Impacts of Phosphorus
on Streams
Wisconsin Department of Natural Resources
Bureau of Water Resources Management
April 1984
...
<3g»-p«=r
\ H*-
-------
IMPACTS OF PHOSPHORUS ON STREAMS
Steven E. Mace, Patrick Sorge, Timothy Lowry
Water Resource Management
Wisconsin Department of Natural Resources
P.O. Box 12436
Milwaukee, Wl 53212
April 1984
Final Report of the Phosphorus and High-Flow Field Studies
EPA Grant No. P0055420 01
Project Officer
Robert To I pa
U.S. EPA - Region V
230 South Dearborn St.
Chicago. IL 60604
-------
TABLE OF CONTENTS
ABSTRACT I
LIST OF FIGURES II
L IST OF TABLES | v
INTRODUCTION AND STUDY BACKGROUND
Phosphorus Water Quality Standards I
Sources of Phosphorus I
Stream Primary Producer Communities I
Impacts of Primary Producers on Stream Environments 2
Need For Study 3
Study Background 3
FIELD STUDY DESIGN 4
METHODS
Study Reach Selection 5
Intensive Study Reaches 5
Synoptic Survey Reaches 9
STUDY REACH CHARACTERISTICS
Reach Chemical Characteristics II
Reach Physical Characteristics 12
STREAM PRIMARY PRODUCERS
MACROPHYTES
IntroductIon 15
Results and Discussion 15
Stream Type DetermInatIons 15
Sediment Nutrients 18
Macrophyte Blomass and In-stream Nutrients 19
Macrophyte Tissue Nutrients and In-stream Nutrients 20
Macrophyte Blomass and Tissue Nutrients 22
Model Selection 23
Summary and Conclusions 26
PERIPHYTON
IntroductI on 27
Results and Discussion 28
Per Iphyton Blomass and In-Stream Nutrients 28
Perlphyton Tissue Nutrients and In-Stream Nutrients 29
PerIphyton Blomass and Tissue Nutrients 31
Other Aspects of Perlphyton Growth and Measurement 32
Summary 33
DIEL DISSOLVED OXYGEN STUDIES
IntroductIon 34
Methods of Estimating P. R. and K2 36
Box Studies 36
-------
Bottle Studies 36
Summary of Box and Bottle Study Results 38
Dlel Curve Analysis 39
Modeling Assumptions and Parameters 39
Single and Double-Station Analysis: Differences In Methodology and Purpose 41
Stream Areas Represented by Dlel Ana Iyses 42
Limit of Reach Length For the Double Station Differential Method 43
Maximum Attainable Dissolved Oxygen Deficit 43
Macrophytes and Reaeratlon 45
Deviations From Modeling Assumptions 47
Further Considerations Important to the Mode 11ng Process 48
Results and Imp 11 cat Ions of Dlel Analyses 49
Comparison of Double and Single-Station Data 54
Area Represented by Single-Station Coefficients 54
Comparison of Mean Single and Double Station Values for the Entire Survey Area .. 55
Double-Station Coefficients and "Average Double-Station" Coefficients 56
Integrated and Differential Coefficients For the Double-Station Technique 56
Discussion 50
Summary of Dlel Studies 51
Potential Method of Allocating Phosphorus 63
Cone I us I ons 53
SUMMARY AND CONCLUSIONS 65
LITERATURE CITED 69
APPENDIX I - Study Reach Descriptions 74
APPENDIX 2 - Stream Reach Species List 80
APPENDIX 3 - Methods for Evaluating Macrophyte Populations In Small Strean Systems 86
APPENDIX 4 - Die! Data Col lection and Analysis 89
-------
ACKNOWLEDGEMENTS
This report represents one element of a larger project. In addition to the authors, project
principals Included project coordinators Jeff Bode and Ouane Schuettpelz, Candy Schronk,
Lynn Persson, Mark Tusler and Barry O'Flanagan of the Wisconsin Department of Natural Resources,
and project officers Bob Tolpa, Madeline Lewis and Mike Phillips of EPA-Reglon V.
The authors thank the many people who contributed their time, efforts and creativity to com-
pletion of the field and laboratory work. Among these are Tim Babros, Dave Haselow,
Amy Ihlenfeld, Kevin Shaver, E. B., and Bob Schuknecht and George Bowman at the State Lab of
Hygiene.
We also thank the project's Advisory Committee and technical work groups for their assistance
and cooperation In the seemingly endless review process. So many contributed In many ways that
It Is Impossible to Individually acknowledge their efforts, but we collectively thank them all.
We are especially grateful to Dick Lathrop, Dr. Sanford Engel, Dr. Eugene Lange, R. Steve Grant,
Steve Skavronek and Dale Patterson for critical reviews and technical assistance.
-------
ABSTRACT
The purpose of this study was to define methods of dealing with phosphorus In setting appropri-
ate stream water quality goals or standards. In-stream and sediment nutrients were compared to
rooted plant and attached algae growth In southern Wisconsin streams, 1981 and 1982. The re-
sultant Impacts on stream die) dissolved oxygen (DO) characteristics were also Investigated.
Three empirical models describing macrophyte blomass, tissue phosphorus content and In-stream
phosphorus are presented. Results of the analyses suggest several different stream types, dif-
fering In the percent contribution of In-stream nutrients as opposed to sediment nutrients.
Stream perlphyton were also collected from glass slide and brick substrates. Models describing
brick perlphyton community blomass, tissue phosphorus content and In-stream phosphorus, similar
to macrophyte models, are also presented. Single-station and double-station dlel DO curve
analyses as well as light/dark productivity studies are compared to In-stream primary producer
blomasses. Maximum night-time DO deficit can be described as Respiration divided by Reaera-
tlon. This estimate, when combined with the ability to predict plant blomass from In-stream
nutrients using the primary producer blomass models may allow prediction of the Impact of
changing phosphorus concentrations on small stream dissolved oxygen minima. In addition to In-
vestigating the Impacts of phosphorus In small stream systems, the study also evaluates methods
of documenting phosphorus Impacts and recommends monitoring strategies.
-------
LIST OF FIGURES
Page
Figure I. Frequency Analysis of Macrophyte Occurrence on Bottom Substrate Sizes In
Type I and 11 Streams 16
Figure 2. Summer Maximum Stable Blomass vs. Mean Summer Phosphate -
Phosphorus Concentration 17
Figure 3. Regression Line of Summer Macrophyte Blomass on Mean Summer Phosphate -
Phosphorus Concentrations for Type I Streams 21
Figure 4. Regression Line of the Macrophyte Tissue Phosphorus-Concentration on the
Mean Summer Phosphate Concentration for Type I Streams 22
Figure 5. Regression Line of Summer Macrophyte Blomass on Macrophyte Tissue
Phosphorus Concentration for Type I & II Streams 24
Figure 6. Regression Line of Model II Including the Estimated Upper 95$
Confidence Limit 25
Figure 7. Regression Line of Monthly Mean Brick Rerlphyton Chlorophyll £_
on Monthly Mean Stream P04P 30
Figure 8. Regression Line of Monthly Mean Brick Perlphyton Tissue P on
Monthly Mean Stream P04P 3la
Figure 9. Regression Line of Monthly Mean Brick Perlphyton Chlorophyll on
Monthly Mean Tissue P 32
Figure 10. Path Diagram of Factors Which Regulate Stream DO 34
Figure II. The Effect of Oxygen Concentration on Plant Respiration Rates 38
Figure 12. Comparison of Single- and Double-Station DO Methods and Purpose 41
Figure 13. Attainment of EqulIIbrlurn DO Deficit 44
Figure 14. Illustration of BOD & Plant-Induced DO "Sags" 49
Figure 15. Theoretical & Observed Maximum Deficits for the Double-Station Method 50
Figure 16. Theoretical & Observed Maximum Deficits for the Single-Station Method 50
Figure 17. Double-Station Respiration Modeling and Plant Blomass Estimates 52
Figure 18. Single-Station Respiration Modeling and Plant Blomass Estimates 52
Figure 19. Double-Station Area! Photosynthesis Modeling & Plant Blomass Estimates 53
Figure 20. Single-Station Area! Photosynthesis Modeling & Plant Blomass Estimates 53
II
-------
LIST OF FIGURES (con't)
Figure 21. Average Differential Photosynthesis vs. Measured Differential
Photosynthesis Coefficients 57
Figure 22. Average Differential K2 vs. Measured Differential K2 Coefficients 57
Figure 23. Average Differential Respiration vs. Measured Differential Respiration
Coefficients 59
Figure 24. Comparison of Differential & Integrated Double-Station Respiration
Coefficients 58
Figure 25. Comparison of Differential & Integrated Double-Station Photosynthesis
Coefficients 59
Figure 26. Comparison of Differential and Integrated Double-Station Reaeratlon
Coefficients 59
Figure 27. Ashlppun River Example of Apparent Light Saturation 62
Figure 28. Bark River Example of Apparent Light Saturation 62
III
-------
LIST OF TABLES
Page
Table I. Water Chemistry Analyses 6
Table 2. Substrate Size Classes 8
TableS. Macrophyte Species Abundance Ratings 8
Table 4. Summary of Stream Reach Chemical Characteristics, 1981-1982 II
Table 5. Breakdown of Stream Reaches by Phosphorus and Nitrogen Characteristics 13
Table 6. Summary of Stream Reach Physical Characteristics 14
Table 7. Substrate Index - Substrate Type & Value Ranges 14
Table 8. Summary of Stream Reach Mapping Data 16
Table 9. Bulk Sediment Nutrient Concentrations 19
Table 10. Sediment Interstitial Water Nutrient Concentrations 19
Table II. Mean Summer Water Chemistry Parameter Values 2la
Table 12. Maximum Stable Summer Macrophyte Blcmass & Macrophyte Tissue Nutrient
ConcentratIons 21 b
Table 13. Brick Data used to Calculate Models IV, V & VI 28
Table 14. Correlation Coefficients for Water Chemistry & Perlphyton Blcmass 29
Table 15. Correlation Coefficients for Water Chemistry & Perlphyton Tissue Nutrients ... 31
Table 16. Correlation Coefficients for Perlphyton Tissue Nutrients & Blcmass 31
Table 17. Autotrophlc Index Values, as Estimated from Bricks and Perlphytometers 33
Table 18. Box Study Summary 36
Table 19. Box Study Results 37
Table 20. Time Required to Satisfy a Given Percentage of the DO Deficit 41
Table 21. Approximate Values of Manning's Roughness Coefficient 46
Table 22. Discharge, Velocity, Depth and Blcmass for POTW-Impacted & Non-Impacted Sites . 46
Table 23. K2 Values for June-August at POTW-Impacted & Non-Impacted Sites 46
Table 24. Dlel Modeling Coefficients for Single- and Double-Station Analyses 51
Iv
-------
LIST OF TABLES (continued)
Table 25. Single- and Double-Station Coefficients for the Bark River Wasteload 55
Table 26. Mean Modeling Coefficients and Standard Deviations for the Entire
Survey Area 56
Table 27. Comparison of Mean Double-Station Coefficients 60
-------
INTRODUCTION AND STUDY BACKGROUND
PHOSPHORUS WATER QUALITY STANDARDS
Any assessment of water quality must be
made relative to water quality guidelines
or criteria. Water quality standards and
allowable concentrations of most chemical
constituents have been developed and demon-
strated based on toxlclty or production of
conditions unfavorable to public health or
aquatic life. With the exception of the
elemental form, no such Federal water
quality standards have been set for phos-
phorus. Inability to define phosphorus as
a pollutant In the conventional sense (e.g.
toxlclty, human health hazard) has hindered
establishment of water quality phosphorus
standards and related wastewater effluent
limits. The U.S. EPA "Red Book" (1976)
does, however, recognize phosphorus as a
contributor to accelerated lake and stream
eutrophlcatlon, and suggests levels of
total and ortho-phosphorus which would slow
enrichment of surface waters.
SOURCES OF PHOSPHORUS
Phosphorus Is delivered to surface waters
from both point and non-point sources.
Generally, non-point source Inputs are
described by the total load of a particular
pollutant. In terms of annual loading,
storm events usually supply the greatest
amount of nutrients. Nutrient concentra-
tions as well as total loads surpass those
of low-flow conditions. Lake or watershed
management and planning are' usually based
on nutrient and sediment loadings. In most
lake assessments, loadings are the criteria
most often used to predict a lake's re-
sponse to changes In nutrient Inputs.
It Is unlikely, however, that the forms of
phosphorus In runoff are readily available
to stream primary producers. Streambed
scouring, light-limitation (Increased tur-
bidity), nutrient sorbtlon on suspended
solids and relatively short contact times
between nutrients and stream primary pro-
ducers would act to limit plant response to
storm-Induced nutrient delivery. Storm
flows may remove stream perlphyton and mac-
rophytes, also reducing potential for
nutrient uptake. In reach-specific
analyses, stream primary production Is
probably Influenced more by low-flow or
sustained nutrient concentrations they ex-
perience rather than seasonal or event
non-point source nutrient loadings.
Based on their ability to assimilate
wastes, streams have served as natural
treatment systems for wastewater. The
rate of enrichment of a stream varies with
the amount of dilution water available.
The rate of biological response to Intro-
duced nutrients. In turn. Is dependent on
physical factors such as light, substrate
type and stability, water temperature,
depth and velocity.
Point source problems have usually been
associated with Inadequately treated waste-
water or Inadequate stream dilution. Ex-
cessive loads of unstabl 11 zed waste
material results In lowered stream dis-
solved oxygen. Modern, efficient treatment
plants are designed to reduce the high oxy-
gen demand component, discharging miner-
alized, biologically available nutrients.
Specific water quality standards have been
applied to streams based on a particular
stream's natural low-flow potential, using
existing and potential aquatic life uses to
determine acceptable levels of a particular
waste discharged. Traditionally, the dis-
charge constituents of concern are dis-
solved oxygen, ammonia, BOD, residual
chlorine and solids.
STREAM PRIMARY PRODUCER COMMUNITIES
As the group directly able to use phos-
phorus, and responsible for many undesir-
able water quality conditions, primary pro-
ducers offer the best opportunity to
evaluate the Impacts of phosphorus In
stream environments. Streams are dynamic
systems that support an extremely complex
and variable biological community. Physi-
cal factors exert considerable Influence on
stream primary producers and modify this
-------
community's response to nutrient enrich-
ment. This makes measurement of stream
biological response to nutrients diffi-
cult. Boor definition of water quality
Impairments relative to system productivity
and nutrient Inputs also Impedes definition
of optimum or desirable nutrient levels.
PhytopIankton have frequently been used to
measure lake and large river response to
nutrient enrichment. Deep, slow-moving
stream reaches or backwater areas may sup-
port substantial short-term phytop lanktonic
populations, which are probably flushed out
by storm flows. While phytop Iankton may be
present In smaller streams, attached algae
are predominant. Suspended algae In
smaller rivers may actually represent de-
tatched perIphyton. Reliance on measure-
ment of phytopIankton production to assess
nutrients In small stream environments Is
therefore not dependable.
Because of their rapid turnover rates and
recovery from catastrophic events, pro-
duction of the perIphyton component has
frequently been used to study nutrients In
streams. Unlike macrophytes, this group
obtains all of Its nutrients from the
water, thus measurement of perlphyton to
assess the effects of nutrients Is not com-
plicated by estimating the contribution of
sediment nutrients to community growth.
A variety of methods and approaches have
been used In the assessment of perlphyton
communities. These Include blomass
measurements. Indicator species, community
diversity Indices, tissue nutrient content
and enzyme activity. Methods of measure-
ment and environmental response of primary
producers In streams have been reviewed as
part of this study (Mace, et al. 1983).
For most macrophytes to become dominant In
streams, the bottom substrate must be able
to support growth, and the nutrient supply
must be adequate. Other factors, such as
light, water depth and velocity must also
be conducive to growth at the proper time
of year. Unlike perlphyton, macrophyte
growth nutrients may be obtained, to
various degrees, from bottom sediments.
This last factor complicates somewhat the
use of macrophytes to determine the nutri-
ent status of a particular stream. Macro-
phytes are also susceptible to tangling and
breakage caused by stream current. Their
large size, abundance and ability for pro-
ducing "nuisance" conditions, however. In-
crease their desirability as study subjects.
IMPACTS OF PRIMARY PRODUCERS ON STREAM
ENVIRONMENTS
Macrophytes can, however, modify the stream
environment to make It more hospitable for
plant production. Primary producers Impact
the stream environment In a variety of
ways. Excessive macrophyte growth can
alter the stream channel by encouraging
sediment deposition. By Increasing sedi-
mentation and water depth, macrophytes
create more favorable conditions for
growth. Through channel modification, pei—
petuatlon of macrophyte communities Is
almost assured. This may lead to a point
where native fish and Invertebrate popula-
tions lose quality habitat.
Filling of stream channels by macrophytes
can cause flooding, navigational, aesthetic
and dissolved oxygen (00) problems. Macro-
phytes can, by retarding water flow and
Increasing the stream depth, lower stream
re-oxygenatlon rates. Coupled with plant
and animal respiratory oxygen demand, this
could lower the night-time 00 concentra-
tions below the desirable level for a
designated stream use.
Standing crop blomass does not necessarily
reflect actual growth and production within
a particular reach. Due to fragmentation
or other causes, blomass export results In
a substantial seasonal downstream loss of
perlphyton and macrophyte growth. Even If
macrophyte growth (standing crop) does not
cause severe problems within a particular
reach, export of produced material can pro-
vide a considerable load of organic, oxy-
gen-demanding materials to downstream lake.
Impoundment or riverine systems.
- 2 -
-------
NEED FOR STUDY
O'Shaugnessy and McDonnell (1973) state
that "trends to control discharges ...
clearly dictate the need to establish ef-
fective procedures for Identifying those
elements primarily responsible for acceler-
ating the rate of eutrophlcatlon In a given
situation." The level of technical exper-
tise Is not always able to quantify ob-
served or perceived problems and their
causes. The public's perception of water
quality conditions, especially use Impair-
ment, Is also not necessarily directly re-
lated to empirical assessment of water
quality. Basic to any classification or
management scheme should be a definition of
what constitutes an undesirable condition.
Tools (such as models or Indices) to evalu-
ate nutrients In stream environments are
also not always available.
The need for simple, reliable methods to
evaluate the Impacts of nutrient loadings
to lakes and streams has led to a large
number of approaches with an equally large
number of methods and Indicators. General
agreements have emerged as to the methods
of approaching and solving lake problems.
Of particular significance are the rela-
tively simple tools and procedures used for
analysis and assessment of lake water
quality — usually water transparency,
algal blomass, and the causative nutrient
levels. Sufficient research, directed at
solving specific problems, has been Inte-
grated Into whole-lake approaches to be
able to address lake water quality problems
with some confidence.
Of most Importance has been a further defi-
nition and refinement of lake use cate-
gories and classifications, nuisance con-
ditions, and application and evaluation of
various lake management and protective
strategies. Many of the concepts developed
and applied to lakes and ponds (lentlc sys-
tems) are not directly applicable to
streams (lotlc systems). The predictive
methodologies developed for lake systems
are based on simple and useful parameters.
No such tools are currently available for
allocating phosphorus In a site-specific
manner to flowing water.
If a defensible position and regulation of
phosphorus discharges to small stream sys-
tems on a site-specific basis Is to be
taken, the following concerns must be
addressed:
- What level of enrichment (or aquatic
plant growth) Is considered objection-
able?
- Which community and level of function
will provide the best tool to evaluate
the Impacts of Increased or decreased
phosphorus levels?
STUDY BACKGROUND
The overall purpose of the Phosphorus
Assessment study Is to define methods of
dealing with phosphorus and non-point
source pollution In setting appropriate
water quality goals or standards. The
study was conducted In phases, beginning
with a review of past efforts to establish
phosphorus and non-point source water
quality standards or objectives. Reviews
Included other state's and agencies'
approaches to defining phosphorus water
quality standards and limiting non-point
source Impacts (Lewis 1980, Warn 1980), the
ability of chemical and biological water
quality Indices to assess sediment and
nutrient Impacts on streams (Warn 1980,
Chantry 1981, Schrank 1982, Wawrzyn and
Randall 1983, Narf In prep.), lake and
stream classification schemes (O'Flannlgan
1980, Schuettpelz 1982, Ball 1982), and
stream primary producer responses to nutri-
ents (Mace et al. In prep.).
The conclusions and recommendations of
other agencies provided a framework to
assess the feasibility of establishing
phosphorus water quality standards and
non-point source control objectives In
Wisconsin. The topical reviews formed the
technical basis for selecting specific
areas where more detailed Investigation was
required. These Investigations constituted
the second study phase. The direction
given for the field studies then focused
project efforts on the phosphorus control
element (Mace et al. 1982).
- 3 -
-------
In Wisconsin, water quality changes due to
point source discharges have been evaluated
through pre and post-operative surveys,
basin-wide assessments, wasteload allo-
cation studies and, recently, stream clas-
sifications. These studies document stream
response to gross changes In wastewater
characteristics of dischargers. These
methods, however, are generally not ade-
quate to evaluate stream response to
changing phosphorus Inputs, nor for allo-
cating phosphorus to small stream systems.
The sensitivity (or level of resolution) of
such methods must be able to assess the
existing situation, responsibly recommend
phosphorus limits to the discharger, and
reliably predict water quality Improvements
resulting from specified phosphorus removal
recommendations. For a particular assess-
ment or a I locative method to succeed, It
must be empirically developed, tested and
defensible.
Currently a small stream, dissolved oxygen
model Is available for use to allocate
certain wastewater constituents discharged
to streams. These allocations are designed
to maintain a specified dissolved oxygen
criterion for a designated stream use under
critical stream conditions (e.g. high
temperature, low-flow). Either directly
through respiratory demand, or Indirectly
through channel modification, macrophyte
growth will Impact stream dissolved oxygen
(DO). Modification of DO models or
development of similar mathematical expres-
sions could establish a link between phos-
phorus, plant growth and the associated
Impacts on stream DO. This might allow
allocation of phosphorus directly through
the Impacts on primary production and In-
directly through this community's Impacts
on DO.
If the data support such, establishment of
phosphorus uptake (P-decay) characteristics
due to stream assimilation, much as BOD Is
now allocated, could be an Important con-
sideration In assigning phosphorus dis-
charge limits. A second alternative for
controlling phosphorus discharges Is to
define levels of acceptable or unacceptable
primary producer growth, based on community
response to phosphorus, aesthetics, or
physical changes. There Is little Infor-
mation available on the use of subjective
limits based on plant density or aesthetic
conditions.
The field program was designed then, to
establish a basis for phosphorus control
(water quality objective or standard), a
methodology for applying standards
(assigning effluent limits) and evaluating
water quality Impacts (monitoring require-
ments). In addition to providing a defen-
sible, scientific basis for phosphorus con-
trol In small streams, the field study re-
sults should also assist In specifying
receiving system classifications or cate-
gories, associated water quality criteria
and methods of applying selected criteria.
FIELD STUDY DESIGN
Specific objectives of the Phosphorus
Assessment field study Included:
- Quantifying the relationship between
phosphorus and plant blomass;
- Recognizing the factors which modify the
response of stream plants to nutrients;
- Quantifying the Impacts of plant blomass
on stream systems;
- Determining a level of acceptable plant
blomass within a particular system; and
- Evaluating and recommending monitoring
strategies for use In small stream water
quality Investigations.
This study was conducted over a two year
period. Due to the uncertainty of control-
ling factors In most small stream systems,
an approach to evaluate a variety of fac-
tors In Intensely studied reaches was Im-
plemented the first year. Study sites were
selected which represented a variety of
stream types and physical conditions,
nutrient and flow regimes, and dominant
blotlc community.
- 4 -
-------
Initially, It was also desirable to collect
data from systems not Impacted by waste-
water discharges to document stream be-
havior In the absence of point sources.
These streams would then serve as bench-
mark, or reference data to compare with
strean behavior In reaches Impacted by
wastewater treatment plants. This would
also provide comparative data on the Im-
portance of non-point vs. point source Im-
pacts In considering the relative contri-
bution of a discharge to a particular
situation In the presence of NPS Inputs.
Based on analysis of the first year's data,
the second year of data collection Included
a larger number of streams to further In-
vestigate relationships and answer specific
questions. In order to Isolate and deal
effectively and responsibly with the phos-
phorus question, a number of other concerns
needed to be resolved. These questions In-
cluded the contribution of sediment nutri-
ents to macrophyte nutrition In streams,
the Influence of bottom substrate type In
macrophyte colonization and growth,
seasonal weather Influences, and physical
changes the primary producers themselves
Impose on the system.
There were no tested approaches (methodolo-
gies) available to evaluate the Impacts of
phosphorus on small stream water quality.
The field studies, then, were also designed
to develop assessment methods that I) Iden-
tify actual or potential "problem" condi-
tions; 2) estimate the potential for Im-
provement or degradation; 3) by comparison
or modeling methods, recommend actions to
remedy or prevent water quality deteriora-
tion due to phosphorus discharges; 4) pro-
ject stream response to phosphorus reduc-
tion; and 5) develop the ability to predict
changes In water quality based on changes
I n phosphorus concentrations.
METHODS
STUDY REACH SELECTION
Study reaches were selected which would
best depict the Impacts of nutrients on
primary producer communities. Criteria
used to select stream reaches were designed
to minimize the effects of physical factors
on the growth of primary producer communi-
ties In small streams. The criteria used
to select the stream reaches were:
-Maximum reach depth 2-3 feet. Shallow
depth would decrease the potential for
light limiting plant growth;
-Mean annual flow less than 60 cfs.
(small stream category);
- Maximum stream top width 60-70 feet
(small stream category);
- Stream reach should be relatively un-
shaded and free of obstructions;
- Stream reach length should be a minimum
of 300 feet to a maximum of 2,000 feet
(1981 stream reaches were less than 300
feet).
Based on these criteria and existing water
quality data 19 stream reaches were
selected In southeastern and southern
Wisconsin (see Appendix I). Seven stream
reaches were Intensive monitoring sites In
1981, four of which were expanded and moni-
tored In 1982. Twelve synoptic stream
reaches were selected and monitored In
1982. Sites were chosen which represented
a variety of In-stream nutrient concen-
trations.
INTENSIVE STUDY REACHES
The major objective of monitoring the In-
tensive study reaches was to evaluate the
environmental factors Impacting primary
producer growth over the growing season.
Water chemistry, macrophytes, perlphyton,
die) oxygen regime, substrate type and
stream flow were monitored at the Intensive
study reaches.
Water Chemistry
Water chemistry samples were collected
every two weeks from May through December
- 5 -
-------
1981 and May through September 1982. Sam-
pling times corresponded with primary pro-
ductivity or diet DO surveys. During the
dlel surveys, samples were collected Just
prior to dawn, and again In late afternoon
at the upstream and downstream limits of
the reach. In 1982, sonples were collected
only at the downstream limit of the reach.
Chemistry samples were collected one foot
betow the water surface where possible,
preserved according to Wisconsin State Lab
of Hygiene (SLoH) and Standard Methods
(APHA. et al. 1981) guidelines. Iced and
shipped within 24 hours to SLoH for analy-
sis. Water chemistry parameters are pre-
sented In Table I.
Table I
Water Chemistry Analyses
Total Phosphorus (PTOT)
Ortho-phosphate (PC^P)
Total KJeldahl Nitrogen (TKN)
Nitrite-Nitrate Nitrogen (N02+N03N>
Ammonia Nitrogen (NH3N)
Turbidity
Total Alkalinity*
Hardness*
PH*
Biochemical Oxygen Demand (B005)
Chemical Oxygen Demand* (COO)
Total Non-FI Itroble Suspended Solids
Total Volatile Non-FI Itrable
Suspended Solids
* 1981 analyses only.
Sediments
The purpose of sediment Interstitial water
and bulk sediment sampling In 1981 and
spring 1982 was to characterize overall
sediment nutrients within an entire reach,
and to determine If differences In nutrient
concentrations occurred between areas
colonized and areas that were unco Ionized
by aquatic macrophytes. Bi-weekly stream
mapping data were used to determine plant
cover and associated bottom materials.
Macrophyte and non-macrophyte samp 11 ng
sites were chosen based on these occur-
rences. In 1982, surveys were also con-
ducted on two streams that received waste-
water treatment plant effluent. These sur-
veys were designed to determine If nutri-
ents were accumulating In the sediments and
Interstitial water at points downstream
from the effluent outfalls.
Interstitial Water Analyses
Interstitial water (IW) nutrient concentra-
tions within and outside of macrophyte beds
were determined monthly In 1981 and once In
the spring during 1982. In September 1982,
the Bark River and White River were sampled
upstream and at points downstream of the
treatment plant outfalls. Sand-gravel sub-
strates with visually similar composition
were sampled at each location on these two
rivers.
Samples were collected by vacuum from a
1/2" (1.25 on) l.d., 9" (22.5 cm) long
well-point In the substrate and filtered
before contact with the atmosphere. Prior
to collecting each sample, the we 11-point,
suction lines. 0.45 urn filter and collec-
tion flask were rinsed with 50-100 ml 10*
HCI and twice with 50-100 ml distilled
water. Sample blanks were collected at
this time.
