MODELING PHOSPHORUS LOADING AND LAKE RESPONSE
UNDER UNCERTAINTY
A MANUAL AND COMPILATION OF EXPORT COEFFICIENTS
By
Kenneth H» Reckhow
Michael N. Beaulac
Jonathan T. Simpson
Department of Resource Development
Michigan State University
East Lansing, Michigan 48824
June, 1980
Project Manager.
Robert J. Johnson, Chief
CLEAN LAKES SECTION
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C 20460
-------
REVIEW NOTICE
This report has been reviewed By the Office of Welter
Regulations and Standards, EPA, and approved for publication.
Approval does not signrfy that the contents necessarily
reflect the vtews and policies of the Environmental Protection
Agency* nor does mention of trade names or commercial products
constitute endorsement or recommendations for use,
-------
ABSTRACT
A procedure is proposed that may be used to quantify the relationship
between land use and lake trophic quality. This methodology, based on an
input-output phosphorus lake model, is presented in a step-by-step manner
and illustrated through example An important part of this procedure is a
section describing the estimation of nonparametric prediction intervals.
These intervals quantify the total prediction uncertainty which is a measure
of information value contained in a prediction.
When the methodology is employed to predict the impact of projected
land use changes, it is necessary to use phosphorus export coefficients
extrapolated from other points in time and/or space. These coefficients re-
present the mass loading of phosphorus to a surface water body per year per
unit of source (e.g., per hectare of forested land) A substantial portion
of this document is devoted to a presentation of carefully screened nutrient
export coefficients. These values are intended for inclusion in the modeling/
uncertainty analysis methodology To that end, criteria are described that
will aid the analyst in the selection of appropriate export coefficients and
in the interpretation of the results of an application of this methodology
m
-------
CONTENTS
Page
ABSTRACT ni
FIGURES v
TABLES vn
ACKNOWLEDGMENTS vm
CHAPTERS
1. Introduction . . 1
2. The Phosphorus Lake Modeling and Uncertainty
Analysis Methodology . 8
2.1 Introduction .... 8
2.2 The Phosphorus Lake Model 8
2.3 The Modeling/Uncertainty Analysis Procedure ... . 12
2.4 Application Summary . 36
3. Nutrient Export Coefficients 37
3.1 Introduction 37
3.2 Criteria Employed in the Selection of Export
Coefficients for this Manual . 38
3.3 Export Coefficient Selection Criteria for the Modeler . 40
3.4 The Phosphorus and Nitrogen Export Coefficients . . 58
4. Concluding Comments 107
REFERENCES 114
APPENDIX
A.I Research Methodology for the Assessment of Nutrient
Runoff in Field Studies .... 137
A.1.1 Watershed Designs 137
A.1.2 Sampling Design and Flux Estimation . ... . 143
A.1.3 Standardization of Results Reported in the
Literature . . 146
A.2 Issues Important in the Determination of Phosphorus
Loading to Lakes . . 147
A.2.1 Phosphorus Fractions and Availability. ..... 147
A.2.2 Variability, Precision, and Accuracy . . 153
A.2.3 Temporal Extent of Sampling .157
A.3 Prediction Uncertainty Estimation for Area! Water
Loading (qs) Error 158
IV
-------
FIGURES
Number Page
1 A Schematic Illustrating Phosphorus Loading
Determinants and Lake Response 4
2 The Basic Configuration of a Box Plot and Comparison of
Two Plots Possessing Significantly Different Medians. . 62
3a Box Plots of Phosphorus Export Coefficients from
Various Land Uses 64
3b Box Plots of Nitrogen Export Coefficients from
Various Land Uses 65
4a Histogram of Phosphorus Export from Forested Watersheds .... 91
4b Histogram of Nitrogen Export from Forested Watersheds ... 92
5a Histogram of Phosphorus Export from Row Crops . . .... 93
5b Histogram of Nitrogen Export from Row Crops 94
6a Histogram of Phosphorus Export from Non Row Crops 95
6b Histogram of Nitrogen Export from Non Row Crops 96
7a Histogram of Phosphorus Export from Grazed and
Pastured Watersheds 97
7b Histogram of Nitrogen Export from Grazed and
