the total network and may provide information about how the drainage density is distributed through the system and therefore, network type classification.

Plots of Box Counts and Sizes With Fractal Dimensions for Three Tidal Creek Basins in Plum Island Estuary

Figure 6. Plots of Box Counts and Sizes With Fractal Dimensions for Three Tidal Creek Basins in Plum Island Estuary.

Table 2. Fractal Dimensions for Three Tidal Creek Basins in Plum Island Estuary

ractal Dimensions for Three Tidal Creek Basins in Plum Island Estuary

The high resolution, fine scale fractal measure, Ds, shows minor differences for all three basins. The low value of Shad may reflect the mapping technique where many of the channels terminated in headwater ponds which were mapped as straight reaches. Club Head reveals the highest value which may reflect the observation that all channel reaches exhibit some curvature. The Upper Parker has a high number of mosquito ditches which should give a relatively low value, but in this case, the value falls in the middle. These values, however, are preliminary because subjectivity within the technique requires modification to insure robust measures.

The low resolution, broad scale fractal measure, Db, shows more notable differences. The low value for the Upper Parker corresponds with a relatively low drainage density basin that has large marsh expanses punctuated with locations of extensive mosquito ditching. The high value for Club Head fits well with a moderate drainage density basin having a more evenly distributed network. The Shad basin’s value reveals a combination of the characteristics described for the other two basins. Given our promising results to date, further investigation of the fractal dimension of tidal creek networks is warranted.

Marsh Sedimentation and Pond Metabolism – Plum Island

We complemented our work in indicators research with two Research Experience for Undergraduates students from the Boston University Marine Program to explore in detail some causes of marsh development and degradation. Morgan Johnston investigated the importance of metabolism in salt marsh pond development within the Plum Island Ecosystems Long-Term Ecological Research (PIE-LTER). She measured dissolved oxygen changes in open pond water (see Figure 7) and oxygen consumption in sediment cores. Her results showed respiration generally exceeded gross primary production, indicating net consumption of former marsh peat. She found that peat respiration can account for 30-60 percent of the increase in pond depth over the past 50 years.

Diurnal Patterns of Dissolved Oxygen (DO) Over a 3-Day Period

Figure 7. Diurnal Patterns of Dissolved Oxygen (DO) Over a 3-Day Period

Jason Cavatorta studied marsh sedimentation patterns in the PIE-LTER site. He measured accumulated sediment along transects from channel edges to the interior marsh and sampled total suspended solids (TSS) in the adjacent channels. He found that TSS (see Figure 8) and sedimentation in the estuary’s main stem decreased from the upper reaches, near presumed sediment sources, towards the sound. He noted that TSS in sub-watershed channel networks generally decreased from low-order to high-order channels, suggesting some internal sediment sources. He also noted that sedimentation in the sub-watersheds decreased with distance from the estuary’s main stem, which is presumably the primary sediment input to the marshes.

Suspended Solids (TSS) Down Estuary From Dam in Upper Reaches

Figure 8. Suspended Solids (TSS) Down Estuary From Dam in Upper Reaches

Geomorphic Analyses of Intertidal Creek Network Structure- North Inlet

Two of our graduate students, Juana Montane and Karyn Novakowski, are investigating questions relating to the relationship between marsh geomorphology, habitat disturbance and stability. The study area is a 8.67 km2 portion of the North Inlet intertidal marsh. In this mapped region the deeper and wider intertidal channels circumscribe 56 marsh islands. We delineated all the intertidal channel networks and found 725 discrete creek systems. Maximum channel lengths and associated network watershed areas ranged from 3-10,654 m, and 127-609,418 m2, respectively. Approximately 94 percent of the networks had a terrestrial network appearance; this means that in the downstream direction channel segments converge toward a main channel. On the other hand, approximately 9.1 km or 6 percent of total channel length was classified as being part of a reticulating drainage system.