The well-point was Inserted Into the sedi-
ment deeply enough to avoid collecting the
overlying stream water, yet within macro-
phyte rooting depth (ca. 6"). Silts for
collection of the pore water were located
In the terminal 2" (5 cm) of the
well-point. Approximately 25-50 ml were
collected by vacuum and discarded as
rinse. Vacuum was reapplled to collect 50
ml of filtrate for the sample. Samples
were Iced and sent to SLoH for dissolved
P04P (total dissolved P In 1982), NH3N
and N02+N03N determinations.
Bulk Sediment Analyses
Bulk sediment nutrient concentrations with-
in and outside of macrophyte beds were also
determined at monthly Intervals In 1981.
Cores collected In 1981 were from the pre-
dominant substrate In stream transects.
- 6 -
-------
those from 1982 were collected from organic
sediment deposits. Similar to the Inter-
stitial water sampling, sediment cores
(composites) were col lected up and down-
stream of treatment plant outfalls on the
Bark River and White River In 1982.
To minimize variability between Individual
col lections, sediment samples were col-
lected from several different areas and
composited from 2-4 cores taken from each
transect. Two composite samples were col-
lected, one representing bulk sediment
nutrient content within macrophyte areas,
the other nutrients from macrophyte-free
areas. In the surveys conducted on the
Bark River and White River In 1982, samples
were composited from 5-10 cores collected
at each site from areas of silt and organic
sediment deposition.
Cores were 3.8 cm (1.5") In diameter and
from 5-15 cm (2-6") In length, depending on
substrate type. Grab samples were col-
lected where substrate did not allow
coring. Individual samples were mixed,
sub-samples Iced and sent to SLoH to be
analyzed for TKN and PTOT.
Stream Mapping
Physical and biological stream character-
istics were mapped every two weeks along
established transects In each stream reach.
In 1981, stream reach lengths were selected
between 100-230 feet and transects were
established at approximately 20 to 40 foot
Intervals. The reach lengths were expanded
In 1982 to 800-1600 feet with 10-20 tran-
sects established per reach. One
square-foot observation points were mapped
at one to three foot Intervals along each
transect. These methods are a modification
of a line-transect method (e.g. Ku11 berg
1974, Wong and Clark 1976, Wright et al.
1981). Depth, velocity, substrate type,
percent macrophyte and/or perlphyton
coverage, and species abundance were
recorded at each observation point.
Depth and velocity were measured using a
wading rod and Marsh-McBlrneyR Model 201
current meter. Stream discharge measure-
ments were calculated from Incremental
depth and velocity measurements (Buchanon
and Saners 1969). In 1982, velocity data
were collected at one transect for dis-
charge calculations.
Stream bottom type was recorded as percent
of the various substrate size classes.
Substrate was classified by visually esti-
mating particle sizes using USGS (1978)
guidelines (Table 2). A detritus class was
added for Incompletely decomposed organic
material. Substrate data for each obser-
vation point were summarized by calculating
a Substrate Index (SI). Weighting of the
various size classes was conducted to pro-
vide a continuum of SI values. The SI was
calculated from the following equation:
SI - (S1I.T) * 2<\SAND) * MVCRAV) » KXRUBBLU * S(XBOULUEIO
5
Percent macrophyte coverage, perlphyton
coverage and type (filamentous or
non-filamentous) were estimated. Macro-
phyte species abundance was given a rating
corresponding to percent coverage (Table 3).
Macrophyte Harvesting
The objectives of the macrophyte harvesting
surveys were to:
- Provide macrophyte blcmass estimates at
the time of die I oxygen surveys;
- Estimate macrophyte blcmass accumulation
throughout the growing season: and
- Determine maximum stable macrophyte bio-
mass (summer standing crop) and when It
occurred.
Macrophyte blomass samples were collected
monthly, within 7 days following the dlel
surveys. Three to five sample quadrats
were selected within 10 ft. up or down-
stream of each stream transect using random
numbers tables: the first two digits
specified the distance from the left
streambank, the third the distance up (even
number) or downstream (odd number) from the
transect. Quadrats were not selected wlth-
- 7 -
-------
In three feet of the transect to avoid .sam-
pling areas disturbed by the mapping acti-
vities.
Table 2
Substrate Size Classes
(based on USGS 1978 guidelines)
Class
Boulder (Large Cobble)
Cobble (Rubble)
Gravel
Sand
Silt
Detritus
Size mm (In.)
256 ( 10")
64 - 256 (2.5 - 10")
2 - 64 (O.I - 2.5")
.062 - 2 (.003 - .1")
.004 - .062
Table 3
Macrophyte Species Abundance Rating and
Corresponding Percent Coverage Estimate
Species
Abundance Rating
Percent Coverage
5
4
3
2
1
80 -
60 -
40 -
20 -
1 -
100
80
60
40
20
In 1981, a record MBS kept of all pre-
viously harvested coordhiates and samples
were not collected within 2 ft. of a pre-
viously sampled location. After the first
harvests In 1982, transects were moved
slightly upstream or downstream of the
original transects to avoid sampling pre-
viously harvested areas. Wright, et al.
(1981) noted, however, that disturbance by
mapping or harvesting did not result In
changes In macrophytes or substrate.
Survey flags were used to locate quadrats
In the stream. Quadrats were delineated
using a one-square foot Surber sampler with
the random number coordinate at Its cen-
ter. Percent macrophyte cover, species
abundance ratings, water depth and velocity
over the sampling point were measured
before plant harvesting.
Mocrophy-tes rooted within the Surber frame
were harvested with roots, and vigorously
washed with stream water. Plants from each
quadrat were sorted Into dominant and other
species, bagged and transported to the lab
on Ice. Samples were briefly re-washed
with tapwater In the lab, placed In dried
pre-welghed paper bags, and dried to con-
stant weight (24-48 hrs) at 65°C In a
forced-air oven to determine sample dry
weight.
Tissue nutrient analyses were conducted on
dried macrophyte samples collected at each
harvesting. Three to eight tissue samples
were prepared. Including one composite sam-
ple for each dominant species and/or three
composite samples. Composite tissue sam-
ples were thoroughly mixed, bagged and sent
to the University of Wisconsin Soil and
Plant Analysis Lab, Madison, for analy-
sis. Plant material was analyzed using
Inductively-coupled plasma spectroscopy for
N, P, K, Ca, Mg, S, Zn. B, Mn, Fe, Cu, Al
and Na.
Perlphyton Harvesting
The primary objectives of the perlphyton
harvest were to characterize community
growth at different nutrient concentra-
tions and compare this with macrophyte
growth characteristics and die I dissolved
oxygen regimes.
Perlphyton were collected every two weeks
from glass-slide samplers (perlphytometers)
and bricks. In 1981, most reaches
contained three perlphytometers and six
bricks. In 1982, as the reaches were
lengthened, four perlphytcmeters and eight
bricks were harvested.
Bricks were scored Into quarters for sample
harvesting and placed on the stream bottom
at perlphytometer placement sites. Areas
of each brick quadrant were measured for
use In computing area! blcmass estimates.
After the first harvest In 1981. exposure
periods for the bricks were effectively
eight weeks. Each sampling date, one quar-
ter of each brick was scraped Into Indivi-
dual vials and Iced for ash-freo weight
- 8 -
-------
determinations. Velocity over each brick
was measured every two weeks, on placement
and soup I Ing dates.
In 1982, bricks were moved to locations In
the stream to ensure minimal macrophyte
shading. Exposure periods for bricks In
1982 were four weeks. Two brick quadrants
were sampled for tissue nutrient analyses,
and one quadrant each for chlorophyll and
ash-free weight analyses. In the field,
chlorophyll samples were washed directly
Into 50 ml centrifuge tubes with 90$ ace-
tone, packed on dry Ice and shipped to SLoH
for analysis. Tissue nutrient and ash-free
weight samples were washed with distilled
water Into appropriate containers. Tissue
nutrient samples were analyzed using In-
ductively-coupled plasma spectroscopy at
the UW Soils and Plant Analysis Lab.
Madison. Ash-free weight was determined
according to Standard Methods (APHA, et al.
1981). Water depth and velocity were
measured at col lection.
Perlphytometers were exposed for two-week
periods, the glass slides suspended 2.5 cm
(I") below the water surface. At the time
of placement and retrieval, velocity over
the sampler was measured and sampler con-
ditions noted. Slides were removed,
drained of excess moisture, placed In
folI-wrapped containers and frozen. Slides
were selected for community composition,
chlorophyll, ash-free dry weight and nutri-
ent analyses.
A study comparing glass-slide perlphyton
parameters to those occurring on natural
substrates was conducted In 1981. Perl-
phytometers were treated similar to the
routine placements. In addition, natural
substrates (rocks) were sampled by placing
a plastic cylinder over the substrate, and
scraping the perlphyton off using a stiff
bristle brush and a razor blade. The col-
lected material was handled In the same
manner as the perlphytoneter and brick col-
lections for chlorophyll and ash-free
weight analyses.
Die I Studies
The primary objective of the dlel studies
was to Investigate relationships between
photosynthesis, respiration and plant bio-
mass. Die I productivity studies were con-
ducted monthly at each stream reach before
macrophytes were harvested. Dissolved oxy-
gen (DO) and temperature were measured at
the upper and lower transects of each
reach. In 1981, measurements were made at
2-3 hour Intervals for 24-27 hrs. DO and
temperature were contInuou sIy recorded I n
1982. The DO and temperature data were
collected using Yellow Springs Instru-
ment^ (YSI) dissolved oxygen meters
(Models 56, 57, 58). Meter calibration was
checked against triplicate WInkier tltra-
tlons at 3-6 hour Intervals.
Light Intensity data were collected using a
Blospherlcal Instrument's** Quantum Scalar
Irradlance system, which measures Photo-
synthetical ly Active Radiation (PAR) In the
400-700 ran range. PAR data were collected
at 2-3 hour Intervals In 1981, and con-
tinuously recorded In 1982.
The 11ght measurements represent In-
stantaneous rates, and were measured In
units of Quanta/sec • cm'2. Light
readings were taken submerged, approxi-
mately 15 cm from the stream bottom and at
the surface (dry reference sensor). [Note:
6 x I017 0/sec • cm'2 = lu
Elnsteln/sec • cm'2; I Watt/cm2 = 4.6
uE/cm2; I Klux = 18 uE/cm2!.
Water chemistry and suspended solids sam-
ples were collected twice during the dlel
period; Just prior to sunrise and late
afternoon.
SYNOPTIC SURVEY REACHES
The primary objective of conducting the
synoptic surveys was to evaluate the
relationship between mean summer In-stream
nutrient concentrations and maximum stable
summer macrophyte blomass. Twelve synoptic
stream reaches were selected using the sane
criteria as were used for the Intensive
- 9 -
-------
monitoring sites, with macrophytes telng
the dominant primary producers.
Water chemistry collections were made at
two week Intervals for 6-8 weeks prior to
macrophyte harvesting to characterize the
mean simmer In-stream nutrient concen-
trations. The synoptic reaches were sampled
a minimum of three times with some reaches
being sampled a fourth time. Collection
methods and water chemistry parameters were
the same as those used at the Intensive
study reaches In 1982. Samples were not
analyzed for BOO at the Synoptic Sites.
Stream mapping and macrophyte harvests were
conducted using the same methodologies as
were used at the Intensive monitoring
sites. Perlphyton harvests and dlel sur-
veys were not conducted at the synoptic
monitoring sites.
Macrophyte harvests and stream mapping were
conducted at the time Judged to be maximum
stable blomass. A number of Investigators
(e.g. Gerloff and Krombholz 1966, Calnes
1965, Stake 1968, Ball et al. 1973) have
recommended sample collection later In a
growing season due to the stability of tis-
sue nutrient concentrations. Prior to the
actual harvest, all sites were visited fre-
quently to Judge the best harvesting time.
The timing of the maximum macrophyte blo-
mass was estimated from the data collected
at the Intensive monitoring sites In 1981
and field observations. This time period
was found generally to be mid to late
August. Macrophyte tissue sampling and
analysis was Identical to methods used for
the Intensive monitoring sites.
Photosynthesis/Respiration Studies
In 1982, reclrculatlng light and dark cham-
bers, and light bottle/dark bottle studies
were conducted at stations on the Bark R.
to determine l_£ situ photosynthesis and
respiration rates. Light, 00 and tempera-
ture were continuously recorded for the
duration of the experiments (usually 2-3
hours).
The reclr«ulatlng plexiglass chambers had a
volume of 80 liters, and used a pump
capable of reclrculatlng 90 llters/mln. In
the field, the boxes were placed over a
macrophyte bed and sealed to the substrate
using bentonlte clay. The macrophytes were
harvested at the end of the testing to de-
termine the dry weight blomass. In all
cases, macrophytes were dried and weighed
In the same manner as the regular harvests.
Macrophytes were also placed In light and
dark BOD bottles and Incubated on the
stream bottom. Incubation times were
varied from 0.5 to 7.0 hours. As with the
enclosure studies, light was continuously
recorded. Initial and final DO concentra-
tions In the bottles were determined by
Wlnkler tltratlons. Macrophytes were re-
moved from the bottles and dried at 65-70°C
to constant weight.
Perlphyton were filtered from a known
volume of sample after the tltratlon for
dissolved oxygen was completed. The fil-
ters were dried at IOO°C for 24 hours, then
ashed at 500°C for 2 hours to provide dry
weight and ash-free weights.
STUDY REACH CHARACTERISTICS
Nineteen strean reaches were selected to
provide as wide a range of chemical condi-
tions as possible. Including reaches
receiving wastewater treatment plant
(POTW) effluent. To Illustrate the range
of parameter values represented In the data
analyses, mean values were ranked, from
lowest to highest. Occurrence of a particu-
lar study reach within the ranked hierarchy
was used to describe or group streams with
similar characteristics. This was done for
chemical and physical parameters. For many
parameters, divisions between groups was
based on frequency distributions. In
others, divisions were used which rep-
resented specific parameter ranges.
For the purpose of the following rankings,
the 1981 data were treated as separate from
the 1982 data, giving 22 data sets for ccm-
- 10 -
-------
par I son (7 sets In 1981 plus 15 sets In
1982). Data from the reaches represent
mean growing season (June-August) values.
REACH CHEMICAL CHARACTERISTICS
Study reaches were ranked according to mean
growing season (June-August) stream total
phosphorus (TOTP) concentrations. Inorganic
nitrogen (NH3 + N02 + NOjN) concen-
trations and the N:P ratio. Although
various In-stream levels have been
suggested as limiting concentrations to
primary producers, study reaches were
grouped based on their distribution within
the ranked hierarchy. Chemical character-
istics of the study reaches are presented
In Table 4.
Phosphorus
Study reaches, ranked by mean growing
season In-stream total phosphorus concen-
trations, were separated Into three groups
representing "low" P «0.05 mg/l TOTP),
"medium" P (0.05-0.20 mg/l TOTP) and "high"
P (0.20-0.50 mg/l TOTP) groups (Table 5).
No streams receiving POTW effluent occurred
In the low P group, which represented 30?
of the study reaches. Fifty-six percent of
the study reaches occurred In the medium P
group, six of which (46$ of this group)
Table 4
Summary of Stream Reach Chemical Characteristics, 1981-1982
Reach
Sugar
Ashlppun-M
Ashlppun-N
Ashlppun-S
Kohlsvllle
Bark-Wolf
Bark-Lurvey
Bark-Masonic
POTW
Impacted
N
N
N
N
N
N
Y
Y
Group*
1981 1982
2
2
2
2
2
2
2
—
1
2
--
—
—
2
2
2
Total P Inor. N
1981 1982 1981 1982
.122
.139
.137
.141
.091
.022
.244
...
.104
.095
- —
___
.51
.05
.08
.08
.14
.021 .09
.159 1.17
. 163
3.24
1.05
—
—
—
.11
.74
.72
Blcmass
1981 1982
151.7
151.7
99.0
151.7
- —
45.2
249.6
— —
134.4
125.7
—
—
—
25.6
289.0
187.5
SYNOPTIC SITES
Bark-Wan I N
Mukwonago N
MlIwaukee
- Campbe 11 sport Y
- East Br. N
Suppernong N
Pewaukee Y**
Cedar N
Mt. Vernon N
Black Earth Y
Fox-Portage Y
3
3
3
3
3
3
3
3
3
3
.033
.013
.493
.040
.027
.140
.050
.050
.107
.147
1.95
.03
.56
.33
.83
.48
.71
4.28
2.15
.62
179.3
57.8
448.5
72.5
262.6
213.8
161.0
365.5
282.1
146.9
*GROUP I = Intensive study site with primary producer harvest and mapping but without dlel studies.
2 = Intensive study site with primary producer harvest and mapping. Including dlel studies.
3 - Synoptic survey, harvest and mapping.
**Thls site has received wastewater discharge from the City of Pewaukee up until the end of 1981.
-------
Mere Impacted by POTWs. (This group could
be further sub-divided Into medium-low and
medium-high P ranges (0.05-0.12 and 0.12 -
0.20 mg/l TOTP respectively) with about 38$
of the group occurring In the lower phos-
phorus group. This group roughly corres-
ponds to most recommended levels of phos-
phorus necessary to control or slow eutro-
phlcatlon. All but one of the POTW-lm-
pacted reaches would occur In the medium-
high range.) Two streams represented the
high P group, both Impacted by POTWs.
Nitrogen
The distribution of streams along the
nitrogen gradient determined the group rank
boundaries (Table 5). The groups repre-
sented "low" N «I.OO mg/l Inorganic
nitrogen), "medium" N (1.00-1.9 mg/l In-
organic nitrogen) and "high" N O2.00 mg/l
Inorganic nitrogen). Almost 50)1 of the
study reaches occurred In the low N group,
5 of which (45$ of the group) were Impacted
by a POTW. Approximately 33$ of the study
reaches occurred In the medium N range, 2
of which (25$ of the group) -were Impacted
by POTWs. Four study reaches occurred In
the "high N" range, only one of which
received POTW effluent.
N:P Ratio
Grouping of reaches by N:P ratios followed
somewhat the ranking of streams by nHrogen
gradient (Table 5), those reaches with very
high nitrogen having the highest N:P
ratios. The N:P ratios ranged from approxi-
mately I to 86. Based on the ranked dis-
tribution of study reaches, reaches were
grouped at N:P ratios of <5, 5-20, and
Approximately 39$ of the reaches had N:P
ratios of less than 5:1, the majority of
the group (61%) being Impacted by POTWs.
Forty-four percent of the reaches occurred
In the middle ranking, two of which receive
WWTP effluent, and 17$ In the high N:P
range, none of which receive an effluent.
REACH PHYSICAL CHARACTERISTICS
Stream physical characteristics may have a
significant Impact on plant growth. Sub-
strate composition, water depth and
velocity and other physical characteris-
tics can modify plant response to nutri-
ents. In addition to affecting die I DO
changes. The study reaches represented a
variety of morphometrlc conditions. Reach
physical characteristics are presented In
Table 6. For comparison, all streams were
ranked by mean reach width, depth,
cross-sectional area, predominant substrate
and current velocity. For the most part,
grouping of these parameters were deter-
mined by their frequency distributions
rather than established criteria.
The data from 1981 reaches are not strictly
comparable to their 1982 data. With the
exception of Sugar Cr., all the 1981
reaches were lengthened In 1982.
Mean reach depths for all sites ranged from
.15 m (0.50 ft) to .65 m (2.16 ft). Mean
reach depths for those reaches continued
from 1981 were generally higher than 1982
values. Of the 22 data points, approxi-
mately 26% were In the .15 -.2 m (0.50-0.75
ft) depth range, 61$ In the 0.30-0.45 m
(1.0-1.5 ft) depth range, and 13* In the
0.5-0.65m (1.65-2.15 ft) depth range
(Table 6). Of those reaches In the shallow
range, one Is Impacted by a POTW. Four of
the middle range (or 28$) and two of the
three In the deep range receive POTW
effluent.
Width and Cross-sectional Area
Mean reach widths ranged from 3 to 20 m
(10-67 ft) (Table 6). Approximately
one-half (48$) of the reaches were 8 m (25
ft) or less across. The other half were 15
m (50 ft) or less, with only two reaches
greater than 15 m across. Most of the
POTW-lmpacted streams were In the 8-15 m
(25-50 ft) range.
- 12 -
-------
Ranking mean cross-sectional areas sug-
gested almost equal division between the
nunber of streams 2.3 square meters (25
square feet) or less and the number of
streams 3-8 m2 (35-85 ft2) (Table 6).
The range was .5-8 m2 (5-85 ft2). As
with mean width, most of the POTW-1 nip acted
reaches were In the larger group.
Mean Velocity
The majority of streams (83$) were within a
mean reach velocity range of 0.08-.20 m/sec
(0.25-0.65 ft/sec) (Table 6). All but one
of the POTW-Impacted streams occurred In
this group. The range of velocities was
about .03-.3 m/sec (0.1-0.9 ft/sec).
Substrate
Table 5
Breakdown of Stream Reaches By
Phosphorus and Nitrogen Characteristics
The majority of stream reach substrates
were composed of gravel and sand. The two
extremes, predominantly silt substrate and
predominantly rubble substrate, were also
represented In the study reaches.
Reaches were ranked by Substrate Index, and
grouped by the SI values presented In Table
7. In this manner, approximately 30? of
the reaches were represented by si It-sand
substrates (SI - 20-40), 52J In the
sand-gravel-rubble group (SI - 40-60) and
13? In the rubble-cobble group. Reaches
receiving POTW effluent were present In
each group. The majority, however,
occurred In the sand-gravel-rubble group.
LOW P . LOW N
Bark-Wolf - 81
Bark Wolf - 82
Mukwonago R.
Cedar Cr.
Milwaukee R.-East Br.
Scuppernong R.
LOW P . HIGH N
Bark-Wan IschIaeger
Mount Vernon Cr.
MID P . LOW N
Bark Lurvey - 82
Bark Masonic
Fox at Portage
Pewaukee R.
MID P . MID N
Ashlppun Main - 81
Ashlppun Main - 82
Ashlppun North
Ashlppun South
Sugar Cr. - 81
Kohlsvllle R.
MID P . HIGH N
Black Earth Cr.
Sugar Cr. - 82
HIGH P . LOW N
MlIwaukee CampbelI sport
HIGH P . MID N
Bark Lurvey - 81
(low N:P)
n n
n n
(mid N:P)
n n
(high N:P)
(high N:P)
n n
(low N:P)
n n
n n
n n
(mid N:P)
it n
n n
n n
n n
n n
(low N:P)
(high N:P)
(low N:P)
(low N:P)
- 13 -
-------
Table 6
Sunroary of Stream Reae» Physical Character I sties, 1981-1982
Sugar
Ashlppun-M
Ashlppun-N
Ashlppun-S
Kohlsvllle
Bark-Wolf
Bark-Lurvey
Bark-Masonic
SYNOPTIC SITES
Bark-Wah 1
Mukwonago
Ml Iwaukee
- Campbell sport
- East Br.
Supper nong
Pewaukee
Cedar
Mt. Vernon
Black Earth
Fox- Portage
Length
1981
60
40
26
32
56
68
67
--
— —
~
—
«
—
~
__
—
--
(m>
1982
60
235
—
—
—
305
480
242
74
300
211
91
91
91
86
120
112
136
Width
1981
7
5
4
3
2
9
12
—
_•
—
—
—
—
—
—
—
— ~
(m)
1982.
7
6
...
—
—
II
12
13
9
16
9
12
5
6
5
6
II
17
Depth
1981
.22
.34
.19
.18
.19
.29
.36
••w
__
—
—
—
_-
—
—
—
•»
(m)
1982..
.35
.39.
—
—
— —
.49
.65
.45
.36
.39
.34-
.32
.32
.15
.16
.35
.43
.44
Velocity
1981
.18
.12
.16
.19
.18
- .15
.12
—
•
— -
—
z
— —
— •
~~
(in/sec)
1982
.24
.27
~" "
— ~~
._•
.16
.11
.14
.14
.22
.08
.08
.12
.0*
.20
.16
.19
.09
1981
53.6
34.0
4&.2
45.0
67.2
44.4
45.2
— ™ ™
~~~
^r^^m-
••^
— —
"^"
SI*
1982
52.2
37.0
— ™~
™~~
46.8
50.0
47.4
37.0
43.0
84.2
30.4
36.4
54.4
64.2
37.0
24.6
37.4
•Method- of calculation In text.
Table 7
Substrate Index Substrate Type and Value Ranges
Substrate C\ass
Silt
Sand-Si It
Sand-Qrave4
Rubble
SouIder
SI Range
20 - 29
30 - 39
40-49
50 - 69
90 - 100
- 14 -
-------
STREAM PRIMARY PRODUCERS
MACROPHYTES
Introduction
The primary objective of the macrophyte
surveys was to determine If a significant
relationship exists between In-stresn nu-
trient concentrations and late summer bio-
mass (summer standing crop). Although
several authors have suggested minimum nu-
trient concentrations which will stimulate
maximum macrophyte growth (e.g. Gerloff
1969; Mulligan and Baranowskl 1969;
Pltcalrns and Hawks 1973) little has been
done to develop a usable predictive re-
lationship defining growth using In-stream
nutrient concentrations. These relation-
ships have been quantified for lake phyto-
pi ankton (Jones and Bachmann 1976; Dillon
and Rlgler 1974; Hoyer and Jones 1983), but
not for macrophytes.
The Interim Technical Report of the Phos-
phorus Assessment Study (Mace, et al. 1982)
reported that a significant relationship
did exist between late summer blomass and
mean summer phosphorus concentrations at
seven study reaches In four southeastern
Wisconsin streams. The 1981 samplings,
however. Involved too few data points to
develop a substantiated predictive model.
Sampling In 1982 Included the 1981 sites
and II additional stream reaches repre-
senting a wider range of In-stream nutrient
concentrations.
Results and Discussion
While correlation coefficients are Indica-
tors of relationships, they are not neces-
sarily Indicators of cause and effect re-
lationships. Least squares regression
equations were calculated for paired para-
meters having significant correlation co-
efficients.
The strength of the regression models was
assessed using R-square values, mean resid-
ual error and confidence limits (p=.05)
expressed as percent of the predicted
values. R-square Is the proportion of the
total variance In the dependent variable
that may be attributed to the regression on
the Independent variable. Mean residual
error Is the absolute difference between
the observed and predicted values as a per-
centage of the predicted values. The con-
fidence Intervals (p=.05) for the predicted
values are given as the difference between
the predicted value and the confidence
limit (p=.05) value expressed as a per-
centage of the predicted value.
These parameters describe how well a re-
gression model fits a particular data set.
These parameters do not test a model to
determine how well It will work as a man-
agement tool. An Independent data set Is
used to test empirical models and verify
their predictive capabilities.
Stream Type Determination
The strewn reach mapping data suggested
that two basic types of stream reaches were
surveyed. Criteria related to macrophyte
distribution and dominant substrate type
were used to classify the stream reaches as
Type I or Type II. Stream reaches were
classified as Type I If macrophyte popu-
lations were relatively homogenous I y dis-
tributed and secondarily If the reach had
substrate dominated by sand, gravel or rub-
ble. Stream reaches were classified as
Type II If macrophyte distribution was
patchy and secondarily was limited to areas
of si It and si It deposition. Based on
these criteria. Type I streams Included
ell streon reaches except Mount Vernon
Creek, Black Earth Creek and Scuppernong
River which were classified as Type II.
The mapping data were evaluated using fre-
quency analysis to assess edaphlc Impacts
on macrophyte occurrence. The frequency
analysis compared the dominant bottom sub-
strate of each stream reach Type (I or II)
and the substrate size over which macro-
phytes were growing. Larger size substrate
classes (Sl>40) were dominant In Type I
stream reaches. Smaller size substrate
classes (SK40) were dominant In Type II
stream reaches (Table 8).
- 15 -
-------
Table 8
Summary of Stream Reach Mapping Data
Stream Reach
Type
1
1 1
Number of
Reaches
16
3
'Percent
Occurrence
On Sl>40
77
28
Percent
Occurrence
On SK40
23
72
2Percent
Substrate
Sl>40
75
31
Percent
Substrate
SK40
27
69
'Values calculated from observation points where macrophytes occurred.
^Values calculated from all observation points.
Macrophytes occurred at approximately the
same percentage of sample points (8IJ-83JI)
I n both Type I and Type 11 stream reaches
(Figure I). The Important difference be-
tween the stream Types, however, was that
In Type I stream reaches raacrophytes occur-
red on sand, gravel and larger substrates
(Sl>40) and In Type II stream reaches
macrophytes were found to occur on si It
dominated substrates (SK40).
All stream reaches were selected using
Identical criteria, designed to standardize
or minimize the physical Impacts exerted on
macrophyte growth. By limiting the physi-
cal Impacts on macrophyte growth the amount
of available nutrients would be the domi-
nant factor controlling macrophyte growth.
The data presented In Figure 2 Indicate
that both Type I and Type 11 streams are
capable of producing a high macrophyte bio-
mass but that Type 11 streams can produce
high macrophyte bloraass at relatively low
I n-stream phosphorus concentrations.
It has been established that macrophytes
can absorb nutrients from either the ambi-
ent water or from the sediment through
roots (Carlgnan and Kalff 1980; McRoy and
Barsdate 1970; McRoy, et al. 1972; Walsel
and Shapiro 1971).
