EPA/620/R-05/005
                                  September 1995
      COASTAL WETLANDS
        INDICATOR STUDY:
         EMAP-ESTUARIES
LOUISIANIAN PROVINCE -1991
                    by

                R. Eugene Turner
                Erick M. Swenson
               Coastal Ecology Institute
               Louisiana State University
              Baton Rouge, Louisiana 70803

                    and

                J. Kevin Summers
            US Environmental Protection Agency
            Environmental Research Laboratory
               Gulf Breeze, Florida 32561
                                 Printed on Recycled Paper

-------

-------
                                     DISCLAIMER


This report represents data from a single year of pilot study operations of the Environmental Monitoring
and Assessment Program (EMAP). Because the probability-based scientific design used by the EMAP
necessitates multiple years of sampling, there may be significant levels of uncertainty associated with
some of these data. This uncertainty will decrease as the full power of the approach is realized by the
collection of data over several years.  Similarly, temporal changes and trends cannot be reported, as these
require multiple years of observation. Please note that this report contains data from research studies in
only one biogeographic region (Louisianian Province) collected in a short index period (July-August)
during a single year (1991).  Appropriate precautions should be exercised when using this information for
policy, regulatory or legislative purposes.

A reference to a specific manufacturer or product does not indicate or imply endorsement of that product
by the Environmental Protection Agency.
EMAP Draft Report -1994                                                          Page Hi

-------
                                      PREFACE


This document is the first pilot study summary for the Coastal Wetlands component of the Louisianian
Province of the Estuaries component of the U.S. Environmental Protection Agency's (EPA) Environmental
Monitoring and Assessment Program for Estuaries (EMAP-E),

The appropriate citation for this report is:

R.E. Turner, E.M. Swenson, and J.K. Summers. 1995.  Coastal Wetlands Indicator Study: EMAP-
Estuaries Louisianian Province - 1991, U.S. Environmental Protection Agency, Office of Research and
Development, Environmental Research Laboratory, Gulf Breeze, PL.  EPA/620/R-95/005,
EMAP Draft Report -1994	Page iv

-------
                                     Table of Contents
DISCLAIMER	.. iii
PREFACE	iv
TABLE OF CONTENTS	v
ACKNOWLEDGMENTS	V	-..	vii

1 EXECUTIVE SUMMARY					1	.9

2 INTRODUCTION	 15
    2.1  Overview of EMAP 	„	,	15
    2.2  Objectives of EMAP-Coastal Wetlands 	,	15
    2.3  EMAP Framework for Indicator Development	 16
        2.3.1  Biological Integrity	16
        2.3.2  Consumptive Uses  ....		16
        2.3.3  Non-Consumptive Uses	19
        2.3.4  Pilot Testing of Indicators of Wetlands Environmental Values	19
    2.4  Purpose and Objectives of Pilot Study 	19
    2.5  Structure of Report	,22

3 STUDY DESIGN AND INDICATOR SELECTION	23
    3.1  Site Selection	,	23
        3.1.1  Purpose of non-random sites - indicator development	23
        3.1.2  Site selection criteria	23
    3.2  Study Design	,	,	25
        3.2.1  Geographic Scope - with rationale	25
        3.2.2  Location and Timing  of Sampling	25
    3.3  Indicator Selection	27
        3.3.1  General criteria	:	27
        3.3.2  Sampling Index Period	31
        3.3.3  List of indicators chosem	31
    3.4  Sampling Scheme	31
    3.5  Sampling Procedure	38
        3.5.1  Soil Parameters - Field 	38
        3.5.2  Soil Parameters - Lab	 39
        3.53  Vegetation Parameters - Field	40
        3.5.4  Vegetation Parameters - Lab	40
        3.5.5  Other Field Measurements	41
    3.6  QA for Field Sampling and Instrumentation	41
    3.7  Data Analysis	43
        3.7.1  Data Base Creation	 43
        3.7.2  QA/QC	46
        3.7.3  Data Analysis	49
EMAP Draft Report -1994     	    	                         	Page v

-------
                             Table of Contents (continued)

 4 RESULTS AND DISCUSSION	53
    4.1. Completion Rate,  Accuracy and Precision	53
    4.2. Site Classification	63
    4.3. Indicator Variability Within and Among Sample Sites	64
    4.4. Relationships among Indicators 	71
        4.4.1 Stem Morphology and Density	71
        4.4.2 Reproductive Tissues	75
        4.4.3 Soil Conditions and Relationships with Other Factors	75
        4.4.4 Biomass and Spectral Reflectance Differences Among Marsh Health Classes 	79
        4.4.5 Relationships Between Plant Vigor and Spectral Reflectance Indicators	79
    4.5  ANOVA RESULTS 	82
    4.6  Multivariate Results	82
    4.7  Hydrology	86


5 CONCLUSIONS AND RECOMMENDATIONS 	91
    5.1  Response Indicator Development	91
    5.2  Sampling Efficacy	92
    5.3  Additional Indicators	92
    5.4  Expansion to Other Regions	93
    5.5  Summary Table 	93


6 REFERENCES    	97
EMAP Draft Report -1994	Page vi

-------
                             ACKNOWLEDGMENTS


The authors acknowledge the assistance of many individuals in the preparation of this report.  The
personnel at ManTech, Inc. in Corvallis, Oregon; particularly Ms. Louisa Squires, Ms. Kate Dwire, and
Mr. Richard Novitski, who provided many suggestions, criticisms and comments throughout the writing of
this report. Their input is greatly appreciated. This project was funded through the U. S. Environmental
Protection Agency, Environmental Research Laboratory in Corvallis, Oregon (ERL-C), through a
Cooperative Agreement (EPA 818409-01, Spencer Peterson - Project Officer).

The following were involved in the 1991 field sampling:
Ms. I. Hess
Mr. J. M. Lee
Mr. T.Lu
Mr. T. A. Oswald
Mr. G. Peterson
Mr. R. Raynie
Ms. L. Squires
Mr. S. Subbian
Mr. C. M. Swarzenski
Ms. Y. Swarzenski
Mr. E. M. Swenson
Dr. R. E. Turner
Mr. M. Ying

The following contributed to the laboratory analyses:

Ms. L. Brunei, Coastal Ecology Institute, LSU
Mr. C. S. Milan, Coastal Ecology Institute, LSU
Mr. T. A. Oswald, Coastal Ecology Institute, LSU
Mr. M. Ying, Coastal Ecology Institute, LSU

The following assisted in the data analyses:

Mr. Ted Ernst, ManTech Inc.
Ms. Lynn McAllister, ManTech Inc.
Dr. B. Moser, Department  of Experimental Statistics, LSU
Mr. M. Ying, Coastal Ecology Institute, LSU

The following reviewed the report:

Dr. Hillary Neckles, National Biological Service
Dr. John Rodgers, University of Mississippi
EMAP Draft Report -1994	Page vii

-------

-------
                            EXECUTIVE  SUMMARY
 This document describes the rationale,
 objectives, approach, and strategy for testing
 biological indicators of ecological condition in
 coastal wetlands. This coastal wetlands program
 is part of the Environmental Monitoring and
 Assessment Program (EMAP) administered by
 the Environmental Protection Agency's (EPA)
 Office of Research and Development.

 The overall goal of EM AP-Coastal Wetlands is
 to provide a quantitative assessment of the status
 and long-term trends in coastal wetland
 condition on regional and national scales with
 known confidence.  The specific, long-term
 objectives of EMAP-Coastal Wetlands are to:

 1)  Quantify the regional  status and monitor
 changes through time of coastal wetlands by
 measuring indicators of biological condition.

 2)  Quantify the change in extent of coastal
 wetlands through time on regional and national
 scales.

 3)  Identify associations between coastal wetland
 condition and hydrologic stress, pollution
 exposure, and other factors affecting wetland
 condition.

 4) Provide timely data and interpretive
 summaries, reports, and assessments of wetland
 condition and trends.

 The purpose of this report is to begin the process
 of indicator selection and testing to produce the
 appropriate field measurements, statistical
 metrics, and reporting indices to assess status or
 condition of coastal wetlands.  In short, how do
 we define and measure  coastal wetland
 condition?

 The use of biological indicators to assess coastal
 wetland condition or "health" is central to the
 EMAP concept. It assumes that meaningful
 information can be obtained for regional and
 national assessments of important coastal
 wetland attributes on a fairly constrained and
 limited set of indicator measurements.  The
 development and selection of indicators for
 EMAP-Coastal Wetlands is viewed as a
 continual process, now in its  early stages.

 This study examined the evaluation of 21
 wetland indicators related to  sediment
 characteristics, vegetation, and hydrology. The
 study focused on the quantification and
 evaluation of five endpoints with regard to these
 indicators:

 1) Spatial and Temporal Variability - Indicators
• exhibiting low natural  temporal and spatial
 variability at the sampling site significantly
 assist in the ability to ascertain differences in
 status and detect trends.

 2) Responsiveness - Indicators exhibiting high
 responsiveness reflect  change in ecosystem
 condition and respond  to either stressors of
 concern or management strategies.

 3) Interpretability and Ambiguity - Indicators
 related unambiguously to a biological endpoint,
 exposure, or habitat increase the clear
 interpretation of findings.

 4) Integration - Indicator integrates numerous
 aspects of environmental stress over time and
 space.

 5) Cost Effectiveness -Indicators can be
 collected and  evaluated at low cost relative to
 information value.

 The objective of the pilot study for coastal
EMAP - Estuaries Draft Report -1994
                                     Page 9

-------
 wetlands was to evaluate 21 indicators in the
 above five categories to ascertain those that
 might be of use to determine condition in a
 regional/national monitoring program.

 The report is in three major sections describing:

 1)  Methodologies used in the design, field
 sampling, laboratory processing, and statistical
 analysis of the  study,

 2)  Results of the indicator evaluation, the
 quality of the collected data, and the
 interpretation of the findings, and

 3)  Recommendations for the further suitability
 of any of these 21 indicators for regional
 monitoring, as well as any new indicators
 determined as candidates but untested by this
 pilot study.

 The indicators evaluated can be grouped into
 three broad categories:

   1. Soil Parameters
        •  Salinity
        •  Bulk Density
        •  Percent Organic
        •  Sulfide
        •  PH
        •  eH
        •  Hydraulic Conductivity
        •  Water Levels
        •  Chemical constituents - trace metals
        •  Chemical constituents - nutrients
        •  Sediment/organic accumulation

   2. Vegetation Parameters
        •  Cover
        •  Biomass
        •  Stem Density
        •  Stem Length
        •  Stem Diameter
        •  Chemical constituents - trace metals
        •  Chemical constituents - nutrients
        •  Species presence
   3. Other
        • Water levels (time series
         measurements) '
        • Spectral Reflectance

The initial selection and classification of sites as
either healthy or impaired were made based on a
basin-scale habitat map, Chabreck (1978), that
showed the extent of salt marsh habitats.
Healthy sites and impaired sites were selected,
using aerial photography, from each of the three
basins (Barataria, St. Bernard, Terrebonrie) in the
Louisiana coastal salt marshes. The judgment
(determining what  was healthy and what was
impaired) was based upon: 1) the rate of recent
land loss, 2) obvious internal marsh breakup, and
3) severe alteration of natural hydrology or
impoundment by canals and spoil banks.

The sampling occurred at "Healthy" and
"Impaired" sites in three hydrologic basins
within the Louisiana Coastal zone (Terrebonne
Basin, Barataria Basin and St. Bernard Basin).
These Basins were formed by various
distributary lobes of the Mississippi River over
the last -5,000 years. The St. Bernard marshes
are the least likely to receive new sediment from
the Mississippi River and are the  most stable salt
marshes in the coastal zone. The  Terrebonne and
Barataria marshes both have some external
sediment input, although the degree of input
varies for each basin. The stratification  of
sampling by drainage basin  was intended to
account for possible variation due to the
different sedimentary history and age of the three
basins sampled.

The general sampling scheme at a site consisted
of a circular sampling cluster with a center
sampling point surrounded by 5 sampling plots
10 m from the center, arranged like spokes on a
wheel. The center point is located 50 m inland to
ensure that edge effects will not influence the
data. This scheme  allowed for the collection of
up to six replicates  within a study site to address
site-level sampling variability.
EMAP - Estuaries Draft Report -1994
                                   Page 10

-------
Within each of the basins, six "Healthy" and six
"Impaired" marsh health classes were sampled,
using the scheme described above.  In addition,
triplicate sampling was conducted at one of the
sample sites within each basin-health class to
address within-site variability at a 50-to 100-
meter scale. The triplicate sampling provided
replicates for the accretion cores, the leaf tissue
and the sediment constituents.

The overall data return (all sites combined) for
the project was 94%. The major data loss was
from the St. Bernard basin, primarily due to
rough weather.

In general, most of the indicators show the
minimum variance at the within-sample site
level, with increasing variance as the spatial
scale increased from sample site to co-located
site to basin or marsh health level. This increase
in variance is small enough (<25% increase) for
some indicators to be unimportant.  Indicators
that exhibit essentially constant or consistent
variance across all spatial scales are:
   Total Biomass
   Spartina alterniflora biomass
   Water cover
   Number of stems
   Mean stem length
   Mean stem diameter
   Wet bulk density
   Dry bulk density
   eH
   Sulfide
   Bottom salinity (>20 cm depth)
   Depth to 1963 137Cs peak
The sampling replication of these indicators,
within a site, could be decreased in favor of
greater spatial coverage. Similar results can be
seen with the sediment and leaf constituent data.

There are reasonable relationships between the
morphology of the plant and total biomass that
may be used to non-destructively estimate
standing live biomass for this species.  In
practice this procedure would, for example,
result in measuring the morphological aspect on
all samples and bringing back some samples
(25%) for biomass determinations. The
empirical  relationships can be established in the
lab and compared to previous measurements,
resulting in a significant increase in efficiency
(i.e., less equipment and fewer samples in the
field and fewer lab measurements).  It will be
useful to investigate morphometric indices for
other species (especially for Juncussp.). - Not all
species are amenable to this approach.
Measurements of plant stem morphology may be
used to distinguish healthy from impaired sites
in this plant community.

There is an apparent relationship  between the
sulfide concentration in the soil at the time of
sampling and the density of tassels. There are no
tassels above a sulfide concentration of 30 ppm
indicating a minimal tolerance for sulfides or
another factor that co-varies with sulfides. The
sulfide measurements are representative perhaps
of soil conditions over the previous half-day to  a
few days.  The tassel density is indicative of
growing conditions for the previous several
weeks.

Soil hydrologic conductivity, sulfide and total
sulfur concentration may be useful indicators to
distinguish between healthy and impaired «£.
alterniflora  marshes. These measurements
should be  continued over a wider area and
expanded  to examine other species-dominant
groups.

There appear to be some statistical relationships
between plant spectral reflectance (particularly
those measured at the 200 ft. and  400 ft.
altitudes) and marsh health. Although the
spectral indices were not well correlated with
plant vigor, a weak relationship among plant
vigor and  some of the spectral indices was
detected.  These results suggest that a more in-
depth investigation of the use of spectral
EMAP - Estuaries Draft Report -1994
                                   Page 11

-------
 reflectance in assessing marsh plant vigor is
 warranted.

 The discriminant model showed that marshes can
 be classified into healthy or impaired with an
 error of-29% for the healthy and 22% for the
 impaired using the following variables:

   • The sum of the stem diameters
   • The log of the number of tassels
     (stems with seed heads)
   • The log of the sulfide concentration
   • The log of the "hydraulic conductivity"
   • The log of the sediment sulfur
     concentration.

 The Canonical Discriminant Analysis model
 showed that healthy and impaired sites separate
 (statistically). Although this model seems
 reasonable, it still needs to be verified, perhaps
 by using either part of the data to develop the
 model then testing it with the remaining data or
 by collecting a new data set. We feel that the
 latter approach should be used, because the data
 set is fairly small. This verification can be
 accomplished by applying the model developed
 during this Indicator Study to the data to be
 collected during the next phase of EMAP.

 Recommendations from the study are:

 • Plant morphology and structure (e.g., stem
   width and reproductive structures) are
   potentially biomass-independent indicators of
   stress.

 • Soil properties (e.g., eH, bulk density, carbon,
   hydraulic conductivity, sulfide and total S)
   are sources or consequences of stress that are
   easily measurable and probably essential
   properties to measure in EMAP.  Interpreting
   the significance of variations in these
   properties requires additional measures that
   may eventually be reduced (e.g., accretion
   rates, water level, etc.).

 • Accretion rates are a valuable addition for
   data interpretation, especially for evaluating
   controlling factors causing plant stress.
   These new data should be used to address
   questions about long-term marsh accretion
   and the relationship between biomass and
   accretion rates. These relationships remain
   prevalent issues for both indicator
   development and resource management.

•  Pre-sampling aerial surveys should be made
   available for site selection, and logistical
   support and a 2-or 3-segment historical
   comparison of the sites is very informative
   for determination of whether the sites are
   healthy or impaired.

•  Installation of water-level gages may be too
   labor intensive to continue for most sites, but
   water-level is an essential measurement to
   continue in some fashion, if only to determine
   important relationships among stressors and
   plant responses. It may be informative to
   examine the tide gage records of nearby field
   sites or to choose field sites for indicator
   development on the basis of their proximity
   to good tide gage records.

•  It is very  cost-effective to collect some soil
   samples for archival purposes. The toxic
   effects  of pollutants are frequently a threat,
   and these data could be integrated with the
   other EMAP studies (e.g., EMAP Estuarine).
   It may be good to include a screening for
   some organic pollutants for the same reason.
   Furthermore, the constituents may  be giving
   us signals to interpret about marsh health and
   indicator responses.

•  This study was initiated as a preliminary
   attempt to identify whether spectral
   reflectance measurements of the marsh
   surface from a helicopter platform could be
   used to assess marsh health and, thus, would
   warrant continued investigation.  The results
   presented above indicate that differences in
   marsh vigor may be definable with this
   technique. However, the sources of variation
EMAP - Estuaries Draft Report -1994
                                   Page 12

-------
   in .the data must be identified, and a larger
   number of sampling stations must be
   employed.

 • pH measurements seem quite useless for the
   program. The variability among sites was
   low and ephemeral where it varies; thus, the
   biological basis for continuation is unclear.

 • Sampling efficacy may be improved by
   investigating the relationship between sample
   frequency and variability. For example, there
   are two ways to improve upon the previous
   sampling efforts for estimating plant biomass.
   One is to sample fewer plots, and the other is
   to further develop morphometric measures for
   non-destructive sampling.  Modification of
   sampling scheme will reduce overall
   sampling effort with a small loss of
   replicability.  Specifically, the number of
   replicates for biomass harvest can be reduced
   from 6 to 5 plots.  This should be examined
   further and may have a potentially long-term
   consequence for field sampling efficiency.

 • EMAP-Wetlands has expanded its scope
   beyond ecosystem health of monocultural
   stands of Spartina alterniflora to include
   ecosystem health and general resource
   condition of coastal wetlands comprised of
   multiple species and habitats. In practice, this
   may mean that indicators of fish habitat
   quality, for example, are appropriate areas for
   indicator development.

 « Non-destructive sampling techniques are
   desirable, especially in view of the
   desirability of long-term landowner
   cooperation.

 • Below ground biomass is a potentially
   important parameter to measure in subsequent
   studies.

 • Indicator development for individual species
   of homogenous macrophyte cover will be
   easier than for development of heterogeneous
   plant cover. There is a drastic change in
   species dominance between salt and
   freshwater marshes. The difficulties involved
   in sampling the brackish marshes are much
   greater than in sampling monotypic salt
   marshes.  Caution is urged in expecting too
   much too soon when expanding the
   vegetation types analyzed from salt marsh to
   other plant communities.

•  The response of plants to a  stressor is not
   necessarily linear. There may be a threshold
   effect (e.g., to tidal energy or submergence)
   or an optimum response level (e.g., a
   pollutant, sulfide  or salinity). The range of
   conditions found  in the Louisiana field trials
   may not represent all ranges of factors
   affecting the status of plant health in Gulf of
   Mexico wetlands. For these reasons and
   others, it is prudent to continue using more
   rather than fewer  of the tested indicators.

•  Soil salinity was never an important
   component of any of the statistical cluster or
   discriminant analyses. However, it may be an
   especially important parameter to include in
   Gulf of Mexico-wide sampling, in view of the
   hypersaline conditions anticipated in Texas
   estuaries.
EMAP - Estuaries Draft Report -1994
                                  Page 13

-------

-------
                               2  INTRODUCTION
2.1  OVERVIEW OF EMAP

This document describes the rationale, approach,
objectives, and strategy for the testing of
biological indicators of ecological condition in
coastal wetlands. This activity is one element of
a larger strategy for the establishment of a
monitoring program to assess the status and
trends in the ecological condition of the Nation's
coastal wetlands. This coastal wetlands
monitoring program is a single element of the
Environmental Monitoring and Assessment
Program (EMAP), a nationwide program
administered by the Environmental Protection
Agency's (EPA) Office of Research and
Development (ORD). EMAP is designed to
characterize the changing conditions of the
Nation's ecological resources on large
geographic scales over long periods of time.
Although EMAP is designed and funded by
ORD, other offices and regions within EPA (e.g.,
Office of Water) and other federal agencies (e.g.,
National Biological Survey, U.S. Fish and
Wildlife Service) have contributed to its
development and will participate in the
collection and use of EMAP data.

The overall goal of EMAP is to monitor the
condition of the Nation's ecological resources, to
evaluate the  success of current policies and
programs, and to identify emerging problems
before they become widespread or irreversible.
In addressing this goal, EMAP has four primary
objectives:

   1)  Estimate the current-status, trends, and
       changes in selected indicators of the
       Nation's ecological resources on a
       regional basis with known statistical
       confidence.

   2)  Estimate the geographic coverage and
       extent of the Nation's ecological
       resources with known statistical
       confidence.

   3)  Identify associations between selected
       indicators of natural and anthropogenic
       stresses and indicators of condition of
       ecological resources.

   4)  Provide annual statistical summaries and
       periodic assessments of the Nation's
       ecological resources.
2.2  OBJECTIVES OF
   EMAP-COASTAL WETLANDS

The overall goal of EMAP-Coastal Wetlands is
to provide a quantitative assessment of the status
and long-term trends in coastal wetland
conditions on regional and national scales. The
specific, long-term objectives of EMAP-Coastal
Wetlands are to:

1) Quantify the regional status and monitor.
changes through time of coastal Wetlands, by
measuring indicators of biological condition.

2) Quantify the change in extent of coastal
wetlands through time, on regional and national
scales.

3) Identify associations among coastal wetland
condition and hydrologic stress, pollution
exposure, and other factors affecting wetland
condition.

4) Provide timely data and interpretive
summaries, reports, and assessments of wetland
condition and trends.

The first objective of the EMAP-Coastal
EMAP - Estuaries Draft Report -1994
                                  Page 15

-------
 Wetlands Program (EMAP-CW) requires the
 identification of biological indicators of coastal
 wetland condition to ascertain status and monitor
 changes.  In other words, we will be reporting on
 selected wetland indicators on regional and
 national scales and on the status and trends in
 wetland extent as important indicators of
 wetland condition. This goal raises many
 challenging questions that converge on two
 general themes: the selection of indicators and
 the sampling/analytical design that would permit
 the extrapolation of specific measurements to
 represent  large spatial regions.

 The purpose of this report is  to begin the process
 of indicator selection and testing to produce the
 appropriate field measurements, statistical
 metrics, and reporting indices to assess status or
 condition of coastal wetlands. In short, how do,
 we define and measure coastal wetland
 condition?
2.3 EMAP FRAMEWORK FOR
   INDICATOR DEVELOPMENT

The use of biological indicators to assess coastal
wetland condition or "health" is central to the
EMAP concept. It assumes that meaningful
information can be obtained for regional and
national assessments of important coastal
wetland attributes on a fairly constrained and
limited set of indicator measurements.
Identification of the best set of indicators to
achieve this objective is critical to the success of
EMAP-CW.

