ABSTRACT
Deborah D. Noltemeier. COMPARISON OF WATER QUALITY IN
COASTAL PLAIN STREA_MS OF NORTH CAROLINA (Under the direction
of Mark M. Brinson) Department of Biology, November, 1984.
A one year study was conducted in the central Coastal
Plain of North Carolina to examine water chemistry in a
Piedmont-draining alluvial river (and its floodplain), and
in ten Coastal Plain-draining streams which were further
subdivided into two general categories based on their
geologic origin: (1) peat-draining streams and (2) mineral
soil-draining streams (including a channelized stream).
Objectives of this research were to demonstrate similarities
and differences in water chemistry and to determine if
streams could be distinguished and categorized on the basis
of the chemical composition of their waters.
The results of this study elucidated differences in the
chemical composition of streamwater v/hich can be used to
distinguish the geologic origin of Coastal Plain-draining
streams. The characteristics of peat-draining streams that
distinctively separate them from, mineral soil-draining
streams are higher concentrations of organic carbon and
lower levels of dissolved oxygen, biochemical oxygen demand,
pH, total alkalinity, conductivity, cations (Ca, Mg, Na, K
and Fe) , inorganic nitrogen, and phosphorus. V7hen compared
to the two categories of streams originating in the Coastal
Plain, the Piedmont-draining alluvial river had higher
levels of dissolved oxygen and cations; lower concentrations
of organic carbon; and levels of biochemical oxygen demand,
pH, total alkalinity, conductivity, iron, inorganic and
organic nitrogen, and phosphorus that were between those of
the other two categories. Categorical separations of
streams, based on their geologic origin, are basically
sound; however, large watersheds often possess diverse
lithology which results in overlapping stream types.
Anthropogenic influence was superimposed on Coastal
Plain-draining streams as a result of v/astewater discharges,
agricultural activities, animal husbandry operations, and
channelization. Localized loadings of nitrogen and
phosphorus in Chicod Creek and Grindle Creek resulted in
high concentrations of these nutrients in streamwater, which
indicated that the assimilative capacity of these streams
had been exceeded. Channelized Grindle Creek had perennial
flow; higher levels of cations, conductivity, and nutrients;
and lower dissolved organic carbon concentrations, which
clearly distinguished it from the natural streams.
COMPARISON OF WATER QUALITY
IN COASTAL PLAIN STREAMS OF
NORTH CAROLINA
A Thesis
Presented to
the Faculty of the Department of Biology
East Carolina University
In Partial Fulfillment
of the Requirements for the Degree
Master of Science in Biology
by
Deborah D. Noltemeier
November, 1984
3. T. iOYNKH LIBSARÜ
M a fw Á’ T7WTVWO
4o Cj:>r
G B
/J ?
COMPARISON OF WATER QUALITY
A]
IN COASTAL PLAIN STREAMS OF
/9Si
NORTH CAROLINA
by
Deborah D. Noltemeier
APPROVED BY:
SUPERVISOR OF THESIS /Ćx 07).
Mark M. Brinson, Ph.D.
DEAN OF GRADUATE SCHOOL
ACKNOWLEDGEMENTS
This study represents the combined efforts of a number
of people to whom I am indebted. I express my sincere
appreciation to my thesis chairman, Dr. Mark M. Brinson, for
his guidance, assistance, and patience throughout this
project. His support and encouragement, as well as critical
reviews of this thesis, have ultimately lead to com.pletion
of this project. I also thank the other members of my
thesis committee. Dr. Lee Otte, Dr. Robert Christian, and
Dr. Graham Davis, for critically reviewing this manuscript
and for providing valuable suggestions. Martha Jones
deserves special recognition and thanks for supervising and
performing the laboratory analyses and for her assistance in
various aspects of this project. David Bradshaw collected
the Tar Swamp rainfall samples and provided technical
advice. Jerry Freeman maintained and repaired equipment
involved in this project. Steve Nelson assisted in
performing nutrient analyses. Randy Creech deserves special
recognition for computer programming. Tim Charles and
Laddie Crisp provided technical assistance. Bruce Payne
drafted the figures. Teresa Smith and Ben Loftin provided
assistance in various aspects of this project. Mary Grace
Pate deserves special recognition for professional
excellence in word processing. This study was supported in
part by the U.S. Environmental Protection Agency, grant
R-805881.
I am especially grateful to my father, Al Noltemeier,
for accompanying me on many field collections and providing
assistance in field work. I dedicate this thesis to my
parents with special thanks for their support and
encouragement throughout this study and in all of my
academic endeavors.
TABLE OF CONTENTS
Page
LIST OF TABLES V
LIST OF FIGURES vi
LIST OF APPENDIX TABLES vii
INTRODUCTION 1
Factors Affecting Water Quality of Streams ... 3
Features of this Study 12
STUDY AREA 17
Coastal Plain-draining Streams 22
Peat-draining streams 23
Mineral soil-draining streams 27
Alluvial River and Swamp 31
METHODS 3 5
Field Measurements, Collections, and Samiple
Preparations 35
Chemical Analyses 41
Data Processing and Analysis 44
RESULTS 47
Dissolved Oxygen and Biochemical Oxygen Demand . 48
pH, Total Alkalinity and Conductivity 51
Cations 58
Nitrogen 66
Inorganic forms 66
Organic forms 70
Phosphorus 7 3
Organic Carbon 77
DISCUSSION 81
Seasonal Trends 81
Characteristics of Peat-draining Streams .... 86
Comparison of Water Chemistry Between North
Carolina and Minnesota Peatlands 92
Characteristics of Natural Mineral Soil-draining
Streams 96
Effects of Channelization on a Mineral Soil-
draining Stream 99
Characteristics of a Mineral Soil-draining
Stream Associated with Limestone Formations . . 103
Characteristics of a Piedmont-draining
Alluvial River 105
iii
TABLE OF CONTENTS
Page
DISCUSSION (cont'd.)
Exports to Downstream Ecosystems 108
Conclusions 112
LITERATURE CITED 116
APPENDIX 124
IV
LIST OF TABLES
Page
1. Stream categories with representative sampling
sites and watershed areas 13
2. Annual arithmetic means (A), with standard errors,
and annual weighted means (W) of concentrations of
rainfall constituents in the Croatan (RGC) and
Tar Swamp (RGT) rain gauge collections 54
3. Comparison of water chem.istry in North Carolina
streams and Minnesota bogs and fens 93
4. Annual runoff and export of nitrate-nitogen,
ammonium-nitrogen, dissolved organic nitrogen,
particulate organic nitrogen, total phosphorus,
total organic carbon, and dissolved organic carbon
from September 1978 through August 1979 109
LIST OF FIGURES
Page
1. Map of North Carolina showing the location of the
study areas 18
2. Preparation of field samples for chemical analyses 39
3. Annual arithmetic means and weighted m.eans for (a)
dissolved oxygen (0_) and (b) biochemical oxygen
demand (BOD ) from September 1978 through August
1979 49
4. Annual arithmetic means and weighted means for (a)
pH and (b) total alkalinity (TA) from September
1978 through August 1979 52
5. Annual arithmetic means and weighted m.eans for (a)
conductivity (COND) and (b) calcium (Ca) from
September 1978 through August 1979 56
6. Annual arithmetic means and weighted means for (a)
magnesium (Mg) and (b) sodium (Na) from September
1978 through August 1979 60
7. Annual arithmetic means and weighted means for (a)
potassium (K) and (b) iron (Fe) from September 1978
through August 1979 63
8. Annual arithmetic means and weighted m.eans for
nitrate-nitrogen (NO^-N), (b) ammonium-nitrogen
(NH .-N) , and (c) nitrite-nitrogen (NO„-N) from
September 1978 through August 1979 67
9. Annual arithmetic means and weighted means for (a)
dissolved organic nitrogen (DON) and (b) particulate
organic nitrogen (PON) from Septem.ber 197 8 through
August 1979 71
10. Annual arithmetic means and weighted means for (a)
total phosphorus (TP), (b) filterable reactive
phosphorus (FRP), and (c) particulate phosphorus
(PP) from September 1978 through August 1979 .. . 74
11. Annual arithmetic means and weighted means for (a)
dissolved organic carbon (DOC) and (b) particulate
organic carbon (POC) from September 1978 through
August 1979 78
12. Seasonal patterns of dissolved organic carbon (DOC)
and discharge (Q) at (a) East Brice Creek (EB), (b)
Island Creek (IS), (c) Palmetto Swamp (PA), and (d)
Tar River (TR) from September 1978 through August
1979 83
VI
1ST OF APPENDIX TABLES
Page
1. Annual arithmetic mean concentrations (A), with
standard errors, and weighted mean concentrations
(W) for nitrogen and phosphorus (in mg/1) from
September 1978 through August 1979 125
2. Annual arithmetic mean concentrations (A), with
standard errors, and weighted means concentrations
(W) for total organic carbon, dissolved organic
carbon, calcium, magnesium, sodium, potassium and
iron (in mg/1) from September 1978 through August
1979 126
3. Annual arithmetic mean concentrations (A), with
standard errors, and weighted means concentrations
(W) for pH, total alkalinity (in meq/1), conducti-
vity (in um.ho/cm), temperature, dissolved oxygen
(in mg/1), biochemical oxygen dem.and (in mg/1) and
discharge (in m^/sec) from September 1978 through
August 1979 127
4. Linear correlations (r) between chemical and
physical factors for East Brice Creek (EB) .... 128
5. Linear correlations (r) between chem.ical and
physical factors for West Brice Creek (WB) .... 129
6. Linear correlations (r) between chemical and
physical factors for Black Swamp (BL) 130
7. Linear correlations (r) between chemical and
physical factors for Mill Creek (MI) 131
8. Linear correlations (r) between chemical and
physical factors for Catfish Lake drainage
canal (CL) 132
9. Linear correlations (r) between chemical and
physical factors for Island Creek (IS) 133
10. Linear correlations (r) between chemical and
physical factors for Creeping Swamp (CS) 134
11. Linear correlations (r) between chemical and
physical factors for Palmetto Swamp (PA) 135
12. Linear correlations (r) between chemical and
physical factors for Chicod Creek (CH) 136
13. Linear correlations (r) between chemical and
physical factors for Grindle Creek (GR) 137
Vll
LIST GF APPENDIX TABLES (continued)
Page
14. Linear correlations (r) between chemical and
physical factors for Tar River (TR) 138
15. Linear correlations (r) between chemical and
physical factors for Tar Swamp (TS) 139
viii
INTRODUCTION
The Coastal Plain of North Carolina is characterized by
a diversity of wetland ecosystems due to the low topography,
poor drainage, and abundant rainfall typical of the area
(Kuenzler et al. 1977). At the time of Wilson's (1962)
report, it was estimated the areas of v/etland types were:
bogs (9,160 km^), sounds and bays (6,050 km^), wooded swamps
(4,000 km2), seasonally flooded bottomlands (1,860 km^),
coastal open fresh water (1,560 km^) , irregularly flooded
salt marsh (406 km^), inland open fresh water (357 km^),
regularly flooded salt marsh (236 km^), and fresh marsh (192
km2 ) .
Wetlands identified by Wilson (1962) as bogs are now
recognized by the descriptive name pocosin, which is derived
from an Indian word meaning "swamp on a hill" (Richardson
1981). Pocosin wetlands comprise over 50% of North Carolina
wetlands in the Coastal Plain; however, they are under
severe developmental pressure with only 31%, or less, left
in their natural state (Richardson 1983). Pocosins are
blocked drainage systems which formed along the Coastal
Plain, approximately 10,000 years ago, as a result of a
complex interaction of climatic, hydrologic, geologic, and
biological processes (Otte 1981). Pocosins typically have a
shrub-dominated vegetational structure situated on a broad
topographic plateau and are characterized by an accumulation
of autochthonous organic matter as peat, a water table vihich
2
remains at the surface of the pocosin throughout most of the
year (resulting in a waterlogged substrate and predominately
anaerobic edaphic conditions), acidic soils, and periodic
fires (Snyder 1978, Christensen et al. 1981, Otte 1981,
Richardson et al. 1981, Sharitz and Gibbons 1982, Ash et al.
1983). Only a few studies (Kirby-Smith and Barber 1979,
Skaggs et al. 1980, Daniels 1981) have been conducted on the
water chemistry of streams draining pocosin wetlands. The
question arises, how does the water chemistry of peat-
draining streams differ from the water chemistry of stream.s
draining mineral soils? This study v./ill attempt to answer
that question.
Why is the study of stream water chemistry important?
Streams are the circulatory system in an ecosystem which
transport nutrients and organic matter from upstream eco-
systems along a gradient to downstream aquatic ecosystems.
In this hydrological circulatory system, tributaries are
capillaries, streams are venules, and rivers represent veins
returning water to the heart of the system--the estuary, and
ultimately, the sea. Water is continuously circulated in
this solar-powered system by evaporation and precipitation.
Knowing that our very existence is intricately linked to the
hydrological cycle, and that the quality of water in streams
regulates our use of this water, makes the study of water
chemistry of paramount importance.
3
Factors Affecting Water Quality of Streams
Lotie ecosystems are open, primarily heterotrophic
ecosystems (Odum 1971, Wetzel 1983) that are energetically
dependent on allochthonous input from longitudinal coupling
(upstream-downstream) and lateral coupling (stream-flood-
plain) (Wharton and Brinson 1978) . Allochthonous energy
input is usually in the form of particulate and dissolved
organic matter (Wetzel 1983).
Sources of stream water are direct rainfall, overland
runoff, and groundwater seepage (base flow) (Wetzel 1983).
The chemical constituents in stream water are influenced by:
(1) precipitation input, (2) the regional geology, (3) the
amount and chemical composition of base flow contributions
to the stream channel, (4) interaction with the biotic
components of the lotie ecosystem, (5) riparian vegetation
(which provides a source of allochthonous energy influx into
the stream and shading), (6) organic productivity and the
photosynthesis-respiration balance in the stream, and (7)
land and water utilization in the watershed and hum.an
activities in the stream channel (Kuenzler et al. 1977).
Aquatic and terrestrial ecosystems are coupled by the
hydrologic cycle such that chemical and physical charac-
teristics of lotie ecosystems are largely dependent upon the
biological, lithological, and topographical attributes of
the watershed. The biotic community of the watershed
contributes allochthonous energy to the stream in the form
4
of litterfall and other organic debris. The soil biotic
community plays an essential role in decomposition and
nutrient cycling in the terrestrial environment of the
watershed which, in turn, affects the chemical composition
of runoff and base flow reaching the stream. Energy flow in
small streams with forested watersheds is primarily depen-
dent upon allochthonous energy input, since shading, as v/ell
as current, tends to limit autochthonous primary production
(Goldman and Horne 1983 , V7etzel 1983). Although the major
energy input pathway in small streams is exogenous, autoch-
thonous primary production by aguatic macrophytes, algae,
aquatic mosses, and encrusted diatoms (in illuminated
reaches of the stream) may also contribute to stream ener-
getics on a seasonal basis (Goldman and Horne 1983).
Detritus serves as the major energy source for heterctrophic
metabolism in lotie ecosystems (Wetzel 1983).
Lithological attributes of the watershed largely
determine the inorganic chemical com.position of streamwater
as well as controlling the amount of runoff reaching the
stream. The chemical composition of streamwater is influ-
enced by weathering and erosion of parent material and soils
in the watershed (Hem 1970). The chemical composition of
streamwater is also influenced by climate, gases and aero-
sols from the atmosphere, activities of humans in the stream
and surrounding watershed, and biogeochemical cycling.
5
Runoff, erosion, and subsurface infiltrating water
(base flow) are agents of physical weathering which contrib-
ute minerals to streamwater. Rates of weathering are
determined by temperature, the amount of precipitation, the
chemical constitution of parent and soil minerals, as well
as the influence of vegetation. The chemical composition of
parent material and soil minerals governs their solubility
and ion-exchange capacities with groundwater and with
streamwater (Hem 1970) .
The infiltration capacity of the watershed for pre-
cipitation partially depends on the composition of the
soils. Sandy soils have a low field capacity and readily
allow precipitation to infiltrate, thereby minimizing runoff
and providing an opportunity for aquifer recharge. Alterna-
tively, clayey soils, with a high field capacity, possess
small pore spaces that restrict rainfall infiltration,
especially when dehydrated (or conversely, saturated). This
results in increased runoff to streams which limits the
potential aquifer recharge in clay-dominated regions (Heath
1980). Peats are organic soils that are characterized by a
great capacity to retain water and maintain an exceptionally
high moisture content (Ingram and Otte 1981). For these
reasons, peat accumulation also involves the accumulation of
a significant amount of water within an ecosystem.
The seasonal precipitation/evapotranspiration balance
controls the v/ater table level and hence, the infiltration
capacity of wetlands for precipitation. The amount of
6
runoff to streams is also affected by the time of year
rainfall occurs due to the influence of évapotranspiration
on water table levels. In summer and fall, vegetation
effectively removes water by évapotranspiration, which
subsequently lowers the groundwater table resulting in
increased infiltration capacity for precipitation. During
this time of the year, overland runoff to streams from the
watershed is minimal. In winter and early spring, when
deciduous vegetation is in a leafless condition, precipita-
tion exceeds évapotranspiration output resulting in high
water tables and hence, rainfall rejection since the soil is
saturated. Runoff rates would be highest during these
seasons.
Local topographic conditions may influence stream, v/ater
quality. Topographic lows, along stream margins, form
floodplains for streams. These seasonally flooded bottom-
lands are important in removing sedim.ent from floodwaters,
thereby enhancing the quality of water transported to
downstream^ ecosystems. Within the floodplain, various
biogeochemical interactions between floodwaters and the
swamp forest ecosystem occur which significantly affect
stream water quality as floodwaters are gradually released
back into the stream (Wharton et al. 1982, Brinson et al.
1983, Yarboro 1983).
7
Conversely, topographic highs along stream margins
minimize interaction with the adjoining terrestrial water-
shed (when compared to streams with broad floodplains). In-
creased water velocity during spates, due to flow contain-
ment by high banks, would result in increased erosion and
hence, increased transport of suspended sediments as in
channelized streams. However, the high suspended sedim.ent
load characteristic of Piedmont-draining streams is more a
function of steeper stream gradients in the Piedmont and the
greater susceptibility of soils there to erode because of
the rolling topography.
The elevated topographic condition of pocosins re-
stricts surface water influx from surrounding terrestrial
ecosystems. Thus nutrient input to pocosins is limited to
precipitation and atm^ospheric depositions. Peat thickness
isolates the water held within the peat from contact with
underlying mineral soils and the regional groundwater
system. As a result the nutrient levels in pocosin drainage
water are low.
Finally, precipitation may have a major influence on
streamwater chemistry. Precipitation is a major water input
pathway in most streams, however indirectly, and therefore
the chemical composition of rainfall is an important factor
to consider in analysis of stream water chemistry. V7hen
precipitation reaches terrestrial ecosystems its chemical
composition becomes modified as it passes over vegetation
and dissolves nutrient-rich exudates from the leaves. When
P
throughfall and stemflow reach the ground surface, inter-
actions with the surface litterfall and soil humus layer
occur before infiltration into the soil further alters
precipitation chemistry.
Cations and anions dissolved in rainfall originate from,
diverse sources including oceanic spray, terrestrial dust,
gaseous pollutants, and volcanic emissions (Likens et al.
1977). Strong acids such as nitric and sulfuric acid
contribute to the acidity of rainfall. Hall et al. (1980)
experimentally demonstrated that concentrations of Ca, Mg,
K, and Fe increased as a result of stream acidifications in
the Hubbard Brook Experimental Forest; however, there was no
change in DOC, Na, NO^-N, or NH^-N levels at the lowered pH.
The strong coupling between streams and the surrounding
terrestrial watershed has been demonstrated in various
studies. Fisher and Likens (1973) prepared an annual energy
budget for Bear Brook in a forested area of northern Nev;
Hampshire. This study revealed that over 99% of the energy
input comes from the surrounding forested watershed or
upstream areas. In Bear Brook 66% of the organic input was
exported downstream, leaving 34% to be utilized locally.
Once this material enters the stream system it undergoes
biochemical and mechanical degradation m.ediated by the
biotic community within the stream (Kaushik and Hynes 1968,
Cummins 1974). In an unpolluted Georgia stream. Nelson and
Scott (1962) discovered that primary consumers derive 66% of
their energy from allochthonous leaf material. Mann (1969)
Q
estimated that the Thames River depends on allochthonous
input for approximately half of its energy flow. Fisher and
Likens (1973) concluded that small, headwater streams in
forested areas are almost totally dependent on input of
course particulate organic matter as an energy source, such
as leaf litter, from the terrestrial ecosystem. These
studies illustrate the importance of a land-water inter-
change in providing lotie ecosystem, energy requirements.
The orientation and branching pattern of streamiS
reflect topographic and geological variations in the water-
shed. The nature of the underlying geologic formations and
surface soils in the drainage basin are important factors
affecting stream water chemistry. Due to the diversity in
lithology and solubility of rocks underlying North Carolina,
differences in regional geology will cause base flow to
contain variable amounts of dissolved constituents in the
water (Sirronons and Heath 1979). Because surface soils
interact with precipitation to produce base flow, soil is an
important component of watersheds that strongly influences
the hydrology and water chemistry of streams. The effect of
base flow influx on stream water chemistry is most pro-
nounced during low flow periods (Wharton et al. 1982).
Lotie ecosystems, and their adjacent floodplains, are
fluctuating water level ecosystems (Wharton and Brinson
1979). Low flows during summer and fall are a result of
high évapotranspiration rates whereas high flows typically
10
occur during winter and spring, when precipitation exceeds
évapotranspiration (Brinson et al. 1981b).
Nutrient cycling and sedimentation in seasonally
flooded bottomlands influence the water quality in Coastal
Plain streams and rivers. Sedimentation represents a sink
of nutrients for streams that would be carried downstream, if
the floodplain ecosystem did not provide depositional space.
Southeastern floodplain swamp forests exhibit a rapid
nutrient cycling rate which serves as a mechanism of nutri-
ent conservation. This efficient recycling component
reduces the probability that allochthonous input of nutri-
ents will be removed from the system by leaching from the
soil and export in drainage waters (Brinson et al. 1981b).
Chemical and biological interactions between streams
and their floodplains affect the quality of water exported
to estuaries (Kuenzler et al. 1977). Nutrient inflow from
upstream lotie ecosystems is essential to estuarine produc-
tivity (Odum 1971). However, excessive amounts of nitrogen
and phosphorus in drainage water can result in eutrophica-
tion in estuarine waters (Robbie 1974, Kuenzler et al.
1979). The water quality of streams and rivers, which drain
into estuaries, potentially influences the chemical, physi-
cal, and biological structure of these estuarine ecosystems.
Examination of water quality characteristics of these
Coastal Plain streams should provide a basis for the eval-
uation, management, and protection of North Carolina coastal
aquatic ecosystems.
11
Today, perhaps the most active biotic component affect-
ing the water quality of streams and rivers is man. Human
activities in the watershed significantly affect the water
chemistry of streams. Streams receive anthropogenic inputs
of domestic and industrial wastewater, agricultural runoff
loaded with fertilizer nutrients, and wastes resulting from
animal husbandry activities which increase nutrient
concentrations in streamflow and contribute to degradation
of water quality (Kuenzler et al. 1977).
One of the most recent alterations of natural lotie
ecosystems is the process of channelizetion--the widening,
deepening, and straightening of a stream channel to provide
better drainage, thereby reducing flooding of adjacent
terrestrial ecosystems. Channelization results in drastic
changes in lotie ecosystem structure and function which
alter the biology and hydrology of the natural stream.
Kuenzler et al. (1977) has demonstrated that channelized
streams differ in water quality characteristics from natural
streams. Their study showed that channelized streams are
characterized by lower color, yet higher levels of pH,
conductivity, turbidity, and also higher concentrations of
nitrate and total phosphorus, than observed in the natural
streams studied. Other studies have shown that biotic
productivity is decreased as a result of channelization due
to the destruction of habitat (Maki et al. 1980).