Stream Order

Applying Horton’s Law’s of drainage composition, order (w) versus number of stream and mean stream length (lm) revealed an exponential trend that was similar to those found in terrestrial systems (see Figure 9). Some of the terrestrial and intertidal trends in these data overlie each other despite differences in scale from which the data points were extracted and processes controlling network development. The high correlation coefficient, r2, of 0.95 (data from North Inlet) indicates that the exponential model adequately describes changes in stream frequency with order.

Relation of Stream Order to the Number of Streams for Terrestrial and Marsh Networks. Marsh data are open symbols; terrestrial data are closed.

Figure 9. Relation of Stream Order to the Number of Streams for Terrestrial and Marsh Networks. Marsh data are open symbols; terrestrial data are closed.

Mean stream segment length exponentially increases with order (Horton’s Law of Stream Order) although there were distinct differences in the trends between terrestrial and coastal networks (see Figure 10). For example, the terrestrial data illustrate a nearly uniform positive slope and higher r2 values. The coastal data on the other hand are different in two ways: First, the Wadsworth (1980), Pestrong (1965) and Pethick (1980) data show a uniform slope between orders 1- 4, well described by an exponential function, with slopes comparable to the terrestrial systems. Secondly, the slope of the Myrick and Leopold (1963) data is greater than the terrestrial trend slope, whereas the slope for the North Inlet data is lower. These exponential relationships (see Figure 9, 10) also are consistent with Horton’s Law of stream numbers and lengths, and they are consistent with expectations from Kirchner (1993) who showed that Horton ratios are insensitive to network structure.

Relationship of Stream Magnitude to Number of Stream Segments

Figure 10. Relationship of Stream Magnitude to Number of Stream Segments

Link Magnitude

Stream frequency declined (see Figure 10) as a power function from magnitude 1 to magnitude 33, but variability increased substantially at higher magnitudes. Over the range of magnitude 34 to 93, there were 14 link magnitude values that had a single occurrence, and 12 channel networks with magnitude greater than 93 had a single occurrence. On the other hand, mean channel length shows an increase over the same range in magnitude, but the scatter about the curve fit increases substantially above magnitude 33. Channel segments of magnitude 1-6 comprise 81 percent of total channel length and 89 percent of total channel segments counted. The rate of change in mean channel length, however, decreases to approximately 1 m per increase in magnitude between magnitude 6 to magnitude 17, but the scatter about the curve fit increases substantially at greater magnitude (see Figure 11). This trend indicates that a “limit” to mean channel length, or distance between tributaries, may exist in networks with link magnitude above 6 to 17.

The topologically random network model developed by Shreve (1966) and Smart (1968) provides the theoretical framework for the prediction of drainage network composition (e.g. Horton’s and Hack’s Laws). As Shreve (1974) stated, a topologically random population of channel networks is one in which all topologically distinct channel networks of the same magnitude are equinumerous. Random topology in North Inlet salt marsh creek networks was evaluated using the equinumerous test. We evaluated the occurrence of each possible topological sequence for a given magnitude. A chi-squared test for goodness-of-fit of observed to expected counts was performed. All chi-square values were less than expected, therefore for magnitudes 3, 4, and 5 we cannot reject the hypothesis that marsh creek networks are topologically random.

Relation of Stream Order With Stream Length. Marsh data are open symbols; terrestrial data are closed

Figure 11. Relation of Stream Order With Stream Length. Marsh data are open symbols; terrestrial data are closed

Channel Length-Watershed Area Analyses

For topologically distinct channel networks the length-area (L-Aw) relationship determined by Hack (1957) is expected from theoretical predictions by Shreve (1974). L-Aw analyses for 725 non-reticulating channel networks were conducted (black diamonds in Figure 12). The data vary over four orders of magnitude in watershed area Aw, and three orders of magnitude in mainstream length L. Scatter in the data declines with increasing basin area. A regression line through the L-Aw data with a r2 of 0.79, gives the power function L = 0.08Aw0.73. For comparison to terrestrial systems, these data are plotted over a shaded region that depicts the upper and lower limits of terrestrial values for slope and y-intercept reported in the literature and projected to our scale of observation. In this case, Willimen (2000) defines these upper and lower limits, with y-intercepts and exponents of 2.3 and 0.5, and 0.073 and 0.60, respectively. Approximately 86 percent of North Inlet marsh data fall within the range of the reported terrestrial values. Most of the data that lie outside of the terrestrial region lie below the terrestrial lower limit. The regression line, however, falls completely within the terrestrial region. Hence, the intertidal creek networks and accompanying watershed areas appear to adhere to L-Aw relations postulated by Hack (1957), and predicted by Shreve (1974). On the other hand, because the intertidal L-Aw exponent is greater than those reported for terrestrial systems, we surmise that intertidal networks are more elongate than terrestrial networks.

