United States National Health Environmental EPA/600/R-96/057
Environmental Effects Research Laboratory May 1996
Protection Agency Corvallis, OR 97333
BACKGROUND AND
RECOMMENDATIONS
FOR ESTABLISHING
REFERENCE WETLANDS IN
THE PIEDMONT OF THE
CAROLINAS AND GEORGIA
-------
EPA/600/R-96/057
BACKGROUND AND RECOMMENDATIONS FOR
ESTABLISHING REFERENCE WETLANDS
IN THE PIEDMONT OF THE CAROLINAS AND GEORGIA
by
Mark M. Brinson1, Wade L. Nutter2, Richard Rheinhardt1, and Bruce Pruitt2
Department of Biology
East Carolina University
Greenville, North Carolina 27858
2Daniel B. Wamell School of Forest Resources
University of Georgia
Athens, Georgia 30602
CONTRACT NUMBER: 3B0978TEX
EPA Project Officer:
Mary Kentula
U.S. Environmental Protection Agency
National Health and Environmental Effects Research Laboratory
Western Ecology Division
Corvallis, OR 97333
U.S. Environmental Protection Agency
Office of Research and Development
National Health and Environmental Effects Laboratory
Research Triangle Park, NC 27711
-------
DISCLAIMER
This project has been funded by the U.S. Environmental Protection Agency
(EPA) and conducted through contract number 3B0978TEX. This document has been
subjected to the Agency's peer and administrative review and approved for publication.
The official endorsement of the Agency should not be inferred. Mention of trade names
or commercial products does not constitute endorsement or recommendation for use.
This document should be cited as:
Brinson M.M., W. L. Nutter, R. Rheinhardt, and B. Pruitt. 1996. Background and
Recommendations for Establishing Reference Wetlands in the Piedmont of the
Carolines and Georgia. EPA/600/R-96/057. U.S. Environmental Research Laboratory,
National Health and Environmental Effects Laboratory, Western Division, Corvallis,
Oregon.
The U.S. Department of Commerce, National Technical Information Service (NTIS)
Accession Number is PB96-176813. Additional copies of the manual are available at
cost from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Phone 1-800-553-NTIS
. in
-------
Acknowledgments
This project benefited from contributions from Dennis Whigham who shared with
us his previous experience with wetlands in the Piedmont of North Carolina and
elsewhere. Mike Schafale and Alan Weakley of the North Carolina Natural Heritage
program provided suggestions for reference sites.
in
-------
CONTENTS
Acknowledgments . . . Hi
List of Tables v
List of Figures vi
Abstract vii
1.0 INTRODUCTION AND PURPOSE OF STUDY 1
1.1 The Southern Piedmont Region 2
1.2 Objectives of a Reference Wetland Approach 2
1.3 Problems in Developing Information and Standards for Wetlands 6
2.0 CHARACTERISTICS OF PIEDMONT WETLANDS 8
2.1 Hydrology and Geomorphic Setting 9
2.2 Vegetation and Soils 11
2.3 Gradients of Natural Variation and Disturbance 14
2.4 Fish and Wildlife Habitat 15
3.0 APPROACHES FOR IDENTIFYING REFERENCE WETLANDS 17
3.1 The Terminology of Reference 21
3.2 Recommended Subclasses for Piedmont Riverine Wetlands 25
3.3 Strategies for Choosing Reference Wetlands 30
3.4 Locating Reference Streams and Wetlands in the Same Place 35
3.5 Problems with Non-jurisdictional Wetlands 36
4.0 DATA COLLECTION AND ANALYSIS 38
4.1 Field Collection 38
4.2 Data Reduction and Statistical Analysis 39
4.3 Implementation and Field Testing 40
5.0 CONCLUSIONS AND RECOMMENDATIONS 42
6.0 LITERATURE CITED 43
APPENDIX 50
IV
-------
List of Tables
Table 1. Parameters that may be useful in characterizing reference wetlands 18
Table 2. Categories and nomenclature for reference 22
Table 3. Major watersheds of the Piedmont 26
-------
List of Figures
Figure 1. Flows of the Congaree River at Columbia, South Carolina 10
Figure 2. Location of the Piedmont in the eastern United States 23
Figure 3. Relative importance of overbank flow to riparian source as a function of
stream size or position in the drainage 29
Figure 4. Sampling design and approach for identification and classification 31
Figure 5. Sampling design and approach for project 32
VI
-------
Abstract
The Piedmont of the Carolines and Georgia is the most rapidly urbanizing
physiographic province of the Southeast, excluding Florida. Bottomland hardwood
forests in river floodplains are the dominant wetland type. These riverine forests are
not only being subjected to urbanizing pressures, but are also in a state of transition
due to sediment redistribution initiated over a century ago. One of the consequences
of this transition is that many floodplain areas are not currently jurisdictional wetlands,
a status that weakens efforts to protect riparian zones for their widely recognized
contributions to the maintenance of water quality. Research activity and technical
information on these ecosystems is quite sparse, thus exacerbating the difficulty of
developing management options and implementing actions that are consistent with
maintaining water quality.
Human activities have had profound effects on riverine wetlands of the region.
By the Civil War, conversion of forest to agriculture was virtually complete. Continued
poor farming practices through the early part of this century caused the highly erodible
upland soils to fill existing floodplains and even bury bridges and dams. Subsequent
recovery of uplands to forest vegetation reduced sediment supplies and shifted streams
to an incising rather than an accreting mode. Stream incision and other hydrologic
alterations have increased channel capacity for water conveyance and correspondingly
have reduced flooding frequency of wetlands due to a diminished source from overbank
flow. Sites that continue to be wetlands are dominated instead by ground water and
surface water from adjacent uplands. Others floodplains are increasingly influenced by
beaver dams. These situations have led us to provisionally classify Piedmont riverine
wetlands into three subclasses: overbank flow-dominated, riparian-source dominated,
and beaver-dam dominated. Except for the latter, all are forested when unaltered.
Where floodplains are well developed, streamside levees are dominated by river birch,
sycamore, hackberry, silver maple, and ironwood. Wettest areas in backswamps are
dominated by overcup oak. Intermediate conditions of wetness support a variety of
hardwood species including ash, willow oak, swamp white oak, red maple, American
elm, and sweetgum. Common soils in floodplains are Congaree, Tawcaw, Chastain,
Wehadkee, and Chewacla.
Reference wetlands can play several roles in the management of riverine
wetland in the Piedmont. The first is that they can serve as standards against which
environmental degradation and compensatory mitigation can be measured through
functional assessment. Without such standards, it is impossible to keep track of
functional changes, positive and negative, from a single project or cumulative changes
from multiple projects over time. Reference wetlands may play other roles such as
serving as endpoints for water quality standards, as templates for restoration, and as
sites for training personnel in functional assessment and mitigation. Terminology for
VII
-------
reference wetlands has been developed to facilitate communication of these various
uses, and to be congruent with current wetland regulatory practices.
Protocols have not yet been developed for choosing reference wetlands. There
are several concepts, however, that are unlikely to change with standardization. First,
separation into hydrogeomorphic subclasses is essential to distinguish natural variation
and disturbance from variation instigated by human-induced impacts. Within each
subclass, the geographic range over which the subclass is applicable should be
restricted likewise to limit variation in species composition of dominants. Second,
stratified random approaches keying on descending size categories — ecoregions,
watersheds, drainages, stream orders, and stream reaches - provide elements for
organizing sampling using maps and field surveillance. Best professional judgement
and local knowledge weigh heavily in any successful identification and sampling
program for reference wetlands. Third, reference wetlands within a defined subclass
should include a range of site alterations for two purposes: to identify those wetland
sites that are considered the most highly functioning and to gain knowledge on the
condition of altered wetlands that may be good candidates for significant functional
gain through restoration. Finally, the most highly functioning wetlands of a subclass
are used to set reference standards that become benchmarks for measuring functional
change or departure from reference standards.
There are both commonalities and differences between the reference concept
for wetlands and reference as applied to biomonitoring of streams. These are
discussed and evaluated in light of the different applications of "reference" in the two
cases. Notwithstanding differences, it is recommended that efforts in establishing
stream biomonitoring sites and reference wetlands sites be coordinated for purposes of
efficiency and gaining perspective on the interaction between streams and riverine
wetlands of the Piedmont. Water quality standards for wetlands can be developed
using reference wetlands in parallel fashion as developed for streams. By
encapsulating physical, chemical, and biotic characteristics of wetlands as ecosystem
functions, functions can be translated directly into designated uses and biocriteria that
are consistent with state and federal environmental statutes.
VIII
-------
1.0 INTRODUCTION AND PURPOSE OF STUDY
Natural wetlands of the Piedmont of the Carolinas and Georgia are found largely
in river bottoms ( Costing 1942, Wharton et al. 1982). The magnitude of the resource
is not trivial. In the Piedmont of Georgia alone, Kundell and Woolf (1986) estimated by
Landsat imagery 143,480 acres (58,066 .ha) of forested wetlands. Bottomland
hardwood forests are the dominant ecosystem type for Piedmont wetland areas in the
Carolinas and Georgia.
The Piedmont is the most industrialized and urbanized of the physiographic
provinces in the Southeast (excluding south Florida). Rapid development has occurred
in spite of the fact that the cropland of the Georgia Piedmont, for example, has
decreased by 1982 to half of what it was in 1935, being replaced largely by forest
(Turner 1987, Odum and Turner 1990). Increases in demand for water, expansion of
impermeable surfaces, construction of additional reservoirs, increases in land use
intensity, and discharges of eutrophying and toxic pollutants are all potential stressors
affecting Piedmont wetlands. Because urbanizing conditions are generally intensified
on small areas of land, they do not show up as significant losses of wetlands relative to
other more extensive uses (Hefner et al. 1994). However, increasing urbanization and
sub-urbanization places greater demands on surface water and increases the volume
of wastewater discharges. The consequences are that water quality suffers, wetlands
are altered, and stream flows become modified. In each case, or a combination of
them, habitats for aquatic organisms are undergoing deterioration and predators
dependent on these aquatic species are losing their food base.
In stark contrast to the high rate of urbanizing activity in this region, research
activity on wetland ecosystems is sparse. Of the few written descriptions of Piedmont
bottomlands that exist, most have focused on species composition of vegetation
(Oosting 1942; Peet and Christensen 1980; see also review by Sharitz and Mitsch
1993). Consequently, the greatest and most widespread impacts to wetlands are .
occurring on ones for which we have the least knowledge. This makes it difficult to
justify their management and regulation in all but the most general way.
Comprehensive ecological and hydrological studies necessary for more exacting
management would require many years to elevate our understanding of Piedmont
floodplain wetlands to the level that exists for the Coastal Plain, Gulf Coast, and
Mississippi lower alluvial valley. State and federal agencies with water resource
mandates cannot wait to make decisions until such studies are proposed, funded, and
carried out. Managers with water protection authority are, therefore, forced to apply
principles of adaptive ecosystem management and environmental protection that have
been developed in other classes of wetlands. If our understanding of Piedmont riverine
ecosystems is to advance rapidly and efficiently, we cannot rely on chance to get the
job done. A carefully thought-out strategy is needed. One potentially fruitful approach
is identifying and using reference wetlands as a possible shortcut for closing the gap
between assembling critical technical information and applying that information in water
quality programs.
-------
This report outlines many of the issues that must be dealt with when identifying
reference wetlands in the southern Piedmont. Reference wetlands are needed to
provide a basis for comparing the functions of wetlands for a variety of management
applications. Foremost in these applications is the Clean Water Act Section 404
program administered by the U.S. Army Corps of Engineer and the Environmental
Protection Agency. Part of the review process for applications to alter wetlands must
determine whether the permitted activity significantly degrades water resources.
Reference wetlands can provide a framework for determining significant degradation
from existing conditions. In the development of reference wetland concepts, we rely
heavily on principles developed for the hydrogeomorphic classification of wetlands
(Brinson 1993a), research conducted elsewhere in the country on bottomland
hardwood forests (Wharton et al. 1982, Brinson 1990, Sharitz and Mitsch 1993), and
the handful of efforts elsewhere on characterizing the structure and function of
wetlands within a reference framework (Kentula et al. 1992, Brinson et al. 1996).
1.1 The Southern Piedmont Region
This report applies primarily to the Piedmont province of North Carolina, South
Carolina, and Georgia, although it may be relevant also to other areas. Most of our
observations are in Georgia and North Carolina. This area is within the Southeastern
Plains Ecoregion of Omernik (1987). Omernik's ecoregion spans the Fall Line to
combine portions of the Coastal Plain and Piedmont. The Fall Line is one of the most
distinctive physiographic features of the Southeast. Above the Fall Line, elevations
range from 150 - 450 m in western portions to 70 - 150 m in the eastern Piedmont.
Further differences across the Fall Line are represented by changes in water table
characteristics, local versus regional sources of groundwater, geologic setting, and soil
morphology (Fetter 1988), species composition of vegetation (Peet and Christensen
1980), and dominance of hydrogeomorphic classes of wetlands. The frost-free season
ranges from 210 to 240 days in the southern Piedmont compared to 230 to 260 days in
the Coastal Plain. The reason for concentrating our efforts to the west of the Fall Line
is to focus attention on the lack of information on Piedmont wetlands relative to ones in
the Coastal Plain. The importance of ecological differences between Piedmont and
Coastal Plain has been demonstrated in establishing reference streams for rapid
bioassessment protocols (Piafktn @i ai. 1383) in the Mississippi/Alabama Ecoregion
Project (Omemik and Griffith 1991; ADEM and MDEQ 1995).
1.2 Objectives of a Reference Wetland Approach
One objective of this report is to explain how reference wetlands might be useful
for state programs mandated to protect the quality of their waters. This wiii be dune by
using the southern Piedmont as the region of focus. Standards for surface waters
routinely are set either by limiting the discharge and concentrations of specific
constituents in point sources or by establishing thresholds below which environmental
-------
conditions are not allowed to fall. Environmental conditions include flow rates as well
as levels of dissolved oxygen, pH, color, and so forth. The development of biotic
indices has become a major tool to determine "ecosystem health" or whether waters are
biologically "impaired" (Loeb and Spacie 1994, Davis and Simon 1995).
Environmental standards for wetlands can be established also. However, there
are important differences between streams and many wetlands. One is that wetlands
tend to be more complex structurally because they often contain both aquatic and
terrestrial components. Because of this structural complexity, a single guild of
organisms such as fish, macroinvertebrates, or diatoms is unlikely to be effective in
encapsulating the condition of a wetland as it may be for streams. Another difference
is that stream ecology has benefitted from a long history of pollution monitoring. While
the measurement of such chemical parameters as dissolved oxygen, pH, and nutrient
content has been the focus over much of this history, biomonitoring is increasingly the
approach of choice. A similar monitoring of wetland ecosystems has not been
conducted, likely because wetlands are not as directly connected to water supply and
wastewater disposal as are streams. Instead, ecosystem characteristics, encapsulated
in the concept of ecosystem functions, provide focal points toward which performance
standards can be set for wetlands.