12
Features of this Study
The purpose of this study is to demonstrate simi-
larities and differences in water chemistry of streams in
the central Coastal Plain of North Carolina. Study sites
were chosen to reflect the diversity of stream types in the
Coastal Plain. This study examines water chemistry in a
Piedmont-draining alluvial river (and the associated flood-
plain) and in ten Coastal Plain-draining streams, which are
further subdivided into two general categories based on
their geologic origin (Table 1). These categories are
mineral soil-draining streams and peat-draining streams.
The designation mineral soil-draining streams refers to
streams primarily associated with sands, silts, and clays,
in varying amounts, throughout the watershed.
The peat-draining streams category refers to streams
with headwaters draining pocosin peat deposits; however,
further downstream, these streams drain predominately
mineral soils. This study investigates the differences in
streamwater chemistry which result from the lithological
variations in the drainage basins. Organic soils have a
number of characteristics which distinguish them from
mineral soils such as: lower bulk density (Dolman and Buol
1967, Ingram and Otte 1981), high water-retention capacity
(Dolman and Buol 1967, Bannister 1976, Ingram and Otte
1981), low hydraulic conductivity (Boelter 1970), high
organic content (Dolman and Buol 1967, Ingram and Otte
13
Table 1. Stream categories with representative sampling sites
and v/atershed areas. Stream, abbreviations are in
parentheses.
Watershed
Stream categories area (km^)
I, Coastal Plain-draining blackwater streams
A. Peat origin
1. Natural drainage
East Brice Creek (EB) 32
West Brice Creek (WB) 28
Black Swamp (BL) 78
Mill Creek (MI) 82
2. Artificial drainage
Catfish Lake drainage canal (CL) 5
B. Mineral soil origin
1. Natural
Island Creek (IS) 18
Creeping Swamp (CS) 80
Palmietto Swamp (PA) 54
Chicod Creek (CH) 132
2. Channelized
Grindle Creek (GR) 183
II. Piedmont-draining alluvial river
Tar River (TR) 7097
Tar Swamp (TS)
14
1981), a high cation exchange capacity (Puustjarvi 1956),
poor thermal conductivity (VJilliams 1970), and low ash
(mineral) content (Ingram and Otte 1981). With these
edaphic differences in mind, one might expect differences in
water chemistry of streams draining peat deposits from
streams which drain predomiinately mineral soils. An objec-
tive of this research is to determine if Coastal Plain-
draining streams can be distinguished and categorized on the
basis of the chemical composition of their waters.
Although categories are useful for data interpretation,
it should be realized that differences in biology, geology,
water chemistry, and hydrology make each stream unique.
This study attempts to elucidate relationships between water
chemistry, regional geology, and watershed land use prac-
tices in each study area.
Kuenzler et al. (1977) conducted a study to examine
water quality in Coastal Plain streams, with emphasis on
comparing water quality between natural and channelized
streams. Four sampling sites in this study were located on
the same streams used by Kuenzler and his co-workers. The
sampling sites were at the same location on Creeping Swamp
(CP-10), Palmetto Swamp (PM-10), and Chicod Creek (CH-20),
but my collection site on Grindle Creek was approximately 10
km downstream from the sampling site location used in the
Kuenzler et al. (1977) study.
The purpose of choosing the same collection sites, in
some cases, was to compare data and to expand the data file
15
of water quality characteristics and changes v/ithin these
streams. Historical reviews of water chemistry and stream
conditions are a valuable basis for detecting water quality
changes resulting from the alteration of land use patterns
in the watershed.
In addition to examining water quality differences
between natural, mineral soil-draining streams (Creeping
Swamp and Palmetto Swamp) and a channelized stream (Grindle
Creek), this study investigates water quality characteris-
tics of: a Piedmont-draining alluvial river (Tar River),
and the waters in the bordering floodplain (Tar Swamp);
streams which originate in and drain peat deposits (Brice
Creek, Mill Creek, Black Swamp); a stream in contact with
Tertiary limestone deposits (Island Creek) and the influence
of this deposit on two other streams; an artificially con-
structed drainage canal associated with an experimental
waterfowl impoundment underlain by peat (Catfish Lake
canal); and a stream undergoing transition from natural to
channelized (Chicod Creek).
Many different physical, chemical, and biological
param.eters are of significance in determining water quality
in lotie ecosystem.s. However, due to time and cost limita-
tions, it is not feasible for most water quality studies to
measure all of these parameters, so general or special
purpose parameters are chosen on the basis of research
goals. The water quality parameters chosen for my research
are examined to detect similarities and differences that
might occur in water quality of Coastal Plain-draining
streams. These include major nutrients (form.s of carbon
nitrogen, and phosphorus) and several cations, as well a
dissolved oxygen, biochemical oxygen demand, pH, total
alkalinity, and conductivity.
ÍTTUDY I^PEA
The study area for this project is located in the
central portion of the North Carolina Coastal Plain. In
addition to the Tar River, streams associated with the
Tar-Pamlico, White Oak, and Trent-Neuse River drainage
basins are included in this study. The watersheds of the
streams examined encompass portions of Pitt, Craven, Jones,
and Beaufort counties (Figure 1).
North Carolina has a warm and humid clim^ate and is
located in the temperate zone. The Piedmiont Plateau is
located in the central region of the state and is charac-
terized by clayey soils and intermediate elevations and
slopes. The central portion of the Coastal Plain has clayey
to sandy soils, low elevations, and gentle slopes (Kuenzler
et al. 1977). Sediments underlying this region are composed
of sands, clays, shell beds, limestones (Heath 1980) , and
peat (Ingram and Otte 1981) . Many elements (such as m.ajor
cations) exhibit low concentrations in the soils of the
Atlantic Coastal Plain when compared to other regions of the
United States (Shacklette and Boerngen 1984).
The climate is seasonal with cool winters and warm
summers. July and August are the warmest months with
30-year average temperatures of 26.0 and 25.5°C, respective-
ly; December and January are the coldest months with mean
temperatures of 7.1 and 6.5°C, respectively (NCAA 1973).
The mean annual temperature of the area (averaged from
18
Figure 1. Map of North Carolina showing the location of the
study areas. The streams examined in this study
and their watershed areas are shaded. The
Catfish Lake station is not illustrated.
19
1941-1970) is 16°C (NOAA 1973). This region receives
abundant rainfall, which has approximately even distribution
throughout the year (Sumsion 1970, Kuenzler et al. 1977).
The mean annual precipitation for the central Coastal Plain,
averaged for 30 years, is 129 cm (NOAA 1973).
The study area is a region of flat to gently undulating
topography which is characterized by low gradient, slow
flowing streams bordered by seasonally flooded bottomlands.
Sediments in the stream bottoms consist of varying amounts
of sand, silt, and clay. These sediments are sometimes
overlain by organic matter, especially in the summer months
when discharge rates are low and no flow periods occur (Maki
et al. 1980).
Coastal Plain-draining streams are closely associated
with wetlands. Peat-draining streams originate in pocosin
wetlands and, as they increase in size downstream, pass
through floodplain wetlands similar to those bordering
mineral soil-draining streams.
Pocosin wetlands may have peat deposits up to 3 m in
thickness. The low nutrient status in pocosins is due to
isolation of the vegetation from mineral soils by peat.
Pinus serótina (pond pine), the only tree species found in
low pocosin vegetational systems, exhibits a deformed and
stunted growth form. These structural characteristics may
be due to a combination of environmental stresses including
long hydroperiods, acidic soils, low nutrient availability,
and periodic fires (Christensen et al. 1981, Otte 1981).
20
The density and height of serótina and shrub species
decreases from the outer margins, where pond pine woodland
and high pocosin occur, to the center of the pocosin, which
is underlain by the thickest peats and dominated by low
pocosin vegetation (Otte 1981) .
Ingram and Otte (1981) reported the following species
were observed at almost every one of their sampling lo-
cations in the Croatan National Forest pocosins: Pinus
serótina, Lyonia lucida, Zenobia pulverulenta, Gordonia
lasianthus, Persea borbonia, Smilax spp., Woodwardia
virginica, Cyrilla racemiflora, and Ilex glabra. Other
frequently encountered species were Magnolia virginiana,
Kalmia augustifolia, Cassandra calyculata, Sorbus arbu-
tifolia, and Vaccinium spp. (Ingram and Otte 1981). An
extensive moss network forms a spongy ground cover through-
out the pocosin; the most comm.on moss encountered is Sphag-
num (Christensen et al. 1981). For a more detailed descrip-
tion of the vegetational composition of the Croatan National
Forest pocosins, see Snyder (1977) and Otte (1981).
Along mineral soil-draining streams, Nyssa aquatica and
Taxodium distichum are common riparian species found in the
wettest habitats in floodplains. Species diversity is
higher in more elevated portions of floodplains and in
upland plant communities within these watersheds (Kuens^ler
et al. 1977). Common floodplain species are Acer rubrum, N.
sylvatica var. biflora, Liquidambar styraciflua, and Quercus
michauxii. These species also occur in the upland regions
21
of the watershed but are much less prevalent. Other
riparian vegetation commonly encountered in the central
region of the North Carolina Coastal Plain are Betula nigra,
Fraxinus caroliniana, Q. nigra, Carya aquatica, Populus
heterophylla, Ulmus americana, Carpinus caroliniana, Ilex
opaca, and Persea borbonia. For a more detailed description
of southern riparian vegetation, see Brinson et al. (1981b)
and Wharton et al. (1982).
The Coastal Plain-draining streams in this study are
located in Geochemical Zone V, and are described below
according to information in Simmons and Heath (1979). The
Coastal Plain region is underlain by sedimentary rocks that
create a complex series of geologically distinct formations.
Five geological units are found in the upper sediments of
the Coastal Plain (arranged from youngest to oldest) are:
(1) Quaternary deposits--dominate in the exposed sediment
and comiposed of stream-deposited beds of yellow (or tan)
sandy silt and clay; (2) Yorktown Formation--of Pliocene Age
composed of marine-deposited clay, silt, and sand inter-
spersed with a few, thin shell beds; (3) Pungo River Forma-
tion--of Miocene age, which consists of phosphatic sand
interbedded with silt and calcareous clay; (4) Castle Ilayne
Limestone--of Eocene age which is composed of fossiliforous
limestone, containing a few calcareous sand bed layers
(Simmons and Heath 1979, Winner and Simmons 1977, Heath
1980); and (5) Cretaceous sediments--composed of a complex
miixture of fluvial estuarine and m^arine sand and mud (L.
22
Otte, personal communication, 1984). These formations
intersect with groundwater in a discontinuous manner over
Geochemical Zone V, which results in variation in the water
chemistry (especially dissolved solids) of streams in this
region (Simmons and Heath 1979). These surficial aquifers
provide base runoff to streams in this study. The upper 30
m of these sediments are hydrologically important to lotie
ecosystem dynamics, since most of the groundwater circu-
lation is restricted to this depth (Simmons and Aldridge
1980). The limestone and shell beds, associated with these
formations, are water soluble and thereby contribute dis-
solved solids to streams in the area (Simmons and Heath
1979). The amount of dissolved minerals in streamwater is a
function of their relative abundance in the watershed and
their solubility.
Ten streams, a drainage canal, and a seasonally flooded
bottomland were studied, with all but one (Tar River)
confined to the central portion of the North Carolina
Coastal Plain (Figure 1) . The stream.s and their watershed
areas are listed in Table 1.
Coastal Plain-draining Streams
The streams examined in this study are collectively
categorized as Coastal Plain-draining streams and are
blackwater streams originating in the central region of the
North Carolina Coastal Plain. The name "blackwater stream"
23
originates from the characteristically dark brown-colored
v/ater found in these streams. Prolonged contact with
organic matter in swamp forests adjacent to these streams
results in leaching of humic and fulvic compounds which
darkly stain the drainage water (Beck et al. 1974, Gjessing
1976). These streams are precipitation-dominated (sensu
Wharton et al. 1982); however, base flow is the primary
source of water to these streams during summer and fall.
Surface runoff, along with base flow, becom.es hydrologically
important to stream water chemistry during the winter and
spring when low évapotranspiration rates result in soil
saturation and hence, increased overland runoff of local
precipitation (Brinson et al. 1981a). Although both base
flow and overland runoff originate from precipitation, base
flow typically has a higher mineral content than surface
runoff waters due to longer contact with mineral soils
(Simmons and Heath 1979). When compared to alluvial rivers,
undisturbed blackwater streams usually have narrower flood-
plains and transport louver sedim.ent loads (Wharton and
Brinson 1979).
Peat-draining streams
The peat-draining streams in this study have headwaters
originating in the peat deposits of the Croatan National
Forest which is located south of New Bern, North Carolina
(Figure 1). The Croatan National Forest contains the third
24
largest peat deposit in North Carolina (Ingram and Otte,
1981). A large portion of the Croatan is covered by pocosin
peat, although mineral soil is exposed in many areas. Peat
is defined as any partially decomposed plant material that
has accumulated, over time, in water-saturated soil where
plant growth and deposition have exceeded the rate of
decomposition.
The designation, peat-draining stream, refers to the
fact that pocosin peats exist in the stream drainage basin;
however, it does not imply that peat is the only substrate
in contact with the stream. The stream water passes through
a gradient of edaphic conditions ranging from predominately
peat, to mixed organic-mineral soils, and gradually to
mineral-rich soils as it travels from the stream, headwaters
to my collection sites.
Two sampling sites were chosen on Brice Creek (Figure
1), a natural, peat-draining stream whose headwaters drain
areas containing a maximum of 1.8 m of peat thickness
(Ingram and Otte 1981). Brice Creek is located in the
northern portion of the Croatan National Forest and drains
northv/ard into the Trent River, which flows into the Neuse
River estuary.
The sampling site on East Brice Creek (EB) is located
on Catfish Lake Road (SSR 1100) and has a watershed area of
32 km2. The sampling site at West Brice Creek (WB) is
located on Little Road (FSR 121-2) and has a watershed area
of 28 km^. Disturbances resulting from forestry practices
25
(i.e., clearcutting, controlled burns, pine monoculture)
were observed in the watershed in close proximity to the WB
sampling station. The vegetation surrounding the collection
site is predominately pond pine forest or pond pine wood-
land. For a description of these plant communities, the
reader is referred to Otte (1981).
The sampling stations at Mill Creek (MI) and Black
Swamp (BL) are located on the western border of the Croatan
National Forest. The headwaters of Mill Creek and Black
Swamp contact regions underlain by 0.6 m of peat (Ingram^ and
Otte 1981). The water chem.istry of these streams is also
influenced by an underlying Tertiary limestone deposit.
The Mill Creek station is located on SSR 1004, on the
edge of Pollocksville's city limit. Mill Creek drains into
the Trent River and has a watershed area of 82 km^. in
addition to agricultural runoff, this stream receives
pollution from a hog farm a short distance upstream from the
sampling site.
The sampling station at Black Swamp is located on State
Highway 58, south of Maysville. Black Swamp has a watershed
area of 78 km^ and drains into the White Oak River, which
flows into the Bogue Sound estuary.
The Catfish Lake (CL) sampling station is located in
the northwestern part of the Croatan National Forest, just
northwest of Catfish Lake. The sampling site is an artifi-
cial drainage canal, associated with a waterfowl impound-
ment, underlain by 1.2 m of peat (Ingram and Otte 1981).
The watershed area is 5 km^, which is much smaller than any
of the other study areas.
Water level in the impoundment is controlled by a
floodgate constructed in the drainage canal. This device
allows artificial manipulation of the water level in the
imipoundment. When closed, the floodgate acts as a barrier
obstructing water outflow. Opening the floodgate would
result in slow drainage of the impoundment area.
Water samples were collected at a location in the
drainage canal approximately 10 m upstream from the flood-
gate. Because discharge in this canal is artificially con-
trolled by the floodgate, no flow data was collected for
this sampling site. During most of the study year, the
floodgate remained closed; therefore, the water in the
drainage canal was usually in a stagnant condition.
The pre-impoundment vegetation of the Catfish Lake
watershed was typical low pocosin growing on waterlogged,
deep peats (L. Otte, personal communication, 1983).
Approximately 20 years ago, the area was cleared of its
pocosin vegetation and burned in an attempt to put the land
into experimental agricultural production. Later, the
farming attempt was abandoned and, at least 10 years ago,
the area was planted with vegetation suitable for v;aterfowl
consumption and flooded to provide an experimental waterfowl
impoundment area. The aquatics v/hich were planted have not
prospered in the acidic impoundment water. The im.poundment
27
area has remained inundated for several years, yet has not
been a successful waterfowl habitat.
Mineral soil-draining streams
Island Creek (IS) is a mdneral-soil draining stream
located in Jones County, east of Pollocksvilie. The samp-
ling site is located on SSR 1004. The watershed is only 18
km^ and is relatively undisturbed above the sampling site.
The stream, flows northward into the Trent River, which flows
into the Neuse River.
Island Creek differs from the other streams in this
study because the stream channel dissects a Tertiary lime-
stone formation of Eocene or Oligocčne age (L. Otte, person-
al communication, 1984). Limestone outcrops (often covered
with liverworts and other bryophytes) are exposed along the
stream banks. These outcrops are visible during low flow
periods, but are sometimes obscured during high flov; stages.
Rare calciphytes occur on the calcareous outcrops (Sears
1966) . The erosive action of flowing water along the stream,
banks results in gradual disintegration of minerals in the
limestone outcrop.
Creeping Swamp (CS) and Palmetto Swamp (PA) represent
natural, mineral soil-draining streams, which are bordered
by seasonally flooded bottomland forests. The two streams
flow southwesterly into Swift Creek, which flows southward
into the Neuse River drainage basin. The sampling stations
28
are located on State Highway 43, southeast of Calico. Both
streams are gauged by the U.S. Geological Survey. The
watershed areas are similar in size, with that of Creeping
Sv/amp (80 km^) being slightly larger than that of Palmetto
Swamp (54 km^) (Kuenzler et al. 1977).
Creeping Swamp and Palmetto Swamp drain watersheds that
are over 60% wooded (Kuenzler et al. 1977). Pine plan-
tations in various stages of development, and mixed hard-
woods in the uplands and floodplain, comprise the woodlands
in the watershed (Kuenzler et al. 1977). For a detailed
description of the vegetation in the Creeping Swamp and
Palmetto Swamp watersheds, see Kuenzler et al. (1977). The
watersheds contain no point sources of pollution, with the
exception of a small tributary entering Creeping Swamp less
than 2 km upstream from my collection site, which receives
effluent from a hog farm (Kuenzler et al. 1977).
Winner and Simmons (1977) prepared an annual water
budget for the Creeping Swamp watershed. In this model,
precipitation provides 100% of the water inflov/. Base
runoff (20%), overland runoff (17%), and groundwater outflow
through deep aquifers (2%) together constitute less than
half of the total water outflow pathway, while evapotrans-
piration is responsible for 61% of the water output.
Chicod Creek (CH) is a mineral soil-draining stream
which flov/s northward into the Tar River. The sampling
station on Chicod Creek was located on Secondary Road 1760.
The watershed area above the sampling site is 132 km^
(Kuenzler et al. 1977).
Chicod Creek is a U.S. Geological Survey gauged stream,
which is equipped with a continuous stage recorder, a
continuous conductance and temperature recorder, and an
automatic sediment sampler. Base flow is the primary source
of stream water during dry conditions (Simmons and Aldridge
1980).
The Chicod Creek watershed is sparsely populated and
50% of the region is forested by mixed pine-hardwood commu-
nities. Approximately 43% of the watershed is used for
cropland and pastures, and the remaining 5% of land use is
occupied by residential areas, roadways, and waterways. A
reconnaissance of the watershed, conducted in 1978, revealed
that poultry and swine farms are the most prevalent animal
husbandry activities in the watershed. Sources of stream
pollution are from several large poultry farms (each with
approximately 80,000 chickens) which are located near the
stream, and from direct outlets to the stream from holding
ponds adjacent to animal shelters. Buffer zones of forest
separate the stream* and farmland in most parts of the
watershed. In addition to the above information, Simmons
and Aldridge (1980) provide a geologic overview of Chicod
Creek.
The sampling site at Chicod Creek provided an oppor-
tunity to study a stream in transition from a semi-natural
system to one undergoing channelization. Local farmers
30
requested help from the U.S, Department of Agriculture in
1963 to solve flooding and drainage problems (Coffey 1982).
Channelization of Chicod Creek was proposed by the Soil
Conservation Service (SCS) to provide better drainage and
thereby reduce flooding of cropland, and roads, in the
watershed.
Phase I of the project began in November 1978, and
consisted of obstruction removal (clearing and snagging) in
the downstream reaches of Chicod Creek (Coffey, 1982). This
phase of the project was completed in January 1979. Because
my sampling site was located upstream from the obstruction
removal project, the disturbance probably did not affect the
water quality of my samples. From February 1979 through
June 1979 all drainage improvement efforts ceased in an
attempt to avoid interference with fish spawning patterns
(A. Coffey, personal communication). Phase II of the
channelization project began in July 1979, and involved
channel excavation using draglines and hydraulic hoes (A.
Coffey, personal communication, 1984). This phase of the
project was conducted approximately 4 km upstream from my
sampling site. Therefore, this disturbance in the v/atershed
and stream channel potentially could affect the water
quality characteristics of the stream water at my collection
site from July through August 1979.
Grindle Creek (GR) is a mineral soil-draining stream
which was channelized in 1966. The watershed area is 183
km2. The sampling site is on U.S. Highway 264, near the
31
small town of Pactolus. Grindle Creek flows southward into
the Tar River. Base flow results in a continuous discharge
throughout the year (Kuenzler et al. 1977). High banks, a
clean, sandy bottom, and relatively fast flowing waters
distinguish Grindle Creek from natural Coastal Plain-
draining streams.
Since channelization nearly 20 years ago, woody vege-
tation has become re-established along the stream banks.
Approximately 59% of the watershed is wooded, while the
remaining 41% is cleared (Kuenzler et al. 1977). Agricul-
ture and animal husbandry activities are prevalent in the
watershed. The town of Bethel, located in the headwaters of
Grindle Creek, is a point source of sewage effluent.
Alluvial river and swamp
Tar River, a Piedmont-draining alluvial river, is the
only stream in this study that does not originate in the
Coastal Plain. The inorganic chemical composition of
alluvial rivers is affected by weathering and leaching of
parent rocks and soil in the Piedmont (Wharton et al. 1982).
Alluvial rivers typically have periods of sustained high
flow, following heavy precipitation, which results from the
cumulative effect of contributing tributaries (Wharton et
al. 1982). In the low elevation of the Coastal Plain,
alluvial rivers tend to form large floodplains which are
seasonally inundated (Wharton et al. 1982). A prolonged
32
hydroperiod occurs during the winter, which results more
from reduced évapotranspiration rates than higher precipita-
tion (Wharton and Brinson 1979, Brinson et al, 1981a).
Tar River is affected by Geochemical Zones I and II
described by Simmons and Heath (1979). Zone I, at the
headwaters of Tar River, is underlain by a large, complex
series of granite rocks and a smaller area of rocks in the
diorite group. Zone II is underlain by metamorphosed
volcanic and sedimentary rocks and by a combination of
cemented conglomerates, sandstones, siltstones, and shales.
This zone is characterized by relatively im.permeable clayey
surface soil which restricts percolation of rainfall through
the soil, thereby inhibiting aquifer recharge. The presence
of clayey soils results in a large amount of overland
runoff, and the easily eroded soils contribute to the high
concentrations of suspended sediments in the Tar River.
The area of the Tar River watershed above the sampling
site on SSR 1565 is 7097 km^. Samples were collected in the
middle of the channel which was accessible from a catwalk
extending from the main bridge crossing the river. Chicod
Creek and Grindle Creek are tributaries of the Tar River
that join it less than 2 km upstream from the samipling site.
Tar River flows eastward into the Pamlico River estuary,
which is located approximately 15 km downstream from my
sampling site.
Tar Swamp is an alluvial, tupelo-cypress swamp forest
that is part of the Tar River floodplain. It was sampled
33
when standing water was present in order to gain information
on possible consequences of floodplain-stream exchanges on
water chemistry. The sampling site in Tar Swam.p is located
150 m from the north bank of Tar River where the river was
sampled. The dominant canopy trees of Tar Swamp are Nyssa
aquatica. Scattered Taxodium distichum occur throughout the
upper canopy and Fraxinus caroliniana dominates the under-
story. Trees are relatively uniform in size as a result of
clearcutting approximately 30 years ago. This area has been
described in more detail by Brinson (1977).