The development and selection of indicators for
EMAP-Coastal Wetlands is viewed as a
continual process, now in its early stages. The
basic framework for indicator identification and
evaluation is described fully in Barber et al.
(1993) and is summarized in Figure 2.1. It is
important to the success of EMAP-Coastal
Wetlands  that the indicators selected, upon
which assessments of status and condition will
be made, establish a foundation for interpretation
 by identifying the primary environmental values,
 assessment endpoints, and assessment questions
 of concern for the resource.  These values,
 endpoints, and questions are the roadwork to the
 selection of indicators appropriate to meet
 EMAP-CW's objectives (Figure 2.2). Once these
 attributes are established, the process of
 indicator selection and evaluation can begin.

 The three primary, common environmental
 values associated with the resources being
 examined by EMAP are:

   1. Biological Integrity
   2. Consumptive Uses
   3. Non-Consumptive Uses.

 2.3.1  BIOLOGICAL INTEGRITY

 Wetlands perform many functions that can
 translate into  the maintenance of biological
 integrity. Wetland habitats offer unique physical
 and biotic features not found in other
 ecosystems. They are productive resources that
 support breeding, nesting, developmental, and
 feeding activities for many species offish and
 wildlife. In addition to providing habitat for
 numerous obligate wetland species,
 approximately 20% of the species listed as
 threatened and endangered depend upon wetland
 habitats during some part of their life cycles.
 Wetland productivity is often greater than that of
 surrounding ecosystems and supports both
 internal trophic relationships and biomass
 export.  Wetlands provide important hydrologic
 functions including water storage and flood
 abatement. Coastal wetlands can also contribute
 to water quality improvement through
sedimentation, pollutant immobilization, and
uptake of various pollutants  and nutrients.
2.3.2  CONSUMPTIVE USES

Coastal wetlands provide critical spawning and
nursery habitat for commercially-and
EMAP - Estuaries Draft Report -1994
                                  Page 16

-------
                  EMAP  INDICATOR EVOLUTION
                       (1)  IDENTIFY
                ISSUES/ASSESSMENT ENDPOINTS  <«l	
      OBJECTIVES
      Develop Indicators
      linked to endpoints
              METHODS
              EExpert Knowledge
              Literature Review
              Conceptual Models
              Criteria
  CANDIDATE INDICATORS   <^-
      Prioritize based
      on criteria         '3'
        - reject, suspend, or
         proceed
              EExpert Knowledge
              Literature Review
              Conceptual Models
                RESEARCH INDICATORS
      Evaluate expected
      performance      (4)
        -quantitative testing
        and evaluation
              Analysis of Existing Data
              Simulations
              Pilot Tests
              Indicator Testing/Evaluation
              Mock Assessments
              Conceptual Models
DEVELOPMENT INDICATORS
      Evaluate actual
      performance on a
      regional scale      '5'
        - build infrastructure
        - demonstrate ability
        - assess logistics
              Regional Demonstration
               Projects
              Regional Statistical
               Summary
                  CORE INDICATORS
      Implement Regional
             and       (6)
      National Monitoring
        - periodic evaluation
              EMAP Data Analysis
              Correlate Old Indicators with
               Proposed Replacements
EVALUATION

Workshops
Criteria
Peer Review
Criteria
Peer Review
Criteria at
 Regional Scale
Peer Review
Agency Review
 of Summary
Feedback from
 Peers and Agencies
Peer Review
                            Assessment Indicators
                            Revisit Assessment Endpoints
      Figure 2-1. Framework for indicator development.
EMAP - Estuaries Draft Report -1994
                                                        Page 17

-------
                            Regional or National Wetland Condition

       Assessment
       Endpoints
Productivity
           Abundance of
            Vegetation
            Vertebrates
            Macroinvertebrates
           Sediment characteristics
Biodiversity
Sustainability
                   Landscape Indicators

                   Community composition of
                   Vegetation
                   Vertebrates
                   Macroinvertebrates

                   [Threatened and Endangered
                   Species]
                          Wetland Extent

                          Landscape Indicators

                          Indicators of Hydrology

                          Vegetation community
                          composition
      Nutrients in    Chemical    Pesticide    Sediment  Hydrologic  Land  Resource  Point and   Climate
      sediments and  contaminants applications accretion  alteration   cover  mgmt.    non-point
      tissues         in sediments                                  type            discharges
Figure 2-2. The basic framework for indicator identification and evaluation for EMAP-Coastal Wetlands.
EMAP - Estuaries Draft Report -1994
                                                             Page 18

-------
 recreationally-important fish and shellfish and
 serve as primary nesting, feeding, and resting
 habitats for many species of birds including
 migrating waterfowl. By providing recreational
 opportunities and serving as a source of
 commercial products, coastal wetlands are
 important economic resources. The sporting
 industry is dependent on the continued
 productivity of coastal wetlands (i.e., biological;
 integrity) for sport fishing and waterfowl
 hunting. Coastal wetlands support an annual.
 harvest of fish and shellfish.

 2.3.3 NON-CONSUMPTIVE USES

 By providing recreational opportunities beyond
;the extraction of consumptive items like
^shellfish, waterfowl, and fish, coastal wetlands
 provide a unique ecosystem for many public   '
 users.  Non-consumptive users of coastal
 wetlands are attracted by their diversity of plant
 and animal life.  Many wetlands provide
^educational and research opportunities that
 provide significant non-consumptive value.

 2.3.4 PILOT TESTING OF
   INDICATORS OF WETLANDS--
   ENVIRONMENTAL VALUES '   .-

 An earlier evaluation of environmental values
 and assessment endpoints produced'a list of
 potential indicators of coastal wetland condition
 (Leibowitz et al., 1991). The result of mat
 activity is in Table 2.1 which lists candidate
 indicators for coastal wetlands.
2.4 PURPOSE AND OBJECTIVES
   OF PILOT STUDY

This present study examined the evaluation of 20
specific metrics (Table 2.2) of several of the
coastal wetland indicators relating to sediment
characteristics, vegetation, and hydrology. This
study focused on the quantification and
 evaluation of six endpoints with regard to these
 attributes:

 1) Spatial and Temporal Variability —
 Indicators exhibiting low natural temporal and
 spatial variability at the sampling site
 significantly assist in the ability to ascertain
 differences in status and detect trends.

 2) Responsiveness --- Indicators exhibiting high
 responsiveness reflect change in ecosystem
. condition and respond to either stressors of.
 concern or management strategies.

 3) Interpretability and Ambiguity --- Indicators
 related unambiguously to a biological endpoint,
 exposure, or habitat increase the clear
; interpretation of findings.

 4) Integration — Indicator integrates numerous
 aspects of environmental stress over time and
 space.

 5) Cost Effectiveness — Indicators that can be
 collected and evaluated  at low cost relative to
• information value.

 6) Regional Applicability — Indicator is
 meaningful over geographic space.

 The evaluation of these indicators was only
 within one type of coastal wetland, making any
 assessment of regional applicability (the last of
 the EMAP indicator selection criteria) pertinent
 only to a single marsh type.

 The objective of the pilot study for coastal
 wetlands is to evaluate each of the tested
 indicators in the above five categories to
 ascertain which ones might be of use to
 determine conditions in  a regional/national
 monitoring program.
EMAP - Estuaries Draft Report -1994
                                   Page 19

-------
Indicator"

Groups
Wetland Extent

Landscape Indicators

Indicators of Hydrology


Sediment Characteristics

Community Composition
and Abundance of Vegetation

Community Composition and
Abundance of Vertebrates
Hcrpctofauna
Mammals
Birds
Community Composition
Abundance of
Macroinvcrtebratcs
Chemical Contaminants
in Sediment
Bioaccumulation in Tissues


Nutrients in Sediment and (or)
V T*
vegetative i issues

S = SustalnabDIty
B = Biodiversity
P = Productivity
' "Category "


Response &
Exposure
Exposure
& Response
Exposure


Exposure
& Response
Response


Response




Response


Exposure

Exposure


Exposure






Relevant


S

S

S
P
B
S
P
B
P
S
B
P



P
B

S

' S
B
P
P
. „ ,c . ', . '
o •




	 Priority' " " Compatibility Endpoints with Other - '
Resources

High High :

' 'High High t .'':•.

High Low


High Low

High

Moderate


Low Moderate
Low Moderate
•',-.. .i . ' ' 	 , »
:!•...,..:• , ,. ' :, 	 i. : .,: 	 , .-:. i, , ,. 	 • , , ,; 1
Low Low, except for adjoining
surface water
I
Low Low, except for adjoining ' ;
surface water !
Low Low " '

'• • ' ": :/ 	 ; '•
" Low Moderate
: / ^', '."I',1 t , ', .f ,

.-.. ••.'..: - .:-..•'•••..>,•«••"- j s- :;•••• : >.,- ,". . . -;.:'. .•>„., V



Table 2-1. Candidate Coastal Wetland Indicators for EMAP.
EMAP - Estuaries Draft Report -1994

-------
 Indicator Category


 Wetland Extent


 Landscape Indicators

 Indicators of Hydrology

 Sediment Characteristics
 Community Composition and
 Abundance of Vegetation
 Community Composition and
  Abundance of Vertebrates

 Community Composition and
  Abundance of Macroinvertebrates

 Chemical Contaminants in Sediments

 Bioaccumulation in Tissues

 Nutrients in Sediment and/or
 Vegetative Tissues	
Specific Indicators Samples


Water Levels


None

Water Levels Time Series

Salinity
Bulk Density
Percent Organic Carbon  :
Sulfide Concentration
pH     -   .  .-
eH ,
Hydraulic Conductivity
Sediment/Organic Accumulation
                                                         Cover
                                                         Biomass
                                                         Stem Density
                                                         Stem Length
                                                         Species Presence
                                                         Number of Tassels
None


None

Trace Metal Concentrations

Trace Metal Concentrations

Nutrients in Sediment
Nutrients in Plant Tissue
: Table 2-2, Coastal wetlands indicators evaluated in this pilot study.
 EMAP -Estuaries Draft Report -1994
                                                    :Page2J::

-------
 2.5 STRUCTURE OF REPORT

 The remainder of this report is organized into
 three major sections describing: 1)
 methodologies used in the design, field
 sampling, laboratory processing, and statistical
 analysis of the data,

 2) results of the indicator evaluation, the quality
 of the collected data, and the interpretation of the
 findings, and

 3) recommendations for the further suitability of
 any of these 20 indicators for regional
 monitoring, as well as any new Indicators
 determined as candidates but untested by this
 pilot study.
EMAP - Estuaries Draft Report -1994	Page 22

-------
      3  STUDY DESIGN AND INDICATOR SELECTION
3.1  SITE SELECTION

3.1.1  PURPOSE OF NON-RANDOM
   SITES-INDICATOR
   DEVELOPMENT

The purpose of EMAP Coastal Wetlands
monitoring is to estimate wetland condition based
on a probabilistic sampling design. However, the
testing of the efficiency of indicator condition (i.e.,
can they differentiate between "healthy" and
"impaired" conditions) requires the use of a
non-probabilistic design to efficiently ascertain
the power of the selected indicators.  The use of
judgmental sites of known condition (based on
previous knowledge) is the more effective design
to test the strength of individual indicators or
groups of indicators. This indicator testing
program sampled sites at the two ends of the
marsh health continuum ("Healthy" and
"Impaired") and developed indicators that could
differentiate between these health conditions.  The
sites were classified as either being healthy or
impaired and were sampled to determine
variability in a wide variety of physical, biological
and geological parameters. These parameters are
called "indicators" within the framework of this
program, because we were looking for ways to
characterize differences among levels of
ecological conditions. If the differences were
strongly differentiated, then these parameters
would be examined for their general applicability
in a regional scale-monitoring program.
3.1.2  SITE SELECTION CRITERIA

The initial selection and classification of sites as
either healthy or impaired were made based on a
basin-scale habitat map, Chabreck (1978), that   '
showed the extent of salt marsh habitats.  Healtliy
sitesand impaired sites were selected, using aenal
photography, from each of the three basins    %  '
(Barataria, St. Bernard, Terrebonne) in the
Louisiana coastal salt marshes. The.judgment   ;
(determining what was healthy and what was
impaired) was based upon: 1) the rate of recent
land loss, 2) obvious internal marsh breakup, and
3) severe alteration of natural hydrology or
impoundment by canals and spoil banks.'
     -    • . "  -       •'''' •-',-- V"    ',-'--    - •  ,
Candidate sites were evaluated using ah inventory
of the NASA overflights for various time periods
within the Louisiana Coastal zone and using
loss/accretion maps. We used the most recent
overflight (1988-1989) and the USACOE land
loss maps (i.e., showing land loss from -1935 to
1978) to determine areas that have remained
stable and areas that are breaking up.  Defining
whether a marsh is healthy or impaired is
somewhat subjective and also complicated by the
varying scales of the available photography. The
1988 aerial photography is high-altitude
photography (scale approximately 1:24,000),
while the U.S. Army Corps of Engineers land loss
maps were at a coarser resolution (1:62,500).
However the ACOE map scale did not permit us
to assess vegetation and open water in the same
manner as they could be assessed from low
altitude overflight or finer-scale photography.

The following steps were used to select field
sampling sites:

1. Using most recent aerial photographs and
vegetation maps, salt marsh areas were located
that were characterized by <40% open water and
those with >60-70% open water.
EMAP-Estuaries Draft Report-1994
                                 Page 23

-------
 2. The recent photos were compared with the
 USACOE maps to determine if the site had
 changed during the time period.

 3. If the site had remained stable at <50% open
 water, it was classified as healthy. If the site
 showed an increase from <40% open water in
 1978 to >60% open water in 1988, then it was
 classified as impaired.

 4. Procedures 1 through 3 were repeated until 6
 healthy and 6 impaired sites were identified within
 the salt marshes of each of the three basins.

 5. The sites were checked to ensure that each
 could be considered a unique site and that no two
 sites of a given classification (healthy, impaired)
 were hydrologically controlled by the same local
 drainage network.

 The intent was to select sites at the two ends of the
 marsh health continuum ("Healthy" and
 "Impaired"). The original classification procedure
 called for the sites to be flown over before
 sampling to confirm that the classification was
 reasonable. However, time constraints did not
 allow for the photos to be collected prior to the
 field sampling. As a result, some sites were
 misclassified. An incorrectly classified site was
 defined as a site that was determined upon
 sampling to be (1) not salt marsh or (2) not  '
 meeting the classification criteria for healthy or
 impaired sites described above.  A re-
 classification scheme was developed, based upon
 the Pilot Study sampling and the analyses  of the
 aerial photos that were obtained after the
 sampling. [If the photos had  been available before
 sampling, we believe that several sites would not
 have been sampled.]

 A reclassification scheme was developed based
 upon the aerial photos and our field sampling
 experience. We reclassified the sites without
 looking at the indicator data to minimize any bias.
 The results of the re-classification are described in
 Table 4.6 which presents the justification used for
 the classification and reclassification for each of
 the sites sampled. The reclassification is
 summarized in Table 4.7.  Of the initial
 classification for the Terrebonne basin, 75% were
 not changed.  We had difficulty actually finding a
 healthy marsh in the Barataria basin. Only one
 site that we initially classified as healthy in
 Barataria turned out to be a healthy site, and two
 turned out to be impaired sites. The percentage of
 sites initially classified correctly for Barataria
 basin was 42%.  In the Bernard marshes, one site
 initially classified as an impaired site was
 reclassified as healthy. The percentage of sites
 initially classified correctly for the St. Bernard
 basin was 75%.  In summary, of the 45 sampled
 clusters, 15 were re-classified as healthy sites, 17
 were reclassified as  impaired sites, and 13 were
 reclassified as "in-between or undetermined" sites.
 Twelve of the 45 sites (27%) required a change in
 classification.

 Sites were selected on the basis of examination of
 these aerial photographs and knowledge of the
 field sites.  A "healthy" site had relatively high and
 constant habitat areas from 1978 to 1988 (the
 most recent photographs available at the time).
 An "unhealthy" site had relatively low and
 constant (low variability) plant cover from 1978
 to 1988 or had declining plant cover from 1978 to
 1988. This was a somewhat subjective analysis
but was also based on personal knowledge of the
field conditions in south Louisiana.  Further, these
decisions were discussed among three experienced
marsh ecologists, and sites with indecisive or
unclear classification results "were not used in this
study.
EMAP - Estuaries Draft Report-1994
                                   Page 24

-------
3.2  STUDY DESIGN

3.2.1  GEOGRAPHIC SCOPE —
   WITH RATIONALE

Sites were selected to provide broad geographic .
coverage, geographic variability, an
-------
            1200
      G.m2
             600-
        G/cc
                                                  MARSH PERCENT ORGANIC
              20 J  Swenson» 1989-91          Marsh Salinity at 15 cm Depth
        PPT
Swenson, 1989-90
        PPT
                                                                 • Nov.'89
                                                                 •B-Mar.'90
                                                                 •e-Jun. '90
                                                , Marsh Salinity at 50 cm Depth
                                                          » 50 cm Nov. '89
                                                          •H- 50 cm Mar. '90
                                                          •9- SO cm Jun. '90
                             25          50         75        100
                                    Distance Inland (m)
                                                         125
    Figure 3.2 Variations in different parameters measured with distance into a salt marsh. (1) Live biomass in August
    and September (from Kaswadji ct al. 1990). (2) Percent organic matter (from Burcsh 1978). (3) Bulk density (from
    Swenson 1983). (4) Marsh salinity at IS cm depth in November, March and June (from Swenson, Peterson and
    Turner, unpublished). (5) Marsh salinity at 50 cm depth in November, March and June (from Swenson, Peterson
    and Turner, unpublished).
EMAP - Estuaries Draft Report -1994

-------
is mostly due to erosion, not fragmentation.
Further, the last marshes left in a natural marsh
are stream-side marshes, because flieir substrates
are at a slightly higher elevation than those in
inland marshes. The National Wetland Inventory
is mapping these erosional losses. Variability in
the stream-side zone confounds sampling and
data interpretation; therefore, it is wise to sample
far enough inland to avoid the "stream-side"
effect. It appears that starting a transect about 50
m into the marsh is sufficient to reduce sampling
variability due to this elevation gradient.
Samples were taken at least 50 m inland to
reduce variability and to thereby improve
statistical comparisons of marshes.

The original classification procedure called for
aerial photography of the sites to be sampled two
months before sampling. The purpose of this
photography was to ensure that the classification
was accurate. Time constraints did not allow for
the photos to be collected before the field
sampling.  As a result, some sites were
misclassifled.

Variability with soil depth is also considerable
but diminishes with depth. Figure 3.3 shows
examples of this variability for soil salinity and
pH. Data on the vertical distribution of salinity
in a marsh (Figure 3.3) indicates that salinity
varies most near the surface. Thus, a sampling
depth > 30 cm should be used to help reduce
some of this upper-level variation and to increase
chances of detecting long-term trends. However,
we measured vertical profiles of marsh salinities
over the upper 50-75 cm of the marsh during the
development of the sampling protocol.
3.3  INDICATOR SELECTION


3,3.1  GENERAL CRITERIA

Indicators of Wetland Condition

The term "indicator" within EMAP refers to the
specific environmental characteristics to be
measured or quantified through field sampling,
remote sensing, or compiling of existing data.
The selection of indicators is viewed as a
multi-year process, now in its fairly early stages.
The indicators proposed in this document are
considered research indicators; each requires
additional field testing and evaluation and, in
some cases, methods, development prior to
full-scale implementation.

It is critical to the success of EMAP-Wetlands
that the characteristics of the environment
monitored are appropriate to meeting the
program's assessment goals. The first step in the
indicator development process, therefore, is to
define a framework for indicator interpretation
by identifying the environmental values,
assessment endpoints, and major stressors of
concern for the resource. The interpretation of
the EMAP-Wetlands monitoring results will
focus around three major assessment endpoints:

1.  Productivity, including both floral and faunal
components.

2.  Biodiversity, defined by  the variety of floral
and faunal species inhabiting the wetland in
terms of both community composition and
structure, as well as by the functional niches that
are represented.

3.  Sustainability, defined as the robustness of
the wetland, its resistance to changes in structure
and function and its persistence over long
periods of time, as measured by both the size and
hydrology of a wetland.
EMAP - Estuaries Draft Report -1994
                                  Page 27

-------
            Marsh Salinity Profile
           D Jul. '89 40m
           o Sept.'8950m
           • Mar.'9040m
         Swenson 1989-90
                   5         10
                SALINITY (PPT)

                Marsh pH Profile
 CM
   -25~
   -50
          Buresh, 1977
                Rep.
            ©  1
            A  2
            •  3
            a  4
       6.5
7.5
8.5
9.5
                    pH
                                   Marsh Salinity S.E. Profile
                                                      40m Inland
                                                   o  50m Inland
                                                   •  4m Inland

                                               Swenson 1989-90
                                 0            20          40
                                   TWO STANDARD ERROR (%)

                                   Marsh  pH 1 Standard (Deviation
                         Buresh, 1977
   0.2
Std. Dev. (pH)
0.4
Figure 3 J Sampling variations in parameters and their statistical variability with depth. (Top left) Interstitial soil salinity vs. Depth for
samples taken In July, September and March, 1990, at a station located 40 m Into the marsh. (Top right) The statistical variability (+/-2S.E.)
for 10 monthly samples at depth for three locations in salt marsh: 4,30 and 40 m into the marsh (from Swenson and Peterson, unpublished).
(Bottom left): pH measurements in a salt marsh taken 4 times with depth. (Bottom right) Standard deviation of the 4 samples for pH, as shown
in the left pane! (from Buresh, 1978, for samples taken in 1977).
EMAP - Estuaries Draft Report -1994
                                                          Page28

-------
 :,?Wetland;coridition. will be judged, therefore, in
 -relatiqn to the productivity, biodiversity, and
  sustainability of the system as inferred from the
  measured EMAP indicators. The objective is not
  to maximize the wetland attribute, such as
 • productivity, but to evaluate the measured
 ; indicator values relative to expected norms for a
 ] wetland of that type andjegion. Natural,
 j wetlands are not always highly productive (e.g.,
 •. ombrotrophic bogs) nor highly diverse (e.g.,
 ; coastal salt maj-shes). The proposed
 ; EMAJP'-Wetlands indicators and their   £
  relationships to these assessment jbndpoihts are
 ' illustrated in Figure 3.4.    -  ;- •       ]._
 , As. a grpup, the se,t of indicators measured for
 -EMAP-Wetlands must provide ah adequate basis
  both tprassess wetland condition and to conduct
  the diagnostic analyses described belpw.  Four
  types of indicators will be monitored:  (1)
  response indicatprs that provide a measure of
  biological condition (e.g., vegetation community
f •composition); (2)'exposurevrndi'catdrs that assess
  the occurrence and magnitude of contact with a
  physical, chemical, or. biological stressor  (e.g.,
\  nutrient concentration); (3) habitat indicators
  that characterize the natural physical, criemical,
[  or biological conditions necessary to support an
•  organism, biological population, or community
  (e.g., wetland hydrology); and (4) stressor
\.  indicators that quantify natural processes,
?  environmental hazards, or management .actions
  that result in changes in exposure or habitat (e.g.,
]  changes in land cover type).