How do ecosystem functions provide a mechanism for setting standards? Most
functions - such as maintenance of site water balance, primary productivity,
biogeochemical cycling, maintenance of plant community species composition, and
support of food webs - are well recognized and quantifiable processes of ecosystems of
all kinds. Hence, quantitative standards can be established for such functions. The
ecological literature is rife with reports on such studies. It is a small step to tailor
functions that are more specific to wetlands, and to gain even more resolution by
tailoring levels of functioning to specific wetlands classes. This "tailoring" simply
recognizes inherent and broadly recognized differences among wetland classes. Thus,
standards for elemental cycling in tidal salt marshes would differ from those in
ombrotrophic bogs. The capacity of broad floodplain wetlands in the southern
Piedmont to store surface water during flooding would differ greatly from that of side-
slope seep wetlands in mountains. Consequently, wetlands should be classified first
within a geographic region before either the functions can be properly identified or the
standards against which the functions are judged can be estimated. Stated another
way, classification reduces the range of variation within functions for reference
wetlands that would otherwise make difficult the separation of natural variation from
that due to stressors from societal impacts.
What is missing from the foregoing logic toward setting standards is
identification of the unit from which standards are derived. The remainder of this
report is an initial attempt to identify these units for the southern Piedmont. Without
reference wetlands, there would be no consistent or recognizable benchmark from
-------
which departure from standards could be .recognized and measured. Just as pollution
dischargers may be in or Out of compliance with respect to a particular water quality
constituent or biotic index, wetlands can be in or out of meeting reference standards
with regard to their level of functioning. The functioning, then, is established from a
group of Wetlands that are considered to have an appropriately high level of functioning
for a particular class.
Reference wetlands also should have utility for meeting the following objectives:
1.2.1 To use reference wetlands as standards for wetland functions in functional
assessments.
A regional set of reference wetlands could be utilized as templates for initial
design toward which degraded wetlands are restored. Progress toward the
reference standard condition (to be defined later) would be part of the success
criteria.
1.1.2 To establish success criteria appropriate to wetland compensatory mitigation.
The US Environmental Protection Agency (EPA) is currently involved in mapping
and assessing functions and values of wetlands on a large scale throughout the
United States in its Wetland Advance Identification (ADID) program. The
program is coordinated with the U.S. Army Corps of Engineers Section 404
permitting program. Reference wetlands that represent each wetland type within
an ADID study area would provide a means of verifying and calibrating the
hydrogeomorphic functional assessment techniques under development at the
U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg,
Mississippi (WES). This would establish a link between planning efforts in
ADIDs and project impact assessment in the regulatory program.
1.1.3 To provide monitoring sites for evaluating the status and trends of the Nation's
wetland base.
EPA:s Environmental Monitoring and Assessment Program (EMAP) was
implemented in 1988 to "monitor the status and trends of ecological condition
and to develop innovative methods for anticipating emerging problems before
they become crises" (Hunsaker et al. 1990). Wetlands are included in EMAP as
one of the important resource categories for long-term monitoring. The success
of the EMAP is dependent on determining what constitutes normal
environmeniai conditions in each of the resource categories. Rsfsrerics
wetlands in the context of hydrogeomorphic wetland classification would provide
crucial data for this program. In addition, establishing normal environmental
-------
conditions in wetlands is applicable to cumulative impact assessment as
proposed by Leibowitz et at. (1992).
1.1.4 To serve as ecological endpoints for establishing water quality standards,
designated uses, and biocriteria applicable to wetlands.
States are required by the EPA to develop water quality standards for wetlands.
Before designated uses of wetlands can be adequately identified, recognition of
normal environmental conditions in wetlands such as conventional water quality
parameters (e.g. temperature, dissolved oxygen, pH, conductivity, and Eh),
hydroperiod (duration, depth, frequency, and season of flooding), and biocriteria
(species diversity, relative abundance, indicator species, and species richness)
is a first step toward formulating standards applicable to wetlands. In addition,
sediment quality and rate of deposition in wetlands has received little attention,
but would likely be important for identifying normal conditions.
1.1.5 To contribute to whole basin studies.
Several states are engaged in whole basin studies which include both aquatic
and terrestrial components of wetlands, hydrology, agroecosystems, forest,
wildlife, and land use. For example, the South Carolina Department of Health
and Environmental Control has embarked on a whole basin effort entitled
'Watershed Water Quality Management Strategy: Savannah-Salkehatchie
Watershed" (Technical Report No. 002-93). In general, the watershed planning
effort is an attempt to address the interdependence of water quality related
activities associated with a drainage basin. Critical information for watershed
management will be derived from problem identification and prioritization,
ambient monitoring, water quality modeling, planning, permitting, and wetland
quantification, and qualification. Methods for defining a sub-set of reference
wetlands specific to a basin and/or physiographic province will provide data
applicable to such studies. The above studies vary widely in approach, but
several are oriented toward determining cumulative effects, and may use the
Synoptic Approach of Liebowitz et al. (1992).
1.1.6 To serve as templates for the physicochemical restoration of wetlands.
Just as Section 1.2.1 suggested the development of success criteria for
mitigation, reference wetlands represent self-maintaining systems that may be
used as background conditions for restoration of wetlands degraded by
contaminated water, altered flows, and modified channel and floodplain
configuration. Fifty-eight percent of Superfund sites nationwide are located
adjacent to or up-gradient of wetlands. In Region IV-EPA, an estimated 72
percent of Superfund sites are located adjacent to wetlands. It is unknown how
-------
many Piedmont wetland sites are actually contaminated by Superfund sites.
Reference wetlands can provide background data important in establishing
normal environmental conditions in feasibility studies on the remediation of
contaminated sites.
1.1.7 To serve as sites for training new personnel in assessment and mitigation of
wetlands.
To meet several of the foregoing objectives, personnel must be trained to
conduct functional assessments, determine thresholds of ecological conditions
for water quality standards, and so forth. Reference wetlands that are
established in protected areas (sites managed by the Forest Service, Fish and
Wildlife Service, Park Service, and similar lands managed by state and local
governments) are least likely to be degraded further. Altered but recovering
wetlands must be included in a reference wetland set to provide contrast with
ones that are functioning at the highest levels.
1.3 Problems in Developing Information and Standards for Wetlands
A number of issues confound the identification of reference wetlands for the
southern Piedmont. First is the lack of information in the scientific and technical
literature. We have cited fewer than a dozen papers or other works that deal in part
with riverine wetlands of the Piedmont, and several of these were conducted decades
ago. Otherwise similar lists for riverine wetlands of the Coastal Plain would exceed this
several-fold and would include more contemporary studies. The principles of
functioning of riverine wetlands in other parts of the world (Brinson 1990) and the USA
southeastern Coastal Plain in particular (Sharitz and Mitsch 1993), as well as
application of these principles to reference wetlands (Brinson et al. 1996), all have
relevance to the functioning of those of the southern Piedmont. We have chosen not to
include in this report a review of this extensive relevant literature because (1) it exists
elsev.tiere and (2) it may detract from the fact that there is still a critical need to conduct
more fundamental research on wetlands of the southern Piedmont. This research is
needed to verify the function assumed to occur in Piedmont riverine wetlands, to
distinguish among the classes proposed in this report, and to provide information for
the scaling of functions along disturbance gradients.
Another problem is the jurisdictional status of lands on Piedmont floodplains.
Most wetlands in the Piedmont are in floodplains and former floodplains of rivers.
However, not all floodplains are jurisdictional wetlands to which regulatory practices
appiy. Even ones that are currently jurisdictions! wetlands may not be ir, the future if
stream incision and reduced flooding frequency cause them to become drier. The
critical landscape position of floodplains between uplands and stream channels makes
these zones important for water quality whether or not they are jurisdictional wetlands.
-------
As pointed out below, large portions of Piedmont floodplains do not meet the hydrology
and soils criteria (Geiger 1994). In the Chesapeake Bay watershed, these streamside
zones are identified as riparian forest buffers which focuses on their water quality
function rather than on whether they happen to be classified as wetlands (Lowrance et
al. 1995). Because wetland delineation currently identifies areas for regulation by
inundation/saturation, soils, and vegetation (in contrast to function in the landscape),
large portions of Piedmont floodplains are not recognized as jurisdictional wetlands.
This report does not attempt to make the distinction, but treats floodplains as a
geomorphic unit for which wetland reference conditions and standards can be
established. This approach should be applicable even if the jurisdicational definition
changes for political reasons.
Disturbances that are severe and relatively permanent must be separated from
those that are transient. For example, natural disturbances like major floods and
tropical storms may disrupt riverine corridors (Costa 1974), but their effects are
temporary and should not be confused with permanent alterations that lower the
potential of ecosystems to function. Reference wetlands provide a basis for separating
temporary disturbances from permanent alterations.
Finally, the establishment of reference wetlands may suffer from the vagueness
of the term "reference" itself. In section 3.1 we offer some definitions of terms that
should reduce ambiguities. These basic terms have been crafted to facilitate functional
assessment as well. Because reference wetlands may be useful in various programs
described above, it is urgent that common terminologies be developed early.
-------
2.0 CHARACTERISTICS OF PIEDMONT WETLANDS
Human activities during the past 200 years have profoundly affected sediment
dynamics and flow. The conversion of forest to agriculture in the Piedmont was
virtually complete by the Civil War. Forests of the region by that time were reduced in
area to only a small fraction of their coverage today. Piedmont soils were highly
erodible, and mass wasting of uplands filled floodplains and river corridors with
sediment (Trimble 1970a). This erosion was likely due to increased exposure of soil to
direct rainfall, which resulted in less infiltration and higher rates of runoff and erosion.
Small scale subsistence farming following colonization by Europeans was replaced by
an overwhelming dominance of cotton farming after the invention of the cotton gin in
1793. Cotton plantations became seriously eroded resulting in the loss of nearly all
topsoil from 47 percent of the uplands, and gullying on 44 percent of the land (Brender
1974). Floodplains in the Piedmont became filled with soil from the eroding uplands,
thus removing floodplains from use as productive crop and pasture land (Barrows et al.
1917). Trimble (1970a) reports that top soil rapidly filled many of the river valleys up to
18 feet (5.5 m) in depth, not only burying floodplains, but also causing the wholesale
burial of bridges and small mill dams. Stream gradients were reduced, further affecting
flooding regimes (Dowd et al. 1993).
Farmland lost fertility and topsoil, resulting in abandonment of cultivation. When
upland forests became re-established, the sediment supply to floodplains and streams
was reduced. Many of these streams have since incised by eroding downward into the
recently deposited alluvium, and in the process have exported sediment downstream
(Burke 1996). These areas still are not in equilibrium, although many in the upper
reaches have incised to the bedrock which halts downcutting. Natural levees that
formed during the high sediment transport phase are still in place and represent
significant floodplain features. Recovery of uplands to forest vegetation has had the
effect of increasing infiltration and changing the timing of storm discharges. As a
result, more of the water budget is allocated to evapotranspiration and groundwater
recharge. Trimble et al. (1967) estimated that annual water yields have been reduced
from 3 to 10 cm between 1919 and 1967 as a result of reforestation. The Saluda River
above Columbia, South Carolina, for example, showed a decrease in annual yield of
6.6 crn, cr 16.1 percent. Greatest reductions occurred during dry years, explained in
part by the greater access of forest to deeper moisture than was available to shallow
rooted crops which previously dominated agricultural land use. Also forests have
greater transpiration and interception loss than does non-forested vegetation.
Major alterations in hydrology have resulted from creation of impoundments.
• ^T*U»**. I *n*tff* Ik £• . ^•mm. * ^\**M* •A«W*1M •* b*A**AW*A *HMA^Atl*NP* A| ^Hf* tft^A OA|| t*4^ DlWAV 1 ••* O <•*« l4l^
I I iv? i_cir\tj iviui i cay L>cnit v»i iOi i n t/Ov/Ch i iO w^>Oi i HBI wi • n »C \>OiwwO i xi»Gi n i WWMU i
Carolina, for example, shifted the 2-year recurrence interval peak discharge to 4.5
years and the 5-year recurrence interval peak discharge to 25 years (Patterson et al.
8
-------
1985). Thus, flooding frequencies of greater than one year have been substantially
altered by impoundments, an occurrence repeated throughout the Piedmont.
Trimble (1970) theorized that many wetlands were created in the Piedmont by
the raising of channel bottoms and filling of floodplains during the intense period of
sedimentation from culturally accelerated erosion of the late 19th and early 20th
centuries. Burke (1996) observed that many of the Oconee River basin (Georgia)
stream sites visited by Trimble in the late 1960's and described to have "swampy"
floodplains with shallow channels in 1995 were deeply incised and had widened
channels. These observations and others (Richter et al. 1995) lead to the conclusion
that wetland area in portions of the Piedmont river system subjected to heavy
sedimentaiton is decreasing. To verify this conclusion, Burke (1996) compared maps
of riverine systems in Clarke County, Georgia by Barrows et al. (1917) and for the
Mulberry River (subbasin of the Oconee River) by McCallie (1911) with current National
Wetland Inventory maps and found the total wetland area had decreased by 9 and
32%, respectively. Although undoubtably different definitions were used to devine
wetlands, this comparison does appear to confirm the observation that wetland area is
decreasing.
2.1 Hydrology and Geomorphic Setting
As for most riverine wetlands, those in the Piedmont receive varying proportions
of water from overbank flow from the stream channel when discharge exceeds the
capacity of the channel to convey flow and various riparian sources. (Riparian sources
are subsurface discharge of groundwater to the floodplain sediments, overland flow
across the upland to the surface of the floodplain, and small tributary flow from the
adjacent upland onto the floodplain). Many Piedmont wetlands are in transition from
being dominated by overbank flow to a dominance by riparian sources due to changes
in channel morphology as sediment is removed (Burke and Nutter 1995). While both
sources normally occur in most floodplains, it is the relative dominance that is
changing. These sources of water profoundly influence the biota, either through
eliminating potential competitors that cannot tolerate soil saturation and flooding, or
providing aquatic habitat to those that require submergence.
Flows characteristic of the Congaree River at Columbia, South Carolina,
exemplify large Piedmont rivers that have had their flows influenced by impoundments.
Flow typically varies greatly between years and within years. Peak annual discharge of
the Congaree River in Columbia, for example, varies more than either mean or
minimum flows (Figure 1a). The short-term nature of these flows is evident from the
extent to which mean flows are skewed toward minimum flows. The seasonal pattern of
mean monthly discharge illustrates the consistency of an annual cycle of pulsed flow
(Figure 1 b). Depending on channel capacity and elevation of floodplains, breaks or
cuts in natural levees begin to carry overflow into low elevations of floodplains before
-------
£ 200,000
&
fc 175,000
S 150,000
| 125,000
^
•= 100,000
•5 75,000
0)
I 50,000
UJ
I
25,000 -
1940
Maximum
Mean
Minimum
1950
1960
1970
1980
1990
V)
u_
o
68 70 72 74 76 78 80 82 84 86 88 90 92 94
Year (1968-1994)
Figure 1. Flows of the Congaree River at Columbia, South Carolina, (a) Annual
maximum, minimum and mean daily discharges from 1940 to 1994. (b) Mean
monthly discharge from 1968 to 1994.