Tar Swamp has an aquic moisture regime; however, water
levels in the swamp vary seasonally. During the winter and
spring, water levels above the sediment surface may reach as
high as 1.3 m when Tar River rises above its natural levee
during flood stages, causing water to flow onto the flood-
plain. During summer and early fall, water levels m.ay drop
to 10 cm or more below ground surface (Brinson et al.
1981a). During this annual drydown phase, a few centimeters
of standing water may occur in the swamp for short periods
following heavy, local rainfall, but it is usually removed
within a few days as a result of high évapotranspiration
rates.
Sources of Tar Swamp water include: (1) precipitation,
(2) overbank flooding from the Tar River, (3) groundwater
seepage, and (4) surface runoff from terrestrial ecosystems.
Overbank flooding of the Tar River results in sustained
hydroperiods in Tar Swamp during winter and spring when
34
évapotranspiration rates are low. Below the leaf litter
layer, the principal components of the surface sediments in
Tar Swamp are silts and clays v/hich have settled out from
the floodwaters of Tar River. Because of the low elevation,
wind tides in the Pamlico River may also affect inundation
and floodplain dynam.ics.
METHODS
Field Measurements, Collections, and Sample Preparations
Eleven streams were chosen to represent the diversity
of stream water quality in the North Carolina Coastal Plain.
The collection of field data and water samples for chemical
analyses was designed to elucidate differences and simi-
larities in water chemistry among streams and to determ.ine
if seasonal patterns exist. One site was chosen for each
stream for biweekly collections of water samples. Data were
collected for 1 year (September 1978 through August 1979).
All of the sampling sites were located near bridges for
accessibility. Sampling site, date, time, weather con-
ditions, sample bottle numbers, field measurements, and
notes concerning stream and watershed conditions were
recorded.
Field measurements included water temperature, water
level (stage height), and flow velocity. Originally, pH
measurements were to be taken in the field; however, due to
the questionable accuracy of the field pH meter, laboratory
pH readings were made instead, promptly after returning from
the field.
Water temperatures were determined by submerging a
thermometer into a water sample. Water levels (stage
heights) were measured by direct reading of U.S.Geological
Survey staff gauges where present. Water levels for the
36
ungauged streams were determined by measuring the vertical
distance, from a reference point on the bridge, to the water
surface, using a field meter tape with a metal v/eight at-
tached to the end.
On each collection date, instantaneous discharge for
the gauged streams was estimated by applying stage height
data to U.S. Geological Survey rating curves. Tar River was
gauged at Tarboro, which is upstream from the sampling site
chosen for this study. The average daily discharge for Tar
River on the sample dates had to be corrected for the
additional watershed area involved. This was accomplished
by dividing the total drainage area of Tar River above
Grimesland (7,097 km^) by the drainage area above Tarboro
(5,540 km2) to produce a ratio (1.28). This value can be
used to calculate Tar River discharge at Grimesland by
multiplying it by the discharge at Tarboro (Barry Adam.s,
personal communication, 1979). This method assumées that the
area between the sites v/ill yield the same amount of water
as the area above Tarboro.
Instantaneous discharge for the ungauged streams was
calculated from the stream cross-sectional area and flow
velocity measurements made during each sampling interval. A
stream profile was constructed for each ungauged streami by
measuring the depth, at set intervals, across the width of
the stream and by measuring the stage height to determine
the cross-sectional area. Stream velocity was measured by
using a stopwatch to time the passage of dye (sodium
fluorescein) across a standardized distance (5 m or the
width of the bridge) in relatively uniform reaches, either
upstream, downstream, or underneath the bridge. Stream
velocity measuremcents were duplicated in the center, and the
right and left sides of the stream. These velocities were
averaged and then used to calculate the discharge rate. The
difference in stage height, multiplied by the stream width,
was appropriately added to, or subtracted from, the original
cross-sectional area. Discharge (Q) was then calculated by
the formula:
Q = A X V
where Q is the instantaneous discharge on a given sampling
date (m^/s) , A is the total cross-sectional area (m.^) , and V
is the mean velocity of streamwater (m/s). Discharge rates
were used to calculate the weighted mean concentrations of
organic and inorganic nutrients, as well as other aqueous
parameters.
Samples were collected for dissolved oxygen analyses
with a van Dorn sampler and transferred to 300 ml BOD
bottles. These samples were fixed in the field, returned to
the lab, and analyzed by the azide modification of the
Winkler method (Golterman and Clymo 1969) within 24 h. In
the field, duplicate samples were collected in the van Dorn
sampler and stored in 300 ml BOD bottles to determine
biochemical oxygen dem.and (BOD^). Samples were aerated to
saturation level using the laboratory air terminals. The
oxygen concentration was measured with a YSI Model 51B
38
dissolved oxygen meter with a YSI Model 5720 bottle probe
prior to incubation. The BOD bottles were capped and placed
in a Fisher Model 300 low temperature incubator at 20°C for
5 days (APHA 1975). After the incubation period, the oxygen
concentration was measured to determine oxygen uptake by
microbial activity within the 5-day period.
Water samples for laboratory analyses were collected in
4-liter polyethylene bottles that were prepared for use by
acid-washing in a 50% HCl solution and then thoroughly
rinsing them in distilled water. In the field, each sample
bottle V7as rinsed with stream water prior to collecting the
sample. Samples were immediately placed in an ice-packed
cooler for transport back to the lab that day. The samples
were transferred from the ice chest to a refrigerator set at
4-5°C upon returning to the lab.
The water samples were prepared for chemical analyses
within 12-48 h of collection time (Figure 2). All polyethy-
lene bottles used for storage of subsamples were acid-washed
in 50% HCl solution and then thoroughly rinsed with dis-
tilled water. Bottles for use in cation analyses were
soaked in a 2% HNO^ wash for 24 h, followed by a thorough
rinsing with deionized water (APHA 1975).
Samples v/ere stored in labeled bottles to await chem.i-
cal analyses. Each sample v/as shaken before pouring, to
insure homogeneity. For filtering water samples, Gelman
magnetic filter funnels were used and rinsed thoroughly with
distilled water after each use. Gelman glass fiber filters
39
Field sample (4 liters)
Total alkalinity and pH (100 ml)
Conductivity (250 ml)
FILTERED UNFILTERED
K jeldahl- Nitrogen and Total phosphorus Total organic
nitrogen phosphorus I carbon
I 1. NH -N (250 ml)
(500 ml) 2. NO^-N refrigerate (50 ml)
freeze 3. NO^-N acidify with
I 2 drops HCl-
Particulate 4. FPP refrigerate
organic 5. FTP
nitrogen
I (250 ml)
(filters) refrigerate
freeze
Dissolved Cations
organic carbon 1. Ca
I 2. Mg
(50 ml) 3 . K
acidify with 4. Na
2 drops HCl- 5 . Fe
re frigerate 1
(50 ml)
Acidify to 1% HNO-
and 2.5% HCl (v/v)^-
freeze
Figure 2. Preparation of field samples for chemical analyses.
40
(type A/E, 47 nun, nominal pore size 0.3 urn) were used for
particulate organic nitrogen analyses. They were prepared
for use by placing the filter on the funnel and rinsing with
5 ml of 5% HCl, followed by addition of 150 ml deionized
water to the filtering apparatus (to thoroughly rinse the
filter). After the filters were prepared by this procedure,
500 ml of the sample were filtered. The filter was removed
with forceps, placed in a plastic container, and frozen
until analysis.
Total alkalinity and pH were measured with a Corning
Model 101 Digital Electrometer equipped with pH and refer-
ence electrodes (APHA 1971) . Conductivity v/as measured by a
Beckman RC-16C Conductivity Bridge, with a 0.1 constant
cell, at 25°C. These measurements were made within 4-18 h
after returning to the lab.
Two rain gauge stations were constructed to collect and
measure bulk precipitation over the study year. One rain
gauge station v/as located in Tar Swam.p and the other one in
the Croatan National Forest. Both gauges were placed in
clearings where no tree canopy interfered with rainfall
R R
collection. A Taylor Clear-Vu rain gauge, with a Tenite
butyrate calibrated cylinder for measuring rainfall volum.e,
was attached to a post. A large plastic trash can was hung
on the opposite side of the post to collect a rainfall
sample for chemical analyses. The trash can was lined with
a polyethylene bag, with several drops of saturated HgCl^
solution added to inhibit microbial growth. The lid of the
41
trash can was placed upside down and secured by handles to
the can. To collect rainfall, a hole was cut through the
center of the lid and stuffed with Pyrex wool (which was
acid-washed in 50% HCl solution then rinsed with deionized
water) which served as a filter to prevent leaf litter and
other debris from entering the rainfall sample. Bulk
precipitation was collected and rainfall volume was measured
at biweekly intervals which corresponded with the stream
sampling schedule. The rainfall sample was transferred from
the collection bag to a clean sample bottle. The collection
bag and glass wool were replaced each sampling time.
Rainfall samples were treated in the same manner as the
stream samples (i.e., packed in ice for transport back to
the lab) and selected chemical analyses followed the same
procedures as used for the stream waters.
Chemical Analyses
Chemical analyses were performed in the Central En-
vironmental Laboratory of the Department of Biology at East
Carolina University. The nutrient analyses consisted of
nitrate-nitrogen, ammonium-nitrogen, nitrite-nitrogen,
dissolved Kjeldahl-nitrogen, dissolved organic nitrogen,
particulate organic nitrogen, total phosphorus, filterable
total phosphorus, filterable reactive phosphorus, particu-
late phosphorus, total organic carbon, dissolved organic
carbon and particulate organic carbon. The cation analyses
42
consisted of calcium, magnesium, sodium, potassium, and
iron. Dissolved organic nitrogen concentrations were
calculated by subtracting ammonium from dissolved Kjel-
dahl-N. Particulate phosphorus levels were calculated by
subtracting filterable total phosphorus from total phosphor-
us.
Total (TOC) and dissolved organic carbon (DOC) analyses
v/ere performed by combustion and infra-red analyses using a
Beckman Model 915 Total Organic Carbon Analyzer (Beckman
Instruments, Inc. 1977). Due to the inaccuracy associated
with derivation of particulate organic carbon (POC) by
subtraction of DOC from TOC, POC values were determined on a
group of samples (collected in November 1979) by dichromate
oxidation. For this collection, analysis of POC from
1-liter samples, from representative streams, was used to
determine POC concentrations for other collection dates by
extrapolation from PON-POC relationships (Banse 1974).
Strickland and Parsons (1979) served as a reference for
reagent and filter preparations and for glassware cleaning
procedures. Sample treatment followed the method outlined
by Wetzel and Likens (1979). POC was determined by "wet
ashing" particulate matter filtered onto a glass filter with
a mixture of potassium dichromate and concentrated sulfuric
acid, and measuring spectrophotometrically the decrease in
the extinction of the yellow dichromate solution after it
had been reduced by organic m.atter. The empirical relation-
ship between POC and PON from replicate subsamples, for all
43
stations except Tar River, was described by the equation POC
= 12.117 PON - 0.075 (r=0.82). POC concentrations from Tar
River were estimated independently because phytoplankton may
have contributed significant amounts of particulate matter
to the POC pool, since the Tar River channel is wider and
more illuminated than the other streams. The empirical
POC/PON ratio of the TR sample was 7.2.
Ammonium was measured by the indophenol spectrophoto-
metric method (Scheiner 1976) with a Bausch and Lomb Spec-
tronic 88 Spectrophotometer. Dissolved Kjeldahl-nitrogen
and particulate organic nitrogen were determined by mercuric
sulfate digestion-distillation (APHA 1971) using a Labconco
12 unit digestion-distillation apparatus followed by indo-
phenol spectrophotometric analysis, using a Turner Model 350
spectrophotometer. Nitrate-nitrogen determination was m.ade
by ultraviolet light absorption (Brown and Bellinger 1977)
using a Coleman-Hitachi Model 124 Spectrophotometer.
Nitrite-nitrogen was measured by a diazotization spectropho-
tom.etric method (EPA 1976) using a Bausch and Lomb Spec-
tronic 88 Spectrophotometer. This instrument was also used
for molybdate spectrophotometer analysis of phosphorus (EPA
1976). Total phosphorus and total dissolved phosphorus
received persulfate digestion prior to analysis.
Unfiltered samples for cation analyses were acidified
to 1% HNO^ and 2.5% HCl (to facilitate solution of cations
and serve as a preservative) and then frozen until time of
analysis. Samples were autoclaved before determining cation
44
concentrations by atomic absorption spectrophotometry
(Perkin-Elmer Corporation 1976) , using a Perkin-Elmer Model
305B Atomic Absorption Spectrophotometer.
Data Processing and Analysis
Annual weighted mean concentrations for organic and
inorganic nutrients were calculated for all stations where
discharge measurements were available. For a given year and
sampling station, the annual weighted mean concentration of
a given species was weighted on the basis of discharge in
the following manner:
Wt. Cone. = Y, [Cone. (Date X) x Discharge (Date X) ]
E Discharge
where Wt. Cone, represents the weighted concentration of the
constituent during the sampling period. Cone. (Date X) is
the measured concentration of the constituent on a given
sampling date (Date X), Discharge (Date X) is the instanta-
neous discharge on a given sampling date (Date X), and E
Discharge is the sum of all the discharge values on all
sampling dates for the given year and station. Annual
inputs of chemicals in rainfall were calculated using
weighted mean average of precipitation. Annual rainfall
volumes for 1978 and 1979 were obtained from. NOAA (1978,
1979) and were averages of values measured at Greenville and
Viashington, North Carolina, for sampling stations close to
these cities, and averages at New Bern and Maysville, North
Carolina, for sampling stations near these cities. Annual
45
volume of precipitation per square kilometer was multiplied
by concentrations to obtain the annual area based inputs of
chemicals delivered by bulk precipitation.
Linear correlations (r) between all of the physical and
chemical factors, measured at each station between 7 Septem-
ber 1978 and 22 August 1979 were calculated by means of a
SAS program using the CORR procedure (Barr and Goodnight
1972). This produces product-moment correlation coeffi-
cients which were used to elucidate interrelationships among
stream constituents representing random, bivariate normal
distributions. Computation of correlation coefficients by
the CORR procedure of a SAS program produces the product-
moment correlation coefficient, its significance probabil-
ity, and the number of observations contributing to the
correlation coefficient (Barr and Goodnight 1972). Corre-
lation coefficients are reported in this paper only if the
significance probability is less than or equal to 0.002
(Kuenzler et al. 1977).
Discharge measurements were not taken frequently enough
to accurately estimate annual runoff. Therefore, runoff at
each of the sampling sites was determined from U.S. Geolo-
gical Survey (1979, 1980) for the gauged streams, estimated
from U.S.Geological Survey gauging stations in proximal
streams, or extrapolated from monthly relationships between
precipitation and runoff. Creeping Swamp and Chicod Creek
had gauging stations at the sampling site. Because the
watershed of Palmetto Swamp is adjacent to that of Creeping
46
Swamp, equivalent runoff was assumed. However, because I
used the Kuenzler et al. (1977) estimate of watershed size
of 80 km^ rather than the 70 km^ assumed by U.S. Geological
Survey, the annual runoff from U.S. Geological Survey was
multiplied by a factor of 0.875 (70/80). Winner and Simjnons
(1977) reported that annual runoff is not altered by chan-
nelization--therefore, runoff for Grindle Creek (chan-
nelized) and Chicod Creek (natural) was assumed to be equal
since they were in relatively close proximity to each other.
Runoff for peat-draining streams was estimated on a
monthly basis as a percentage of precipitation that occurred
at weather stations closest to the study sites. The per-
centages were calculated from precipitation-runoff relation-
ships derived from the average of Greenville and Washington,
North Carolina precipitation data (NOAA 1978, 1979) and from
U.S. Geological Survey mean runoff values for Creeping Swamp
and two tributaries of Chicod Creek. These monthly runoff
percentages of precipitation were applied to the monthly
precipitation averaged for New Bern and Maysville, North
Carolina to calculate runoff for peat-draining streams.
Annual export of selected chemical constituents of
streamwater was calculated by the formula;
E = R X W
where: E is the annual export of a specific chemical
parameter (g/m^’yr), R is the annual runoff rate for the
watershed (m/yr), and W (g/m^) is the discharge weighted
mean of the specified chemical parameter.
RESULTS
The figures in this section are histograms that show
both the annual arithmetic mean and the weighted mean of
water chemistry analyses for the study year (Figures 3-11).
Comparison of arithmetic and weighted means is used as an
interpretive tool in this section to elucidate the relation-
ship between elemental concentrations and flow regime. For
example, when the weighted mean exceeds the arithmetic mean,
this is interpreted to suggest higher concentrations during
high flows. The limitations of this approach must be
realized, especially when the differences between the two
means is small. When differences are small, a relationship
between elemental concentration and flow pattern may or may
not actually exist. Correlation coefficients are used to
further determine the reliability of this relationship.
However, many other factors beside flow may be involved in
concentration regulation.
Unless otherwise indicated, arithmetic means will be
used for comparing stream categories. The arithmetic m.ean
is preferred because it represents the average of the actual
biweekly m.easurements at the sampling site and because its
standard error can be used as an index of variation. The
weighted mean is used later to estimate exports from, the
watershed. Concentrations are referred to as "intermediate"
when they tend to fall between the highest and lowest group
of concentrations on the figures.
48
Stream abbreviations used in this section are listed in
Table 1. The streams are arranged vertically on the histo-
gram following the categories outlined in Table 1. Peat-
draining streams appear first, following by mineral soil-
draining streams, and lastly the Piedmont-draining alluvial
river and swamp.
Dissolved Oxygen and Biochemical Oxygen Demand
Mean dissolved oxygen (0^) levels for the streams were
very similar, with no distinguishing pattern evident betv/een
peat-draining streams and mineral soil-draining streams
(Figure 3a, Appendix Table 2). Peat-draining CL, WB, and EB
had low 0^ concentrations ranging between 5.6 and 6.7 mg/1.
MI and BL, in the peat-draining category, had intermediate
0^ concentrations of 7.1 and 7.3 mg/1, respectively.
Mineral soil-draining CS, PA, and CH had low 0^ concen-
trations between 5.6 and 6.3 mg/1. However, mineral soil-
draining IS and GR (channelized) exhibited the highest
concentrations of any of the Coastal Plain-draining streams
studied, with 8.4 and 9.0 mg/1, respectively. Piedmont-
draining TR also had a high concentration of 8.5 mg/1
which sharply contrasted with the extremely low 0^ level
observed in TS (3.5 mg/1). TS had a lower average annual 0^
content than any of the streams examined.
Weighted mean 0^ levels were generally higher than the
arithmetic means for most of the streams sampled (Figure 3a,
O2 (mg/liter) BOD5 (mg/liter)
o 2 4 6 8 10
Figure 3. Annual arithmetic means (shaded) and weighted means (solid) for (a) dissolved
oxygen (0„) and (b) biochemical oxygen demand (BODj.) from September 1978
through August 1979.
50
Appendix Table 2). This would suggest that 0^ concen-
trations in natural Coastal Plain-draining streams tend to
be highest during high flows. However, positive corre-
lations between discharge (Q) and 0^ concentrations were not
significant in any of the streams. Channelized GR had a
higher arithmetic mean than weighted mean and a negative
correlation between Q and O2 concentrations.
The 0^ content of streamwater is negatively correlated
with temperature in most of the Coastal Plain-draining
streams examined in this study (Appendix Tables 5-12).
However, temperature has only an indirect influence on the
oxygen content of streamwater. Reid and Wood (1976) identi-
fy other factors contributing to the status of stream-
water as: (1) the amount of physical aeration created by
turbulent flow, (2) the photosynthesis-respiration balance
of the stream biotic community, and (3) oxygen-consuming
chemical reactions within the stream.
Peat-draining EB, WB, BL, and MI exhibited the lowest
biochemical oxygen demand (BOD^) along with mineral soil-
draining IS (Figure 3b, Appendix Table 2). The BOD^ in
these streams ranged between 0.60 and 1.1 mg/1. CL, the
pocosin drainage canal, had a high to intermediate average
BOD^ of 1.8 mg/1. Two of the mineral soil-draining streams,
GR (channelized) and PA, also had intermediately ranged
body's of 1.4 and 1.7 mg/1, respectively. Highest BOD^
levels were observed in mineral soil-draining CS (2.3 mg/1)
and CH (2.3 mg/1). These concentrations were exceeded by
51
the BOD^ for TS, which was 4.1 mg/1. The TS level was
almost three times greater than the intermediate BOD^
recorded for Piedmont-draining TR (1.6 mg/1).
Arithmetic mean BOD^ levels exceeded the weighted means
in peat-draining EB and VTB, in Piedmont-draining TR, and in
most of the mineral soil-draining streams, with IS and GR
the exceptions (Figure 3b, Appendix Table 2). However, none
of the streams had significant correlations between dis-
charge and BOD^ which indicates that other factors m.ay
control BOD^ levels. positively correlated with
POC in CS (r = +0.85), PA (r = +0.65), and CH (r = +0.64)
(Appendix Tables 10, 11, and 12).
pH, Total Alkalinity, and Conductivity
A clear difference existed between pH levels of the
peat-draining and the mineral soil-draining streams; how-
ever, two of the streams categorized as peat-draining (MI
and BL) did not have the characteristic low pH of the other
streams in the peat-draining category (Figure 4a, Appendix
Table 2). This deviation is attributed to the influence of
Tertiary limestone in their drainage basins. The solution
of limestone deposits is responsible for the high pH levels
exhibited in mineral soil-draining IS (7.3), as well as in
peat-draining MI (6.8) and BL (5.9). Peat-draining EB, WB,
and CL exhibited low mean pH's ranging from 3.7 to 4.5.
Mineral soil-draining CS, PA, CH, and GR had interm.ediate
pH TA (meq/liter)
gure 4 Annual arithmetic means (shaded) and weighted means (solid) for (a) pH and (b)
total alkalinity (TA) from September 1978 through August 1979.
53
mean pH levels between 5.8 and 6.5. Intermediate mean pH
levels were also observed in Piedmont-draining TR (6.6) and
TS (6.3).
Although the arithmetic and weighted means for pH were
similar in most cases, the arithmetic means surpassed
weighted means (Figure 4a, Appendix Table 2). This trend
indicates that pH levels are highest during low flows.
However, a negative correlation between pH and discharge was
found only in IS (r = -0.62, Appendix Table 9).
The mean pH of precipitation over the study year was
4.1 for the Croatan rain gauge, and 4.8 for the Tar Sv/amp
rain gauge station (Table 2). These data suggest that
rainfall contributes to the acidity of Coastal Plain-
draining streams.
Peat-draining EB, WB, and CL exhibited the lowest total
alkalinity (TA) of all the streams sampled (Figure 4b,
Appendix Table 2). Mean TA for these streams were practi-
cally nil, with only 0.01 meq/1 recorded for CL, 0.02 meq/1
for EB, and 0.04 meq/1 for WB. The other streams in the
peat-draining category (MI and BL) did not have the same
trend as the other peat-draining streams due the influence
of Tertiary limestone within their watersheds. Mineral
soil-draining IS had the highest TA of 1.7 meq/1, followed
by peat-draining MI with 0.88 meq/1, and BL with 0.58 m.eq/1.
The high buffering capacity of these streams results from
solution of limestone deposits in their drainage basins.
Mineral soil-draining CS, PA, CH, and GR displayed
54
Table 2. Annual arithmetic means (A), with standard errors, and
annual weighted mean (W) of concentrations of rainfall
constituents in the Croatan (RGC) and Tar Swamp (RGT)
rain gauge collections. Measurements are expressed in
mg/1 for all constituents, except pH and total
alkalinity (which is in meq/1).
Constituent Mean RGC RGT
pH A 4.08±0.17 4.8210.26
W 3.60 3.90
TA A 0.01±0.004 0.0410.02
W 0.004 0.02
Ca A 1.26+0.26 0.8010.23
W 0.95 0.58
Mg A 0.30±0.07 0.2010.04
W 0.21 0.15
Na A 2.21±0.45 1.2010.28
W 1.68 0.90
K A 0.52±0.12 0.4710.09
W 0.37 0.40
NO.-N A 0.72+0.16 0.4210.08
W 0.56 0.58
NH, -N A 0.15±0.02 0.1010.024
W 0.13 0.10
TON A 0.23±0.04 0.3310.05
W 0.19 0.28
TP A 0.04±0.01 0.0610.02
W 0.03 0.05
TOC A 4.7510.75 4.03 10.51
W 4.03 3.70
55
intermediate mean TA ranging between 0.24 and 0.39 meq/1.