A predictive empirical relationship was
developed by Carlgnan (1982) which Indl-
FIgure
Frequency of occurrence of macrophytes on
substrate sizes In Type I and Type II
Stream Reaches
I,M I
I.M N
II < 40 SI > 40 SI < 40 SI > 40
IIIITIITI
cates the probable source of macrophyte
nutrients. Carlgnan's (1982) model esti-
mates the percentage of plant tissue phos-
phorus taken up by the roots. This model
suggests that macrophytes obtain the nutri-
ents they require from the most plentiful
and readily available source.
- 16 -
-------
Figure 2
Plot of late summer macrophyte stable blomass (gm/m2) against the mean summer phosphate-
phosphorus conentratlon (POA In mg/l). Open circles are Type I stream data points and boxes
are Type II stream data points
600
500
01
ra 400 -
o
m
01
O 300
u
CO
200 -
100 -
O O
oo
o o
o
o
0.0
0.1
0.2 0.3 0.4 0.5
Mean Summer PO4P (mg/l)
0.6
0.7
Vaux (1962, 1968) studied the Interchange
of Intragravel (Interstitial) water with
the overlying water In streams. He found
this Interchange Is affected by stream bed
permeability (substrate size), the depth of
the material, the configuration of the
stream bed surface and channel slope.
If the permeability changes (e.g. de-
creases) down-welling of surface water will
occur Immediately downstream of a low-per-
meablllty area. This occurs on small
(point measurements) as well as larger
scales. Examples of localized exchange
were Illustrated by the Interchange re-
sulting from placement of rocks on the
streambed (Vaux 1968).
Stream channels, especially those with
larger bottom substrate sizes, usually con-
tain alternating channel slopes due to geo-
morphlc factors. Irregular substrate size
and distribution, shifting of substrates
and animal activities. This would result
In rapid and extensive exchange of Intra-
gravel and overlying surface water through-
out a stream reach In larger substrate
areas. The larger substrates In Type I
streams would then permit Intragravel flow
as well as rapid Interchange between over-
lying and Interstitial waters.
Based on the premise that macrophyte growth
Is a function of nutrient availability. It
appears, then, that In those streams clas-
- 17 -
-------
si fled as Type I, mecrophytes are essen-
tially offered overlying stream water as
the primary nutrient source, either through
shoot absorption or due to Intimate root
contact with percolating overlying water.
These data also suggest that macrophytes In
Type II reaches are utilizing an alternate
nutrient source. It seems probable then,
that sediments can fulfill a significant
portion of macrophyte phosphorus nutrition
In these streams which produce a large bio-
mass at low In-stream phosphate phosphorus
concentrations.
Evaluation of the stream reach mapping data
and the pore water/ambient water relation-
ship Indicated that two distinct Types of
streams were surveyed. Type I strean
reaches were found to have substrate domi-
nated by larger particle sizes (Sl>40),
macrophytes were relatively homogenousIy
distributed and the emblent water Is the
probable primary nutrient source being
utilized by the macrophytes. Type II
stream reaches had small size dominated
substrate (Sl<40), macrophyte occurrence
was often limited to zones of silt or silt
deposition and the sediments are the prob-
able primary nutrient source.
Sediment Nutrients
The 1981 sediment data were analyzed to
determine If significant relationships
existed between sediment nutrients and
stream macrophyte blomass. Multiple cor-
relation analyses Including pore water and
bulk sediment phosphorus and nitrogen con-
centrations, macrophyte bloreass, macrophyte
percent coverage, macrophyte tissue nitro-
gen and phosphorus, and macrophyte tissue
nitrogen to phosphorus ratio were conducted.
The only sediment parameters to correlate
significantly with macrophyte blomass
measurements were Interstitial phosphorus
In non-macrophyte areas with macrophyte
blomass per square meter (r = .638, p =
.002) and mean reach percent coverage (r =
.836, p = .0001). In-stream P04P concen-
trations also correlated with macrophyte
area and non-mocrophyte area Interstitial
phosphorus concentrations (r = .798, p =
.0001; and r = .859, p = .0001 respective-
ly). Reviewing the discussion of sediment
Interstitial water and overlying water ex-
change (Vaux 1962, 1968) It Is probable
that this relationship Is responsible for
the correlation between In-stream phos-
phorus and Interstitial water phosphorus
concentrations.
In general, sediment nitrogen did not cor-
relate wall with either In-stream nitrogen
or plant blomass parameters. Exceptions
were the correlations between stream In-
organic nitrogen with macrophyte and
non-macrophyte area Interstitial
N02-N03N concentrations (r = .959, p =
.0001; and r = .675, p - .0009 respective-
ly). The relationship between stream water
and Interstitial water nutrients again Is a
probable cause for these correlations.
Sediment nutrient data were also evaluated
to determine If there were significant dif-
ferences between sediment nutrient concen-
trations In macrophyte-populated areas and
non-macrophyte areas. The purpose of this
analysis was to determine If macrophytes
were colonizing areas that had higher con-
centrations of sediment phosphorus and
nitrogen or If the macrophytes had an ob-
vious Impact on pore water nutrient concen-
trations. If macrophytes were found to
colonize areas of higher nutrient concen-
trations, this would have provided sup-
portive evidence that the sediments may
have been the dominant macrophyte
nutritional source.
Growing season (June-September) mean sedi-
ment nitrogen and phosphorus concentrations
were computed for each reach from monthly
samples collected In and out of macrophyte
areas In 1981. Mean sediment phosphorus
and nitrogen concentrations within macro-
phyte (MSEDP and MSEDN) and out of macro-
phyte areas (NSEDP and NSEDN) were compared
using a t-test (significance level .05).
T-test values, sediment nitrogen and phos-
phorus concentrations In and out of macro-
phytes are given In Table 9. No signifi-
cant differences In nutrient concentrations
within and outside of macrophyte beds
within each stream were Indicated by the
analyses.
- 18 -
-------
Mean sediment Interstitial water (IW)
nutrient concentrations within and outside
of macrophyte beds are given In Table 10.
Correlations between Interstitial water and
bulk sediment parameters also showed no
clear relationships between these two sedi-
ment measurements.
The results of the sediment nutrient
analysis provided no clear relationships
between bulk sediment nutrients and stream
macrophytes. This suggests that In the
reaches studied macrophyte nutritional
needs are satisfied either directly through
shoot absorption or Indirectly through
water percolating through the substrate and
Into the root system.
Macrophyte Blomass and In-stream Nutrients
Based on the concepts of agriculture and
horticulture, that plant growth Is propoi—
tlonal to the amount of nutrients aval(able
for growth. It would be expected that
macrophyte growth could be modeled most
accurately for streon reaches where the
primary nutrient source has been quanti-
fied. That Is, the relationship between
macrophyte blomass and available nutrients
can be defined best for Type I streams
where the amount of aval (able nutrients has
been quantified (I.e. ambient water), and
not In Type II streams where the primary
nutrient source was not quantified (I.e.
sediments).
Table 9
Bulk Sediment Nutrient Concentrations. All Concentrations are
Annual Means, Expressed In tng/kg Sediment Dry Weight
Stream
•MSEDP
•NSEDP
tP(0.5)
•MSEDN
•NSEDN
tN(0.5)
Ashl p pun-Mai nstem
Ash! ppun-North Branch
Ashlppun-South Branch
Bark-Lurvey
Bark-Wolf
Sugar
565
507
452
no
185
572
487
325
332
133
157
374
0.58
1.44
l.ll
-0.71
0.64
1.17
3725
2825
2200
356
1280
2610
2225
1127
1330
246
925
1556
1.63
1.31
l.ll
0.72
0.51
0.89
* MSEDP and MSEDN represent phosphorus and nitrogen concentrations within macrophyte areas.
NSEDP and NSEON represent phosphorus and nitrogen concentrations outside of macrophyte areas.
Table 10
Sediment Interstitial Water Nutrient Concentrations.
All Concentrations are Expressed In mg/l
Stream
MIWP
NIWP
MNH3
NNH3
MN03
NN03
Ashlppun-Malnstem .003
As hi ppun-North Branch .049
Ashlppun-South Branch .012
Bark-Lurvey .950
Bark-Wolf .368
Sugar .026
.019
.024
.180
.250
.027
.032
.500
.220
.107
.075
.017
.176
.250
.090
.065
.075
.017
.176
.530
.760
.850
.380
.013
1.410
.760
.900
1.040
1.020
.660
1.210
- 19 -
-------
A preliminary correlation analysis In-
cluding all streams (data In Tables II and
12) Indicated that significant relation-
ships existed between late summer macro-
phyte blomass and mean summer (June-August)
In-stream nutrient concentrations. Total
phosphorus (TOTP) and phosphate-phosphorus
(P04P) concentrations had correlation co-
efficients of .642 (p=.003) and .686
(p=.OOI), respectively, with macrophyte-
blomass (59010). The natural log of
SOMBIO correlated significantly with the
natural logs of TOTP (r=.633, p=.004) and
P04P (r=.783. p=.OOOI). Total kjeldahl
nitrogen (TKN) and Inorganic-nitrogen
(INORN) were Insignificantly correlated
with SOMBIO, (r=.2!9, p=.367 and r=.403,
p=.088, respectively). The logarithmic
transformation of INORN and SQMBIO, how-
ever, did correlate significantly (r=.689,
p=.OOI).
The macrophyte bIamass/In-stream nutrient
concentration relationship Improved when
only the Type I streams were Included In
the correlation analysis. This analysis
was conducted under the premise that If the
two Types of streams (Type I and Type II)
exist, the relationships between plant blo-
mass and In-stream nutrients will Improve
when the streams were categorized. This
analysis Is also used as supportive evi-
dence for classifying the streams as Type I
or Type II. With Type II stream reaches
deleted (Mount Vernon Creek, Black Earth
Creek and Scuppernong River), the correla-
tion between SOMBIO, TOTP and P04P In-
creased to .889 (p=.OOOI) and .901
(p=.OOOI), respectively. The highest cor-
relation was found between the logarithmic
transformations of SOMBIO and P04P (r=.907,
p=.OOOI). Although the correlation between
SOMBIO and total kjeldahl nitrogen Improved
(r=.624, p=.OI), the relationship between
SOMBIO and INORN did not.
The correlation analysis Indicated that
significant relationships existed between
I n-stream phosphorus concentrations and
late summer blomass In Type I streams.
Predictive equations were developed from
these relationships, by regressing TOTP,
P04P and their logarithmic transformations
against macrophyte blomass. The most
statistically significant least squares
regression model (R-square-.823) was de-
veloped by regressing the natural log of
the maximum summer blomass on the natural
log of the mean summer P04P concentration
(Figure 3). The equation describing this
relationship Is:
Model I
SQMBIO = 546.8 (P04P)-415
where: SQMB'IO = Late summer blomass (grams
per square meter)
P04P = Mean summer (June-August)
phosphate-phosphorus
(ml 111 grans per liter)
The equation was developed from the data In
Tables II and 12 with a P04P concentration
range-of .002 to .430 ml 11 Igrms per liter
and macrophyte blomass from 25.6 to 448.5
grams per square meter. The mean residual
error for this regression Is 24.7 percent
and ranged from 1.8 to 87.0 percent of the
predicted values. The ninety-five percent
confidence limits for the predicted values
ranged from 53 to 114 percent of the pre-
dicted values. This equation appears to be
a good predictive tool for the assessment
of macrophyte communities In streat! reaches
where macrophytes derive phosphorus from
the water.
Macrophyte Tissue And In-Strean Nutrients
It has been shown that macrophyte growth Is
dependent upon tissue phosphorus concentra-
tions (Garloff and Krombholtz 1966; Wilson
1972). If the relationship between macro-
phyte tissue nutrients and In-stream nutri-
ent concentrations can be quantified. It
will provide supportive evidence that
macrophyte growth may be limited by con-
trolling In-stream nutrient concentra-
tions. This relationship would only be
quantifiable for streams where the amount
and source of available nutrients (I.e.
Type I streams) has been determined.
Macrophyte tissue phosphorus (PHOS) and
nitrogen (N) concentrations were highly
correlated with In-stream phosphorus and
- 20 -
-------
Figure 3
Regression line of late summer blomass (gm/m2) on mean summer phosphate-phosphorus
concentration (PO^ Inmg/l) for Type I streatis
600
c
-------
Table II
Stream
Ashlppun River
Bark Rlver-Lurvey
Bark River-Masonic
Bark R. -Uallschlaegar
Bark River-Wolf
Cedar Creek
Fox River-Portage
Mllw. R.-Campbellsport
Mllw. R.-Enat Branch
Hukuonago River
Pevaukee River
Sugar Creek
Aa. hlppun River
Bark Rlver-Lurvey
Bark River-Wolf
Sugar Creek
Black Earth Creek
Mount Vernon Creek
Scuppernong River
Given fc
Year
82
82
82
82
82
82
82
82
82
82
82
82
81
81
81
81
82
82
82
Stream
Type
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
II
II
II
Total Phosphate
Phosphorus mg/1 Phosphorus mg/1
Mean
.095
.159
.163
.033
.021
.050
.147
.493
.040
.013
.140
.104
.139
.244
.022
.122
.107
.050
.027
Std.Oev. N
.026
.053
.061
.012
.004
.017
.025
.186
.020
.006
.069
.026
.085
.051
.004
.024
.025
.010
.012
7
7
5
3
7
3
3
3
3
3
4
7
21
7
7
6
3
3
3
Mean
.032
.125
.128
.016
.003
.029
.078
.430
.015
.002
.074
.041
.050
.214
.002
.043
.052
.037
.009
Scd.Dev. N
.011
.055
.064
.006
.002
.012
.018
.174
.007
.000
.048
.021
.031
.051
.000
.008
.011
.008
.004
7
7
5
3
3
2
3
3
3
3
4
7
21
7
7
6
3
3
3
Total Kjeldahl Inorganic
Nitrogen mg/1 Nitrogen mg/1
Mean Std.Dev. N
1.03
.79
.78
.60
.74
.50
1.10
1.67
.80
.57
1.05
.81
1.15
.83
.69
.78
.43
.20
.83
.21
.02
.11
.20
.10
.10
.17
.31
.20
.06
.30
.26
.34
.11
.14
.09
.15
.00
.45
7
7
5
3
3
3
3
3
3
3
4
7
.21
7
7
6
3
3
3
Mean
TToT
.74
.72
1.95
.11
.71
.62
.56
.33
.03
.48
3.24
1.07
1.18
.09
1.51
2.15
4.28
.83
Std.Dev. N
.27
.34
.26
.05
.11
.21
.24
.07
.01
.19
.80
.28
.38
.04
.13
.12
.12
.05
7
5
3
3
3
3
3
3
3
4
7
21
7
7
6
3
3
3
-------
Table 12
Summer
Stream
Ashlppun River
Bark Rlver-Lurvey
Bark River-Masonic
Bark Rlver-Wallschlaeger
Bark River-Wolf
Cedar Creek
Fox River-Portage
Milwaukee Rlver-Campbellsport
Milwaukee River-East Branch
Mukwonago River
Pewaukee River
Sugar Creek
Ashlppun River
Bark Rlver-Lurvey
Bark River-Wolf
Sugar Creek
Black Earth Creek
Mount Vernon Creek
Scuppernong River
Stream
Year Type
82 I
82 I
82 I
82 I
82 I
82 I
82 I
82 I
82 I
82 I
82 I
82 I
81 I
81 I
81 I
81 I
82 II
82 II
82 II
can), Standard ueviacion taco. uev.j. and Number of Samples (N) .
Maximum Blomaas Tissue Pho aphonia
go/in* Km/Kg
Mean
12577
289.0
187.5
179.3
25.6
161.0
146.9
448.5
72.5
57.8
213.8
134.4
98.9
249.8
45.9
151.4
282.1
365.5
262.6
Std.Dev.
147.7
240.9
108.3
217.3
41.7
85.3
121.0
508.9
60.0
58.3
317.2
130.6
74.5
253.1
86.5
132.4
334.5
503.8
252.8
N
44
67
36
36
44
44
50
SO
48
55
40
24
26
34
30
24
46
44
48
Mean
5757
4.83
4.63
2.75
1.42
2.40
4.47
6.68
3.73
1.13
4.01
4.10
3.81
5.41
1.38
3.60
5.81
5.48
2.44
Std.Dev.
.45
.87
.41
.20
.22
.41
.25
.92
.52
.31
.70
.41
.96
.25
.24
.96
.33
.58
N
5
7
5
5
3
4
6
3
5
3
4
3
6
7
4
4
6
4
5
r Miogramj
.
Tissue Nitrogen
gm/Kg
Mean
29722
31.66
32.08
27.58
21.37
24.57
26.88
30.93
26.12
15.67
23.80
29.83
28.47
27.03
20.05
25.92
26.17
34.12
22.14
Std.Dev.
3.04
1.66
2.73
3.01
2.77
1.65
2.03
2.79
3.93
1.78
.97
2.48
1.85
3.06
1.03
2.75
.64
.58
N
5
7
5
5
3
4
6
3
5
3
4
3
6
7
4
4
6
4
5
-------
This equation (Figure 4) has an R-square
value of .921 and a mean residual errer of
9.5 percent of the predicted value. The
ninety-five percent confidence Intervals
for the predicted values range from 30 to
43 percent of the predicted.values.
Model II Indicates that macrophyte tissue
phosphorus Is a function of the In-stream-
phosphate-phosphorus concentration In Type
I streams. Based on the concept that macro-
phyte growth Is dependent upon macrophyte
tissue nutrient concentrations, this model
provides supportive evidence that In Type I
streams macrophyte growth Is a function of
In-strem phosphate-phosphorus concentra-
tions.
Macrophyte Blomass and Tissue Nutrients
Establishing an empirical relationship be-
tween late summer macrophyte blcmass and
Figure 4
Regression line of the macrophyte tissue phosphorus concentration (gmAg) on the mean summer
DhosDhate-ohosDhorus concentration (P04 In mg/.l) for Type I streams
Modelll 310
PHOS= 9469P04P
R2= .921
N= 16
0.0
0.2
0.3 0.4 0.5
Mean Summer P04P (mg/l)
0.6
0.7
- 22 -
-------
macrophyte tissue nutrient concentrations
could provide a methodology to estimate
macrophyte blomass Independent of nutrient
source. This type of model would be of
significant utility In streams where It
would be difficult to determine the sources
and amounts of nutrients available for mac-
rophyte growth.
Macrophyte tissue phosphorus and tissue
nitrogen correlated significantly with mac-
rophyte blomass. The correlation analysis
evaluating the relationships between macro-
phyte tissue phosphorus, macrophyte tissue
nitrogen, and plant blomass Included the
data from all streams surveyed (Table 12).
The correlation coefficients are .798
(p=.OOOI) and .634 (p=.004) for the log-
arithmic transformations of SQMBIO:PHOS and
SQMBIO:N relationships, respectively.
Based on these correlations, an equation
was developed to describe the relationship
between macrophyte tissue phosphorus con-
centration at late summer blomass and sum-
mer maximum blomass. All the data points
In Table 12 were Included In this least
squares regression analysis. The equation
describing this relationship Is (Figure 5):
Mode I 11 I
SQMBIO = 36.06 (PHOS)1'161
where: SCMBIO = Late summer macrophyte
blomass (grams per square
meter)
PHOS - Macrophyte tissue
phosphorus concentration
(grams per kilogram dry
weight)
The R-square for the model Is .637 and the
mean residual error Is 36.4 percent of the
predicted values. The mean ninety-five
percent Intervals for the predicted values
were 63.7 to 175.9 percent of the predicted
values. This model has the widest con-
fidence limits of the three equations de-
veloped by these analyses. A probable
cause for the wide confidence limits for
this model may be that macrophyte tissue
phosphorus concentrations rapidly decrease
In senesclng macrophytes. It has been de-
termined that from 20 to 50 percent of the
tissue phosphorus can be rapidly lost from
decaying macrophytes, and 65 to 85 percent
may be lost over longer periods (Nichols
and Keeney 1973; SoI ski 1962, In Wetzel).
It was noted In sampling that a few of the
macrophyte populations were In the process
of senescence at the time of harvest. This
model may have Improved substantially If
all harvesting had been conducted before
any macrophyte populations began to
deteriorate.
A number of Investigators (e.g. Gerloff
1975. Gerloff & Krombholz 1966) have
suggested tissue nutrient concentrations at
which a particular nutrient becomes limit-
ing to growth. Gerloff (1975) has sug-
gested approximately ,\% tissue phosphorus
as limiting. Schmidt and Adams (1981) have
reported P limitation at about .32. The
tissue nutrient concentrations reported
here are somewhat above the .1 level, even
though significant relationships are de-
scribed between water P, tissue P. and
plant blomass. This may be due to dif-
ferences In the tissue nutrients of dif-
ferent parts of the plant. Gerloff's
(1975) values are taken from apical meri-
stem tissue whereas this study used whole
plants (Including roots) for analysis.
Model Selection
The primary objective of this portion of
the study was to develop an empirical
relationship predicting macrophyte blomass
(summer standing crop) In small streams.
This objective was accomplished with the
development of Models I and III. The dif-
ferent variables used to derive these
models make their applicability dependent
upon stream Type (Type I or Type II).
Model I was developed from the relationship
between In-stream phosphate-phosphorus and
late summer blomass. Model III Is derived
from the relationship between macrophyte
tissue phosphorus and plant blomass, and Is
Independent of nutrient source (I.e. water
or sediment). Model I should be used when-
ever It Is applicable as It Is a much more
- 23 -
-------
Figure 5
Regression line of summer macrophyte blomass (gm/m2) on the macrophyte tissue phosphorus
concentration (gm/kg) for Type I and Type II streans
600
- 500
CM
-I
y.
1
o
m
400-|
a
2 300
200 •
100 •
Model III
SQMBIO= 3606 PHOS
R2= 637
N = 19
1 161
o o
1.0
2.0
—I—
3.0
5.0
6.0
7.0
Macrophyte Tissue Phosphorus (gm/kg)
accurate mode I than I s Mode I III. It Is
obvious that Model I will most accurately
predict macrophyte growth In Type I streams
(Figure 3) and Model III (Figure 5) Is the
best available method to evaluate blomass
I n Type 11 streams.
For Model I or Model 111 to become accepted
strean management tools, a methodology must
be developed which determines the primary
nutrient source for stream macrophytes.
The best methodology for determining pri-
mary nutrient source would be to use the
model developed by Carlgnan (1982). This
method though has not been evaluated for
streams and would require verification be-
fore It could be widely used.
Field observations describing the distri-
bution of macrophytes could also be used to
Indicate the most probable macrophyte
nutrient source. Macrophyte populations
with a relatively homogeneous distribution
In a stream reach having a high percentage
of large size bottom substrates (Sl>40) are
believed to be Indicative of Type I
streans. Streams having macrophyte popu-
lations limited to zones of silt deposition
are characteristic of Type II streans.
Tissue analysis Is based on the assumption
that nutrient concentrations In an organism
are representative of the onount of nutri-
ents available to the organism for growth.
The theory of tissue analysis has been dis-
cussed by several authors (Lundegardh 1951;
Ulrlch 1952; Bould. et al. I960; Smith
1962; Chapmen 1966). Based on these as-
sumptions, the results of this study may
provide a suitable alternative method for
- 24 -
-------
determining the primary macrophyte nutrient
source. Model II Indicates that macrophyte
tissue phosphorus concentration (PHOS) Is
dependent upon the In-strean mean summer
phosphate-phosphorus concentration (P04P)
In Type I streams. This relationship
suggests that Model II may be used to
Identify the conditions where the ambient
water Is the primary nutrient source.
Figure 6 Indicates that In Type II streans.
macrophytes can have high tissue phosphorus
at low stream P04P, which suggests they are
obtaining tissue phosphorus from the sedi-
ments. If macrophytes are using ambient
water as a primary nutrient source they
should belong to the relationship described
by Model II. Stream data points lying near
or outside of the average upper confidence
limits of Model II would be a strong Indi-
cation that these streans do not belong to
this relationship and the ambient water Is
not the major source of tissue phosphorus.
Figure 6
Regression line of Model II Including the estimated upper ninety-five percent confidence limit.
Ooen circles are Type I stream data points and boxes are Type 11-stream data points
Model ii 210
PHOS = 9.469 P04P '
o.i
0.2
0.3 0.4 0.5
Mean Summer PC^P (mg/l)
0.6
0.7
- 25 -
-------
The above methodology can be used to deter-
mine If Model I or Model III would be the
best management tool to predict late summer
iracrophyte blomass for a given stream
reach. If the data for a given stream
reach, when plotted on Figure 6, fall with-
in the upper confidence Interval estab-
lished for Model II and macrophyte distri-
bution and substrate type are characteris-
tic of Type I streams, then It would be
strong evidence that Model I would be the
best model available to predict macrophyte
blomass. If the conditions for Model I are
not met, then Model III should be used to
predict macrophyte blomass, as It was de-
veloped under the premise that blomass Is
predictable Independent of nutrient source
(I.e. water or sediment). The confidence
Intervals for Model III, however, are much
wider than those for Model I and the pre-
dicted values from this model would have a
lesser degree of accuracy associated with
them.
Summary and Conclusions
Empirical relationships were developed that
describe the responses of macrophyte com-
munities to a range of In-stream phosphorus
concentrations. The primary purpose for
quantifying these relationships was to de-
velop a predictive tool to estimate late
summer macrophyte blomass In selected
stream reaches.
It was evident from the data analysis that
macrophyte growth could not be predicted
from In-stream nutrient concentrations for
all stream reaches that were surveyed.
Streams were classified by substrate type,
macrophyte distribution and apparent macro-
phyte nutrient source. Type I stream
reaches are characterized by having
sand-gravel-rubble bottom substrate, shal-
low depth, and relatively homogenous macro-
phyte distribution. Macrophytes In Type I
streams are believed to be utilizing the
ambient water as their primary nutrient
source. Type 11 stream reaches had
silt-sand substrate, shallow depth and mac-
rophyte occurrence was limited to zones of
silt deposition. Bottom sediments are
believed to be the primary nutrient source
In Type II streams. All streams received
little or no shade.
The results of the 1981 sediment surveys
suggest that Interstitial water nutrients
are closely related to the nutrient concen-
trations In the overlying water. This
relationship has been demonstrated by Vaux
(1962) for streams with large substrates.
The analyses of the 1981 sediment and
Interstitial water data Indicated that
there were . not significant differences In
the phosphorus and nitrogen concentrations
In and out of areas colonized by macro-
phytes. This analysis suggests that for
Type I streams macrophytes are not
colonizing areas of nutrient rich sub-
strate. These findings are supportive of
the hypothesis that for Type I streams mac-
rophytes are utilizing the ambient water as
the primary nutrient source.
It was apparent that macrophyte growth
responses could be modeled best for
situations where the major nutrient source
was quantified. Model I and Model II were
developed with the data collected from the
Type I stream reaches. Model I Is linear
regression equation that estimates macro-
phyte blomass from mean summer In-stream
phosphate-phosphorus concentrations. The
equation describing Model II was developed
by regressing the average macrophyte tissue
phosphorus concentration on the mean summer
I n-strean phosphate-phosphorus concen-
tration. Model I and Model II fit the data
sets they were developed from very well,
with R-square values of .823 and .921,
respectively.
Model III Is a linear regression equation
developed to describe the relationship be-
tween macrophyte blomass and macrophyte
tissue phosphorus concentrations. Data from
both Type I and Type 11 stream reaches were
used to calculate this model as this
relationship was considered to be un-
affected by the source of nutrients that
the macrophytes were utilizing. Model III
does not fit the data as well as Model I or
Model II, It has an R-square value of .638.
- 26 -
-------
Model I and Model III are predictive
equations that estimate late summer blomass
In selected small stream reaches. Model II
may be used along with stream reach mapping
data to determine which model Mill provide
the best estimate of macrophyte blomass.
Model I will provide the best blomass es-
timates for stream reaches classified as
Type I stream reaches. Model III can pro-
vide a methodology to estimate blomass In
Type 11 stream reaches.
In order to provide useable management
tools, we wanted to derive empirical models
which were at least as statistically sig-
nificant as models currently being used by
lake managers. An analysis of the pre-
cision of various lake phosphorus loading
models was conducted by Canfleld and
Bachmann (1981) using a data set of 704
lakes. Models evaluated were Canfleld and
Bachmann (1981), Larsen and Mercler (1976),
Jones and Bachmann (1976), Reckhow (1979),
Klrchner and Dillon (1975), Chapra (1975)
and Vollenwelder (1975).
The most precise model evaluated In this
group was that of Canfleld and Bachmann
(1981) which had an R-square of .69.
average residual error of 38 percent and
confidence limits (p=.05) of 31-288 per-
cent. The range of the average error for
the rest of the models was 42 to 63 percent
with confidence limits (p=.05) ranging from
15 to 599 percent. The precision of the
models developed from the preceding data
analyses compares favorably with lake
models currently being used.
Of the models developed from this study.