Pastured Watersheds .... 98
8a Histogram of Phosphorus Export from Animal Feedlots
and Manure Storage 99
8b Histogram of Nitrogen Export from Animal Feedlots
and Manure Storage 100
9a Histogram of Phosphorus Export from Mixed
Agricultural Watersheds . . . . . 101
9b Histogram of Nitrogen Export from Mixed
Agricultural Watersheds . . 102
lOa Histogram of Phosphorus Export from Urban Land Use 103
lOb Histogram of Nitrogen Export from Urban Land Use 104
11 Total Phosphorus Export from Two Corn Cropped Watersheds
Illustrating Variability Over Time. . 105
12 Surface Water Runoff from Two Corn Cropped Watersheds
Illustrating Variability Over Time 106
-------
APPENDIX FIGURES
Number Page
Al Dissolved and Particulate Nutrient Response to the
Storm Hydrograph 152
A2 Hydrographic Response of Varying Land Uses to a
Storm Event ' 155
VI
-------
TABLES
Number Page
1 Proposed Relationships Among Phosphorus Concentration,
Trophic State and Lake Use 11
2 Minimum and Maximum Values for the Data Set Used to
Develop the Phosphorus Model 16
3 Higgins Lake Data Necessary for the Estimation of q 16
4 Land Use Areas in the Higgins Lake Watershed 16
5 High, Most Likely, and Low Export Coefficients Selected
for Higgins Lake 22
6 Nutrient Export from Forested Watersheds 67
7 Nutrient Export from Row Crops 70
8 Nutrient Export from Non Row Crops 73
9 Nutrient Export from Grazed and Pastured Watersheds 75
10 Nutrient Export from Animal Feedlots and Manure Storage .... 77
11 Nutrient Export from Mixed Agricultural Watersheds. . . . . .79
12 Nutrient Export from Urban Watersheds 83
13a Forest Atmospheric Inputs 86
13b Agricultural-Rural Atmospheric Inputs 87
13c Urban-Industrial Atmospheric Inputs 88
14 Nutrient Loads for Household Wastewater discharged
into Septic Tanks 89
15 Magnitude and Variability for Phosphorus Loading from
Wastewater Treatment Plant Effluent 90
VI1
-------
APPENDIX TABLES
Number Page
Ala Phosphorus Export from Forested Watersheds 162
Alb Nitrogen Export From Forested Watersheds 167
A2a Phosphorus Export from Row Crops 172
A2b Nitrogen Export from Row Crops .178
A3a Phosphorus Export from Non Row Crops . ... 183
A3b Nitrogen Export from Non Rwo Crops 186
A4a Phosphorus Export from Grazed and Pastured Watersheds 189
A4b Nitrogen Export from Grazed and Pastured Watersheds ... .192
A5a Phosphorus Export from Animal Feedlots and Manure Storage . . 194
A5b Nitrogen Export from Animal Feedlots and Manure Storage . . .196
A6a Phosphorus Export from Mixed Agricultural Watersheds 198
A6b Nitrogen Export from Mixed Agricultural Watersheds 203
A7a Phosphorus Export from Urban Watersheds 208
A7b Nitrogen Export from Urban Watersheds 212
vm
-------
ACKNOWLEDGMENTS
A number of people assisted in the development and preparation of this
document. Appreciation is extended to Robert J. Johnson, Chief of the Clean
Lakes Section, for providing the opportunity to prepare this report. Thanks
are also due to Janine Niemer, Patsy Johnston, and Barb Visser for typing
the document, to Paul Schneider for graphics work, and to David Lee, Ralph
Ancil, and Robert Montgomery for editorial assistance and proofreading
IX
-------
-------
Chapter 1
INTRODUCTION
The planner or engineer concerned with the issue of lake trophic
quality management must be conversant across several disciplines. Eutro-
phication is fundamentally a problem in chemical and biological limnology,
so the analyst must be familiar with the natural sciences Methods and
processes for managing watershed characteristics and human activities that
determine water quality have an engineering basis with strong economic and
sociologic features. Finally, techniques that project the link between
the engineered management strategies and the limnological water quality are
in the planning arena yet require a good foundation in statistics and mathe-
matics. It is this latter set of planning methods, with a firm basis in
the statistical understanding and description of empirical relationships,
that is the focus of this manual.