10
-------
higher flows overtop natural levees and continue to contribute to inundation. Typically,
flooding occurs late in the dormant season after low evapotranspiration reduces the
capacity for additional soil storage, thus contributing to stream discharge. During
extreme floods, the floodplains themselves may carry significant portions of the total
flow.
There are many variations on this general pattern, including more unusual flood
events midway into or late in the growing season, sites that infrequently flood because
channel incision has increased conveyance and channel capacity, and altered flows
and flooding both below and above impoundments. Smaller streams differ by having
sharper flood peaks and shorter periods of floodplain inundation. Seasonality of flows
in small streams does not differ from larger ones.
Flows from riparian sources to the floodplain surface are less well understood.
Overland flow directed to the floodplain can occur sporadically when precipitation
during rainstorms exceeds soil infiltration. Subsurface flow from adjacent saturated and
unsaturated soil horizons is likely to be highly site-specific, depending on such
variables as hydrostatic head, permeability of both alluvial fill and the adjacent upland
matrix, the dimensions of the cross section through which flow occurs, and the change
in slope at the upland/wetland interface (Nutter 1973).
2.2 Vegetation and Soils
Except where beaver ponds or impoundments create high water tables for
extended periods into the growing season or where floodplains have been converted to
agricultural use, Piedmont riverine wetlands are usually forested. In floodplains broad
enough to have a natural levee bordering a relatively flat floodplain, and a lower
elevation next to the upland slope, the vegetation may display several distinct
associations. Oosting (1942) identifies these as birch and sycamore, which would
correspond roughly to the levees, and bottomland mixed hardwood which would
correspond roughly to backswamps and forests of smaller Piedmont streams described
below.
Levees of rivers dominated by overbank flow support canopy trees such as
sycamore (Platanus occidentalis), river birch (Betula nigra), hackberry (Celtis laevigata),
silver maple (Acer sacchahnum), and boxelder (A. negundo). These species are able
to withstand periodic flooding and are able to quickly regenerate sites disturbed by
erosive actions of flooding.
A variety of habitats, influenced by variations in flooding frequency and duration,
exist in back swamps behind levees. The wettest areas of backswamps (oxbows and
abandoned channels) are dominated by overcup oak (Quercus lyrata). Less wet areas
are more species diverse and support a variety of hardwood species, including ash
11
-------
(Fraxinus spp.), willow oak (Q. phellos), swamp white oak (Q. michauxii), and red maple
(-4. rubrum).
Many smaller Piedmont streams (first to third order) are rarely flooded by
overbank flow and instead obtain most of their water from riparian sources and
precipitation. Well-drained portions of former floodplains and riparian slopes are
dominated by mesic species such as tulip poplar (LJriodendron tulipifera), walnut
(Juglans nigra), beech (Fagus grandifolia), hop hornbeam (Ostrya virginiana), wild black
cherry (Prunus serotina), and white oak (Q. alba). Components of the understory
include spicebush (LJndera benzoin), pawpaw (Asimina triloba), flowering dogwood
(Cornus florida), and southern sugar maple (A. barbatum). In the alluvial sites of the
Duke Forest in North Carolina, Peet and Christensen (1980) mention the presence of
the following additional species: sweetgum (LJquidambarstyraciflua), ironwood
(Carpinus caroliniana), beech (F. grandifolia), red elm (Ulmus rubra), and swamp red
oak (Q. falcata var. pagodaefolia).
Depressions on floodplains and seepage (groundwater discharge) areas at the
foot of upland slopes are generally wetter than the more well-drained areas of relic
floodplains. These depressions and seepages support species tolerant of prolonged
saturation such as American elm (U. americana), sweetgum, red maple, sycamore, and
ash. Seepage areas also tend to support a diverse array of herbaceous vegetation,
including such species as lizard's tail (Saururus cernuus), false stinging nettle
(Boehmeria cylindrica), jumpseed (Tovara virginiana), jack-in-the-pulpit (Arisaema
triphyllum), and wild rye grass (Elymus virginicus). Other herb species common in
apparently similar areas surveyed by Peet and Christensen (1980) include Gallium
obtusum, Carex rosea, C. blanda, C. complanata, Solidago caesia, and Festuca
obtusa.
Many Piedmont floodplains, particularly those near farms and urban areas, have
been overrun by exotic species (Costing 1942). In many places, Japanese
honeysuckle (Lonicerajaponica), Japanese privet (LJgustrum sinense), and
Microstegium virmineum (a grass) are preventing the re-establishment of native shrub
species and ground cover after disturbance.
Soils cover the full range of texture from coarse particle sizes (sand and gravel)
to dense clays. County Soil Surveys do not always adequately characterize soil in
wetlands at a scale that is useful for establishing correspondence with vegetation.
However, when a study area has been thoroughly mapped, pertinent information can
be extracted from the survey such as texture, structure, permeability, drainage class,
degree of wetness, reduced matrix, organic matter content, etc. in addition, soii
classification is a excellent means of making soil interpretations, and ultimately,
establishing reference wetlands. For instance, the formative element aqu indicates the
degree of wetness depending on where it occurs in the taxonomic name. For example,
12
-------
Wehadkee, a common soil series in southern Piedmont floodplains, is classified as
fine-loamy, mixed, nonacid, thermic Fluventic haplaqi/epts. Since aqu occurs at the
suborder level, the soil has aquic conditions within 40 cm (16 in.) of the soil surface. In
contrast, Cartecay series is classified as a coarse-loamy, mixed, non-acid, thermic
Aquic Udifluvents. Aqu occurs at the subgroup level, therefore the soil has aquic
conditions within 75 cm (30 in.). Several functions can be inferred from the above
interpretation including the capacity for subsurface storage of water, degree of
moderation of groundwater flow, nutrient cycling (e.g., denitrification), and capacity to
sequester and sometimes transform toxic elements and compounds.
In general, Piedmont floodplain soils are characterized by poorly developed
horizons resulting from recent vertical accretion during episodal flood events.
Predominant taxonomic orders are Entisols and Inceptisols. Clay minerals are
dominanted by mica, kaolinite, quartz, sometimes hydroxy-interlayed vermiculite, and
various iron minerals such as hematite and goethite.
In the southern Piedmont of Georgia, three floodplain soil series occur in
association with each other depending on landscape position (listed in order of
increased drainage class): Wehadkee, Chewacla, and Congaree.
Wehadkee - Wehadkee Series are poorly drained alluvial soils subject to frequent
flooding and prolonged saturation near the surface. The prolonged saturation is
evidenced by the highly gleyed matrix and fain masses in the surface horizon.
Common micaceous material in both the surface and subsurface horizons is evidence
of alluvial deposition.
Chewacla - Chewacla Series are deep, somewhat poorly drained soils on first bottoms
formed from recent alluvium. The water table is high. However, this series is not
flooded as frequently as Wehadkee as evidenced by the brighter matrix (higher
chroma).
Congaree - Congaree series are moderately well drained or well drained and formed
from alluvium. Relative to Wehadkee and Chewacla Series, Congaree has a lower
water table, is less frequently flooded, and, its matrix has a higher chroma than the
previous two.
Other common soils found in floodplains are:
Tawcaw - Tawcaw soils frequently contain small Chastain inclusions and may be found
between the natural levees and uplands. They are composed of silty clays and have
less fine sands than Congaree soils. Tawcaw soils are poorly drained but are very
productive for bottomland hardwood tree species.
13
-------
Chastain - Chastain soils are subjected to prolonged flooding throughout the year.
These poorly drained, silty clay loams are limited to depressions and sloughs toward
the floodplain interior.
2.3 Gradients of Natural Variation and Disturbance
Recommendations for classifying reference wetlands appear in Section 3.1
below. Within these regional subclasses, there are many natural sources of variation
and disturbances that may influence their functioning. These can be categorized into
non-disturbance variation, natural disturbance variation, and human-induced
disturbance variation. It is important to be able to separate the last of these in the
identification of reference wetlands. If this cannot be accomplished, the exercise
cannot meet the goals for which reference wetlands are developea.
Non-disturbance variation includes landscape setting; flooding frequency, water
table position, and groundwater flow; stream gradient, order, and discharge; soil
texture, fertility, and permeability; floodplain geomorphic complexity (ridge and swale
topography, etc.), microtopographic relief, ratio of floodplain to channel width, and
floodplain slope; plant community composition and physiognomy; and climate. Some of
these parameters, such as climate and plant species composition, will be useful in
identifying the geographic area over which reference domain (see definitions in section
3.1) is relevant. Such variation is then isolated by subclassification whereby reference
wetlands of one subclass separated from those of another. This brings up the issue of
how much non-disturbance variation can be tolerated in a group of reference wetlands.
There is no appropriate answer to this because it depends, in part, on the purpose for
which reference wetlands are to be applied. A list of purposes was provided in the
introduction (section 1.0) ranging from use in impact assessment to detect functional
change to broad scale application to whole basin studies.
Natural disturbance variation includes the effects of tropical storms and tornado
strikes, drought related changes in animal and plant population abundance, and flood
related damage by extreme rates of erosion and deposition as well as "damage" to
vegetation. In contrast to the spatial issue of isolating variation due to geographic
variation in non-disturbance variation, natural disturbance should be subsumed in the
characteristics of reference wetlands themselves. In other words, the natural
disturbance variation must be accommodated in the same sense that non-disturbance
variation becomes a property of defining reference. If this were not done, it would imply
that reference wetlands are to be used to detect natural disturbances rather than man-
initiated ones. To elaborate, a relevant example of this is illustrated by the natural
variation in vegetation in the depressional wetlands of the northern prairie region (van
der Valk and Davis 1978, Kantrud et al. 1989). Multi-year periods of drought in the
northern prairie, while not necessarily a "disturbance" for that climate, is analogous to
the time-varying natural disturbances listed above. During these droughts, upland
14
-------
vegetation colonizes the deepest portions of some of the seasonal wetland
depressions, giving the superficial appearance that they might not be good candidates
for reference wetlands. This can even lead to extreme opinions that such sites are not
even wetlands during years when they are dry! In order to accommodate drought
conditions as part of the natural variation of reference wetlands, the interannual
climatic variation must be subsumed in the conditions that depict the wetland class of
depressional wetlands of the northern prairies.
Variation caused by human-induced alterations includes stream channelization
to increase capacity and conveyance; stream impoundment, water withdrawals and
additions (e.g., wastewater discharge), and other flow alterations; levee construction
and fills for roads and other structures; flooding by reservoirs; removal of vegetation
and animal populations; and invasion or introduction of exotic plant and animal species.
It is alterations such as these that cause wetlands to function differently from their
natural conditions, and, obviously, are the kinds of alterations to which reference
wetlands must be calibrated in order to meet goals for which they are developed. This
is why it is important to include such altered sites in the sampling of reference wetlands
as discussed in section 3.1.
2.4 Fish and Wildlife Habitat
Riverine wetlands in the Piedmont play an important role in supporting
vertebrate and invertebrate communities. The level of support is highly site specific
and related strongly to adjacent land use. In urban situations, riparian corridors may be
the only areas with enough food and cover to support populations, especially birds
dependent on trees for nesting, cover, and feeding. Where impervious surfaces in the
watershed are extensive, stream channels may be very unstable, and the vegetation
restricted to saplings, shrubs, and graminoids. Accordingly, animal populations will be
altered, but may represent the only remaining quasi-natural assemblage in highly
modified landscapes.
In rural landscapes, the floodplain itself and surrounding upland area vary
among mixtures of forest, pasture, and annual row crops. Along headwater streams, it
is common for farm fields to be ditched, with small drainages feeding directly into
intermittent stream channels. Such sites are potentially important control points for
nonpoint sources of sediment and nutrients. A fairly typical situation is row crops on
the terrace, pasture on the active floodplain, and a corridor of forest between the edge
of the pasture and the stream channel (Lowrance et al. 1995).
Piedmont riverine wetlands do not support as much habitat for aquatic
organisms as many Coastal Plain streams (personal observations of the authors). This
may be due in part to lower floodplain gradients and more extensive floodplains in the
Coastal Plain which are conducive to longer term flooding and water detention.
15
-------
However, headwater streams and their associated floodplains are critical in production
and transport of coarse participate organic matter important in food web support
downstream. In addition, Piedmont floodplains of third and higher order streams often
have backswamps where high water tables are maintained by groundwater discharge
and where both rainfall and overbank flooding causes water to be ponded. In ponded
areas, surface water and high soil moisture creates potential habitat for amphibians
and reptiles.
Although fish seldom directly use the floodplain surface in a manner equivalent
to coastal plain streams, the stream channel environment is nevertheless sensitive to
the status of the riverine wetland. For low order perennial streams, not only does
shade from forest canopies control stream temperature and light penetration, but tree
roots and forest structure in general are important to stream bank and floodplain
stability. In both headwater and larger streams, turbidity generated by the highly
erodible upland soils, and the entrainment of sediment from cutbanks and channels, is
a persistent stressor on aquatic organisms. Where stream channels have incised
sufficiently to reestablish bedrock control, they may still receive high suspended
sediment loads during stormflow from extensive watershed development activities
and/or re-entrainment of channel sediments, but stream channels offer complex riffle
and pool habits conducive to supporting a variety offish species..
In many areas the mid-20th century trend of sediment delivery and transport
trend is reversing. Extensive urban and suburban development is increasing the rate of
upland erosion. This increases sediment delivery to channels which results in
increased suspended sediment and bed loads. Where sediment is delivered directly to
the wetland surface, excessive rates of sedimentation occur, thus increasing elevation
of the surface.
16
-------
3.0 APPROACHES FOR IDENTIFYING REFERENCE WETLAND SITES
Before describing the terminology used for reference wetlands, and describing
the probable subclasses that will be necessary for southern Piedmont wetlands, we
should explain some of the differences between the reference wetland approach and
the use of reference condition for streams. There are differences in both what is
measured as the metrics for reference condition and how reference sites are chosen for
measurement in the field.
Much of the stream work is oriented toward biotic indicators of stream condition
relative to some standard. The standards would appear superficially similar to
reference wetlands described thus far. There are some important differences,
however. Aggregate indices, such as the index of biotic integrity (IBI) for fish (Karr
1991) and the invertebrate community index (ICI) for macroinvertebrates (DeShon
1995) in streams, have been used as standards for ecosystem condition. Such indices
rely on the assumption that the relative abundance of several key species populations
is reflective of an ecosystem that is relatively intact and functioning at an optimal level.