Piedmont-draining TR also had an intermediate mean TA of
0.36 meq/1, which was lower than the TA observed for TS
(0.45 meq/1).
Total alkalinity was highest at low flow periods for
all of the streams sampled (Figure 4b, Appendix Table 2).
*
In most cases the arithmetic mean was more than double the
weighted mean. The relationship between high TA and low
flow periods can be attributed to the mineral contributions
of base flow, v/hich becomes the primary source of stream-
water during dry conditions. A hydrologic shift to surface
runoff as the main source of streamwater, during winter and
spring, resulted in base flow dilution since rainfall
typically contains low concentrations of dissolved minerals
(Kuenzler et al. 1977). The result of this shift would be
an inverse relationship between TA and discharge rates.
However, negative correlations between TA and discharge were
significant only for IS (r = -0.67) and TR (r = -O.^p.)
(Appendix Tables 9 and 14, respectively).
Total alkalinity was low in the rainfall samples.
Precipitation from the Croatan rain gauge station had a mean
TA of 0.01 meq/1, and the Tar Swamp rain gauge sample had a
mean TA of 0.04 meq/1 (Table 2).
The peat-draining streams, geographically removed from
the influence of limestone, had lower conductivities than
the mineral soil-draining streams (Figure 5a, Appendix Table
2). Peat-draining EB and WB had the lowest conductivities.
COND (jjmho/cm) Ca (mg/liter)
Figure 5 Annual arithmetic means (shaded) and weighted means (solid) for (a)
conductivity (COND) and (b) calcium (Ca) from September 1978 through August Ul
1979.
57
with 51 and 47 umho/cm, respectively. Peat-draining CL (83
umho/cm) and BL (86 umho/cm) ranked in the intermediate
range of mean conductivity measurements. MI, of the peat-
draining category, had a high mean conductivity of 131
umho/cm, which reflects the influence of limestone deposits
on streamwater chemistry. This high value for MI is more
than twice the mean conductivities recorded for EB and WB.
Conductivity levels for the mineral soil-draining streams
showed a large degree of variation among streams. CS had
the lowest conductivity of the mineral soil-draining
streams, with 69 umho/cm. PA had an intermediate ranged
conductivity of 88 umho/cm. The mineral soil-draining
streams, IS, CH, and GR, had high conductivities between 116
and 188 umho/cm. The highest conductivity for all of the
streams sampled was observed in IS (188 umho/cm) and result-
ed from solution of limestone formations along the stream
channel. Piedmont-draining TR had an intermediate average
conductivity of 90 umho/cm. TS, however, exhibited a much
higher mean conductivity than TR, with 164 umho/cm.
Conductivity was generally lower at high flow periods
in most of the streams investigated, since the arithmetic
mean consistently exceeded the weighted mean (Figure 5a,
Appendix Table 2). Exceptions to this trend were observed
in peat-draining EB and WB, and also in mineral soil-
draining CS. Conductivity was positively correlated with
major cations in most of the m.ineral soil-draining streams
(Appendix Tables 9-13).
58
Cations
Lov/ mean calcium (Ca) concentrations were observed in
the peat-draining streams EB, WB and CL with levels between
1.0 and 2.1 mg/1 (Figure 5b, Appendix Table 3). The highest
Ca concentrations over the study year were observed in
mineral soil-draining IS, which dissects a Tertiary lime-
stone deposit. The water chemistry of MI and BL were also
affected by this deposit, giving them much higher Ca concen-
trations than the other streams in the peat-draining catego-
ry. Island Creek had a weighted mean Ca concentration of 17
mg/1 and an arithmetic mean of 43 mg/1. The strong negative
correlation (r=-0.79) between Ca concentrations and
discharge illustrates this relationship. Peat-draining MI
had a weighted mean similar to IS with 17 mg/1, but exhibit-
ed a lower arithmetic mean value of 26 mg/1. Peat-draining
BL had a weighted mean Ca content of 7.5 mg/1, making it
intermediate in Ca concentrations. However, the higher
arithmetic mean value of 15 mg/1 reflects the small influ-
ence of the limestone deposit on BL water chemistry. Base
flow probably accounts for the main influx of Ca into BL.
Mineral soil-draining CS, PA, CH, and GR displayed interme-
diate mean Ca levels ranging between 6.3 and 12 mg/1.
Piedmont-draining TR and TS are also within the intermediate
range of Ca concentrations, with 6.6 and 11 mg/1, respec-
tively.
59
Arithmetic mean Ca concentrations exceeded the weighted
mean for all the streams sampled, with two exceptions (V7B
and CS) (Figure 5b, Appendix Table 2). Generally, Ca
concentrations are highest at low discharge, except in WB
and CS where discharge rates did not appear to affect Ca
levels, since the means were equal. The dominance of Ca
among cations and its high correlation with conductivity
suggests that Ca is an important contributor to conductivity
levels in Coastal Plain streams.
The Ca concentrations measured in the bulk precipita-
tion samples were lower than in any of the streams sampled,
except for CL. The mean Ca content for both the Croatan
rain gauge and Tar Swamp rain gauge samples was less than
1.3 mg/1 (Table 2).
The lowest mean magnesium (Mg) levels were found in
peat-draining EB (0.69 mg/1) and WB (0.47 mg/1) (Figure 6a,
Appendix Table 2). Peat-draining CL, BL, and MI, and
mineral soil-draining IS and CS, displayed intermediate mean
Mg concentrations between 1.0 and 1.6 mg/1. The highest Mg
levels were observed in mineral soil-draining PA, CH, and
GR, with means between 1.9 and 2.6 mg/1. Piedmont-draining
TR is also ranked in the high mean category, with a Mg
concentration of 2.1 mg/1. TS had the highest overall Mg
level over the study year with a mean of 4.7 mg/1. This
value is double the concentration determined for TR and is
almost seven times greater than the average annual Mg
Mg (mg/liter) No (mg/liter)
Figure 6. Annual arithmetic means (shaded) and weighted means (solid) for (a) magnesium
(Mg) and (b) sodium (Na) from September 1978 through August 1979.
o
61
concentrations recorded for EB and WB of the peat-draining
streams.
In most of the stream.s, the arithmetic means for Mg
surpassed the weighted means (Figure 6a, Appendix Table 2),
indicating that Mg levels generally were highest at low
discharge. Negative correlations between Mg and discharge
were found in MI (r = -0.71), IS (r = -0.82) and GR (r =
-0.74) (Appendix Tables 7, 9, and 13). Exceptions to this
trend were found in EB [in which Mg levels were positively
correlated with discharge with r = +0.71 (Appendix Table 4)]
and WB (where the means were equal and no significant
correlations existed, disclosing that flow patterns do not
significantly affect Mg concentrations).
The Mg content of the precipitation samples was lower
than the average amount found in any of the Coastal Plain-
draining streams. The rainfall sample at Tar Swamp (0.20
mg/1) had a lower annual mean Mg concentration than the
Croatan rain sample (0.30 mg/1) (Table 2).
Peat-draining EB, WB, and CL had the lowest mean sodium
(Na) concentrations, ranging from 2.8 to 3.3 mg/1 (Figure
6b, Appendix Table 2). BL (4.0 mg/1) and MI (5.0 mg/1) had
higher Na levels than reported for the other peat-draining
streams. Sodium concentrations were between 4.0 and 4.9
mg/1 in mineral soil-draining IS, CS, and PA, giving them an
intermediate ranJcing. The highest mean Na levels were found
in mineral soil-draining CH (6.9 mg/1) and channelized GP.
(7.5 mg/1). Piedmont-draining TR also had a high mean Na
62
concentration of 7.4 mg/1. TS exhibited the highest overall
Na level (20 mg/1); however, the standard error of the m.ean
was also high (3.8) indicating a large amount of variation
in Na concentrations. The annual mean Na concentration in
TS was three times higher than the amount recorded for TR.
Weighted mean Na levels were higher than arithmetic
means in peat-draining EB and WB, and in mineral soil-
draining CS (Figure 6b, Appendix Table 2). This trend
suggests that Na concentrations are highest during high
discharge. BL and MI, in the peat-draining category, along
with the remaining mineral soil-draining streams (IS, PA,
CH, and GR) and Piedmont-draining TR, exhibited higher
arithmetic means than weighted means indicating that Na
levels are highest during low flow stages in most Coastal
Plain streams.
Rainfall samples were lower in Na than streamwater
concentrations. There was a fairly large difference in the
Na levels between rain gauge stations (Table 2). The Tar
Swamp rainfall sample had a mean Na concentration of 1.2
mg/1 whereas the Croatan rainfall had a mean of 2.2 mg/1.
Comparison of potassium (K) concentrations between
peat-draining and mineral soil-draining streams reveals that
peat-draining streams generally exhibit lower annual K
levels (Figure 7a, Appendix Table 2). The lowest K concen-
trations are found in peat-draining EB (0.35 mg/1) and V7B
(0.27 mg/1). Peat-draining MI, BL, and CL have intermediate
mean K, levels from 0.63 to 1.2 mg/1. Mineral soil-draining
K (mg/liter) Fe (mg/liter)
Figure 7. Annual arithmetic means (shaded) and weighted means (solid) for (a) potassium
(K) and (b) iron (Fe) from September 1978 through August 1979.
64
IS, CS, and PA displayed intermediate ranged K concen-
trations. IS has an annual mean K content of 0.94 mg/1, and
CS and PA have mean K levels of 2.5 mg/1 and 2.4 mg/1,
respectively. The highest mean K concentrations were in
mineral soil-draining CH (7.2 mg/1) and GR (3.8 mg/1). CH
exhibited the highest annual mean K level of all of the
streams studied, with a value that was 20 times greater than
the average K concentrations reported for peat-draining EB
and WB. Piedmont-draining TR exhibited an intermediate K
content of 2.7 mg/1. This value was similar to the mean K
level reported for TS, which was 2.9 mg/1.
The arithmetic means for K exceeded the weighted means
in all of the mineral soil-draining streams (except for CS) ,
in peat-draining BL and MI, and also in Piedmont-draining TR
(Figure 7a, Appendix Table 2). This trend indicates that
the highest K levels in these streams are associated with
low flow stages. In mineral soil-draining CS, and peat-
draining EB and WB, the weighted mean was greater than the
arithmetic mean, suggesting that high K concentrations occur
most frequently during high discharge; however, K and
discharge were not significantly correlated in any of the
streams examined.
Mean annual K levels for the two rain gauge stations
were generally lower than the concentrations reported for
streamv/ater (except in peat-draining EB and VJB) . The Tar
Swamp rainfall sample had a mean K concentration of 0.47
65
iîig/1 and the Croatan rainfall sample had a similar K level
of 0.52 mg/1 (Table 2) .
The lowest mean iron (Fe) concentrations occurred in
the peat-draining EB, WB, BL, MI, and CL with values betv/een
0.41 and 1.2 mg/1 (Figure 7b, Appendix Table 2). IS, from
the mineral soil-draining category, also had low Fe levels
(0.55 mg/1). Mineral soil-draining PA, CH, and GR exhibited
intermediate ranged Fe concentrations between 1.3 and 1.5
mg/1. The highest annual mean Fe level occurred in m.ineral
soil-draining CS (3.08 mg/1). Piedmont-draining TR had an
intermediate Fe concentration of 1.9 mg/1. TS exhibited a
higher mean Fe level (10.0 mg/1) than reported for any of
the streams. The m^ean Fe concentration in TS was five times
higher than the average Fe content found in TR. The high
arithmetic mean for Fe in the waters of TS is associated
with a high standard error (2.16), which reflects a large
degree of variation among collections.
Weighted means and arithmetic means for Fe were similar
in most cases, except in EB and CS, where the arithmetic
means were significantly greater than the weighted means,
and in TR, where the weighted mean also surpassed the
arithmetic mean (Figure 7b, Appendix Table 2). In the
streams where the arithmetic mean exceeded the weighted m.ean
(EB, WB, BL, and CS), Fe levels were highest during low
discharge. East Brice Creek exhibited a negative corre-
lation (r = -0.59) between Fe and discharge (Appendix Table
4). Identical weighted means and arithmetic means in
66
mineral soil-draining IS and PA, suggest that Fe concen-
trations were not affected by flow patterns. This is
supported by the lack of significant correlations between Fe
and discharge for these streams. The weighted means were
higher than the arithmetic means in MI, CH, GR, and TR,
indicating that high Fe levels correspond with high dis-
charge rates.
Nitrogen
Inorganic forms
The lowest nitrate-nitrogen (NO^-N) levels were found
in peat-draining EB, WB, and BL, with mean concentrations
between 0.09 and 0.12 mg/1 (Figure 8a, Appendix Table 1).
MI and CL had higher NO^-N levels than the other peat-
draining streams with intermediate means of 0.25 mg/1
reported for both of these streams. IS had the lowest
average NO^-N concentration (0.08 mg/1) in any of the
streams sampled. Mineral soil-draining CS and PA had
intermediate NO^-N levels of 0.24 and 0.23 mg/1, respective-
ly. The highest concentrations of NO^-N were found in
mineral soil-draining CH (1.2 mg/1) and in channelized GR
(1.5 mg/1). Piedmont-draining TR had an intermediate ranged
NO^-N level of 0.59 mg/1. The average NO^-N concentration
in TS (0.31 mg/1) was lower than the average annual amount
reported for TR.
NOj-N (mg/liter) NH^-N (mg/liter) NOj-N (mg/liter)
Figure 8. Annual arithmetic means (shaded) and weighted means (solid) for nitrate-
nitrogen (NO_-N), (b) ammonium-nitrogen (NH^-N), and (c) nitrite-nitrogen <J\
(NO2-N) from'^September 1978 through August 1979.
68
Weighted means and arithmetic means for NO^-N are
similar in most of the streams examined (exceptions: EB and
PA) (Figure 8a, Appendix Table 1). East Brice Creek and PA
were the only two streams that had significant correlations
between NO^-N and discharge (Appendix Tables 4 and 11).
The rain gauge stations had relatively high NO^-N
levels. The mean NO^-N concentration for Tar Swamp rainfall
samples was 0.42 mg/1, and Croatan rainfall had an average
NO^-N level of 0.72 mg/1 (Table 2). Therefore, precipita-
tion contained more NO^-N than found in streamwater samples
from most of the Coastal Plain-draining streams examined in
this study (exceptions: CH and GR) .
Peat-draining EB, WB, BL, and MI, exhibited low
ammonium-nitrogen (NH^-N) concentrations, along with mineral
soil-draining IS and PA. NH^-N levels in these streams were
between 0.01 and 0.04 mg/1 (Figure 8b, Appendix Table 2).
CL did not show the same trend as the other peat-draining
streams; instead, it was characterized by a high mean NH_^-N
level of 0.25 mg/1, which was 5 times higher than the
average amount reported for streams in the peat-draining
category. Mineral soil-draining CS had an intermediate mean
of NH^-N of 0.07 mg/1. However, mineral soil-draining CH
and channelized GR displayed high average annual NH^-N
concentrations of 0.20 mg/1 and 0.56 mg/1, respectively. CH
also had the highest weighted mean (0.42 mg/1) of any
stream studied. Piedmont-draining TR had an intermediate
NH^-N level of 0.14 mg/1. The average annual NH^-N content
69
of TS (0.10 mg/1) was slightly lower than the amount report-
ed for TR.
Ammonium-N concentrations had notably higher arithmetic
means than weighted means (Figure 8b, Appendix Table 1) in
peat-draining EB, WB, and BL, in mineral soil-draining CS,
and in Piedmont-draining TR. However, the lack of negative
correlations in any of these streams make the relationship
of questionable significance.
Precipitation samples contained intermediate NH^-N
concentrations in comparison to streamwater NH^-N levels.
The mean NH^-N level recorded for Tar Swamp rainfall was
0.10 mg/1 (Table 2). The Croatan rainfall sample had a
similar mean NH^-N concentration of 0.15 mg/1. Precipita-
tion input contained more NH^-N than streamwater outflow
from pocosins.
Peat-draining EB, WB, BL, MI, and CL exhibited extreme-
ly low annual mean NO2-N concentrations between 0.001 and
0.003 mg/1 (Figure 8c, Appendix Table 1). Mineral soil-
draining IS, CS, and PA also had low NO^-N levels within
this range. The highest NO^-N concentrations were found in
mineral soil-draining GR (0.01 mg/1) and CH (0.03 mg/1).
Piedmont-draining TR also had a relatively high NO^-N level
(0.01 mg/1). TS, however, had a lower m.ean NO^-N concen-
tration of 0.004 mg/1.
Nitrite-nitrogen (NO2-N) concentrations were not
related to flow patterns in peat-draining BL and MI, and in
m.ineral soil-draining IS and PA, since weighted means and
70
arithmetic means are equal (Figure 8c, Appendix Table 1) and
correlations with discharge were lacking. The arithmetic
means exceeded the weighted means in the remaining Coastal
Plain-draining streams and in Piedmont-draining TR, indicat-
ing that the highest NO^-N levels occurred during low flows.
Organic forms
Dissolved organic nitrogen (DON) concentrations were
similar for most Coastal Plain-draining streams (Figure 9a,
Appendix Table 1). Peat-draining EB, WB, BL, and MI
exhibited intermediate mean DON levels (between 0.40 and
0.56 mg/1) when compared to the low concentration reported
for mineral soil-draining IS (0.26 mg/1). The highest
annual mean DON level was found in the pocosin drainage
canal station CL with 1.17 mg/1. Intermediate DON levels
were recorded for the mineral soil-draining CS, PA, and GR.
The mean DON concentration for CS was 0.65 mg/1. PA and GR
both had DON levels of 0.51 mg/1. CH had the highest mean
DON level in the mineral soil-draining stream category (0.82
mg/1). TR exhibited an intermediate DON concentration of
0.45 mg/1. The mean DON level in TS (0.92 mg/1) was twice
as high as the average amount reported for TR.
DON displayed no distinguishing trend between peat-
draining and mineral soil-draining streams. Weighted means
were similar to arithmetic means in most cases (Figure 9a,
Appendix Table 1).
DON (mg/liter) PON (mg/liter)
0 0.2 0.4 0.6 0.8 1.0 1.2 0 .04 .08 .12 .16 .20 .24 .28 .32 .36 40
Figure 9. Annual arithmetic means (shaded) and weighted means (solid) for (a) dissolved
organic nitrogen (DON) and (b) particulate organic nitrogen (PON) from
September 1978 through August 1979.
72
Peat-draining EB, WB, BL, and MI exhibited the lowest
average annual particulate organic nitrogen (PON) concen-
trations between from 0.05 and 0.08 mg/1 (Figure 9b, Appen-
dix Table 1). Mineral soil-draining IS also fell within
this range. CL had a much higher PON level than the other
peat-draining streams with a mean of 0.24 mg/1. Intermedi-
ate average PON concentrations were reported for mineral
soil-draining PA (0.14 mg/1), CH (0.13 mg/1) and GR (0.17
mg/1). It is interesting to note that most nitrogen forms
(i.e., NO^-N, NH^-N, and DON) are generally highest in CH,
but this trend is not repeated for PON. The highest PON
level was observed in CS, with 0.38 mg/1. The arithmetic
m.ean for this stream is three times larger than the weighted
mean (0.13 mg/1), emphasizing the effect of flow on PON
concentrations. Piedmont-draining TR had an intermediate
ranged PON level of 0.14 mg/1. The mean annual PON content
of TS (0.36 mg/1) was greater than twice the amount reported
for TR.
Arithmetic means and weighted means for PON were
noticeably different in m.ost of the streams examined (Figure
9b, Appendix Table 1). However, there were no significant
correlations between PON and discharge (Appendix Tables
4-14).
Total organic nitrogen (TON) was determined for the
rain gauge samples over the study year (Table 2). The Tar
Swamp rain gauge samples had the highest average annual TON
concentration of 0.33 mg/1. Croatan rainfall had a mean TON
73
level of 0.23 mg/1. These values are intermediate between
DON and PON concentrations for streamwater.
Phosphorus
Peat-draining EB, WB, and BL exhibited the lowest total
phosphorus (TP) concentrations between 0.02 and 0.C4 mg/1
(Figure 10a, Appendix Table 1). MI and CL had intermediate
ranged mean TP values of 0.14 and 0.17 mg/1. The mineral
soil-draining streams displayed a wide range of TP levels.
PA had the lowest average TP content (0.06 mg/1). IS (0.11
mg/1), CS and GR (both with 0.16 mg/1) exhibited intermedi-
ate TP concentrations. CH had an extremely high mean TP
level of 0.41 mg/1 which was 20 times larger than the
concentrations reported for peat-draining EB and WB. CH
also had the highest TP weighted mean (0.32 mg/1) observed
in the streams examined. Piedmont-draining TR exhibited an
intermediate TP content of 0.20 mg/1. The average annual TP
level in TS (0.34 mg/1) was higher than the mean concen-
tration reported for TR.
Concentrations of TP were highest at low discharge for
most of the streams examined, since arithmetic means
generally exceeded weighted means (Figure 10a, Appendix
Table 1). Exceptions to this trend were in observed mineral
soil-draining PA and channelized GR. However, the only
significant correlation (r = -0.69) between discharge and TP
concentrations and discharge was for IS (Appendix Table 9).
TP (mg/liter) FRP (mg/liter) PP (mg/liter)
0 .04 .08 .12 .16 .20 .24 .28 .32 .36 .40 .44 0 .04 .08 .12 .16 .20 .24 .28 0 .04 .08 .12
Figure 10. Annual arithmetic means (shaded) and weighted means (solid) for (a) total
phosphorus (TP), (b) filterable reactive phosphorus (FRP), and (c) particulate
phosphorus (PP) from September 1978 through August 1979.
75
The rainfall samples contained low levels of TP. The
Tar Swamp rainfall sample (0.06 mg/1) had a slightly higher
average annual TP content than the Croatan rainfall sample
(0.04 mg/1).
Streams had similar ranking in concentrations of
filterable reactive phosphorus (FRP) as those observed for
TP, with MI the only exception (Figure 10b, Appendix Table
1). Peat-draining EB, WB and BL displayed the lowest mean
FRP concentrations, between 0.01 and 0.02 mg/1. Mineral
soil-draining PA was also included within this range.
Peat-draining CL had an intermediate FRP level of 0.08 mg/1
whereas MI (which had an intermediate TP) ranked in the high
range for mean FRP content with 0.14 mg/1. Mineral soil-
draining IS, CS, and GR had intermediate FRP concentrations
between 0.05 and 0.08 mg/1. CH, also a mineral soil-
draining stream, had the highest average annual FRP level of
0.26 mg/1. Piedmont-draining TR possessed an intermediate
mean FRP concentration of 0.08 mg/1. TS, however, had an
average annual FRP level that v/as double the mean amount
reported for TR, giving TS a high ranked FRP concentration
(0.16 mg/1).
Concentrations of FRP repeated the trend for TP of the
arithmetic means exceeding the weighted means for MI, IS,
CS, CH and TR (Figure 10b, Appendix Table 1). In these
situations, FRP levels would be highest at low discharge.
Both MI (r = -0.59) and IS (r = -0.75) had negative corre-
lations between FRP and discharge (Appendix Tables 7 and 9,
76
respectively). In PA and GR, the weighted means surpassed
the arithmetic means, indicating that the highest FRP levels
occur during high flows; however, no positive correlations
between FRP and discharge were found. No significant
correlations between FRP and discharge were observed in EB,
WB, and BL.