Field observations show that well-developed and discrete channel networks and basin areas develop, and persist in intertidal environments. During tidal inundation, when water may be exchanged between discrete tidal basins, exact basin or watershed boundaries defined by topographic highs are irrelevant to geomorphic analysis. Hence, topographic flow divides, a well-established feature of terrestrial landscapes, typically do not exist in intertidal landscapes. Consequently, we developed an alternative procedure to further refine the Hack relationship in the intertidal zone; the procedure is based on island area as opposed to basin area. We propose that in some intertidal environments, the unit of geomorphic analysis should be the island instead of the watershed or basin area. Using this approach, a semi-logarithmic plot of island area (AI) as the dependent variable against maximum mainstream length (Lmax) of an island, for all 56 islands (see Figure 13) are related by the power function L = 0.11A0.65. These data exhibit a power function relationship that is similar to the Hack relationship for terrestrial landscapes, with a r2 equal to 0.77.

Relation of Stream Length With Watershed Area, Gray Area Depicts Terrestrial Range

Figure 12. Relation of Stream Length With Watershed Area, Gray Area Depicts Terrestrial Range

Relation of Stream Length With Island Area

Figure 13. Relation of Stream Length With Island Area

Drainage Density

Drainage density, Dd, is defined as l/A, where l = total channel length of the region of interest, and A = the corresponding area of the region. Drainage density values for all 725 marsh watersheds ranged from 0.0008 m/m2 to 0.069 m/m2 with an average value 0.013 ± 0.009 m/m2 (see Figure 14). For comparison, some terrestrial drainage density values range between 0.0023-0.0137 m/m2 ( Ritter, et al., 1995); 57 percent of the marsh drainage density values fell within this terrestrial range (see Figure 12). It is important to note the effect of map scale on drainage density calculations ( Gardiner, et al., 1977; Tucker, et al., 2001) because stream lengths and areas determined on large scale maps may only include larger streams while neglecting the smaller tributaries. This may result in an underestimation of the total length for a given area, and therefore, artificially low values of drainage density. The North Inlet drainage density data were collected on an image with 0.7 m spatial resolution. Because of this fine-scale, it is unlikely that a significant number of creek segments in a given area were overlooked. If the terrestrial data were collected on finer scale maps, the terrestrial drainage density range might widen, possibly including more of the marsh drainage density values.

Pethick (1992) hypothesized that intertidal drainage density is inversely related to watershed area. This relationship is supported by the intuitive argument that a plot of drainage density versus basin area amounts to a plot of A-1 versus A, giving rise to an inverse relationship. The relationship between total length for each creek network plotted against watershed area can reveal specific information about the variability in drainage density and rate of network development (e.g. Marani, et al., 2003). Our data show a great degree of variability in drainage density for a given watershed area (see Figure 15). While scatter on this plot increases as area increases, for basins with area greater than approximately 2,000 m2 the upper limit to the plot defines a nearly straight line with a slope of 1.0 and uniform drainage density of 0.025 m/m2. This indicates that for larger basins, total channel length can be expected to increase up to a predictable length determined by a power function with a unit exponent. Moreover, the regression through these data gives the power function l = 0.03Aw0.88. The slope of the line, 0.88, is similar to the values reported by Steel and Pye (1997) and the assumed unit value of Marani, et al. (2003).

Drainage Density for all Watersheds

Figure 14. Drainage Density for all Watersheds

The Relationship Between Drainage Density and Watershed Area in North Inlet

Figure 15. The Relationship Between Drainage Density and Watershed Area in North Inlet.