Consequently, the presence and relative abundance of key species populations
become the performance standard rather than the ecosystem itself, and, in a sense, is
used as a surrogate for other ecosystem characteristics that might be measured,
especially the physical and chemical properties. The rationale is that the biotic
community subsumes the physical and chemical characteristics within the IBI or ICI
index. Not all efforts are oriented toward biotic indices, however. Habitat indices that
represent multiple hydrologic and geomorphic variables of streams are being
developed as well (Rankin 1995). In fact, most if not all methods for biotic indices
require some collection of habitat data that can be used in characterizing reference
conditions.
Parameters that may be used as variables for establishing reference standards
for wetlands range from biotic to geomorphic (Table 1). While not all of those listed in
Table 1 would be used for a given set of reference standards of a particular wetland
class, there is a heavier reliance on physical features, especially hydrologic features,
than on a biotic community consisting of animals. In fact, many animals are
inappropriate for use in efficiently characterizing wetlands because of their short-term
use of sites (e.g., fish during spring floods or waterfowl during migration) or because of
difficulties in sighting during surveys (e.g., nocturnal animals and those that are difficult
to survey because of heavy cover). While each of these problems could be overcome
with well timed and intensive sampling strategies, the costs soon outweigh the return
on investment. For vegetated wetlands, especially forested ones, plant community
species composition and physiognomy provide both biotic and physical information.
Examples of variables and how to measure them are given below, and reported
elsewhere (Brinson et al. 1996, Brinson and Rheinhardt 1996).
17
-------
Table 1. Parameters that may be useful in characterizing reference wetlands. Not
all parameters are practical to directly measure for functional assessment. The
list includes variables, indicators, and processes. This list was developed by
EPA-ESD Region IV (Advance Identification WET template - B. Pruitt,
unpublished) to meet the objectives of the CWA. Additional functions and
attributes of riverine wetlands can be found in Brinson et al. (1996).
CLASSIFICATION:
Stream classification (Rosgen 1994)
Wetland and deepwater habitat classification (Cowardin et al. 1979)
Hydrogeomorphic classification (Brinson 1993a)
Soil classification (SCS County Soil Surveys)
STREAM/VALLEY GEOMORPHOLOGY:
watershed size/shape
stream order
network pattern
drainage density
meander wavelength
sinuosity
channel gradient
stream width/depth
valley width
pediment aggradation/degradation
landscape position
watershed position
HYDROLOGY:
stream gage data
FEMA/FIRM maps
surface and groundwater monitoring data
floodplain drainage patterns
roughness
water source(s)
inlet/outlet morphology
macro- and microtopography
vegetation signature of hydroperiod
scoured/deposition sequences (e.g., buried root collars)
18
-------
Table 1. Continued.
WATER QUALITY:
conventional parameters (temperature, pH, conductivity, dissolved oxygen)
color
tannins/lignins
suspended solids
turbidity
nitrogen series
organic carbon
analysis of metals and organics
SOIL AND SEDIMENT FEATURES:
particle size distribution
structure
depth to abrupt textural change
consistency
clay mineralogy
organic matter content
humus stratum
Fe, Al and Ca complexes
drainage class/moisture regime
permeability/hydraulic conductivity
temperature regime
redoximorphic features (accumulations, concretions, masses, pore linings, iron
depletions, odor)
redox potential
PLANT COMMUNITY STRUCTURE AND HABIT A T FEA TURES:
biomass/net primary productivity
density/basal area
submerged aquatic vegetation
plants with aerenchymous tissue
buttressing/shallow root system
multi-trunking/stooling
pneumatophores/adventitious roots
hypertrophied lenticels
vertical strata
taxa
mast trees
snags/cavity trees
decomposing stages
19
-------
Table 1, Concluded.
PLANT COMMUNITY STRUCTURE AND HABITAT FEATURES, CONTINUED:
contiguous with other ecosystems
vegetation/hydrogeomorphic class interspersion (banded, mosaic, etc.)
fishery survey
macroinvertebrate survey
mammal survey or "sign"
habitat suitability
IMPACTS:
channelization/diking
drained/filled
nutrient enrichment
siltation
NPSP sources
point sources
agricultural/silvicultural practices
mining
introduced species
impoundments
20
-------
Both similarities and differences exist in the choice of reference sites. The
general approach for streams is to sample sites in relatively unaltered watersheds
where the biotic indices can be calibrated to a maximum or optimal level (Hughes et al.
1986). There are several variations of this, including upstream-downstream reference
conditions, near field-far field reference condition, paired watershed approaches, and
ecoregional reference conditions (USEPA 1990). Historical, paleoecological, and
experimental laboratory data may also be used (Hughes 1995). The objective is to
develop a series of metrics for fish or macroinvertebrate communities that are
representative of the least disturbed condition for an ecoregion. Metrics may consist of
the general categories of community structure, taxonomic composition, individual
condition, and biological processes (Barbour et al. 1980).
As yet, there are no definitive methods for establishing reference conditions for
wetlands. For a published example for wet flats (Brinson and Rheinhardt 1996), an
equal or greater amount of emphasis was placed in sampling relatively degraded or
altered sites as the ones that were initially believed to be reference standard sites (see
definitions below). This was necessary for flats in part to identify wetland subclasses
(i.e., wet pine flats maintained by fire in contrast to wet hardwood flats where fire was
excluded) at the same time that reference standards were being identified. In part, the
sampling of altered sites provides the scaling necessary for functional assessment
using the hydrogeomorphic approach (Smith et al. 1996, Brinson 1995,1996). Without
scales for measuring departure from reference standards, it would be difficult to
establish thresholds of tolerance for significant degradation due to project impacts. In
addition, the regulatory environment dealing with wetlands places heavy emphasis on
the mitigation of impacts through restoration when a project is permitted. In order to
know whether a mitigation effort of wetland restoration is achieving success, it is
necessary to have goals established in restoration projects. The logic is that greater
success of restoration projects will be achieved by projects that use reference wetlands
as templates for design and targets for gauging success. Thus, the use of a different
type of reference standard is needed for wetlands than for streams in order to have
utility in quantifying functional change, in dealing with environmental impacts in a 404
enforcement mode, and in measuring progress toward project targets in restoration.
3.1 The Terminology of Reference
The process of identifying reference wetland sites requires that the
terminology be defined to clarify the relationship between the sites themselves and the
criteria that make sites appropriate for various uses as reference wetlands. Table 2 is
a list of the terms with brief definitions. Each of the terms are described below in the
context of southern Piedmont riverine wetland subclasses, although they are
applicable also to reference wetlands in general.
21
-------
Table 2. Categories and nomenclature for reference.
Reference domain - All wetlands within a defined geographic region that belong to a single
hydrogeomorphic subclass.
Reference wetlands - Wetland sites within the reference domain that encompass the known
variation of the subclass. They are used to establish the range of functioning within the
subclass. Reference wetlands may include: (1) former wetland sites for which restoration to
wetland is possible and (2) characteristics.of sites derived from historic records or published
data.
Reference standard sites - The sites within a reference wetland data set from which
reference standards are developed. Among all reference wetlands, reference standard sites
are judged by an interdisciplinary team to have the highest level of functioning.
Reference standards - Conditions exhibited by a group of reference wetlands that
correspond to the highest level of functioning (highest sustainable capacity) across the suite
of functions of the subclass. By definition, reference standard functions receive an index
score of 1.0.
Site potential - The highest level of functioning possible given local constraints of
disturbance history, land use, or other factors. Site potential may be equal to or less than
levels of functioning established by reference standards.
Project target - The level of functioning identified or negotiated for a restoration or creation
project. The project target must be based on reference standards and/or site potential and
be consistent with restoration or creation goals. Project targets are used to evaluate whether
a project is developing toward reference standards and/or site potential.
Project standards - Performance criteria and/or specifications used to guide the restoration
or creation activities toward the project target. Project standards should include and specify
reasonable contingency measures if the project target is not being achieved.
The reference domain for Piedmont riverine wetlands would include all of those
wetlands within a single hydrogeomorphic (HGM) subclass (Smith et al. 1996, Brinson
1995. 1996). As the subclasses have not been definitively identified, the geographic
range over which they occur cannot be finalized at this time, (n general, however, the
southern Piedmont of the Carolines and Georgia (Figure 2) would represent an
appropriate physiographic unit to define an initial reference domain for Piedmont
riverine wetlands because of similarities in topography, climate, and species
composition. We anticipate that smaller reference domains will be more practical. For
example. Wharton (1978) recognized four physiographic provinces for the Georgia
Piedmont alone. If differences in species composition are shown to be consistent and
substantial, then reference domain could be redefined to accommodate the source of
22
-------
PIEDMONT PHYSIOGRAPHIC PROVINCE
EASTERN UNITED STATES
—i ic»ir
H = PIEDMONT
Figure 2. Location of the Piedmont in the eastern United States. The Carolinas and
Georgia are major areas in this physiographic region.
23
-------
variation. Use of species composition of the plant community, and not the distribution
of a single species, can be a powerful tool for detecting edaphic and climatic variations.
Reference wetlands are those sites within a subclass of Piedmont riverine
wetlands that are used for characterizing the reference domain from qualitative
descriptions and quantitative data. Several examples of potential reference wetlands in
the southern Piedmont, with qualitative descriptions only, are listed in the Appendix.
Reference wetlands include not only the highest functioning sites, but also ones that
have been altered through channelization, logging, filling, and other activities. The
altered ones are useful for confirming, by differences, which wetland sites are the
highest functioning sites.
Reference standard sites are those sites in a subset of reference wetlands that
function at the highest, sustainable level across a suite of functions. They are most
likely to be found in relatively unaltered landscapes (national forests, state parks, etc.)
where fairly mature forest stands have developed. However, few sites have completely
escaped alterations because of modified flows by impoundments, historic sediment
deposition, alteration of channel morphology, or current human activity in uplands. The
mass erosion and valley filling during the past 150 years, followed by stream incision in
the past 25 years or less, poses special problems for identifying reference standards.
This is not unlike determining "normal circumstances" in wetland delineation.
Reference standards are conditions exhibited by reference standard sites.
Determining the levels of functioning for reference standards, along with site potential,
will be a major challenge for determining reference standards in the Piedmont.
Site potential represents the highest level of functioning that can be sustained
for a particular location in a riverine wetland. For Piedmont riverine wetlands, site
potential must be adjusted to differing levels of functioning within the subclasses that
we identified: overbank flow dominated, riparian source dominated, and beaver
dominated riverine wetlands (see Section 3.2). For example, sites that are not
dominated by overbank flow (because of historical channel downcutting and
enlargement) would rank low in functions that rely on overbank flow relative to those
sites that are dominated by overbank flow. However, sites not dominated by overbank
flow should not be compared with those dominated by overbank flow because they are
in different subclasses. For this reason it is critical that sites are properly classified
according to current hydrologic conditions before undergoing functional assessment,
restoration, or other steps in the mitigation process.
Project target is the level of functioning identified for a restoration project. It is
not particularly relevant for the purpose cf this report. However, site potential is an
integral component of project target in the sense that acceptable targets must have the
potential to develop over time to realize the capacity of the site (e.g., site potential).
24
-------
Locations of the basins and study sites for reference wetlands should be
modified to meet goals and directives of the EPA, state and local governments.
However, when possible, the study area should be bound by natural landforms (e.g.,
geology, drainage basins, ecoregions) and not political boundaries. Fifteen major
drainage basins have been identified in the Piedmont physiographic province (Table 3).
3.2 Recommended Subclasses for Piedmont Riverine Wetlands
A major reason for classification is to separate variation from natural sources
that relate to functioning from variation due to disturbance, particularly disturbance
caused by human activity. The following three subclasses are proposed as defining the
range of natural variation within the reference domain for Piedmont riverine wetlands.
Below each of these subclasses, additional stratification could be invoked from their
position in the drainage basin, such as headwater tributaries, middle reaches, and
large river/floodplain systems. Further variation derives from the geomorphic structure
of floodplains due to underlying lithology. This separates regions with steep
topography, such as the Uwharrie National Forest, North Carolina, from areas with
more muted topographic relief, such as the Triassic basin in North Carolina. Within this
range of lithic control, land use has had additional influences on riverine wetlands.
3.2.1 Overbank flow-dominated
These are floodplains where stream channel incision has not proceeded to the
extent that overbank flow is rare or completely absent at frequent recurrence intervals
of 2 years or less. The presence or absence of riparian source does not affect the
criterion because the subclass is dominated by overbank flow. Large alluvial rivers are
most likely to occur in this subclass because they continue to have large sediment
supplies in contrast to floodplains of headwater streams or those below dams. Riverine
wetlands above knick point (base level) controls lack the capacity to incise and may
have extremely low channel capacity, thus maximizing the interchange of water
between stream channel and wetland surface.
Evidence for this subclass includes fresh sand deposits on levees or sand
splays behind levees. Fine sediment layers on leaves in low velocity areas of the
wetland are also indicative of this subclass. In high velocity areas, leaves may be
swept away before sediment deposition has an opportunity to prevent leaves from
floating.
3.2.2 Riparian source-dominated
If a riverine wetland has virtually no overbank flow or infrequent recurrence
intervals, it is dominated by riparian sources of water. These sources are a
combination of overland flow from uplands during storm periods, subsurface discharge
25
-------
Table 3. Major watersheds of the Piedmont. In many cases, the watersheds
actually extend into the Blue Ridge Mountains or Appalachian physiographic
provinces.: Watersheds designated as "CP & P" indicate the river has segments
within both the Coastal Plain and Piedmont.
(CP = Coastal Plain, P = Piedmont)
State: Alabama
Drainage Basin: Tombigbee (CP)
Watershed: Alabama (CP)
Subwatershed: Coosa (CP & P)
Subwatershed: Tallapoosa (CP & P)
Little Tallapoosa (P)
Watershed: Black Warrior (P)
State: Alabama & Georgia
Drainage Basin: Apalachicola (CP)
Watershed: Chattahoochee (CP & P)
GA Watershed: Flint (CP & P)
GA Watershed: Little (P)
State: Georgia
Drainage Basin: Altamaha (CP)
Watershed: Ocmulgee (CP & P)
Subwatershed: Little Ocmulgee (CP)
Subwatershed: Towaliga (P)
Subwatershed: Alcovy (P)
Subwatershed: South (P)
Subwatershed: Yellow (P)
Watershed: Oconee (CP & P)
Subwatershed: Little (P)
Subwatershed: Apalachee (P)
Subwatershed: North Oconee (P)
Subwatershed: Middle Oconee (P)
Watershed: Ohoopee (CP)
Drainage Basin: Ogeechee (CP & P)
Watershed: Little Ogeechee (P)
State: Georgia and South Carolina
Drainage Basin: Savannah (CP & P)
GA Watershed. Little (P)
GA Watershed: Broad (P)
SC Watershed: Keowee (P)
26
-------
Table 3, Concluded.