Lake eutrophication is both a natural and culturally-induced phenomenon.
Natural eutrophication is a slow, largely irreversible process associated
with the gradual accumulation of organic matter and sediments in lake
basins Cultural eutrophication is an often rapid, possibly reversible
process of nutrient enrichment and high biomass production stimulated by
cultural activities causing nutrient transport to lakes. Eutrophication is
a complex process, and hence the reference is made above to the importance
of chemical and limnological knowledge for proper lake management. Each
lake is unique, and a study and understanding of the unique features are
(
essential for good planning Yet there are also characteristics of water-
shed and lake behavior that are, if not universal, certainly shared by many
lake systems. The planner can exploit this commonality in management
1
-------
studies, providing he/she is aware of where the commonality ends and the
uniqueness begins. In fact, to the degree that planning studies must
surely depend upon the efficient use of resources, all planning is probably
a compromise between unique and common features.
This manual is based on the aforementioned premise of similar trophic
behavior among lakes. This is both a strength and a weakness. Its strength
lies in the fact that the methodologies are not necessarily lake-specific,
so that models and data are transferrable, keeping analysis costs low. Its
weakness is that the ease of application of the methodology and statistics
can foster inappropriate use of the techniques described herein and incom-
plete study of the unique features of a lake. This lowers planning costs
but increases risks associated with poor planning decisions. Again, the
lesson is to know the limitations of the general methodology.
The techniques presented in this manual for lake trophic management
planning are based on control of the nutrient phosphorus. There are two
reasons for this reliance on phosphorus:
1. Phosphorus is often the major nutrient in shortest supply
relative to the nutritional needs of algae and aquatic
plants. This means that the concentration of phosphorus
is frequently a prime determinant of the total biomass
in a lake.
2. Of the major nutrients, phosphorus is the most effec-
tively controlled using existing engineering technology
and land use management
In general terms, with phosphorus as the controlling mechanism, the reasons
suggest that proper management of human activities in the watershed can
often be effective in the maintenance of desirable biomass levels in the
lake.
-------
Figure 1 presents a schematic diagram of the watershed and lake eco-
system as viewed from the perspective of phosphorus movement. Human activ-
ities (including land use), watershed characteristics, and climate are the
general determinants of phosphorus mass transport to lakes. The phosphorus
loading to a lake is empirically related to the phosphorus concentration in
a lake as a function of the hydro!ogic and geomorphologic characteristics
of the lake. Phosphorus concentration, in turn, is causally linked with
biomass levels, water clarity, dissolved oxygen concentrations, and fish
populations, which are all empirically interrelated.
The methodology presented herein is based on the phosphorus flow sche-
matic in Figure 1. Historically, it is derived from the work of Vollen-
weider (1968, 1975) on the phosphorus loading concept. Vollenweider's
contribution to this field was the recognition of the simill an ties among
lakes in trophic response to nutrient input. He defined nutrient loading
criteria for lakes as a function of selected hydrologic and geomorphologic
characteristics (e.g., mean depth or area! water loading). Vollenweider
and others (Dillon and Rigler, 1975, Larsen and Mercier, 1976, Chapra, 1977,
Walker, 1977, and Reckhow, 1979c) modified Vollenweider's initial approach
and empirically derived simple input-output models for phosphorus One of
these cross-sectional empirical phosphorus lake models (Reckhow, 1979d) is
incorporated in the procedure described in the next chapter.
The empirical phosphorus lake model mathematically describes the sec-
tions of Figure 1 from phosphorus loading to lake phosphorus concentration.
The appropriate method for determination of the phosphorus loading depends
in part upon whether the application is primarily descriptive or predictive.
Descriptive use of the model implies an assessment of current lake and
watershed conditions. Direct measurement of the phosphorus loading would
therefore be possible. Alternatively, predictive use of the model suggests
the estimation of the impact of projected land use on lake water quality.