Particulate phosphorus (PP) concentrations were derived
by subtraction of filterable total phosphorus from, total
phosphorus and, hence, weighted means were not calculated
for this parameter. However, arithmetic means for PP were
examined (Figure 10c, Appendix Table 1). Peat-draining EB,
WB, BL, and MI exhibited the lowest mean PP concentrations
(0.003 to 0.02 mg/1). PP levels for mineral soil-draining
IS and PA also occurred within these extremes. CL exhibited
the highest PP concentration reported for the peat-draining
stream category with 0.04 mg/1. Mineral soil-draining CS,
CH, and GR possessed relatively high mean PP levels when
compared to the peat-draining streams. CS and CH had the
same mean PP level of 0.08 mg/1, which was unexpected due to
the extremely high concentrations of TP and FRP reported for
CH. Channelized GR exhibited a slightly lower PP level of
0.07 mg/1. Piedmont-draining TR had the highest annual
average PP concentration of all of the streams examined v^ith
0.09 mg/1. TR exhibited the highest overall mean PP level
of 0.11 mg/1. It should be noted that during the summier dry
months no water was collected in TS; therefore, the high
77
annual mean reflects the greater relative abundance of PP in
TS when water was present for collection.
Organic Carbon
Total organic carbon (TOC) and dissolved organic carbon
(DOC) concentrations were analyzed for all of the study
sites; however, only the results of the DOC analyses are
presented graphically and in the discussion of results,
because TOC and DOC display similar concentrations and
trends (Appendix Table 2).
Peat-draining streams exhibited the highest DOC content
of all of the stream types sampled (Figure 11a). EB, V7B,
BL, and MI had mean annual DOC concentrations between 19 and
28 mg/1. The peat-draining CL canal displayed the highest
average DOC concentration (51 mg/1). Channelized, mineral
soil-draining GR had the lowest mean DOC concentration of
7.6 mg/1. The remaining mineral soil-draining streams
displayed intermediate DOC levels. IS and CH both had
average annual DOC concentrations of 13 mg/1; PA possessed a
mean DOC content of 14 mg/1; and CH had the highest DOC
level of the mineral soil-draining streams with 16 mg/1.
Piedmont-draining TR exhibited a low mean DOC concentration
of 8.7 mg/1, which places it among the lowest DOC values
along with channelized, mineral soil-draining GR. TS had a
high mean annual DOC level of 28 mg/1.
DOC (mg/liter) POC (mg/liter)
Figure 11. Annual arithmetic means (shaded) and weighted means (solid) for (a) dissolved
organic carbon (DOC) and (b) particulate organic carbon (POC) from September
1978 through August 1979.
79
Arithmetic means exceeded the weighted means in the
peat-draining streams EB, WB, and BL, and also in mineral
soil-draining CS, indicating that the highest DOC concen-
trations occur during low flows (Figure 11a, Appendix Table
2). Negative correlations between DOC and discharge were
found in EB (r = -0.61) and WB (r = -0.60) (Appendix Tables
4 and 5, respectively). The reverse trend of higher weight-
ed mieans than arithmetic means was observed in the mineral
soil-draining IS, PA, CH, and GR, in peat-draining MI, and
in Piedmont-draining TR. This suggests that high DOC
levels, in these streams, occur during high discharge.
Positive correlations between DOC and discharge were ob-
served in IS (r = +0.66), MI (r = +0.59), and GR (r = +0.71)
(Appendix Tables 9 and 13, respectively).
Analyses of TOC were performed on the rainfall samples
collected at the rain gauge stations. Precipitation in the
Croatan (4.8 mg/1) had a slightly higher mean annual TOC
concentration than Tar Swamp rainfall (4.0 mg/1) (Table 2).
Precipitation input of organic carbon was generally lower
than streamwater concentrations throughout the study year.
Particulate organic carbon (POC) concentrations were
not directly analyzed on each collection date; instead, a
group of samples were collected, analyzed by dichromate
oxidation, and then used to extrapolate POC concentrations
(for each collection date) from PON-POC relationships (see
methods section). This method of was assumed to be more
reliable than subtraction of DOC from TOC to derive POC
80
values. Because of this, POC trends parallel those of PON
already discussed (Figure 11b).
The lowest mean POC level was found in mineral soil-
draining IS, with 0.90 mg/1. Mineral soil-draining PA, CH,
and GR (channelized) exhibited intermediate POC concen-
trations of between 1.6 and 2.0 mg/1. CS had the overall
highest mean POC level of 4.5 mg/1, which was associated
with a high standard error (1.68) indicating large variation
in POC content.
Piedmont-draining TR exhibited a low mean POC concen-
tration of 1.0 mg/1. Since both DOC and POC levels are low
in TR compared to the other stream examined, it may be
concluded that this Piedmont-draining river system annually
transports less organic carbon per unit of watershed than
natural. Coastal Plain-draining streams. Conversely, TS had
a high average POC level of 2.6 mg/1, with a relatively high
standard error (0.94) reflecting variation among samples in
POC concentrations. TS probably serves as a carbon source
for TR during flood stages.
DISCUSSION
Seasonal Trends
Chemical constituents of streamwater exhibit varia-
tions--some to a greater and others to a lesser degree--
which are evident from differences in the standard error of
the arithmetic mĆan (Appendix Tables 1, 2, and 3). However,
it is not apparent whether the standard error is an index of
seasonality or simply reveals unsystematic variations among
samples. A complex relationship among hydrological, litho-
logical, climatological, and biogeochemical factors, as well
as anthropogenic influences, may determine whether a season-
al response is present. Kuenzler et al. (1977) demonstrated
seasonal differences in chemical and physical parameters
between natural and channelized streams. Here a comparison
of seasonal trends is extended to include peat-draining
streams and a Piedmont-draining alluvial river.
Mean annual dissolved organic carbon (DOC) concen-
trations serve as a criterion for distinguishing peat-
draining streams from mineral soil-draining streams (Figure
11a, Appendix Table 2). Thus, trends in seasonal variation
of DOC concentrations may serve to further distinguish
stream types. Four streams (EB, IS, PA, and TR) representa-
tive of each stream category were selected to illustrate
trends in DOC concentrations and discharge between stream
types on a seasonal basis.
82
DOC concentrations in peat-draining East Brice Creek
(EB) were relatively high during all seasons of the year
compared to the mineral soil-draining streams and Piedmont-
draining TR (Figure 12a-d). DOC concentrations are nega-
tively correlated with discharge rates (r = -0.61) in EB
(Appendix Table 4). Peat-draining streams are characterized
by prolonged no-flow conditions in the summer and fall
(Figure 12a) when base flow is so low that stream velocity
is not measurable. Groundwater that eventually reaches
peat-draining streams contains high concentrations of DOC.
This is due to increased contact time between infiltrated
rainfall and peat, a process which leaches DOC from the
peat, and from removal of surface water by evapotranspira-
tion in summer and fall, which concentrates the DOC in
groundwater. Conversely, during the winter and spring
rainfall rejection occurs when the peat becomes saturated.
This results in lower DOC concentrations in streamwater
because the contact time between rainfall and peat is de-
creased (Figure 12a).
DOC concentrations in Island Creek (IS) were positively
correlated (r = +0.66) with discharge rates (Appendix Table
9). The highest DOC concentrations, as well as the highest
discharge rates, occurred during the spring and early summer
(Figure 12b). Apparently, high DOC levels during this time
were a result of increased contact time between organic
matter in the watershed and runoff waters. DOC contribu-
tions to the stream also include terrestrial runoff which
(Dm(Dmg/OligtOe/rlC)i.tCer) D(m'(IImSCVHAsReGcE)
figure 12. Seasonal patterns of dissolved organic carbon (•) and discharge (o) at (a) East
Brice Creek (EB), (b) Island Creek (IS), (c) Palmetto Swamp (PA), and (d) Tar
River (TR) from September 1978 through August 1979.
84
leaches organic carbon from organic soil horizons and the
forest floor during the spring when water tables are high.
Discharge rates are low and periods of no flow occur in late
summer and fall when évapotranspiration exceeds precipita-
tion. This corresponds with low DOC concentrations in IS
when surface runoff would be minimal and base flow becomes
the primary source of streamwater (Simmons and Heath 1979).
Base flow to IS is chemically different from base flow
contributions to the other mineral soil-draining streams in
this study because IS is underlain by Tertiary limestone.
Apparently, base flow delivers to IS only a minimal amount
of DOC which is either recently leached from the shallow
surface organic soil layers by rainfall or represents
residual DOC in the groundwater. During the winter when
discharge rates are highest, DOC concentrations in IS
increase due to increased overland runoff in contact with
organic layers, and the annual cycle is repeated.
There was no significant correlation between DOC
concentrations and discharge for mineral soil-draining
Palmetto Swamp (PA) (Appendix Table 11). Discharge rates
fluctuated widely during winter, spring, and early summer.
The discharge pattern stabilized in late summer and through-
out the fall between low discharge and no-flow conditions.
Although correlation between DOC concentrations and dis-
charge was not statistically significant, the highest DOC
concentrations in PA. occurred in late summer and throughout
the fall when discharge rates were lowest (Figure 12c). The
85
higher DOC concentrations may be controlled by two factors;
(1) long contact time of the nearly still water in a channel
with a bottom of high organic content, and (2) seepage of
base flow through the floodplain swamp into the stream
channel. In any case, DOC values fall intermediate between
base flow concentrations of peat-draining EB and mineral
soil-draining IS. Although not actually measured, it was
observed that the organic composition of the stream beds
roughly corresponds to the low flow DOC concentrations in
EB, IS, and PA.
Piedmont-draining Tar River (TR) exhibited a positive
correlation between DOC concentrations and discharge (r =
+0.69) (Appendix Table 14). Discharge patterns in TR
displayed erratic pulses of high and low flows during the
winter and spring months. Flow stabilized at lower levels
during the summer and fall (Figure 12d). TR flow was
perennial during the study year, unlike small, natural
Coastal Plain-draining streams, but characteristic of
channelized. Coastal Plain-draining streams, such as Grindle
Creek (GR). DOC concentrations showed less seasonal varia-
tion in TR than in the Coastal Plain-draining streams
examined in this study (Figure 12a-d). The broad, deep
channel of TR transports more water at a greater velocity
than any of the other streams sampled. Due to the wider
channel and larger flows, allochthonous sources of organic
carbon would have less influence on the chemical composition
of this large river system than in small stream.s. The
86
highest DOC concentrations occurred during winter and spring
peak discharges (Figure 12d). During some of these high
flow stages, floodwaters from TR overflowed into adjoining
floodplains. These seasonally flooded bottomlands, such as
Tar Swamp (TS), have high DOC concentrations (Appendix Table
2) which may serve to augment the DOC content of floodwaters
that gradually return to the river system. Moreover,
mineral soil-draining streams that are tributaries to TR
have higher DOC concentrations (see PA, Appendix Table 2 and
Figure 12c) . This latter source is probably the major
reason for the occasionally higher DOC peaks in TR during
the winter and spring period of high flows. Discharge is
lowest in the summer and fall with corresponding lower DOC
concentrations. Runoff from tributaries during these
seasons is minimal and would reduce the amount of organic
matter entering TR. It is likely that much of the discharge
at this time is due to flows from upstream that originate in
Piedmont and upper Coastal Plain tributaries. Som.e of the
DOC during low flow periods m.ay be derived from situ
algae production.
Characteristics of Peat-draining Streams
East Brice Creek (EB) and West Brice Creek (WB) can be
considered the best examples of peat-draining streams in
this study and will be used to characterize peat-draining
streams. These streams normally have high concentrations of
87
DOC but lov/ levels of 0^, (Ca, Mg, Na,
K and Fe), inorganic nitrogen (NO^-N, NH^-N, and NC^-N),
organic nitrogen (DON and PON), phosphorus (TP, FRP, and
PP), and POC. The parameters which most clearly distinguish
peat-draining streams from mineral soil-draining streams are
DOC, pH, TA, conductivity, and cations.
The low pH of pocosin soils (Dolm.an and Buol 1967,
Snyder 1977) and hence, of peat-draining streams, results
from a combination of factors. Pocosin vegetation produces
acid substances as by-products of their oxidative metabolism
(Moore and Bellamy 1974). Carboxyl-associated hydrogen
ions, which are released by freshly produced plant growth,
are actively exchanged for nutritionally important cations
that occur in rainfall (or in groundwater), thereby increas-
ing acidity of the soil and groundwater in pocosins (Wetzel
1983). Organic compounds such as humic, fulvic, and tannic
acids produced by pocosin plants will tend to accumulate in
the ecosystem due to continual deposition of autochthonous
primary production as peat and can be produced secondarily
from decomposition products (Moore and Bellamy 1974).
Leaching of these organic acids from peat occurs as water is
exported fromi the ecosystem, via runoff or base flow into
peat-draining streams, thereby contributing to the acidity
of pocosin drainage water. Increased acidity and reduction
of cations are especially evident in areas dominated by
mosses of the genus Sphagnum., which behave as cation ex-
changers (even when dead) because of their high
88
concentrations of polymerized unesterified uronic acids
which comprise up to 30% of the plant's dry weight (V7etzel
1983). The common pocosin moss Sphagnum, acting as cation
exchange sites, may have a significant effect on the ionic
composition of precipitation that passes through it (Moore
and Bellamy 1974). Cations, such as Ca and Mg, are absorbed
by the moss and hydrogen ions, and organic acids are subse-
quently released into the environment (Clymo 1963). By ion
exchanges. Sphagnum spp. regulate the pH of their surround-
ings and affect the availability of cations in the pocosin
ecosystem. Therefore, the cation exchange properties of
Sphagnum^ spp. indirectly contribute to the low pH of peat-
draining streams. Another significant factor contributing
to the acidity of pocosin drainage water is related to
precipitation, which serves as the main source of water
influx into pocosins (Otte 1981). The pH of precipitation
analyzed from the Croatan rain gauge station was low
throughout the study year. Also, waters with low dissolved
solids, such as rainwater, tend to become increasingly
acidic in contact with peat (Gosselink and Turner 1978).
The combination of these factors explains the low pH levels
characteristic of peat-draining streams.
The low TA of peat-draining streams is a measure of
their low buffering capacity. Contributions of hydrogen
ions from the sources mentioned above are especially effec-
tive in maintaining low pH's of pocosin waters. Low conduc-
tivities in peat-draining streams reflect the low dissolved
89
ionic content in the water and, thus, the low concentration
of cations in the pocosin ecosystem.
Low cation concentrations, along with low levels of
inorganic nutrients, partially result from autochthonous
deposition of organic matter within the pocosin ecosystem.
Peat accumulation causes the water table within the peat to
become elevated above the regional groundwater system and
isolates the surface waters from, the influence of underlying
mineral soils thereby restricting influx of minerals into
the pocosin ecosystem (Moore and Bellamy 1974, Daniel 1981,
Otte 1981). Due to the convex topography resulting from
peat accumulation and paludification, water flows periph-
erally from the pocosin ecosystem (Otte 1981). Therefore,
inorganic nutrients are not supplied by allochthonous
sources as they are in the other stream categories. As om-
brotrophic ecosystems, pocosins receive water and minerals
primarily from precipitation (Moore and Bellamy 1974, Otte
1981, Ash et al. 1983, Richardson 1983). Low cation levels
in pocosin drainage water may be attributed partly to the
low influx of cations into the pocosin ecosystem by pre-
cipitation (Table 2).
The reasons for high concentrations of DOC and low
inorganic nutrient levels are evident, in part, from the
source of streamwater which is the carbon-rich, nutrient-
poor pocosin ecosystem. The process of peat accumulation
removes nutrients from circulation by "locking up" nutrients
in organic matter. Peat accumulation occurs where the
90
decomposition rate is slower than the rate of plant primary
production (Snyder 1977). Decomposition is dependent on the
intensity of microbial activity within the peat substrate
(Moore and Bellamy 1974). This activity is influenced by
the presence of a fluctuating water table, responding to
évapotranspiration and precipitation, which allows periodic
aeration of the upper layers of peat providing both aerobic
and anaerobic microsites. Slow decomposition rates in
pocosins can be attributed to limited aerobic microbial
activity, because long hydroperiods cause anaerobic edaphic
conditions to predominate (Moore and Bellamy 1974). Only
inefficient, anaerobic microbes are active in waterlogged
soils (Brock 1979, Atlas and Bartha 1981). Low soil pH also
reduces microbial activity (Brock 1979). Inhibition of
decomposition results in a slow, incomplete turnover of
organic matter which retards the rate of nutrient cycling
due to the delayed release of organically bound nutrients
(Maimer 1975). Pocosins, in spite of their actual shortage
of available plant nutrients, serve as an accumulation
center not only for organic carbon but also for mineral
nutrients. The slow decomposition of organic carbon and the
slow circulation of minerals results in a buildup of stored
energy and nutrient pools in the pocosin ecosystem. There-
fore, peat-draining streams export more DOC than mineral
soil-draining streams because, quite simply, there is more
organic carbon in their watersheds. The high DOC concen-
trations observed in peat-draining streams reflect the
91
ability of pocosins to function as both a carbon sink for
the atmosphere and a source of DOC in drainage water. The
prolonged contact of groundwater with peat allows the water-
extractable fraction of organic matter to be leached into
pocosin drainage water.
Chemical analyses from the Catfish Lake (CL) sampling
station were not included in this description of charac-
teristics of peat-draining streams because CL does not
represent a natural system. It deviated due to its impound-
ment and the watershed use history (see "Stream De-
scriptions"). This artifically constructed pocosin drainage
canal, with impounded water, was chemically different from
peat-draining streams. For example, it had higher concen-
trations of organic carbon and lower levels of pH and TA.
Also, the unusually high NH^-N levels in CL could have
resulted from inhibition of nitrification due to anaerobic
conditions that may have developed in its stagnant waters.
Very few correlations between chemical and physical
parameters were found for peat-draining WB and CL (Appendix
Tables 5 and 8). An interesting pattern observed in these
streams was the paucity of correlations between the various
cations. The mineral soil-draining streams typically had
positive correlations between cations, the most common being
between Ca and Mg and Mg and Na ; however, various com.bina-
tions of significant cation correlations occur depending on
the stream examined. Total alkalinity, conductivity, and pH
were generally positively correlated with cations in most of
92
the streams. However, in the peat-draining streams this
relationship was not apparent due to the low cation status
of their waters (Appendix Table 2).
Comparison of Water Chemistry in North Carolina
and Minnesota Peatlands
Water chemistry of two North Carolina peat-draining
streams examined in this study (East Brice Creek and West
Brice Creek) are compared to drainage waters from Minnesota
peatlands in Table 3. The values presented are all arith-
metic mean concentrations but represent different numbers of
observations. Verry (1975) provides data from a 5 year
study, Clausen and Brooks' (1983) study was conducted over a
2 year interval and this study presents data from. 1 year of
biweekly collections.
Peatlands, also referred to as mires, can be subdivided
into distinctive types of wetlands dominated by peat. The
classification of a peatland as a fen or bog (the latter
includes pocosins) serves to elucidate differences between
these ecosystemis which have peat accumulation in common.
Fens are minerotropic peatlands which receive water and
nutrients from groundwater inflow and are therefore ion-
rich, with only slightly acid to slightly basic conditions
(Clausen and Brooks 1983). Bogs, which would be m.ore
similar to pocosins, are ombrotrophic peatlands which are
isolated from the influence of regional groundwater and thus
Table 3. Comparison of water chemistry in North Carolina streams and Minnesota bogs and
fens. These arithmetic mean concentrations are expressed in mg/1, with the
exception of conductivity (COND) values which are expressed in umho/cm.
Site NO^-N TCNTPFHPPP CaMgNa KFepH OI®
North Carolina
This study
Plei±nCTit-dralning TO 0.59 0.14 0.59 0.20 0.06 0.09 6.61 2.08 7.36 2.72 1.87 6.64 90
Peat-draining EB 0.12 0.03 0.46 0.02 0.01 0.01 2.01 0.69 3.31 0.35 1.21 4.50 51
WB 0.09 0.03 0.50 0.02 0.01 0.003 2.14 0.47 2.78 0.27 0.72 4.40 47
Mineral soil-draining CS 0.24 0.07 1.03 0.16 0.05 0.08 6.27 1.60 4.28 1.56 3.08 5.80 69
PA 0.23 0.03 0.65 0.06 0.02 0.02 9.57 1.88 4.91 2.07 1.52 6.40 88
CH 1.20 0.56 0.95 0.41 0.26 0.08 12.05 2.48 6.19 7.22 1.32 6.50 116
Channelized GR 1.A7 0.20 0.68 0.16 0.07 0.07 U.04 2.57 7.46 3.75 1.50 6.20 124
Knenzler ct al. 1977
Natural CS 0.16 0.62 0.39 0.16 0.07 0.08 7.10 1.60 5.10 3.20 2.10 5.70 78
PA 0.11 0.08 0.34 0.04 0.01 0.02 11.70 1.90 6.10 2.30 1.60 6.20 97
Ol 0.57 0.28 0.38 0.24 0.09 0.10 13.80 2.20 6.70 4.30 2.10 6.30 122
Channelized l.AO 0.08 0.36 0.13 0.03 0.08 13.90 2.50 8.20 3.70 0.80 6.50 139
Minnesota
Clau-sen and Brooks 1983
Bog 0.06 0.10 1.40 0.06 6.76 2.37 1.84 1.24 2.64 5.60 45
Transition 0.05 0.20 1.40 0.05 11.57 3.93 1.32 0.99 2.07 6.50 72
Fen 0.08 0.20 1.60 0.08 27.27 7.90 2.74 1.15 2.37 7.00 171
Verry 1975
Perched bog 0.20 0.45 0.69 0.19 2.40 0.97 0.60 1,30 1.35 3.60 51
Fen (groundwater) 0.10 0.15 0.33 0.09 16,60 2.88 2.00 1,10 0.98 6.50 125
94
dependent upon precipitation and atmospheric deposition as
the main sources of water and nutrients (Moore and Bellamy
1974). Hence, bogs are ion-poor and exhibit characteris-
tically acidic edaphic conditions (Sharitz and Gibbons
1982). Another distinguishing feature between these peat-
land types is that fen vegetation is domiinated by sedges and
tall shrubs, whereas bog (or pocosin) vegetation is dominât-
ed by Sphagnum moss and short-stature shrubs (Clausen and
Brooks 1983). Fens have a greater influx of nutrients and,
therefore, larger shrub species. Otte (1981) addresses the
relationship between peat depth and vegetational structure
in the Croatan National Forest pocosins. Many peatlands
possess characteristics of both bogs and fens; Clausen and
Brooks (1983) call these ecosystems transitional (some
researchers refer to them as poor fens).
Clausen and Brooks (1983) reported higher concen-
trations of NH^-N in fen runoff than in bog runoff, which
was not the case in the study by Verry (1975). Fens typi-
cally have more well-decomposed peats than bogs (Moore and
Bellamy 1974). The quantity of nutrients in peat is related
to their degree of decomposition (i.e., highly decomposed
peat contains more nutrients than undecomposed peat) (Lucas
and Davis 1961). Therefore, fens might be expected to
display greater concentrations of inorganic nitrogen than
bogs. Verry (1975) reported lower pH, and lower levels of
Ca, Mg, Na, Fe, and TON in bog runoff than observed by
Clausen and Brooks (1983). Also, Verry reported higher
95
concentrations of TP yet lower concentrations of TON than
reported by Clausen and Brooks. These differences may be
attributed to variations in the size of the watersheds
examined. Clausen and Brooks sampled runoff from large
peatlands (2.0-170 km^), whereas Verry sampled runoff from
much smaller watersheds (0.02-0.25 km^). Undoubtedly there
are other differences between study sites that may explain
contrasting water chemistry.
The watershed sizes of the peat-draining streams in my
study were intermediate between those of Clausen and Brooks
(1983) and Verry (1975). However, the values reported in my
study for EB and WB were similar to Verry's observations in
bog runoff for TON, Ca, Mg, Fe, and conductivity. On the
other hand, my results agreed with those of bog runoff
reported by Clausen and Brooks for NO^-N and TP. Amjnonium-N
concentrations in my study were lower than either of the
values reported for bog runoff in the Minnesota studies.
The pH levels of EB and WB were intermediate between the pH
levels reported by Clausen and Brooks (1983) and Verry
(1975). Potassium levels in EB and WB were lower than in
bog runoff in Minnesota. However, higher Na levels in North
Carolina peat-draining streams may be attributed to the
proximity of the Croatan National Forest pocosins to the
Atlantic Ocean.