State: South Carolina
Drainage Basin: Edisto (CP)
Watershed: South Fork Edisto (CP & P)
Watershed: North Fork Edisto (CP & P)
State: South and North Carolina
Drainage Basin: Cooper (CP)
Watershed: Congaree (CP)
Subwatershed: Broad (P)
Tyger(P)
Pacolet (P)
Subwatershed: Saluda (P)
Subwatershed: Catawba (P)
State: South and North Carolina (Cont)
Drainage Basin: Great Pee Dee (CP & P)
NC Watershed: Rocky (P)
NC Watershed: Yadkin (P)
State: North Carolina
Drainage Basin: Cape Fear (CP & P)
Watershed: Deep (P)
Watershed: Haw(P)
Subwatershed: New Hope (P)
Watershed: South (CP & P)
Drainage Basin: Neuse (CP & P)
Drainage Basin: Tar (CP & P)
State: North Carolina & Virginia
Drainage Basin. Roanoke (CP & P)
Watershed: Banister (P)
Watershed: Staunton (P)
Drainage Basin: Meherrin (CP & P)
State: Virginia
Drainage Basin: Nottoway (P)
Drainage Basin: Appomattox (P)
27
-------
from the aquifer or lateral subsurface (saturated and/or unsaturated) flows from the
upland to the footslope or within the alluvium of the floodplain, and input from small
truncated tributaries. It is apparent from Figure 3 that riparian source dominated
wetlands are more prevalent in headwaters than in streams of higher order, all other
things remaining equal. This tendency is made even more extreme in situations of
stream channel incision which reduces the opportunities for frequent overbank flow.
Burke and Nutter (1995) and Burke (1996) identified floodplains in the Oconee River
Basin of the middle Georgia Piedmont where return intervals for bankfull discharge
were greater than 500 years (2.33 years is expected) because channel entrenchment
and/or widening has increased channel capacity in recent times.
Evidence for the riparian source-dominated subclass is the presence of seepage
areas at the transition from the base of the upland slope to the floodplain. Good
evidence for subsurface discharge of anoxic water to the surface is oxidized iron
deposits or organic sheen in ponded areas or'soil saturation with no evidence of
overbank flow. Rills and even channels on the slope leading to the wetlands are
evidence of contributions to overland flow. Deeply incised active gullies typify this
condition. Channels associated with these wetlands have bankfull recurrence of
greater than the annual event, and often much greater than that. The fact that the
overbank flow source is infrequent does not mean that it should be excluded as a
variable influencing functions. Infrequent events are often those that effect the greatest
changes in terms of sediment deposition, changes in channel morphology, and
redistribution of detritus on the floodplain surface.
3.2.3 Beaver dam-dominated
If water levels in a wetland are influenced by a beaver dam, it is classified as
beaver dam-dominated. This subclass is justified because hydrologic, biogeochemical,
and habitat features differ dramatically from the other two classes of wetlands. Beaver
prefer a seasonably stable water level supplied by a permanent stream (Slough and
Sadleir 1977). However, larger rivers are generally unsuitable for the species (Murray
1961, Slough and Sadleir 1977). Generally, the authors have not observed beaver or
the result of their activities in Piedmont streams greater in size than third order
(Strahler 1952). Fourth order streams roughly correspond to a watershed size of 30
km2 or larger (Burke and Nutter 1995).
Indicators include abnormally high water on the floodplain during normally low
flows, dead and dying trees suggesting changes in hydrology, and evidence of beaver
activity including bark removal from trees, dams and lodges, and related signs. Only
where dams have been most recently built will non-hydrophytic vegetation persist in
abnormally wet conditions created by beaver dams. Denney (1952) reported that the
food of North American beaver, in order of preference, were aspen (Populus
tremuloides), willow (Sa//x spp.), balsam-poplar (P. balsamifera), and alder (Alnus
28
-------
^i^'^jp^
OVERBANK TRANSPORT
DEEP
GROUND-WATER^
DISCHARGE
RIPARIAN
TRANSPORT
Figure 3. Relative importance of overbank flow to riparian source as a function of
stream size or position in the drainage. From Brinson (1993b).
29
-------
spp.). In the southern Piedmont, however, the authors have observed that sweetgum
(Uquidambar styraciflua) is a preferred food of the species as indicated by cuttings.
Key diagnostic indicators of beaver habitat include percent tree canopy closure,
percent trees sapling size category, percent shrub crown cover, average height of
shrub canopy, dominance of woody vegetation by food preferences, and degree of
edge irregularity. Beaver dams greatly influence water chemistry in the Adirondack
Mountains (Cirmo and Driscoll 1993) as they do elsewhere (Johnston and Naiman
1987).
3.3 Strategies for Choosing Reference Wetlands
The strategy for choosing reference wetlands depends in part on the purpose to
which they will be used (section 1.0). If reference wetlands are used to characterize
conditions over a large region, such as for an ADID project, a stratified random
approach may be most appropriate. Alternatively, reference wetlands for functional
assessment in evaluating Section 404 permits in a particular Corps district may be
approached much more selectively over a smaller geographic area. From this smaller
group, the determination of reference domain becomes an iterative process of site
selection and determining whether new sites fall within variation that is acceptable for
contributing toward refence standards of the subclass. Other purposes such as
mitigation banks and reserves, and identification of potential restoration sites, may
suggest alternative approaches for choosing sites. A two-phased sampling design and
approach is outlined in Figures 4 and 5.
3.3.1 Stratified random approaches
We suggest a hierarchical approach for identifying reference sites in each of the
three riverine wetland subclasses: a regional scale, a watershed scale, and a stream
reach scale. The regional scale is already implicit in the scope of this report, i.e., the
Southern Piedmont, a well recognized physiographic province (Wharton 1978).
However, Wharton recognized four distinct subregions within the Georgia Piedmont
alone, although it is not clear how much these regions correspond to Variations in
riverine wetland conditions.
The second scale at which to identify reference sites is to subclassify by
watersheds. Drainage basins and major watersheds in the Piedmont as tabulated in
Table 3. Watersheds should be identified that represent a wide range of natural and
anthropogenic modifications (e.g., relatively unaltered wetlands versus those affected
by point and non-point pollution, impoundments, channelization, urbanization, and
silviculture).
30
-------
PROJECT GOAL
LU
DEFINE PERSONNEL &
BUDGETARY CONSTRAINTS
I
IDENTIFY SPECIFIC
OBJECTIVES
PEER REVIEW
DEFINE
HYPOTHESES
IDENTIFY STUDY AREA(S)
*
UTERATURE REVIEW
IDENTIFY REFERENCE
DOMAIN
i
DEVELOP/TEST CLASSIFICATION
STRATIFY WETLANDS
BY ECOREGION
STRATIFY WETLANDS
BY WATERSHED
STRATIFY WETLANDS
BY CLASS
REFERENCE SET
DE
Figure 4. Sampling design and approach for identification and classification.
31
-------
RANDOMLY SELECT
REFERENCE SET
WITHIN EACH CLASS (STRATA)
DEFINE SAMPLING/MONITORING FRAMEWORK
DOES PLAN
MEET DQO's?
OQISTICAL SUPPORT
AVAILABLE?
CONDUCT PRELIMINARY OR
PILOT SURVEY
ANALYZE 4 EVALUATE DATA
(PEER REVIEW
PILOT
SURVEY MET
SAMPLING OBJECTIVES
INITIATE INTENSIVE SURVEY PHASE
DATA REDUCTION. ANALYSES &
INTERPRETATION
FINAL REPORT &
RECOMMENDATIONS
Figure 5. Sampling design and approach for project.
32
-------
At the third scale, stream reach can be chosen to determine the location of
reference wetlands. Reaches could be further stratified by drainage area or stream
order. There are several choices at this scale on how to allocate reference wetland
sites. One is by wetland area, such that stream orders with the largest surface area are
allocated the greatest number of reference sites. For watersheds as a whole, wetland
area remains relatively constant among stream orders in contrast to rapid increases in
stream length with increasing stream order (Brinson 1993b). To allocate sites by
stream length would require that the greatest number of sites be located on headwater
streams. Also, the size of the watershed drained could be used to further categorize
subclasses, or, alternatively, if there are large disparities in discharges from streams of
the same stream order, discharge categories could be used to subclassify or stratify
sites. Finally, it should be recognized that these are merely logical recommendations
that are not based on extensive practical experience.
3.3.2 Comparison of strategies for reference wetlands and reference streams
Although some of the objectives for choosing reference sites for stream
biomonitoring may differ from those for wetland functional assessment, the approaches
have considerable overlap. For example, the Alabama Department of Environmental
Management (ADEM and MDEQ 1996), as part of the Mississippi and Alabama
ecoregions and southeastern plains subregions effort (Omemik and Griffith 1991),
provides details that relate to map work, field reconnaissance, equipment lists, and
data forms. A number of suggestions provided at the level of map work would be useful
for identifying reference standard sites. Topographic maps at 1:100,000 scale can be
used. Relevant features that can be identified and evaluated on these maps are: (1)
that the candidate watershed is restricted to a single ecoregion, (2) that the area is
forested, or, if not mostly forested, the watershed contains a National Forest or other
land cover type that is typical for the ecoregion, (3) that the site has geologic structure
typical of the region, (4) that stream channels are natural and without impoundments,
(5) that a municipality or active mine is not in the area, (6) that herbicide usage is not
present because of its application to pipelines or powerlines at stream crossings, (7)
that few roads are in the watershed, and (8) that the site is accessible by road (ADEM
and MDEQ 1996). National criteria (USEPA1990) for biomonitoring of forest streams
call for such site characteristics as (1) extensive, old, riparian vegetation, (2) relatively
high heterogeneity of the channel bottom, (3) abundant large woody debris, coarse
bottom substrate, or extensive aquatic or overhanging vegetation, (4) relatively clear
waters with natural color and odor, (5) abundant diatom, insect, and fish assemblages,
and (6) the presence of piscivorous birds and mammals. Because the objective in
cases of biomonitoring is to identify minimally altered streams, reference streams are
more in synchrony with reference standard sites, defined above, than they are to
reference wetlands. Of course, site visits and data collection in both cases are needed
in order to characterize both reference wetlands and reference streams. The map
33
-------
identification exercise is simply a screening tool that allows efficient reconnaissance
over large geographic areas.
In contrast to reference streams, we think it is equally important to identify site
potential of reference wetlands within watersheds that are altered. Because the
mitigation policies involving wetlands place emphasis on restoration, site potential
becomes a very critical consideration to determine. One must recognize that altered
watersheds cannot be expected to support riverine wetlands that function equally to
those in minimally altered watersheds. In fact, riverine wetlands, because they are
more closely connected hydrologically to the remainder of the landscape than other
wetland classes, are likely to be the most affected by land use activities distant from the
wetland being assessed.
The purpose identifying site potential is to recognize that the end point of
restoration of wetlands cannot be at the level of reference standards. If reference
standards for wetlands are to be acheived and maintained in highly altered watersheds
where they are not self sustaining, active maintenance of these sites must be an
ongoing process and thus a component of mitigation plans. Judging from the difficulty
in tracking mitigation projects (Kusler and Kentula 1990), it is unlikely that such
continuous monitoring and vigilant stewardship can be expected, especially in areas
that are undergoing rapid urbanization.
3.3.3 Best professional judgement and local knowledge
In the course of this project, we contacted the Natural Heritage Program in North
Carolina for recommendations of sites that have characteristic vegetation and other
features. We were directed to both riverine and depression/slope wetlands, the latter
occurring in very low abundance. The rarity of slope and depressional sites and their
vulnerability to alteration suggest that they should be considered "waters of special
concern" and thus be preempted from evaluations and assessments for most purposes
described in Section 1.0. One should be aware that the mission of groups like the
Natural Heritage Program is not to identify representative wetlands, but to select sites
based on criteria such as exemplary plant communities that lack disturbance, rare and
endangered species, special geomorphic features, and other similar "unique"
characteristics.
Most of the sites that we visited were identified by the Natural Heritage Program.
These are described in the Appendix. The Heritage Program's classifications have a
strong component of geomorphology, and include both riverine and depressional
classes. Two sites were designated as Piedmont Alluvial Forest (Eno River and New
Hope Creek) in the Durham-Onapei Hiii area, two as Piedmont Levee Forest (naw
River and Uwharrie River), one Piedmont Swamp and Bottomland Forest (Brown
34
-------
Creek, a tributary of the Pee Dee River), and two "Upland" Depression Swamp Forests
(Duke Forest and Uwharrie National Forest).
We also considered including sites near streams used as reference stations for
biomonitoring by the NC Division of Environmental Management. We anticipated that
the least polluted of these sites, as identified through biomonitoring, would conform to
conditions associated with reference standards for adjacent floodplains. The lack of
information on wetlands associated with these sites does not provide a basis for
choosing one over another. The spatial focus on characteristics of stream channels
alone is apparently too narrow to reveal potential wetland reference sites. Emphasis
on purely aquatic rather than wetland conditions appears to be the main reason that
work on reference streams does not provide the necessary information for choosing
reference wetlands.
Other knowledge from consulting, academic, and regulatory professionals can
be useful in the choice of sites for reference standards. It may be possible to quickly
find wetland sites that would establish reference standards by contacting these groups
because they are able to not only direct one to the least altered sites, but they have
information on access to private land.
3.4 Locating Reference Streams and Wetlands in the Same Place.
It would be useful to establish reference wetlands and stream monitoring sites at
the same locations. First, there are functional connections between the two due to the
.exchange of water, materials (organic matter, sediments, etc.), and organisms.
Further, both streams and wetlands are considered "waters of the United States," so
there is administrative and enforcement utility to recognizing the functional
connections. Combining the sites should provide complementary information not
available from one or the other. For example, wetlands are often devoid of surface
water thus lacking habitat for benthic invertebrates needed in streams for
bioassessment. On the other hand, wetlands contribute to stream bank stabilization,
water storage, base flow, shading, and other "stream support functions" that might be
missed in a typical stream biomonitoring program. Finally, co-locating reference
wetland sites and stream monitoring sites would emphasize that the same standards
would not be appropriate for both streams and wetlands. For example, thresholds used
for water quality standards may be markedly different within a wetland than in a flowing
stream. In cypress-tupelo swamps, filamentous algal blooms are normal wintertime
events, while dissolved oxygen concentrations may be extremely low under the shaded
conditions of a flooded forest floor in the summer. Neither of these conditions would be
as extreme in unaltered stream channels. Thus, such environmental standards
associated with biomonitoring for streams would not be appropriate for wetland portions
of waters. The development of reference standards for wetlands would reduce the
35
-------
tendency to inappropriately transfer standards from the stream channel to the
floodplain and vice versa.