3
-------
CLIMATE
WATERSHED
CHARACTERISTICS
HUMAN
ACTIVITIES
LAKE
MORPHOMETRY
PHOSPHORUS.
INPUT
PHOSPHORUS
-> CONCENTRATION
IN LAKE
HYDROLOGY
ALGAE AND
AQUATIC PLANTS
DISSOLVED
OXYGEN
WATER
CLARITY
FISH
POPULATIONS
Figure I. A Schematic Illustrating Phosphorus Loading Determinants and Lake Response
-------
Here direct measurement is obviously impossible, so phosphorus loading in-
formation must be extrapolated from other, similar watersheds It Is for
the predictive function of empirical phosphorus lake models that this manual
is designed, although the methodology is equally applicable to descriptive
applications (with perhaps an increase in planning risks, see Reckhow and
Chapra, 1980, Chapter 1).
A significant portion of this manual is devoted to a discussion and
presentation of the phosphorus export coefficients that are used to estimate
phosphorus loading for the predictive mode of the modeling methodology
The phosphorus export coefficients presented are taken largely from a com-
prehensive study (Beaulac, 1980) of the literature on the phosphorus mass
transported to surface water bodies from various land uses. An earlier
survey of this topic (Uttormark et a!., 1974) has been frequently cited in
lake nutrient loading studies. However, the paucity of well-designed phos-
phorus export studies pre-dating Uttormark's thorough compilation resulted
in the unavoidable inclusion of inaccurate export values in Uttormark's
work. Cognizant of this problem, Beaulac screened the now considerable
literature on phosphorus export to water bodies for accurate and reliable
sampling designs. The result is a set of phosphorus export coefficients
that are generally representative of the watershed conditions described.
They are presented in Chapter 3 and in the Appendix with watershed descrip-
tions (location, precipitation, soil type, fertilizer application, etc.)
designed to facilitate selection of the appropriate export coefficients for
lake phosphorus loading.
The concern expressed above for representative export coefficients is
grounded in the emphasis in this manual *on uncertainty. For planning to be
effective, it must be based upon reliable information. Predictive models
provide information for planners, so it is vital to the planning process
-------
that the reliability of this predictive information be estimated Without
a reliability measure, the planner has no basis for weighing model predic-
tive information against other planning information. Inefficient and un-
popular decisions can be the result.
Fortunately, it is possible to incorporate an uncertainty analysis into
the modeling methodology. This is presented in a step-by-step manner, along
with the phosphorus lake model, in Chapter 2. The end product is an esti-
mate of total prediction uncertainty, which should be extremely useful to
a planner as a measure of the value of the information provided by the model.
It must be underscored that the estimation of uncertainty does not
obviate the need for consideration of the limitations of the modeling/error
analysis methodology and for care in the selection of the phosphorus export
coefficients. Geomorphologic and climatic constraints on the phosphorus
lake model are mentioned in Chapter 2 These limitations are associated
with the general rule that empirical models are developed for only a sub-
population of lakes represented by the model development data set. Applica-
tion of the methodology to lakes not belonging to this subpopulation can
increase uncertainty and prediction bias. Since there is no mechanism for
the inclusion of this additional error term in the existing methodology,
hidden planning risks may result.
The description of the modeling/uncertainty analysis methodology in
Chapter 2 offers guidance on the selection of phosphorus export coefficients.
Failure to adhere to the criteria for export coefficient choice or failure
to carefully match the application lake watershed characteristics with the
candidate export coefficient watershed characteristics (described in the
tables in Chapter 3 and the Appendix) can again lead to hidden error and bias.
This, too, may increase planning risks
-------
This manual is organized around the modeling/error analysis methodology
in Chapter 2. The phosphorus export coefficients are presented in Chapter 3
and in the Appendix. This is accompanied by criteria used by Beaulac (1980)
in the selection of published coefficients and criteria to be employed by
users of this methodology when the export coefficients are selected and
applied. At the end, some concluding thoughts are offered in Chapter 4 on
the use of this manual for lake trophic management planning.