96
Characteristics of Natural Mineral Soil-draining Streams
Creeping Swamp (CS) and Palmetto Swamp (PA) were chosen
for this discussion as representative of natural, mineral
soil-draining streams to provide a basis for comparison of
stream categories. The following characteristics attributed
to mineral soil-draining streams are based on the annual
arithmetic mean concentrations for CS and PA, in relation to
other streams, presented in Appendix Tables 1-3. These
stream.s exhibited low levels of NO^-N; intermediate-to-low
concentrations of 0^/ NH^-N, and phosphorus; intermediate
levels of pH, TA, conductivity, cations, NO^-N, and DOC; and
intermediate-to-high levels of POC and BOD^.
Water chemistry in natural, mineral soil-draining
streams is influenced by the interactions between stream-
water and the floodplains adjoining these streams when
overbank flooding occurs (Brinson 1977, Kuenzler et al.
1977, Wharton et al. 1982, Brinson et al. 1983, Yarboro,
1983). For example. Creeping Swamp's floodplain was ob-
served to retain net amounts of phosphorus through active P
uptake by algae (and incorporation into their biomass),
uptake by forest floor components (e.g., litter), and
sedimentation of particulate phosphorus during floodplain
inundation (Yarboro 1983).
During the summer and fall, when low to zero flow
conditions predominate, base flow becomes a significant
contributor to streamwater chemistry in natural, mineral
97
soil-draining streams (Kuenzler et al. 1977, Simmons and
Heath 1979). Groundwater usually carries more minerals than
surface water due to water acquiring CO^ as it passes
through the soil. This process produces a weak carbonic acid
solution that acts to solubilize carbonate and silicate
minerals to ions such as Ca, Mg, and Na (Simmons and Heath
1979) .
Annual arithmetic means from this study are compared to
those from the Kuenzler et al. (1977) study for stations at
the same location on Creeping Swamp (CS) and Palmetto Swamp
(PA) (Table 3). Chicod Creek (CH) is included in this
discussion since it was sampled in both studies. Higher
nitrogen concentrations are reported for this study than in
Kuenzler's. However, TP concentrations were comparable
between studies (with the exception of TP and FRP in CH
which were higher in this study). Cation concentrations
were in general agreement between studies. However,
Kuenzler et al. reported higher K levels in CS and lower K
concentrations in CH than reported for this study. Kuenzler
et al. reported lower Fe levels in CS and higher Fe levels
in CH than observed in this study. Conductivity and pH
levels between the two studies were similar and showed the
same basic relationships among streams (CS had the lowest pH
and conductivity levels, PA exhibited intermediate levels,
and CH had the highest levels). In general, there was good
agreement between results from the two studies, especially
considering the possibility of year-to-year variations and.
98
to some degree, the influence of using different analytical
methods.
Chicod Creek (CH) exhibited different water quality
characteristics from CS and PA (Appendix Tables 1-3) and
provides an example of the influence of anthropogenic
activities on natural streamwater chemistry. High nitrogen
and phosphorus levels in this stream presumably result
miainly from animal husbandry operations in the watershed.
Chicod Creek is a direct outlet for wastes from holding
ponds adjacent to poultry and livestock shelters (Simmons
and Aldridge 1980) . Intensive agricultural activities in
the drainage basin also contribute to nitrogen and phospho-
rus loading in CH by runoff of fertilizer nutrients. High
levels of conductivity and cations were also observed in
this stream which may be due to livestock waste influx and
exchangeable cations associated with high sediment loads
transported in terrestrial runoff (especially from plowed
fields or unpaved roads) during spates.
In the latter part of the study year, Chicod Creek was
undergoing channelization upstream from my collection site
(see discussion of CH in "Stream Descriptions" section).
Examination of water quality parameters during this transi-
tional phase did not reveal any changes in water chemistry
that could be directly attributed to channelization opera-
tions. Seasonal variations in stream.water chemistry during
this time may have overshadowed possible effects of channel
modification or perhaps there was not enough time for the
99
effects of channelization to start showing up in the down-
stream reaches of the creek.
Watkins and Simmons (1984) measured hydrological
conditions in Chicod Creek prior to and during channel
modification until the project completion in December 1981.
In comparison to the pre-channelization condition they found
higher base flow, higher suspended-sediment concentrations
during high flows, and higher concentrations of calcium,
sodium, and phosphorus.
Effects of Channelization on a Mineral
Soil-draining Stream
Channelized streams have several characteristics which
distinguish them from natural streams (Kuenzler et al.
1977). The results of this study support the findings of
Kuenzler et al. (1977) (Table 3). Grindle Creek (GR) was
the channelized stream examined in both studies. The
following description is based on the annual arithmetic
means in GR reported for my study (Appendix Tables 1-3).
V7hen compared to natural, mineral soil-draining streams,
channelized GR had higher concentrations of 0^, conduc-
tivity, cations (Ca, Mg, Na, and K), inorganic nitrogen, and
phosphorus; lower concentrations of DOC; and similar levels
of pH, TA, Fe, organic nitrogen, and POC.
Channelized streams typically have higher flow veloci-
ties than natural streams (Kuenzler et al. 1977). This is
100
due partially to containment of water flow by higher banks
than natural streams and also results from straightening,
widening, and deepening the stream channel (which is accom-
panied by obstruction removal). Another difference is in
the annual discharge pattern. Channelized streams typically
flow throughout the year, since the depth of the channel
allov/s continual base flow which sustains flow velocity
(Kuenzler et al. 1977). In contrast, natural streams
maintained prolonged no-flow periods during summer and fall
because increased transpiration by growing vegetation
effectively removes incoming precipitation. In Grindle
Creek, I observed that discharge increased rapidly in
response to heavy rainfall. Sediment loading of GR was also
most obvious during spates. I observed rapid transport of
allochthonous organic matter from the watershed (i.e,,
leaves, twigs, fruits, etc.) through the stream channel
during these high flow stages. When the waters receded and
turbidity decreased, the sandy bottom of the stream channel
was exposed, indicating that particulate organic matter was
swept downstream by rapidly flowing waters instead of
settling out and being incorporated into the sediments. The
total amount of organic matter in the bottom sediments of
slow-moving natural streams generally exceeds the quantity
of organic matter found in streamwater (Stumm and Morgan
1981). This was evident by the leaf litter layer observed
on the bottom of the stream channel in the natural streams
examined. Therefore, the influx of allochthonous organic
101
matter does not contribute to the chemical composition of
Grindle Creek in the same manner as in natural stream.s due
to its rapid export through the system. Lower concen-
trations of DOC would be expected in GF than in natural
streams because organic matter is the ultim.ate source of DOC
in aquatic ecosystems. This trend was verified in my study
(Figure 11a, Appendix Table 2). The low DOC content of GR
also may be attributed to the contained flow imposed by
channelization which restricts floodwater interaction v/ith
the natural floodplain (Kuenzler et al. 1977). The lower
concentrations of DOC in the waters of GR is also reflected
in the low color visually observed in these waters. Comp-
ared to the other blackwater streams in this study, GR had
the lightest colored water, which indicates a low aquatic
humus content. This characteristic was also observed by
Kuenzler et al. (1977).
The high levels of nitrogen (NO^-N, NH^-N, and NO^-N),
phosphorus, conductivity, and cations observed in GR m.ay be
attributed to: (1) inflow of sewage effluent from the town
of Bethel, North Carolina, (2) animal husbandry operations
in the watershed, (3) runoff from farmland loaded with
fertilizer nutrients, and (4) continuous base flow influx
throughout the study year. Water chemistry in GR was more
similar to that of CH than the other mineral soil-draining
streams possibly because of extensive anthropogenic activi-
ties within the watersheds of both streams. Chicod Creek
exhibited higher annual mean levels of NH^-N and phosphorus
102
than GR even though CH receives no point source of sewage
influx from a municipality. This may be due to higher
levels of non-point sources of livestock waste inflow.
Higher O2 concentrations in this channelized stream
could result partially fromc increased flow velocities (which
improves aeration). However, this relationship is not
supported by the correlation analysis which revealed that 0^
was negatively correlated with discharge (r - -0.59) in GR
(Appendix Table 13) . This relationship miay be attributed to
higher biological oxygen demand during high flows, which
might result from increased runoff from animal husbandry
activities in the watershed. The annual m.ean BOD^b of GR was
lower than most of the natural streams examined but BOD^
would be expected to fluctuate in response to organic
loading. Much higher BODY'S would be expected upstream in
GR near the sewage outfall than at the sampling site in this
study.
Grindle Creek had positive correlations between NH_^-N
and TP, FTP, and FRP (Appendix Table 13), which may be a
result of the combined effects of sewage influx and agricul-
tural runoff. The water chemistry in GR more closely
resembles that of Tar River than any other Coastal Plain-
draining stream, examined (see Appendix Tables 1-3) . This
may be due to the fact that both streams (1) have larger
sources of pollution from urban and rural wastes, (2)
exhibit perennial flow, and (3) have a higher export to
103
assimilation ratio of organic matter than natural streams
due to their continuous discharge.
Characteristics of a Mineral Soil-draining Stream
Associated with Limestone Formations
The natural, mineral soil-draining Island Creek (IS)
was chemically distinct from other mineral soil-draining
streams examined. Island Creek dissects a Tertiary lime-
stone formation which is exposed at the surface in parts of
the watershed. Limestone is a sedimentary rock which
consists mostly of calcite (CaCO^), intermixed with magne-
sium and other impurities (Hem 1970). Solution of limestone
into streamwater results in the extremely high Ca concen-
trations observed in this stream (Figure 5b, Appendix Table
2). Calcium carbonate has an important influence on the pH
of water through its carbon dioxide-bicarbonate equilibrium
system (Hem 1970, Goldman and Horne 1983, Wetzel 1983),
When compared to other Coastal Plain-draining streams.
Island Creek is characterized by: higher levels of 0^, pH,
TA, conductivity, and Ca ; lov/er levels of BOD^, Fe, and
inorganic and organic nitrogen. Its intermediate levels of
Mg, Na, K, phosphorus, and organic carbon are not diagnos-
tic. The high pH, TA, conductivity, and Ca levels reported
for IS are directly related to the solution of limestone in
streamwater. Alkalinity is an index of the lithology of the
watershed and the degree to which these rocks are weathered
(Cole 1983), so the high alkalinity of IS reflects a sub-
stantial amount of weathering of limestone in the watershed.
Base flow contributions to IS differ from that of the
other mineral soil-draining streams due to groundwater
interaction with limestone. This difference is most pro-
nounced during summer and fall low flow periods when base
flow becomes the principal source of stream water. Base
flow to IS viould contain low concentrations of organic
nitrogen and carbon for reasons already explained in the
section "Seasonal Trends." Also, IS represents a relatively
pristine stream which is not influenced by man's activities
to the extent observed in the other mineral soil-draining
streams (i.e., agriculture and animal husbandry operations
are scarce, if they even exist, in the watershed). This
would also account for the relatively low levels of phospho-
rus observed in IS.
Due to the influence of limestone formations in their
watersheds. Black Swamp (BL) and Mill Creek (MI) differed
from the other streams that originate in the Croatan Nation-
al Forest pocosins. Although both streams had high organic
carbon concentrations characteristic of peat-draining
streams, they do not fit well into this category when other
water quality parameters are examined. Black Swamp and Mill
Creek had higher levels of 0^/ pH, TA, conductivity and
cations than the other peat-draining streams. Mill Creek is
apparently more affected by solution of limestone in its
streamwater than BL which is evident from the higher Ca
105
concentrations (Appendix Table 2). This, in turn, is
associated with higher pH, and conductivity in MI than BL.
Correlation coefficients in BL and MI deviated from
patterns observed in the other peat-draining streams. The
correlations observed in these streams are similar to IS,
from the mineral soil-draining category, whose streamwater
is even more greatly influenced by limestone since the
stream channel in IS directly dissects a Tertiary limestone
formation. BL is the least affected of these three streams
by solution of limestone in its streamwater; therefore it
does not exhibit the same pattern of correlations observed
for IS and MI. An interesting trend observed in IS and MI
that was not present in any of the other Coastal Plain-
draining streams (except isolated, rare cases in EB, WB and
GR) was a negative correlation between organic carbon (TOC
and DOC) and cations (Ca, Mg, Na, and K) (Appendix Tables 7
and 9). Phosphorus, pH, TA, and conductivity were also
negatively correlated with organic carbon in IS and MI;
however, this trend was not observed in the other Coastal
Plain-draining streams. A negative correlation between DON
and cations, as well as phosphorus was observed only in IS.
Characteristics of a Piedmont-draining Alluvial River
Tar River (TR) originates in the Piedm.ont and passes
through the central Coastal Plain as it flows toward the
Pamlico River estuary. Approximately 78% of the TR drainage
106
basin is located in the undulating topography of the Pied-
mont, where extensive floodplains are not as prevalent as in
the Coastal Plain. When TR reaches the more extensive
floodplains such as Tar Swamp (TS) in the Coastal Plain,
there is an opportunity for greater stream channel-flood-
plain interaction. In addition. Coastal Plain-draining
tributaries contain significant am.ounts of swamp forests in
their watersheds which influence streamwater chemistry.
Chicod Creek and Grindle Creek are two tributaries to TR
that influence the water chemistry of this river system.
When compared to the two categories of Coastal Plain-
draining streams examined in this study, TR is characterized
by higher levels of 0^ and cations (Ca, Mg, Na, and K);
lower concentrations of organic carbon; and intermediate
levels of pH, TA, conductivity, Fe, BOD^, inorganic and
organic nitrogen, and phosphorus.
Because floodplains serve as an extension of the river
channel during high flows (Brinson et al. 1981b), flood-
plains function as temporary water retention basins, slowly
releasing floodwaters back to the river during low flow
periods. While water is retained in the floodplain, part of
the sediment load is deposited due to the reduced competence
of slow-moving to standing water and also because of sedi-
ment entrapment within litter and debris on the forest
floor. Flooding provides allochthonous inputs of minerals,
nutrients, and organic matter, along with sediment, from the
floodwaters to the swamp forest ecosystem. My data show
107
that water chemistry in the swamp often differed from that
in the adjoining river system (Appendix Tables 1, 2, and 3).
This difference results from chemical and biological inter-
actions between floodwaters and the biotic and abiotic
components of the swamp forest ecosystem (Brinson et al.
1983). The water chemistry of standing water collected in
TS (when available) over the study year, when compared to
the water chemistry of TR and other Coastal Plain-draining
streams, exhibited higher levels of cations (Mg, Na, K, and
Fe), conductivity, BOD^, DON, phosphorus, and DOC; lower
concentrations of 0^; and intermediate levels of Ca, pH, TA,
and inorganic nitrogen.
An explanation for the intermediate levels of inorganic
nitrogen reported for TS results from nitrogen cycling
within this ecosystem. During the characteristic sumirier and
fall drydown phase which occurs in southeastern túpelo-
cypress swam^p forests, ammonification, nitrification, and
subsequently denitrification in the sediments are stimulated
(Brinson et al. 1981a). Through these microbial transforma-
tions, sedimentary nitrogen reserves are depleted. This
restores the swamp's capacity to assimilate nitrogen, and
therefore, potentially reduces the amount of inorganic
nitrogen available for transport to the estuary. Kuenzler
et al. (1979) and others have suggested that nutrient
loading of estuaries should not be increased because of the
potential for damaging algal blooms. Floodplains such as TS
may significantly affect the inorganic nitrogen content of
108
TR. When early spring rains produce high runoff loaded with
sediment and fertilizer nutrients, TR distributes them
across the TS floodplain when overbank flooding occurs.
This may reduce the nutrient load that the TR delivers to
the Pamlico River estuary.
High Mg and Na levels in TS may be related to the
effects of brackish water intrusion into the swamp when the
water table is low. During the dry summer months, a salt
wedge may move up the Tar River, propelled by wind tides.
Seepage from the river channel into the groundwater of TS
may occur due to the low water table created by evapotran-
spiration. This seems to be the most reasonable explanation
for higher Mg and Na concentrations in the swamp.
Exports to Downstream Ecosystems
Annual weighted mean concentrations of nitrogen,
phosphorus, and carbon were multiplied by estim.ated runoff
values (see "Methods") to obtain annual exports of various
streamwater constituents (Table 4). Export of inorganic
nitrogen (NO^-N and NH^-N) was similar in peat-draining
streams and in some of the mineral soil-draining streams,
with notable exceptions in CH and GR (which show effects of
known point and non-point sources of nitrogen). Nitrate-N
export from PA falls between those of the nitrogen enriched
and other mineral soil-draining streams. Dissolved organic
nitrogen (DON) and particulate organic nitrogen (PON)
109
Table 4. Annual runoff and export of nitrate-nitrogen,
ammonium-nitrogen, dissolved organic nitrogen,
particulate organic nitrogen, total phosphorus, total
organic carbon, and dissolved organic carbon from
September 1978 through August 1979.
Runoff Export (g/m^.yr)
Stream (m/yr)
NO^-N NH4, -N DON PON TP TOC DOC
EB 0.474 0.11 0.01 0.19 0.04 0.005 9.06 8.24
WB 0.474 0.05 0.01 0.20 0.03 0.005 11.26 10.67
BL 0.474 0.06 0.01 0.25 0.05 0.01 12.22 12.12
MI 0.474 0.10 0.02 0.23 0.03 0.06 11.48 10.99
IS 0.474 0.03 0.01 0.16 0.02 0.03 8.50 8.13
CS 0.648 0.14 0.02 0.30 0.08 0.03 9.75 9.10
PA 0.648 0.29 0.02 0.35 0.03 0.04 10.40 9.03
CH 0.491 0.58 0.21 0.33 0.03 0.16 7.49 7.03
GR 0.491 0.71 0.10 0.29 0.11 0.10 6.78 5.32
TR 0.491 0.30 0.05 0.20 0.05 0.09 5.39 5.11
lio
exports were similar in Coastal Plain-draining streams from
all categories, although the channelized GR appears higher.
Export of TP was lowest in the peat-draining streams and
highest in mineral soil-draining CH and GR, which have
localized loading of phosphorus from point and non-point
sources within their watersheds. Predictably, peat-draining
streams generally export more organic carbon than mineral
soil-draining streams. Channelized GR and Piedmont-draining
TR exhibited the lowest organic carbon export values.
In a study conducted in pocosins of the Albemarle-
Pamlico peninsula in northeastern North Carolina, Skaggs et
al. (1980) examined the TOC content of runoff waters from
organic-rich soils under naturally vegetated conditions.
The export of TOC (average of 2 years) in runoff waters was
26 g/m^.yr from deep organic soils and 22 g/m^.yr from
shallow organic soils. These values are about twice those
observed in the peat-draining streams examined in my study
(Table 3). This difference may be a result of two factors.
The first is watershed size. Watersheds of peat-draining
streams in the Croatan ranged from 28 to 32 km^ while those
of Skaggs et al. (1980) were only 0.07 km^. The larger the
watershed, the greater the possibility that sources other
than deep-peat drainage are included in export water (i.e.,
mineral soil areas and discharge from groundwater). The
other factor is the possibility of an artifact introduced by
ditching small watersheds so drainage can be controlled and
measured, as was necessary in the study of Skaggs and
Ill
co-workers. Ditching could provide an avenue for export of
deeper groundwater that would, under natural drainage
patterns, contribute little to export. DOC concentrations
in peat groundwater tend to increase with depth (Laurie
Loftin, personal communication, 1984), presumably as a
function of lower turnover rate (and export) of the deeper,
as compared with shallower groundwater. Although the
Catfish Lake station in my study was not monitored for
export, DOC concentrations were generally higher than in
peat-draining streams (Appendix Table 2). This may be
partially attributed to ditching and thus intersecting
deeper and more organic-rich groundwater and partially to
the stagnation of water in a peat-lined ditch.
Schlesinger and Melack (1981) summarized the rate of
loss of organic carbon from terrestrial watersheds (less
than 10,000 km^) in the United States. The export of TOC
from various temperate forests range from 1.52 to 6.14
g/m^.yr. These values are generally lov;er than the export
of TOC observed in Coastal Plain-draining streams in North
Carolina (Table 3). Comparison of river systems in North
Carolina reveals that TR exports more TOC (Table 4) than the
Neuse River (2.72 g/m^.yr).
Mulholland and Kuenzler (1979) also provide a compari-
son of organic carbon export for various upland and swamp-
draining watersheds. Export of DOC from the Coastal Plain-
draining streams examined in this study is higher than DOC
112
export in the upland streams, and most of the swamp-draining
streams, used in their study comparison.
Conclusions
Streams are much more than transporters of water and
materials; they are complex biogeochemical processing
systems and are also responsive to biogeochemical conditions
in their watersheds. Lithological attributes of the water-
shed, biochemical and physiochemica1 competition for nutri-
ents, the decompositional release rates and recycling phases
of nutrients, as well as ecological turnover times of
individual nutrients, to a degree, regulate nutrient avail-
ability and hence, the nutrient status of lotie ecosystems.
Moreover, water chemistry not only varies from stream to
stream but also from station to station on each stream
(Kuenzler et al. 1977). The chemical composition of stream-
water is subject to considerable changes as it passes
through the watershed due to localized biogeochemical
interactions. Therefore, the chemical composition of
streams also reflects the headwater origin, the nature of
water sources, and the drainage basin geology as streams
flow toward the sea. However, it is not possible at this
time to separate the importance of each of these variables
because, in part, these variables tend to be inseparable in
complex natural ecosystems.
113
In spite of the complexity of the relationship between
watersheds and streams, this study demonstrates that water
chemistry can be used to detect differences in landscape
characteristics of watersheds in one region of the North
Carolina Coastal Plain. Limitations are clearly apparent.
For example, some water quality characteristics are better
indicators than others. Also, although the Coastal Plain
watersheds used in this study were relatively small, they
are by no means homogeneous in lithological characteristics,
even if land use is eliminated as a variable. For example,
the peat-draining category designated for some stream.s (MI
and BL) actually turned out to have strong signals of
mineral soil influence, such as higher levels of cations,
pH, TA, and conductivity. The presence of limestone in
watersheds further complicates attempts to identify rigid,
well-defined categories. Finally, Piedmont-draining
streams, because of their large catchment area, are really
an amalgam of many stream categories by the time they
traverse the Coastal Plain.
The results of this study have implications for water
quality monitoring efforts by state and federal agencies to
miaintain water quality in streams by controlling sources of
pollution. Specifically, it should be recognized that the
quality of stream water in moderately altered landscapes
varies among watersheds with different biogeochemical
attributes. Streams not only have different background
water quality characteristics, but also have varying
114
capacities to assimilate products of human activities.
Major conclusions from this study and their relevance to
water quality management are summarized as follows.
(1) Coastal Plain-draining streams have distinguishing
characteristics which can be used to detect the geolog-
ic origin of streamwater.
(a) Water chemistry in peat-draining streams differs
from that in mineral soil-draining streams by
having higher DOC concentrations and lower levels
of pH, total alkalinity, conductivity, cations,
nitrogen, and phosphorus.
(b) Streams with limestone formations in their water-
sheds have higher levels of Ca, pH, TA, and
conductivity than streams that are not affected by
limestone.
(c) Catégorial separation of streams based on their
geologic origin is basically sound—however, large
watersheds often possess a wide range of edaphic
conditions which results in overlapping stream
types (such as BL and MI streamwaters which were
influenced by both peat and limestone in their
watersheds).
(d) Seasonal trends in DOC concentrations distinguish
stream types.
(e) Small, natural streams export more organic carbon
than small, channelized streams and large river
systems.
115
(f) Water in the river channel is lower in levels of
most chemical constituents (except dissolved
oxygen) than water collected in the adjoining
swamp forest.
(2) Water quality characteristics of Coastal Plain-draining
streams are different, but converge when anthropogenic
influences are superimposed on these systems. Nutrient
concentrations are higher in streams which receive
sewage discharge fromi municipalities, livestock wastes,
and fertilizer runoff, than in streams with no major
point source pollution.
(a) Coastal Plain-draining streams that received
localized loading of nitrogen and phosphorus (such
as CH and GR) showed indications that the assimi-
lative capacity of the stream for processing
wastes has been exceeded.
(b) Channelized Grindle Creek differed from natural
streams by having perennial flow; higher levels of
cations, conductivity, and nutrients; and lower
dissolved organic carbon concentrations.
(c) It is anticipated that peat-draining streams would
be most sensitive to acid rain due to the low
buffering capacity of their waters.