Model I may have the most significant Im-
pact on the management of water quality In
small streams. This model has the pre-
dictive capabilities to estimate changes In
macrophyte blomass when mean summer
In-stream phosphate-phosphorus concen-
trations are changed. At this point how-
ever. Model I as well as the other models
are untested and therefore their appli-
cation should be limited to the stream
reaches that the models were developed
from. Before these models can become
accepted water quality management tools.
they should be substantiated by applying
them to Independent data sets. The test
data set must be collected using Identical
criteria as were used to collect the data
set the models were developed from.
PERIPHYTON
Introduction
Research directed at defining perlphyton
response to stream enrichment has utilized
a variety of approaches and methods. These
Include plant pigment, gravimetric and en-
zyme analyses, cell counts, community
species composition, and occurrence of In-
dicator species. Employing many of these
types of analyses In routine water quality
management activities Is usually not prac-
tical, due to time or budget constraints
and the complexity of many of the analyses.
Perlphyton analyses such as chlorophyIl-a
and gravimetric (e.g. ash-free weight blo-
mass) estimates have been used In routine
monitoring programs. These analyses have
the advantage of being relatively Inexpen-
sive and commonly used. Collection and
analysis techniques have also been de-
veloped to the extent of providing rela-
tively uniform sample quality assurance and
comparability of data. This Includes sam-
pling equipment (such as glass-slide sam-
plers). Incubation (exposure) times and
sample handling and preservation. Data
evaluation and Interpretation, however. Is
still dependent on the Investigator's
skill, experience and personal preferences.
The Influence of physical factors such as
light, temperature and water velocity on
the resulting blomass estimates Is, how-
ever, poorly documented. This could result
In Inaccurate or poor correlation with
growth nutrients.
The purpose of the perlphyton element In
the Phosphorus Assessment study was mainly
to characterize growth of perlphyton com-
munities In streams and compare this to
stream chemical and physical characteris-
tics. Specific objectives Included:
- 27 -
-------
- Augmenting macraphyte collections to
estimate stream primary production (In-
cluding photosynthesis and respiration
estimates)(see die I section!; and
- Evaluating the ebllWy of sanpUng
methodologies and conventional para-
meters as tools to assess stream nutri-
ent status. This Included glass-slide
samplers as well.as samples collected
from brick substrates.
As with the macrophyte data, correlation
matrices and least squares regression
models were calculated to evaluate which
physical and chemical parameters correlated
best with the perlphyton data. Regression
equations were only calculated for paired
parameters having significant correlation
coefficients.
The data used In the development of- the
perlphyton models (brick collections) are
presented In Table 13.
Results and Discussion
Perlphyton Blomass and In-stream Nutrients
Perlphytometers
Monthly means were calculated from the 1981
and 1982 perlphy*oneter -data. • Perlphyto-
meter chlorophyll-a concentrations cor-
related positively with In-strean phos-
phorus and nitrogen concentrations. These
relationships Improved with transformation
to their natural logs (Table 14). The!
natural log (In) of perlphytometer chloro-
phyl l-a was correlated most strongly with
In TOTP (r*.635, p=.OOOI. n=76) and
Table 13
Brick Data Used to Calculate Models IV. V and VI
I I
PO..P Values are In mg/l. Brick Chla In mg/mz. and
Bark - Wolf
Bark - Lurvey
Bark - Masonic
Kohlsvllle
Sugar
Brick
6
7
8
9
10
6
7
8
9
10
7
8
9
10
6
7
6
7
8
9
Tissue Phosphorus
PO*P
.003
.002
.002
.002
.002
.095
.116
.119
.202
.260
.108
.110
.208
.237
.040
.045
.053
.050
.043
.031
In mg/gm Dry Weight
Chi -a
20
13
13
18
15
33
68
63
96
140
121
224
146
400
54
60
103
62
59
55
Tlssue-P
.67
.51
.46
.39
.28
.96
1.10
1.83
1.84
1.20
2.22
2.28
2.56
1.50
1.06
1.37
1.57
1.65
1.42
1.41
1
2 |
2
3
2
1
2
2
3
2
i ;
i ,
3 '
2
1
2
2
2
2
3
1
- 28 -
-------
lnP04P (r=.603, p=.OOOI. n=76). Perl-
phytometer In chlorophyll-e : In Inorganic
nitrogen and In total nitrogen correlation
coefficients were r=.485 (p=.OOOI, n=79)
and r=.337 (p=.003, n=76) respectively.
Rerlphytcmeter Dry Weights and Ash-free
Weights did not significantly correlate
(p .05, n=76) with In-stream nutrient
values.
Bricks
Monthly mean values were calculated from
the 1981 and 1982 data. Ch I orophy 11 -a and
nutrient data were collected only In 1982.
In general, brick perlphyton blomass esti-
mates exhibited higher correlation coeffi-
cients with In-stream nutrients than perl-
phytometer estimates. As with the perl-
phytcmeter values, correlations Improved
with natural log transformations. The
natural log of brick chlorophyll-a col-
lections with InTOTP and lnP04P corre-
lates were r=.879 (p=.OOOI, n=2l) and
r=.875 (p=.OOOI, n=2l) respectively (Table
14). The natural logs of brick chlorophyll
correlated significantly
Table 14
Correlation Coefficients for Monthly Mean
Water Chemistry Values and Perlphyton
Blomass Estimates
InTOTP lnP04P InTKN InlNORN
InPAFWT .453 .413 .273 .413
InPUCCHLA .635 .603 .343 .479
InBAFWT -.203 -.060 -.405 .387
InBUCHLA .879 .875 .036 .648
Perlphytometer Values Included 1981-1982
Data and Brick Values Included 1982 Data
Perlphyton parameters are noted as:
Perlphytometer Ash-free Weight (PAFWT).
Perlphytometer Chlorophyll-a (PUCCHLA),
Brick Ash-free Weight (BAFWT), and Brick
Chlorophyll-a (BUCHLA)
with In Inorganic nitrogen (r=.648, p=.004,
n=2l), but not with In total nitrogen
(r=.090).
Model Development - Perlphyton
Blomass and In-Stream Nutrients
The correlation analyses Indicated that a
significant relationship existed between
In-stream P04P and perlphyton chloro-
phyl l-a concentrations. Between the perl-
phytcmeter and brick harvests, the brick
chlorophyll a: P04P relationships were
most significant. A least squares regres-
sion equation was then calculated for brick
chlorophyl-l-a and stream P04P. The curve
representing this relationship Is presented
In Figure 7. The equation describing this
model Is:
MODEL IV
BRICK CHLOROPHYLL-A = 258.68(P04P)«453
where: BRICK CHLOROPHYLL-A Is In mg/m2
P04P • stream P04P concentration
In mg/l
This model has an R-square of .766
(p=.OOOI, n-20). Mean residual error Is
36.24$, and ranged from .60 to 214.40$ of
the predicted values. The mean upper and
lower 95$ confidence limits are 267$ and
37$ respectively.
Perlphyton Tissue Nutrients and In-stream
Nutrients
Perlphytometers
In general, per Iphytcmeter tissue nutrients
were positively correlated with In-stream
phosphorus and poorly correlated with
In-stream nitrogen. Relationships Improved
when natural log transformations were made.
The best correlations occurred between
Perlphytometer tissue phosphorus with
InTOTP (r=.375, p=.0006, n=74) and lnP04P
(r=.352, p=.0008, n=74). Perlphyton tissue
N was not significantly correlated with
either In-stream nitrogen or phosphorus
values (Table 15).
- 29 -
-------
Figure 7
Regression line of monthly mean brick perlphyton chlorophyll-a (mg/m2) on monthly mean stream
POAP concentrations (ing/1)
250
(400)
Chi a =25B.«8(P04P)
.455
r2s .766
n a 21
.2
.3
P04P
The correlations between perlphytcmeter
tissue nutrient concentrations
-------
Table 15
Correlation Coefficients for Monthly Mean
Mater Chemistry Values and Perlphyton
Tissue Nutrient Concentrations
InTOTP lnPOAP InTKN InlNORN
InPCAFW .375 .352 .184 .194
InNCAFW .226 .179 .092 -.019
InBRIKN .848 .868 .130 .738
InBRIKP .839 .864 .286 .819
Perlphyton parameters are as noted:
Perlphytometer Tissue Phosphorus
Concentration (PCAFW), Perlphytometer
Tissue Nitrogen Concentration (NCAFW),
Brick Tissue Nitrogen Concentration (BRIKN)
and Brick Tissue Phosphorus Concentration
(BRIKP)
Tissue phosphorus and nitrogen had higher
correlation coefficients with In-strean
phosphorus than In-stream nitrogen. Brick
tissue phosphorus correlated best with
lnP04P (r=.864. p=.OOOI, n=20), InTOTP
(r=.839, p=.OOOI, n=20) and InlNORN
(r=.8l9. p=.0009, n=20). Similarly, Brick
tissue N correlated best with lnP04P
(r=.868, p=.OOOI, n=20>. InTOTP (r=.848.
p=.OOOI, n=20) and InlNORN (r=.738, p-.OI,
n=20).
Based on these correlates a least squares
regression equation was calculated
describing the relationships between brick
tissue phosphorus and In-strean P04P con-
centrations (Figure 8). This equation Is:
MODEL V
BRICK TISSUE PHOSPHORUS = 3.07(P04P)-230
where: TISSUE PHOSPHORUS Is In mgP/g dry
weight
P04P = Instrem P04P concentrations In mg/l
This model has an R-square of .747. Mean
residual error Is 26.76$, ranging from 6.25
to 53.27$ of the predicted values. The mean
upper and lower 95$ confidence limits are
203.66 and 49.031 respectively.
Perlphyton Blomass and Tissue Nutrients
The correlation coefficients of the perl-
phyton blomass: tissue nutrient concen-
tration parameters are listed In Table 16.
The correlation coefficients derived from
brick parameters were substantially higher
than those derived from perlphytometers.
The rather large negative correlations
between nutrient concentrations derived
from perlphytometer and brick ash-free
weights are anomolles. The negative cor-
relations could be due to shading of the
bricks by macrophytes or, as mentioned
above. Inaccurate approximation of the
perlphyton nutrient concentrations.
Table 16
Correlation Coefficients for Perlphyton
Nutrient Concentration and
Blomass Estimates
InPCAFW InNCAFW InBRIKP InBRIKN
InPAFWT .335 .258 .234 .109
InPUCCHLA .529 .415 .637 .540
InBAFWT -.869 -.930 .713 .531
InBUCHLA .793 .652 .831 .826
Perlphytometer data Included 1981-1982,
Brick data Included 1982 only.
Perlphyton parameters are noted as those In
Tables 14 A 15.
The best reasonable (I.e. positive)
relationship between a tissue nutrient con-
centration and a blomass measurement was
between brick chlorophylI-a and brick
tissue phosphorus. The model calculated
for this relationship Is (Figure 9):
- 31 -
-------
Ffgure 8
Regression line of monthly mean brick perlphyton tissue P on monthly mean stream PO^P
concentratIons (ing/I)
2.5-
O
a
2.0-
1.5 -
u 1.0
m
.5 -
Tissue PC 3.07(PO4P)
.230
r2s.747
n a 21
.2
I
.3
P04P (mg/,)
- 3la -
-------
Figure 9
Regression line of monthly mean brick perlphyton tissue phosphorus concentrations (mg/gm) on
monthly mean stream PO^P concentrations (mg/l)
250-
200 -
ISO -
o>
u
so -
(400)
Chla= 47.73 (Tissue P)
.1.245
's 691
n a 21
oo
10
MODEL VI
BRICK CHL-A = 47.73! (BRICK TISS-P)'-245
where: BRICK CHL-A Is In mg/m2
BRICK TISS-P Is In mgP/g dry weight
This model has an R-square value of 0.69I.
Mean residual error Is 49.82, and ranges
from 5.2J to 406$ of the predicted values.
The mean upper and lower 95? confidence
limits of the predicted value are 3252 and
I 1.42 respectively.
20
30
Tissue P (mg/ )
Other Aspects of Perlphyton Growth and
Measurement
The ratio of ash-free weight (AFW gm/m2) to
uncorrected chlorophyll-a (UCCHLA, mg/m2)
Is known as the Autotrophlc Index (AD.
Autotrophlc Index values are generally In-
terpreted as Indicators of the trophic
level of perlphyton communities. High Al
values (>200) are found In heterotroph-
domlnated communities and lower values are
found where autotrophlc organisms are
dominant (APHA, et al. I98I).
- 32 -
-------
The' per I phytcmeter-der I ved AI values were
negatively correlated with the natural logs
of several nutrient parameters (e.g. InPOAP
r=»-.55, p<.OI and InlNORN, r=-.48,
p<.OI). This Indicates that perlphyton
communities become more autotrophlc as
nutrient levels Increase. The
brick-derived AI values were negatively
correlated with PTOT (r=-.46, p<.OI) and
P04P (r»-.46. p<.OI), but were Insignifi-
cantly correlated with In stream nitrogen
concentrations (INORN, TOTN, and NH3N).
The brick-derived AI values were substan-
tially larger than those derived from perl-
phytometers (Table 17). The reason for
this Is uncertain, but may be related to
colonization time (two weeks for the perl-
phytometers, four weeks for the bricks) or
other factors such as depth, velocity or
substrate carrying capacity.
A study comparing the perlphyton on artifi-
cial and natural substrates was conducted
(Babros 1981). The study was carried out
on the Kohlsvllie River (perlphyton-dcm-
Inated stream), and concluded that perl-
phytometers estimate chlorophyll-a accept-
ably, but underestimate the AFW of natural
substrates. The Al values for the natural
substrate samples were much higher than any
found on perlphytometers. The mean of
three AI estimates (using pheophytln-cor-
rected chlorophylls) was 950. The use of
unconnected chlorophy11 values would de-
crease this value to approximately 500,
which Is still much higher than the cor-
responding per Iphytometer estimates of
125-228.
Table 17
Mean Values for Autotrophlc Index Values,
as Estimated from the Bricks (BAD and
Perlphytometers (PAD (The values given
are the means of all samples available)
Site
BAI
PA I
Bark-Wolf
Bark-Lurvey
Bark-Masonic
Kohlsvllle
Sugar Creek
388
231
123
368
382
261
85
59
173
137
Velocities were determined at perlphyto-
meters when the slides were placed and
collected. Correlation of ash-free weight
and chlorophyll values with velocity (using
all available data) did not yield signifi-
cant results. On two dates, however, the
perlphytcmeters did show a negative cor-
relation of ash-free weight with velocity.
R-squared values for the regression of
velocity against ash-free weight values
were 0.95 and 0.49, n = 6 In both cases.
Each of these regressions used perlphyto-
meters which were exposed to Identical
nutrient and light conditions.
Some concern has been expressed as to
whether or not ambient phosphate concen-
trations In small streams are naturally
high enough to "saturate" perlphyton
growth. The positive correlations of ortho
and total phosphate concentration with both
brick and perIphytcmeter estimates of
chlorophy11-a would seem to Indicate other-
wise. Other studies and measurements
designed to estimate a "saturation level"
offer further evidence that Increased phos-
phorus levels will lead to Increased perl-
phyton growth In small streams (Auer and
Canale 1982; Rosemarln 1982; Lehman, et al.
1976).
Summary
Horner and Welch (1981) have shown that
equations can be developed to predict
chlorophyIl-a from temperature, velocity
and phosphorus concentrations. A total of
six equations were developed, each of which
was only applicable to a particular sample
period and velocity range. The coeffi-
cients of the parameters (temperature,
velocity and phosphorus concentration) ex-
hibited substantial differences between
colonization periods and velocity ranges.
Intensive studies designed to describe the
growth of a single species of algae
(Cladophora glomerulata) In the littoral
region of the Great Lakes further demon-
strate the complexity of perlphyton growth
(Auer, et al. 1982).
The results of this study and others (Auer,
et al. 1982; Horner and Welch 1981)
-------
strongly suggest that Increased nutrient
levels, particularly phosphorus, will
stimulate the growth of perlphyton In small
streams. The Increase In blomass will be
modified by many other factors. Including
temperature, light, velocity and Inherent
characteristics of the dominant species or
community (e.g. resistance to sloughing).
Because of the variability of the Impor-
tance of these factors. It does not appear
feasible to develop a "universal model"
capable of accurately predicting area!
perlphyton blomass In small streams. Site
specific models, however, appear to be
relatively accurate and easily obtainable.
DIEL DISSOLVED OXYGEN STUDIES
INTRODUCTION
It Is convenient to divide the factors
responsible for dlel stream dissolved oxy-
gen (DO) fluctuations Into two general
categories; metabolic and physical. The
metabolic category consists of plant and
animal metabolic activity. The net effects
can be positive or negative (DO production
or consumption) depending on time of day
and biological community composition. The
physical category Includes the effects of
reaeratlon, loading from tributaries and
groundHater Inflow. Loadings from tribu-
taries and ground water Inflow were not
obvious at the study sites, and were not
considered In this study. The Instantaneous
contribution of reaeratlon to stream DO
content Is positive when the water Is
undersaturated and negative when the water
Is oversaturated with respect to DO. The
magnitude of the deficit or surfeit will
determine the magnitude of the Instan-
taneous rate.
Figure 10 represents a simple path by which
the Impact of phosphorus on stream dis-
solved oxygen levels can be examined. The
substances which are'dissolved In the water
are generally expressed as concentrations
(e.g., mg/l dissolved oxygen). The blomass
quantities are usually expressed as "weight
per unit area" (e.g. grams of macrophyte
dry weight per square meter). The rates of
photosynthesis (P) and respiration (R) can
be measured In light and dark enclosures
(bottles or boxes), or approximated from
dlel surveys. These rates can be expressed
as units of oxygen produced or consumed
per unit blomass per unit time (grams oxy-
gen Ag dry welght/hr). For the measurement
of photosynthesis, a light level must be
defined. The primary producer blomass
Includes both macrophytes and perlphyton.
Figure 10
Path diagram of factors which regulate stream dissolved oxygen concentration.
Phosphorus & -
Other Nutrients
nutrient
Light.
Substrate
Primary Producers
Death,
grazing
Stream DO
K2
Atmospheric
Oxygen
recyc11ng
Detrltal and Consumer Blomass
(e.g. bacteria, fungi. Invertebrates,
fish, etc. and the organic matter
they consume)
- 34 -
-------
and the consumer blomass Includes a rather
diverse group of organisms, such as bac-
teria. Invertebrates and fish. The respi-
ration or photosynthesis rates per unit
blomass can be expected to vary for dif-
ferent organisms, as well as with life
stage, light, temperature, etc.
Over the course of a die I survey (about 24
hrs), much of this variation can be Ignored
If the community composition Is assumed to
remain constant. Community respiration Is
the sum of all types of respiration, and
community photosynthesis Is the sum of all
types of photosynthesis. The measured rate
of change of strean dissolved oxygen Is due
to community photosynthesis and respiration
and therefore dlel curve analysis yields
community rates.
The other major process which Impacts
stream dissolved oxygen Is the reaaratlon
rate. K2 Is a constant which expresses the
proportion of a deficit which will be
satisfied per unit time. In order to
specify a rate (quantity per unit time) due
to reaeratlon, this constant must be multi-
plied by a deficit (Cs-Co). Cs represents
the saturation concentration of dissolved
oxygen, which Is calculated from tempera-
ture data assuming normal atmospheric pres-
sure (760 mm Hg). Co Is the measured con-
centration of dissolved oxygen. When Cs >
Co, a deficit exists, and the product
K2(Cs-Co) should be positive. Indicating
that oxygen Is being gained by the stream.
When Co > Cs, the product K2(Cs-Co) should
be negative. Indicating that oxygen Is
being lost to the atmosphere.
The most notable aspect of Figure 10 Is
that phosphorus only directly Impacts the
primary producers (macrophytes and perl-
phyton). This relationship has been ex-
plored In detail and has resulted In
various models (Canale and Auer 1982,
Dillon and Rlgler 1974, Jones and Bachman
1976) as well as the phosphorus and macro-
phyte blomass model presented In this re-
port. Figure 10 then represents a simple
path by which the Impact of phosphorus on
stream dissolved oxygen levels can be
examined'.
Primary producers are capable of removing
oxygen from the water as we 11 as add I ng
It. Other factors, such as reaeratlon and
non-photosynthetlc organisms, also
Influence stream dissolved oxygen content.
The relative Importance of each of these
factors Is site specific, but some
generalizations are possible.
Several authors have remarked that small
streams tend to be net consumers of dis-
solved oxygen (e.g. Hynes 1970, Vannote et
al. 1980). .On a dally basis, respiration
tends to exceed production. This suggests
that respiration In small streams Is not
strictly a function of primary producer
blomass, since a positive net production Is
required for the accumulation of plant blo-
mass. If the primary producer blomass Is
getting larger (as demonstrated by growth
on perlphytcmeters and seasonal Increases
In the harvested macrophyte blomass) and
community respiration Is larger than com-
munity photosynthesis, a significant part
of the community respiration must be due to
non-photosynthetlc (consumer) organisms.
The relationship between phosphorus and
community respiration thus seems likely to
be naturally variable due to differing
amounts of respiration attributable to con-
sumer organisms (Invertebrates, fish, bac-
teria, etc.) which are not likely to be
phosphorus limited.
Photosynthesis, since It Is only a function
of primary producers, would seem more like-
ly to correlate well with phosphorus. How-
ever, due to seasonal changes In photosyn-
thetlc efficiency, variable amounts of blo-
mass, self shading, daylength and community
composition, this relationship Is likely to
be difficult to define.
Reaeratlon, the third major factor respon-
sible for stream dissolved oxygen fluc-
tuations, has no direct relationship with
phosphorus. The only possible Impact re-
sults from ponding of the strean due to
macrophyte growth. Ponding (Increased
depth and decreased velocity) could be ex-
pected to decrease the reaeratlon rate.
- 35 -
-------
If the quantities and rates In Figure 10
can be approximated from field studies and
modeling efforts, a good approximation of
stream dissolved oxygen content should be
possible.
METHODS OF ESTIMATING P, R, AND K2
Box Studies
Light and dark box studies were conducted
on the Bark River In late July and early
August. The average values of net photo-
synthesis (Pnet) and respiration (R) for
three days are presented In Table 18. The
light values (photosynthetlcally active
radiation IPAR, 400-700 nMl) for the In-
cubation periods were around 60-91 x I015
quant a/sec/cm2, well In excess of the
"typical" light saturation levels of
25-30 x 10l5 quanta/sec/cm2 reported by
West lake (1966). The Pnet values are far
below the 10 g02/kgDW/hr' reported by
West lake (1966). Self shading could
account for at least part of the differ-
ence. The plant densities under the boxes
were at or above Westlakes1 calculated
plant densities for optimum dally net pro-
duction. Other possible explanations In-
clude Internal storage of oxygen, decreased
productivity due to senescence, bubble for-
mation and possibly leakage of the boxes.
The dark box estimates of respiration were
In better agreement with Westlakes' esti-
mate of 1.5 g02/kgDW/hour. The values from
the box studies were slightly higher, which
Is not surprising since the boxes would
also Include sediment oxygen demand (SOD),
as well as perlphyton and Invertebrate res-
piration. For example, on 29 July, the DO
concentration In the box decreased 1.18
mg/l In the first hour. Volume of the box
was 80 liters, so total consumption of DO
was 80 x 1.18 = 94.4 mg 02 In one hour.
The macrophyte blomass enclosed In the box
was 51.94 gDW. Straight division
(94.4/51.94) gives 1.82 mg02/gDW/hr. If
the "true rate" was 1.5 mg02/gDW/hr, the
consumption due to macrophytes would be 1.5
x 51.94 - 77.91 mg02. This leaves 16.49
mg02 (94.4 - 77.91) unaccounted for. If
this remainder Is entirely attributed to
SOD, the SOD rate would be 16.49/0.28 m2
•> .059 g02/m2/hr, which Is within the
range of .0125 -.125 g02/m2/hr reported
by Edberg and Hofsten (1973). The results
also seem comparable to those of Owens and
Edwards (1962).
Table 18
Box Study Summary
NET
DATE
29 JULY, 82"
10 AUGUST, 82
12 AUGUST, 82
PHOTOSYNTHESIS RESPIRATION
3.00
2.68
3.25
2.31
1.82
1.51
1.76
2.42
Pnet and R values in g02/kgDW/hour.
"Dal ly totals" of P and R from the box
studies can be calculated If photosynthesis
Is presumed directly proportional to light
and respiration Is assumed constant. At
4 g02/kg DW/hr (« BO x I015 Quanta/sec/
cm2), an average day In July (3 x I02'
0/sec/on2) would give a gross production
value of 42 g02Ag DW/day. Respiration
would be 1.5 g/kgOW/hour, or 36 g02/kg
DW/day. Net production then, would be
about 6 g02/kg DW/day. On a very overcast
July day, total light could be as low as
I x I021 Q/sec/cm2, reducing gross
photosynthesis to 14 g02/kg DW/day, and
Pnet would be -22 g02/kg DW/day.
Each blomass sample harvested In the box
studies was also analyzed for nutrient con-
tent. Nutrient levels were low. and typi-
cal of the Bark River-Wolf Road site. One
box study was conducted at the Masonic Home
site, but It failed to show a substantially
higher net photosynthesis rate.
Bottle Studies
Bottle studies were conducted as another
method of approximating productivity. The
results are listed Table 19. Net and gross
productivities from the bottle studies are
- 36 -
-------
Table 19
Bottle Study Results. In gQ2Ag Dry Welght/hr
Dqte Site Pnet R
28 July BM 20.0
15.7
15 Aug B 10.3
1.0
7.3
1.5
15 Aug BL 11.2
2.7
9.6
2.3
20 Aug 0.47
0.62
0.86
1 Sept BM 1.3
1.7
3 Sept BL 2.3
8.1
2.4
Replicates
3
3
9
4
6
2
5
2
8
2
5
5
5
12
12
4
4
4
Species
Potamogeton spp.
Heteranthera dubla
Valllsnerla amerlcana
n n
Heteranthera dubla
n n
V. amerlcana
n n
H. dubla
n n
Myrlophyl lum spp.
H. dubla
Potanogeton spp.
V. amerlcana
H. dubla
H. dubla
"epiphytes" (perlphyton dislodged
from macrophytes)
H. dubla and epiphytes
substantially higher than- those of the box
studies. Respiration va-lues are approxi-
mately the same.
There are at least three possible expla-
nations for the discrepancy between the box
and bottle Pnet estimates. The first Is
that the bottles Included primarily leaf
and/or stem tissue, whereas the boxes con-
tained entire plants l_n situ. Some whole
plants were Included In the bottle studies,
but due to Individual bottle variability,
small sample* numbers and obviously artifi-
cial conditions, no conclusions are war-
ranted. The second possible explanation
Is Internal storage of oxygen (Metzel
1975). This, effect could be pronounced In
Intact plants, such as In the boxes. Leaf
and stem fragments In the bottles could be
expected to store less oxygen. The third
explanation Is that self shading could be
more Important In the boxes than the bot-
tles. Field observations pertaining to the
box studies do not support the self shading
hypothesis, but further, more carefully
constructed experiments would be needed
before self shading could be discounted.
On August 15, with one exception, plants
from the high nutrient (Lurvey) site showed
significantly higher (statistically) meta-
bolic rates than plants taken from the low
nutrient (Wolf Road) site. The Hateranthera
dubla Pnet values were significantly higher
(t-test, p < .05) for the high nutrient
plants. Respiration rates were also sig-
nificantly higher at the high nutrient site
(•b-test, p < .001. all values). Due to the
small numbers of each species Involved,
species specific respiration rates were not
tested.
- 37 -
-------
Epiphytic algae were present on essentially
all of the macrophytes, but were particu-
larly dense on those from the high nutrient
(Lurvey) site. Gentle rinsing removes most
of the (loose) epiphytes, but some (tight)
Inevitably remain. Cattaneo and Kalff
(1980) studied the relative productivity
of epiphytic algae ("loose" and "tight")
and macrophytes In lakes. They concluded
that the relative production depended on
season and nutrient levels. Epiphyte pro-
duction was found to exceed macrophyte pro-
duction during spring and fall In me so-
trophic portions of the lake, but exceeded
macrophyte production all year In the
eutrophlc portions of the lake. This sug-
gests that perlphyton can significantly
affect community photosynthesis and res-
piration terms, despite their relatively
small blomass, and that high nutrient con-
centrations may enlarge the contribution of
perlphyton to total community rates.
The dark bottle experiments of August 20
showed much lower respiration rates than
the other dark bottle studies. This Is
undoubtedly due to the low Initial DO con-
centration (4.85 mg/l), and the fact that
the ending concentrations averaged 0.62
mg/l. Owens and Marls (1964) used a series
of short Incubations to demonstrate that
the respiration rate varied with the dis-
solved oxygen concentration (Figure II).
McDonnell and Weeter (1971) also found a
decrease In respiration with decreasing DO
levels. Unlike Owens and Marls, however,
the relationship they found was linear
(R = «* + B (DO)), where R = Respiration
(mg02/gDW/hr), and «*-and B are constants.
Summary of Box end Bottle Study Results
Both the box end bottle studies Indicate
that macrophyte respiration should be about
1.5 g02/kg DW/hr. This Is In good agree-
ment with the literature estimates of
West lake (1966) and Owens and Marls
(1964). The box-derived estimates of
photosynthesis seemed loner, and the bottle
Figure II
The effect of oxygen concentration on plant respiration rates (from Owens and Marls 1964).