-------
Chapter 2
THE PHOSPHORUS LAKE MODELING AND
UNCERTAINTY ANALYSIS METHODOLOGY
2. 1 Introduction
The lake modeling/uncertainty analysis procedure presented below is a
variation of the methodology developed by Reckhow and Simpson (1980) Be-
cause of the emphasis attached to uncertainty, the analyst is urged to
follow the directions carefully. Further, it is suggested that the analyst
read Chapter 1 (and possibly Chapter 4) before beginning the modeling, and
read Chapter 4 before preparing a final report documenting the methodology
and the application. These chapters offer valuable guidance on modeling,
error analysis, and the interpretation of results for lake quality management
planning.
2. 2 The Phosphorus Lake Model
In order to plan for the management of phosphorus in a lake watershed,
mathematical models describing phosphorus loading and lake trophic response
can be quite useful. Given the state of knowledge regarding phosphorus
cycles, and the limited funds available to most planning agencies, often the
most practical mathematical model for phosphorus management is the simple
input/output or "black box" empirical model (Reckhow and Chapra, 1980) This
type of model contains terms for the input, the output, and the settling (to
the lake bottom) of phosphorus, but it does not explicitly include any bio-
logical or chemical reactions. The term "black box" is employed because the model
treats the lake like a magician's black box, one is aware only of objects that
enter and exit the box, as the contents and internal processes remain a mystery
8
-------
The left side of Figure 1 (Chapter 1) is a schematic of a "black box"
phosphorus model and represents the conceptual foundation for the mathematical
model used in this procedure. The figure shows that phosphorus input (load-
ing) to a lake is a result of climate, watershed characteristics, and human
activities. This input is modified by environmental factors and yields an
output: the lake's average phosphorus concentration.
The mathematical model proposed herein was developed by Reckhow (1979d)
from 47 north temperate lakes included in the Environmental Protection Agency's
NationalEutrophication Survey This model expresses phosphorus concentra-
2
tion (P, in mg/1) as a function of phosphorus loading (L, in g/m -yr), area!
water loading (q , in m/yr), and apparent phosphorus settling velocity (v ,
' 5 j
in m/yr) in the form:
Using least squares regression, it was found that the apparent settling velo-
city could be fit using a weak function of qg. This resulted in the fitted
model•
11.6 + 1.2qc
S
A few limitations on the use of this model should be mentioned now.
Since the model was constructed only from lakes within the north temperate
climatic zone, it should be applied only to lakes within this zone. Further-
more, the model should not be applied to a lake with variable values more
than the maximum values, or less than the minimum values, specified in Table
2. This is because an empirical model should not be used on lakes different
from those used to develop the model, without prior testing. Finally, the
-------
model may be used to predict the average phosphorus concentration throughout
a lake during the growing season. It cannot be used, as developed, to pre-
dict nearshore or short-term concentrations.
An important yet often overlooked aspect of the application of models
is the fact that the model itself is a simplification of the real world,4 and
thus the prediction from a model is inherently uncertain. Therefore, quanti-
fication of the prediction uncertainty should be a required step when a
mathematical model is applied. This estimate of the prediction uncertainty
could then be used by a modeler or planner as a weight indicating the value
of the information contained in the prediction.
Uncertainty, or error, in this modeling exercise may arise from three
primary sources: the model, the model parameters, and the model variables
Errors in the model and model parameters are derived from the procedure
(regression analysis) used to empirically fit the model. For the phosphorus
lake model (Equation 2), the model error (smloq) for the log transformed
model is .128; parameter error was found to be quite small for most applica-
tions so it was ignored. Model variables uncertainty is estimated for the
application lake, and Steps 2 and 4 in the procedure presented herein illu-
strate the necessary calculations. The separate error terms are combined
for an estimate of total prediction uncertainty in Step 4
Once phosphorus concentration is predicted through the application of
the empirical model, it is useful to interpret this prediction in the context
of expected water quality characteristics in the lake of interest. One ex-
ample of a trophic state ranking scheme or index was proposed by Chapra and
Reckhow (1979) based on average phosphorus concentration. In an earlier work,
Dillon and Rigler (1975) also devised a trophic classification scheme which
related general water quality and lake use features to the traditional trophic
states. These two indices are combined in Table 1 to link phosphorus concen-
tration to potential lake use.