LITERATURE CITED
American Public Health Association (APHA). 1971. Standard
methods for the examination of water and wastewater.
13th edition. APHA, Washington, District of Columbia.
8 74 pp.
American Public Health Association (APHA). 1975. Standard
methods for the examination of water and wastewater.
14th edition. APHA, Washington, District of Columbia.
1193 pp.
Ash, A. N., C. B. McDonald, E. S. Kane, and C. A. Pories.
1983. Natural and modified pocosins: literature
synthesis and management options. FWS/OBS-83/04,
United States Fish and Wildlife Service, Division of
Biological Services, Washington, District of Columbia.
156 pp.
Atlas, R. M. and R. Bartha. 1981. Microbial ecology:
fundamentals and applications. Addison-Wesley Publish-
ing Co., Reading, Massachusetts. 559 pp.
Bannister, P. 1976. Introduction to plant physiological
ecology. Blackwell Scientific Publications, Oxford,
England. 273 pp.
Eanse, K. 1974. On the interpretation of data for the
carbon-to-nitrogen ratios of phytoplankton. Limnology
and Oceanography 19:695-699.
Barr, A. J. and J. H. Goodnight. 1972. A users guide to
the statistical analysis system. North Carolina State
University, Raleigh, North Carolina. 260 pp.
Beck, K. C., J. H. Reuter, and E. M. Perdue. 1974. Organic
and inorganic geochemistry of some coastal plain rivers
of the southeastern United States. Geochimica et
Cosmochimica Acta 38:341-364.
Beckman Instruments, Inc. 1977. Instruction m.anual -
conductivity bridge RC 16C. Cedar Grove, New Jersey.
19 pp.
Boelter, D. H. 1970. Important physical properties of peat
materials. Pages 150-154 ^ Proceedings of the Third
International Peat Congress, Quebec, Canada.
Brinson, M. M. 1977. Decomposition and nutrient exchange
of litter in an alluvial swamp forest. Ecology
58:601-609.
117
Brinson, M. M., H. D. Bradshaw, and E, S. Kane. 19Sla.
Nitrogen cycling and assimilative capacity of nitrogen
and phosphorus by riverine wetland forests. Report No.
167, Water Resources Research Institute of the Univer-
sity of North Carolina, Raleigh, North Carolina. 90
pp.
Brinson, M. M., B. L. Swift, R. C. Plantico, and J. S.
Barclay. 1981b. Riparian ecosystems: their ecology
and status. United States Fish and Wildlife Service,
Biological Services Program FWS/OBS-81/17. Eastern
Energy Land Use Team, Washington, District of Columbia.
155 pp.
Brinson, M. M., H. D. Bradshaw, and R. N. Holmes. 1983.
Significance of floodplain sediments in nutrient
exchange betv/een a stream and its floodplain. Pages
199-221 ^ T. D. Fontaine, III and S. M. Bartell,
editors. Dynamics of lotie ecosystems. Ann Arbor
Science Publishers, Ann Arbor, Michigan.
Brock, T. D. 1979. Biology of microorganisms.
Prentice-Hall, Inc., Englewood Cliffs, New Jersey. 802
pp.
Brown, L. and E. G. Bellinger. 1977. Nitrate determination
in fresh and some estuarine waters by ultravioJ.et light
absorption: a new proposed method. Water Research
12 : 223-229.
Christensen, N. L., R. B. Burchell, A. Liggett, and E. L.
Simms. 1981. The structure and development of pocosin
vegetation. Pages 43-62 C. J. Richardson, editor.
Pocosin wetlands. Hutchinson Ross Publishing Co.,
Stroudsburg, Pennsylvania. 364 pp.
Clausen, J. C. and K. N. Brooks. 1983. Quality of runoff
from Minnesota peatlands: I. A characterization.
Water Resources Bulletin 19:763-767.
Clymo, R. S. 1963. Ion exchange in Sphagnum and its
relation to bog ecology. Annals of Botany New Series.
27:309-324.
Coffey, A. 1982. Stream improvement: the Chicod Creek
episode. Journal of Soil and Water Conservation
37:80-82.
Cole, G. A. 1983. Textbook of limnology. 3rd edition.
C. V. Mosby Company, St. Louis, Missouri. 401 pp.
Cummins, K. W. 1974. Structure and function of stream
ecosystems. Bioscience 24:631-641.
118
Daniel, C. C., III. 1981. Hydrology, geology, and soils of
pocosins: a comparison of natural and altered systems.
Pages 69-108 C. J. Richardson, editor. Pocosin
vćtlands. Hutchinson Ross Publishing Co., Stroudsburg,
Pennsylvania.
Dolman, J. D. and S. W. Buol. 1967. A study of organic
soils (histosols) in the Tidewater region of North
Carolina. North Carolina Agricultural Experimental
Station Technical Bulletin No. 181, Raleigh, North
Carolina. 52 pp.
Environmental Protection Agency (EPA). 1976. Methods for
chemical analysis of water and wastes. Environmental
Monitoring and Support Laboratory, Office of Research
and Development, United States Environmental Protection
Agency, Cincinnati, Ohio. 430 pp.
Fisher, S. G. and G. E. Likens. 1973. Energy flow in Bear
Brook, New Hampshire: an integrative approach to
stream ecosystem metabolism. Ecological Monographs
43:421-439.
Gjessing, E. T. 1976. Physical and chemical characteris-
tics of aquatic humus. Ann Arbor Science Publishers,
Inc., Ann Arbor, Michigan. 113 pp.
Goldman, C. R. and A. J. Horne. 1983. Limnology.
McGraw-Hill Book Company, New York, New York. 464 pp.
Golterman, H. L. and R. W. Clymo, editors. 1969. Methods
for chemical analysis of fresh waters. International
Biological Program Handbook No. 8, Blackwell Scientific
Publications, Oxford, England. 166 pp.
Gosselink, J. G. and R. E. Turner. 1978. The role of
hydrology in freshwater wetland systems. Pages 63-78
in R. E. Good, D. R. Whigham, and R. L. Simpson,
editors. Freshwater wetlands. Academ.ic Press, Inc.,
New York, New York.
Hall, R. J., G. E. Likens, S. B. Fiance, and G. R. Hendrey.
1980. Experimental acidification of a stream in the
Hubbard Brook Experimental Forest, New Hampshire.
Ecology 61:976-989.
Heath, R. C. 1980. Basic elements of groundwater hydrology
with reference to conditions in North Carolina. U.S.
Geological Survey Water-Resources Investigations
Open-File Report 80-44, United States Geological
Survey, Raleigh, North Carolina. 86 pp.
119
Hem, J. D. 1970. Study and interpretation of the chemical
characteristics of natural water. Geological Survey
Water-Supply Paper 1473. United States Government
Printing Office, Washington, District of Columbia. 363
pp.
Robbie, J. E. 1974. Nutrients and eutrophication in the
Pamlico River estuary. North Carolina, 1971-1973.
Report No. 100, Water Resources Research Institute of
the University of North Carolina, Raleigh, North
Carolina. 239 pp.
Ingram, R. J. and L. J. Otte. 1981. Peat deposits of the
Croatan Forest: Craven, Jones, and Carteret counties.
North Carolina. Contract No. DE-AC01-79ET-14693,
United States Department of Energy, North Carolina
Energy Institute, Department of Geology of the Univer-
sity of North Carolina at Chapel Hill, Chapel Hill,
North Carolina. 20 pp.
Kaushik, N. K. and H. B. Hynes. 1968. Experimental study
on the role of autumn shed leaves in aquatic environ-
m^ents. Journal of Ecology 56:229-243.
Kirby-Smith, W. W. and R. T. Barber. 1979. The water
quality ramifications in estuaries of converting forest
to intensive agriculture Report No. 148, Water Re-
sources Research Institute of the University of North
Carolina, Raleigh, North Carolina. 70 pp.
Kuenzler, E. J., P. J. Mulholland, L. A. Ruley, and R. P.
Sniffen. 1977. Water quality in North Carolina
Coastal Plain streams and effects of channelization.
Report No. 127, Water Resources Research Institute of
the University of North Carolina, Raleigh, North
Carolina. 160 pp.
Kuenzler, E. J., D. W. Stanley, and J. P. Koenings. 1979.
Nutrient kinetics of phytoplankton in the Pamlico
River, North Carolina. Report No. 130, Water Resources
Research Institute of the University of North Carolina,
Raleigh, North Carolina. 163 pp.
Likens, G. E., F. H. Bormann, R. S. Pierce, J. S. Eaton, and
N. M. Johnson. 1977. Biogeochemistry of a forested
ecosystem. Springer-Verlag, New York, New York. 146
pp.
Lucas, R. E. and J. F. Davis. 1961. Relationship between
pH value of organic soils and availabilities of 12
plant nutrients. Soil Science 92:177-182.
120
Kaki, T. E., A. J. Weber, D. W. Hazel, S. C. Hunter, B. T.
Hyberg, D. M. Flinchum, J. P. Lollis, J. E. Rognstad,
and J. D. Gregory. 1980. Effects of stream
channelization on bottomland and swamp forest
ecosystems. Report No. 147, Water Resources Research
Institute of the University of North Carolina, Raleigh,
North Carolina. 135 pp.
Maimer, N. 1975. Developm.ent of bog mires. Pages 85-92
A. D. Hasler, editor. Coupling of land and water
systems. Springer-Verlag, New York, New York. 309 pp.
Mann, K. H. 1969. The dynamics of aquatic ecosystems.
Pages 1-81 J. B. Cragg, editor. Advances in ecolog-
ical research, Vol. 6, Academic Press, New York, New
York.
Moore, P. D. and D. J. Bellamy. 1974. Peatlands.
Springer-Verlag, New York, New York. 221 pp.
Mulholland, P. J. and E. J. Kuenzler. 1979. Organic carbon
export from upland and forested wetland watersheds.
Limnology and Oceanography 24:960-966.
National Oceanic and Atmospheric Administration (NOAA).
1973. Monthly averages of temperature and precipita-
tion for state climatic divisions 1941-70, Climato-
graphy of the U.S. No. 85. U.S. Department of Comi-
merce, National Climatic Center, Asheville, North
Carolina. 6 pp.
National Oceanic and Atmospheric Administration (NOAA).
1978, 1979. Climatological data for North Carolina.
Annual Summary. United States Government Printing
Office, Washington, District of Columbia. 12 pp. each.
Nelson, D. J. and D. C. Scott. 1962. Role of detritus in
the productivity of a rock-outcrop community in a
Piedmont stream. Limnology and Oceanography 7:396-413.
Odum, E. P. 1971. Fundamentals of ecology. W. B. Saunders
Co., Philadelphia, Pennsylvania. 574 pp.
Otte, L. J. 1981. Origin, development and maintenance of
pocosin wetlands of North Carolina. Unpublished report
to the North Carolina Natural Heritage Program. North
Carolina Department of Natural Resources and Comjnunity
Development, Raleigh, North Carolina. 51 pp.
Perkin-Elmer Corporation. 1976. Analytical methods for
atomic absorption spectrophotometry. Environm.ental
Section, Publication No. 303-0152. Perkin-Elmer
Corporation, Norwalk, Connecticut. 4 p.
121
Puustjarvi V. 1956. On the cation exchange capacity of
peats and on the factors of influence upon its forma-
tion. Acta Agriculture Scandinavia 6:410-449.
Reid G. K. and R. D. Wood. 1976. Ecology of inland waters
and estuaries. 2nd edition. D. Van Nostrand Company,
New York, New York. 485 pp.
Richardson, C. J. 1981. Pocosins: ecosystem processes and
the influence of man on system response. Pages 135-151
in C. J. Richardson, editor. Pocosin wetlands.
Hutchinson Ross Publishing Co., Stroudsburg, Pennsyl-
vania. 364 pp.
Richardson, C. J., R. Evans, and D. Carr. 1981. Pocosins:
an ecosystem in transition. Pages 3-19 C. J.
Richardson, editor. Pocosin wetlands. Hutchinson Ross
Publishing Co., Stroudsburg, Pennsylvania. 364 pp.
Richardson, C. J. 1983. Pocosins: vanishing wastelands or
valuable wetlands? Bioscience 33:626-633.
Scheiner, Dora. 1976. The indophenol spectrophotometric
method for ammonium. Water Research 10:31-36.
Schlesinger, W. H. and J. M. Melack. 1981. Transport of
organic carbon in the world's rivers. Tellus
33:172-187.
Sears, M. N. 1966. A floristic study of the limestone
along Island Creek in Jones County, North Carolina.
Unpublished Masters thesis. Department of Botany,
University of North Carolina at Chapel Hill, Chapel
Hill, North Carolina. 96 pp.
Shacklette, H. T. and J. G. Boerngen. 1984. Element
concentrations in soils and other surficial m.aterials
of the counterminous United States. United States
Geological Survey Professional Paper 1270, United
States Government Printing Office, Washington, District
of Columbia. 105 pp.
Sharitz, R. R. and J. W. Gibbons. 1982. The ecology of
southeastern shrub bogs (pocosins) and Carolina bays:
a community profile. United States Fish and Wildlife
Service, Division of Biological Sevices, Washington,
District of Columbia. FWS/OBS-82-04. 93 pp.
Simmons, C. E. and M. C. Aldridge. 1980. Hydrology of the
Chicod Creek basin. North Carolina, prior to channel
improvements. United States Geological Survey Open-
File Report 80-680, United States Geological Survey,
Raleigh, North Carolina. 27 pp.
122
Simmons, C. E. and R. C. Heath. 1979. Water-quality
characteristics of streams in forested and rural areas
of ^7orth Carolina. United States Geological Survey
Water-Resources Investigations 79-108, United States
Geological Survey, Raleigh, North Carolina. 49 pp.
Skaggs, R. W., J. W. Gilliam, T. J. Sheets, and J. S.
Barnes. 1980. Effect of agricultural development on
drainage waters in the North Carolina Tidewater region.
Report No. 159, Water Resources Research Institute of
the University of North Carolina, Raleigh, North
Carolina. 274 pp.
Snyder, J. R. 1977. Report on the vegetation and site
factors of the Pond Pine-Pocosin Research Natural Area,
Croatan National Forest. Submitted to the Southeastern
Forest Experiment Station, Asheville, North Carolina.
6 2 pp.
Snyder, J. R. 1978. Analysis of vegetation in the Croatan
National Forest, North Carolina. M.S. thesis. Univer-
sity of North Carolina, Chapel Hill, N.C. 151 pp.
Strickland, J. D. H. and T. R. Parsons. 1972. A practical
handbook of seawater analysis. Journal of Fisheries
Research Board of Canada, Bulletin 167:1-310.
Stumm, W. and J. J. Morgan. 1981. Aquatic chemistry. 2nd
edition. Wiley-Interscience, New York, New York. 780
pp.
Sumsion, C. T. 1970. Geology and groundwater resources of
Pitt County, North Carolina. U.S. Geological Survey
Groundwater Bulletin No. 18, Raleigh, North Carolina.
7 5 pp.
United States Geological Survey (U.S.G.S.). 1979. Water
resources data for North Carolina, water year 1978.
United States Geological Survey Water-Data Report
NC-78-1. Raleigh, North Carolina. 515 pp.
United States Geological Survey (U.S.G.S.). 1980. Water
resources data for North Carolina, water year 1979.
United States Geological Survey Water-Data Report
NC-79-1. Raleigh, North Carolina. 541 pp.
Verry, E. S. 1975. Streamflow chemistry and nutrient
yields from upland peatland watersheds in Minnesota.
Ecology 56:1149-1157.
123
Watkins, S. A. and C. E. Simmons. 1984. Hydrologic
conditions in the Chicod Creek Basin, North Carolina,
before and during channel modifications, 1975-81.
United States Geological Survey Water-Resources
Investigations 84-4025, United States Geological
Survey, Raleigh, North Carolina. 36 pp.
Wetzel, R. G. 1983. Limnology. 2nd edition. CBS College
Publishing, Philadelphia, Pennsylvania. 858 pp.
Wetzel, R. G. and G. Likens. 1979. Limnological analyses.
W. B. Saunders Co., Philadelphia, Pennsylvania. 357
pp.
Wharton, C. H. and M. M. Brinson. 1979. Characteristics of
southeastern river systems. Pages 32-40 R. R.
Johnson and J. F. McCormick, technical coordinators.
Strategies for protection and management of floodplain
wetlands and other riparian ecosystems. Forest Service
General Technical Report WO-12, United States Depart-
ment of Agriculture, Washington, District of Columbia.
Wharton, C. H., W. M. Kitchens, and T. W. Snipe. 1982. The
ecology of bottomland hardwood swamps of the southeast:
a community profile. FWS/OBS-81-37, United States Fish
and Wildlife Service, Biological Services Program,
Washington, District of Columbia. 133 pp.
Williams, G. P. 1970. The thermal regime of a Sphagnum
peat bog. Pages 195-200 Proceedings of the Third
International Peat Congress, Quebec, Canada.
Wilson, K. A. 1962. North Carolina wetlands: their
distribution and management. North Carolina Wildlife
Resources Commission, Raleigh, North Carolina. 169 pp.
Winner, M. D., Jr. and C. E. Simmons. 1977. Hydrology of
the Creeping Swamp watershed. North Carolina, with
reference to potential effects of stream channel!-
zation. United States Geological Survey Water Re-
sources Investigations 77-26. United States Geological
Survey, Raleigh, North Carolina. 54 pp.
Yarboro, L. A. 1983. The influence of hydrologic
variations on phosphorus cycling and retention in a
swamp stream ecosystem. Pages 223-245 dm T. D.
Fontaine, III and S. M. Bartell, editors. Dynamics of
lotie ecosystems. Ann Arbor Science Publishers, Ann
Arbor, Michigan.
Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall,
Inc., Englewood Cliffs, New Jersey. 620 pp.
APPENDIX
Table 1 Annual arithmetic mean concentrations (A), with standard errors, and weighted
mean concentrations (W) for nitrogen and phosphorus (in mg/1) from September
1978 through August 1979.
Station Mean NO3-N NHA,-N NO^-N DON PON TP FTP FRP PP
EB A 0.12+0.02 0.03+0.005 0.001+0.0002 0.90+0.02 0.06+0.01 0.02in.009 0.0110.002 0.0110.001 0.0110.003
W 0.23 0.02 0.000 0.91 0.09 0.01 0.01 0.01
WB A 0.09±0.01 0.03±0.002 0.001+0.0001 0.95+0.03 0.05+0.01 0.02+0.002 0.01+0.002 0.01+0.001 0.003+0.001
w 0.10 0.02 0.000 0.92 0.06 0.01 0.01 0.01
BL A 0.12+0.02 0.09+0.009 0.001+0.0002 0.56+0.05 0.08+0.02 0.09+0.002 0.0310.002 0.0210.002 0.0110.002
W 0.12 0.03 0.001 0.52 0.11 0.03 0.02 0.02
HI A 0.25+0.02 0.03+0.003 0.003+0.0009 0.99+0.09 0.07+0.02 0.17±0.02 0.1510.01 0.1910.01 0.0210.01
W 0.21 0.09 0.003 0.99 0.07 0.13 0.13 0.11
CL A 0.25+0.03 0.25+0.07 0.002±0.0003 1.17+0.12 0.29+0.05 0.19+0.03 0.1010.03 0.0810.02 0.0910.02
IS A 0.08+0.01 0.01+0.002 0.001+0.0001 0.26+0.03 0.08±0.02 0.11+0.01 0.0910.01 0.0810.01 0.02+0.01
W 0.07 0.02 0.001 0.33 0.09 0.07 0.06 0.05
CS A 0.24±0.07 0.07+0.03 0.002+0.0009 0.65+0.05 0.38±0.19 0.16+0.09 0.08+0.01 0.0510.01 0.0810.09
W 0.22 0.03 0.001 0.97 0.13 0.05 0.09 0.03
PA A 0.23+0.0A 0.03+0.01 0.002+0.0009 0.51+0.02 0.19+0.05 0.06+0.01 0.0910.009 0.0210.003 0.0210.01
W 0.A4 0.03 0.002 0.59 0.05 0.07 0.05 0.03
CH A 1.20+0.19 0.56+0.16 0.03+0.01 0.82+0.11 0.13+0.03 0.91+0.05 0.3310.09 0.2610.03 0.0810.1
W 1.19 0.92 0.02 0.68 0.07 0.32 0.26 0.21
GR A 1.A7+0.09 0.20+0.05 0.01+0.001 0.51+0.07 0.17+0.03 0.1610.03 0.0910.02 0.0710.02 0.0710.02
W 1.95 0.20 0.006 0.59 0.22 0.21 0.12 0.08
TR A 0.59±0.09 0.19+0.02 0.01+0.001 0.95+0.05 0.19±0.03 0.20+0.03 0.1110.01 0.0810.005 0.0910.02
W 0.61 0.10 0.005 0.91 0.11 0.18 0.10 0.07
TS A 0.31+0.09 0.10±0.03 0.009±0.001 0.92+0.10 0.36+0.13 0.3910.05 0.2310.09 0.1610.03 0.1110.02
Table 2. Annual arithmetic mean concentrations (A), with standard errors, and weighted
means concentrations (W) for total organic carbon, dissolved organic carbon,
calcuim, magnesium, sodium, potassium and iron (in mg/1) from September 1978
through August 1979.
Station Mean TOC DOC Ca Mg Na K Fe
EB A 22,.8+0.87 21..8±1.05 2. 01+0.,22 0.,69+0,,06 3.,31+0.,10 0,.35+0,.04 1,,21+0..17
W 19..12 17.,39 1. 28 0..98 3..57 0..46 0..52
WB A 28..12+1. 15 27,,1 + 1.:20 2. 14+0,.33 0.,47+0,.02 2,,78+0,,05 0.,27+0,.03 0,.72+0,,06
w 23,, 75 22..50 1. 59 0..54 2,.88 0.. 30 0..55
BL A 28..47±1. 34 27.,57 + 1 .54 15. 26+3,.26 1..16+0..17 4,.00+0,,24 0..63±0..12 0,.79±0. 07
W 25..79 25.,56 7. 50 0,.87 3..44 0.,41 0,.50
MI A 20..23+1. 79 19..29+1 .68 26. 46+3.,47 1,.42+0.,11 4,.98+0,.29 0..92±0,. 12 0,,63+0..07
W 24.,21 23..19 16. 79 1,,14 4.,32 0.,68 0,.67
CL A 54,.63±4. 36 50,,67 + 3 .96 1. 03±0..07 1,,04±0,.04 3..23+0,,14 1..23±0,.09 0.,41+0.,06
IS A 13..44±1. 31 13..23+1 .25 42. 96+6,.90 1,,41+0,,11 3.,98+0,,10 0,.94+0..12 0.,55+0,.06
w 17.,94 17,.16 17. 20 0,.95 3..69 0..58 0.,56
CS A 18,,74±1. 24 16,. 46+0 .91 6. 27+0..72 1..60+0,,16 4.,28+0..30 2..46+0.,31 3..08+0..86
W 15..04 14.,05 5. 52 1,,53 4,,62 ] .,56 0..75
PA A 14..85+0. 68 13,.53+0 .62 9. 57±1,,21 1,.88+0,.16 4.,91+0..28 2,,40±0,,23 1..52±0,,17
W 16..05 13..94 5. 19 1,.32 3..76 2..07 1,.50
CH A 14,.81+0. 73 13..41+0 .62 12. 05±1..17 2,.48+0,,23 6,.19+0..60 7,,22+1..14 1..32+0.,14
W 15..25 14,,32 7. 97 1..66 3..95 3..49 1..56
GR A 9..65+0. 90 7..57+0 .59 11. 04 + 0..59 2,,57+0,. 11 7..46+0..76 3,,75+0,,33 1.,50+0.,31
W 13.,80 10.,84 7. 78 1 ,.96 4,.86 3..30 1.,92
TR A 10,.n±o. 52 8..66+0 .52 6. 61+0,,36 2,.08+0,.12 7..36+0.,89 2,.72+0,.18 1..87±0.,20
W 10..97 10,.41 5..32 1,.78 4,,77 2.,44 2..68
TS A 33,.30+4. 47 27,.91±3 .36 10. 56±1 ,.21 4..69+0,.79 20,.23+3.,83 2.,85+0,.12 10,,00±2,,16
126
Table 3 . Annual arithmetic mean concentrations (A), with standard errors, and weighted
means concentrations (W) for pH, total alkalinity (in meq/1), conductivity (in
umho/cm), temperature, dissolved oxygen (in mg/1), biochemical oxygen demand (in
mg/1) and discharge (in ra^/sec) from September 1978 through August 1979.