  Assessing Wetland Health

 - The assessment of ecosystem condition or, by
  human analogy, "health" requires both (1) the
  occurrence of certain criteria considered
  indicative of a healthy  sustainable resource and
  (2) the absence of known stressors and
  detectabl,e'symptoms of ecosystem stress. .The ,:
••: challenge for EMAP-Wetlands is to conduct
  such an assessment using the types of
  information and measurements that can be
 collected within the"fohitraiirit&bf the EMAP
 design.  No indices of wetland condition. ;>
 currently exist that are widely acceptedfin the
 scientific literature and tested and applied on a
 regional scale. The development of techniques
 for assessing wetland health will require,    , , ;?
 therefore; innPVatiy.ekpproaches to data analysis
 and interpretation4nd are the subject of,
 substantial future research within the  *
 EMAP-Wetlands program.            >£,i - -

 In general, for each ^vetland plass in each region,
 wetland condition will Tj.e judged by comparing
 the measured indicator Values sjWitti:    v,,

 • expected nprjmal ranges for^each response
, yafiable, derived from rneasurements preference
 sites, historical records, the available literature,
 and (or) expert judgment; and

 • information on stress-damage thresholds for
 each exposure indicator, obtained from the
 literature andtavailable data.—    • - - •   .

 The terms nominal and subnominal within
 EMAP refer to "healthy" and "unhealthy"
 conditions, respectively. Wetlands classified as
 nominal are assumed, by definition, to be    . . „
 performing as expected for a wetland of .that'
 type, within that region, and for the specific^
 assessment endpoint of interest. Classificlifcidh
 of a wetland as nominal or subhominal|will rely
 not on any single indicator, but on the full set of
 monitored response, exposure, habitat,
-------
 I
                            Regional or National Wetland Condition
                            	A	
       Assessment
       Endpoints
Productivity
           Abundance of
            Vegetation
            Vertebrates
            Macroinvertebrates
           Sediment characteristics
Biodiversity
Sustainability
                   Landscape Indicators

                   Community composition of
                    Vegetation
                    Vertebrates
                    Macroinvertebrates

                   [Threatened and Endangered
                   Species]
                           Wetland Extent
                          Landscape Indicators

                          Indicators of Hydrology

                          Vegetation community
                           composition
      Nutrients in    Chemical     Pesticide    Sediment  Hydrologic  Land  Resource Point and  Climate
      sediments and  contaminants  applications accretion   alteration   cover  mgmt    non-point
      tissues         in sediments                                  type            discharges
Figure 3.4 Conceptual model showing linkages to salt marsh values and assessment questions.
          EMAP - Estuaries Draft Report -1994
                                                               Page 30

-------
has demonstrated that plant biomass (Hardisky et
al., 1984) and plant stress (Mendelssohn, McKee
and Ewing, 1990) can be determined with high-
resolution spectral measurements of the marsh
canopy and leaf tissue, respectively. Stress is
typically manifested as higher reflectance spectra
in the 400-600 nm and 800-1100 nm ranges. An
advantage of this bio-indicator is the potential to
correlate the results of these measurements taken
at a very low altitude with high altitude remote
sensing techniques such as the Landsat thematic
mapper and Airborne Imaging Programs of
NASA.       '                          •    •
3.3.2  SAMPLING INDEX PERIOD

Seasonal and annual variability exist with
biomass, salinity, water levels, as well as with
other factors. Figure 3.5 contains examples of
this variability.  Annual variations occur in salt-
marsh biomass (Morris et al., 1990), salinity
(Wiseman el al., 1990), average water level and
rnonlhly variations in water level (Turner 1991),
and water-column nutrients. Although August is
the peak month in the accumulated
above-ground live biomass of Spartina
alterniflora, that peak does not last for long and
the standing crop of live biomass declines
quickly within a 1-2 month period (Figure 3.5).
Reproductive structures begin to appear in
August; therefore, it is important to sample both
before the decline in biomass and after the peak
production period. However, the appearance of
reproductive structures may be indicative of
healthy plants, and their presence/absence can be
used as an indicator of stress.

Soil chemistry differences in stressed salt
marshes, as indicated by eH measurements
(Figure 3.5), are likely to be greatest during the
period of maximum soil flooding and highest
plant biomass. Further, it is important to sample
in as short a time period as possible, perhaps
within one week, to minimize the variable
impacts of seasonal changes in flooding that are
common during late summer. Samples were,
therefore, taken in the index period between late
August and September.
3.3.3  LIST OF INDICATORS
  .'CHOSEN ..;•:.      ".;..',  '_   ....;,. ••...

The indicators evaluated can be grouped into
three broad categories:

   1. Soil Parameters
          Salinity
          Bulk Density
          Percent Organic
          Sulfide                   .
          pH
          eH
          Hydraulic Conductivity
          Water Levels
          Chemical Constituents - trace metals
          Chemical Constituents - nutrients
          Sediment/Organic Accumulation

   2. Vegetation Parameters
       • Cover
       • Biomass
       • Stem Density
       • Stem Length
       • Stem Diameter
       • Chemical Constituents - trace metals
       • Chemical Constituents - nutrients
       • Species Presence

   3. Other
       • Water Levels (lime series
         measurements)
       • Spectral Reflectance
3.4  SAMPLING SCHEME

The sampling occurred at "Hcallhy'Vand
"Impaired" siles in Ihree hydrologic basins
wilhin Ihe Louisiana Coastal zone (Terrebonne
EMAP - Estuaries Draft Report -1994
                                  Page 31

-------
  CM
        20
        10-
          0-
              Monthly Mean 1950-79
Open Water Level
              Recording gage 1989-90
   Hr  300.
       1000
               Kaswadji, 1980-81
 GM2 50°-
              Monthly Sampling 1989-
 PPT  10'
          0
          0
 MV  -250 -
      -500
              1974
                                                           Live Biomass
    Marsh Salinity
                                                                S -15 cm
                                                                X-50 cm
                                                                -» Bayou
       Marsh Eh
                                                           -»  Natural
                                                           S  Weir
                                                          i
                                                         11
           13
15
                                         Month
Figure 3.5 Monthly variations in the marsh for different parameters to be measured in this project. (1) Water
level at Hayou Rigcau on Grande Isle. Monthly deviations from the mean level from 1950-1979 are plotted to
compensate for sea level rise and subsidence (adapted from Turner 1991). (2) Hours of marsh flooding at a salt
marsh near Cocodrlc, Louisiana (from Swcnson, Peterson and Turner, unpublished).  (3) Monthly standing crop
of live .Sport/no atlerntflora near Airplane Lake in 1970-1971 (Kirby 1971) and in 1980-81 (adapted from
Kaswadji ct al. 1990). (4) Salinity in a salt marsh near Cocodric in the bayou and at 15 and 50 cm depths, in the
Inland marsh (from Swcnson, Peterson and Turner, unpublished). (5) Monthly cH values for a salt marsh
measured at 10 cm in a salt marsh with and without a weir (from Hoar 1975).
EMAP - Estuaries Draft Report -1994
                               Page 32

-------
Basin, Barataria Basin and St. Bernard Basin).
These Basins were formed by various
distributary lobes of the Mississippi River over
the last -5,000 years. The St. Bernard marshes
are the least likely to receive new sediment from
the Mississippi  River and are the most stable salt
marshes in the coastal zone.  The Terrebonne
marshes and the Barataria marshes both have
some degree of possible sediment input,
although the degree of input for each may be
different.  The stratification of sampling by
drainage basin accounts for possible variations
due to the different sedimentary histories and
ages of the three basins sampled.

The general sampling scheme at a site consisted
of sampling a cluster of points located in the
marsh. The cluster was circular with a center
point, surrounded by 5 sampling plots that
radiated out from  the center with a constant
distance of 10 m,  like spokes on a wheel (Figure
3.6).  The center point was 50 m inland to ensure
that any edge affects would not influence the
data.  This distance (50m) was measured from
the back edge of the natural berm or spoil bank.
Thus, the sampling was 50 m into the interior
marsh vegetation. This scheme allowed for the
collection of the required number of replicates
(up to a maximum of six) within a study site to
address site-level  sampling variability.

The number of replicates within a plot was based
upon literature estimates of the effect  of sample
size on the estimated mean weight of Spartina
alter niflor a biomass (Kaswadji et al.,  1990).
This study showed that the variation began to
level off around seven samples and was
unchanged for 8 (or greater) samples.  Similar
results were obtained for the measurements of
bulk density and eH (Figure 3.7). Based upon
these results, six sampling plots within a sample
site were used for Biomass, Cover, Bulk Density,
Percent Organic and Water Depth. Leaf Tissue
samples and Sediment Constituent samples were
collected at all six plots but were combined (in
the field) into a single sample. The cost of the
analyses precluded the testing of more than ~50
samples for each of these variables. The same
was true for the sediment accretion cores. The
time involved in sample processing and analysis
limited the number of samples to one per site.

Within each of the basins,,six "Healthy" and six
"Impaired" marsh health classes were sampled
using the scheme described above,  m addition,
triplicate sampling was conducted at one of the
sample sites within each basin-health class to
address within-site variability at a 50-to 100-
meter scale.  In  addition, the triplicate sampling
gave site replicates for the accretion cores, the
leaf tissue and the sediment constituents. The
sites where the triplicate sampling was made
were chosen randomly from the sites to be
sampled within  a basin-health class. At a
triplicate site, the sampling cluster was set up,
sampled, then a  second sampling cluster was
established 50 meters from the first cluster. This
was accomplished by extending the line 90
degrees from Replicate A another 50 meters and
making this point the center of the new cluster
(Replicate B). If the Replicate B was in an area
that was closer than 50 meters from the back
edge of the berm or spoil, Replicate C was
established 50 meters inland and was used
instead (this was repeated using Replicates D, E,
F, if needed). This second cluster was set up and
sampled. A third cluster was then set up and
sampled using the same procedures (Figure 3.8).
The overall sampling design is shown in Figure
3.9.
EMAP - Estuaries Draft Report -1994
                                   Page 33

-------
              I
              o
              •
              D
              D
Bulk Density, % Organic Cores
Sediment Accretion Core
Sediment Contaminant/Grain Size Core
Hydrologic Conductivity Test
Vertical Salinity Samples
Sulphides, eH
Clip Plot, Leaf Tissue Sample
Belowground Biomass
Benthic Core   _
                    50 Meters to
                    Vegetation Edge
                                                                    Radius = 10 Meters
              Figure 3.6 The sampling distribution and frequency within the Reid plot
EMAP - Estuaries Draft Report -1994
                                                                  Page 34

-------
  1100

•**


11000
•S


J 900-
I  800-
o
        0
s
 C
 0

 o
 0
 o
30-


20-


10-
          B
                                              LIVE BIOMASS
                           4          6

                          SAMPLE SIZE
10
                                                 BULK DENSITY
                    246

                         NUMBER OF SAMPLES
                                                  8
10
Figure 3.7 The effect of the number of samples on (A) the mean weight uffiparlina alterniflara biomass (Kaswadji ct al. 1990), and (B) the coefficient

of variation (%) of bulk density.
EMAP - Estuaries Draft Report -• 1994
                                                    Page 35

-------
                             WATER
                                                               REPC
                                              REPB

                                                 50 m center to center
                     SAMPLING,              m   ,   m
                     ENTRY   j ^50 m from edge *>4^-?__

                     POINT   **	*<*•  REPA
                                                              MARSH
              Figure 3.8 The layout of the triplicate sites.
EMAP - Estuaries Draft Report -1994
Page 36

-------
       BASIN
                            MARSH
                            CONDITION
SAMPLE
SITE
SUB-SAMPLE
ITERREBONEI
I BARATARIA \
{ST. BERNARD!
Figure 3.9 The overall sampling design. "H" refers to a "healthy marsh" and "I" refers to an "impaired marsh."
EMAP- Estuaries Draft Report -1994
                                         Page37

-------
 3.5 SAMPLING PROCEDURES

 Details of the sampling methods used in the
 1991 sampling season are provided in the
 Coastal Wetlands Pilot Study Quality Assurance
 Project Plan (Swenson  et al., 1992a).  The
 methods are summarized in general terms below.
 The indicators were grouped into field and
 laboratory procedures.
 3.5.1 SOIL PARAMETERS - FIELD

 Salinity Samples

 Vertical salinity profiles were measured using
 sampling pipes made from 1.3-cm (1/2") diameter
 PVC plumbing pipe.  The pipe was cut to the
 desired length, a PVC point was cemented on the
 end, and a scries of small holes were drilled in the
 pipe about 10 cm above the point. The pipes were
 inserted into the marsh sediment until the holes
 were at the desired depth for sampling and were
 allowed to stay in place for about 30 minutes. The
 pipes were then withdrawn from the marsh, and
 the water that had collected in the pipe was
 withdrawn and placed in small vials. Samples
 were collected at depths of 0,10,20, 35, and 50
 cm. The amount of sample collected was too
 small (~1.5 ml) for the use of a field conductivity
 probe. The samples were returned to the
 laboratory for salinity determination using a
 digital chloridimctcr that only required -100
 microlitcrs of sample.

 Bulk Density/Percent Organic Cores

 Near-surface cores (11 cm length) were collected
 using a small piston corer which collected an
 uncompacted core with a volume of 50 cc. The
 corcr base was placed on the marsh surface in a
 relatively flat area (avoiding the tops of clumps, if
 possible), and the core barrel was pushed into the
 marsh substrate using the attached handle. When
 the proper depth was reached (there was a depth
stop on the corer), the barrel was withdrawn from
the marsh substrate and the cores were extruded
into clean, pre-weighed plastic centrifuge tubes.
The samples were taken to the laboratory for
analysis.

Sulfide Samples

Interstitial water samples were collected using a
Teflon sampling tube connected to a syringe. The
tube was inserted to the sampling depth, (30 cm);
then a sample was carefully withdrawn from the '
substrate (the first sample is used to rinse out the
system and is discarded). The sample was fixed in
the field with an antioxidant buffer solution and
stored on ice until taken to the Laboratory for
analysis.

eH and pH Measurements

Soil eH was measured in the field at 30 cm depth
using five replicate probes (brightened platinum)
calibrated in the laboratory before and after the
field trip.  Soil eH was measured (using a digital
voltmeter) as the potential (in  mV) of a calomel
electrode against the eH probe. The half-potential
of the calomel electrode (+244 mV) was added to
the measured potential to calculate eH.

Interstitial water pH was measured using a hand
held pH meter with a sensing well. Two drops of
water collected from the sulfide sample were
placed in the sensor well and a reading were made.
The pH meter was calibrated using pH 4.0 and 7.0
buffers before use.

Hydraulic Conductivity Measurements

A simple and inexpensive method, developed to
measure soil  infiltration, was employed at each
field site. The device was a plastic tube whose
lower end was pushed into the marsh.  Water was
put into the top end and allowed to settle. After a
few seconds, a valve was opened, allowing water
to flow from  the tube into the marsh through slits
EMAP - Estuaries Draft Report - 1994
                                  Page 38

-------
in the buried end of the lube.  The water fall rate
was limed by watching water in a transparent tube
connected on the outside of the larger tube holding
water and was marked at 0.1 m intervals. The
technique was easily learned, and the equipment
was simple and reliable.

Water Depth Measurements

Water depth was measured in the center of each
plot to the nearest centimeter, using a meter stick.

Sediment Constituent Samples

Soil samples were collected as a composite
sample, using the bulk density corer.  The
material was placed in a clean (acid washed)
Nalgcne sample container. The sample was stored
on ice and relumed to the laboratory for analysis.
Sediment/Organic Accumulation cores

Cores were collected using a 10 cm (4") diameter
x 50 cm long PVC core tube.  The tube was
inscrled carefully into the marsh with a twisting
motion, with a minimum of compaction.  The
distance the core was inserled inlo the marsh and
Ihc amounls of core collccled were measured in
the field to determine the amounl of compaction, if
any. The cores were capped, sealed with tape,
then returned to Ihc laboratory for analysis. The
cores were kept  upright during handling and
transport.

3.5.2  SOIL PARAMETERS - LAB

Salinity Analysis

Salinity samples were titrated using a
Haake-Buchler Digital Chloridometer. This
device measured the amounl of chloride in the
sample by titrating it with silver. The
corresponding salinity was then calculated. The
machine was calibrated with a manufacturer-
supplied standard during use.

Bulk Density/Percent Organic Cores

The bulk density cores were returned to the lab
where they were cleaned, wiped dry and weighed
(to the nearest 0.01 g). The caps were removed
from the sample tubes and the cores were placed
in an oven at 60° centigrade until dry. The cores'.'
were then removed from the oven, rc-cappcd and
re-weighed. The weights were used to calculate
the wet and dry bulk densities (in g/cc). The cores
were homogenized using a Wiley mill (with a #40
mesh screen). A sub-sample (-1.0 g) of the
homogenized core was used to determine percent
organic content by loss on ignition at 550
centigrade.

Sulfide Analysis

Soil Sulfide was measured (using the interstitial
water fixed in the field with the anli-oxidanl
buffer solution) with a sulfide electrode (La/ar
Research Laboratory, Los Angeles, CA). The
electrode was calibrated before use by the
preparation of laboratory standards.

Sediment  Constituents Analysis

The caps were removed from the sample
containers and the samples were placed in an oven
at 60 centigrade until dry.  The samples  were then
removed from the oven, homogenized using a
Wiley mill (with a #40 mesh screen) and placed
into numbered and labeled containers for delivery
to the analytical laboratory. Sediment constituents
(micro-nutrients, trace metals, sulfur) were
analyzed by an outside contract laboratory
(Dynatech, Inc., now Benchmark Laboratories of
Baton Rouge).  The analysis techniques included
either Flame or Furnace AA or ICP depending  •
upon the element being analyzed.  The samples
were digested (using EPA Method 3050A) in
nitric acid and hydrogen peroxide prior to
analysis. The digestate was then refluxcd with
EMAP - Estuaries Draft Report -1994
                                   Page 39

-------
either nitric acid (for furnace AA) or hydrochloric
acid (for ICP and Flame AA). The samples for
TKN were digested in hydrochloric acid only.  ,
                             -        : r  r •/..'•.;.„'
Sediment/Organic Accumulation Core
Analysis

The accretion cores were frozen upon return to the
laboratory.  The fro/en qorcs were extruded from
the core tube using a thawing box that melted trie
outer edge of the core enough to allow it to be
pushed from the core tube.. The extruded core was
measured, then placed in a labeled plastic bag and
relumed to the freezer to harden. The frozen cores
were sectioned at 1 cm intervals, using a band
saw. As the sections were cut, they were placed
in numbered and weighed dishes.  The wet •
weight of each sample was determined (to the .. .
nearest 0.1 g).  The sections were dried,
re-weighed, homogenized using a Wiley mill
(with a #40 mesh screen) then placed into
numbered and labeled containers. A sub-sample
(~1.0 g) of each of the homogenized core
sections was taken to determine percent organic
content by loss on ignition at 550 ° centigrade.  .
During core sectioning, the thickness of every
fifth sub-section was measured with a digital
micrometer to ensure the accuracy of the
sectioning. The dried and ground samples were
then counted for 137Cs using a 40% efficiency.,
Germanium detector.  137Cs, a residuai.of bomb
fallout, first appeared in 1954, peaked in the
spring of 1963 with additional large amounts in ,
1964, and has declined since, with minor
fluctuations. The activity of the 137Cs can be
used to locate the 1964 horizon.

3.5.3  VEGETATION PARAMETERS
   -FIELD

Cover

Cover was visually estimated in the field for
each species. Cover was estimated as absolute
percentages and equaled  100% for each plot
(when water was included).     ,            ,

Biomass Samples

Aboveground Spartina alterntflorq biomass was
harvested from ,0.25m2 plots.  All standing live.
and dead culms and litter were removed and
placed into pre-labeled plastic bags. The bags
were kept in either a walk-in cooler or an air
conditioned room until they were returned to the
laboratory.                  >     ••,.-,..-.? ?


3.5.4 VEGETATION PARAMETERS
   -LAB

Biomass and  Stem Morphometrics

Upon return, the biomass samples were placed in
a walk-in cooler. During processing, the
standing live portion of the samples was sorted
by species, the standing dead was separated into
standing,dead.Spartina alierniflpra and standing
dead other, and the litter from the surfacie was .,
rinsed. The sorted samples were placed into
labeled Kraft paper bags, then dried at 75 °
centigrade (-72 hours). During sorting, the
length (to the  nearest 1.0 cm)  and diameter (to
the nearest 1.0 mm) of the stems of Spartina
alterhiflora we're measured and recorded. The'
number  of Spartina. altefniflora stems with seed
heads (tassels) was noted and  recorded. The live
standing stems for other species were counted
and recorded.

Leaf Tissue Analysis

The leaf tissue samples were refrigerated upon    '
arrival at the laboratory. The samples were
prepared by washing in distilled water to remove
surface salt and mud, then placing them in Petri
dishes.  The samples were placed in an oven at
60 C° until dry (-36 hours). The samples were
then removed  from the oven, homogenized using
a Wiley  mill (with a #40 mesh screen), then ••
EMAP - Estuaries Draft Report -1994
                                 Page

-------
placed into numbered and labeled containers for
delivery to the analytical laboratory. Leaf tissue
constituents (micro-nutrients, trace metals,
sulfur) were analy/ed by an outside contract
laboratory (Dynalech, Inc., now Benchmark
Laboratories of Baton Rouge). The analysis
techniques used were the same as those used for
the sediment constituents (Table 3.1).
3.5.5 OTHER FIELD
   MEASUREMENTS

Water Levels (time series measurements)

Water levels above and below the marsh surface
were monitored for ~2 month periods at eight of
the sample sites and for ~6 month periods at four
of the sample sites. The gages used, "Stevens
Type A/F," arc a float-counterweight system with
a digital data logger. The gages were set to record
water levels at 15-minule intervals. The gages
were deployed on platforms built on the marsh
surface, with the sensing float located in a PVC
stilling well that was dug into the marsh surface.
This deployment scheme allowed for measurement
of water levels over a range from 50 cm below the
marsh surface to 150 cm above the marsh surface.
The gage elevations were surveyed to obtain the
relative elevation of the gage in reference to the
local marsh surface (mud surface and vegetation
clump surface).
Spectral Reflectance

We utilized measurements of salt marsh
vegetation spectral reflectance, measured with a
portable Li-Cor spectroradiometer, as a potential
indicator ofplant vigor and marsh health. This
instrument provides a spectral curve in the range
of 400 to 1100 nm.

Previous research demonstrated that plant
biomass (Hardisky et al., 1984) and plant stress
(Mendelssohn, McKee and Ewing, 1990) can be
determined with high-resolution spectral
measurements of the marsh canopy and leaf tissue,
respectively.  Measurement protocols followed
that of Hardisky el al. (1984).   '  ''•
3.6  QA FOR FIELD SAMPLING
   AND INSTRUMENTATION

QA methodology, as set forth in the QA Project
Plan (Swenson et al., 1992a) was used to insure
that the QA objectives of the study were met. All
participants were impressed from the beginning
with the importance of maintaining a commitment
to quality control throughout the project. The
training of field personnel was an important part
of QC. All personnel were familiar with the
procedures used and confident in their ability to
use the equipment. The use of standard methods
among teams minimized operator error associated
with the data.  Field personnel were given the
opportunity to assess procedures and to suggest
improvements, since this was a pilot study.

Field data forms were designed to prompt the field
teams to follow the field standard procedures.
Team leaders were supplied with a check-list to
ensure that all data were collected.  During the
field, laboratory, and analysis portions of the
study, internal QC checks were used to ensure
data reliability, identify potential problems and
identify sources of error.