3.0
0. 2'°
O
o>
1 1.5
a.
E
o
u
s
o
I 1.0
O.8
C = Callitriche sp.
H = Hippuris sp.
R = Ranunculus sp
I I I I I I
2 34
Oxygen concentration (p.p.m.)
8 9
- 38 -
-------
estimates higher, than the estimate given
by West lake (1966) of 10 g02Ag DW/hr at
IIght saturation.
The box study estimates were not numerous
enough to statistically test for dif-
ferences In productivity between low-nutri-
ent and high-nutrient conditions. Bottle
studies Indicated higher net photo synthetic
rates for Heteranthera dubI a under high
nutrient conditions. Respiration rates,
overall, were higher for plants grown under
high nutrient concentrations.
DIEL CURVE ANALYSIS
This study also used two modeling approaches
to estimate community photosynthesis, com-
munity respiration and reaeratlon. These
are referred to as single-station and
double-station analyses. When the DO con-
centration at a single station Is monitored
over the course of a day, the results of
dlel changes In the magnitude of both meta-
bolic and physical factors Is observable.
Analysis of the dlel curve for a given
station can give estimates of photosyn-
thesis, respiration and reaeratlon which
represent upstream averages. The area for
which these averages apply Is not clearly
defined In the literature, and will be
dealt with In a later discussion. A single
station method Is currently used for stream
modeling of WONR waste load allocation sur-
veys.
The double-station method analyzes the
change In DO between two stations to derive
estimates for P, R, and K2. The values for
photosynthesis, respiration and reaeratlon
which result from double-station analysis
are applicable to the area between
stations, but may or may not be represen-
tative of the stream as a whole. Both the
single-station and double-station methods
were used to analyze the data In an attempt
to approximate the Impact of phosphorus on
stream dissolved oxygen levels.
Modeling Assumptions and Parameters
A few assumptions will be made to simplify
the modeling process. These are as follows:
- Photosynthesis Is directly proportional
to light Intensity. Some constant (P)
when multiplied by light Intensity (<*>
should equal the rate of gross photo-
synthesis.
- The respiration rate (R) and reaeratlon
coefficient (K2) are constant throughout
the day. I.e. that DO fluctuations are
not so wide that they significantly In-
fluence R and that temperature does not
change enough to significantly Influence
either R or K2 for a given reach.
- The area which Is responsible for DO
variation at a single-station sample
point Is homogeneous, and all areas con-
tribute equally, or the area between the
two double-station sample points Is
homogeneous, and causes a linear re-
sponse In the DO concentration of a mass
of water as It moves through the reach.
For modeling purposes, the major terms re-
sponsible for stream dissolved oxygen fluc-
tuations (ADO/At) can be characterized as:
•GROSS CCmJNITY PHOTOSYNTHESIS (<*P>.
due to both macrophytes and perlphyton.
* COMMUNITY RESPIRATION (R), due to all
forms of bacteria, fungi, algae, macro-
phytes. Invertebrates, etc., as well as
chemical oxygen demand (COD) (anything
which removes oxygen from the water).
* REAERATION, usual ly represented as
K2(Cs-Co). In which K2 Is a physical
constant which Indicates a proportion of
the deficit (Cs-Co) which Is satisfied
per unit time. Cs represents the satu-
ration concentration of dissolved oxy-
gen, which Is calculated from tempera-
ture data assuming normal atmospheric
pressure (760 mm Hg). Co Is the mea-
sured concentration of dissolved oxygen.
The sum of these three rates (P, R and
K2(Cs-Co)l should account for the rate at
which the stream concentration of dissolved
oxygen Is changing, £DO/&t. These terms
can be combined to approximate a differen-
tial equation:
- 39 -
-------
ADO/At
+ R + K2 (Cs-Co)
It should be noted that the resulting
A DO/A t remains constant only as long as all
terms l°<, P. R, and K2(Cs-Co)l remain con-
stant. If the A DO/A t term Is expressed In
terms of the deficit (A(Cs-Co)/At) and the
differential equation Is Integrated, the
result Is:
(C8-Co)t +Ac - (Cs-C0)t
ll<*P
This lest equation was derived under the
assumption that the sum («P+R) remains
constant, and this must be considered when
applying the result. The differential form
of the equation was used In the double-
station analyses, and the Integrated form
was used In the single-station analyses.
If all of the foregoing assumptions and
restrictions are satisfied, the differen-
tial and Integrated equations should return
the same coefficients.
A few simple observations can help clarify
these equations, and hopefully represent a
simple set of guidelines which define the
behavior of dlel curves.
First, the ADO/At term will be positive
only when the sum of <*P, R and K2(Cs-Co)
Is positive. In a strictly mathematical
sense, all of these terms are Independent,
and any one of the terms could be the major
factor determining the magnitude and sign
of the A DO/At term. More realistically,
however, we expect P, K2, Cs and Co to be
positive, and R to be negative. Over the
course of a day, the ex P term should In-
crease with rising light (<=*) levels, and
when the Increase Is sufficient, the
A DO/At term will become positive, which
means that the DO concentration will rise.
Towards evening, when the product ot. P Is
becoming smaller, the ADO/ At term falls
through zero (the highest DO concentration
for the day Is reached at this point), and
then becomes negative which means that the
DO concentration Is falling. As the DO
concentration falls below saturation, the
deficit (Cs-Co) begins to Increase, which
Increases the product K2(Cs-Co). Since R
Is assumed constant and negative, and
K2(Cs-Co) Is positive and becoming
larger, the two will balance each other and
the system will be at "equilibrium" (I.e.
ADO/ At will become zero). If a dlel
curve attains Its minimum DO concentration
(maximum deficit) prior to dawn (ADO/At =
0, and ot =0, so«»
-------
Tab-le 20
Time Required to Satisfy
90 Percent of the Deficit
Time Required (hours)
I
2
5
10
15
55.2
27.6
II.0
5.5
3.7
At this point, two observations deserve
emphasis:
I. The speed at which the maximum deficit
Is approached Is determined solely by K2.
2. The absolute magnitude of the maximum
deficit depends on the R/K2 ratio.
Single and Double Station Analyses;
Differences In Methodology and Purpose
The major difference between the single and
double-station methods Is best Illustrated
by their different Interpretations of the
A DO/At term of the differential
equation. The single-station method uses
the slope of a single die I curve as an
estimator of A DO/ At. The double-station
method uses the change In the 00 concen-
tration of a mass of water as It flows from
one station to another, divided by the time
of travel (TOT) between stations (Figure
12). In case "A" (Fig. 12) the double-
station value for A DO/ At would be -0.5
mg/l/hr. The single-station values would
be 0.64 and 0.57 mg/l/hr for the upstream
and downstream stations, respectively. In
case "B", the single and double-station
values are essentially equal.
Figure 12
Comparison of single and double-station die I dissolved oxygen methods and purpose
10 •
o
a
O
5--
o o
Key
• upstream station
o downstream nation
Time of Travel 1 hour. 17 minutes
O .
O
1046 1446 1846 2246 2646 3046
Time (hours)
Bark River, August 30, 1982
3446
3846
- 41 -
-------
The goal of double-station analysis Is to
produce a set of coefficients which will
accurately predict the DO concentration at
a downstream station, If the upstream DO
concentration Is given, and the temperature
and light conditions which prevail for the
duration of the time of travel between
stations are known. The values of P, R and
K2 from double-station calculations are a
function of processes within the area be-
tween the stations, and may or may not be
characteristic of the strew as a whole.
If the double-station coefficients are used
to generate a si ng le curve (I n the sane
manner that single-station coefficients are
used to generate a single curve), the re-
sultant curve may differ from both the up-
stream and downstream curves from which the
double-station estimates were derived. The
double-station coefficients will accurately
predIct the downstream curve on I y If the
upstream curve Is given. If the upstream
and downstream curves are Identical, the
curve and coefficients produced by
double-station analysis will be Identical
to the results of the single-station
analysl s.
If a very good fit Is obtained In
single-station analysis, the coefficients
(P, R, and K2> wl 11 reproduce the curve
which was obtained for that station. The
end result of single-station analysis Is a
set of coefficients which allows prediction
of a DO concentration at any time for a
particular station.
Stream Areas Represented by Die I Analyses
Within a homogeneous reach, the area Im-
mediately upstream of the sample point will
have a greater Impact on the dissolved oxy-
gen fluctuations at the sample point than
an equivalent area farther upstream. The
relative Importance of each area, according
to the Integrated equation. Is determined
by the magnitude of the K2 value for each
upstream area. The larger the K2 value,
the smaller the area represented by the
single-station method, and the larger the
difference between the effect of two up-
stream areas.
Dl agrammatlcal ly..<
I-
A
-Area 2-
FLOW
•I-
B
-Area I-
-I
C
C = single-station sample point
A, B, C = double-station sampling points
For the purposes of this discussion;
I. Area (will have a greater Impact than
Area 2 at point C.
II. The magnitude of this difference Is
dependent on the K2 value of the entire
reach (A - C). A larger K2 value wl11
result In a greater difference between
the relative contributions of Area I
and Area 2 at point C.
The above discussion assumes that Areas I
and 2 are similar. If there Is a great
discrepancy between conditions (e.g. bio-
mass, K2, etc.), and the time of travel
between the points Is short (again relative
to the K2 value), the single-station
calculations could be more representative
of Area 2 than of Area I. The exact con-
tributions of each area are dependent on
the product (K2xTOT). If no
differences exist between areas (I.e. the
diet curves at each point A. B, and C are
Identical) the calculated rates could be
correctly applied to both areas and the
single-station and double-station analyses
would be expected to produce the same co-
efficients (P, R, and K2).
Since the double-station method calculates
P, R and K2 values for the area between
sample points, the correlation of measured
plant blomass (from harvesting and mapping
procedures) Is potentially straight-forward.
The difference between the single and
double-station techniques becomes Important
when we try to model what will happen to
the die I curve at point "C" If a sewage
treatment plant discharges at point "B".
The only way to predict what will happen to
the die I curve at point "C" Is by
- 42 -
-------
quantifying the Impact of each area. In
order to apply coefficients obtained from
single-station analysis then. It Is
necessary to know what area the
coefficients were derived from. This
concept also has a bearing on the choice of
sampling locations (distance or time
between samp-le points) and the question of
"how far downstream11 the Impact of the
discharge wl11 reach.
Limit of Reach Length for the Double Station
Method
The double-station differential method, as
presented above, makes certain assumptions
which limit the length of the reach (or the
amount of travel time between sample
points) to which the method can be ap-
plied. The primary assumption Is that none
of the measured parameters (ADO/At, <*. or
(Cs-Co)) changes significantly during the
time Interval over which ADO/At Is
measured, or ttiat the variation Is such
that approximation by an average value Is
Justifiable. For example, If photosynthesis
Is directly proportional to light Inten-
sity, and the light Intensity varies from
10 to 20 to 30 over a two hour time In-
terval (the "20" value occurring after
exactly one hour has elapsed) the total
amount of oxygen produced should be
correctly predicted by using a single
average value of "20" over a two hour
period.
If the Integrated equation Is used, a dif-
ferent set of restrictions on travel time
Is appropriate. The Integrated equation
removes the requirement that A DO/At re-
mains constant, but the requirements re-
garding constancy of at., P, R and K2 re-
main. If the Integrated equation Is used
In the double-station technique, the time
of travel and stream character within the
reach must remain short enough to assure
constancy of these terms.
Maximum Attainable Dissolved Oxygen Deficit
According to both the Integrated and dif-
ferential equations, the maximum attainable
deficit Is R/K2. If the length of the night
Is multiplied by an approximated K2 value,
we can use the Integrated equation to cal-
culate what proportion of the difference
between the observed deficit at sunset
(Cs-Co) and the calculated maximum attain-
able deficit at sunrise (R/K2) will be
satisfied. The actual concentration of DO,
of course, depends on the stream
temperature and the magnitude -of the
deficit at sunset. If night-length Is 12
hours, and K2 Is greater than 5, more than
90$ of the maximum attainable deficit will
be satisfied.
The concept of a "ha If- life" for a deficit
Is pertinent at this point. For those
familiar with the fundamental decay equa-
tion, K2 Is a decay constant, and Is the
only factor that controls the rate at which
the maximum deficit Is approached. After
the elapse of one half-life (K2 x t =
.693), 50? of the maximum deficit has been
achl«ved, after two half-lives (K2 x t =
1.386), 75* of the maximum deficit has been
achieved, after three half-lives, 87.5* of
the maximum deficit has been achieved, etc.
Examples which Illustrate the Importance of
K2 In controlling the rate at which the
maximum deficit Is approached follow.
Case I :
If, for example, the DO deficit at sunset
(Cs-Co)ss, Is 2 mg/l, R/K2= 4, the length
of night Is 12 hours and K2 Is 5/day, then
the deficit after 12 hours (I.e. the
deficit at sunrise (Cs-Co)sr) can be
calculated as follows:
(Cs-Co )sr = (Cs-Co )|
R/K2
= 2 (.08) + 4 (.92)
= . 16 + 3.68
- 3.84 mg/l or 96$ of R/K2
Case 2:
If the Initial DO was 2 mg/l higher
(deficit at sunset = 0), (Cs-Co)sr would
be O(.OB) + 4(.92) = 3.67 (92* of R/K2).
- 43 -
-------
Case 3:
If the R/K2 ratio was 5, (all other para-
meters as In the Initial case) (Cs-Co)sr
= 2(.08) + 5(.92> - .16 + 4.6 - 4.76 mg/l
(95* of RA2>.
If the temperature at sunrise was 20°C, Cs
would be 9.07 mg/l. and of the above
examples, only the third would result In a
violation of the 5 mg/l criteria. It Is
clear from these examples that "R/K2" Is
very Important In determining the maximum
deficit, and that according to the model,
where K2 Is large enough (I.e. K2 > 5),
R/K2 essentially specifies the maximum
deficit, regardless of the concentration at
sunset.
Since a drop In temperature will Increase
the saturation oxygen level, Cs, normal
nighttime cooling will Increase the rate at
which the maximum deficit R/K2 Is
approached. Once a deficit of R/K2 Is
reached, any further drop In stream tem-
perature will cause the dissolved oxygen
level, Co, to rise as the stream maintains
the equilibrium deficit of R/K2. This Is
Illustrated In Figure 13 which shows the
die I curves for 18 August 1981 In the
Kohlsvllle River. Due to a very high re-
aeration rate (15/day), the equilibrium
level Is reached almost Immediately after
sunset. The DO concentration rises
throughout the night because reaeretlon and
respiration maintain their "equilibrium"
(I.e. (Cs-Co)= R/K2 ), and temperature de-
creases cause the saturation concentration
to rise. (A decrease of 3°C will Increase
Cs approximately 0.5 mg/l. If this change
Is "added" to the above examples (case I -
case 3), all of them would have reached
Figure 13
Attalrcnent of equilibrium DO deficit (R/K2) and the effects of
changing DO saturation In the Kohlsvllle River (August 18 1981).
1 1
en
—•
2
>
O
O
w
(/>
Q
10.8 -
10.6 -
10.4 -
10.2 -
10 -
9.8 -
9.6 -
9.4 -
9.2 -
9 -
8.8 -
8.6 -
8.4 -
8.2 -
8 -
7.8 -
a
D.O. UP
—r—
12
16
—I—
20
~T—
24
28
32
TIME (hours)
D.O. DN.
SATURATION
- 44 -
-------
"equilibrium". There Is sane question as
to when (I.e. under-what conditions)'this
"adding" trearhnent Is valid, however.
MACROPHYTES AND REAERATION
A potential for change In K2 because of
Increased macrophyte growth must also be
dealt with. Increases In macrophyte den-
sity will cause an increase In the drag
felt by the water as It flows over the
stream bed. Such an Increase In drag will
cause greater depths (I.e. ponding) of
water for equal flow when weeds are pre-
sent. Mathematically this can be expressed
in the Manning flow formula as an Increase
In the roughness coefficient, "n".
The Manning formula Is...
0 = 1.486/n x A x (R2/3) x (Sl/2)
By dividing both sides by cross-sectional
area (A), we can obtain...
V = 1.486/n x (R2/3) x (Sl/2)
Where... Q = discharge. In cubic feet per
second (cfs)
A = cross sectional area. In
square feet
R = hydraulic radius (A/wetted
perimeter)
S = slope of the water surface
n = Mannings "roughness co-
efficient"
V = mean velocity (time-of-travel)
In the first equation, we can see that If Q
Is held constant, and "h" Is Increased, the
product [A x (R2/3) x (Sl/2)l must also
Increase. At least part of this Increase
could be expected'to-trans I ate Into an In-
crease In depth.
Similarly, In the second equation, an In-
crease In "h" would result In a pro-
portional decrease In mean velocity If "R"
and "S" remain relatively constant.
Table 21 gives values of Mannings' "n" for
various substrate types (from Corbett
1945). It Is Important to note that the
highest values are associated with "very
weedy reaches".
Table 22 shows values of stream discharge
(Q), mean velocity (V), mean depth (0), and
macrophyte blcmass for the Lurvey site
(Impacted) and the Ashlppun site (non-Im-
pacted) for the months of June, July and
August.
The mean velocity and mean depth change
substantially at the Lurvey site (Im-
pacted), but not the Ashlppun site. If
changes In mean velocity and depth due to
macrophyte blcmass: "n" relationships are
to be quantified, factors such as condition
(shape?) of the stream bank, character of
the stream bed, and slope of the channel
would also have to be taken Into account.
In the present example, however, we can
approximate the effect of ponding due to
macrophytes by examining a few equations
which were developed to predict K2 from
mean depth (D) and velocity (V). The fol-
lowing three equations ranked highest among
those equations which used mean depth and
velocity to predict K2 (Grant and
Skavroneck 1980). All three equations
predict K2/day at 25°C.
Padden-Gloyna (1971)*
K2 = 7.73 (V-703) x (D-l-054)
Bansal (1973)*
K2 • 5.26 (V-6) x (D'1-4)
Negulescu-RoJanskl (1969)*
K2 - 12.29 (V/D)'85
*from Grant & Skavroneck, 1980
Table 23 shows the K2 values calculated
from the above equations for each month and
each stream. The decrease In K2 from
July to August at the Lurvey site Is ap-
parently due to macrophyte growth.
Values of Mannings' "n" taken from tables
are not exact, and the Impact of weed
growth upon mean depth and velocity would
be even less exact. The various equations
- 45 -
-------
Table 21
Approximate Values of Manning's Roughness Coefficient, "n" (from Corbett 1945).
Channel Conditions
Channel Description
1.
2.
3.
4.
5.
6.
7.
8.
Clean, straight bank, full stage, no rifts
or deep pools
Same as (1). but with sane weeds and stones
Winding, some pools and shoals, clean
Same as (3), lower stages, more Ineffective slopes
and sections
Same as (3), some weeds and stones
Same as (4), stony sections
Sluggish river reaches, rather weedy or with very
deep pools
Very weedy reaches
Perfect
0.025
0.030
0.035
0.040
0.033
0.045
0.050
0.075
Good
0.030
0.033
0.040
0.045
0.035
0.050
0.060
0.100
Fair
0.035
0.035
0.045
0.050
0.040
0.055
0.070
0.125
Poor
0.040
0.040
0.050
0.055
0.045
0.060
0.080
0.150
Table 22
Discharge (Q In cfs). Velocity (V In ft/sec). Depth (D In ft) and Macrophyte
Bl amass (In gm/mz DWT) for POTW- Impacted (Lurvey) and No n- Impacted (Ashlppun) Sites
Lurvey
June,
July,
August
1982
1982
, 1982
0
36
28
29
V
1.2
.71
.38
D Bl amass
1.2
1.3
2.03
75
151
289
Q
16.5
11.7
11.2
Ashlppun
V
.86
.72
.72
D Blomass
I.I
1.2
1.3
30
109
125
Table 23
Values (/day) for June-August at POTW-Impacted (Lurvey) and Non-Impacted (Ashlooun) Sites
Q
(cfs)
Blomass
(gm/m2)
Padden
Gloyna
Bonsai
NeguIescu-
Rojanskl
LURVEY
June, 1982
July, 1982
August, 1982
ASHIPPUN
36
28
29
75
151
289
7.25
4.61
1.86
4.54
2.97
1.09
12.29
7.34
2.95
June. 1982
July, 1982
August, 1982
16
12
II
30
109
125
6.28
5.06
4.65
4.20
3.35
2.99
9.97
7.96
7.43
- 46 -
-------
which predict K2 from mean depth and
velocity do not show good agreement In many
Instances. All of this seems to Indicate
that an attempt to model this process at
this time would be frustrating. Although a
study specifically designed to relate K2
and blomass at a given location could yield
quantitative results, the results would be
applicable only to the study site, and
prove no more usef u I than the ex amp les pre-
sented above. From the present study, the
only conclusion to be drawn Is that for a
given stream, a large Increase In blomass
could lead to a substantial decrease In K2.
The change In depth could also be expected
to lead to a decrease In the rates of res-
piration and photosynthesis (mg/l/hr). If
the depth doubles, the volume of water over
a square meter would also be expected to
double. This would lead to a sharp decrease
In the volumetric (mg02/l/hr) metabolic
rates ascrlbable to the blomass on the
square meter.
The decrease In respiration and the dec-
rease In reaeratlon In this scenario could
be off-setting. The actual Increase In
R/K2 predicted by assuming that all R Is
allocatable on an area! basis (no signifi-
cant BOD or plankton populations), and the
change In reaeratlon Is entirely due to the
change In depth, varies for the equation
used to predict K2. For the Padden-Gloyna
equation, the "new" R/K2 would be four
percent higher (new R = .5 times old R, and
new K2 = old K2 times 2'1-054). Similar
calculations with the Bansal equation would
Indicate a 32* Increase. Decreases In mean
velocity would be certain to accompany In-
creases In depth, so these estimates must
be considered conservative.
DEVIATIONS FROM MODELING ASSUMPTIONS
There are cases where K2 doesn't seem large
enough to explain the apparent attainment
of equilibrium. Among examples are cases
where the K2 approximated from the die I
data Is significantly larger than the K2
predicted by various equations. One possi-
ble explanation for this situation lies In
the failure of the assumption that
respiration remains constant over the
course of the day. In all likelihood, res-
piration will at some point become limited
by the availability of oxygen, or even (po-
tentially) be proportional to oxygen
availability throughout the day (see
Box/Bottle study section).
The question of how much the respiration
rate can vary Is at least partially an-
swered by the dark bottle experiments. In
the range of 5 to 9 mg/l DO, a value of 1.5
mg 02/g DW/hr seems valid. Below 5 mg/l,
the rate appears lower. Since we are con-
cerned with keeping the DO levels above 5
mg/l In the stream. It seems prudent to
choose the value applicable to DO levels
above 5 mg/l.
The variation of respiration with tempera-
ture has been explored In some detail
(Lasslter 1975, Canale et al. 1982). The
general relationship has been expressed
as...
T2-T,
R2 » R| x 0
Where... R2 = respiration at T2°C
R| = respiration at TI°C
0=a constant which Is
specific for a specified
process (In this case
respiration) over a given
temperature range.
If 0 Is assumed to be 1.07, a three degree
drop In temperature would lead to a nine-
teen percent decrease In the R value. This
would lead to a nineteen percent decrease
In the R/K2 ratio.
For those who prefer to use OJQ. « °io
of "2.0" corresponds to 9 = 1.07) where
R2 = R, x
-------
"community R", the term which Is actually
approximated In die I surveys, Is likely to
vary less than a species specific 0 would
predict (Odum 1973, McDonnell 1982).
If the R-value from the die I curve analysis
Is an "average" for a period where a sub-
stantial amount of time Is spent In the
5-10 mg/l DO range, the use of 1.5 g 02/kg
DW/hr rate seems reasonable. If the average
night-time DO Is substantially below 5
mg/l, the application of a lower rate would
be defensible. The ultimate solution of
course, would be to alter the dlel equa-
tions so that they would produce tempera-
ture and DO sensitive coefficients. Con-
struction and validation of such a model,
however, would be a major project.
The reaeratlon coefficient K2 Is presumed
constant as we 11, but has been shown to
vary with temperature according to the
equation:
K2(25.C) =K2(T) I.024<25-T>
According to the equation, a temperature
decrease of 3°C (e.g. 23°C to 20°C) would
lead to a seven percent decrease In the K2
value, as follows:
K2(25'> = 5
K2(23") = (5) l.024('2) = 4.77
K2(22") = <5) l.024('3) = 4.44
% change = .33/4.77 = .07 x I00| = 1%
This would cause a slight (+7$) Increase In
the R/K2 ratio. In the night period, the
decrease In R/K2 caused by decreased respi-
ration and the Increase In R/K2 caused by
decreased K2 are thus opposed and poten-
tial I y compensatory. The decrease I n R/K2
due to decreased respiration Is likely to
be more Important. (See Box/Bottle Study
section.)
FURTHER CONSIDERATIONS IMPORTANT TO THE
MODELING PROCESS
The question of whether or not respiration
and/or K2 remain constant with decreasing
temperature and DO levels does not serious-
ly alter the discussion regarding R/K2 and
the maximum deficit for at least two
reasons. First, a change In the blomass
should provide an Incremental change In
respiration, and second, the respiration
rate which results from the dlel curve
analysis Is likely to be an "average" for
the site, I.e. It will over-estimate the R
value applicable to the period during which
DO Is lowest, and because of this be more
typical of higher DO levels. Thus, al-
though the magnitude of projected changes
Is not likely to be predicted with absolute
accuracy, there Is a good degree of cer-
tainty that changes within a certain range
will occur.
The area (or time of travel) downstream of
a point source to which the modeling pro-
cess should be applied Is another con-
sideration. Attempts to document a "phos-
phorus decay curve" In this study met with
limited success. A decrease In phosphorus
concentration over the study area was ob-
vious on some days, and lacking on others.
When attempting to specify a region of Im-
pact, an Important distinction must be made
between the true decay observed for
non-conservative pollutants, and the de-
crease In concentration ("decay") observed
for nutrients such as phosphorus. O300 may
be considered a non-conservative pollu-
tant. It Is oxidized to water and carbon
dioxide, and lost from the system. Phos-
phorus may show an Initial "decay" due to
uptake by plants and physical adsorption,
but It Is not eliminated from the system.
It may be released at the end of the
growing season, re-dissolved through
grazing, re suspended during storm flows,
etc. Whl le the present study does not
rigorously define the area Impacted, It
does suggest that the area Is well In
excess of several stream ml les.
The actual area Included In a model of dlel
oxygen curves should Include the area which
Is Impacted, which must be decided on a
case-by-case basis. If phosphorus levels at
the lower end of the modeled area
approached those encountered upstream of
the source, the model could be considered
to have accounted for most of the Impact.
- 48 -
-------
This will not always be practical, but
should be a desired goal. If no decay Is
obvious, or other sources or factors
(tributaries. Impoundments, etc.) become
Important enough to masK the Impact, the
modeled area will have to be defined and
Judged on the basis of preliminary studies.
The Impact of BOO and the Impact of phos-
phorus are separate entitles. BOO may be
responsible for depleted oxygen levels at
one point, and phosphorus at another. A
good example of this occurs on the White
River below the Lake Geneva POTW. The BOO
sag occurs within the modeled area, but
oxygen depletion due to mecrophytes Is
obvious at the last station (Figure 14).
RESULTS AND IMPLICATIONS OF DIEL ANALYSES
The results of the double-station and
single-station analyses are supportive of
the dissolved oxygen model and the
assumptions under which It was developed.
Plots of R/K2 versus the maximum deficit
(Figures 15 and 16) show good agreement
with theoretical considerations. It should
be pointed out that the values returned
from the diet curve analyses are "first
Figure 14
Illustration of BOD oxygen "sag" and "sag" attributable to
p I ant growth In the white River
,J
-
WHITE RIVER
SEPTEMBER 3-4,1980
X
en
£
Z
UJ
o
X
0
Q
UJ
0
1/1
i/i
Q
12 -
1 1 -
10 -
9 -
8 -
7 -
// V
/ >
/
*
^£b*
f^^'^^ ^^~~~~~i
D^_ — — ~"^~--^
A ~^Si
>»
6~
5 -
4 -
3 -
2 -
10
n = STATION 1
1 1 1
14 18 22 26 30
TIME OF
+ = 2 o
DAY (hrs.mins)
= 3 A » 4 x=5
1
34
v = 6 •= 7
- 49 -
-------
o
I
(1 1 slope)
345
Maximum Deficit (mg/l)
(1 1 slope)
23456
Maximum Deficit (mg/l)
Figure 15
Theoretical maximum deficit (R/K2) and observed
maximum deficit for the double-station method.
Figure 16
Theoretical maximum deficit (R/K2) and observed
maximum deficit for the single-station method.
-------
approximations". In a true modeling set-
ting, the coefficients are subject to
adjustment In validation and verlflcatl«n
procedures. Such adjustments would be
ejected to Improve the relationship be-
tween R/K2 and the maximum deficit.