Topographic Analyses of the Intertidal Marsh Landscape-North Inlet

We measured the topography of a marsh island using RTK GPS and airborne LiDAR data using an Optech ALTM 11210 system flying at 1,190 m. Instead of using an interpolation approach for testing LiDAR elevations at reference points, we georeferenced the LiDAR data and used the RTK GPS to navigate to a statistically significant number of points necessary to attain 95 percent confidence in LiDAR derived elevations. Assuming the GPS data represent ground truth, we evaluated the accuracy of LiDAR and created an empirical error budget for “last return” signals. RMSE of LiDAR elevation measurements are 0.06 m. These observations show that LiDAR remote sensing of elevation in salt marsh environments is a robust method for creating DEMs and subsequent geomorphic analyses. Currently, we are evaluating the three-dimensional structure of the North Inlet salt marsh.

Contributions to the State of Knowledge

The indicators that we are developing will provide new tools for evaluating the condition of coastal wetlands. The actual products will be indicators that are based on measurements made in the field. All the indicators being developed have a significant potential for being developed as applications that can be calibrated using remotely sensed data. To date, (1) Progress has been made using pigments and reflected light as indicators of the condition of vegetation. (2) Neural networks have proven to be effective tools for classifying remote sensor data and quantitatively classifying phytoplankton pigment data. (3) Significant trends in the productivity of coastal wetlands have been observed. (4) We have documented that we are able to discern interannual changes in the relative elevation of the marsh surface.

Products

Databases. All research results are posted on an annual basis on the Plum Island Ecosystems LTER database. Each data file is fully documented in separate metadata files. Work on the GLEI/EPA metadata site has now commenced, and all metadata for the Coastal Wetlands group will shortly be up on the GLEI server. (Metadata manager – Dr Helen Marshall)

Software. Our project has generated a powerful software package that generates synthetic data for training neural networks and for testing the Chemtax and Neural Network models.

Future Activities:

We willcontinue to explore the use of pigment and geomorphic indicators for a broader suite of conditions and systems. Collaborations are underway with EaGLe Consortium for Estuarine Ecoindicator Research for the Gulf of Mexico and Atlantic Slope Consortium scientists to compare and expand the indicators being evaluated in this project. Future activities will emphasize these collaborations in the context of the EaGLe and other (e.g. NSF-LTER) research efforts aimed at long-term, cross-ecosystem indices of coastal wetland condition and change. Our future activities are focused on investigating the development of optimal channel patterns and form that facilitate water and nutrient exchange between subtidal and intertidal environments.