Table 3.2 summarizes the QA/QC samples and
procedures used. Table 3.3 presents the project
QA goals in terms of data completion, accuracy
and precision.
EMAP- Estuaries Draft Report - 1994
                                  Page 41

-------
Sample Digestion
Leaf Tissue Trace Metals
Sediment Trace Metals
Leaf Tissue TKN
Sediment TKN
Leaf Tissue Analysis
I -lament

TKN'
N
K
Ca
Mg
S
P
N'a
Fc
Mn
Al
B
Cu
Zn
Mo
Ba
Pb
V
Sediment Constituent Analysis
TKN
N'
K
Ca
Mg
S
P
Na
PC
Mn
AI
B
Cu
Zn
Mo
Ba
Pb
V

HN03
HN03
H2SO4
H2SO4

Method


ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
AA (Graphite furnace)
ICP


ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
ICP
AA (Graphite furnace)
ICP

EPA Method 3050A
EPA Method 3050A
SM 421
SM 421

Method Reference
A
SM421
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA

SM 421
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA ;
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 60] OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
EPA Method 601 OA
Table 3.1. Analytical techniques used for the Leaf Tissue and Sediment Constituents. SM = Standard Methods
EMAP - Estuaries Draft Report -1994
Page 42

-------
   Type of QC Frequency
   Purpose

   Field Replicates at a sample location
             6 clip plots per location
             6 bulk density cores per location
             3 hydrologies conductivity tests per location
             3 vertical salinity profiles per location
             3 eH samples per location
             5 eH probes per sample
             2 pH samples per location
             5 pH probes per sample
             6 leaf tissue samples per location
   Sampling' site replication
   Re-measurements          '
             2 measurements oT eH per probe
             2 measurements of pH per probe
             re-measure every tenth stern
             re-weight 10% of .samples
   Lab Replicates
   Standards
   Other
             2 tilrations/salinity sample
             3 sub samples every 6th core
             for percent organic -"
             3 per organic batch
             3 every tenth salinity sample
             Vegetation Team members compare
             themselves al lest plots during the
             sampling phase
R
R
R
R
R
R
R
R
R
             Three sampling locations in one healthy
             and one impaired marsh site (selected randomly)
             in each of the three basins                 R
P

R
P,A
P,A
Figure 3.2 Summary of QA/QC samples and procedures used. Indicated
for each type of QC is the frequency with which it was used and the
intended purpose of the measurement. (R = Representativeness, A =
Accuracy, P = Precision, C = Comparability). The sample site was the
marsh that was defined as either "healthy" or "impaired"; the sample
location is the area within that site where samples were collected (the
cluster location).
3.7 DATA ANALYSIS

3.7.1  DATA BASE CREATION

The data reduction details are discussed in depth
in the Project QAPP and the Project Data Report
(Swenson, et al, 1992a, b). The general
procedure used for data base creation is outlined
here.  Data were entered into a computer data
base on a Macintosh computer using
commercially available word processing and/or
spreadsheet programs (MS WORD 5.0®, EXCEL
4.0®).  The data sets were printed out, then
verified for data entry errors by comparing the
printouts to the data sheets.  Any corrections
were noted (in red) on the printouts. These
corrections were then made to the data sets.
These data files are referred to as the "raw" data
files and are Hie "machine form" of the  field
and/or laboratory data sheets.  Thus, a field or
lab sheet could be compared with the raw data
file to verify that the data were entered  correctly.

After the raw data sets had been corrected, they
were transferred to the mainframe computer for
permanent storage and analysis. The mainframe
computer was used because data sets stored on it
are routinely backed up by the operating system
— a distinct advantage over storing them on the
desk-top computer where data could be lost if the
users forget to back it up on a regular basis. The
only valid raw data files are those stored on the
mainframe computer. This is to eliminate
confusion that may result if there were multiple
copies.  These raw files were used as input to  a
program that merged all of the data, applied any
calibration factors and or unit change factors
(English to metric), then constructed the data
into a final file for analysis. The raw data files
always remained unchanged. The program
documented all of the processing applied to the
raw data to obtain the final data set that was used
for analysis.
        EMAP - Estuaries Draft Report -1994
                                                     Page 43

-------
I. Vegetation Data
Indicator or
Data Type
Taxonomic ID
Biomass
Stern Diameter
Siem Height
Tissue analysis
X
K
Ca
tMg
S
P
Xa
Fb
Mn
Al
B
Cu
7ji
Mo
Ba
Pb
V
11. Soils/Hydrology
dl
pH
Soil Salinity
Bulk Density
Percent Organic
Sulfides
I [ydraulic
Cbnduciivily
Water Levels
Accretion


Units
species
gAn2
cm
cm

ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm

mV
0-14
ppt
g/cm3
% dry wt.
ppm
cm/sec

cm
cm

Expected
values

1000
1-2
40-100

-10000
-10000
-1500
-3000
-6000
-1000
-100
-100
-25
-100
—5
~5
*
~5
-1000
-10000
-10

-150
7-8
10-20
0.5-1,1
50-80
-100
1

-10-100
-1

Accuracy
Goal
10%*
NA
NA
NA

±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%

10 mV
0.1 pH
0.3 ppt
0.1 g/cm3
10%
100 ppm
1 crn/sec

0.5 cm
0.5cm

Precision
Goal
NA
±20%
±20%
±20%

±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%

±20%
±20%
±15%
±15%
±15%
±25%
±30%

±20%
±40%

Completeness
Goal
95%
95%
95%
95%

95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%

95%
95%
95%
95%
95%
95%
95%

95%
95%
    Table 33 Quality Assurance goals for the project. Accuracy is given in absolute units where possible; precision is the Relative
    Percent Difference between replicated measurements. The precision goal refers to individual measurements as well as the
    precision between sampling crews.
EMAP - Estuaries Draft Report -1994
Page 44

-------
II. Soils/Hydrology (can/.)
Indicator or
Data Type
Sediment constituents
N
K
Ca
Mg
S
P
Na
Ps
Mn
Al '
B
Cu i
7i\
Mo
Ba
Pb
V
* Only 10% mis-identification is allowed,
pg/g'iefers to dry weight.


Units

Pg/g
Pg/g
Mg/g
M8/S
Pg/6
M8/8
Pg/g
Pg/g
ps/g
W/8
P8/g
ppm
P8/g
Pg/g
Pg/g
PS/8
t"g/g


Expected
values

-100
-10
-0.5
-0.5
-100
-100
-10000
-20000
100-1400
10000
-1
25
-25
-5
-100
-100
-100


Accuracy
Goal

±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
'±15%
±15%
±15%
±15%
±15%
±15%.


Precision
Goal

±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%
±15%


Completeness
Goal

95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%
95%

     Table 3,3 (conl.) Quality Asrarance goals for the project. Accuracy is given in absolute units where possible; precision is the
     Relative Percent Difference between replicated measurements. The precision goal refers to individual measurements as well as
     the precision between sampling crews.
EMAP - Estuaries Draft Report -1994
Page 45

-------
Thus, il was possible to trace an indicator from
the field or lab sheet to the final data set.  Copies
of the mainframe final data files were
downloaded to the desk-top computer for ,
analysis and preparation of final graphics. This
final data file included a date in the file name to
ensure that it was the correct file.  If an error was
discovered in the mainframe files that'resulted in
a change to the file, the new  file was downloaded
and replaced the file being maintained on the
desk-top computer. The date of the file was
changed when this was done. Figure 3.10
illustrates the data-base creation process.
Inconsistencies in the final data file were
checked using STATVIEW II® on a Macintosh
Computer.  This was accomplished by plotting
various indicator variables and looking for
outliers and/or "impossible" combinations.  ,
Outliers were points that plotted outside the
main data distribution.  These were determined
by inspection of cumulative distribution plots for
each of the indicator variables. For example, on
a plot of cover against biomass, there can be no
data points that show biomass with zero cover.
Any points of this type were noted, then checked
against the original data sheets to verify them.
All outliers were also checked.  These points
v ?re *hei> ver>1cxl by checking back to the.
o ..'trial field and/o. 'ab G^ia .sheets to ensure that
there was not a data entry error. If the point was
a valid entry, it remained in the data set; if it was
a data entry error, the error was corrected.
Outliers were not deleted from the data set.
(There were very few data points that were
considered outliers after verification.)

A similar procedure was followed for the water-
level data.  However, because all the data were
digitally recorded, data base creation consisted
of off-loading the data from the cartridges and
then transferring it to the mainframe computer.  •
Final data base creation consisted of putting the
data in time-series format, creating station ID , .;•
variables and computing water levels relative to
the local marsh surface (using the elevation
 survey data).  The data set was then ready for
 final analysis.

 Spectroradiorneter scans were conducted at each
 altitude within each marsh site.  Stress is
 typically manifested as higher reflectance spectra
 in the 400-600 run and 800-1100 nm ranges. An
 advantage of this bio-indicator is the potential to
 correlate the results of these measurements at a
 very low, altitude, using high altitude remote
 sensing techniques. The operation of the
 spectroradiometer and the storage of spectral
 measurements were automatic or automatically
 controlled by a Licor 1800-01A portable
 terminal. Data were downloaded to a Zenith
 portable computer with Terminal Emulator and
 Graphics software (Licor 1800-14) for transfer
 onto, floppy disks and for printing spectral
 responses, respectively.  These data were later
 transferred to a Macintosh cqmputer for further
 analyses.
 3.7.2  QA/QC

 The five general Measurement Quality
 Objectives listed below were observed during
• this project-

 1.  Accuracy-the degree that a measured value
 agrees with an accepted known value (Taylor,
 1988).  Accuracy was estimated by measuring a
 reference sample with a known value. Bias is the
 systematic error inherent in a method or caused
 by a particular measurement device. Accuracy
 was assessed through the use of standards
 whenever such standards existed. Laboratory
 standards (manufacturer supplied or from NBS)
 were used in chloride (salinity) analyses, leaf
 tissue nutrients, and soil constituents. The
 accuracy of the eH and pH measurements was
 ensured by calibrating the meter and probes with
 pH buffer solutions.
EMAP - Estuaries Draft Report - 1994
                                   ' Page 46

-------
                  DATA SHEETS


                  DATA ENTRY
                       1
         DATA ENTRY CHECK
              MACINTOSH COMPUTER  __
             WITH SPREADSHEET AND
               WORD-PROCESSING

               FINAL RAW DATA SETS
        "ELECTRONIC" IMAGE OF DATA SHEETS
                       t
        TRANSFER TO MAINFRAME COMPUTER
   USE PROGRAM TO MERGE AND EDIT RAW DATA SETS
                  t
             PRINT OUT DATA
             FILES
     MASTER DATA SET ON MAINFRAME COMPUTER
                       t
      CREATE SUMMARY DATA SETS FOR ANALYSIS
MAINFRAME COMPUTER ««
    MEANS AS
 -^OUTLINED
    IN QA/QC

    TIME SERIES
 ->PLOTS FOR
    WATER LEVELS

 -^REGRESSIONS

 -^ANOVA

 -^DISCRIMINANT
MACINTOSH COMPUTER

   + PLOTS TO LOOK FOR
     OUTLIERS AND/OR
     "IMPOSSIBLE" DATA
     POINTS

   > REGRESSION

   * DATA CV PLOTS

   * DATA PLOTS

PRODUCE FIGURES,
TABLES FOR THE
FINAL REPORT
                          FINAL
                          VERIFICATION,
                          CORRECTION
Figure 3.10 Outline of data entry, reduction and analysis procedures.
EMAP - Estuaries Draft Report -1994
                              Page 47

-------
2. Precision—a measure of scatter among
repeated independent observations of the same
property under controlled similar conditions
(Taylor, 1988). Precision was assessed by
replicate measurements. Replicate field
measurements were made on hydraulic
conductivity, eH, and pH.  Lab precision was
measured by repeating measurements of a
sample or by  sample splits. Repeated laboratory
measurements were made on stem length and
diameter, biomass, and chloride (salinity).
Sample splits were used on percent organic,
sulfides, chloride (salinity), tissue nutrients and
soil constituents.

3. Representativeness—or how well data truly
characterize a population or environmental
condition (Stanley  and Verner, 1985; Smith et
al., 1988)-was assessed by the use of the
sub-sample sites within each of the basins. In
the lab, representativeness of a sub-sample was
assessed by taking  multiple sub-samples and
analyzing each one. This procedure was used for
chloride analysis, percent organic, tissue
nutrients, accretion, and soil constituents.

4. Comparability—the degree of confidence
with which data sets may be compared.
Comparability among the data sets was ensured
by using standardized methods for the collection
of all the data. The team members received
training prior to the start of field data collection.

5. Completeness—or the ratio of the amount of
valid data obtained to the amount expected
(Stanley and Verner, 1985; Smith et al.,  1988)-
was used as an overall index for the project. If
the completeness is not high enough  (many
missing data sets), the entire project is
compromised. Completeness for the project is
defined as the number of field samples actually
collected as a percentage of the number of
samples assigned to the sampling teams  when
sampling begins.
Field QA checks included discussions with the
sampling teams to ensure that all team members
were following the standard field procedures.
Team members were assigned to collect certain
measurements based upon their performance
during training. Thus, measurements were
collected by the "team expert" for each of the
measurement techniques. The "team expert" for
a particular measurement was the person who
demonstrated consistency and accuracy for the
measurement technique during training.  (A
person may qualify as  "team expert" for several
categories.)  Thus, each team had a "vegetation
expert," a "sediment core expert," an "accretion
core expert," etc. The  use of these assigned
duties, based upon performance, ensured
comparability among measurement teams and
sample sites. In addition, replication of
vegetation, water, and  soil samples allowed for
an estimate of precision in the field and lab
procedures.  Table 3.4  summarizes the QA
checks used.

The following formulas were used to calculate
each of the five QA objectives:

1. Accuracy  was assessed by the relative
percent difference between the measured.
parameter and the true value as set by a standard,
using the following formula:

% difference =
true value - measured value, 100
       true value

In cases where more than two samples were
involved (multiple readings of a standard), the
Relative Standard Deviation (RSD), that is, the
coefficient of variation (CV) expressed as a
percentage, was used (Taylor, 1990):

   CV=  standard deviation / mean

   RSD = CV * 100
EMAP - Estuaries Draft Report -1994
                                  Page 48

-------
2. Precision. Representativeness and
Comparability were based on analyses of the
replicate samples, using the following formula
for comparing two samples (or two subsamples
of a given sample), A and B:

% difference =
 (A-B)   . 100
(A + B)/2

In cases where there were more than two
replicates, the coefficient of variation was used.

3. Completeness will be assessed by the percent
of data collected as a percentage of the number
of proposed samples to be collected and will be
determined by the following formula:

% complete =
samples collected - proposed samples . 100
       proposed samples
3.7.3  DATA ANALYSIS

Variability Assessment

The variability associated with the indicator
measurements was assessed at the following
levels:

1.  Sampling error - by comparing the six
   replicated plots in a sampling cluster

2.  Variability within a sample site - by
   comparing the triplicate sites

3.  Variability within a basin - by using all sites
   in a basin

4.  Variability among the basins - by comparing
   sites from among basins

5.  Total variability - by using all sites
6. Marsh health class variability - by comparing
   sites by marsh health class

7. Analytical variability - by comparing
   laboratory replicates and/or standards

The variability was assessed at each of the above
levels by computing the means and standard
deviations, using the Statistical Analysis System
(SAS, 1990a,b,c,d,e).                   .

Exploration

The distribution of data from all indicators
measured was plotted in the form of cumulative
percentile plots, using commercially available
software (Statview Ilr, Abacus Concepts, 1987)
on a Macintosh computer. These plots were
inspected visually to look for large departures
from a normal data distribution.  Simple linear
correlations were performed among the indicator
variables to ascertain which indicators were
closely related. These correlations were
performed on the  mainframe computer using the
Statistical Analysis System (SAS, 1990 a, b,
c,d,e).

Analysis of Variance Modeling

All ANOVA modeling was conducted using the
Statistical Analysis System (SAS, 1990 a, b, c).
The following discussion of the method is based
upon the description of the procedure found in
the SAS/STAT Users guide (SAS, 1990 e).

ANOVA, using linear models* calculates the
variance components from ratios using the
expected mean square error. The general form
of the linear model is:

Y = XB + e

where:                                '

Y represents the univariate data,
EMAP - Estuaries Draft Report -1994
                                  Page 49

-------
B is an unknown vector of fixed-effect
parameters with a known model matrix X,
e  is an unknown vector of independent
   random variables.

The standard linear model is used to model the
mean of Y using the fixed effects B. The
variance of each element of e is assumed to be
constant.

The mixed model approach is a modification of
the standard linear model.  The general form of a
mixed model is:

Y = XB + Zv + e

where:

Y represents the univariate data,
B is an unknown vector of fixed-effect,
parameters with a known model matrix X,
v  represents an unknown vector of random
   effects with a known model  matrix Z,
e  is an unknown vector of independent
   random variables.

The variance of each element of e is not required
to be independent. The mixed model approach
can model both the mean of Y as well as the
variance of Y. In this case the variance
components can be estimated by a maximum
likelihood method, a restricted maximum
likelihood method (REML), or  a minimum
variance quadratic unbiased estimation
(MIVQUEO). In our analysis we used the
REML method.

Discriminant Analysis

All Discriminant analyses were conducted using
the Statistical Analysis System  (SAS, 1990a, b,
c).  The following discussion of the methods
employed is based upon the description of the
procedure found in the SAS/STAT Users Guide
(SAS, 1990 e).

Discriminant analysis is used to classify data
into groups by developing a classification
function (Discriminant function) based upon
measured quantitative data.  The development of
the function can be accomplished with
parametric  methods, if the data is multivariate
normal. In the case of non-normally distributed
data, non-parametric methods can be used.
Discriminant analysis differs from cluster
analysis in  that cluster analysis is used to derive
a classification, whereas in Discriminant analysis
the classification is known beforehand. Thus,
this method seems well  suited to the EMAP data
where the marsh health classification has already
been assigned.

Discriminant analysis classifies the data by
developing either a linear or a quadratic
Discriminant function (for parametric  methods).
This function classifies  the data through the uses
of the generalized squared distance between
points.  The data are placed into the group from
which they have the smallest squared distance.
We analyzed our data using both a linear and a
quadratic function and compared the results.
The Discriminant function (also referred to as the
classification criterion) is developed using either
the individual wilhin-group covariance matrices
or the pooled covariance matrix. The procedure
also allows for the specification of prior
probabilities for each of the classes being used.
The prior probabilities are used to specify the
probability  of a sample falling into one of the
classes.  In  our analysis we set the probabilities
equal  to the proportion of the original  data that
was in each of the classes being considered.

We also analyzed the data using Canonical
Discriminant Analysis.  In this technique, linear
combinations of the variables are derived, based
upon quantitative measures made on several
groups of observations.  These combinations are
derived  to have the highest possible multiple
EMAP - Estuaries Draft Report - 1994
                                  Page 50

-------
correlation within the groups. The maximum
multiple correlation is referred to as the first
canonical correlation (the coefficients of the
linear combih'atibn are referred to as the
canonical coefficients or canonical weights).
The second canonical correlation is derived by
finding a second linear combination of the
variables, uncorrelated with the first canonical
variable, that has the highest multiple
correlation. This process can be repeated to find
higher order canonical correlations up to a
maximum of the number of original variables or
classes minus one, whichever is smaller. The
correlation can be calculated either from the
pooled (within class) correlations or from the
total sample correlation. In either case, the
resulting canonical variables are un-correlated.
We derived two canonical variables for the
analysis performed for this project. The
canonical variables were then plotted with the
points identified as either coming from a healthy
or an impaired marsh. The resulting plots can
then be inspected to ascertain whether or not the
marsh health classfes are separately identified.

Regression Analysis

Regression analysis among various indicate^
variables was performed on a desktop computer
(Macintosh) using a commercial software
product (Statview Ilr, Abacus Concepts,  1987)!
Regression analysis was used primarily as an
exploration tool to investigate the relationships
among various indicators.  The results of the
desktop analysis helped  to determine which
indicators to look at in greater detail.

Other Analyses        '

Water levels

The water level data (on solid-state data
cartridges) were read using an IBM PC-XT
computer. The resulting data files were
transferred to the mainframe computer for'
analysis using the Statistical Analysis System
(SAS 1990 a, b, c, d, e).  Because all data were in
time-series format, the same techniques were
used for all sites. A preliminary analysis to
check the data for missing values and/or outliers
was performed after the data were transferred.
During this check, any needed correction factors
(for calibration) were applied. The data were
then ready for final analysis. The final analysis
consisted of the following:

   1.  Time series plots of the data
   2.  Computation of flooding statistics
   3.  Comparison of flooding data among
       sites

The flooding statistics were computed by
calculating the length of time (in hours) the
marsh was flooded relative to (1) the mud
surface, and (2) the vegetated surface (top of
vegetation clumps). The length of time flooded
was summarized as the percent of time the marsh
was flooded, on a weekly basis.  These data were
then used to estimate the total percent of time the
marsh was flooded, at each site, over the gage
deployment period (November 1991 through
June 1992).

Spectral Radiometer Measurements

Indices of plant vigor were derived from the
spectral reflectance data as described in Tucker
(1954) and McKee, Mendelssohn and Ewing
(1990). These 20 indices, presented in Table 3,4,
were determined for each spectroradiometer scan
conducted at each altitude within each marsh ,
site.  The values for green, red and near-infrared
reflectances required to calculate the spectral
reflectance indices were derived from spectral
scans by averaging the reflectance values
(determined at 2 nmjntervals) for the spectral
bands equivalent to those of the Landsat
multi-spectral scanner: green=500-600 nm (Band
1), red=600-700 nm (Band 2), and near
infrared=800-l 100 nm (Band 3). A FORTRAN
EMAP -Estuaries Draft Report - 1994
                                   Page 51

-------
computer program was developed for this
purpose. These data were then put into the JMP
statistical program, and one-way ANOVA's were
performed on these indices to determine if any of
them were significantly different healthy and
unhealthy marsh classes. Also, these indices
were regressed against total live biomass and
total plant cover (ground-truth estimates of plant
vigor) to determine if the reflectance indices
could provide a statistically significant estimate
of plant vigor.
    I, VI - maximum planl pigment reflectance
    2. Y2 - pigment reflectance (integrated area between 500 and 678 nm)
    3. Y3 - near infrared plateau (height of plateau between 770 and 900 run)
    4. Red radiance (600 - 700 nm)
    5. Infrared (IR) radiance (800- 1100 nm)
    6. IR/Red
    7. Squire rool (SQRT) IR/Red
    8, 1R minus Red
    9. IRpiusRed
    10. (IR-Red)/(IR-i-Red)
    ll.(IR + Red)/(IR-Rcd)
    12. SQRT (IR - RcdV(IR + Red) + 0.5
    13. Green radiance (500-600 nm)
    14. Green/Red
    15. SQRT (Green/Red)
    16. Green minus Red
    17. Green plus Red
    IS. (Green-Red)/(Green +Red)
    19, (Green + Red)/(Green - Red)
    20. SQRT (Green - Red)/(Green + Red) * 0.5
Table 3.4. Reflectance variables derived from the spectral wavelength scans.
EMAP • Estuaries Draft Report -1994
Page 52

-------
                      4  RESUJLTS AND DISCUSSION
4.1  COMPLETION RATE,
   ACCURACY, AND PRECISION

Data completeness for each of the indicators is
presented in Tables 4.1 through 4.3. These
tables present the amount of data obtained as a
percentage of the expected amount of data to be
collected. Table 4.1 summarizes the number of
sites sampled.  The overall data return (all sites
combined) for the project was 94%. This is very
close to the project goal of 95%. All of the sites
were sampled in Barataria and Terrebonne
Basins. The major data loss was in the St.
Bernard basin, where rough weather precluded
the last day of sampling with the result that three
of the impaired sites could not be sampled.
Table 4.2 presents the data return for each
indicator as a percentage of the target number of
sites. Table 4.3 presents the data return for each
indicator as a percentage of the sites sampled.  A
comparison of Tables 4.2 and 4.3 shows that
when we sampled a site, we were able  to sample
all of the indicators (except pH and salinity)
within the 95% completeness project goal.