If the estimates of respiration from die I
curve analysis and the respiration expected
on the basis of 1.5 g 02/kg DW/hr are
compored (Figures 17 and 18), It Is obvious
that the dlel estimates are much higher.
This suggests that other factors (BOO. SOD.
Invertebrates, perlphyton, etc.) are Im-
portant when calculating the total
community respiration estimate.
Graphs of gross photosynthesis versus bio-
mass are presented In Figures 19 and 20.
Double-station estimates are generally much
higher than single-station estimates. The
1982 double-station estimates, which were
derived from larger reaches with longer
travel times, are generally lower than the
1981 estimates when corrected for volume
(Figure 20). The statistical fits of the
1982 data sets were also Improved. The
1982 double-station data and both year's
single-station data are below the estimates
of photosynthesis derived from Westlake's
(1966) blomass approximations. The
photosynthesl s approximations from the box
studies are In better agreement with the
dlel estimates, but the scatter of the dlel
estimates still precludes serious modeling
effort.
Table 24
Dlel Modeling Coefficients for
Single and Double-Station Analyses
SINGLE STATION (INTEGRAL) COEFFICIENTS
Table 24 (con't)
DOUBLE .STATION INTEGRAL COEFFICIENTS
Data
Set
1
7
9
13
16
18
K2
(I/day)
1.0+0.5
0.9+0.8
0.8+0.7
1.4+1.0
2.2+1.5
3.9+2'.5
P
(mg/l/day)
10.9
9.2
7.0
10.7
19.4
1.3
R
(mg/l/day)
13.3
13.3
11.3
19.5
29.7
10.3
Data
Set
1
2
3
4
5
6
7
8
9
10
II
12
13
14
15
16
17
18
DOUBLE
Data
Set
,
2
3
4
5
6
7
8
9
10
II
12
13
14
15
16
17
18
K2
(I/day)
2.6*0.4
2.8+0.3
2.7~
5.7
9.3+3.8
••""
3.2
2.6+0.5
1.7+0.8
3.7+0.7
6.3+1.3
7.6+1.1
4.3+J.O
0.6+1.2
6.3+1.0
5.9+2.9
7.2+2.7
7.3*"
STATION
K2
(I/day)
2.7+0.5
2.6+0.3
2.7+0.2
4.8+0.7
6.0+1.1
6.5+2.4
3.2+0.4
2. 1+0.4
1.5+0.7
3.4+0.7
5.47|.0
6.2+0.7
4.1+1.0
4.8+_l.2
5.7+0.7
4.4+"|.7
5.7+J.O
6.6+1.4
P
( ing/ 1 /day)
9.9
6.7
8.2
12.0
19.4
**
6.9
8.5
7.5
10.3
13.3
12.2
10.7
8.8
12.4
15.7
13.9
10.7
DIFFERENTIAL
P
(mo/ I/day)
10.6
9.4
10.2
11.2
13.0
12.7
8.3
8.7
9.6
10.7
13.9
12.6
11.4
15.0
13.9
16.4
14.4
11.3
R
(mq/ I/day)
14.9
17.8
16.5
24.4
35.9
**
20.6
17.0
10.8
17.1
24.7
26.7
19.5
10. 1
23.9
22.1
24.1
20.8
COEFFICIENTS
R
(mg/ I/day)
15.0
17.9
16.5
21.3
24.1
22.7
21.0
14.5
10.6
16.8
22.7
23.2
19.4
21.9
22.5
21.1
22.0
21.3
- 51 -
-------
01
N
35
•> 30-
1
f »•
|
| »•
>
2
s
^ 15-
1
a
5-
0-
C
15.0-
0 - ll.S-
m
O
i
? 10.0-
I
2
oo § ,.s.
o o o S
o £
1
c 5.0-
o 1
000 g
f
0 * Z.S-
o
50 100 150 100 150 30C Q
Macrophyle Bmnm (gm/m2)
Figure 17
Double-station respiration estimates
o
0
o
o
0 0
o
o o
o
o
o
0 O
0
0 0
o o o
0 0
0 0
o
50 100 ISO 100 250 300
Microphyte Biomra (gm/m2)
Figure 18
Single-station respiration estimates
from die I and plant blonass data.
from dlel and plant blomass data.
-------
III
u
so
45-
I"
C*
i»
1
c 30~
1*
c
8 „
•
• I
o g
I -
O j
|
_
i
0 0 0 *
t »•
» a
0 0
0
o
o
e
0
o
o
0
0
0 0
0 O
o
0 0
0 0
O 0 0
o
o o
o
o o o
E«.mal«l Biomra R«p,ral,on (9mO2/n.2/d.V) En,m««i B,om,,, R«p,ra,,on (9mO2/n,2/d.y)
Figure 19 Figure 20
Double-station area! photosynthesis estimates Single-station area! photosynthesis estimates
from die I and plant blomass data.
from dlel and plant blomass data.
-------
Comparison of Double and Single Station Data
A demonstration waste load allocation survey
was conducted on the Bark River on 30
August-2 September,1982. The Intent was to
gather data which would allow comparison
of the single and double-station tech-
niques, and hopefully corroborate some of
the assumptions of the model.
Seven diet stations were established on the
Bark River, downstream of the Dela-Hart
POTW outfall. The locations of these
stations. In miles downstream of the out-
fall, were 0.0, 0.3, 0.6, 0.85, 1.15. 1.42
and 1.75. Dissolved oxygen (DO) was
monitored at all stations for about 48
hours. Light was monitored at the 0.85
mile station. Temperature was monitored at
the 0.3, 1.15, 1.42 and 1.75 mile
stations. The last three stations (1.15,
1.42 and 1.75) were sampled for dissolved
oxygen and temperature at two to three hour
Intervals. All other data were con-
tinuously recorded. An Interpolation
routine (AISPIN/AISPEV, available at the
Madison Academic Computing Center) was used
to generate "continuous" data for the last
three stations.
Three separate analyses of the data set
were performed (Table 24):
I. The single-station analysis was per-
formed on each of the seven die I
curves, using the Integrated equation.
2. Double-station analysis was performed
on all possible pairs of dlel curves
(0.0-0.3, 0.0-0.6, 0.0-0.85, etc.),
using the differential equation.
3. Double-station analysis was repeated,
using the Integrated equation.
The resultant coefficients, and the "fit"
of the Individual determinations, can be
used to test some of the assumptions of
dlel curve analysis, and answer some
questions which pertain to the choice of a
modeling method. The following section
shall examine the following questions:
• What upstream area does the single-
station analysis represent?
• Do the differential and Integrated
equations return the same coefficients
when "short" reaches are analyzed?
* How long must a reach be before the
differential equation falls, and whan
Is It necessary to employ the
Integrated equation?
* How does Increasing reach length Impact
the accuracy of the double-station
technique?
Area Represented By Single Station
Coefficients
The relative positions of the DO stations
for a wasteload allocation survey on the
Bark R. (In miles downstream of the out-
fall), the coefficients returned by the
single and double-station analyses, and the
time of travel between stations are shown
In Table 25. The P, R and K2 values are
arranged to correspond with the area or
position from which they were derived. The
double-station values listed between the
0.0 and 0.3 mile stations were derived for
the reach between these two points.
Single-station values are adjacent to the
station from which they were derived, and
represent averages applicable to some area
upstream of the station for which they were
derived.
As might be expected, the double-station
and single-station values do not seem to
closely agree. The determination of P, R
and K2 for longer reaches (e.g. the 0.0 to
0.85 or 0.0 to 1.15 reaches) failed to
yield sufficient data with which to deter-
mine and validate the area represented by
the single-station analyses.
If we assume an "average K2" for the 0.0 to
1.85 mile reach of 2/day, and use the ob-
served travel time of 7.2 hours (0.30
days), we can see that the Incoming deficit
at the 0.0 mile station still has a sub-
stantial Impact on the deficit that arrives
(7.2 hours later) at the 1.85 mile station.
- 54 -
-------
Table 25
Single and Double-Station Coefficients
for the
Al
Single-Station
K2
1.8
1.6
1.8
2.7
2.'7
4.5
4.8
P R
11.3 14.5
10.0 14.8
7.2 12.9
8.6 12.7
7.7 11.4
9'.2 15.4
11. 1 17'. 4
Bark. River Waste load
location
Stream
Miles
(TOT In
hours)
0.0
(1.65)
0.3
(1.27)
0.6
(1,10)
0.85
(1.16)
1.15
( 1 .07)
1.42
(0.97)
1.85
Survey
Double-station
K2 P
2.7 10.6
3.2 8.3
1.5 9.6
4.1 11.4
4.4 16.4
6.6 11.3
R
15.0
21.0
10.6
19.4
21.1
21.3
The calculations are as follows:
(Cs - Co), 85 • (Cs - Co)0 0 i
-(C. - Co)0 „ e'° 6 » ((-P * R)/2)(l - e'° 6)
• (Cs - Cs)0 „ (.55)
P and R values are In mg02/liter/day. K2
values In I/day.
If we assume a larger K2 value (4/day), the
Importance of the Incoming deficit
diminishes, but Is still significant.
(Cs - Co), as • (Cs - Co), „ (.30) * (Iff » *)/4)( 70)
From this perspective It would appear that
the area represented by "the 1.85 mile
single-station analysis Is In excess of
1.85 miles. These calculations are simi-
lar to assuming that an "average P, R and
K2" can be uniformly applied to the area
upstream of the 0.0 mile station and would
produce an "average (Cs-Co) ", and that
another set of "average P, R and K2" values
uniformly applies to the 0.0-1.85 reach,
and should produce an "average (Cs-Co)".
This Is a "steady-state" viewpoint. Under
these conditions, little or no variation In
DO occurs over time at a given station.
Under these conditions, the time of travel
(or distance) which must elapse before the
"new" P, R and K2 values are fully ex-
pressed Is primarily a function of the new
K2 value. In reality, the Incoming deficit
Is not constant, and In this Instance fluc-
tuates around "0". It Is clear that the
magnitude of both the Incoming deficit and
the ((«*P + R)/K2) term wl 11 be Important.
When the Incoming deficit Is large. Its
Impact may mask the effects of the area
Immediately upstream. If the Incoming
deficit Is small «!?). however. It Is
likely to have a small (relative to the
Impact of the ((«*? + R)/K2) term) Impact
on the deficit at a downstream station.
The term (txP + R1/K2) Is also presumed
constant In this discussion. Like (Cs-Co),
however, the value of this term could be
expected to fluctuate around zero, and
exert a variable Impact upon actual deficit.
In a real stream, natural variations In the
magnitude of P, R or K2 can be expected to
play equally Important roles, and compli-
cate matters even more.
Comparison of Mean Single and Double
Station Values for the Entire Survey Area
The single and double-station techniques
generated two separate estimates of the
coefficients which are characteristic of
the study reach. A t-test can be used to
determine whether the "average" single-
station estimate for each of the coeffici-
ents Is significantly different from the
"average" double-station estimate. The
probabilities that the "true means" are
equivalent are 0.42, 0.11 and 0.09 for K2,
P and R respectively. Only the six
double-station reaches shown In Table 26
were Included In this test.
Casual Inspection of Table 25 seems to In-
dicate that both estimates of K2 Increase
In the downstream direction (with the ex-
ception of the 0.6-0.85 mlle double-station
estimate), but there are no apparent
patterns In the variations of P or R.
- 55 -
-------
Table 26
Mean Modeling Coefficients and their Standard Deviations for the Entire Survey Area
Coefficient
Single-Station Values
K2 P R
Double-Station Values
K2 P R
Mean
Standard Deviation
Number of Samples
2.83
1.30
7
9.3
1.6
7
14.2
2.0
7
3.75
1.74
6
11.3
2.8
6
18.1
4.4
6
The above data and discussion suggest that
the area represented by the single-station
method Is smaller than the area (time of
travel) required to reduce the Incoming
deficit to a small quantity. If a study of
this nature was conducted where the reaera-
tlon rate was higher, a better definition
of the area represented by the single-
station method might be possible. A com-
puter simulation of the effects of changing
values of P, R and K2 would also be useful.
Double Station Coefficients and "Average
Double Station Coefficients"
Figures 21 through 23 represent the com-
parison of calculated coefficients and co-
efficients determined by averaging the co-
efficients returned for the shortest
reaches. This exercise was carried out to
demonstrate the failure of the "double
station differential" method when reach
length, or the product "K2TOT" becomes too
large. The plots demonstrate that as TOT
Increases, the estimates of R and K2 fall
further from the theoretical 1:1 line which
would be expected If no failure occurred.
The estimates of P seem relatively un-
affected by reach length. R and K2 co-
efficients returned from the longest
reaches are much higher than would be ex-
pected from the averages of the short
reaches. Even the shortest "averages",
I .e. those returned from the average of two
one hour reaches, lie above the 1:1 line
that would be expected If only random
errors In the determinations were the
cause. This suggests that a systematic
error Is Involved, and that the error In-
creases with Increasing reach length.
A slightly anomalous point Is that the
"statistical fit" of the determinations
generally Improves with Increases In travel
time. The above diagrams and figures In-
dicate that coefficients derived from the
long travel times are Incorrect. This sug-
gests that "statistical fit" (e.g. large
r-square values and small confidence
limits) ts not necessarily a good Indicator
of "correctness".
Integrated and Differential Coefficients for
the Double Station Technique
The Integrated equation was used to cal-
culate a second set of double-station co-
efficients for the WLA data set. The re-
sultant coefficients are compared with
those returned by the differential equation
In Figures 24 through 26. A t-test was
performed to test whether or not the two
methods (differential and Integrated) re-
turned mean values for P, R, and K2 which
were significantly different. The mean
values, their standard deviations and the
probability that the mean values are
statistically equal are presented In
Table 27. (A value less than 0.05 Is
commonly Interpreted to Indicate
statistical Inequality.)
The results of these tests Indicated that
the "whole-reach" mean values returned by
the two analytical methods (differential
and Integral) were not significantly dif-
ferent at the five percent level.
A paired t-test was also performed to
determine whether the reach-by reach dif-
ferences between analytical methods were
significant. The results Indicated that P,
- 56 -
-------
65
B 10
AVE. ESTIMATE OF PHOTOSYNTHESIS
Figure 21
Average differential photosynthesis coefficients versus
measured differential photosynthesis coefficients
(gm 0?/fo*/dny). The line drawn represents a "perfect"
relationship (slope = I. Intercept = 0).
3-
4.5-
4 -
3.5 -
3 -
2.3 -
2
AVE ESTIMATE OF REAERATION
Figure 22
Average differential K2 versus measured differential
K2 coefficients (gm 09/mVday-1). The line drawn
represents a "perfect" relationship.
Note: Symbols represent different travel times:
D » 2-3 hours. + - 3-4 hours. O= 4.5 hours, 4=6 hours, x = 7.2 hours
-------
I
$
I
a
i/i
a:
17 19
AVE ESTIMATE OF RESPIRATION
21
12
14 16 IB 2O
DIFFERENTIAL. RESPIRATION COEFFICIENTS
22
24
Figure 23
Average differential respiration versus measured
differential respiration coefficients (gtn 0?/m*/day).
The line represents a "perfect" relationship.
Figure 24
Comparison of differential and Integrated
double-station coefficients for respiration (gm 07/m2/day).
The line represents a "perfect" relationship.
Note: Symbols represent different travel times:
0= 2-3 hours, + = 3-4 hours. O = 4.5 hours. A « 6 hours, x = 7.2 hours
-------
I
S
I
IU
u
s
s
DIFFERENTIAL COEFFICIENTS
Figure 25
Comparison of differential and Integrated double-station
coefficients for photosynthesis (gm 0?/mz/dayT7
The line represents a "perfect" relationship.
DIFFERENTIAL COEFFICIENTS
Figure 26
Comparison of differential and Integrated
double-station coefficients for reaeratlon (I/day).
The line represents a "perfect" relationship.
Note: Symbols represent different travel times:
a " 2-3 hours. + = 3-4 hours. «= 4.5 hours, A» 6 hours, x = 7.2 hours
-------
Coefficient
Table 27
Comparison of Mean Double-Station Coefficients
Double-Station Coefficients
Integrated Differential
Mean
S.D.
N
Mean
S.D.
P (mg/l/day)
R (mg/l/day)
K2 (mg/l/day)
10.59
20.06
4.67
2.69
4.38
2.02
15
15
15
11.85
19.75
4.37
2.28
3.72
1.62
18
IB
18
0.39
0.83
0.65
R and K2 estimates from the differential
method were significantly different from
the Integral estimates of P, R and K2
.9) reaches.
The results of the "differential" :
"average differential" comparison were
mixed. The differential equation failed to
return "average coefficients" (for R and
K2) for long reaches, as expected. The
fact that the differential estimates of P
for the longest reaches were not apprecia-
bly different from the expected average
values Indicates that linear processes can
be accurately determined In very long
reaches. The failure for K2 (and therefore
R) Is possibly related to Inappropriate
determination of the "average deficit" to
be applied over the time of travel. The
results suggest that time of travel for the
double-station differential method be
limited to one to two hours. It Is un-
certain as to whether the time limit should
be shorter In streams where reaeratlon Is
higher.
- 60 -
-------
The Integrated and differential comparison
showed unexpectedly good agreement be-hteen
the two methods. The results were similar
to the "differential" : "average differen-
tial" comparison, with best agreement
occurring between differential and Integral
estimates of P. The estimates of R and K2
from the long reaches seemed to substan-
tially overestimate the true (as Indicated
from weighted averages of short reaches)
values. The Integral equation, as Imple-
mented, did not alleviate problems associ-
ated with long travel times. The reason
for this Is unclear. Further Investigation
In this area Is warranted.
SUMMARY OF DIEL STUDIES
error, according to the bottle and box
studies. Further studies would be needed
to define the magnitude of the error, and
If Indeed It would be possible to correct
this error on a routine basis.
Finally, each of these dlel surveys covered
approximately twenty-four hours. Day to
day variations In light levels are likely
to add variability to the determinations of
"P".
From the results of the simulated wasteload
study. It appears that continuously re-
corded data (light, temperature and DO)
collected over a period of about 48 hours
provided the best (In a statistical sense)
data.
There are many possible reasons for the
scattered dlel estimates of P, R and K2.
Inaccuracy of data collection techniques
could be a major source of error. The
double-station technique suffers more from
slight Inaccuracies In DO measurement than
does the single-station technique. Season-
al changes In the photosynthetlc efficiency
of the plants, self-shading, and the pres-
ence or absence of perlphyton as other
photosynthetlc agents could all lead to the
observed results.
In addition, It appears that the assump-
tions upon which the model Is based are
sometimes violated. Figures 27 and 28 show
several examples where light and photo-
synthesis are apparently not linearly
related. These graphs are from the
double-station analysis, and represent the
change In DO across a reach, corrected for
reaeratlon, at different light levels. It
Is worth noting that the curves "flatten
out" around 30x10'5 Q/cm2/sec, which Is
near the range of 25-30xl015 0/cm2/sec
quoted In the literature as being a
"typical light saturation level". This
non-11 near I ty may or may not lead to an
error In the estimate of "P" and "R". (In
most cases It does not appear to be a
serious error.) It Is an Inaccuracy In the
technique, and deserves attention.
The assumption that "R" remains constant
throughout the night Is probably also In
The choice of the modeling method used to
determine the K2 and R value for a given
stream Is complicated by the fact that the
exact relation between single and double-
station coefficients has not been adequate-
ly explored. It Is clear, however, that
the results of the two methods will differ
In most cases, and that they do so because
of their different determinations of the
term.
It seems prudent to conduct waste- loads
with a design conducive to double-station
analysis wherever possible. The small
amount of extra effort and expense will
provide additional (and possibly better)
data. If problems are encountered with the
double-station analysis, the data are still
available for single-station analyses.
If the differential double-station method
Is used, care must be taken to choose
stations which are relatively close, with
less than two hours travel time between
them. Travel times shorter than two hours
may be advisable where K2 Is expected to
exceed 5 /day. (Further research Is needed
to determine more exact time of travel re-
strictions.) When travel time Is less than
one half hour, very accurate determinations
of the ^00/£f- term are necessary. Poor
resolution of the &00/£$ term may make
analysis very difficult and prone to In-
accuracy.
- 61 -
-------
Figures 27 and 28
Examples from the Ashlppun and Bark rivers demonstrating non-11 near Ity of
the relationship between light and photosynthesis (apparent light saturation).
CM
o
o>
0.1
0
-0.1 -
-O.2 -
» -0.4 -
> -0.5 -
-0.6 -
-0.7 -
a:
Q. -0.8 -
Ashippun R.
July 8, 1982
2O 40
LIGHT (Q/cm2/s«c X 1015)
60
1
1
"5
o
E
v^
\
O
o>
E
Q£
0.
1 8
1.6 -
1.4 -
1.2 -
1 -
0.8 -
0.6 -
0.4 -
0.2 -
0 -
-0.2 -
-0.4 -
-0.6 -i
O 8 -
A A
A
A A
A *
Bark R.
A July 6, 1982
A
A
u'° ' i i i 1 1 1 1 1 1 |
0 20 -40 60 8O 10
LIGHT (Q/cm2/sec X 1015)
— 62 —
-------
The Integrated double-station method, as
Implemented In this study, remains re-
stricted by the assumption that sum
(frfP+R)/K2) remains constant. It thus
seems reasonable to restrict the travel
time to less than two hours In this method.
POTENTIAL METHOD OF ALLOCATING PHOSPHORUS
The foregoing data and discussion suggest a
simple way In which phosphorus may be In-
corporated Into the waste toad allocation
process. Field surveys or existent data can
provide estimates of reaeratlon and respi-
ration. If a change In the amount of phos-
phorus which will be discharged Is esti-
mated, and the upstream phosphorus concen-
tration Is known, the macrophyte blomass
can be estimated from the models presented
earlier In this paper. The change In bio-
mass projected by the model (I.e. "Blomass
predicted at present phosphorus level"
minus "Blomass predicted at the new (pro-
jected) phosphorus level") can be used to
estimate a change In the respiration rate.
Experiments presented elsewhere In this
paper, as well as In the literature, sug-
gest that macrophytes respire at the rate
of 1.5 g 02/kg DW/hr, which Is equivalent
to 36 g 02/kg DW/day. If the blomass Is
projected to Increase by 100 g/m2, this
would Increase the area! respiration rate
by 3.6 g 02/m2/day, which can be translated
Into a mg/l/day estimate by multiplying by
average depth and dividing by the number of
liters present over a square meter area at
a depth of one foot (I f the average depth
was one foot, 3.6 g 02/2/day x I ft x I
m2/304.8 I = 11.8 mg 02/l/day). The new
respiration rate, the "measured respiration
rate + projected change", can be divided by
the "estimated or measured K2 rate" to
estimate what the new maximum deficit will
be If K2 Is greater than 5/day. If K2 Is
less than 5/day, the projected change can
be added to the respiration term In the
model to estimate the maximum deficit.
The advantage of the "R/K2" methodology
lies In Its ability to predict the maxi-
mum deficit regardless of daytime DO fluc-
tuations. Application of the Integrated
model should yield approximately the sane
result, yet Is much more cumbersome.
In streams where K2 exceeds 5/dey, the
equilibrium concept will also allow the
Impact of non-point sources to be quanti-
fied, If an Increase In phosphorus or sedi-
ment oxygen demand can be linked to
non-point sources. Difficulties Inherent In
projecting a new photosynthesis rate make
application of the Integrated model diffi-
cult In situations where K2 Is less than
5/day.
This procedure would augment the present
BOD allocation process. In conjunction
with CBOD and NBOD, some level of phos-
phorus with attendant Increases In plant
blomass and respiration would result In
violation of the 5 mg/l criteria.
CONCLUSIONS
The dissolved oxygen concentration of small
streams has been modeled as a function of
three basic terms: P, R, and K2.
Under high reaeratlon conditions
(K2>5/day), the maximum deficit Is essen-
tially specified by the quotient R/K2, and
phosphorus Induced Increases In blomass
(see Macrophyte section) can be expected to
Increase the maximum deficit (decrease the
minimum DO level).
Under lower reaeratlon conditions
(K2<5/day), the result of Increasing blo-
mass Is less clear, but should still be
mode I able through the present "BOD allo-
cation process". If the modeled area In-
cludes less than twelve hours of travel
time, then analysis of the night time
changes can give the "expected deficit at
sunrise" (the Incoming deficit at sunset
(upstream of discharge) would be expected
to remain unchanged). Where more than 12
hours of travel time are Included In the
modeled area, some approximation of the
deficit at sunset Is necessary. This
approximation would have to be done on a
site specific basis.
- 63 -
-------
The allocation of phosphorus under these
guidelines Is very similar to, yet separate
from the present BOO allocation pro-
cedures. Some Increase In phosphorus
levels could be expected to result In vio-
lation of the 5 mg/l stream DO standard,
even I f BOD leve I s are reduced to
negligibly small quantities.
- 64 -
-------
SUMMARY AND CONCLUSIONS
The purpose of the Phosphorus and High-Flow
Assessment study was to define methods of
dealing with phosphorus In setting appro-
priate water quality goals or standards.
The overall study Initially evaluated the
feasibility of both point and non-point
source control of phosphorus. The field
study reported here addressed the nutrient
control objective, and was aimed at de-
fining low-flow or sustained stream phos-
phorus contributions rather than high-flow,
or event-related phosphorus loadings.
In addition to Investigating the Impacts of
phosphorus In small stream systems, the
field study also evaluates methods of docu-
menting phosphorus Impacts In streams and
recommends monitoring strategies.
Stream and sediment nutrients were compared
to rooted plant and attached algae growth
In selected southeastern Wisconsin stream
reaches In 1981 and 1982. The Impacts of
In-stream nutrients and plant growth on
stream dlel dissolved oxygen (DO) charac-
teristics were also Investigated. These
reaches provided a variety of physical and
biological characteristics as well as a
wide range of water and sediment nutrient
conditions. Streams receiving wastewater
treatment plant effluents were also In-
cluded In the study.
STREAM MACROPHYTES
Based on frequency of occurrence of macro-
phytes on specific substrate types, sedi-
ment Interstitial water/stream phosphorus
concentration ratios and macrophyte tissue
nutrient concentration data,, the study
reaches were categorized Into two groups.
Type I and Type II. Various Investigators
have reported rapid exchange between sedi-
ment pore water and overlying stream water
In larger substrate sizes. Macrophytes In
Type I streams, characterized as growing
over larger substrate sizes, are suspected
of obtaining growth nutrients from the
overlying water. Significant relationships
were described between stream phopsphorus
concentrations and macrophyte blomass In
Type I streams. Based on these relation-
ships, a predictive equation was developed
which predicts maximum summer blomass from
mean summer (June-August) P04P concen-
trations (Model I).
In these stream types, the model predicting
maximum plant blomass from In-stream P04P
Is most applicable and may provide a good
predictive tool for assessing phosphorus
Impacts In streams. Macrophytes In Type II
streams, occur Ing over primarily silt sub-
strates, are suspected of deriving growth
nutrients from the sediments.
Macrophyte tissue nutrients were also sig-
nificantly related to In-stream nutrient
concentrations. A predictive equation was
developed which describes macrophyte tissue
phosphorus concentrations as a function of
mean summer In-stream P04P concentrations
(Model II). As with the macrophyte blo-
mass/stream P04P model, tissue nutrient
concentrations of macrophytes In those
stream reaches Identified as Type II, were
higher than that which could be obtained
from the ambient water alone. These Type
II streams (where sediments are suspected
of being the primary nutrient source) were
clearly Identified as not belonging to the
the Type I stream relationship. This model
(Model II) Indicates that the maximum
tissue phosphorus concentration Is depen-
dent on the In-stream mean summer P04P con-
centration In Type I streams.
A third equation (Model III) was developed
to describe the relationship between macro-
phyte tissue phosphorus concentration and
maximum stable summer plant blcmass. This
model Is somewhat sensitive to timing of
the harvesting as tissue nutrients are
rapidly lost from senesclng plants.
The three least squares regression models
presented may provide an alternative method
for determining the primary macrophyte
nutritive source, and specify the proper
model to assess stream macrophyte pro-
duction and phosphorus Inputs In different
stream Types. This provides a basic tool
- 65 -
-------
with which to determine existing levels of
macrophyte blomass and project changes In
stream macrophyte populations due to
changing phosphorus Inputs. It Is sug-
gested, however that the results and the
macrophyte models be further tested to Im-
prove their applicability to a larger
number of situations.
The macrophyte mapping and harvesting
methods developed and refined during the
study appear to adequately describe stream
macrophyte communities. These methods are
similar to those employed In present Waste-
load Allocation Surveys. Based on the re-
sults of this study, a monitoring protocol
with recommendations for Its use In
P-assessment surveys Is appended to the
study report.
STREAM PERIPHYTON
Stream perlphyton were harvested from
glass-slide samplers and bricks, exposed
for two and four weeks, respectively. The
two-week exposure of glass-slide perlphyto-
meters was employed to test what Is usually
considered an optimum or "standard" ex-
posure period. The longer brick exposure
periods were designed to approximate a
naturally occurring perlphyton population's
response to nutrients.
Chlorophyll-a was positively correlated
with In-stream phosphorus and Inorganic
nitrogen concentrations. Brick chlorophyll,
however, was most strongly related to
In-stream P04P. Brick values were also
more strongly correlated with stream
nutrients than perlphytometer values.