Combining stream biomonitoring sites and reference wetland sites would also
facilitate and promote a more integrated approach to watershed and ecosystem
management. Bottomland hardwood forests in the Piedmont are critical to maintaining
water quality in adjacent streams. This is especially true for nonpoint sources of
pollution where water does not enter streams directly through discharge pipes, but
through streamside riparian zones. However, if these streamside zones are not
maintained at high levels of function, they may actually contribute to the nonpoint
source problem rather than providing a solution to it. As described earlier, the channel
downcutting that has been occurring in Piedmont stream channels contributes to the
high suspended sediment loads as a consequence of historical land use of the
Piedmont. Allowing these floodplain areas to deteriorate further will only exacerbate an
already serious problem.
3.5 Problems with Non-jurisdictional Wetlands.
Geiger (1994) observed that in the urban Piedmont of North Carolina, only
13.5% of the area indicated as wetland on National Wetland Inventory maps were
jurisdictional wetlands. Many of the areas chosen in the Geiger study was in a riverine
hydrogeomorphic setting and were identified as bottomland hardwoods on floodplains.
It is likely that even non-jurisdictional portions of these floodplains are functioning
ecologically as wetlands in providing hydrological, biogeochemical, and plant and
animal habitat functions. Assessment of functioning of ecological wetlands is critical for
meeting the goals of the Clean Water Act in such landscapes where wetlands are as
reduced in surface area as they are in the Piedmont, and partly an artifact of the
disequilibrium condition of channel incision. The proposal of the riparian-source
dominated riverine subclass is partly in response to the disequilibrium that has isolated
floodplains from flooding by annual flow regimes. The latest revision of the hydric soil
indicator list by the National Technical Committee for Hydric Soils (1995) needs to be
tested within representatives of each wetland subclass in order to determine if
jurisdictional wetlands are being excluded because of peculiarities of the young
alluvium.
Once jurisdictional uncertainties are removed, there is still a disparity between
riverine wetlands functioning as hydrogeomorphic units and jurisdictional wetlands
which are driven more by local water table dynamics. The riparian source-dominated
riverine subclass provides a good example of why it is necessary to consider the
floodplain as an integral unit. In such cases, the backswamp or footslope wetlands
may be separated from the stream by a wide floodplain that no longer receives annual
flow because of channel enlargement. The placement of levees on the narrow zone of
jurisdicational upland along the channel would interfere with the potential of the
36
-------
floodplain to store water during overbank flow. This would interfere with the function of
dynamic surface water storage on the jurisdictional wetland in the backswamp. As a
further complication, beaver-dominated riverine wetlands can appear on the landscape
over very short periods of time and convert uplands into wetland jurisdictional status.
37
-------
4.0 DATA COLLECTION AND ANALYSIS
It is not the objective of this report to present detailed directions on how to
collect and analyze data for riverine reference wetlands. We provide some guidelines,
however, that are applicable in forested wetlands in contrast to methods documented in
Magee et al. (1993) for emergent wetlands and open-water sites.
4.1 Field Collection
The strategy for choosing reference wetlands depends in part on the purpose to
which reference wetlands will be applied. Methods used in collecting data to
characterize sites, however, can be much more standardized. While there is some
potential to adjust methods to funding levels, the actual time, methods, and effort for
sampling sites can be relatively well defined.
Most Piedmont riverine wetlands are forested, so vegetation sampling methods
for forest vegetation are appropriate. This includes measuring diameters and densities
of trees in larger (e.g., 10m radius) plots and densities or cover of shrubs and
subcanopy species in smaller plots. Rectangular plots may be required to quantify
vegetation in narrow floodplains, but circular plots can usually be delineated and
measured more quickly than square or rectangular plots (Linsey et al. 1958, Levy and
Walker 1971). A plotless technique (e.g., Bitterlich technique) can be used to rapidly
measure basal area (by species) of canopy trees, but may not be practical in
floodplains less than about 30 m across. Small quadrats (e.g., 1 m2), belt transects,
and line-intercept methods can be used to estimate cover of herbaceous species, litter,
standing water, and other microsite parameters.
The number of plots required to characterize a site depends upon the diversity
and heterogeneity of its species and other measured attributes. More heterogeneous
or diverse habitats require more intensive sampling. A species-area curve can be used
to determine at which point additional sampling is no longer needed to characterize a
site. Usually, if an additional sampling point (or an increase in area sampled) fails to
provide more than a 5% addition in new species (or of some other attribute being
measured), then a site can be considered to have been adequately sampled.
Procedures for sampling sites should be developed so that a reference site can be
quantitatively sampled between 1.0 and 1.5 hours by experienced field workers.
The number of reference wetlands needed to characterize a given geomorphic
subclass depends upon the variability in the defined subclass. If sites are carefully
chosen to represent a variety of age classes and disturbance patterns, 20-40 sites
should suffice for most subclasses. This is based on having at least 10 sites that are
considered reference standard sites with the remainder representing some type of
alteration.
38
-------
Several of the altered sites should be those whose only alteration has been
removal of vegetation so that a series of stages of natural succession can be observed.
Data on successional conditions can be used to establish short-term target conditions
for compensatory mitigation projects. Conditions exhibited by other types of
anthropogenic alterations can be used to calibrate variable indices (see section 3.3). If
funds or time preclude obtaining 20-40 reference sites, then the subclass for which
reference is collected may be even more narrowly defined to reduce the potential range
in natural variability.
Data quality objectives (DQOs) should be established. DQO is a measure of the
magnitude of error associated with a data set (Peters 1988). DQOs are statements that
specify the quality of data needed from a particular data collection activity. They allow
flexibility in the quality control necessary to meet the goal and objectives of the project.
For example, if the DQO is simply to determine if a flood event occurred in the past
without establishing frequency or duration, diagnostic field indicators such as drift lines,
silt lines, sediment deposition and scour are adequate to meet the DQO. Not only will
DQOs provide guidance for quality control, but they serve also as instructions for
including additional sites in the reference wetlands data set.
When data are collected at larger scales, say for mesoscale analysis of the
acreage of wetland classes, it is critical to collect remote and on-site data in a
consistent and reproducible format. By using aerial photographs, maps, etc., it is
possible to stratify the wetlands according to the variables indicated in section 2.3.
NWI maps, county soil surveys, and aerial photography can be used to determine the
following: geographic and physiographic setting; landscape and watershed position
including aspect and slope; connectivity with other landforms and land uses (e.g. other
wetlands, forested uplands, pasture, open water); size and shape; classification
(vegetation, geomorphic, and hydrodynamic); surface water flow patterns including
inlet/outlet configuration; and stream geomorphology (e.g. channel, braided or sheet
flow, channel alterations, degrading or aggrading, sinuosity, pool/riffle ratio).
4.2 Data Reduction and Statistical Analysis
Data should be keyed into a spreadsheet for verification with field sheets and
analysis. At a minimum, statistical analysis should include number of samples,
minimum, maximum, arithmetic mean, geometric mean (when appropriate), standard
deviation, and variance. Data distribution should be tested for normality, skewness,
and kurtosis. The variation of each wetland subclass within each reference domain
should be determined statistically. Standard deviation and variance about the
population mean for each variable should be used to determine the precision and
accuracy of the classification and if bias has been introduced. If variance is observed
to exceed anticipated amounts in any strata, the classification scheme and/or the
39
-------
cluster analysis should be reconsidered and revised if warranted. Factors that
contributed to the variability in stratification within subclasses of wetlands should be
identified.
In the analysis of data, there are usually two goals that must be approached
simultaneously: that of identifying reference domain and that of identifying reference
standard sites. There are no fixed rules for either at this time, and both require a
considerable amount of "best professional judgement." If the classification step has
resulted in a group of wetlands that are sufficiently homogeneous to allow separation of
natural variation from impact variation, then the reference domain can be expanded
geographically with data from sites incrementally more distant from the original data
set. For example, beaver dominated wetlands of the southern Piedmont, due to
hydrologic features from damming, may be quite similar to beaver dominated wetlands
of the northern Piedmont, or even the coastal plain. In such cases the reference
domain could be expanded. On the other hand, it may be useful to keep as separate
reference domains the riparian source dominated wetlands of the northern and
southern Piedmont regions.
Identification of reference standard sites can be partly accomplished by
recognizing variables that suggest maturity or old age in stands, such as density and
species composition. This task is made easier by including wetlands that have been
altered in the analysis. Various multivariate techniques, such as CANOCO designed
for vegetation analysis, are useful for determining the degree of dissimilarity among
stands (Ter Braak 1988). Such techniques allow the graphical illustration of stands in
relationship to one another through ordination. Familiarity with the stands and the data
often allow the analyst to recognize clusters of points that suggest similar
physiognomy, species composition, soil conditions, and so forth. An analysis of such
data should also reveal whether reclassification or further subclassification would be
useful. Once sites are identified as reference standard sites, data for individual
variables can be examined to determine reasonable amounts of variation to accept for
reference standards. The relative amount of variation may differ among variables. For
stand density and basal area of live trees, for example, the amount of variation about
the mean may be quite low relative to an inherently more variable metric such as
density of snags.
4.3 Implementation and Field Testing
It is premature to make suggestions on methods for scaling up from a pilot scale
at this time. We recommend that projects start small by establishing initial reference
wetlands and use them to field test the methods for which they are adapted before an
elaborate sampling design is developed and routine field work is initiated. For an ADiD
project, ground truth data on wetland subclass and verifying site potential may be the
principal objectives. Reference wetlands developed for functional assessment
40
-------
procedures used in permits and enforcement will require testing to determine if
reference standards are realistic for the reference domain. Questions to ask are: How
sensitive is the method for detecting gains and losses in functions due to impacts
regardless of area? Are certain types of impacts more amenable to being detected
than others? What is considered significant degradation?
If reference wetlands are to be involved in establishing state standards, it is
imperative to determine whether preliminary data generated from them are able to
facilitate the establishment of standards for wetlands in the state. Sediment standards
for streams, for example, would be inappropriate for wetlands if they did not prevent
their deterioration.
41
-------
5.0 CONCLUSIONS AND RECOMMENDATIONS
Floodplains of streams of the Piedmont province are located in landscape
positions that are critical to the maintenance of stream water quality. Wetlands occur
on many but not all of these floodplains. Most states have developed water quality
standards as a method of determining environmental condition and regulating
discharges of pollutants that would degrade water quality. Wetlands are also waters of
the United States. However, no known water quality standards have been developed
for wetlands in spite of the Piedmont being one of the most rapidly developing regions
of the U.S.A. We recommend that standards be developed for wetlands in the
Piedmont of the Carolinas and Georgia.
Water quality standards for wetlands can be developed through the use of
reference wetlands. The reference wetland concept requires that wetlands first be
classified by hydrogeomorphic characteristics. For riverine wetlands, we provisionally
recommend three: overbank flow dominated, riparian source dominated, and beaver-
dam dominated. Within each of these, reference standards must be developed from
the ones that are in the least altered condition. This is done through a combination of
quantitative data collection and analysis, and application of best professional
judgement. From wetlands that are considered to be least altered, reference standards
are developed that can be used in a rapid assessment mode to judge whether a
particular wetland meets reference standards. Not only can reference wetlands be
used to develop standards for rapid assessment, but the concept can be applied to
determining short-term targets for restoration. Reference wetlands can serve a broad
range of needs including whether projects cause significant environmental degradation,
determining the status and trends of wetland conditions, serving as standards for
establishing designated uses and biocriteria, incorporating information into whole basin
planning, and serving as training sites for personnel involved in functional assessment
and restoration.
States that have biomonitoring programs for streams are encouraged to extend
their efforts toward a more comprehensive approach in monitoring, assessing, and
managing by including riverine wetlands in their network of stream reference sites.
While many similarities exist for the choice of reference sites for both streams and
wetlands, there are significant differences that depend partly on the goals of the
programs. Where program goals justify combining reference sites for streams and
wetlands, protocols should be explored for standardizing and coordinating site
selection, sampling, and methods of analysis. This approach may strengthen the
scientific basis for resource management, and if feasible, could more efficiently utilize
limited resources available for water quality rnsnagsm^nt.
42
-------
6.0 LITERATURE CITED
ADEM and MDEQ. 1995. Alabama/Mississippi Pilot Reference Site Project, 1990-
1994. Alabama Department of Environmental Management, Field Operations
Division, Montgomery, AL; Mississippi Department of Environmental Quality,
Office of Pollution Control, Pearl, MS, USA. 37 pp. + Appendices.
ADEM and MDEQ. 1996. Standard Operation Procedure, Volume II: Freshwater
Macroinvertebrate Biological Monitoring. Alabama Department of Environmental
Management, Field Operations, Montgomery, AL, USA.
Barrows, H.H., J.V. Phillips, and J.E. Brantly. 1917. Agricultural Drainage in Georgia.
Bulletin 32, Geological Survey of Georgia, Atlanta, Georgia, USA.
Barbour, M.G., J.H. Burk, and W.D. Pitts. 1980. Terrestrial Plant Ecology. The
Benjamin/Cummings Publishing Company, Menlo Park, California, USA.
Brender, E.V. 1974. Impact of past land use on the lower Piedmont forest. Journal of
Forestry. Volume 72.
Brinson, M.M. 1990. Riverine forests. Pages 87-141 in A.E. Lugo, M.M. Brinson, and
S. Brown (editors). Forested Wetlands, Elsevier Scientific Publishers,
Amsterdam, The Netherlands.
Brinson, M. M. 1993a. A Hydrogeomorphic Classification for Wetlands. U.S. Army
Corps of Engineers Waterways Experiment Station. Wetlands Research
Program Technical Report WRP-DE-4. Vicksburg, Mississippi, USA.
Brinson, M.M. 1993b. Gradients in the functioning of wetlands along environmental
gradients. Wetlands 13:65-74.
Brinson, M.M. 1995. The HGM approach explained. National Wetlands Newsletter
17:7-13.
Brinson, M.M. 1996. Assessing wetland functions using HGM. National Wetlands
Newsletter 18:10-16.
Brinson, M.M., F. R. Hauer, L. C. Lee, W. L. Nutter, R.D. Rheinhardt, R. D. Smith, D.
Whigham. 1996. A National Guidebook for Application of Hydrogeomorphic
Assessments to Riverine Wetlands: an Operational Draft. Technical Report
WRP-DE-11, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
Mississippi, USA.
43
-------
Brinson, M.M. and R. Rheinhardt. 1996. The role of reference wetlands in functional
assessment and mitigation. Ecological Applications 6:69-76.
Burke, M. A. 1996. Historic Evolution of channel Morphology and Riverine Wetland
Hydrologic Functions in the Piedmont of Georgia. M.S. thesis, University of
Georgia, Athens, GA, USA.
Burke, M. A. and W. L. Nutter. 1995. Channel morphology evolution and riverine
wetland hydrology in the Georgia Piedmont. Pages 463-475 in K. L. Campbell
(editor). Versatility of Wetlands in the Agricultural Landscape. American Society
of Agricultural Engineers, St. Josephs, Ml, USA.
Cirmo, C.P, and C.T. Driscoll. 1993. Beaver pond biogeochemistry: acid neutralizing
capacity generation in a headwater wetland. Wetlands 13:277-292.