The estimate of accuracy for each of the
indicators is presented in Table 4.4. The
standard used, the number of measurements of
the standard made, and the mean value (± 95%
confidence interval) obtained from these
measurements are listed in the table. The mean
Relative Percent Difference (RPD) between the
standard and the measurements, along with the
95% confidence interval, is also given. The last
column in the table indicates whether or not the
project accuracy goals were met. The project
accuracy goals were met for all of the indicators
except accretion core storage compaction. The
original quality objective for accretion core field
compaction was less than 20% but did not state a
value for an acceptable storage compaction.  The
storage compaction was not considered in the
Number of Clusters
Basin
Barataria Healthy
Barataria Impaired
St. Bernard Healthy
St. Bernard Impaired
Terrebonne Healthy
Terrebonne Impaired
All Sites
Target
8
8
8
8
8
8
48
Sampled
8
8
8
5
8.
8
45
Percent Complete
Target
95
95
95
95
95
95
95
Actual
100
100
100
62
100
100
94
Table 4.1 Percent of sites sampled during the 1991 EMAP Pilot Study. The target
completeness goal (number of samples, percent complete) for each indicator (as defined
in the QAPP) is listed along with the actual project completeness (number of samples,
percent complete).
     Quality Assurance Project Plan but was noted
     during data analysis. However, our total
     compaction (field plus storage) was within the
     project goal.  The precision estimates for each of
     the indicators are presented in Table 4.5.  This
     table presents the results of cases where multiple
     measurements of an indicator were made.  These
     multiple measurements were either replicated
     field measures, replicated laboratory measures or
     sample splits. A description of the replication
     used, the number of measurements made, and the
     mean Relative Percent Difference (RPD) of the
     measurements, along with the 95% confidence
     interval, are given. The last column in the table
     indicates whether or not the project precision
     goals were met. The project precision goals were
     met for all of the indicators except hydraulic
     conductivity, some of the sediment constituents
     and some of the leaf tissue constituents. In the
     case of the hydraulic conductivity, sampling
EMAP Draft Report -1994
                                       Page 53

-------
                 Indicator
  Number of Clusters

Target       Sampled
Percent Completeness

 Target          Actual
      Biomiss                             288         •  270               95

      Cover                               288           270               95

      Bulk Density Cores                   288           270               95

      Accretion Cores                       48           45                95

      Sediment Constituent Samples           48           45                95

      Leaf Tissue Samples                    48           45                95

      cH Readings                          1440          1215               95

      pH Readings                          288           239               95

      Sulfidos                              192           176               95

      Salinity                              960           816               95

      Hydraulic Conductivity                1152          1051               95

      The target number of samples was calculated using the following formulas:

      Sample Cluster (sample site) = 6 plots (in all cases)

      Biomass = 6 samples/cluster

      Percent cover = 6 estimates/cluster

      Leaf tissue = 6 samples/cluster

      Constituent cores = 1 core/cluster

      Accretion cores = 1 core/cluster

      Bulk Density cores = 6 cores/cluster

      ell = 3 plots/cluster x 5 readings/plot x 2 (replicate readings) 30 readings/cluster

      pll = 3 plots/cluster x 2 readings/plot = 6 readings/cluster

      sulfidcs = (3 samples/cluster +• 1 replicate sample) = 4 samples/cluster

      salinity = (3 samples/cluster + 1 replicate sample) x 5 depths = 20 samples/cluster

      Hydraulic Cond. = 3 samples/cluster x 2"(replicate reading) x 4 depths = 24 readings/cluster
                                                  94

                                                  94

                                                  94

                                                  94

                                                  94

                                                  94

                                                  84

                                                  83

                                                  92

                                                  85

                                                  91
     Table 4.2 Percent completeness for indicator variables measured during the 1991 EMAP pilot study
     based on expectations for all sites. The target completeness goal (number of samples, percent
     complete) for each indicator (as defined in the QAPP) is listed along with the actual project
     completeness (number of samples, percent complete).
EMAP Draft Report -1994
                                                                     .   Page 54

-------
Number of Samples
Indicator Target , Sampled
Biomass 270 269
Cover 270 270
Bulk Density Cores " 270 270
Accretion Cores 45 45.
Sediment Constituent Samples 45 45
Leaf Tissue Samples 45 45
eH Readings 1350 1215'"
pH Readings . 270 239
Sulfldes 180 175
Salinity 900 816
Hydraulic Conductivity 1080 1051
The target number of samples was calculated using the following formulas:
Sample Cluster (sample sile) = 6 plots (in all cases)
Biomass = 6 samplesAtoter
Percent cover = 6 estimates/cluster
Leaf tissue = 6 samples*'uste1'
/:-/
Constituent cores g core/cluster
Accretion cores = J coteMuster
Bulk Density cores = 6 cores/cluster
eH = 3 plots/cluster x 5 readings/plot x 2 (replicate readings) 30 readings/cluster
pH = 3 plots/cluster x 2 readings/plot = 6 readings/cluster
sulfides = (3 samples^Iuster + 1 replicate sample_ = 4 samples/cluster
salinity = (3 samp/cluster -t- 1 replicate sample) x 5 depths = 20 samples/cluster
Percent Completeness
Target Actual
95 99
95 100
95 100
95 100
95 100
95 100
95 90
95 88
95 97
95 91
95 97











Hydraulic CondS' samples/cluster x 2 (replicate reading) x 4 depths = 24 readings/cluster
>=— «^
    Table 43 Percent completeness for variables measured during the 1991 EMAP pilot study for indicator variables versus
    total expected as percent of sites actually sampled. The target completeness goal (number of samples, percent complete) for
    each Indicator (as Mined in the QAPP) is listed along with the actual project completeness (number of .samples, percent
    complete).       F
                     t
                     t
EMAP Draft Repvt -1994                                                                       Page 55

-------
Indicator
Variable
Biomass
Biomass
Biomass
Bulk Density
Bulk Density
Bulk Density
Percent Organic
Percent Organic
Salinity
cH
cH
Accretion Core Field
Accretion Core
Storage
Accretion Core sub-
sample thickness
Water Level gage 1
Water Level gage 2
Water Level gage 3
Water Level gage 4
Water Level gage 5
Water Level gage 6
Water Level gage 7
Water Level gage 8
Sediment AI
Sediment Al
Sediment Cu
Accuracy
Standard
Used
lOOg
500 g
lOOOg
•10g
20g
50 g
16.8%
Blank
6.4ppt
41mV
218mV
comp. (%)
comp. (%)
(1cm)
7 point cal
7 point cal
7 point cal
7 point cal
7 point cal
7 point cal
7 point cal
7 point cal
200 mg/1 Std
lOOmg/lStd
10 mg/1 Std
Accuracy
Goal
(RPD)
±5%
±5%
±5%
±5%
±5%
±5%
±10%
±10%
±5%
±20%
±20%
±20%
±0%
±10%
±1%
±1%
±1%
±1%
±1%
±1%
±1%
±1%
±15%
±15%
±15%
n
13
13
13
31
31
31
36
36
230
519
508
44
36
45
21
21
21
21
21
21
21
21
3
9
3
Mean of
Measurements
±95% C.I.
100±0.03 g
499.6±0.04 g
999.30±0.05g
10.0±0.00g
20.00±0.00 g
50.0±0.01 g
16.9±0.07 %
0.0±0.00 %
6.3±0.02 ppt
47.7±0.4 mV
214.6±0.8 mV
14.1±2.5%
3.7±1.3%
1.0±0.03cm
NA
NA
NA
NA
NA
NA
NA
NA
195.8±9.0 mg/1
99.1±5.8mg/l
9.8±0.66 mg/1
Mean
RPD
±95% C.I.
0.1 ±0.0%
0.0±0.0%
0.1 ±0.0%
0.0±0.0%
0.0±0.0%
0.0±0.0%
0.6±0.4%
0.0±0.0%
1.6±0.3%
16.3±1.0%
1.6±0.4%
NA
NA
0.0±3.0%
NA
NA
NA
NA
NA
NA
NA
NA
2.1 ±4.5%
0.9±5.8%

Was
Goal
Met?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
2.0±6.6%
   Table 4.4 Accuracy for marsh health indicator variables measured during the 1991 EMAP pilot study. Results are given for each
   Indicator variable where accuracy was assessed by means of comparison to a standard. The measurement accuracy goal [(standard
   value and expected relative percent difference (RPD) between the standard and the measurement (as defined in the QAPP)], the
   number of measurements made (n), the mean value measured for the standard (±95% Confidence Interval), and the mean RPD of the
   measurements (± the 95% Confidence Interval) are listed. The last column states whether or not the measurement accuracy goal was
   met. Although not defined as a goal in the QAPP, the percent spike recovery for the sediment and leaf tissue analyses are also listed.
EMAP Draft Report -1994
Page 56

-------
Indicator
Variable
Sediment Cu
Sediment Mn
Sediment Mn
Sediment Mo.
Sediment Mo
Sediment Zn
Sediment Zn
Sediment V
Sediment V
Sediment P
Sediment P
Sediment Pb
Sediment Pb
Sediment B
Sediment B
Sediment K
Sediment K
Sediment Ba
Sediment Ba
Sediment Fe
Sediment pe
Sediment Mg
Sediment Mg
Sediment Ca
Sediment Ca
Sediment Al
Accuracy
Standard
Used
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10mg/l Std
5 mg/1 Std
lOmg/i Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
200 mg/1 Std
100 mg/1 Std
10 mg/1 Std
5 mg/1 Std
200 mg/1 Std
100 mg/1 Std
, 200 mg/1 Std
lOOmg/IStd
200mg/IStd
100 mg/1 Std
blank
Accuracy
Goal
(RPD)
15%
15%
15%
15%
15%
15%
15%
' 15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
. 15%
15%
15%

-------
Indicator
Variable
Sediment Cu
Scdimcm Mil
Sediment Mo
Sediment Zn
Sediment V
Sediment P
Sediment Pb
Sediment B
Sediment K
Sediment B»
Sediment Fc
Sediment Mg
Sediment Ca
Sediment Cu
Sediment Mn
Sediment Mo
Sediment Zn
Sediment V
Sediment P
Sediment Pb
Sediment B
Scdimcm Ba
Accuracy
Standard
Used
blank
blank
blank
blank
blank
Wank
blank
blank
blank
blank
blank
blank
blank
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
Accuracy
Goal
(RPD)
<1 mg/1

-------
Indicator
Variable
Tissue Al
Tissue Al
Tissue Cu
Tissue Cu
Tissue Mn
Tissue Mn
Tissue Mo
Tissue Mo
Tissue Zn
Tissue Zn
Tissue V
Tissue V
Tissue P
Tissue P
Tissue Pb
Tissue B . ,
Tissue B
Tissue K
Tissue K
Tissue Ba
Tissue Ba
Tissue Fe
Tissue Fe
Tissue Mg
Tissue MR
Accuracy
Standard
Used
10 mg/1 Std
5 mg/1 Std
, ,. 10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
10 mg/1 Std
5 mg/1 Std
200 mg/1 Std
100 mg/1 Std
10 mg/1 Std
5 mg/1 Std
10 mg/1 Std
5 mg/1 Std
200mg/IStd -
100 mg/1 Std
Accuracy
Goal
, (RPD)
, 15%
... . 15%
15%
, 15%
15%
, 15%
15%
., ;. -. 15%
15%
15%
15%
15%
15%
15%
•15%
15%
15%
15%
15%
15%
15%
15%
15%
-..-,.. , 15%
• . . ' 15%
n
5 "
9
6
11
6
11
6
11
(,'
11
6
11
6
11
6
6
11
6
11
6
,11
' 5
9
• 6
11
Mean of
Measurements
+95%C.I.
10.8±0.5 mg/1
5.310.2 mg/1
9.82±0.23 mg/1
4.9310.06 mg/1
9.9510.21 mg/1
4.9410.08 mg/1
9.9210. 13 mg/1
4.8410.08 mg/1
9.8910.30 mg/1
5.0710.08 mg/1
9.8010.30 mg/l
4.8910.08 mg/1
/ 9.8810.13 mg/I
4.9110.09 mg/1
9.87±0.11mg/l
9.8310.27 mg/1
4.8510. 11 mg/1
201. 012.6 mg/1
108.014.0 mg/1
9:8710.26 mg/1
5.0110.06 mg/1
9. 810:5 mg/1
4.910.1 mg/1
195.61 1.6 mg/1
106.913.2 mg/1
Mean
RPD
+95% C.I.
8.015%
6.014.0%
1,812.3%
1.411.2%
0.512.1%
1.211.6%
0.811.3%
1.2±1.6% .
1.113.0%
1.411.6%
2.0±3.0% .
2.211.6%
1.211.3% .
1.811.8%
1.311.1%
1.712.7%
3.012.2%
0.511.3%
8.0±4.0%
1.312.6%
0.211.2%
2.010.5% .
2.012.0%
. 2.2±0.8%
'6.913:2%
Was
Goal
Met?
• Yes
Yes
., Yes
. Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
.. Yes
Yes
Yes
Yes
     Table 4.4 (con/.) Accuracy for marsh health indicator variables measured during the 1991 EMAP pilot study. Results are given for
     each indicator variable where accuracy was assessed by means of comparison to a standard. The measurement accuracy goal
     [(standard value and expected relative percent difference (RPD)  between the standard and the measurement (as defined in the QAPP)],
     the number of measurements made (n), the mean value measured for the standard (±95% Confidence Interval), and the mean RPD of
     the measurements (1 the 95% Confidence Interval) are listed. The last column states whether or not the measurement accuracy goal
     was met.  Although not defined as a goal in the QAPP, the percent spike recovery for the sediment and leaf tissue analyses arc also
     listed.
EMAP Draft Report - 1994
Page 59

-------
Indicator
Variable
Tissue Ca
Tissue Ca
Tissue Al
Tissue Cu
Tissue Mn
Tissue Mo
Tissue Zn
Tissue V
Tissue P
Tissue Pb
Tissue B
Tissue K
Tissue Ba
Tissue Fc
Tissue Mg
Tissue Ca
Tissue Al
Tissue Cu
Tissue Mn
Tissue Mo
Tissue Zn
Tissue V
Tissue P
Tissue Pb
Tissue B
Tissue K
Accuracy
Standard
Used
200 mg/1 Std
lOOmg/lStd
blank
blank
blank
blank
blank
blank
blank
blank
blank
blank
blank
blank
blank
blank
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
spike recovery
Accuracy
Goal
(RPD)
15%
15%,
<1 mg/1
<1 mg/1
<1 mg/1
<1 mg/1
<1 mg/1

-------
Indicator
Variable
Biomass
Bulk Density
Percent Organic
Percent Organic
Percent Organic
Salinity
Stem Diameter
Stem Length
eH
PH
sulfide
Hydraulic Cond.
Water Level gage 1
Water Level gage 2
Water Level gage 3
Water Level gage 4
Water Level gage 5
Water Level gage 6
' Water Level gage 7
Water Level gage 8
Sediment TKN
Sediment Al
Sediment Cu
Sediment Mn
Sediment Mo
Precision
Standard
Used
sample re-weighing
sample re-weighing
sample re-weighing
sample splits
batch differences
sample splits
sample re-measure
sample re-measure
probe re-measure
sample re-measure
replicate analysis
sample re-measure
standard re-measure
standard re-measure
standard re-measure
standard re-measure
standard re-measure
standard re-measure
standard re-measure
standard re-measure
sample splits
sample splits
sample splits
sample splits
sample splits
Precision
Goal
(RPD)
20%
15%
15%
15%
15%
15%
20%
20%
20%
20%
25%
30%
20% '
20%
20% '
20%
20%
20%
20%
20%
15%
15%
15%
15%
15%
n
76
37
14
44
7
41
924
924
618
85
174
468
6
6
6
6
6
6
6
6
6
6
6
6
6
Mean
RPD
±95% C.I.
1.11 ±0.04
0.351±0.06
0.10±0.18
1.63±0.40
2.0±0.42
0.91 ±0.28
2.95±0.24
0.00±0.00
7.30±2.20
2.13±0.67
9.63±2.23
30.56±3.45
0.00±0.00
0.00±0.00
0.00±0.00
0.00±0.00
0.00±0.00
0.00±0.00
0.00±0.00
0.00±0.00
7.4±6.4
20.5±15.4
17.0±18.1
3.3±3.1
33.3±85.7
Was
Goal
Met?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
. Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
No
Table 4.5 Estimated precision for marsh health indicator variables measured during the 1991 EMAP pilot study. Results are given for each
indicator variable where precision was assessed by means of multiple measurements. The measurement precision goal [(standard value and
expected relative percent difference (RPD) between measurements (as defined in the QAPP)], the number of measurements made (n), and the
mean RPD of the measurements (+ the 95% Confidence Interval) are listed. The last column states whether or not the measurement precision goal
was met.
EMAP Draft Report - 1994
Page 61

-------
Indicator
Variable
Sediment Zn
Sediment V
Sediment P
Sediment Pb
Sediment B
Sediment K
Sediment Ba
Sediment Fe
Sediment Mg
Sediment Ca
Sediment S
TisiueTKN
Tissue Al
Tissue Cu
Tissue Mn
Tissue Mo
Tissue Zn
Tissue V
Tissue P
Tissue Pb
Tissue B
Tissue K
Tissue Ba
Tissue Fe
Tissue Mg
Tissue Ca
Tissues
Accuracy
Standard
Used
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
sample splits
Accuracy
Goal
(RPD)
15%
15%
15%
15%
15% '
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
n
6
6
6
6
5
6
6
6
6
6
6
6
6
5
6
6
. 6
6
6
3
4 .
. 6
6
6
6
6
6
Mean
RPD
±95% C.I.
6.214.4
11.1±10.6
3.211.9
26.8124.5
35.5140.3
9.818.8
15,6128.5
8.216.0
5.313.3
26.0133.7
5.414.3
10.519.4
10.7112.4
0.812.2
23.6135.5
0.010.0
31.0128.3
12.6±10.8
' 3.315.7
30.5160.1
13.9±25.8
4.313.7
. 25.7124.1
11.917.8
6.817.8
6.615.9
3.912.2
Was
Goal
Met?
Yes
Yes
Yes
No
No
Yes
No
Yes
' .' Yes
No
Yes
Yes
Yes
Yes
• No
Yes
'No
No
Yes
No
No
Yes
No
No
Yes
Yes
Yes
     Table 4.5 (con/.) Estimated precision for marsh health indicator variables measured during the 1991 EMAP pilot study.  Results
     arc given for each Indicator variable where precision was assessed by means of multiple measurements. The measurement
     precision goal [(standard value and expected relative percent difference (RPD) between measurements (as defined in the
     QAPP) J, the number of measurements made (n), and the mean RPD of the measurements (+ the 95% Confidence Interval) are
     listed. The last column states whether or not the measurement precision goal was met.
EMAP Draft Report - 1994
Page 62

-------
resulted in an RPD of 30.6±3.5%, while the
quality goal was 30%. It is probable that the
method can be refined somewhat, resulting in a
lower RPD. In the case of sediment and tissue
contaminants, 14 of 32 (44%) split samples did
not meet the ±15% precision goal.  In the case of
the sediment constituents, 6 out of 51 tests (11%)
did not meet the accuracy goal of 15% RPD. In
the case of the leaf tissue constituents, 7 out of 51
tests (14%) did not meet the precision goal of
15% RPD. Most test failures (70%) were related
to spike recoveries, and 10 of the 13 failures were
within ±20% and the remaining 2 tests were
within ±25%.
4.2 SITE CLASSIFICATION

The initial selection and classification of sites as
either healthy or impaired were made based on a
basin-scale habitat map, Chabreck (1978), that
showed the extent of salt marsh habitats.

Healthy sites and impaired sites were selected,
using aerial photography, from each of the three
basins (Barataria, St. Bernard, Terrebonne) in the
Louisiana coastal salt marshes. The judgment
determining what was healthy and what was
impaired was based upon: 1) the rate of recent
land loss, 2) obvious internal marsh breakup, and
3) severe alteration of natural hydrology or
impoundment by canals and spoil banks.

Candidate sites were evaluated using an inventory
of the NASA overflights for various time periods
within the Louisiana Coastal zone and using
loss/accretion maps. We used the most recent
overflight (1988-1989) and the USACOE land
loss maps (i.e., showing land loss from ~1935  to
1978) to determine areas that have remained  -
stable and areas that are breaking up.  Defining
whether a marsh is healthy or impaired is
somewhat subjective and also complicated by the
varying scales of the available photography. The
1988 aerial photography is high-altitude
photography (scale approximately 1:24,000),
while the U. S. Army Corps of Engineers land loss
maps were at a coarser resolution (1:62,500).
However, the ACOE map scale did not permit us
to assess vegetation and open water in the same
manner as they could be assessed from low
altitude overflight or finer-scale photography.

The following steps were used to select field
sampling sites:

1. Using most recent aerial photographs and
vegetation maps, salt marsh areas were located
that were characterized by <40% open water and
those with >60-70% open water.

2. The recent photos were compared with the
USACOE maps to determine if the site had
changed during the time period.

3. If the site remained stable at <50% open water,
it was classified as healthy. If the site showed an
increase from <40% open water in 1978 to >60%
open water in 1988, then it was classified as
impaired.

4. Procedures 1 through 3 were repeated until 6
healthy and 6 impaired sites were identified within
the salt marshes of each of the three basins.

5. The sites were checked to ensure that each
could be considered a unique site and that no two
sites of a given classification (healthy, impaired)
were hydrologically controlled by the same local
drainage network.

The intent was to select sites at the two ends of the
marsh health continuum ("Healthy" and
"Impaired"). The original classification procedure
called for the sites to be flown over before
sampling to confirm that the classification was
reasonable. However, time constraints did not
allow for the photos to be collected prior to the
field sampling. As a result, some sites were
misclassified.  An incorrectly classified site was
EMAP Draft Report -1994
                                   Page 63

-------
defined as a site that was determined upon
sampling to be (1) not salt marsh or (2) not
meeting the classification criteria for healthy or
impaired sites described above. A
re-classification scheme was developed, based
upon the Pilot Study sampling and the analysis of
the aerial photos that were obtained after the
sampling. [If the photos had been available before
sampling, we believe that several sites would not
have been sampled.]

A reclassification scheme was developed based
upon the aerial photos and our field sampling
experience. We reclassified the sites without
looking at the indicator data to minimize any bias.
The results of the reclassification are described in
Table 4.6 which presents the justification used for
the classification and reclassification for each of
the sites sampled.  The re-classification is
summarized in Table 4.7. Of the initial
classification  for the Terrebonne basin, 75% were
not changed.  We had difficulty actually finding a
healthy marsh in the Barataria basin. Only one
site that we initially classified as healthy in
Barataria turned out to be a healthy site, and two
turned out to  be impaired sites. The percentage
of sites initially classified correctly for Barataria
basin was 42%. In the St. Bernard marshes, one
site initially classified as healthy was later
reclassified as an impaired site, and one site
originally classified as impaired was reclassified
as healthy. .The percentage of sites initially
classified correctly for the St. Bernard basin was
75%. In summary, of the 45 sampled clusters, 15
were reclassified as healthy sites, 17 were
reclassified as impaired sites, and 13 were
reclassified as "in-between or undetermined"
sites. Twelve of the 45 sites (27%) required a
change in classification.
4.3 INDICATOR VARIABILITY
        WITHIN AND AMONG
        SAMPLE SITES, BASINS,
        AND HEALTH CLASSES -

The environmental variability for all indicators
is summarized in Tables 4.8 through 4.10. These
tables present summaries of indicator variance
[coefficient of variation (CV)] at different
measurement scales (within-site, among sites,
among basins, among marsh health classes and
total). The ratio of scale-specific variance to total
variance is shown in Table 4.8.