Based on these relationships, a least
squares regression model was calculated
describing the Brick chlorophyll and
In-stream P04P relationship (Model IV).
Ash-free weight blomass measurements did
not appear directly related to stream
nutrient concentrations.
Perlphyton tissue nutrients were also
highly correlated with In-stream
nutrients. Similar to the chlorophyll-a:
In-stream P04P relationship, brick collec-
tions were more strongly correlated with
nutrients than glass-slide collections. A
least squares regression model was calcu-
lated describing brick perlphyton tissue
nutrients as a function of In-stream P04P
(Model V).
Although perlphytometer chlorophyll did
correlate with In-stream phosphorus,
In-stream sampling variability and the In-
fluence of physical factors (e.g. current
velocity, shading, temperature) precluded
serious modeling effort.
Brick perlphyton chlorophyll was also cor-
related with tissue phosphorus and a model
calculated to express this relationship
(Model VI). The perlphyton results suggest
that bricks, placed for four-week exposure
periods, more closely reflect nutrient Im-
pacts than glass-slide collections exposed
for two weeks. The correlations of brick
nutrients with water nutrient concen-
trations, similar to macrophyte tissue and
In-stream nutrients suggest that brick
perlphyton collections, representing a
naturally occurring perlphytlc community,
support the marophyte study results.
SEDIMENT
Sediment Interstitial water nutrients and
bulk sediment nutrient samples were
collected within macrophyte-popuIated areas
and outside of macrophyte areas.
Macrophyte blcmass was positively corre-
lated with sediment Interstitial P04P. No
correlation was apparent between Inter-
stitial nitrogen and macrophyte para-
meters. Sediment Interstitial P04P was
also correlated with In-stream P04P concen-
trations. Inorganic nitrogens less so.
Statistical T-tests showed no significant
differences between nutrients In macrophyte
beds and concentrations outside of these
areas within stream reaches. Bulk sediment
nutrient content was also not clearly re-
lated to Interstitial nutrient concen-
trations.
These results substantiate the relation-
ships described by the stream macrophyte
- 66 -
-------
results, suggesting that macrophyte
nutritional needs In Type I streams are
satisfied primarily through shoot absorp-
tion or Indirectly through stream water
exchange through the substrate.
DIEL STUDIES
Single-station and doOble-statlon dlel dis-
solved oxygen analyses were conducted
monthly In 1981 and 1982. The purpose of
the modeling was to determine photosyn-
thesis, respiration and reaeratlon values
for the date each stream was monitored. In
situ light and dark bottte and box studies
were also conducted to Independently
measure photosynthesis and respiration.
Estimates of photosynthesis from the
modeling results generally agreed with In
situ measurements. Light saturation was
also demonstrated In many of the dlel
curves. Measured respiration rates from
both the bottle and box studies show good
agreement with values reported In the
literature. Modeling estimates of respira-
tion, however, were usually higher than
could be accounted for by measured plant
blomass alone. This Is attributed to other
forms of biological respiration and sedi-
ment oxygen demand.
Results from the study and theoretical
developments Indicate that by Increasing
primary producer populations, phosphorus In
streams will Impact photosynthesis,
respiration and reaeratlon capacity. Plant
growth In streams will result In Incremen-
tal Increases In community photosynthesis
(oxygen production) and respiration (oxygen
consumption) and decreased stream re-
aeration. The effect which substantial
macrophyte growth can have on stream re-
aeration capability Is potentially severe.
This Is due to ponding of stream water by
macrophytes, decreasing surface area to
volume ratios, and Increasing channel
roughness.
Using the macrophyte blomass/P04P model
presented In this report, the maximum
stream dissolved oxygen (DO) deficit
(mlnlmuii night-time DO) at sunrise appears
mode I able as a function of stream PO4P.
This should be workable as long as the area
modeled has a time of travel which Is less
than the night length. If time of travel
Is greater than night length, the "deficit
at sunset" must be specified.
Stream reaeratlon will also Influence the
ability to specify or predict the
night-time maximum DO deficit. Where re-
aeration is low (less than 5/day) the small
stream wasteload modeling process can be
used to project the DO deficit at sunrise.
Where reaeratlon Is high (greater than
5/day), the small stream model predicts the
deficit will equal respiration divided by
the reaeratlon rate (R/K2). Plots of R/K2
against the maximum observed deficit showed
that the predicted equilibrium deficit
level (R/K2) Is commonly achieved.
SITE-SPECIFIC PHOSPHORUS ALLOCATIONS
The overall results of the Phosphorus
Assessment field studies Indicate a need
for phosphorus control In small stream sys-
tems. Inability to define phosphorus as a
pollutant In a traditional sense has
hindered establishment of phosphorus water
quality standards In Wisconsin. Relation-
ships between stream phosphorus, macrophyte
blomass and stream DO characteristics
appear to provide a method to approach
phosphorus control.
The results of the primary producer and
dlel studies suggest that phosphorus could
be allocated to streams, on a site-specific
basis. This would be done In a manner
similar to current methods of allocating
Biochemical Oxygen Demand (BOD). Given the
macrophyte/phosphorus relationships
developed In this study. It appears pos-
sible, at least In streams where plants
obtain phosphorus from the water, to pro-
ject changes In macrophyte blomass based on
projected changes In In-stream P04P concen-
trations.
If minimum night-time dissolved oxygen can
be described by R/K2, then the additional
- 67 -
-------
respiratory oxygen demand of the projected
Increase due to macrophyte blomass can be
added as an additional form of BOO. In
addition, theory predicts that reaeratlon
(K2) will also decrease as macrophytes be-
come more abundant In the channel. This
effect would serve to drive the minimum
night-time DO concentration even farther
downward.
Other areas of key concern, attributable to
phosphorus. Include stream ponding, alter-
ation of channel characteristics, changes
In the ability of streams to maintain
"healthy" night-time dissolved oxygen con-
centrations, alteration of natural stream
habitat and production of undesirable
aesthetic conditions due to macrophyte
growth. Other agencies In the United
States have adopted phosphorus standards
based on the above concerns. Appropriate
criteria were generally applied through
water use classifications (US EPA I960).
In addition to the above Incorporation of
"Macrophyte BOO" Into the current WDNR
Waste load Allocation process, considering
these other concepts In development of
phosphorus control strategies for Wisconsin
Is recommended.
- 68 -
-------
LITERATURE CITED
Alba, S.. and H. Ohtake. 1977. Simulation of P04-P balance In a shallow and polluted river.
Wat. Res. 11: 159-164.
American Public Health Association, American Water Works Association and Water Pollution Control
Federation. 1981. Standard Methods for the Examination of Water and Wastewater. 15th ed.
1134 pp.
Auer, M.T. 1982. Ecological studies and mathematical modeling of Cladophora In Lake Huron:
3. the dependence -of growth rates on Internal phosphorus pool size. J. Great Lakes Res.
8(1): 93-99.
Auer. M.T., and R.P. Canale. 1982. Ecological studies and mathematical modeling of Cladoohora
In Lake Huron: 2. phosphorus untake kinetics. J. Great Lakes Res. 8(1): 84-92.
Babros, T.E. 1981. Comparison of two perlphyton col lection methodologies. Intern Report. 12
p. Southeast District, Wisconsin Department of Natural Resources, Milwaukee.
Ball. J. 1982. Stream Classification Guidelines for Wisconsin. WDNR, Madison, Wl.
Ball, R.C., N.R. Kevern and T.A. Halnes. 1973. An ecological evaluation of stream
eutroptiicatlon. Michigan State University, Institute of Water Resources, Research Rept. No.
36.
Barko, J.W., and R.M. Smart. 1981. Sediment-based nutrition of submersed macrophytes. Aquatic
Botany 10: 339-352.
Blaln, W.A. and A.J. McDonnell. 1967. Reaeratlon Measurements In a eutrophlc stream. Proc.
22nd. Ind. Waste Conf., Purdue 22: 1044-1058.
BouId, C., E.G. Bradfleld, and G.M Clarke. I960. Leaf analysis as a guide to the nutrition of
fruit crops. I. General principles, sampling techniques, and analytical methods. J. Scl.
Food Agr. II: 229-242.
Buchanon, T.J. and W.P. Somers. 1973. Discharge measurements at gaging stations. Techniques
of Water Resource Investigations of the United States Geological Survey. Book 3, Chapter
A3. U.S. Government Printing Office, Washington , D.C.
Calnes, L.A. 1965. The phosphorus content of some aquatic macrophytes with special reference
to seasonal fluctuations and application of phosphate fertilizers. Hydroblol. 25: 289-301.
Canale, R.P. 1982. Ecological studies and mathematical modeling of Cladophora In Lake Huron;
7. model verification and system response. J. Great Lakes Res. 8(1): 134-143.
Canale. R.P., and M.T. Auer. 1982. Ecological studies and mathematical modeling of Cladophora
in Lake Huron: 5. model development and calibration. J. Great Lakes Res. 8(1): 112-125.
Canale, R.P., M.T. Auer, and J.M. Graham. 1982. Ecological studies and mathematical modeling
of Cladophora In Lake Huron: 6. seasonal and spatial variation In growth kinetics. J.
Great Lakes Res. 8(1): 126-133.
- 69 -
-------
Canfleld. D.E. Jr., and R.W. Bachrnann. 1981. Prediction of total phosphorus concentrations,
chlorophyll a, and Secchl depths In natural and artificial lakes. Can. J. Fish. Aquat. Scl.
38: 414-423.
Carlgnan, R. 1982'. An empirical model to estimate the relative Importance of roots In
phosphorus uptake by aquatic macrophytes. Can. J. Fish. Aquat. Scl. 39: 243-247.
Carlgnan, R. and J. Kalff. 1980: Phosphorus sources for aquatic weeds: water or sediments?
Science 207: 987-989.
Cattaneo, A. and J. Kalff. 1980. The relative contribution of aquatic macrophytes and their
epiphytes to the production of macrophyte beds. Llmnol. and Oceanogr. 25(2): 280-289.
Chantry, W.A. WDNR. 1981. "Analysis of Recreational Water Quality Criterion." Interim Project
Document.
Chapman, H.D. 1966: Diagnostic criteria for plants and soils. Univ. California, Dlv. Agr. Scl.
793.
Chapra, S.C. 1975. Comment on "An empirical method of estimating the retention of phosphorus
In lakes" by W.B. Klrchner and P.J. Dillon. Water Resour. Res. II: 1033-1034.
Corbett, D.M'. 1945. Stream Gaging Procedure. U.S. Gaol. Surv. Water-Supply Paper 888.
Dillon, P.J., and F.H. Rlgler. 1974. The phosphorus-chlorophylI relationship In lakes.
Llmnol. Oceanogr. 19: 767-773.
Edberg, N., and B.V. Hofsten. 1973. Oxygen uptake of bottom sediments studied In situ end In
the laboratory. Water Research 7: 1285-1294.
Edwards, R.W'., and M. Owens. 1962. The effects of plants on river conditions IV. the oxygen
balance of a chalk stream. J. Ecol. 50: 207-220.
Gerloff, G.C'. 1969. Evaluating nutrient supplies for the growth of aquatic plants In natural
waters. 537-555. In; Eutrophlcatlon; Causes, Consequences, Correctives. National Academy
of Sciences, Washington, D.C.
Gerloff, G.C., and P.H. Krcmbholtz. 1966. Tissue analysis and measure of nutrient availability
for the growth of anglosperm aquatic plants. Llmnol. Oceanogr. II: 529-537.
Graham, J.M., M.T. Auer, R.P. Canale, and J.P. Hoffman. 1982. Ecological studies and
mathematical modeling of Cladophora In Lake Huron: 4. photosynthesis and respiration as
functions of light and temperature. J. Great Lakes Res. 8(1): 100-111.
Grant, R.S., and S. Skavroreck. 1980. Comparison of tracer methods and predictive equations
for determination of stream reaeretlon coefficients on three small streams In Wisconsin.
U.S. Geological Survey Water-Resources Investigations,80-19.
Holdren, G.C. Jr., O.E. Armstrong and R.F. Harris. 1977. Interstitial Inorganic phosphorus In
lakes Mendota and Wlngra. Water Research II: 1041-1047.
Hornberger, G.M., and M'.G. Kelly. 1975. Atmospheric reaeratlon In a river using productivity
analysis. J. Env. Engr. Dlv. ASCE 101: 729-739.
- 70 -
-------
Horner, R.R., and E.B. Welch. 1981. Stream perlphyton development In relation to current
velocity and nutrients. Can. J. Fish, and Aqua. Scl. 38: 449-457.
Hoyer, M.V., and J.R. Jones. 1983. Factors affecting the relation between phosphorus and
chlorophyll a In mldwestern reservoirs. Can. J. Fish. Aquat'. Scl. 40(2): 192-200.
Hynec, H.B.N. 1970. The Ecology of Running Waters. Univ. Toronto Press. 555p.
Jones, J'.R., and R.W. Bachmann. 1976. Prediction of phosphorus and chlorophyll levels In
lakes. J. Water Pollut. Contr. Fed. 48: 2176-2182.
Klrchner, W.B., and P.J'. Dillon. 1975. An empirical method of estimating the retention of
phosphorus In lakes. Water Resour. Res. II: 182-183.
Kullberg, R.G. 1974. Distribution of aquatic macrophytes related to paper mill effluents In a
southern Michigan stream. Aner. Midi. Nat. 91 (2): 271-281.
Larsen, D.P. and H.T. Mercler. 1976. Phosphorus retention capacity of lakes. J. Fish. Res.
Bd. Can. 33: 1742-1750.
Lasslter, R.L. 1975. Modeling dynamics of biological and chemical components of aquatic
ecosystems. Southeast Environmental Research Laboratory, National Environmental Research
Center-CorvolI Is. U.S.-EPA. Athens, Georgia.
Lehman, J.T., D. B. Botkln & G.E. Likens. 1975. The assumptions & rationales of a computer
model of phytopIankton population dynamics. Llmnol. Oceanogr. 20: 343-364.
Lundegardh, H. 1951. Leaf analysis. HI Iger and Watts, London. I76p.
McDonnell, A'.J. 1971. Variations In oxygen consumption by aquatic macrophytes In a changing
environment. Proc. 14th Conf. Great Lakes Res., Int. Assoc. Great Lakes Res. 52-58.
McDonnell, A.J. and D.W. Weeter. 1971'. Variable respiration In aquatic plant communities.
J. Sanlt. Engng. Dlv. Am. Soc. Civ. Engrs. 98 SA2: 325-339.
McDonnell, A.J. 1982. Oxygen budgets In macrophyte Impacted streams. Water Research 16:
1037-1046'.
McRoy, C.P., and R'.J. Barsdate. 1970'. Phosphate absorption In eel grass. Llmnol. Oceanogr.
15: 6-13.
McRoy, C.P., R.J. Barsdate and M. Nebert. 1972. Phosphorus cycling In an eel grass (Zostera
marina L.) ecosystem. Llmnol. Oceanogr. 17: 58-67.
Mace, S.E., T.F. Lowry and P'.W. Sorge. 1981. Methods of Assessing the Impacts of phosphorus on
stream water quality - PHARTS field study design. Project document, Milwaukee, Wl: 15pp.
Mace, S.E., T.F. Lowry and P.W. Sorge. In Prep. Literature review, phosphorus and primary
production In streams.
Mulligan, H.F., and A. Baranowskl. 1969. Growth of phytopIankton and vascular plants at
different nutrient levels. Verb. Int. Vereln. Theor. Angew. Llmnol. 17: 802-810.
Narf, R. WDNR. In Prep. Blotlc Index Quality Assurance Report.
- 71 -
-------
Nichols. D.S., and D.R. Keeney. 1973. Nitrogen and phosphorus release from decaying water
milfoil. Hydroblol. 42: 509-525.
O'Shaugnessy. J.C. and A.J. McDonnell. 1973. Criteria for estimating limiting nutrients In
natural streams. Pensylvanla State University. Research Publ. No. 75. 91 pp.
Owens, M1. 1961. The effects of plants on river conditions III. Further crop studies and
estimates of net productivity of macrophytes In chalk streams. J. Ecol. 49: 119-126.
Owens, M., and R.W. Edwards1. 1962. The effects of plants on river conditions III. Crop
studies and estimates of net productivity of macrophytes In four streams of southern
England. J. Ecol. 50: 157-162.
Owens, M'., and P.J. Marls. 1964. Some factors affecting the respiration of seme aquatic
plants. Hydroblol. 23: 533-543.
Pltcalrn, C'.E.R., and H.A. Hawkes. 1973. The role of phosphorus In the growth of Cladophora.
Water Res. 7: 159-171. c
Reckhow, K.H. 1979. Uncertainty applied to Vollenwelder's phosphorus criterion. J. Water
Pollut. Cont. Fed. 51: 2123-2128.
Rosemarln, A'.S. 1982. Phosphorus nutrition of two potentially competing filamentous algae,
Cladophora glomerata
-------
United States Geological Survey. 1978. National Handbook of Recommended Methods for Water-Data
Aqulsltlon. Chapter 3, Sediment.
Ulrich, A. 1952. Physiological bases for assessing the nutritional requirements of plants.
Ann. Rev. Plant Physlol. 3: 207-228.
Vannote, R.L., 6.W. Mlnshall, K.W. Cummins, J.R. Sedell and C.E. Gushing. 1980. The river
continuum concept. Can. J. Fish. Aquat. Scl. 37: 130-137.
Vaux, W'.G. 1962. Interchange of stream and Intragravel water In a spawning riffle. Special
Scientific Report—Fisheries No. 405. United States Fish and Wildlife Service, Washington,
D.C. II p'.
. 1968. Intragravel flow and Interchange of water In a streambed. U.S.
Fish and Wildlife Service, Fisheries Bulletin 66: 479-489.
Vollenwelder, R.A. 1975. Input-output models with special reference to the phosphorus loading
concept In limnology. Schwelz. Zelt. Hydro). 37: 53-84.
Walsel, Y., and Z. Shaplra. 1971. Functions performed by roots of some submerged hydrophytes.
Israel J. Bot. 20: 69-77.
Warn, T. U.S. EPA-V. Nov 1980. "A Review of Chemical Water Qua!Ity Indices." Interim Project
Document.
Wawrzyn, W. and R. Randall. WDNR-SED. 1983. "Application of the National Sanitation
Foundation Water Quality Index and the Syracuse Water Use Index In the Root River,
Wisconsin." Interim Project Document.
Westlake, D.F'. 1966. A model for quantitative studies of photosynthesis by higher plants In
streams. Air and Water Pollut. Int. J. 10: 883-896.
Wetzel, R.fi. 1975. Primary production. J^B.A. Whltton, ed. River Ecology. Berkley, CA:Unlv.
of California Press.
Whltton, B.A. (ed.). 1975. River Ecology. Berkley: University of Cat IfomI a Press. 725p.
Wilson, D.C. 1972. Phosphate nutrition of the aquatic anglosperm, Myrlophyllum exalbescens,
Fern. Llmnol. Oceanogr. 17: 612-615.
Wong, S.L. and B. Clark. 1976. Field determination of the critical nutrient concentrations for
Cladophora In streams. Jour. Fish. Res. Bd. Can. 33: 85-92.
Wright, J.F., P.O. HI ley. S.F. Ham and A.D. Berrle. 1981. Comparison of three mapping
procedures for river macrophytes. Freshwater Blol. II: 369-379.
- 73 -
-------
APPENDIX I
STUDY REACH DESCRIPTIONS
INTENSIVE SURVEY REACHES
SUGAR CREEK, WALWORTH CO., T4N-RI7E, 515
Sugar Creek (SC), a medium-gradient stream, drains predominantly agricultural and muck-farmed
land. The headwaters are ditched with extensive agricultural tile drainage. Sugar Creek Is the
most heavily nonpolnt source Impacted stream In the study.
The area upstream of the study reach Is predominantly wetlands and fresh-meadow, providing a
good buffer along the stream length. Other than overhanging terrestrial grasses and brush,
there Is virtually no shading of the reach. Increased turbidity during summer low-flow was
attributed to rough fish activity.
The Sugar Creek study area was about 60 m (195 ft) In length with an average width of 6.7 m (22
ft). Sample transects were numbered consecutively from downstream to upstream with approxi-
mately II m (40 ft) between transects. The mean Instream depth of this reach was 0.2 m (0.8
ft), with a mean annual flow of 15.4 cfs. Substrate was predominantly sand and gravel with silt
overlay once macrophytes became stab 11 shed. During the growing season (June-September), sub-
merged macrophytes were restricted to the shallower water and gravel/rubble substrate of the
left side of the channel.
ASHIPPUN RIVER. WAUKESHA CO., T8N-RI7E-S32
The Ashlppun River Is a low-gradient stream draining agricultural land. There Is little shading
of this section of the stream, with good buffer along the length Immediately above the study
area. Cattle pasturing adjacent to the river In the upper watershed appeared the most common
nonpolnt source problem. This reach Is also heavily Impacted by agricultural NPS pollution.
The watershed area Is roughly one-half the size of Sugar Creek's.
The Ashlppun study area Is divided Into three reaches, delineated by an old berm which at one
time served a mill (Ashlppun-Malnstem), and a small Island which divides the downstream flow
Into the North and South Branch reaches. The water serving the reaches Is essentially of the
same qualIty.
Ashlppun-Malnstem
The malnsteam reach (AM) In 1981 was approximately 40 m (130 ft) long with an average width of
5m (17 ft), and divided Into 5 transects, 9m (30 ft) apart. Submersed macrophyte growth was
restricted to the deeper, center channel from June-September. Emergent vegetation (Sparganlum
eurycarpum) occurred along the right bank. In 1982, mean depth of this reach was 0.4 m (1.2 ft)
with a mean annual flow of 15.5 cfs. Bottom materials were predominantly sand and gravel except
In the area associated with the Sparganlum, which Is a thick silt bed. There Is little direct
shading of the reach.
This reach was lengthened In 1982 for a total reach length of 235 m (780 ft), with a mean width
of 6 m (20 ft). Mean depth of the Ashlppun reach In 1982 was .5 m (1.3 ft). Predominant sub-
strate was sand and gravel.
- 74 -
-------
Ashlppun-North Branch
The Ashlppun-North Branch reach (AN) was approximately 26m (87 ft) long and averaged 4m (13
ft) wide. Submersed macrophyte growth occurred throughout the reach In June-September, 1981.
Substrate In this reach was primarily rubble/gravel with little silt. This area received shade
from bank willows for a brief period In the morning. This reach averaged 0.2m (0.7 ft) In
depth with a mean annual flow of 9.3 cfs, or 60$ of the Malnsteam flow. This reach was dis-
continued In 1982.
Ashlppun-South Branch
The South Branch reach (AS) was roughly 32m (104 ft) long and 3 m (10 ft) wide. Bottom
materials varied from sand and gravel to gravel and rubble. Macrophytes occurred primarily In
the gravel and rubble substrate. Mean depth of this reach was 0.2 m (0.6 ft) with a mean annual
flow of 6.0 cfs, with represents 40$ of the Malnsteam discharge.
Both the North and South Branch reaches were divided Into 4 transects, separated by 9 m and II m
(30 and 35 ft) respectively.
KOHLSVILLE RIVER, WASHINGTON 00., TI2N-RI8E-S35
The Kohlsvllie River (KR) Is a high gradient low order stream. The study reach was located at
CTH "D", upstream of the Impoundment at Kohl svl lie. For most of Its length above this point.
Kohl svl Me River Is shallow, limiting the fishery to forage fish. The water upstream of the
study reach Is predominantly agricultural, mostly In hay and grain crop production. The
KohlsvlMe study reach Is one of the smallest watersheds of the study.
The Kohlsvllle study reach was 56 m (185 ft) long and averaged 2 m (7.5 ft) In width. The reach
Is generally shallow, with rubble and gravel substrate. Mean depth during the 1981 study period
was 0.2 m (0.7 ft) with a mean annual flow of 4.2 cfs. The reach was divided Into 9 transects,
approximately 6 m (20 ft) apart. Perlphyton and mosses are the dominant primary producers,
sometimes growing In a thick, felt-like mat on the larger substrate classes; no macrophytes were
observed.
This reach Is shaded In the early morning by an Oak lot. The area Immediately upstream of the
study area, however. Is almost totally shaded.
This reach was discontinued In 1982.
BARK RIVER. WAUKESHA 00.
There were two study reaches on the Bark River In 1981 and three In 1982, one upstream and two
downstream of the Dela-Hart POTW outfal I.
Bark-Wolf, T7N-RI7E-S26
This study reach (BW), upstream of the Dele-Hart POTW outfall. Is roughly I.I km (0.7 ml) down-
stream of Crooked Lake. Land use upstream and adjacent to this study site Is primarily agri-
cultural, however there Is also a large percentage of recreationaI/open space land. Because of
the short distance between the study reach and the Crooked Lake outlet, there Is very little
land area contributing directly to the stream at this point, and there Is a good buffer area
along the stream.
- 75 -
-------
Substrate within the Bark-Wolf reach Is predominantly sand and gravel. There Is little macro-
phyte growth In the study area and for a distance up and downstream. There are, however, areas
upstream of the reach which support moderate perlphyton and macrophyte growth.
In 1981, the Bark-Wolf study reach was 68m (226 ft) long and averaged 9m (30 ft) In width.
Mean depth over the 1981 study period was 0.4 m (1.3 ft) with a mean annual flow of 26.0 cfs.
Bark-Wolf had the lowest nutrient concentrations of all the study sites, reflecting the In-
fluence of the upstream lake.
This reach was lengthened to about 300 m (1000 ft) In 1982.
Bark-Lurvey, T7N-RI7E-S35
The Bark-Lurvey study reach (BL) Is located 1.4 km (0.85 ml) downstream of the Dela-Hart out-
fall. The Dela-Hart treatment plant, which went on-line In August I960, has a design capacity
of 2.2 mgd and summer effluent limits of 10 tng/l for BOD5 and suspended solids, 2 mg/l
ammonia, 6 mg/l DO and pH of 7.6; there are no phosphorus limits. Incorporated Into the treat-
ment process are rotating biological contractors (blo-dlscs), sand filtration and cascade-type
final effluent aeration. Low ammonia concentrations, high N^-NC^N and P04P concentrations
are discharged. Mean discharge at the outfall was estimated at 2 cfs In 1981 and 1.6 cfs In
1982.
As with Bark-Wolf, a very small watershed area contributes to the Bark between the upstream
(Wolf) and the Bark-Lurvey reaches, agriculture being the dominant land use. The section of
stream below the outfall Is characterized by numerous gravel riffles and runs. There Is little
shading of this area and there Is a good buffer along the stream length.
In 1981, the Bark-Lurvey study reach Itself was 67 m (220 ft) long with an average width of 12 m
(40 ft). There were 12 transects, separated by 6 m (20 ft). Mean reach depth during the 1981
study period was 0.4 m (1.3 ft). The mean annual flow was 28.4 cfs. During the growing season
(June-September), reach depth Increased without a corresponding increasing In flow due to the
ponding effect of macrophytes.
A farm bridge splits the reach Into two sections. The downstream section Is characterized by
shallow water depth and predominantly rubble/gravel substrate; the upstream portion by deeper
water and gravel/sand substrate. The entire reach is dominated by submerged macrophytes, with
emergent vegetation (Sparganlum eurycarpum) along the right bank. The study reach Is mostly
unshaded.
This reach length was expanded to about 480 m (1600 ft) In 1982.
Bark-Masonic, T6N-RI7E-S3.
This study reach, 3 km (1.85 ml) downstream of the Dela-Hart POTW outfall (1.6 km or I ml below
the Lurvey site) was added In 1982. Predominantly agriculture and residential land drain to the
reach between this and the Lurvey site. The reach was added to Bark-Wolf and Bark-Lurvey as
permanent stations on the Bark River.
Substrate within the Bark-Masonic reach Is predominantly sand and gravel. There Is little
direct shading of this reach. Adjacent land supports Tamarack and shrub growth.
The Masonic reach Is about 242 m (800 ft) long with an average width of 13 m (44 ft). Nine
transects were established for mapping and macrophyte harvesting.
- 76 -
-------
SYNOPTIC SURVEY REACHES
Bark-Wahlschlaeger, T8N-«I8E-S23.
This site, located upstream of Nagawlcka Lake, Is approximately 74 m (245 ft) long with a mean
width of 9 m (30 ft). Mean depth at the time of harvesting was .4 m (1.2 ft). Substrate was
predominantly sand and gravel. There Is little direct shading of this reach.
Land use upstream and adjacent to the site Is predominantly wetlands and low-density residen-
tial. This reach Is not Impacted by a POTW discharge.
Mukwonaga River, T5N-RI8E-S25.
This site Is-approximately I km (0.5 ml) downstream of the Lower Phantom Lake dan. The reach
sampled was approximately 300 m (1,000 ft) long with a mean width of 16 m (53 ft). Substrate
was predominantly gravel. Mean depth at the August II, 1982 macrophyte harvest was .4m (1.3
ft).
This reach Is similar to the Bark-Wolf reach both In water chemistry and physical characteris-
tics. There Is no direct shading of the reach. Adjacent land use Is primarily open space/sedge
meadow.