Cowardin, L.M., V. Carter, F.C. Golet and E.T. LaRoe. 1979. Classification of Wetlands
and Deepwater Habitats of the United States. FWS/OBS- 79/31. Office of
Biological Services, Fish and Wildlife Service, US Department of the Interior,
Washington, DC, USA.
Costa, J.E. 1 974. Response and recovery of a Piedmont watershed from tropical
storm Agnes, June 1972. Water Resources Research 10:106-1 12.
Davis, W.S. 1995. Biological assessment and criteria: Building on the past. Pages 15
-29 in W.S. Davis and T.P. Simon (editors). Biological Assessment and Criteria.
Lewis Publishers, Boca Raton, Florida.
Denney, R.N. 1962. A Summary of North American Beaver Management. Current
Report #28. Colorado Game and Fish Department, Denver, CO, USA.
DeShon, J.E. 1995. Development and application of the invertebrate community index
(ICI). Pages 217-243 in W.S. Davis and T.P. Simon (editors). Biological
Assessment and Criteria. Lewis Publishers, Boca Raton, FL, USA.
Dowd, J., D. Wenner, and M. Carpenter. 1993. Piedmont geohydrology: implications
for flow and transport. Proceedings of the 1993 Georgia Water Resources
Conference, University of Georgia, Athens, GA, USA.
Fetter, C.W. 1988. Applied Hydrology, third edition. Macmillan College Publishing
>» Voi-Lo MV I ICA
vw • %>i rv, i % • | ** w/~t.
Geiger, J.G. 1994. An Assessment of Urban Piedmont Wetlands in North Carolina.
Master's Thesis. School of the Environment, Duke University, Durham, NC, USA.
44
-------
Hefner, J.M., B.O. Wilen, T.E. Dahl, and W.E. Prayer. 1994. Southeast Wetlands;
Status and Trends, Mid 1970's to Mid-1990's. U.S. Department of the Interior,
Fish and Wildlife Service, Atlanta, GA, USA.
Hughes, R.M. 1995. Defining acceptable biological status by comparing with
reference conditions. Pages 31 - 47 in W.S. Davis and T.P. Simon (editors).
Biological Assessment and Criteria. Lewis Publishers, Boca Raton, Florida.
Hughes, R. M., D. P. Larsen, and J. M. Omernik. 1986. Regional reference sites: a
method for assessing stream potentials. Environmental Management 10:629-
635.
Hunsaker, C.T. and D.E. Carpenter (editors). 1990. Ecological Indicators for the
Environmental Monitoring and Assessment Program. EPA 600/3-90/060. U.S.
Environmental Protection Agency, Office of Research and Development,
Research Triangle Park, NC, USA.
Johnston, C.A. and R.J. Naiman. 1987. Boundary dynamics at the aquatic-terrestrial
interface: the influence of beaver and geomorphology. Landscape Ecology
1:47-57.
Kantrud, H. A., G. O. Krapu, and G. A. Swanson. 1989. Prairie Basin Wetlands of the
Dakotas: A Community Profile. U.S. Fish and Wildlife Service Biological Report
85 (7.28), Office of Biological Services, Washington, DC, USA.
Karr, J. R. 1991. Biological integrity: A long neglected aspect of water resource
management. Ecological Applications 1:66-84.
Kentula, M.E., R.P. Brooks, S.E. Gwin, C.C. Holland, A.D. Sherman and J.C. Sifneos.
1992. An Approach to Improving Decision Making in Wetland Restoration and
Creation. U.S. Environmental Protection Agency, Environmental Research
Laboratory, Corvallis, OR, USA.
Kundell, J.E. and S.W. Woolf. 1986. Georgia Wetlands: Trends and Policy Options.
Carl Vinson Institute of Government, University of Georgia, Athens, GA, USA.
Kusler, J.A. and M. E. Kentula (editors). 1990. Wetland Creation and Restoration:
The Status of the Science. Environmental Protection Agency EPA 600/3-89/038,
Environmental Research Laboratory, Corvallis, OR, USA.
Leibowitz, S.G., B. Abbruzzese, P.R. Adamus, L.E. Hughes and J.T. Irish. 1992. A
Synoptic Approach to Cumulative Impact Assessment: A Proposed Methodology.
45
-------
EPA/600/R-92/167. U.S. Environmental Protection Agency, Environmental
Research Laboratory, Corvallis, OR.
Levy, G. F. and S. W. Walker. 1971. The combined Bitterlich-rangefinder-circular
quadrat method in phytosociological studies. Jeffersonia 5: 37-39.
Linsey, A. A., J. D. Barton, and S. R. Miles. 1958. Field efficiencies of forest sampling
methods. Ecology 39:428-444.
Loeb, S.L. and A. Spacie (editors). 1994. Biological Monitoring of Aquatic Systems.
Lewis Publishers, Ann Arbor, Ml, USA.
Lowrance, R. and 12 coauthors. 1995. Water Quality Functions of Riparian Forest
Buffer Systems in the Chesapeake Bay Watershed. Chesapeake Bay Program
Technology Transfer Report EPA 903-R-95-004. U.S. Environmental Protection
Agency, Annapolis, MD, USA.
Magee, T.K., S.E. Gwin, R.G. Gibson, C.C. Holland, J.E. Honea, P.W. Shaffer, J.C.
Sifneos and M.E. Kentula. 1993. Research Plan and Methods Manual for the
Oregon Wetlands Study. Document Production by K. Miller. EPA/600/R-93/072.
U.S. Environmental Protection Agency, Environmental Research Laboratory,
Corvallis, OR, USA.
McCallie, S.W. 1911. A Preliminary Report of Drainage Reclamation in Georgia.
USDA Bulletin No. 25. U.S. Department of Agriculture, Washington, DC.
Murray, D.F. 1961. Some Factors Affecting the Production and Harvest of Beaver in the
Upper Tanana River Valley, Alaska. M.S. Thesis, Univ. Alaska, Anchorage, AK,
USA.
National Technical Committee for Hydric Soils. 1995 Field Indicators of Hydric Soils
in the United States, Version 2.1. United States Department of Agriculture,
National Resources Conservation Service, Washington, DC, USA.
National Wetlands Policy Forum. 1988. Protecting America's Wetlands: An Action
Agenda. Conservation Foundation, Washington, DC, USA.
Nutter, W.L. 1973. The role of soil water in the hydrologic behavior of upland basins.
Pages 181-193 in The Field Soil Moisture Regime, Soil Science Society of
America Special Publication No. 5. Msdiscri, Wl, USA.
46
-------
Odum, E.P. and M.G. Turner. 1990. The Georgia landscape: a changing resource.
Pages 137-164 in I:S. Zonneveld and R.T.T. Forman (editors) Changing
Landscapes: An Ecological Perspective: Springer-Verlag, New York, NY, USA.
Omernik, J.M. 1987. Ecoregions of the conterminous United States. Annals of the
Association of American Geographers 77:118-125.
Omernik, J.M. and G.E. Griffith. 1991. Draft Ecoregions and Southeastern Plains
Subregions of Mississippi and Alabama. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Corvallis, OR. ManTech Environmental
Technology, Inc., Corvallis, OR, USA.
Oosting, H. J. 1942. An ecological analysis of the plant communities of Piedmont,
North Carolina. The American Midland'Naturalist 28:1-126.
Patterson, G.G., G.K. Sperry, and B.J. Whetstone. 1985. Hydrology and its Effects on
Distribution of Vegetation in Congaree Swamp National Monument, South
Carolina. U.S. Geological Survey Water-Resources Investigations Report 85-
4256, Washington, DC, USA.
Peet, R.K. and N.L. Christensen. 1980. Hardwood forest vegetation of the North
Carolina Piedmont. Veroff. Geobot. Inst. ETH, Stiftung Rubel, Zurich 69:14-
39.
Peters, J.A. 1988. Quality Control Infusion into Stationary Source Sampling. Pages
317-333 in L.H. Keith (editor). Principles of Environmental Sampling. American
Chemical Society, Professional Reference Book.
Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K. Gross, and R.M. Hughes 1989. Rapid
Bioassessment Protocols for Use in Streams and Rivers: Benthic
Macroinvertebrates and Fish. EPA/444/4-89-001. U.S. Environmental
Protection Agency, Office of Water, Washington, DC, USA.
Rankin, E.T. 1995. Habitat indices in water resource quality assessments. Pages
181-208 in W.S. Davis and T.P. Simon (editors). Biological Assessment and
Criteria. Lewis Publishers, Boca Raton, FL, USA.
Richter, D.D., K. Korfmacher, and R. Nau. 1995. Decreases in Yadkin River Basin
Sedimentation: Statistical and Geographic Time-trend Analysis, 1951 to 1990.
Report No. 297. Water Resources Research Institute, University of North
Carolina, Raleigh, NC, USA.
47
-------
Rosgen, D. 1995. Applied Fluvial Geomorphology. Wildland Hydrology Consultants,
Pagosa Springs, Colorado, USA.
Schafale, M.P. and A.S. Weakley. 1990. Classification of the Natural Communities of
North Carolina: Third Approximation. North Carolina Natural Heritage Program,
POB 27687, Raleigh, NC, USA.
Sharitz, R. R. and W. J. Mitsch.. 1993. Southern floodplai.n forests. Pages 311-372 in
W. H. Martin, S. G. Boyce, and A. C. Echtemacht (eds.), Biodiversity of the
Southeast: Lowland Terrestrial Communities. John Wiley and Sons, Inc., New
York, NY, USA.
Slough, B.C., and R.M.F.S. Sadleir. 1977. A land capability classification system for
beaver (Castor canadensis Kuhl). Can. J. Zool. 55:1324-1335.
Smith, R.D., A. Ammann, C. Bartoldus, and M. Brinson. 1996. A Procedure for
Assessing the Functional Capacity of Wetlands. Technical Report WRP-DE-9,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, USA.
Strahler, A.N. 1952. Hypsometric (area-altitude) analysis of erosional topography.
Bulletin of the Geological Society of America 63:1117-1142.
Ter Braak, C.J.F. 1988. CANOCO: a FORTRAN Program for Canonical Community
Ordination by Partial Canonical Correspondence Analysis, Principle
Components Analysis, and Redundancy Analysis (version 3.1). Agricultural
Mathematics Group, Wageningen, The Netherlands.
Trimble, S.W. 1970. The Alcovy River swamps: The result of culturally accelerated
sedimentation. Bulletin of the Georgia Academy of Science 28:131 -144.
Trimble, S. W., F. H. Weirich, and B. L. Hoag. 1987. Reforestation and the reduction
in water yield on the southern Piedmont since circa 1940. Water Resources
Research 23: 425-437.
Turner, M.G. 1987. Land use changes and net primary production in the Georgia,
USA, landscape: 1935-1982. Environmental Management 11:237-247.
USEPA (United States Environmental Protection Agency). 1990. Biological Criteria:
National Program Guidance for Surface Waters. EPA-440/5-90-004. Office of
VVaier Regulations and Standards (WH 585), »A'ashington, DC, USA.
van der Valk, A.G. and C.B. Davis. 1978. The role of seed banks in the vegetation
dynamics of prairie glacial marshes. Ecology 59:322-335.
48
-------
Wharton, C.H. 1978. The Natural Environments of Georgia. Geologic and Water
Resources Division and Resource Planning Section, Office of Planning and
Research, Georgia Department of Natural Resources, Atlanta, GA, USA.
Wharton, C.H., W.M. Kitchens, and T.W. Sipe. 1982. The Ecology of Bottomland
Hardwood Swamps of the Southeast: A Community Profile. FWS/OBS-81/37.
Biological Services Program, Fish and Wildlife Service, Washington, DC, USA.
49
-------
Appendix: Descriptions of sites visited by authors and Dennis Whigham in the Piedmont of
North Carolina on 6-10 June 1994. This list is not meant to imply that only these sites
would be sufficient for developing reference standards for Piedmont riverine wetlands, but
are provided to illustrate the ranges of conditions encountered.
Site No. 1: Eno River
Hydrologic source:
Eno River
USGS 7.5 min. Quad:
Northwest Durham, NC
Location:
Eno River State Park, upriver of picnic area off Route 1401 (Cole Mill Road); northwest of
Raleigh
NC Heritage Program designation(s):
Piedmont Alluvial Forest, Rocky Bar and Shore
Geomorphology and Soils:
The water of the Eno River in this reach is clear and is well-aerated by the flowing water. Only
one small reservoir is located upstream and much of the watershed is protected by Eno River
State Park. The river supports an alternating series of pools and riffles (gravel bars) and appears
to be bedrock controlled, i.e., the river flows over bedrock.
The alluvial floodplain of the river is located 3-4 m above the bedrock base and so is probably
only flooded on rare occasions. Because of this, the floodplain supports both mesic and
hydrophytic vegetation. The alluvial soils are unconsolidated and recently deposited. Therefore,
hydric characteristics are poorly developed, except in low-lying depressions in the floodplain
where rainwater and surface run-off collect. In such low spots, some mottling was found in soil
samples.
Vegetation:
Lizard's Tail (Saururus cemuus) and water willow (Justicia americana) dominate many of the
rocky bars, indicating that there has been relatively little fluctuation in water levels recently and
that the channel is relatively stable.
The alluvial floodplain harbors a specicse assemblage of tree species, including willow oak
(Quercus phellos), river birch (Betula n/g/a), ash (Fraxinus spp.), sycamore (Platanus
occidentalis), southern sugar maple (Acer bartatum), hackberry (Celtis laevigata), tulip poplar
(LJriodendron tulipifera), elm (Ulmus americana and U. alata), walnut (Juglans nigra), and
sweetgum (LJquidambar styraciflua).
The alluvia! subcanopy supports ironwood (Carpinus caroliniana), mulberry (Moms sp.),
deciduous holly (Ilex decidua), eastern red cedar (Juniperus virginiana), dogwood (Comus
florida), mountain laurel (Kalmia latifolia), hop hornbeam (Ostrya virginiana), sourwood
(Oxydendron arboreum), bladdemut (Staphylea trifolia), dwarf buckeye (Aesculus sylvatica), and
southern arrowwood (Viburnum dentatum).
50
-------
The herbaceous stratum is relatively well-developed. We encountered Uniola sp. (U. latifolia or
U. laxa), Botrychium virginianum, Lycopus sp., Polystichum acrostichoides, Tovara Virginians,
Arisaema dracontium, Galium sp., Hexastylis virginica, and Elymus virginicus.
Prevalent vines include Virginia creeper (Parthenocissus quinquefolia), wild grape (Vitus spp.),
poison ivy (Rhus radicans), cross vine (Bignonia capreolata), trumpet creeper (Campsis
radicans), and greenbiar (Smilax spp.).
Exotic pest species were also encountered, including Microstegium virmineum (a grass) and the
shurbs, Japanese honeysuckle (Lonicera japonica) and Japanese privet (Ugustrum sinense).