Station Mean pH TA COND “C “2 BODj Q
EB A 4.50+0.14 0.0210.01 51.1812.70 15.6711.16 6.6710.19 0.7510.09 0.5710.15
w 3.96 0.000 63.35 10.39 7.01 0.50
WB A 4.35+0.14 0.0410.02 46.8012.03 14.8511.19 6.2610.33 0.6010.07 0.6110.15
w 3.97 0.001 49.66 10.17 7.38 0.40
BL A 5.93+0.20 0.5810.16 85.87112.27 14.6511.19 7.3310.32 0.8410.08 0.5610.17
W 5.33 0.08 53.76 10.99 8.10 0.80
MI A 6.84+0.10 0.8810.16 130.78113.27 14.1811.21 7.0510.32 1.0510.14 1.1110.18
W 6.59 0.51 91.95 13.45 7.57 1.11
CL A 3.7110.05 O.OllO.Ol 83.016.50 15.8311.37 5.6310.48 1.8310.37
IS A 7.28+0.11 1.7210.29 187.71120.93 13.74+1.05 8.3910.18 0.8810.11 0.3710.07
u 6.94 0.78 108.05 13.43 8.42 1.06
CS A 5.8310.10 0.3210.17 69.0015.59 15.6211.36 5.5610.60 2.2910.40 3.59+2.18
w 5.08 0.05 74.55 8.84 6.84 0.89
PA A 6.3810.06 0.3910.07 87.7315.76 15.5611.31 5.6010.50 1.6510.34 0.6610.21
w 6.08 0.13 58.92 15.20 6.98 1.36
CH A 6.4810.06 0.3910.05 116.3819.85 15.3611.29 6.3010.54 2.3110.42 1.78+0.51
u 6.29 0.21 82.34 14.76 7.34 2.02
OR A 6.1610.10 0.2410.07 124.2515.28 16.2811.23 9.0110.29 1.3510.31 2.6010.82
w 5.85 0.10 92.09 15.75 7.65 1.67
TR A 6.6410.08 0.3610.03 89.7914.59 16.8511.49 8.5210.39 1.5510.17 99.22119.57
W 6.41 0.25 71.73 13.82 8.32 1.27
TS A 6.2610.07 0.4510.06 163.90120.02 14.4211.38 3.5310.48 4.1110.51
127
Table 4. Linear correlations (r) between chemical and physical factors for East Brice
Creek (EB). Correlations are shown only if P < 0.002.
NOj-N m^-N NO^-N DCNKH TPFTPFRP PPTnCIXKPOCCal^NaK FepHTA ON) “C 0^ KB^ CTHT Q
NOj-N 0.80 0.80 0.65 0.76 0.64
NH,-N 0.66 0.70 0.62 0.67 0.70 0.65 0.63 0.65A
NO^-W 0.68 0.74 0.61 0.62
IXU
P(W 0.80
TP 0.66 0.68 0.75 0.91 0.85 0.83 0.68 0.76 -0.54
FTP 0.70 0.74 0.75 0.88 -0.68 0.82 0.81 0.74 0.69 -0.68 0.72 -0.57
FPP 0.62 0.61 0.88 0.60
PP 0.91 0.80 0.63 0.58
Tu; o.% -0.68
noc 0.96 -0.77 -0.66 0.65 -0.61
POC 0.80
Ca 0.85 0.80 0.86 0.75 0.66 0.70
Mg 0,65 -0.68 -0.64 -0.82 -0.77 0.77 -0.66 -0.66 0.93 -0.71 -0.84 0.71
Na -0.68 0.77 0.70
K 0.67
'
Fe 0.70 0.62 0.83 0.82 0.73 0.63 0.85 -0.66 0.90 0.85 0.74 -0.65 0.78 -0.59
pH 0.65 0.68 0.81 0.66 0.75 -0.66 0.90 0.91 -0.62 0.69 -0.61 0.77
TA 0.63 0.76 0,7A 0.58 0.66 0.85 0.91 -0.60
awD 0.76 -0.70 -0.66 0.93 0.70 -0.62 -0.71 -0.80 0.66
“c 0.69 0.60 -0.71 0.74 0.69 -0.71 0.74 -0.66
°2 -0.68 -0.62 -0.65 -0.61 -0.00
BCl)^ 0.65
STirr 0.72 0.73 0.69 0.65 0.70 0.84 0.78 0.77 -0.80 0.74 -0.86
Q 0,64 -0.62 -0.61 0.71 -0.59 0.66 -0.66 -0.86
128
Table 5 Linear correlations (r) between chemical and physical factors for West Brice
Creek (WB). Correlations are shown only if P < 0.002.
NO^-N NH^-N NO^-N IXW rON TP FTP FKP PP 1Ď1C rxr POC Cn Mg Na K Fe pH TA OWD *’C 0^ fTlT Q
NO^-N
NH4,-N
-0,58
NO^-N 0.77 0.80 0.74 n.67 0.60
IXK
PCN
TP 0.77 0.74 0.60 0.57
FTP 0.80 0.74 0.71
FKP 0.74 0.60 0,71
PP 0.59
TOC n.97 -0.67 -0.59
IXIC 0.97 -0.65 -0.60
Ca 0.67 0.R6
Mr 0.70
Na -0.67 -0.65 0.70
K
Fe 0.57 0.59 -0.7^4
pH 0.60 0.86 0.68
TA
COND
"C 0.59 -0.71 0.64 -0.61
-0.58 -0.7/i -0.71 -0.63
5
ÎTKT 0.68 O.W, -0.63 -0.85
Q -0.59 -0.60 -0.61 -0.85
129
Table 6. Linear correlations (r) between chemical and physical factors for Black Swamp
(BL). Correlations are shown only if P < 0.002.
NO^-N NO^-N PfN PTN TP FTP FKP PP TOC IXIC POC Cn Mp, Nn K Fe pH TA ÍTM) "C STtiT 0
NO3-N
NHA,-N
Ntl^-N 0.59 0.71 O.riO
DON
PON
TP 0.72 0.A6 0.60 0.64 -0.60
FIP 0.59 0.72 0.78 0.73 0.74 O.iV. 0.60 0.57 0.69
FPP 0.71 0.66 0.78 0.75 0.67 0.61
PP
TOC 0.57 0.94
ixx: 0.72 O.V. -0.57
POC
Cn 0.98 o.% 0.93 0.75 0.98 0.97
0.98 o.% 0.94 0.68 0.% 0.95
Na 0.% 0.% 0.9? 0.67 0.94 0.95
K 0.93 0.94 0.9? 0.70 0.93 0.88 -0.65
Fe 0.60 0.60 0.73 0.75 0.70 0.59 0.80 -0.62 0.70
pH 0.64 0.74 0.75 0.68 0.67 0.70 0.70 0.8? 0.75 -0.62 0.69
TA 0.64 0.98 0.% 0.94 0.93 0.59 0.82 0.97 -0.64 0.61
COND 0.60 -0.57 0.97 0.95 0.95 0.88 0.75 0.97
“C 0.57 0.67 0.80 -O.fO 0.58
'’2 -0.60 -0.65 -0.6? -0.62 -0.6/. -0.60
Bfin^
^IHT 0.69 0.61 0.70 0.69 0.61 0.58 -0.84
-0.8/i
130
Table 7. Linear correlations (r) between chemical and physical factors for Mill Creek
(MI). Correlations are shown only if P < 0.002.
N03.,-N Mi,-Na NOj-il DCK PON TP FTP FPP PP TO DOC per Ca Mg Na K Fe pi! TA OCKD ®C ”2 BfD^ smrr Q
NO^-N 0.68
NH,-N \ 0.59 0.59 -0.66 -0.70 -0.70 -0.63
NOj-N 0.66
IKK 0.68 0.71 -0.71 -0.65
PCW
TP 0.96 0.87 0.81 0.83 0,85 0.77 0.86 0.73 0.58 0.71 -0.71
FTP o.% 0.91 0.62 -0.66 -0.67 0.86 0.R9 0.79 0.87 0.77 0.61 0.80 -0.69
FPP 0.87 0.91 -0.69 -0.71 0.85 0.90 0.80 0.86 0.76 0.73 0.85 -0.76 -0.59
PP 0.66 0.81 0.62
TO 0.59 0.68 -0.66 -0.69 0.99 -0.88 -0.85 -0.80 -0.75 -0.75 -0.88
DOC 0.59 0.71 -0.66 -0.71 0.99 -0.90 -0.86 -0.81 -0.7A -0.76 -0.90 0.59
POC
Ca -0.66 -0.71 0.83 0.86 0.85 -0.88 -0.90 0.95 0.92 0.86 0.81 0.89 0.97 -0.70 -0.75
0.85 0.89 0.90 -0.85 -0.86 0.95 0.96 0.88 0.78 0.83 0.96 -0.71
Na -0.70 0.77 0.79 O.RO -0.80 -0.81 0.92 0.96 0.86 0.ÍW 0.93 0.87
K 0.86 0.87 0.86 -0.75 -0.76 0.86 0.88 0.86 0.66 0.75 0.81 -0.74
Ko
pH 0.73 0.77 0.76 0.81 0.78 0.86 0.66 0.80 0.80
TA 0.58 0.61 0.73 -0.75 -0.76 0.89 0.83 0.93 0.75 0.80 0.83 -0.61
(TKD -0.63 -0.65 0.71 0.80 0.86 -0.88 -0.90 0.97 0.96 0.87 0.81 0.80 0.83 -0.65 -0.75
”c 0.68 -0.59
-0.71 -0.69 -0.76 -0.70 -OJt* -0.65 -0.59
BOD^
SWT -0.60
Q -0.59 0.59 -0.75 -0.71 -0.61 -0.75 -0.60
131
Table 8. Linear correlations (r) between chemical and physical factors for Catfish Lake
drainage canal (CL). Correlations are shown only if P < 0.002.
NO^-N NH,-N NO^-N DCN PCN TP FTP FPP PP TOC DOC Mg Na K Fe pH TA OJ® 'C 0^ BOD^ SlWr
N03-N -0.59
'«4-” 0.70 0.83 0.83 0.62
NO^-N 0..58 0.59
0.57
DCN 0.59 0.64 0.71 0.71 0.66 0.71 0.67 -0.63
PCN 0.59
TP -0.59 0.70 0.64 0.86 0.B6
FTP 0.83 0.71 0.85 0.99 0.58 0.68
FRP 0.83 0.71 0.86 0.99 0.58 0.68
PP
TOC 0.58 0.56 0.58 0.58 0.99 0.63 0.77 -0.77 0.91
DOC 0.62 0.59 0.71 0.68 0.68 0.99 0.59 0.77 -0.77 0.88
POC 0.59
Ca
Mg
Na -0.70
K
Fe 0.63 0.59 0.70
pH -0.72
TA
at® -0.72
"c 0.57 0.67 0.77 0.77 -0.77 0.74
°2 -0.63 -0.77 -0.77 -0.70 -0.77 -0.77
BCD5?
smr 0.91 0.88 0.70 0.7A -0.77
132
Table 9. Linear correlations (r) between chemical and physical factors for Island Creek
(IS). Correlations are shown only if P < 0.002.
NH^-N DCN PON TP FIT FPP PP ITT DOC P(X Ca Mg Na K Fe pll TA OOND “C 0^ STIff Q
NOj-N
NH,-N
NO^-W
DON -0.63 -0.61 -0.62 0.76 0.67 -0.72 -0.75 -0.70 -0,69
PON
TP -0.63 0.78 0.79 0.70 -0.65 -0.64 0.81 0.75 0.61 0.77 0.76 0.85 0.72 -0.69
FTP -0.61 0.70 0.96 -0.62 0.88 0.77 0.77 0.83 0.77 0.91 0.80 -0.72
FFP -0.62 0.79 0.% -0.67 -0.60 0.92 0.82 0.79 0.83 0.84 0.% 0.82 -0.75
PP 0.70
TOC 0.76 -0.65 -0.62 -0.67 0.92 -0.82 -0.90 -0.65 -0.74 -0.57 -0.66 -0.78 0.71
DOC 0.67 -0.64 -0.60 0.92 -0.75 -0.85 -0.69 -0.72 -0.70 0.66
POC 0.63
Ca -0.72 0.81 0.88 0.92 -0.82 -0.75 0.94 0.88 0.82 0.83 0.97 0.81 -0.75
Mg -0.75 0.75 0.77 0.82 -0.90 -0.85 0.94 0.90 0.73 0.77 0.90 0.77 -0.82
Na -0.62 -0.65 -0.69
K -0.70 0.63 0.77 0.77 0.79 -0.74 -0.72 0.63 0.88 0.90 0.77 0.63 0.84 0.67
Fe 0.57
pH -0.78 0.76 0.83 0.83 -0.57 0.82 0.73 0.77 0.70 0.82 0.71 -0.62
TA 0.77 0.84 -0.66 0.83 0.77 0.63 0.70 0.82 0.72 -0.67
OONP -0.69 0.84 0.91 0.96 -0.78 -0.70 0.97 0.90 0.8/i 0.82 0.82 0.83 -0.79
“C 0.57 -0.70
°2 -0.70
STwr 0.72 0.80 0.82 0.81 0.77 0.68 0.71 0.72 0.83 -0.79
Q -0.69 -0.72 -0.75 0.71 0.66 -0.75 -0.8? -0.62 -0.67 -0.79 -0.79
133
Table 10. Linear correlations (r) between chemical and physical factors for Creeping Swamp
(CS). Correlations are shown only if P ^ 0.002.
N03,-N NILA-N N02.,-H D(U mN TP FTP. FTIP FPTOCDOCPnCCaMgNaK
Fp pH TA OfT® '’C 0^ ^fT Q
NO^-N 0.83
NH,-N 0.73 0.69
i*
NOj-N 0.83 0.66 0.57
PON 0.66 0.69 0.57 0.59 -0.66
PON 0.95 0.97 0.64 0.96 0.85
TP 0.95 o.% 0.75 0.95 0.% 0.90
FTP 0.73 0.69 0.90 0.61
FPP 0.69 0.57 0.90
PP 0.97 0.96 0.70 0.97 0.97 0.84
TOC 0.64 0.75 0.70 0.74 0.64 0.82 0.58 0.65 0.80
DOC 0.59 0.74 0.83 -0.75
POC 0.95 0.97 OM 0.94 0.85
Ca 0.97 0.80 0.70 0.93
0.97 0.89 0.98
Na 0.80 0.89 0.91
K 0.70 0.7)
Fe 0.94 0.96 0.97 0.82 0.94 0.89 -0.61
pH 0.61 0.58 0.59 -0.69
TA
OTND 0.93 0.98 0.91
0.65 0.83 0.59 -0.72
-0.75 -0.72
BCDg 0.85 0.90 0.84 0.80 0.85 0.71 0.89 -0.69
snrr -0.61 -0.69 -0.69 0.76
q 0.76
134
Table 11. Linear correlations (r) between chemical and physical factors for Palmetto Swamp
(PA). Correlations are shown only if P _< 0.002.
NOj-N NO^-^ DCN PON TP FIT FHP PP TOC ar POC Ca Mr Nn K Fe pH TA aWH “C 0^ BTT^ Sim Q
NOj-N -0.63 0.66 0.66
NH,-N 0.79 0.61 0.(A
NOj-N 0.79 0.59 0.66 0.64 0.67 0.66
DON
PON 0.76 0.65
TP 0.59 0.85 0.70 0.92 0.72 0.62 0.89 0.61 -0.66 0.74
FTP 0.66 0.85 0.87 0.57 0.68 0.83 0.68
FPP 0.64 0.70 0.87 0.57 0.67
FP 0.76 0.92 0.57 0.61 0.76 0.75 -0,69 0,82
TOC 0.72 0.68 0.57 0.61 0.88 O.M
DOC 0.62 0.88
POC 0.76 0.65
Ca 0.90 0.6A 0.65 0.64 0.73 -fl.65
Mg 0.90 0.83 0.69
Na 0.64 0.83 0.68
K 0.71
Fc 0.61 0.67 0.89 0.83 0.67 0.75 0.64 0,80 -0.71
pll -0.63 0.65 0.78 0.69 -0.73 -0.86
TA 0,6A 0.78 -0.72 -0.86
OnND 0.73 0.69 0.68 0.69 0.68 -0.77 -0.62
"c 0.64 0.66 0.61 0.68 0.80 -0.76
"2 -0.66 -0.69 -0.71 -0.73 -0.72 -0.76 0.69
BCD 0.65 0.74 0.82 0.65
SIHT 0.66 -0.65 -0.86 -0.86 -0.77 0.69 0.73
q 0.66 -0.67 0.73
135
Table 12. Linear correlations (r) between chemical and physical factors for Chicod Creek
(CH). Correlations are shown only if P ^ 0.002.
NOj-N DCNPCIN TPFTPIW PPTOCDOCPIXCaUgNaK Fe pH TA OnND ”C 0^ BdD^ ?IHr Q
0.67 0.71
NH,-N 0.67 0.784| 0.60
N0,-N 0.71 0.78
DfN
P(TJ 0.57 0.62 0.70 0.71 0.71 -0.57 0.64
IP 0.98 0.90 0.79 0.60 -0.65
FTP 0.98 0.89 0.65 0.57 0.58 -0.62
FRP 0.57 0.90 0.89 0.68 -0.64
PP 0.62 0.79 0.65 0.68 0.61 -0.57
TOC 0.R8 0.65 0.62
noc 0.88 0.72
POC 0.57 0.62 0.70 0.71 0.71 -0.57 0.64
Ca 0.70 0.70 0.99 0.77 0.87 0.77 0.95 0.67
Mg 0.71 0.71 0.99 0.77 0.87 0.73 0.95 0.70
Na 0.77 0.77 0.84 0,75
K 0.71 0.71 0.87 0.87 0.8A 0.77 0.88 -0.69
Fe 0.65
pH 0.68 0.74 0.59 -0.62
TA 0.60 0.58 0.77 0.73 0.77 0,67 -0.83 -0.85
aw 0.60 0.95 0.95 0.75 0.88 0.67 -0.60
"c 0.65 0.72 0.65 -0,69
Oj -0.57 -0.65 -0.62 -0.64 -0.57 -0.57 -0.83 -0.69 0.74
B(D^ 0.66 0.62 OM 0.67 0.70
SIHT -0.69 -0.62 -0.85 -0.60 0.74
Q 0.75
136
Table 13. Linear correlations (r) between chemical and physical factors for Grindle Creek
(GR). Correlations are shown only if P < 0.002.
NDj-N NH,-N NO^-N DCW PON TP FTP FPP PP TOC noc POC Ca Mg Na K Fp pH TA OHNO “C Oj BOD^ ^rrQ
NO^-N
NH^-N 0.70 0.62 0.60 0.61
NOj-N 0.65 0.63 0.74 0.72
DCW 0.70 0.65 0.70 0.91 0.92 0.65 0.67 0.64
PON 0.62 0.65 0.73 0.59 0.60 0.76
TP 0.67 0.70 0.73 0.77 0.73 0.75 0.70 0.73 0.90 -0.60 0.82
FTP 0.72 0.65 0.91 0.59 0.77 0.98 0.59 0.76 0.66 0,83
FPP 0.75 0.63 0.92 0.60 0.73 0.98 0.60 0.75 0.61 0.80
pp 0.75 0.71
TOC 0.70 0.88 -0.78 -0.74 0.73 -0.86 -0.71 0.73 -0.60 0.58
DOC 0.88 -0.76 -0.68 -0.8? -0.75 -0.70 0.71
POC 0.62 0.65 0.73 0.59 0.60 0.76
Ca -0.78 -0.76 0.92 0.92 0.76 -0.74
Mg -0.74 -0.68 0.92 0.75 0.91 0.77 -0.74
Na 0.75
K 0.74 0.67 0.76 0.75 0.65
Fe 0.60 0.76 0.90 0.66 0.61 0.71 0.73 0.76 0.79
pH 0.61
TA
am -0.60 -0.86 -0.82 0.92 0.91 0.65 0.80 -0.78
“C 0.61
"2 -0.71 -0.75 0.65 -0.59
BCT)^ 0.61 0.72 0.64 0.82 0.83 0.80 0.73 0.65 0.79
^rr -0.60 -0.70 0.76 0.77 0.80 -0.99
Q 0.71 -0.74 -0.74 -0.78 -0.59 -0.99
LJ
Table 14. Linear correlations (r) between chemical and physical factors for Tar River
(TR). Correlations are shown only if P _< 0.002.
ND3,-N m.4-N N02,-N IXUPON IFFTPraP PPTOCnOCPOCCaMgNaK FepHXA aW) °C 0. BCD, STHT QÍ J
NOj-N
0.60 0.66 0.67 0.65
4
N02'^ 0.60 0.62 0.70 0.59 0.67 0.65 0.69
DfW 0.66
RW 0.63
TP 0.62 0.91 0.98
FIP 0.72 0.91 0.68 0.82
rap 0.59 0.68 0.57 -0.60
pp 0.98 0.82
TOC 0.75 0.62
DOC 0.75 -0.75 -0.64 0.76 -0.78 -0.61 0.69
POC 0.63
Ca -0.75 0.82 -0.70 0.76 0.68 -0.78
Mg -0.64 0.82
Na 0.67 0.63 0.63 0.63 -0.66 0.88 0.87
K 0.63
Fe 0.62 0.76 -0.70 -0.66 -0.67 -0.82 0,81
pH 0.75 0.61 0.63 -0.62
TA 0.67 0.88 -0.67 0.75 0.81 -0.72
OCN) 0.65 0.65 -0.78 0.76 0.87 -0.82 0.61 0.81 0.61 -0.80
“C 0.69 0.57 0.62
«2 -0.60
0.63 0.62
SIHT -0.61 0.68 0.61 -0.70
0.69 -0.78 0.81 -0.62 -0.72 -0.80 -0.70
138
Table 15. Linear correlations (r) between chemical and physical factors for Tar Swamp
(TS). Correlations are shown only if P Ł 0.002.
ND3,-N m,4-N N02,-N IXKFCHTPFn>EWPPTOCDOCPOCCaMgNaK FepllTA OOB °C 0ż, HI), !?IWrJ
NOj-N
4
N02-*^
DCN 0.77 0.76 0.66 0.92 0.93 0.79 0.77 0.76 0.72 0.62 0.76 -0.82 0.87
PtW 0.69 0.62 0.75
TP 0.77 0.92 0.81 0.81 0.83 0.81 0.90 0.84 0.63 -0.75 0.68
FTP 0.76 0.92 0.92 0.60 0.83 0.77 0.66 -0.67
FRP 0.65 0.81 0.92 0.68 0.72 0.62
PP 0.69 0.81 0.69 0.82 0.84 -0.65
TOC 0.92 0.62 0.83 0.80 0.68 0.98 0.62 0.84 0.84 0.87 0.85 0.68 0.82 -0.79 0.84
DOC 0.93 0.81 0.83 0.72 0.98 0.81 0.82 0.88 0.82 0.81 -0.77 0.84
POC 0.69 0.62 0.75
Ca 0.79 0.84 0.81 0.93 0.77 0.69 0.88 0.77
Mg 0.77 0.84 0.82 0.93 0.84 0.81 0.73
Na 0.76 0.87 0.88 0.77 0.84 0.75 0.86 0.75
K
Fe 0.72 0.75 0.90 0.77 0.62 0.82 0.85 0.82 0.75 0.69 0.75 0.72 0.78 -0.68 0.68
pH
TA 0.62 0.84 0.66 0.84 0.68 0.72 -0.73
OM) 0.76 0.63 0.82 0.81 0.86 0.88 0.81 0.86 0.78 -0.65 0.80
"C
«2 -0.82 -0.75 -0.67 -0.65 -^.79 -0.77 -0.71 -0.68 -0.68 -0.73 -0.65 -0.80
BOD 0.87 0.68 0.84 0.84 0.71 0.77 0.73 0.71 0.68 0.80 -0.80
STifT
CaJ