In general, most of the indicators show the
minimum amount of variance at the within
sample site level, with increasing variance as the
spatial scale is increased from sample site to
co-located site to basin or marsh health level.
This increase in variance is small enough (<25%
increase) for some indicators and thus is
unimportant. Indicators that exhibit this
behavior of essentially constant variance across
all spatial scales are:

1.  Total biomass
2.  Spartina alterniflora biomass
3.  Water cover
4.  Number of stems
5.  Mean stem length
6.  Mean stem diameter
7.  Wet bulk density
8.  Dry bulk density
9.  eH
10.  Sulfide
11.  Bottom salinity (>20 cm depth)
12.  Depth to 1963 137Cs peak

These are indications that the replication within a
site could be decreased in favor of greater spatial
coverage.
EMAP Draft Report -1994
                                  Page 64

-------
     BH1      Original Classification was healthy; the modified classification was healthy.  Broken Marsh nearby but site is in unbroken area.

     BH2      Original Classification was healthy; the modified classification was undetermined. Area showed break-up on project photos that
               was not visible on'1988-1989 photos.                                                    .                  '

     BH3      Original Classification was healthy; the modified classification was undetermined. Area showed break-up on project photos that
               was not visible on 1988-1989 photos.'     <                                                                      •

     BH4      Original Classification was healthy; the modified classification was impaired. Only the area along the edge of the lake is not
               breaking up, the inland marsh where we sampled is breaking up based upon project photos.  .

     BH5      Original Classification was healthy; the modified classification was undetermined. Area showed break-up on project photos that
               was not visible on 1988-1989 photos.

     BH6      Original Classification was healthy; the modified classification was impaired. Only the area along the edge of the lake is not
               breaking up, the inland marsh where we sampled is breaking up based upon project photos.                              '

     BI1       Original Classification was impaired; the modified classification was healthy. An intact area of marsh in an area that is breaking
               Up-
     BI2       Original Classification was impaired; the modified classification was undetermined.  An intact area of marsh in an area that is
               breaking up. The breaking up area is much further inland (>200 m).

     BI3       Original Classification was impaired; the modified classification was impaired. The last surviving remnant of a former more
               extensive marsh. This site is now a small marsh island.

     BI4       Original Classification was impaired; the modified classification was impaired. This site appears to be in 'the last stages of
               conversion to all ppen water.

     BIS       Original Classification was impaired; the-modified classification was impaired. This site is an area that has become open water,
               except near natural levees or spoil banks.                                        '

     BI6       Original Classification was impaired; the modified classification was impaired. This site is in the process of becoming open water
               for areas near natural levees or spoil banks.

     SHI      Original Classification was healthy; the modified classification was impaired. The random sampling placed this cluster in a large
               area of dead standing S. allern.ifl.ora and mud flats in an area that was mostly solid marsh.  This was quite evident on the project
               photos but not on the 1978-1979 photo?.

     SH2      Original Classification was healthy; the modified classification was undetermined. Area showed break-up on project photos that'
               was not visible on 1988-1989 photos.

     SH3      Original Classification was healthy; the modified classification was healthy.

     SH4      Original Classification was healthy; the modified classification was healthy.

     SH5      Original Classification was healthy; the modified classification was healthy.                                  ,   ••  •

     SH6      Original Classification was healthy; the modified classification was healthy.	
     Table 4.6  Description of the original, the modified site classification, and comments explaining the classification for sites sampled
     during the 1991 EMAP Wetlands Southeast Pilot Study, (xyz where x = Basin, y = class, z = site number; B = Barataria, S = St.
     Bernard, T = Terrebonne, H = healthy, and I = impaired).
EMAP Draft Report -1994                                      	                         Page 65

-------
      SI1       Original Classification was impaired; ihe modified classification was healthy. Maps showed recent land loss in the general area,
                but this site fell into an area that was not breaking up. This was not visible on the 1978-1979 photos but was on the project photos.

      SI2       Original Classification was impaired; the modified classification was impaired.

      SI3       Site not sampled

      S14       Original Classification was impaired; the modified classification was undetermined. Maps showed recent land loss in the general
                area, bul this site fell into an area that was not breaking up. This was not visible on the 1978-1979 photos but was on the project
                photos.

      SIS       Site not sampled

      SI6       Site not sampled

      TH1      Original Classification was healthy; the modified classification was undetermined.  Area showed break-up on project photos that
                was nol visible on 1988-1989 photos.

      TH2      Original Classification was healthy; the modified classification was undetermined.  Area showed break-up (large interior pond) on
                project photos that was not visible on 1988-1989 photos.

      TH3      Original Classification was healthy; the modified classification was healthy.

      TH4      Original Classification was healthy; the modified classification was healthy.

      T1I5      Original Classification was healthy; the modified classification was healthy.

      TH6      Original Classification was healthy; the modified classification was healthy.

      Til       Original Classification was healthy; the modified classification was impaired. Site is in area of vast conversion of marsh to open
                water.

      T12       Original Classification was impaired; the modified classification was impaired. Site is in area of vast conversion of marsh to open
                water, although part of this site included a densely-vegetated natural levee.

      T!3       Original Classification was impaired; the modified classification was undetermined. Area showed recovery on project photos thai
                was not visible on 1988-1989 photos.

      TI4       Original Classification was impaired; the modified classification was impaired: Site may be an example of an impoundment with
                •llerod hydrology.

      TI5       Original Classification was impaired; the modified classification was impaired.  Site is in area that is deteriorating rapidly.

      T!6       Original Classification was impaired; the modified classification was impaired.  Site borders a large open water area that was
     	marsh in the recent past.	
     Table 4.6 (con(.)  Description of the original, the modified site classification, and comments explaining the classification for sites
     sampled during the 1991 EMAP Wetlands Southeast Pilot Study, (xyz where x = Basin, y = class, z = site number; B = Barataria, S =
     St. Bernard, T a Terrcbonnc, H = health, and I = impaired).
EMAP Draft Report - 1994                                                                               Page 66

-------
ORIGINAL CLASSIFICATION
BASIN
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
BARATARIA
ST BERNARD
ST BERNARD
ST BERNARD
ST BERNARD
ST BERNARD
ST BERNARD
ST BERNARD
ST BERNARD
ST BERNARD
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
TERREBONNE
MARSH HEALTH
HEALTHY
HEALTHY
HEALTHY
HEALTHY
HEALTHY
HEALTHY
IMPAIRED
IMPAIRED
IMPAIRED
IMPAIRED
IMPAIRED
IMPAIRED
HEALTHY
HEALTHY
HEALTHY
HEALTHY
HEALTHY
HEALTHY
IMPAIRED
IMPAIRED
IMPAIRED
HEALTHY
HEALTHY
HEALTHY
HEALTHY
HEALTHY
HEALTHY
HEALTHY
' IMPAIRED
, IMPAIRED
IMPAIRED
IMPAIRED
IMPAIRED
SITE ID
BH1
BH2
BH3
BH4
BH5
BH6
BI1
BI2
BI3
BI4
BIS
BI6
SHI
SH2
SH3
SH4
SH5
SH6
SI1
SI2
. SI4
TH1
TH2
TH3
TH4
TH5
TH6
Til
TI2
« TI3
TI4
TI5
TI6
MODIFIED CLASSIFICATION
MARSH HEALTH
HEALTHY
UNDETERMINED
UNDETERMINED
IMPAIRED
UNDETERMINED
IMPAIRED
HEALTHY
UNDETERMINED
IMPAIRED
IMPAIRED
IMPAIRED
IMPAIRED
IMPAIRED
• UNDETERMINED
HEALTHY ,
HEALTHY
HEALTHY
HEALTHY
HEALTHY
IMPAIRED
UNDETERMINED
.UNDETERMINED
UNDETERMINED
HEALTHY
HEALTHY
HEALTHY
HEALTHY
IMPAIRED
IMPAIRED
UNDETERMINED
IMPAIRED
• IMPAIRED
IMPAIRED
NEW SITE
ID
BH1-H
BH2-U
BH3-U
BH4-I
BH5-U
BH6-I
BI1-H
BI2-U
BI3-I
BI4-I
' BI5-I
BI6-I
SH1-I
SH2-U
SH3-H
SH4-H
SH5-H
SH6-H
SI1-H
SI2-I
SI4-U
TH1-U
TH2-U
TH3-H
TH4-H
TH5-H
TH6-H
TI1-I .
TI2-I ,
..' ..TI3-U.V
TI4-I
TI5-I
TI6-I
Table 4.7 Listing of the original site classification and the modified classification for sites sampled during the 1991EMAP Wetlands
Southeast Pilot Study.
EMAP Draft Report -1994
Page 67

-------
Coefficient of Variation as a Percentage
Indicator Variable
Totri biomass (g m*)
Spartina atterni/lora biomass (g ma)
Sparlina patens biomass (g ma)
Juncus raemerianus biomass (g ma)
Disiichilis spicaia biomass (g m2)
Spartlna atterntflora cover (%)
Sparlima patens cover (%)
Juncus roemeriania cover (%)
Distichilis spicaia cover (%)
Water co ver(%)
Number of stems (cm m-2)
Mean stem length (cm m-2)
Mean stem diameter (cm m-2)
Total stem length (on m-2)
Total stem diameter (cm m-2)
Number of tassels (fa-*)
Wet bulk density (g ce'1)
Dry btiBc density (g co'1}
Percent organic
cH(mV)
pH (pH units)
SulfiJc(ppin)
Water depth in plot (cm)
Hydraulic conductivity (s cm'1)
Surface (<20 on depth) substrate
salinity (ppt)
Bottom (>20 cm depth) substrate
salinity (ppt)
Core compaction (%)
Depth to 1963 (cm)
Within Site
100.7
110.0
233.6
180.5
252.2
125.8
237.7
171.5
233.8
23.3
54.8
18.9
12.9
• 54.1
53.0
158.8
10.9
20.1
14.8
-18.5
34.9
30.9
61.6
72.1
6.9
12.3
NA
NA
Among Sites Among Basins Among Health
Class
91.1
103.4
258.0
219.4
336.6
56.1
71.0
87.5
116.9
30.6
69.2
21.0
13.9,
67.1 ,
6S..9
165.2
10.6
25.1 ,
18.4
-40.7
48.9
30.4
97.0
111.3
11.2
18.4
28.4
19.8
89.2
109.5
533.6' '
486.2
490.6
. ' 155.9
612.7
498.8
498.1
26.4
67.9
25.6
17.1
70.8
67.6 .
236.4
14.2
40.2
34.1
-24.6,
54.4
55.6
103.5
187.3
27.9
• .. 29.5
59.3
36.1
92.5
110.8
536.7
367.3
583.4
157.2
560.2
408.1
437.0
26.4
69.0
27.3
18.3
70.7
69.3
164.7 ,
14.8
42.0
37.1
-25.3
60.9
• 54.9
96.1
206.2
28.9
29.6
67.3
39.5
All Data
99.0
111.6
589.4
, 290.3
484.9
155.0
621.5
333.2
565.7
26.5
71.6
27.2
18.3
70.8
68.7
201.3
14.4
44.4
37.5 .
•-25.7
'66.7
58.6
'104.7
353.9
35.2
31.3
63.2
42.0
      Table 4.8 Summary of indicator variable variance for 1991 EMAP Wetlands Southeast Pilot Study. This table represents the
      coefficient of variation (CV) for each of the indicators measured for various measurement levels (within site, among sites, among
      basins, among marsh health class and total). Parentheses after each indicator list the measurement units.
EMAP Draft Report -1994
Page 68

-------

Indicator Variable
Tissue TKN(mg I"1)
Tissue Al (mg kg'1)
Tissue Ba (mg kg'1)
Tissue Bo (mg kg'1)
Tissue Ca (mg kg'1)
Tissue Cu (mg kg'1)
Tissue Fe (mg kg'1)
Tissue Pb (mg kg'1)
Tissue Mg (mg kg"1)
Tissue Mn (mg kg'1)
Tissue Mo (mg kg"1)
Tissue K (mg kg'1)
Tissue P (mg kg'1)
Tissue Na (mg kg'1)
Tissue V (mg kg'1)
Tissue Zn (mg kg'1)
Tissue S (mg kg'1)
Sediment TKN (mg kg'1)
Sediment Al (mg kg'1)
Sediment Ba (mg kg"1)
Sediment Bo (mg kg"1)
Sediment Ca (mg kg"1)
Sediment Cu (mg kg"1)
Sediment Fe (mg kg'1)
Sediment Pb (mg kg'1)
Sediment Mg (mg kg"1)
Sediment Mn (mg kg'1)
Sediment Mo (mg kg'1)
Sediment K (mg kg"1)
Sediment P (mg kg"1)
Sediment Na (mg kg"1)
Sediment V (mg kg"1)
' Sediment Zn (mg kg"1)
Sediment S (mg kg"1)
Coefficient of Variation as a Percentage
Among Triplicate Sites
14.6
66.2
31.4
89.3
18.1
106.0
47.8
107.0
14,4
31.8

8.7
11.3
4.8
15.7
30.4
19.1
13.2
15.2
29.1
37.4
49.0
27.8
15.1
31.0
5.3
19.0

29.5
28.3
13.4
6.0
17.7
11.5
Among Basins
21.3
151.1
48.7
144.5
21.8
179.7
62.4
140.7
20.6
59.6

20.1
21.2
16.2
20.6
67.2
30.3
37.2
22.4
85.2
72.8
70.2
45.5
26.9
107.9
11.3
41.4
360.5
21.9
44.5
31.3
12.7
22.7
16.8
Among Health Classes
22.5
129.3
56.6
150.8
20.6
186.1
69.6
128.9
21.6
60.8

20.5
21.9
24.8
23.5
61.9
34.3
41.3
30.9
112.2
103.2
82.4
49.2
31.4
196.3
22.5
40.2
387.3
21.8 ,
41.3
49.5
13,8
28.5
23.1
Among Total
\ 23.4
157.7
59.8
135.8
21.9
257.5
75.3
135.5
23.5
59.3

25.9
22.9
24.0
23.9
82.6
, 35.5
40.1
30.1
103.8
105.2
81.7
49.8
31.1
230.7
23.1
45.5
670.8
25.1
99.8
51.3
/ ' • . 16.1
26.7
22.5.7
Table 4.9 Summary of leaf tissue and sediment constituent indicator sample site variance for 1991 EMAP Wetlands, Southeast Pilot
Study. This table presents the coefficient of variation (CV) for each of the indicators measured for various measurement levels (among
sites, among basins, among marsh health class and total). Parentheses after each indicator list the measurement units.
EMAP Draft Report - 1994
Page 69

-------
Ratio of Total Variance to Variance

Indicator Variable
Tola! biomass (g m"2)
Sparllna alterniflora biomass (g m"2)
Spartina patens biomass (g m'2)
Juncus roemerianus biomass (g m"2)
Distic/iilis spicala biomass (g m'2)
Spartina alterniflora cover (%)
Spartina patens cover (%)
Juncus roemeriania cover (%)
Distichitis spicata cover (%)
Water cover (%)
Number of Slems (cm m"2)
Mean stem length (cm m"2)
Mean stem diameter (cm m'2)
Total stem length (cm m"*)
Total stern diameter (cm mj)
Number of Tassels (m"z)
Wet bulk density (g ma)
Dry bulk density (g m"2)
Percent organic
eH (mV)
pH (pH units)
Sulfide (ppm)
Water Depth in plot (cm)
Hydraulic Conductivity (s cm"1)
Surface (<20 cm) Substrate Salinity
Surface (>20 cm) Substrate Salinity
Core Compaction (%)
Depth to 1963 layer (cm)
Within Site

0.98
1.01
2,52
1,61
1.92
1.23
2.61
1.94
2.42
1.14
1.31
1.44
1.42
1.31
1.30
1.27
1.32
2,21
2.53
1.39
1,91
1.90
1.70
4.91
5.10
2,54
NA
NA
Among Sites

1.09
1.08
2,28
1.32
1.44
2.76
8.75
3.81
4.84
0.87
1.03
1.30
• 1.32
1.06
1.04
1.22
1.36
1,77
2.04
0.63
1.36
1.93
1.08
3,18
3.14
1.70
2.23
2.12
Among Basins

1.11
1.02
1.10
0.60
0.99
0.99
1.01
0.67
1.14
1.00
1.05
1.06
1.07
1.00
1.02
0.85
1.01
1.10
1.10
1.04
1.23
1.05
1.01
1,89
1.26
1.06
1.07
1.16
Among Health

1.07
1.01
1.12
.079
0.83
0.99
1.11
0.82
1.29
1.00
1.04
1.00
1,00
1.00
0.99
1.22
0.97
1.06
1,01
1.02
1,10
1.07
1.09
1.72
1.22
1,06
0.94
1.06
            Table 4.10 Comparison of the ratio of total variance to variance at the different sampling levels (within a
            sample site, among triplicate sites, among basins and among marsh health classes) for the vegetation and soil
            parameters measured for the EMAP 1991 Pilot Study. The ratio of the total CV to the CV at each of the levels
            is presented.
EMAP Draft Report -1994
Page 70

-------
The coefficients of variation for the spectral
reflectance indices generally ranged from 0 to
28%. Spectral indices that contained green
minus red reflectance values exhibited the
highest coefficients of variation. Generally, the
indices became less variable with increasing
altitude.  For example, the infrared/red index
stabilized at approximately 150-200 feet. The
majority of the indices stabilized at 200 feet.
The coefficients of variation at the 200 foot
altitude were exceptionally small, ranging from
0.5 to 4.7 % for all indices except those
containing the green minus red reflectance
values. We assumed that the viewing area of the
marsh surface that was scanned at 200 feet was
relatively homogeneous; thus, variation in
viewing area due to movement of the helicopter
was minimal. These low coefficients of variation
are probably near the minimum for helicopter-
based measurements of spectral reflectance of
Louisiana salt marshes.

Twelve sites were sampled for the marsh water-
level study (using 8 gages). Four gages were
deployed at sites  in Terrebonne Basin for the
entire field experiment (November 1991 through
June 1992).  The other eight gages were
deployed at four sites in the Barataria Basin, then
were moved to four new sites in the St. Bernard
Basiu. The water level data are summarized in
Figure 4.1 which presents the percent of time the
marsh was flooded for each of the water-level
gage sites.  The upper plot presents the percent
of time the marsh was flooded above the top of
the surface of the vegetation clumps. The lower
plot presents the percent of time the marsh was
flooded above the top of the surface of mud.
These plots were made based upon all the data
collected. The healthy:impaired comparison
from Terrebonne should be the most reliable,
because it was based upon the longest record. In
general, however, the gages showed no
consistent flooding difference between the
healthy and the impaired sites.
4.4  RELATIONSHIPS AMONG
       INDICATORS

4,4.1 STEM MORPHOLOGY AND
       DENSITY

Non-destructive morphometric estimates of
Spartina alterniflora standing biomass may be
obtained using stem density, stem length and
plant cover.  The relationships among standing
biomass of live S. alterniflora and total culm
diameter and total stem diameter are shown in
Figures 4.2 and 4.3, respectively.  A multiple
regression, including total stem diameter and the
number of stems, predicts the biomass value with
a coefficient of determination >0.8 as shown in
Table 4.11. This relationship is quite good for
both healthy and impaired sites (Table 4.12).
Thus, non-destructive sampling can be used to
estimate live biomass for this species.

There is variability in the relationships among
morphometric measurements of culms and the
total biomass. Healthy and impaired sites differ
in the relationships among live plant biomass
and both the total culm length and the total culm
diameter in each sample plot (Figures 4.2 and
4.3). There is only  one. impaired sample that
could t>e considered an outlier in these plots.
This sample was reclassified from healthy to
impaired during the study and has one of the
highest biomass values of all sites. It is possible
that this site was misclassified, but we have not
adjusted the data following this analysis in an
attempt to  be objective.
EMAP Draft Report - 1994
                                  Page 71

-------

•o
0) C
T3 O
o yz
O OJ 75"
LL G>
§0
15 50-
•»- *•*
o o
4- >
C +3
gffl 25-
OQC
Q.
Mean = 27.5, S = 6.6 Mean = 28.9, S = 23.1 Mean = 22.9, S = 1 5.1







;•* V ;.
•'•'"'. *• *•*
1 n Fl 0 1 I
% & ,1 '*- 11
| ^ ' Vi 1 1







M
I
I

I







PI







f








p
/
 •o  o
 o  o
 T3  OJ
        100'
 O  3   75-
 LLCO
 0) TJ

 |l   50-
 4—  jj,
 SI   25'
 o .2
Mean = 50.6, S = 11 .3 Mean = 78.5, S = 20.1 Mean = 66.6, S = 37.3
™

i

sl
<

\ f

\

\

\



\

/

s

t
                      Neither
    Healthy
Site ID
             BH2   BH3   BI2   TI5     SH5   SH6    SI1   TH4   TH6      BI6   SI4    TI4
              I	1    I	1       I-	1
Impaired
Figure 4.1 Summary of marsh flooding relative to the vegetation surface (top) and relative to the mud surface (bottom) for the 1991 EMAP
Wetlands, Southeast Pilot Study. The horizontal axis is the original site ID, with the re-classification assignment indicated. The vertical axis
is the percent of time the marsh was flooded during the gage deployment period. The mean and standard deviation for each marsh health class
is Indicated at the top of the plot.
EMAP Draft Report -1994
                                        Page 72

-------
     1000


^   800
CM


      600
 o
 m
       400
       200
              •  Healthy
              o  impaired
                    50      100      150      200      250
                      Total Culm Diameter (cm/m2)
                                                                   300
Figure 42 The relationship between standing biomass of live S. alternifloraand total stem diameter at the healthy and impaired sites sampled
in 1991.
 CO
 E
 o
 CQ
     1000


      800


      600


      400


      200
             • Healthy
             o Impaired
                            10000            20000             30000
                  Average Total Culm Length (cm/m2)
Figure 43 The relationship between standing biomass of live S. alterniflora and total culm length at the healthy and impaired sites sampled
in 1991.
EMAP Draft Report -1994
                                                               Page 73

-------

% Cover
Live Biomass
Total Length
Toul
Diameler
%
Cover
-
0.81
0.68
0.66
Live
Biomass
0.81
-
0.79
0.74
Total
Length
0.68
0.79
-
0.91
Total
Diameter
0.66
0.74
0.91
-
     Table 4.11 Correlation matrix of the adjusted coefficient of
     determination (R1) for a polynomial regression of four
     morphometrlc measures of S. alierniflora.

Data Set
All data
Healthy Sites
Impaired Sites
Neither Category of
Sites

n
177
70
54
53
Adjusted R2
2 variables
0.82
0.91
0.87
0.74
3 variables
0.85
0.91
0.90
0.76
    Table 4.12 Correlation matrix of the adjusted coefficient of
    determination (R*) for a multiple linear regression of morphometric
    measures of S. alierniflora that may be used to predict standing live
    blomass. The 2 variable linear model uses total culm diameter and total
    culm length. The 3 variable linear model uses total culm diameter, total
    culm length and % cover.
Each of these plots shows a divergence in the
impaired  and  healthy  sites  as  the  biomass
increases.   In  effect,  the density of  stems is
apparently decreased with length or diameter at the
impaired sites. A likely reason for this is increased
aerchyma tissue, because the concentration of N, P
and other tissue  elements  showed  no  higher
concentrations of elements that could explain these
differences. There were no apparent differences in
the relationships between percent cover and  live
biomass at healthy and impaired sites (Figure 4.4).
We found no indices of stem density, number of
stems or size frequency of stems to discriminate
          between healthy and impaired sites.

          Discussion:    There  are  reasonable
          relationships between the morphology
          of the plant and total biomass that may
          be  used to non-destructively estimate
          standing live biomass for this species.
          In  practice  this procedure  would, for
          example, result  in  measuring  the
          morphological aspect on all samples
          and in bringing back some samples
          (25%) for biomass determinations. The
          empirical   relationships   can    be
          established  in  the  lab and  compared
          with   previous  measurements.    A
          significant increase in efficiency would
          result (i.e.,  less equipment and fewer
          samples in the  field and  fewer  lab
measurements).