Additional observations:
Wrack from recent flooding was evident approximately 3-4 m above current stream level. Also,
some course woody debris in part due to beaver activity. Beaver signs fairly old. Frogs
abundant and tadpoles rearing in pools. Pools in streambed associated with possible hypomeic
flow. Some sand accumulation was found behind trees that had fallen into stream channel. Iron
deposits at base of slope near stream channel indicate discharge point.
Well developed levee upstream from bridge has behind it a backswamp wetland. This is where
most of the plant species listed above are found.
Site No. 2: New Hope Creek
Hydrologic source:
New Hope Creek
USGS 7.5 min. Quad:
Chapel Hill
Location:
Duke Forest, upstream of Route 1734 (Erwin Road), between Chapel Hill and Durham
NC Heritage Program designation(s):
Piedmont Alluvial Forest
Geomorphology and Soils:
New Hope Creek flows from out of a Triassic Basin at this location. The stream and floodplain
narrow considerably just before the creek exits the basin at Route 1734 and the remains of an
old dam are still in place at the constriction. Rocks and boulders are prevalent in and along the
streambed. Although New Hope Creek is bedrock controlled, the sediment load in the river was
high at the time of our visit. Except for one small reservoir on one tributary of New Hope Creek,
the watershed is not impounded at this time.
The alluvial floodplain is several meters above the river-bed and the soil does not show well-
developed hydric characteristics, except in low-lying depressions.
Vegetation:
Because the floodplain forest along this section of New Hope Creek has been protected for many
years (it is a part of the Duke Research Forest), trees are relatively large, with many in excess of
51
-------
1 m in diameter at breast height (dbh). The canopy consists of sycamore, American elm,
sweetgum, tulip poplar, willow oak, walnut, ash, elm, hackberry, red maple (Acer rubrum), river
birch, and others.
The subcanopy harbors boxelder (Acer negundo), spicebush (LJndera benzoin), dwarf buckeye,
iron wood, American holly (Ilex opaca), bladdemut, southern sugar maple, and southern
arrowwood. Prevalent vines include Virginia creeper, wild grape, poison ivy, cross vine, trumpet
creeper, and greenbiar. The herb stratum is rich with Impatiens capensis, Polystichum
acrostichoides, Uniola (U. latifolia or U. laxa), and Urtica dioica.
Site No. 3: Meadow Flats Depression Swamp
Hydrologic source:
Infiltration from upslope into low-lying basin
USGS 7.5 min. Quad:
Chapel Hill
Location:
Duke Forest, north of Route 1727 (Eubanks Road) and east of Route 1009 (Chapel Hill
Road), northeast of Chapel Hill
NC Heritage Program designation(s):
Upland Depression Swamp Forest
Geomorphology and Soils:
This depression swamp is in a basin with higher elevation hills on three sides. The area of the
depression swamp appears to drain to the southeast to Bolin Creek and to the northeast toward
Mountain Creek, but the topography is extremely flat throughout. Soils are reduced with much
mottling; a hardpan underlies the depression.
The depressional swamp is part of Duke Forest and so is currently protected. The NC Natural
Heritage Program recognizes this forest as the best example of a basin depression forest in
North Carolina.
•Vegetation:
The canopy is dominated by willow oak and swamp chestnut oak (Quercus michauxii), with
sweetgum, shagbark hickory (Carya ovate), and American elm prevalent as well. Also present in
the canopy is red mapie, winged eirn. eastern red cedar, and Florida basswood (Jllia fioridanaY
The subcanopy harbors ironwood, southern arrowwood, deciduous holly, highbush blueberry
(Vaccinium corymbosum), and black haw (Viburnum prunffolium). The herbaceous stratum
includes Carex lupulina, Saurums cemuus, Arisaema dracontium, Boehmeria cylindrica, and
Sphagnum spp. Prevalent vines include greenbriar, wild grapes (Vitus sppj. and poison ivy.
Additional observations:
On the way into the site, the ruins of a house with concrete steps indicates that the site was
probably highly disturbed at one time. The site is also described in the following thesis:
52
-------
Whigham, D.F. 1971. An ecological study of Uvularia perfoliata in the Blackwood Division of
Duke Forest, North Carolina. Ph.D. thesis, Unviersity of North Carolina, Chapel Hill, NC.
Site No. 4: Haw River Floodplain
Hydrologic source:
Haw River
USGS 7.5 min. Quad:
Merry Oaks
Location:
Immediately south of Route 64 on the west bank of the Haw River, just upriver of the
effects of the Jordan Lake Reservoir
NC Heritage Program designation(s):
Piedmont Levee Forest, Rocky Bar and Shore
Geomorphology and Soils:
This site is located on the alluvial floodplain of the Haw River before the river enters Jordan Lake.
The river has eroded down to bedrock since reforestation of the Piedmont, but the alluvial
floodplain is 2-3 m above the river bed. Like much of the Piedmont alluvial forests, the soils are
composed of red alluvium, with little sign of being hydric except in the low-lying depressions on
the floodplain. The river bed is extremely rocky with large boulders scattered throughout the river
bed; the river probably representative of larger rivers on the eastern flank of the Piedmont.
The Haw River has supports 4 dams upriver and so probably receives little sediment from far
upstream. However, bank erosion is prevalent as evidenced by the highly discolored water. This
site and the next one (#5) provides a good example of Piedmont rivers hydrologically modified by
impoundments.
Vegetation:
The vegetation of this levee forest includes both mesic and hydrophytic vegetation. The canopy
contains sycamore, hackberry, river birch, and sweetgum. Mesic species such as beech (Fagus
grand/folia) and tulip poplar are also present, suggesting that the floodplain is now rarely flooded.
The subcanopy is dominated by bladdemut, spicebush, and ironwood, the herb stratum by Uniola
sp., Solidago spp., Geum canadense, and Microstegium virmineum. The vine stratum includes
Lonicera japonica and Rhus radicans.
Site No. 5: Jordan Lake
Hydrologic source:
Jordan Lake
USGS 7.5 min. Quad:
Merry Oaks
Location:
East of Route 1943 and north of the now flooded Roberson Creek at the Corps of
Engineers canoe landing; flooded section of the Haw River
NC Heritage Program designation(s):
NA
53
-------
Geomorphology and Soils:
This site is one of the upper arms of Jordan Lake, which now intersects the old Haw River
floodplain. Fluctuations in the lake level now control the hydrologic regime of the site. Flow is
almost imperceptible and so the area is a depositional environment. Soils were not examined at
this site.
Vegetation:
Relic floodplain species such as sycamore, hackberry, and river birch grow along the shore. It is
difficult to predict the future composition of the forest, since much will depend upon the
manipulation of lake levels and infilling by sediment.
Site No. 6: Rocky River
Hydrologic source:
Rocky River and small earthen dam
USGS 7.5 min. Quad:
Siler City
Location:
On east bank of Rocky River north of Route 64, 2 miles east of Siler City, NC
NC Heritage Program designation(s):
NA
Geomorphology and Soils:
This river is another deeply incised Piedmont river with a relic floodplain 2-3 m above the river
bed. However, at this location, the river bottom is composed of mud and sand (rather than being
rocky like the previously described rivers). Thus, this section of the Rocky River possibly
represents a fourth order river in the interior Piedmont. Soils on the relic floodplain are
composed of unconsolidated alluvium. Upstream of Route 64, the river is impounded by a small
earthen dam. Relic floodplain soils adjacent to the impoundment are highly reduced and
possess a low chroma.
Vegetation:
The canopy of the high banks and relic floodplain supports ash, sycamore, beech, and wild black
cherry (Prjnus serstins}. The subcanopy supports spicebush and Japanese privet. Prevalent
vines include poison ivy and Japanese honeysuckle. The herbaceous stratum shows high cover
of Uniola sp. and Urtica dioica, along with scattered Carex lupulina. The saturated area adjacent
to the impoundment supports a much more luxuriant herbaceous layer and includes more
hydrophytic plants, including Impatiens capensis, Boehmeria cylindrica, Bidens spp., Arisaema
triphyllum, and Saururus cemuus.
54
-------
Site No. 7: Loves Creek (urban wetland)
Hydrologic source:
Loves Creek
USGS 7.5 min. Quad:
Siler City
Location:
On the east side of Route 421 about 1/2 mile south of Siler City, just downstream from a
lumber yard
NC Heritage Program designation(s):
NA
Geomorphology and Soils:
This stretch of Loves Creek is near town and just downstream from a lumber yard. The
floodplain and stream are narrow. The creek is about 20 m wide and the tributary about 2 m
wide. The creek at this reach of its course is straight and has steep banks (about 1 m high);
thus, it appears that the creek has been straightened and now probably rarely overflows it banks.
This site likely represents many small, urban, Piedmont streams.
Vegetation:
The canopy includes sycamore, ash, sweetgum, and tulip poplar. The subcanopy is dense with
much Japanese privet and some ironwood. Japanese honeysuckle, poison ivy, and greenbriar
are also dense in the understory.
Site No. 8: Uhwarrie River
Hydrologic source:
Uhwarrie River
USGS 7.5 min. Quad:
Badin
Location:
About 1 mile west of Uwharrie, NC on the west bank of the Uwharrie River just
downstream of Moccasin Creek and to the east of FR (Forest Road) 555.
NC Heritage Program designation(s):
Piedmont Levee Forest
Geomorphology and Soils:
The Uwharrie is a free-flowing (unimpounded) Piedmont river; most of its watershed is now
within National Forest. Like other Piedmont rivers, the relic floodplain is elevated several meters
above the river bed and the steep banks are eroded during high flows. Also typical, the soil is
unconsolidated alluvium with few typical hydric indicators present except in the lowest-lying
depressions. A low elevation levee is located along this stretch of the river (just downstream of a
major bend and the mouth of Moccasin Creek). The floodplain is quite narrow except at the
sharp bend in the river where it becomes wide and appears to slope to lower elevations with
increasing distance from channel.
55
-------
Vegetation:
The canopy supports loblolly pine (Pinus taeda), tulip poplar, sweetgum, ash, and American elm.
The subcanopy is rather open with spicebush and ironwood in the understory. The herb layer
includes Geum canadense, Uniola sp., Aster divaricatus, Viola spp., and Arisaema triphyllum.
Vines include poison ivy and Virginia creeper. The forest appears to be fairly young as it does
not contain shade-tolerant canopy species.
Site No. 9: Forest Service Tributary
Hydrologic source:
small forest stream
USGS 7.5 min. Quad:
Badin
Location:
Unnamed tributary to Badin Lake that crosses FR 576 just to the west of PR 597 in
Uwharrie National Forest
NC Heritage Program designation(s):
NA
Geomorphology and Soils:
This tributary is a small, first order stream flowing through the National Forest. It flows north and
crosses through a culvert under FR 576. During rains, runoff from the unpaved Forest Service
road carries so much sediment that the creek turns red. Upstream of the culvert, where no roads
cross the creek, the stream runs clear even during heavy downpours.
This site shows how easily Piedmont soils are eroded when vegetative cover is removed and
how protecting riparian zones adjacent to streams (both first and higher order) is essential for
maintaining water quality in the Piedmont. Numerous four-wheel drive roads have been
established in recent years for recreational driving and hunting access. During rains, these roads
carry a large amount of sediment to the streams in the forest.
Vegetation:
The riparian zone supports beec!., ash, white oak (Quercus alba), and red maple. The
subcanopy supports spicebush, ironwood, blueberry species, and flowering dogwood (Comus
florida). The herb stratum includes Arisaema triphyllum, Boehmeria cylindrica, Polystichum
scrostichoides. Carex spp., Aster spp., and Geum canadense.
Site No. 10: Hilltop Depression Swamp
Hydrologic source:
Mostly rainwater, some infiltration from upslope
USGS 7.5 min. Quad:
Badin
Locslicr,:
Uwharrie National Forest, top a hill at 750' elevation about 1/2 mile east of Sadin Dam.
NC Heritage Program designation(s):
Upland Depression Swamp
56
-------
Geomorphology and Soils:
Two depression swamps are located on this hill; both are about 1-2 ha in size and are underlain
by a clay hardpan that inhibits drainage. Depression swamps and the communities they support
are rare across the Piedmont landscape, but it is likely that many are still unrecorded. These
systems are very unlike the surrounding xeric landscape in supporting hydrophytic and mesic
vegetation.
Vegetation:
The canopy harbors willow oak, sweetgum, red maple, and overcup oak (Quercus lyrata). The
subcanopy, which is sparse, primarily supports Vaccinium spp. The herb stratum is dominated
by Carex spp. and Sphagnum spp.
Additional observations: Parts of these depressions are sloping and one has a non-channelized
outlet. Debatable as to whether these are depression or slope wetlands. Stakes and flagging
indicate that areas are being studied. Drift fence for amphibians along one site with pitfall traps.
It rained a couple of inches while we were in the Uwharrie. Leaves rather thick in some of the
lower microelevations. Mottles appear very near the surface.
Site No. 11: Brown Creek
Hydrologic source:
Brown Creek, a tributary of the Pee Dee River
USGS 7.5 min. Quad:
Ansonville
Location:
Bennett Bridge on Route Pinkston-River Road in the Pee Dee National Wildlife Refuge.
NC Heritage Program designation(s):
Piedmont Swamp and Bottomland Forest
Geomorphology and Soils:
Brown Creek is a tributary of the Pee Dee River located within a broad Triassic Basin floodplain.
However, the floodplain is a relic of deposition that occurred during an earlier period of severe
erosion in the late nineteenth century. As a result of the reforestation of agricultural lands
(primarily cotton plantations), the creek incised its channel back to its original bedrock elevation.
Thus, today the floodplain lies approximately 3 m above the streambed.
During major floods, the creek probably erodes the banks and conveys sediment downriver, but
flooding of the relic floodplain probably rarely occurs today. Soils on the floodplain fail to show
hydric conditions except in lower-lying depression that pond during rains. In such ponded areas,
slight mottling can be observed, but chromes tend to be high. Leaves accumulating in one of the
depressions that receives some drainage from the roadside ditch.
57
-------
Vegetation:
The canopy tends to represent relic conditions when the floodplain flooded. Water oak (Quencus
nigra), sweetgum, loblolly pine, willow oak, blackgum (Nyssa biflora), elm, hackberry, and red
maple occur in the canopy. Ironwood, Japanese privet, pawpaw (Asimina triloba), and spicebush
are scattered throughout the subcanopy, while Uniola spp., Carex spp., Saurvrvs cemuus, and
Boehmeha cylindrica inhabit depressions in the floodplain.
Additional observations:
'*'
At bridge, the site is being excavated to uncover old ford-bridge that represented original location
of floodplain before considerable upland erosion around the turn of the century. Some beaver
activity, but seems to be old. Wrack on first bottom indicates flooding within past year.
Floodplain has low sites that pond with rainfall which results in mottling very near the surface in
comparison with areas adjacent and maybe only 10-20 cm higher in elevation. Consequently,
hydric soil characteristics are now due to ponding of precipiation rather than overbank flooding.
58
------- |