Journal Articles on this Report : 9 Displayed | Download in RIS Format

Other subproject views: All 89 publications 19 publications in selected types All 17 journal articles
Other center views: All 383 publications 99 publications in selected types All 88 journal articles
Type Citation Sub Project Document Sources
Journal Article Cavatorta JR, Johnston M, Hopkinson C, Valentine V. Patterns of sedimentation in a salt marsh-dominated estuary. Biological Bulletin 2003;205(2):239-241. R828677 (Final)
R828677C003 (2003)
R828677C003 (Final)
  • Full-text: Biological Bulletin-Full Text HTML
    Exit
  • Other: Biological Bulletin-PDF
    Exit
  • Journal Article Huang X, Morris JT. Trends in phosphatase activity along a successional gradient of tidal freshwater marshes on the Cooper River, South Carolina. Estuaries 2003;26(5):1281-1290. R828677C003 (2003)
    not available
    Journal Article Jensen JR, Olsen G, Schill SR, Porter DE, Morris J. Remote sensing of biomass, leaf-area-index and chlorophyll a and b content in the ACE Basin and National Estuarine Research Reserve using sub-meter digital camera imagery. Geocarto International 2002;17(3):27-36. R828677C003 (2002)
    R828677C003 (2003)
    R826944 (2000)
    R826944 (Final)
  • Abstract: Taylor & Francis-Abstract
    Exit
  • Journal Article Johnston ME, Cavatorta JR, Hopkinson CS, Valentine V. Importance of metabolism in the development of salt marsh ponds. Biological Bulletin 2003;205(2):248-249. R828677 (Final)
    R828677C003 (2003)
    R828677C003 (Final)
  • Abstract from PubMed
  • Full-text: Biological Bulletin-Full Text HTML
    Exit
  • Other: Biological Bulletin-PDF
    Exit
  • Journal Article Morris JT, Sundareshwar PV, Nietch CT, Kjerfve B, Cahoon DR. Responses of coastal wetlands to rising sea level. Ecology 2002;83(10):2869-2877. R828677 (2001)
    R828677 (Final)
    R828677C003 (2003)
    R828677C003 (Final)
    R826944 (2000)
    R826944 (2001)
    R826944 (Final)
  • Full-text: ESA Ecology-Full Text HTML
    Exit
  • Abstract: ESA Ecology-Abstract
    Exit
  • Other: South Dakota School of Mines-Full Text PDF
    Exit
  • Journal Article Mwamba MJ, Torres R. Rainfall effects on marsh sediment redistribution, North Inlet, SC. Marine Geology 2002;189(3-4):267-287. R828677C003 (2002)
    R828677C003 (2003)
    not available
    Journal Article Noble PA, Tymowski RG, Fletcher M, Morris JT, Lewitus AJ. Contrasting patterns of phytoplankton community pigment composition in two salt marsh estuaries in southeastern United States. Applied and Environmental Microbiology 2003;69(7):4129-4143. R828677C003 (2003)
    R826944 (2000)
    R826944 (Final)
    R829458 (2005)
    R829458C004 (2003)
    R829458C004 (2005)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: Applied and Environmental Microbiology-Full Text HTML
    Exit
  • Abstract: Applied and Environmental Microbiology-Abstract
    Exit
  • Other: Applied and Environmental Microbiology-Full Text PDF
    Exit
  • Journal Article Sundareshwar PV, Morris JT, Koepfler EK, Fornwalt B. Phosphorus limitation of coastal ecosystem processes. Science 2003;299(5606):563-565. R828677C003 (2003)
    not available
    Journal Article Valentine V, Hopkinson, Jr. C. Investigating drainage density and fractal dimension as geomorphometric indicators of tidal marsh condition using remotely sensed data and geographical information systems. International Journal of Remote Sensing (in submission, 2004). R828677C003 (2003)
    not available

    Supplemental Keywords:

    coastal wetlands, marsh habitat, higher aquatic plants, photopigments, geomorphology, tidal ecosystems, regional indicators, LIDAR, nutrient status, physiology, sea level rise, neural network analysis, wetland management,, RFA, Scientific Discipline, Air, Water, ECOSYSTEMS, Ecosystem Protection/Environmental Exposure & Risk, RESEARCH, estuarine research, Hydrology, Ecosystem/Assessment/Indicators, climate change, Air Pollution Effects, Aquatic Ecosystems, Monitoring, Ecological Monitoring, Atmosphere, Ecological Indicators, environmental monitoring, remote sensing, coastal ecosystem, bioindicator, plant indicator, coastal watershed, estuaries, coastal environments, diagnostic indicators, ecosystem indicators, environmental indicators, coastal ecosystems

    Relevant Websites:

    http://www.aceinc.org Exit

    http://www.geol.sc.edu/chapman Exit

    http://ecosystems.mbl.edu/PIE Exit

    Progress and Final Reports:

    Original Abstract
  • 2001
  • 2002 Progress Report
  • Final Report

  • Main Center Abstract and Reports:

    R828677    EAGLES - Atlantic Coast Environmental Indicators Consortium

    Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
    R828677C001 Phytoplankton Community Structure as an Indicator of Coastal Ecosystem Health
    R828677C002 Trophic Indicators of Ecosystem Health in Chesapeake Bay
    R828677C003 Coastal Wetland Indicators
    R828677C004 Environmental Indicators in the Estuarine Environment: Seagrass Photosynthetic Efficiency as an Indicator of Coastal Ecosystem Health