It will be  useful  to  investigate morphometric
indices for other species (especially  for Juncus
sp.). Not all species are amenable to this approach.
Measurements of plant stem morphology may be
used to distinguish healthy from impaired sites in
this plant community at the sites sampled. It is a
promising approach to non-destructively estimate
plant condition and evaluate site condition.
EMAP Draft Report -1994
                                    Page 74

-------


CM
E
S
(O
(0
CO
E
o
CO

1000-
800-

-
600-

-

400-


200-^
                   •  Healthy
                   o  Impaired
                                       20        30         40
                                              % Cover
                          50
60
Figure 4.4 The relationship between standing biomass of live S. aherniflora and percent vegetation cover at the healthy and impaired sites
sampled in 1991.                              .•	   .  ..  .              ....             -
4.4.2  REPRODUCTIVE TISSUES

The reproductive tissues of S'. aherniflora form
over several weeks during the end of the growing
season. The tall, elongated tassels bearing the
flower head do not form on all plants. Plants
would not be expected to form reproductive
structures if carbohydrate reserves below ground
were not available. Thus, the absence or
presence of tassels may indicate recent metabolic
changes affecting plant production.

There is an apparent relationship between the
sulfide concentration in the soil at the time of
sampling and the density of tassels (Figure 4.5):
There are no tassels above a sulfide,
concentration of 30 ppm. This suggests that this
plant has a minimal tolerance for sulfides that
may not be exceeded.  However, due to the small
sample size, it is premature to construct a
relationship between influorescences and sulfide
concentration.  The sulfide measurements are
representative,  perhaps, of soil conditions over
the previous 0.5 to a few days.  The tassel
density is indicative of growing conditions for
the previous several weeks.

Discussion: Decreased tassel density may
indicate poor conditions due to elevated soil
sulfur concentration or a covariate indicator.
4.4.3  SOIL CONDITIONS AND
       RELATIONSHIPS WITH
       OTHER FACTORS

Wetland flooding (data were collected at 11 sites
using water level gages) was positively related to
soil sulfide concentration and cumulative
inorganic accumulation (Figures 4.6 and 4.7).
Wetland flooding was inversely related to the
total sulfur concentration in the soil (Figure 4.8).

Sulfides should form and accumulate during
flooding, as soil reducing conditions develop
under anaerobic conditions. The observed eH
values were generally between -100 to -200 mV,
that is, sufficiently low to suggest that sulfide
formation could occur (observed, but not
shown).
EMAP Draft Report-1994
                                 Page 75

-------
CM
E
"73
0
W
W
CO
O
0)
2
0>
TOO-
M
80-


60-

_

40-
•
20-
n.
: ' £
•2r

• • '•
1

f . ~ ;,
,,K.J
A
• •* C* *
                      10     20     30    40     50     60
                                Average sulfide (ppm)
70     80
Figure 4.S The average sulfide concentration in replicate plots and the average number of reproductive structures (tassels) of Spartina
allemiftora in those plots.                                      ;. •             .,.,-'
       50

       40
£
CL
&9%) are included.    ,   •  ,   . ,,  , , ...
EMAP Draft Report -1994
   Page 76

-------
 o
 o
5


4


3


2
 CO
 S>    1
 o
          0           20          40          60          SO         10O

              % Flooding (relative to mud  surface)

Figure 4.7 The relationship between the percent of time the site is flooded (from water level gage records) and the accumulation of inorganic
matter at the sampling sites. Only sites dominated by 5. aUerniflora (cover>80%) are included and inorganic accumulation Was determined
by sediment cores with high-quality dating using '"Cs.
         24000
         20000
 ~      16000
 CD
         12000
 -                0         20        40         60        80         100

                     % Flooding (relative to  mud surface)
Figure 4.8 the relationship between the percent of time the site is flooded (from water level gage records) and the soil sulfur concentration
at the sampling sites. Only sites that are dominated by S:aUerniflora (cover>80%) are included.
EMAP Draft Report - 1994
                                                                 Page 77

-------
 Increased flooding could result in more
 fluctuations in water level, hence more sediment
 deposition events and therefore higher inorganic
 accumulation rates. If flooding frequency
 increases sedimentation rates (an observed
 relationship in other studies), then the soil matrix
 may become less permeable to oxygen when
 suspended particles, especially fine sediments,
 accumulate.  The relative lack of organic matrix
 would decrease soil porosity under these
 circumstances. Alternatively, flooding may
 reduce soil oxidation (i.e., greater reducing
 conditions) and the organic particles may clog
 the pore spaces, instead of being decomposed to
 a gas.

 The concentration of sulfides irj soils and soil
 sulfur is inversely related and apparently
 different among healthy and impaired sites
 (Figure 4.9). The general decline in soil sulfur
 with increasing sulfide concentration may be
 related to mobilization of the soil S into gaseous
 form and release of the gas through the soil pore
 waters during tidal cycles or through the plant
 tissues. The reasons for differences between the
 healthy and impaired sites are not clear. The
 healthy sites have lower concentrations of soil
                    sulfide per total sulfur in the soil than do the
                    impaired sites. Alternatively, the impaired sites
                    have greater concentration of total sulfur for the
                    same concentration of soil sulfide. This result
                    could be a consequence of higher rates of gas
                    transport (of H2S) from soil to vegetation at the
                    healthy sites, or to greater retention of sulfur in
                    the soils at the impaired sites.  Retention could
                    be favored, for example, by anaerobic conditions
                    lower than -250 mV (e.g., under a long period of
                    flooding). While cause-and-effect relationships
                    remain to be uncovered, percent sulfur as sulfide
                    appears to be a potential discriminating factor of
                    site condition.

                    Discussion:  Soil hydrologic conductivity,
                    sulfide and total sulfur concentration may be
                    useful indicators to distinguish between healthy
                    and impaired S. alterniflora marshes.  These
                    measurements should be tested over a wider
                    geographic area and expanded to examine other
                    species dominance groups.

E
&
a
3
CO
•J±
Q)
E
1
25000

20000
15000


10000'


5000
n
• Healthy
° o Impaired
* •'."•.
• 0 ' * °
^u^ •.o 4) o
^QA
Ofl» 0 ^
O
•

20          40           60
            Sulfide (pprn)
                                                                                          100
Figure 4S The relationship between sediment sulfur concentration and sulfide concentration at the healthy and impaired sites sampled in 1991.
EMAP Draft Report -1994
                                                      Page 78

-------
4.4.4  RIOM^SS A|N(R SPECTRAL
    "r •REFLECTANGE :  ' .'
       DIFFERENCES AMONG
       MARSH HEALTH CLASSES
No significant differenc.es (P>0.15) in total live
bion^ass or.total ^over among the three .marsh
health categories were observed when the marsh
sites were reclassified into healthy, impaired and
undetermined.  However, statistical contrasts
between.only the healthy and impaired classes
revealed significance differences. There was a  ,
consistent tendency (P=0.20) for mean total
biomass to be greater in the healthy; marshes
(727 ± 170 g m2) compared with that of the
impaired marshes (353 ± 11,2 g m?) with the
undetermined marsh class being intermediate
(581 ± 14,8 g mi). No spectral indices at the 100
ft altitude were significajitly (different among
these marsh health classes. .However, at 200 ft
and 400 ft the following spectral indices showed
significant (P<0.15) differences among the
marsh classes: 200 ft: Y3 (P=0.07), infrared
(P=0.12), infrared/red (P=0.06), infrared minus
red (P=0.10), green plus red/green minus red
(P=0.09) (Figure 4.10); 400 ft: Y3 (P=0.06),
infrared (P=0.03), infrared/red gp-p.14), infrared
minus red (P=0.04), infrared plus red (P=0.02),
infrared minus red/infraired plus red (P=0.10)
(Figure 4.11).

Discussion: There appears to be some
reasonable relationships between spectral
reflectance (particularly at the 200 ft. and 400 ft.
altitudes) and marsh health, the latter assessed
from either empirical data on plant biomass (or
plant cover) or more subjectively from aerial
photographs and wetland loss records.
4.4.5 RELATIONSHIPS BETWEEN
       PLANT VIGOR AND
       SPECTRAL REFLECTANCE
       INDICATORS
In addition to analyzing differences in spectral
indices between marsh health classes,  live
above-ground biomass and plant cover were   ,
correlated with the 20 spectral reflectance:
indices to determine if statistical relationships
exist between plant vigor, as estimated by
biomass and cover, and the spectral indices. At
an altitude of 100 ft., neither total live above-
ground biomass nor plant cover were
significantly (P<0.05) correlated with  any of the
spectral indices. However, at the 200 foot
altitude, live above-ground biomass and plant
cover were weakly associated (P<0.15) with
certain spectral indices i.e., biomass correlated
with: infrared reflectance (r=0.42,P=0.14), •>
infrared plus red (r=0.43, P=0.12), and infrared
minus red (r=0.40,P=0.16); cover correlated
with: infrared reflectance (r=0.-44, P=0.11),
infrared plus red (r=0.49, P=0.07), infrared   .
minus red (r=0.39, P=0.17). At the 400 foot
altitude, no correlations were significant even at
the 15% level.  .

Discussion: Although the spectral indices were
not well correlated with plant vigor, a weak
relationship between plant vigor and some of the
spectral indices was detected.  These results
suggest that a more" in-depth investigation of the
use of spectral reflectance in assessing marsh
plant vigor might be warranted.       : ;
EMAP Draft Report -1994
                                 Page 79

-------
I 600-
| 400-
o
? 200-
m
*" n-



»
- "\



'% 'x V'',
, ">!




•.v V:
.• v :


                                                      20-
                                                  >
Q
UJ
QC
+
QC
                                                       5-

                                                       0
            Healthy Undetermined Impaired
              Marsh Health Classification
           Healthy Undetermined  Impaired
             Marsh Health Classification
20-
15-


E 10-
5-














i









„, -1
••
^ ^
f S * j













V








o-
4-

Q o_
UJ 0
DC
la
E 2-









•• ••
s

'
. / f ff
1 1 — _lll_..






1 1_



" '',
- -
'







_ J





'" '/
• i






•
             Healthy Undetermined Impaired
               Marsh Health Classification
           Healthy Undetermined Impaired
             Marsh Health Classification
       20'
       15-
       10-
        5-
                \^         \
            Healthy  Undetermined Impaired
K
§
Sw*^

!
                                                       3-
                                                       2-
               Marsh Health Classification
Figure 4,10 Selected Spectral Reflectance Indices acquired at an altitude of 200 feet as a function of marsh health class.
           Healthy Undetermined Impaired
             Marsh Health Classification
EMAP Draft Report -1994
                                  Page 80

-------
 CO


15-
in
lu-

5-
n-










f

^
' :










'
,
\ - \
1 ,













1 %/








on



10-
5-









i-
^ «•.

;
:*•












",
'-













1 /







           Healthy Undetermined Impaired

            Marsh Health Classification
Healthy  Undetermined Impaired

  Marsh Health Classification

15-
O
UJ
DC m_
*— IU —
DC
5-
0-









^'S^


" -
'











"* "•
: ^
*" •"













' \
OR.

Q 20-
LU
DC -15

£ 10-

5-
fl.








? ••
; •"

ffffS
fff f
-










',
S
fff f
f ' ':






ws «.
           Healthy Undetermined Impaired

            Marsh Health Classification
Healthy  Undetermined Impaired

  Marsh  Health Classification
     20'
     15-

 Q
 oi   ._
 DC   10-
      5-
x-s
Q
"5 0.4-
pc
LU
Of 0.2-
pc
0.0-


\t
V



•. '
""' ?~\


1
i i t
           Healthy Undetermined Impaired               Healthy Undetermined Impaired

            Marsh Health Classification                    Marsh Health Classification

Figure 4.11 Selected Spectral Reflectance indices acquired at an altitude of 400 feet as a function of .marsh health class.
EMAP Draft Report -1994
                       Page 81

-------
 4.5 ANOVA RESULTS

 Table 4.13 presents the analysis of variance
 results for the vegetation and soil parameters,
 Table 4.14 presents the analysis of variance
 results for the leaf tissue and sediment
 constituents. The only indicators that showed
 statistical significance (at the 0.05 level) for
 Marsh Health effects were the following:

        1. Total biomass
        2, Water cover
        3. Mean stem length
        4. Surface substrate salinity
        5, Bottom substrate salinity
        6. Bulk density

 ANOVA may not be the most appropriate
 analysis technique for these data. Discriminant
 analysis seems an appropriate technique because
 the Pilot Study was designed to classify
 observations into groups ("healthy" or
 "impaired") based upon quantitative measures
 collected from members within each group.
4.6 MULTIVARIATE RESULTS

We attempted to classify the sites into healthy
and impaired by discriminant analysis (see
Chapter 4 for discussion of methods). The table
of results lists the variables used, the results of a
test of the homogeneity within the covariance
matrix, the re-substitution error rate summary
and the cross-validation error rate summary.  The
results from the test for homogeneity of the
covariance matrix determine which type of
discriminant function will be used. If the matrix
was not homogeneous (at the 0.10 level), a
quadratic discriminant function was used. If the
matrix was homogeneous (at the 0.10 level),  a
linear discriminant function was used. The
re-substitution summary is a summary of the    "
classification results. The cross-validation
option that was used with the analytical
procedure (PROC DISCRIM: SAS, 1990e) is  a
 bias-reducing technique.  In this process, n-1 of
 the observations is used to develop the  - *,»,-.
 classification. This classification is then used>to
 classify the one observation left out. This
 procedure was repeated for the n observations,
 and the results were used to calculate the
 misclassification rate. Although this technique
 resulted in a nearly unbiased estimate, the
 variance was large.  However, the
 cross-validation is a more accurate estimate of
 how the classification function will work on
 future data.

 The first model run was to see how well the
 analysis could discriminate, using.only water
 cover and total biomass.  Because the sites were
 picked based upon land/water change over time
 (see project QAPP, Swenson et. al., 1992a and
 the project Data Report Swenson et. al., 1992b),
 cover and biomass essentially defined marsh
 health. This classification (see Section 3.2) was
 done on a macro- to meso-scale level, using
 aerial photographs.  The discriminate analysis
 was run on data collected at a microseale (0.25
 m2 plot) level. Thus, this discriminate analysis
 was an estimate (although crude) of the  ,
 agreement between the indicators at two spatial
 scales. Table 4.15 presents the results of the
 analysis using only Total  Biomass and Water
 Cover, The results of the cross-validation,    ;
 summary indicate that these two variables can,
 classify the sites with an error of ~18% for the
.healthy sites and ~23% for the impaired sites.
 The results of the Canonical Discriminant
 analysis are shown in Figure 4.12.  In this case,
 all three marsh health classes (Healthy, Impaired,
 Undetermined) were used to allow for the
 extraction of two canonical variables. In general,
 the healthy and impaired sites did separate, with
 the first canonical variable having the most
 discriminatory power. We then attempted to
 Develop a model based upon ojther variables that
 would yield similar Results.  The analyses were
 performed using several combinations:pf' '  v
 variables. .In, general, most o| me cl^ssifigatiions
 were of moderate success, with classification -
 errors about 50%.               ;    ••:;•••'<::
EMAP Drqft Report -1994
                                   Page-82

-------
...... • • : . ' • • - ' • •-.,;•,..•
Variability Scale '

Indicator
Total Biomass • ' ,
S. Ait. Biomass
S. Pat. Biomass
J.Rom., Biomass
D. Spi. Biomass
S. Alt. Cover
S. Pat. Cover
J. Rom. Cover
D. Spi. Cover
Water Cover
Stem Number
Sum. Length
Sum. Diameter
Length Mean
Diameter Mean
Tassels '
Number 0-10
Number 0-25
Number 125-150
Number >150
Wet Density
Dry Density
Percent Organic
eH
PH
Sulfide
Water Depth
Hydraulic Cond.
Top,Saiinity
Bottom Salinity
Compaction
Depth to 1963
.Cum, Organic
Cum. Inorganic i !.
Organic Accum.
Cesium Accum.
Min. Ace.
Max. Ace.

Basin
' .0.4136
0.5510
0.5822,
0.0478 ,
,0.8060
0.4547
0.9813
0.0
0.8059
0.8577
0.3058
0.4409
0.4523
0.0373
0.4516
0.0565
0.2659
0.2165
0.5210
0.5918
0.6994
0.0671
0.1449
0.1732
0.4942
0.5054
0.2199
0.4783
0.0022
0.0161
0.1997
0.0714
0.2469
•0.4747. . •••
0.0809
0.4645
0.1099
0.0467

, Health
0.0090
0.1043
0.2825
0.3342
0.6298
0.3514
0.1465
0.0
0.1620
0.0249
0.7248
0.5637
0.7796
0.0341
0.4955
0.1714
0.7117
0.8758
0.8186
0.5739
0.1746
0.0428
, 0.2234
0.3713
0.8521
0.5222
0.9204
0.2966
0.0002
0.0014
0.2966
0.0965
0.0973
0.2762
0.2236
0.1604
0.1298
0.4405

Health*Basin
0.3047
0.9207
0.4329
0.3714
0.9934
0.9705
0.9853
0.0
0.7686
0.3343
0.4996
0.4626
0.5741
0.0820
0.3367
0.4750 •
0.6926
0.7737 :
0.4262
0.9525
0.3381
0.6082
0.2165
0.3380
0.8339
0.5342
0.1313
0.6036
0.0013
0.0289
0.1904
0.2654
' 0.7001
• 0.7162
0.1403
0.8762
0.3187
0.2937

Site(Basin, Health)
0.4057
0.3093
0.0194
0.6240
0.9749
0.0089
0.0095
0.9994
0.0427
0.5357
0.1554 . •
0.2595
0.1762
0.3252
0.3365
0.0693
0.0002
0.0074
0.8906
0.9718
0.0096
0.0091
0.0257
0.0414
0.0086
0.0125
0.5129
0.0001
0.0006
0.0084
0.1144
0.0289
0.2155
0.0022
0.0332
0.0924
0.0674
0.2556
Site Rep (Site
Basin, Health)
0.0454
0.0125
0.9290
0.0019
0.0066
0.5528
0.9696
0.0001
0.8849
0.0404
0.0097
0.0017
0.0014
0.1017
0.0659
0.2404
0.8386
0.4892
0.0001
0.0001
0.3697
0.0125
0.0017
0.2380
0.1143
0.0023 .
0.0001
0.9981
0.6213
0.4841
0.9933
0.9890
0.9785
0.8307
0.9929
0.9710
0.9887
0.9959
Table 4-13.  Summary of ANOVA on soil and vegetation indicator variables to look at scales of variability. The probability level is listed for each
indicator for each of the variability scales (Basin, Marsh, etc.). Bold numbers indicate that the probability is significant at the 0.05 level. Results are
based upon Type III sums of squares. Site and Site Rep are considered to be random effects, Basin and Marsh Health are considered to be fixed effects.
The general model is: Indicator = Basin, Health, Basin*Health, Site(Basin Health), Site Rep(Site Basin Health).  Parentheses indicate nesting; asterisk
indicates interaction. Only Healthy and Impaired marsh health classes were used. These are the results from the. General Linear Model Procedure
[(PROC GLM) (SAS) 1988)].                                                                                              ;
EMAP Draft Report -1994
Page 83

-------
Variability Scale

Indicator
LEAF TISSUE
TKN
Al
Ba
Bo
Ca
Cu
Fc
Pb
Mg
Mn
Mo
K
P
N
V
z
s
SEDIMENTS
TKN
Al
Ba
Bo
Ca
Cu
Fe
Pb
Mg
Mn
Mo
K
P
Na
V
Zn

Basin

0.2332
0.8075
0.0980
0.1237
0.5980
,
0.0180
0.1430
0.0488
0.8041

0.0254
0.1963
0.0001
0.0556
0.4275
0.0155

0.2124
0.0112
0.1415
0.1661
0.1035
0.0407
0.0915
0.2359
0.0001
0.4120
.
0.2185
0.0015
0.0009
0.1089
0.0489

Health

0.2358
0.4050
0.1549
0.4678
0.4082
.
0.1088
0.9881
0.1276
0.9036
,
0.8466
0.3658
0.6579
0.1212
0.3085
0.2842

0.3841
0.3134
0.9464
0.8140
0.4848
0.5868
0.4342
0.2849
0.3029
0.1807

0.1803
0.1108
0.2661
0.4336
0.5642

Hcalth*Basin

0.4011
0.4668
0.2749
0.9733
0.4384
.
0.1546
0.8725
0.8898
0.0740

0.4205
0.6137
0.8334
0.7620
0.5478
0.7091

0.2570
0.7006
0.5502
0.6932
0.8260
0.4870
0.1057
0.4162
0.9757
0.4727
.
0.4659
0.2058
0.0903
0.4279
0.2550

Site(Basin, Health)

0.4297
0.0165
0.8663
0.5244
0.8429
.
0.0011
0.4947
0.6210
0.0786
• .
0.0587
0.0148
0.0122
0.8206
0.0018
0.0717

0.0294
0.1987
0.0001
0.3016
0.7399
... 0.2784
0.0685
0.0001
0.0775
0.0046

0.9355
0.6054
0.2589
0.0666
0.2045
Site Rep (Site
Basin, Health)

0.3649
0.9996
0.0297
0.9076
0.6527

0.9754
0.2546
0.4907
0.6467

0.8587
0.8170
0.9924
0.2001
0.9955
0.6010

0.9978
0.3691
0.9802
0.9997
04166
0.9452
0.5916
0.5332
0.9990
0.9551

0.0765
0.9867
0.9476
0.5855
0.0326
TaWe 4-14. Summary of ANOVA on soil and vegetation trace constituent indicators to look at scales of variability. The
probability level is listed for each indicator for each of the variability scales (Basin, Marsh, etc.). Bold numbers indicate
that the probability is significant at the 0.05 level. Results are upon Type HI sums of squares. Na = level not applicable
(only one accretion core per site). Site and Site Rep are considered to be random effects; Basin and Marsh. Health are
considered to be fixed effects. The general  model is: Indicator = Basin, Health, Basin*Health, Site(Basin Health), Site
Rep(Slte Basin Health).  Parentheses indicate nesting; asterisk indicates interaction.  These are the results from the
General Linear Model Procedure [(PROC  GLM), (SAS, 1988)].
EM.AP Draft Report -1994
Page 84

-------
I. Indicators Used;

  1. Total Biomass          -          •         ,  •.
 2. Water Cover

II, Test of Homogeneity of Within Covariance Matrices:

 Chi-square value = 7,11 with 3

 Chi-square significant al the O.KHevel, within matrices used

 Classification based on Quadratic Discriminant Function

III. Re-substitution Summary:
                                        	_HeaJthv_
                 From
                                Healthy
                               Impaired
IV. Cross-Validation Summary:,:
                 From
                                Healthy
                            '  ' Impaired
                                                                  To
                                                   81.8
                                                    0.0
                        .JffiESMSL

                            18.2
                           100.0
                                                       Total Error Count=0.0833
                                                                  To
81.8
23.1
                            18.2
                           76,9
                                                       Total Error Count=0,2083
Table 4.15  Discriminant Analysis results, using total biomass and water cover. The results of the test
of the homogeneity within the co variance matrices are listed under heading II, along with the method
used, linear or quadratic. The classification results are shown under headings III and IV which
present the percent of observations assigned to each class for (1) the re-siibstitutiori classification    .
(Heading III) and (2) the cross-validation (Heading IV).  The total error count rate is also indicated for
each of the classifications.,    '   '
                     -1994
                                                      Page 85

-------
 The best classification model, without cover and
 biomass, is presented in Table 4.16. This
 discriminant model gives classification results
 that are comparable to using biomass and cover.
 The marshes can be classified into healthy or
 impaired with an error of ~29% for the healthy
 and 22% for the impaired using the following
 variables:

 1.      The sum of the stem diameters
 2      The log of the number of tassels (stems
        with seed heads)
 3.      The log of the sulfide concentration
 4      The log of the "hydraulic conductivity"
 5.      The log of the sediment sulfur
        concentration.