MlIweukee River - Main Branch, TI3N-RI9E-SI8.
The Milwaukee River-Wain Branch reach was sampled September 2, 1982, approximately I km (.8 ml)
downstream of the Campbell sport POTW discharge. The satnpJe reach was 211 m (695 ft) long with
ten transects selected for macrophyte harvests. At the time of the survey, mean reach width was
9 m (31 ft) and mean depth was .3 m (I.I ft). Land, use .-adjacent to the reach was predominantly
meadow/open space with little direct shading of the reach.
Present WPDES permit effluent limits are 30 mg/l for BOD5 and suspended solids.
Milwaukee River - East Branch, TI2N-RI9E-S2.
The study site of the East Branch of the Milwaukee River Is approximately 91 m (300 ft) long
with a mean width of I2.m (38 ft). Twelve transects within this reach-were selected for mapping
and macrophyte harvesting. Tne reach was sampled on August 16, 1982. At that time, mean depth
was .3m (I.I ft). Substrate was predominantly gravel and sand with silt along the banks.
Land use adjacent to the sampling site Is sedge-meadow. There Is a mill-pond approximately 3 km
(.9 ml) upstream at New Fane.
White River, T2N-RI8E-SI7.
The White River reach is located approximately 2.6 km (1.6 ml) downstream of the city of Lake
Geneva POTW. Present WPDES Interim permit limits are 45 mg/l B005 and suspended solids and
1.0 mg/l total phosphorus. The existing POTW has trouble meeting the phosphorus limit. Dam
manipulation and seiches In Lake Geneva, which frequently cause water to flow over the spillway,
cause fluctuations In stream depth and velocity.
Macrophytes were harvested August 25, 1982. At this time, mean depth-was .5:m (1.7 ft). Reach
length was 288 m (950 ft) with a mean width of 9 m (31 ft). Substrate In the reach was pre-
dominantly gravel and sand.
- 77 -
-------
Land use above the site Is primarily grass meadow floodplaln. The stream Is generally unshaded,
however portions of the reach sampled receive shade during part of the day.
Scuppernong River, T5N-RI7E-SI9.
This site was approximately 91 m (300 ft) long, with a mean width of 5 m (17 ft). Mean depth at
the time of harvesting was .3m (I.I ft) with sand and gravel substrate. Harvesting was con-
ducted September 7. 1982.
Land use upstream and adjacent to the study reach Is predominantly agriculture. Portions of the
upper watershed are extensively ditched.
Pewaukee River, T7N-RI9E-S26.
This reach Is approximately 6.5 km (4 ml) downstream of Pewaukee Lake and the City of Pewaukee.
Land use adjacent to the reach Is primarily agriculture and low-density residential. The City
of Pewaukee, however, contributes substantial storm drainage to the river during wet weather.
The Pewaukee POTW, which wend off-line In October, 1981, also discharged to the Pewaukee River.
The study reach was approximately 91 m (300 ft) long with a mean width of 6m (19 ft). Ten
transects were selected for mapping and macrophyte harvesting, August 24, 1982. Mean depth at
the time of harvest was .2m (.5 ft). Substrate was predominantly gravel with sand.
Cedar Creek, TION-RI9E-SI3.
The Cedar Creek reach was located approximately 6.5 km (4 ml) downstream of Little Cedar Lake.
This reach was 86 m (290 ft) long with a mean width of 5 m (16 ft). Mapping and harvesting were
conducted at II transects within the reach on September 8, 1982. At that time, mean depth was
.2m (.5 ft). Substrate was predominantly gravel with rubble.
Land use adjacent to the reach is primarily agriculture with some wetlands contribution.
Mt. Vernon Creek, T5N-R7E-S2.
Mount Vernon Creek Is a groundwater-fed stream, supporting an excellent trout fishery. The
study reach was 120 m (400 ft) long with a mean width of 6 m (18 ft). Substrate was pre-
dominantly gravel with sand. Macrophyte harvests and stream mapping were conducted
September 10, 1982. At the time of harvest, mean reach depth was .4m (1.2 ft). Eleven tran-
sects within the reach were sampled.
Land use upstream and adjacent to the reach was predominantly agriculture and pasture.
Black Earth Creek, T8N-R6E-S36.
Black Earth Creek Is a groundwater-fed stream which also supports an excellent trout fishery.
The Black Earth study reach was approximately 112 m (370 ft) long with a mean width of 11 m (37
ft). Stream mapping and macrophyte harvesting were conducted at ten transects within the reach
on September 10, 1982. At that time mean depth of the reach was .4m (1.4 ft).
Black Earth Creek receives effluent from the Cross Plains POTW, approximately 11 km (7 ml) up-
stream of the study reach. Present permit limits are 30 mg/l BOD and suspended solids and
2/7 mg/l ammonia (summer/w Inter). Land use adjacent to and upstream of the reach was primarily
pasture and wetlands.
- 78 -
-------
Fox River (Upper Fox). TI2N-R9E-S4.
The Fox River study reach was sampled approximately I km (.6 ml) downstream of the Portage POTW
discharge. The reach was 136 m (450 ft) long with a mean width of 17 m (57 ft). Mapping and
macrophyte harvesting was conducted at ten transects within this reach. The Fox River study
reach was sampled August 26, 1982. At that time, mean reach depth was .5m (1.5 ft).
- 79 -
-------
APPENDIX 2
STREAM REACH SPECIES LIST
A species list Is presented for each stream reach at the time maximum blomass was harvested.
Species are given In order of percent occurrence. Percent occurrence was calculated from the
mapping data which was collected at the time of harvesting. The value given for percent
occurrence Is the number of times a species occurred divided by the total number of sample
points mapped In a reach. Mean relative abundance Is given for each species. Species abundance
ratings were assigned to each species for each occurrence of a given species according to the
criteria In Table 4. Mean relative abundance Is the mean of the species abundance ratings for
each species'.
Emergent species were not Included In the harvesting surveys. For this reason, some species may
occur on the species list but which were not harvested.
Ashlppun River
August 1981
Percent Relative
Occurrence Abundance
Potamogeton pectlnatus 28.7 1.63
Saglttarla rlglda 16.3 1.62
P. zoster I form Is 14.0 1.80
P. ampllfollus 10.I 1.56
SparganI urn eurycarpum 7.9 1,71
Ceratophy11 urn demersum 3.9 1.00
Lemna minor 1.7 1.00
August 1982
S. rlglda 40.7 2.73
S. eurycarpum 14.8 1.88
P. zoster I formIs 8.3 1.89
P. amp IIfollus 6.5 1.71
L. minor 1.9 2.00
P. pectlnatus 0.9 2.00
Heteranthera dubla 0.9 2.00
C. demersum 0.9 1.00
- 80 -
-------
Bark Rlver-Lurvey Farms
August
Percent
Occurrence
Relative
Abundance
H. dubla
L. minor
Valllsnerla amerlcana
P. nodosus
C. demersum
Nymphaea sp.
P. pectlnatus
Anacharls canadensls
MyrlophylI urn sp.
Sclrpus sp.
P. crlspus
August 1982
H. dubla
V. amerlcana
P. nodosus
P. pectlnatus
C. demersum
L. minor
S. eurycarpum
Myrlophylturn sp.
Sclrpus sp.
A. canadensls
P. zosterI form Is
P. crlspus
64.9
43.2
40.5
20.7
20.7
9.9
7.2
5.4
4.5
3.6
0.9
67.4
28.3
10.5
9.7
1.9
1.9
1.6
1.6
1.2
0.8
0.4
0.4
1.82
1.85
2.10
1.17
1.00
1.09
1.25
1.00
1.00
1.00
1.00
3.17
2.97
1.41
1.36
1.80
2.80
1.25
1.25
1.66
1.00
1.00
2.00
Bark River-Wolf Road
August 1981
Percent
Occurrence
Relative
Abundance
Potamogeton sp.
C. demersum
P. pectlnatus
MyrlophylI urn sp.
V. amerlcana
H. dubla
P. zosterI form Is
Sclrpus sp.
24.6
18.8
17.4
15.9
14.5
13.0
4.3
2.9
1.00
1.31
1.08
1.09
1.10
1.00
1.00
1.00
- 81 -
-------
Bark River-Wolf Road (con't)
Percent
Occurrence
Ausust 1982
Relative
Abundance
P. pectlnatus
MyrlophylI urn sp.
V. amerlcana
Potamogeton sp.
H. dubla
Najas flexllls
C. demersum
34.8
28.0
26.1
24.8
6.2
3.1
1.2
1.43
I.II
1.10
1.43
3.40
1.00
1.00
Sugar Creek
August 1981
P. amerlcanus
S. rlglda
August 1982
P. amerlcanus
S. rlglda
Percent
Occurrence
58.1
9.3
64.9
17.6
Relative
Abundance
2.90
2.38
3.81
3.15
Black Earth Creek
August 1982
P. zosterI form Is
A. canadensls
Ranunculus sp.
P. pectlnatus
L. minor
Sclrpus sp.
P. crlspus
Percent
Occurrence
63.4
17.2
13.4
5.2
3.7
1.5
0.7
Relative
Abundance
2.86
2.70
1.83
1.57
2.80
1.00
1.00
- 82 -
-------
Bark Rlver-Walschlaeger Road
August 1982
C. demensum
P. zosterIHormIs
P. pectfrtshis
S. eurycerpun
L. minor
Sagltterla -spu
Nymphaee sp.
Percent
Occurrence
45.7
39.4
26.8
7.1
6.3
3.9
3.1
Relative
Abundance
3.21
>l.90
1.29
1.67
I.SB
2.00
1.50
Cedar Creek
September 1962
P. pactl narhus
Percent
Occurrence
92-2
Relative
Abundance
2.56
Fox River
August 1982
Percent
Relative
Abundance
H. dubla
Potamoycitun sp.
V. onterlcana
L. minor
Myrlophy-l I urn sp.
Zlzanla aquatlca
C. demersum
63.4
53.6
26.1
5.2
2.6
1.3
0.7
2.92
1.99
2.63
1.25
2.25
1.00
1.00
September 1982
P. pectlnetus
C. demersum
L. minor
S. rlglda
A. canadensls
P. zosterI form Is
•HI luaukee -R1*er-Campbel I sport
Percent
Occurrence
87.8
5.2
4.3
1.7
0.9
0.9
Relative
Abundance
4.52
3.00
2.60
1.00
1.00
1.00
- 83 -
-------
August 1982
Milwaukee River-East Branch
S. rlglda
P. pectlnatus
Potamogeton sp.
H. dubla
S. eurycarpum
Iris sp.
C. demersum
Percent
Occurrence
45.8
30.1
25.9
4.2
2.4
1.2
0.6
Relative
Abundance
2.90
1.54
1.79
1.29
2.25
1.50
1.00
September 1982
Mount Vernon Creek
Ranunculus sp.
A. canadensls
Hyper Icum elllptlcum forma aquatlcum
Zannlchellla palustrls
Percent
Occurrence
55.5
24.5
17.3
16.4
Relative
Abundance
3.75
3.48
1.68
2.72
Mukwonago River
August 1982
Najas flexllls
P. pectlnatus
V. amerlcana
MyrlophyIlum sp.
Char a
H. dubla
Potamogeton sp.
A. canadensls
Sclrpus sp.
P. zosterform Is
Percent
Occurrence
49.0
48.5
40.2
24.5
22.5
11.8
8.3
2.9
2.0
1.0
Relative
Abundance
1.35
1.15
1.07
1.14
1.09
1.04
1.35
1.00
1.50
1.00
- 84 -
-------
Pewaukee River
August 1982
P. pectlnatus
L. minor
Myrlophyllure sp.
P. crlspus
C. demersum
Percent
Occurrence
93.2
26.5
5.4
3.4
0.7
Relative
Abundance
3.95
1.36
2.38
3.60
1.00
Scuppernong River
September 1982
S. eurycorpum
A. canadensls
S. rlglda
Potamogeton sp.
Z. aquatlca
P. amp 11 foil us
Percent
Occurrence
33.6
32.7
27.4
24.8
5.3
5.3
Relative
Abundance
2.40
3.70
1.84
2.57
1.67
1.00
- 85 -
-------
APPENDIX 3
METHODS FOR EVALUATING MACROPHYTE POPULATIONS IN SMALL STREAM SYSTEMS -
APPLICATION OF PHARTS METHODS TO ROUTINE WATER QUALITY INVESTIGATIONS.
INTRODUCTION
The PHARTS data analyses have Indicated that maximum stable summer macrophyte blomass can be
accurately predicted In Southeastern Wisconsin streams. Two linear regression models were
developed which predict summer macrophyte blomass'. These models are based on macrophyte percent
coverage estimates and mean summer phosphorus concentrations within streams. Further
development and testing of these models for use In other parts of the state requires additional
data over a wider range of stream types and conditions.
Collecting data to use In the model Involves a limited amount of field work. To use both
models, stream reach macrophyte blomass, percent of the streambed with macrophyte cover and mean
growing season In-stream phosphorus and nitrogen concentration data need to be collected.
Mapping provides the percent coverage estimates for the reach and harvesting provides
quantitative plant blomass data to compare with the percent macrophyte coverage and water
chemistry data. Substrate class and distribution within the stream reach Is used to evaluate
macrophyte substrate preference and their potential to supply macrophyte nutrients. This work
Is conducted over a relatively short stream reach.
There are, then, two separate elements of macrophyte assessment; mapping and harvesting. The
harvesting element Is designed to provjde corroborative data for model development and
refinement. The mapping element provides the data to use In the model. It Is probable that
mapping will be the only element routinely conducted.
Criteria for selecting stream reaches as well as sample collection requirements and the
macrophyte mapping and harvesting methodologies are discussed.
SITE SELECTION
The following criteria should be followed In selection of a stream reach which will provide the
best obtainable data:
- Reasonably uniform distribution of macrophytes and substrate type within the reach;
- Maximum reach depth of 2-3 ft. The stream must be workable with waders. Greater depth wll I
also Increase the potential for light-limited growth which will obscure any
macrophyte/nutrIent reI atIonshIps;
- Annual mean flow of less than 60 cfs;
- Maximum stream top width of 60-70 ft;
- Stream should be relatively unshaded and free of constructions (e.g. trees, boulders, pools,
logs);
- Stream reach length should be a minimum of 300 ft and maximum of 2,000 ft.
- 86 -
-------
METHODS
Stream macrophyte community assessment Involves conducting a three-part survey; water chemistry
collections, macrophyte mapping, and macrophyte harvesting. Mapping and harvesting are
conducted as close to the time of maximum blomass as possible. Typically, this Is In late
August or early September In the SE District. Equipment required to conduct the surveys
Includes; wading rod; flow meter, tag-llne(s), Surber bottom sampler, survey flags or pins,
mapping and harvesting field data forms (attached), plastic bags, random number tables, and
plastic wash tubs with 1/4 Inch mesh bottoms.
Stream mapping procedures are similar to those used In collecting top-width and stream
cross-section data for wasteload allocation surveys. The primary difference Is In more accurate
descriptions of substrate types and macrophyte percent coverage. A sample reach Is selected
which Is representative of the portion of stream to be characterized. Transects within the
reach are established, equidistant If possible, for mapping and plant harvesting.
The following summarize mapping and harvesting procedures.
Macrophyte Mapping Procedures
Macrophyte and stream channel mapping Is conducted at each reach transect prior to selecting
harvest quadrats and plant removal. Collection of these data provide percent mecrophyte cover
and substrate composition of the stream channel. Procedures are similar to stream cross-section
or flow measurements (without velocity) and Involve the following:
I) Ten to fifteen transects within the reach should be selected with a minimum of 30 feet
between each transect. Distance between transects should be uniform, measured parallel to the
thalweg. Transects are placed perpendicular to the direction of stream flow. Transects should
not be located near major obstructions In the stream (e.g. trees, boulders, deep pools and logs).
2) Ten to twenty observation points should be taken along each transect with a maximum of three
feet between each point. The observation points are one square-foot quadrats. Transect widths
Include open-water areas and do not Include zones of emergent bank vegetation (e.g. cattails,
bulrush, burreed).
3) The observations recorded at each quadrat (observation point) are; the distance of that
point from the left streambank, depth at that point, estimates of the percent composition of
each bottom type (substrate type), percent of quadrat covered by macrophyte and percent of each
species present. It Is convenient to "Imagine" a one square-foot area around the observation
point to estimate the percent macrophyte cover (or percent open area) and substrate types. It's
also convenient and quicker to use macrophyte species codes rather than writing the full species
name, and the number rating corresponding to a given percent coverage (Attachment I). Data
forms are provided for recording this Information. These data will provide the estimated
macrophyte percent coverage values to be used In the model.
Macrophyte Harvesting Procedures
Macrophyte harvesting Is accomplished by re-establishing or using the original mapping
transects. Transects should be marked when mapped so that the harvesting will be conducted at
the same location as the mapping'. It Is usually more efficient to have two tapes and two crews,
one mapping and the other harvesting. This Is not always necessary. A minimum of 40-50 samples
should be collected per reach, with a minimum of 3-4 samples per transect.
- 87 -
-------
I) A random number toble Is used to select the sample quadrats within each zone or along the
transect1. The sample quadrats should be selected no closer than three feet from the tag-line to
avoid areas disturbed by mapping activities. The maximum distance from the tape should be 20
feet or 1/2 the distance to the next transect, whichever Is the least.
2) A 2-4 digit random number Is used to pick each sample quadrat with the transect. The first
one or two digits (depending on total transect width) Is the distance from the left stream
bank. The third and/or fourth dlglt(s) Is the distance up- or downstrean of the transect. If
the digit Is even, the quadrat Is placed that distance upstream of the transect. If the digit
Is odd. the quadrat Is placed downstream of the transect. Different random numbers or a
different column of numbers are used for each transect. If macrophyte growth occurs In distinct
zones within the stream channel, sample locations should be weighted by the size and occurrence
of the zones.
3) A Surber sampler (one square-foot) Is placed on the stream bed with the random number
coordinate at Its center. Percent macrophyte coverage of plants rooted In the quadrat, species
abundance, depth and water velocity should be recorded at each quadrat prior to plant
harvesting. All plants originating within the frame are harvested with the roots. A small hand
garden cultivator works best to get the roots. The harvested plants are dumped Into plastic
washtubs (with the screen mesh bottom) and thoroughly washed with stream water. The sample Is
sorted In the field to remove stones, sticks, fish and Invertebrates. The sorted and rinsed
sample Is placed In a plastic bag, labeled with the transect and sample number. Samples are
transported to the lab on Ice and refrigerated (do not freeze) until processed.
4) In the lab, samples are separated by species, placed In numbered, pre-welghed 20t paper bags
and dried at 60 degrees C to constant weight In a forced-air oven. Drying approximately 5-10
grams (dry) of plant material (about one handful) per bag should take 24 hrs. Larger portions
In each bag will lengthen the drying time.
Water Samp 11ng
Mean growing season phosphorus and nitrogen concentrations provide the best data for predicting
summer macrophyte blomass. The object of the collection Is to characterize the nutrient
concentrations occurring over the greater part of the season. Ideally, grab samples should be
collected every three weeks from mid-May through the end of August. Grab samples should also be
collected at near-normal flow (not necessarily 07-10 or 07-2). This best represents the
conditions that plants have experienced during the growing season. For this reason, high-flow
samples. If collected, are generally not Included In calculation of the mean growing season
stream chemistries.
- 88 -
-------
APPENDIX 4
DIEL DATA COLLECTION AND ANALYSIS METHODS
A variety of methods have been developed to approximate the dlel fluctuations of dissolved oxy-
gen In streams, lakes and rivers. These have generally been developed In response to a specific
need, and the usefulness and applicability of each of these methods Is a direct function of the
needs of the Investigator. In the present study, two specific goals required consideration.
First, an attempt to determine stream photosynthesis and respiration rates, and relate these to
measured plant blomess quantities, and second, to determine whether or not plant blornass could
be expected to significantly Impact stream dissolved oxygen (DO) concentrations, especially the
minimum concentration.
Since the double station method provides estimates of photosynthesis, respiration and reaeratlon
for a defined area (the area between stations), It was thought to be the best tool for obtaining
photosynthesis and respiration estimates, which could then be compared with areaI blcmass
estimates from harvest data.
Assumptions necessary to the modelling process restrict the length of the reach which Is
mode I able through the double station technique. Travel times of one half to two hours gave good
results where K2 ranged from 1.0-6.0/day. Natural variations In the model parameters tends to
hinder attempts to correlate double and single station estimates of photosynthesis, respiration
and reaeratlon.
The single station method determines coefficients which are "upstream averages" for an area
which Is determined by the "average upstream" reaeratlon rate (K2). The larger the value of K2,
the smaller the area represented by single station analysis. The coefficients which result from
single station analysis should reproduce the curve observed for the site from which the co-
efficients were derived. The single station method may be preferable where the goal Is pre-
diction of dissolved oxygen at a given time and place.
Two different equations were used In the present study. The differential equation (I) was
originally proposed by Odum (1956). The Integrated equation (II) was advanced In part by Blaln
and McDonnell (1967) and Independently derived In a form that Included photosynthesis by WDNR
staff. Either equation Is suitable for single or double station analysis.
I. 4DO/At =«*P + R + K2 (Cs-Co)
II. (Cs-Co)t + 4t = (Cs-Co)te~K2*t + l(«*p + R)/K2HI - e~K2At)
The major difference between the single and double station methods lies In their determination
of the A D0/4t term of the differential equation, or the (Cs-Co) terms of the Integrated
equation. In the single station method, the data Is taken from a single dlel curve. ADO/£t
Is the slope of the dlel curve at a given time. The (Cs-Co) terms are the deficit at the
specified times ("t" or "t + At").
The double station technique requires that two dlel curves be obtained, one for an upstream
station and one for a downstream station. It also requires that the time of travel between
stations Is known. In the double station method, A DO/At Is the difference between the
upstream DO concentration at time "t" and the downstream concentration and time "t + time of
travel (TOT)1', divided by the time of travel.
- 89 -
-------
Values for light and temperature are necessary for both equations and both methods. Either can
be approximated, but It Is best If both are measured. The temperature data Is used to calculate
what the "saturation concentration" of DO Is. Several equations are available for this pur-
pose. One which Is commonly used Is as follows (J. Sanlt. Eng. Dlv., Am. See. Civ. Eng., I960):
Cs = 14.652 - .4I022(°C) + .00799K°C)2 - 0.000077774(°C)3
The concentration which results Is that which would occur If the atmospheric pressure Is 760
tnmHg. Variations In atmospheric pressure are assumed to cause only negligible variations In the
saturation concentration.
Light should be measured In a way that closely approximates the light levels which the plants
experience. The best technique appears to be measuring light 'below the water surface, at a
depth that Is similar to the mean depth of the stream. Although this Is not an exact measure-
ment. It does eliminate problems associated with surface reflectance, partially compensates for
attenuation of light with depth and shading due to bank vegetation or the horizon. Where plants
have grown enough to reach the surface, this method probably underestimates the amount of light
the plants actually receive. It should be realized that the main purpose of this measurement Is
to provide a means by which photosynthesis can be proportioned.
An example Is presented below to help clarify these statements, and show how these measurements
are used to calculate double station P, R, and K2 coefficients. In the example, DO and tempera-
ture were continuously recorded at the upstream station, 00 was continuously recorded at the
downstream station, and light was continuously recorded at a point near the downstream station.
A dye study revealed that the time of travel between stations was 18 minutes. The upstream DO
and temperature data were read off the strlpchart at one hour intervals. Then the downstream DO
data were read off the downstream strlpchart at times which corresponded to "upstream times +
18 minutes". Temperature was assumed to be constant within the reach. Light for each sample
period (8:00-8:18, 9:00-9:18, etc) was also read from a strip chart. Upstream DO readings could
be taken at more frequent Intervals If more sample points were desired.
Upstream Station
Downstream Station
TIME
0400
0500
0600
0700
0800
0900
1000
1 100
1200
1300
1400
DO
(mg/l)
5.56
5.58
5.57
5.59
5.80
6.12
6.32
6.70
6.82
7.03
7.04
TEMPERATURE
CO
22.6
22.4
22.2
22.1
22.1
22.1
22.4
22.8
23.1
23.5
23.8
LIGHT
0/sq cm/sec x I015)
00.0
00.0
00.0
02.0
13.0
11.0
20.0
42.0
45.0
38.0
19.0
TIME
0418
0518
0618
0718
0818
0918
1018
1118
1218
1318
1418
DO (mg/l)
5.70
5.70
5.70
5.85
6.10
6.55
6.78
7.20
7.40
7.55
7.40
Some preliminary calculations are necessary before the data Is In a form that Is amenable to
analysis. Specifically, we need to determine A DO/At and (Cs-Co) for each sample Interval.
For the double station method, 4 DO/dt Is simply the difference between the upstream and
downstream values. (If the single station method was used, a rough approximation of 4 D0/4t
- 90 -
-------
could be obtained by subtracting consecutive readings In the upstream or downstream columns.
The t value would then be the time Interval between readings.) For the period 0400-0418, the
double station ADO/At Is 5.70 - 5.56 = 0.14 mg/l. The time component Is not Included at this
point. The next step Is to calculate (Cs-Co). Cs Is calculated from the temperature data and
the equation presented earlier. At 22.6°C. Cs would be 8.31 mg/l. To determine the "average
deficit" for the period, we need to average the upstream and downstream DO values before
determining the deficit. For the 0400-0418 Interval, the average Co value Is 5.63 mg/l, and the
average deficit (Cs-Co) Is 8.31-5.63 or 2.68 mg/l. The light value for the period Is 0.0. If
this process Is repeated for each sample Interval, we end up with three columns: A DO/dt,
(Cs-Co) and light «*).
If A DO/At Is specified as the Independent variable, and (Cs-Co) and light (•<) are specified
as dependent variables, a multiple regression may be performed on the data set. The
coefficients which are returned from the regression analysis are "P" (the coefficient of the
light «X) term), K2 (the coefficient of the (Cs-Co) term) and "R" (the Intercept, or
constant). Since no time correction was made to the A DO/ At term, all coefficients are In
units of "per time of travel". To convert the R and K2 terms to units of "per day", simply
multiply by the number of time of travel units which occur In a twenty four hour period. (For
the 18 minute time of travel specified, the coefficients would be multiplied by 80.0.) The
photosynthesis term Is In units of •Sng/l/unlt llght/TOT". As a raw coefficient It specifies the
number of mg/l that would be produced at a light Intensity of 1.0 for the specified time of
travel. To convert this to a per day rate, the coefficient should be multiplied by the total
quantity of light received In a day, and divided by the quantity of light received during the
time of travel at a light Intensity of "1.0". The total quantity of light received In a day can
be determined graphically (the area under a light curve), or through mathematical subroutines.
An example of converting the raw coefficient to a value proportionate to light follows.
P = X (mg/l)/(!.0 x I015 quanta/sq cm/sec)/18 minutes
= X (mg/l)/|080 x I015 quanta/sq cm
Where "X" Is the coefficient of the light variable (from the regression analysis)
If this value Is multiplied by the total quantity of llght/sq cm received In a day, the product
would represent the amount of oxygen produced In a day. (An average July day would be about 3 x
I021 quanta/sq cm/day at the water surface.) If other light units are used, the appropriate
alterations must be made to these calculations.
To convert from "mg/l/day" to mg/sq meter/day estimates, the mg/l/day estimates were multiplied
by the average number of liters per square meter (304.8 IIters/sq meter/I foot depth x average
reach depth In feet).
Single station analysis could proceed In a similar manner, except that the ADO/Al term would
be taken from a single curve. The ^t would be the time between DO readings on a single curve.
Calculation of per day and per unit area coefficients could proceed In a similar manner.
If the Integrated equation Is used, the Initial and final deficits must be used In a nonlinear
regression routine. (Blaln and McDonnell (1967) used the night-time data to calculate R and K2,
and then calculated P from the daytime data.] P, R and K2 values can be similarly determined
from the nonlinear regression coefficients.
If recorded data Is not available, a method of approximating Intermediate data points may be
necessary. "Approximation and Interpolation" subroutines are available through the Madison
- 91 -
-------
Academic Computing Center (MACC). The- particular routines used In this study were cubic poly-
nomial spline Interpolations (AISPIN/AkSPEV).
Regression routines usually provide an estimate of the "goodness of fit" of the data to the pro-
posed equation and resultant coefficients (e.g. r-squared, F-tests, t-ratlos of coefficients,
etc.) The behavior of the data and the assumptions Inherent In the model should be examined
before the results are accepted. An example of a situation where statistical fit does not
properly Indicate erroneous results Is presented In the "Differential vs Average Differential
Coefficient" discussion.
Violations of the assumptions of the model are common (see "Deviations from Assumptions" sec-
tion). The-effect of these violations may or may not be severe, but they should be examined-on
a case by case basis, .and dealt with If necessary. Plots of &DO/4t (corrected for reaeratlon
by subtracting K2(£s-Co» versus light are useful In confirming the linearity of the
I lght:photosynthesls relationship. Similarly, plots of A DO/A t (corrected for reaeratlon,
again) during the night time hours can be used to confirm the constancy of respiration. These
two simple checks can add to the Information to be gained from die) curve analysis.
- 92 -
------- |