 A log transform was used for those indicators
 that showed a log-type distribution, based upon
 inspection of the data distribution.

 The results of the Canonical Discriminant
 Analysis for this model are shown in Figure
 4.13. The healthy  and impaired sites separated
 using the first two  Canonical variables, with the
 first canonical variable again having the most
 discriminatory power. Although this model
 seems reasonable,  it still needs to be verified.
 This verification can be accomplished by either
 using part of the data to develop the model, then
 testing it with the remaining data, or by
 collecting a new data set. We feel that the latter
 approach should be used, because the data set is
 fairly small and to  split it would not leave much
 data for the analysis. This verification can be
 accomplished by applying the model developed
 during this Pilot Study to the future data
 collection efforts.
measurements of water levels, based upon results
from a previous study (Wisemann and Swenson,
1988). The results from this study which
analyzed water levels measured in a brackish
marsh system along a transect stretching from
the bayou to 75 meters inland are shown in Table
4.17. The analyses indicated that the water
levels within the internal marsh were highly
correlated with each other (R>0.94) but had a
weaker correlation (R<0.80) with the water
levels in the bayou at the time scales  used in the
analysis (half-hour sampling intervals). A more
detailed time-series analysis of the data indicated
that, although there were weak coherences (-0.6)
at short time scales (tidal period and shorter),
there was also an indication of higher coherences
at longer time scales (weeks); however, the time
frame was too short (1 month) to assess this with
any degree of confidence. Clearly, spot
measurements of water levels over short time
scales are not adequate to characterize the water
level regimes in these marshes. We monitored
water levels in the marsh and adjacent bayou for
a longer time period (up to 8 months) to obtain a
reliable estimate  of marsh inundation.
4.7 HYDROLOGY

The hydrologic parameter of interest was marsh
inundation.  This was estimated based upon
analysis of time-series water level data collected
at 8 sites. Time-series measurements at the
sample sites were used rather than spot
EMAP Draft Report - 1994
                                  Page 86

-------
                                 CDA:  Total Biomass, Wafer Cover


0>
JQ
.2
1
m
.2
"S
o
to
O





3.0-

2.0-
1.0-

0.0-

-1.0-

-2.0-

*

* / ' *
m *
' o ' ' ' • *
'o *SL!%-
o 3
* o . •- . .
O ^RJj
^r* ^J
1 1 1,1 1
-3.0 -2.0 -1.0 0.0 1.0 2.0 3
                                          Canonical Variable 2
            Figure 4.12 Results of Canonical Discriminant Analysis (CDA) on the 1991 EM AP Wetlands Southeast Pilot
            Study data. The plots show the distribution of the first two canonical variables as a function of marsh health
            class. The indicators used to derive the discriminant models are indicated at the top of the figure.
EMAP Draft Report -1994
Page 87

-------
I. Indicators Used:
1. Log(Diameter)
2. LogCTassels)
3. LogCsulfidc)
4. Log(scdiment sulfur)
5. Log(HQ
II. Test of Homogeneity of Within Covariance Matrices:
Chl-square value = 19.5 with 15 DF; Prob>Chl-square = 0.19
Chl-square not significant at the 0.10 level, pooled matrices used
Classification based on Linear Discriminant Function
III. Re-substitution Summary:
Healthy
From
Impaired
Healthy
71.4

0.0
To
Impaired
28.6

100.0
Total Error Count=0.125
IV. Cross- Validation Summary:
Healthy
From
Impaired
Healthy
71.4

22.2
To
Impaired
28.6

77.8
Total Error Count=0.2500
     Table 4.16 Discriminant Analysis results, using a combination of vegetation and soil indicators. The
     indicators used in the models are listed under heading I. The results of the test of the homogeneity
     within the covarlance matrices are listed under heading II, along with the method used, linear or
     quadratic. The classification results are shown under headings III and IV which present the percent of
     observations assigned into each class for (1) the re-substitution classification (Heading III) and (2) the
     cross-validation (Heading IV). The total error count rate is also indicated for each of the
     classifications.
•EM.AP Draft Report -1994
Page 88

-------
                               CDA: Total Diameter, Log (Tassels),
                           Log (Sulfide), Log (HC), Log (Sediment S)
                                      HEALTHY
O  IMPAIRED

^_
J>
JO
eg
1
o
s

3.O-1
2.O-
1.0-
o.o-
-1.O-

-2.0-

O
o
o
° 0 °
* o • • '
o • ' ' '°

• ••*
*
-d.o-i | 	 i 	 | 	 - 	 i • •• | 	
-3.O -2.O -1.O- O.O 1.O 2.0 3.
                                        Canonical Variable 2
           Figure 4.13 Results of Canonical Discriminant Analysis (CDA) on the 1991EMAP Wetlands, Southeast Pilot
           Study data. The plots show the distribution of the first two canonical variables as a function of marsh health
           class. The indicators used to derive the discriminant model are indicated at the top of the figure.
EMAP Draft Report -1994
                             Page 89

-------

LEVEE

5 METERS

10 METERS

35 METERS

75 METERS

BAYOU
0.68
.nd
0.69
0.41
0.82
0.19
0.80
0.38
0.75
0.48
LEVEE


0.80
.nd
0.88
.nd
' 0.88
Jid
0.86
.nd
5 METERS




0.91
0.33
0.94
0.47
0.95
0.72
10 METERS






0.99
0.13
0.98
0.29
35 METERS








0.99
0.68
Table 4.17 Correlation matrix of water level and salinity signals as a function of distance into the marsh.
Indicated for each distance arc the Pearson Correlation Coefficients for water levels (top number) and for
salinity (bottom number). The data are from time series deployment in a Louisiana brackish marsh
(Raccourci Bayou) from OSMayS? thru 04June87. The sampling interval was 0.5 hours (Wiseman and
Swcnson, 1988).
EMAP Draft Report -1994
Page 90

-------
        5  CONCLUSIONS AND  RECOMMENDATIONS
The evaluation of indicators and stud}' design
were successfully used for the first time in .the
1991 field season. We exceeded the minimum
sampling scheme in 1991, determined the
indicator variability, the sampling error, and
found several potential indicators of plant health
for S. alterniflom. The intended low level of
sample variance was achieved in almost all
sampling. Most of the original indicators were
justifiable choices, based on the literature and
ongoing research results. A few indicators were
inadequate or likely to be too difficult to
implement with the framework of a regional
sampling scheme.

A first-order goal of EMAP is to! address
questions of inventory, i.e., is the resource there?
For wetlands, this includes  an inventory of the
areal extent and the biomass of the resources. To
a certain extent, the presence or absence of
biomass is a stress indicator.  Plant cover,
biomass and morphometric indicators can also
be used in this effort. A second set of questions
for EMAP involves the health of the biomass
present. We have shown that there are additional
indicators of change or stress that simple
biomass parameters do not reveal. Soil
conditions, stem morphology and the density of
reproductive structures can be monitored to
follow plant condition. To  the extent that
ecosystem health is important, these factors may
themselves prove useful as  indicators of faunal
community health.

Soil reducing conditions have an important
effect on plant health in laboratory and field
experiments.  We have uncovered some
interesting relationships among plant conditions
and both soil sulfide and total S concentration
that are undoubtedly responsive to physical,
biological and geological factors.  It may be
enough to find reasonable indicators of
long-term changes, while understanding only
some of these cause and-effect relationships. If
long-term changes are identified through EMAP,
then a more thorough investigation of the causal
mechanisms of change may be warranted.  Some
new or additional indicators can be developed
within the context of the emerging and evolving
goals of this rather young EMAP program.

Below is a brief summary of key
recommendations resulting from this first year's
field sampling.
5.1 RESPONSE INDICATOR
   DEVELOPMENT

»  Plant morphology and structure (e.g., stem
width and reproductive structures) are
potentially biomass-independent indicators of
stress.

•  Soil properties (e.g., eH, bulk density, carbon,
hydraulic conductivity, sulfide and total S) are
sources or consequences of stress that are easily
measurable and probably essential properties to
measure in EMAP. Interpreting the significance
of variations in these properties  requires
additional measures that may eventually be
reduced (e.g., accretion rates, water level, etc.).

•  Accretion rates are important for evaluating
controlling factors causing plant stress. These
soil property data should be used to address
questions about long-term marsh accretion and
the relationship between biomass and accretion
rates. These relationships remain prevalent
issues for both indicator development and
resource management.
EMAP Draft Report - 1994
                                  Page 91

-------
 • Pre-sampling aerial surveys (photography)
 should be made available for site selection and
 logistical support, and a 2 or 3 segment historical
 comparison of the sites is very informative for
 determination of whether the sites are healthy or
 impaired.

 » Installation of water-level gages may be too
 costly and time-intensive to continue for all
 sites, but water level is an essential measurement
 to continue in some fashion, if only to determine
 Important relationships among stressors and
 plant responses. It may be informative to
 examine the tide gage records of nearby field
 sites or to choose field sites for indicator
 development on the basis of their proximity to
 good tide gage records.

 • It is very cost-effective to collect some soil
 samples  for archival purposes. The toxic effects
 of pollutants are frequently a threat, and this data
 could be integrated with the other EMAP studies
 (e.g,, EMAP Estuarine). It may be good to
 include a screening for some organic pollutants
 for the same reason.  Furthermore, the
 constituents might provide a direct signal
 concerning marsh health and/or a basis for
 interpretive indicator responses.

 • This study was initiated as a preliminary
 attempt at identifying whether spectral
 reflectance measurements of the marsh surface
 from  a helicopter platform could be used to
 assess marsh health and would, thus, warrant
 continued investigation.  The results presented
 above indicate that differences in marsh vigor
 may be definable with this technique.  However,
 the sources of variation in the data must be
 identified and a larger number of sampling
 stations must be employed.

 » pH measurements appear to have little use in
 the program. The variability among sites of
 contrasting conditions was low.
5.2 SAMPLING EFFICACY

• Sampling efficacy may be improved by
investigating the relationship between sample
frequency and variability. For example, there are
two waysio improve upon the previous sampling
efforts estimating plant biomass. One is to
sample fewer plots, and the other is to further
develop morphometric measures for
non-destructive sampling. Modification of
sampling scheme will reduce effort with a small
loss of replicability. Specifically, the number of
replicates for biomass harvest can be reduced
from 6 to 5 plots. This should be examined
further and may have a potentially IongTterm
consequence for field sampling efficiency.
5.3 ADDITIONAL INDICATORS

»  EMAP-Wetlands has expanded its scope
beyond monocultural coastal wetlands to include
ecosystem health and general resource condition
of more diversified coastal wetlands.  In practice,
this may mean that indicators of fish habitat
quality (for example) are appropriate areas for
indicator development.

»  Non-destructive sampling techniques are
desirable, especially in light of the desirability of
long-term landowner cooperation.

»  Below-ground biomass is a potentially
important parameter to measure in subsequent
studies.
EMAP Draft Report - 1994
                                  Page 92

-------
5.4  EXPANSION TO OTHER
   REGIONS  ''    '"; '':   "      '"'  '  '

• .Indicator development for individual species
of homogenous maerophyte cover will be easier
to expand than it will be for heterogeneous plant
cover.  There is a drastic change in species
dominance  going, from salt to freshwater
marshes. The difficulties involved in sampling
the brackish marshes are much greater than those
involved in monotypic salt marshes. Caution is
urged in expecting too much top soon when
expanding the vegetation types analyzed, from
salt marsh to other plant communities. The
end-members conditions  (healthy and impaired)
for S. alternifloramaynot.bv estimated by the
same parameters for all species. In fact, it is
unlikely that is the case.

•  The Louisiana province is not necessarily
representative of all  salt marsh sites. This means
that coastal wetland  monitoring activities in
other Gulf states are likely to present different
geophysical conditions affecting plant
community health.

•  The response of plants to a stressor is not
necessarily linear. There may be a threshold
effect (e.g., to tidal energy or submergence) or an
optimum response level (e.g., to a pollutant,
sulfide or' salinity).  The range of conditions "
found in the Louisiana field trials may not
represent all ranges of factors affecting the status
of plant health.  For  these reasons and others, it
is prudent to continue investigation of any
indicators showing even minimal likelihood of
success.

•  Soil salinity was never an important
component of any of the  statistical cluster or
discriminant analyses. However, it may be an
especially important parameter to include in Gulf
of Mexico-wide sampling, in view of the
hypersaline conditions anticipated in Texas
estuaries.
5.5 SUMMARY TABLE   ...    ......  v

A summary of.the utility of the indicators
selected to reflect wetland condition is shown in
Table 5.1.
EMAP Draft Report -1994
                                  Pcige93

-------
SALT MARSH INDICATOR DEVELOPMENT RESULTS

1. Soil Parameters
a. Salinity
b. Bulk density
c. Percent organic
d. Sulfidc
c. pH
f. cH
g. Hydraulic conductivity
h. Water levels
i. Chemical constituents - trace metals
j. Chemical constituents - nutrients
k Sediment/organic accumulation
2. Vegetation Parameters
a. Cover
b. Biomass
c. Stem density
d. Stem length
c. Stem diameter
f. Chemical constituents -trace metals
g. Chemical constituents - nutrients '
h. Species presence
3. Other, new approaches
Indicators
of
Condition

?(s. Tx.)

X
X

?
X

?
X
?

X
X
X
X
X
?
?
?

Useful
for
Interpretation

X
X
X
X

X
X
X

X
X

X
X
X
X
X
? ..
?
X

Probable
Community
Change Indicator

X
X
X
X

7
X
X
X
X
X

X
X
X.
X
X
?
:••>
XX

New
Direction
Possible







refine method?

S fractions



light wand
light wand


weight/x-sec.



macrobenthos dendritic network fish
PRES./ABS. below ground biomass grain size
Table 5.1 Summary of indicator evaluations.
EMAP Draft Report -1994
Page 94

-------
SALT MARSH INDICATOR DEVELOPMENT RESULTS
'
1. Soil Parameters
a. Salinity
b. Bulk density
c. Percent organic
d. Sulfide
c. pll
f. eH
g. Hydraulic conductivity
h. Water levels
i. Chemical constituents - trace metals
j. Chemical constituents - nutrients
k. Sediment/organic accumulation
2. Vegetation Parameters
a. Cover
b. Biomass
c. Stem density
d. Stem length
e. Stem diameter
f. Chemical constituents - trace metals
g. Chemical constituents - nutrients
h. Species presence
Non-destructive
sampling
Method

?(s. Tx.)

X
X

7
X

?
X
7

X
X
X
X
X
7
, ?
7.
Not recommended
for
Regional Stage II

X
X
X
X

X
X
X

X
X

X
X
' X
X
X
?
?
X
Archive
for
Possible Use

X
X
X
X

7
X
X
X
X
X

X
X
X
X
X
?
7
XX
New
Direction
Possible







refine method?

S fractions



light wand
light wand


wcight/x-sec.



Table 5.1 (cont.) Summary of indicator evaluations.
EMAP Draft Report -1994
.Page 95

-------

-------
                                 6  REFERENCES
Abacus Concepts, 1987. StatView II, StatView SE+Graphics, Abacus Concepts, Inc.  Berkeley, CA, 234
   pp.

APHA-AWWA-WPFC. 1975. Standard Methods for the Examination of Water and Wastewater,
   Fourteenth Edition, American Public Health Association.  Washington, DC.  1193  pp.

Barber et al., 1993 This reference was used by Kevin Summers - who will provide the reference

Buresh, R. J. 1978.  Nitrogen transformations and utilization by Spartina alterniflora in a Louisiana salt
   marsh. Ph.D. dissertation, Dept. Marine Sciences, Louisiana State University, Baton Rouge, LA. 118
   pp.
                                _(
Chabreck, R. H. 1972. Vegetation, water and soil characteristics of the Louisiana coastal region.
   Louisiana State Univeristy  Agricultural Experiment Station Bull. 664.

Chabreck, R. H. G. Linscomb.  1978. Vegetative type map of the Louisiana coastal marshes.  Louisiana
   Dept. of Wildlife and Fisheries and Louisiana State University.  Baton Rouge, LA

Evers, E. D., J. G. Gosselink, C. E. Sasser, and J. M. Hill. 1992. Wetland loss dynamics in southwestern
   Barataria basin, Louisiana (USA), 1945-1985. Wetlands Ecology and Management 2:103-118.

Gosselink, J. G., C.  S. Hopkinson, and R. T. Parrondo. 1977. Common marsh plant species of the Gulf
   coast area. Vol. I Productivity, Vol. II  Growth dynamics. Environmental Effects Laboratory, U.S.
   Army Engineers  Waterways Expt. Sta.,  Vicksburg, MS.  Tech. Rept. D-77-44.

Hardisky, M.A., F. C. Daiber,  C. T. Roman and V. Klemas. 1984.  Remote sensing of biomass and annual
   net aerial primary productivity of a salt marsh. Remote Sensing of Environment 16:91-106.

Hoar, R. J. 1975. The influence of weirs on soil and water characteristics in the coastal marshlands of
   southeastern Louisiana, unpubl. M. S. thesis, School of Forestry and Wildlife Management, Louisiana
   State University, Baton Rouge, LA. 94 pp.

Hopkinson, C. S., J. G. Gosselink, and R. T. Parrondo. 1978. Aboveground production of seven marsh
   plant species in coastal Louisiana. Ecology, 59:760-769.

Hopkinson, C. S., J. G. Gosselink, and R. T. Parrondo. 1980. Production of coastal Louisiana marsh plant
   calculated from phenometric techniques. Ecology, 61:1091-1098.
EMAP Draft Report -1994	Page 97

-------
 Kaswadji, R., J. G. Gosselink and R. E. Turner 1990. Estimation of primary production using five different
   methods in a Spartina alternifiora salt marsh. Wetlands Ecology and Management  1:57-64.

 Kirby, C. L 1971. The annual net primary production and decomposition of the salt marsh grass Spartina
   allcrniflora Loisel. in the Baralaria Bay estuary of Louisiana. Ph.D. diss., Dept. Botany, Louisiana
   State University, Baton Rouge, LA. 74 pp.

 Lcibowit/., N. C., L. Squires and J. P. Baker  1991. Research Plan for Monitoring Wetland Ecosystems.
   U.S. Environmental Protection Agency, Office of Research and Development, Environmental
   Monitoring and Assessment Program,  Environmental Research Laboratory, Corvallis, Ore. 157 pp.
   -f-Appendices.

 McKcc.K.L., I. A. Mendelssohn,  and K. Ewing. 1990. An evaluation of methods to identify and quantify
   sublclhal stress in wetland vegetation.  Final report to the National Wetlands Research Center, U.S.
   Fish and Wildlife Service, Department of Interior, Lafayette, LA.  100pp.

 Morris, J. T., B. Kjcrfve, and J. M. Dean 1990. Dependence of estuarine productivity on anomalies in
   mean sea level.  Limnol. Oceangr. 35:926-930.

 SAS Institute Inc. 1988. SAS/STAT User's Guide:  Release 6.03 Edition. SAS Institute, Inc. Carry, NC
   1028 p.

 SAS, 1990a.  SAS Procedures Guide, Version 6, Third Edition. SAS Institute, Inc., Gary, North Carolina.
   705 pp.

 SAS, 1990b.  SAS Language Reference, Version 6, First Edition. SAS Institute, Inc., Gary, North
   Carolina.  1042pp.

 SAS, I990c.  SAS/Graph Software Reference, Version 6 First Edition, Volume 1. SAS institute, Inc.,
   Gary, North Carolina.  1341 pp.

 SAS, 1990d.  SAS/Graph Software Reference, Version 6 First Edition, Volume 2. SAS Institute, Inc.,
   Gary, North Carolina. 605pp.

 SAS, 1990s.  SAS /STAT User's  Guide, Release 6.03 Edition. SAS Institute, Inc., Gary, North Carolina.
   1028 pp.

 Smith, F., S. Kulkami, L. E. Myers and J. J. Messner.  1988.  Evaluating and presenting quality assurance
   sampling data, pp. 157-168. in: L.H. Keith, ed. Principles of Environmental Sampling. American
   Chemical Society, Washington, DC.

 Stanley, T. W. and S. S. Vemer.  1985. The U.S. Environmental Protection Agency's quality assurance
   program, pp. 12-19.  In J. K. Taylor and T. W. Stanley, eds. Quality Assurance for Environmental
   Measurements. American Society for Testing and Materials, STP 867, Philadelphia, Pennsylvania.

 Sliven, A. E. and E. J. Kuenzler.  1979.  The response of two salt marsh molluscs, Littorina irrorata and
   Guekensia demissa, to field manipulations of density and Spartina litter.  Ecol.  Monogr. 49(2):151-171.


EMAP Draft Report-1994                                                         Page 98

-------
Swensbh, E. 1983. Marsh hydrological studies 1982-1983 data report. Coastal Ecology and Fisheries
   Institute, Center for Wetland Resources, Louisiana State University, Baton Rouge, La.
... LSU-CEFI-83-18.    ,                          .                                    '

Swenson, E. M. 1982.  A report on the Catfish Lake, Louisiana backfilling study. Coastal Ecology
   Laboratory, Louisiana State University.  Prepared for National Marine Fisheries Service, Southeast
   Region,  St. Petersburg, Florida. Contract Number NA81-BA-P00006. Coastal Ecology Publication
   LSU-CEL-82-25

Swcnson, E.M., J. M. Lee, and R. E. Turner 1992a. Quality Assurance Project Plan, 1991 EMAP Wetlands
   Southeastern Pilot Study. US EPA, Corvallis, 1992.

Swenson, E.M., J. M. Lcc, and R. E. Turner 1992b. Field sampling data report, 1991 EMAP Wetlands
   Southeastern Pilot Study. US EPA, Corvallis, 1992.

Taylor, J. K. 1988. Quality assurance of chemical measurements. Lewis Publishers:  Chelsea, MI 328 pp.

Taylor, J. K. 1990. Statistical Techniques for Data Analysis. Lewis Publishers, Inc.  Chelsea, MI, 48118.
   200pp.

Thomas, J.  D.  1976. A survey of gammarid amphipods of the Baralaria Bay, Louisiana region.  Conlr.
   Mar. Sci. 20:87-100.

Tucker, CJ. 1979. Red and photographic infrared linear combinations for monitoring vegetation.  Remote
   Sensing  of Environment 8:127-150.

Turner, R. E. and J. G.  Gosselink 1975.  A note on standing crops of Spartina alterniflora in Texas and
   Florida.  Contr. Mar. Sci. 19:113-118.

Turner, R. E. 1990. Landscape development and coastal wetland losses in the northern Gulf of Mexico.
   American Zoologist 30:89-105.

Turner, R. E. 1991. Tide gage records, water level rise and subsidence in the northern Gulf of Mexico.
   Estuaries 14:139-147.                           ,

Turner, R. E. and Y. S. Rao 1990. Relationships between wetland fragmentation and recent hydrologic
  , changes  in a deltaic coast. Estuaries 13:272-281.

Wiseman, W. J. Jr., and E. M. Swenson.  1988. Measurement of Saltwater movement in a Marsh System.
   In: Causes of Wetland Loss in Coastal Central Gulf of Mexico. Volume II:  Technical Narrative. (R. E.
   Turner and D. R Cahoon, eds).. Final Report submitted to Minerals Management Service, New
   Orleans, LA. Contract No. 14-12-0001-30252. DCS Study/MMS 87-0120. 400 pp.

Wiseman, W. J. Jr., E. M. Swenson, and J. Power. 1990.  Salinity Trends in Louisiana Estuaries. Estuaries
   13:265-271.
EMAP Draft Report -1994                                                         Page 99
                                 «t).S. GOVERNMENT PRINTING OFFICE: l 995-650-006/2208 3

-------

-------