10
-------
Table 1: Proposed relationships among phosphorus concentration, trophic state, and
lake use for north temperate lakes. (Adapted from Chapra and Reckhow, 1979;
Dillon and Rigler, 1975).
Phosphorus Concentration (mg/1) Trophic State
< 0 010 Oligotrophic
0.010 - 0 020
0.020 - 0.050
> 0.050
Mesotrophic
Eutrophic
Hypereutrophic
Lake Use
Suitable for water based recreation and pro-
pagation of cold water fisheries, such as trout
Very high clarity and aesthetically pleasing
Suitable for water-based recreation but often not
for cold water fisheries. Clarity less than
oligotrophic lake.
Reduction in aesthetic properties diminishes
enjoyment from body contact recreation.
Generally very productive for warm water fisheries.
A typical "old aged" lake in advanced succession.
Some fisheries, but high levels of sedimentation
and algae or macrophyte growth may be diminishing
open water surface area.
-------
2. 3 The Modeling/Uncertainty Analysis Procedure
The method proposed herein has as its basis a procedure developed by
Dillon and Rigler (1975) that may be used to calculate the capacity of a
lake for development based upon the relationship between phosphorus input
and water quality. The procedure presented below has two major improvements
over that of Dillon and Rigler. First, the most important improvement is
the addition of an error estimation procedure. This permits the quantifica-
tion of prediction uncertainty, and it indicates to the user how valuable
(certain) the information is that is provided by the model. Second, the
phosphorus lake model imbedded in this procedure has a wider range of appli-
cability than does the Dillon-Rigler model. Dillon and Rigler derived their
model from a highly homogeneous set of lakes, the model development data set
for the model presented herein includes a fairly wide range of lake types
(see Table 2).
In order to facilitate understanding of the impact assessment procedure,
it is presented in conjunction with an application to Higgins Lake in Michi-
gan. It should not be inferred from this that the procedure, as presented,
is applicable only to Higgins Lake On the contrary, the procedure is quite
general and can be easily applied to most lakes (subject to north temperate
location and data set constraints such as those in Table 2) simply by sub-
stituting the appropriate lake data for the Higgins Lake data in the example.
Higgins Lake, located in the northern, lower peninsula of Michigan, is
a deep, cool, oligotrophic lake with a well-oxygenated hypolimnion. The
lake has a maximum depth of 41 meters and a mean depth of 15 meters. Some
agricultural activity occurs in the watershed but most of the area is for-
ested.
For the sake of this example, assume that an estimate of average lake
phosphorus concentration in Higgins Lake, along with an assessment of water
12
-------
quality and recreation potential, is needed for a planning study A measure
of prediction uncertainty is also needed to evaluate the results and to com-
pare them with alternative studies.
The method presented below will be used to solve this problem. Recall
that the model is
P =
11.6 + 1.2qs
This analysis is structured so that the variables are estimated in the
following order: Step 1) areal water loading (qs)» Step 2) area! phosphorus
loading (L), Step 3) lake phosphorus concentration (P), and Step 4) phosphorus
prediction uncertainty (ST).
Step 1: Estimation of q (areal water loading)
The estimation of q involves the solution of two equations;
(A. x r) + (A x Pr) (3)
d o
y-v
(4)
where:
q = Areal water loading (m/yr)
Q = Inflow water volume to lake (m3/yr)\
2
A, = Watershed area (land surface) (m )
2
AQ = Lake surface area (m )
r = Total annual unit runoff (m/yr)
Pr = Mean annual net precipitation (m/yr)
13
-------
Ideally, Q should be determined from direct measurement of inflow or
outflow, since use of any equation like Equation 3 will result in uncertainty
in the predicted variable. When data for Q are not available, it becomes
necessary to estimate A., A , r, and Pr and substitute them into Equation 3
to find Q.
Step 1A: Estimation of A, (area of the watershed)
The highest points of lake and the lake outlet bound the watershed.
In many situtations, all the precipitation that falls on the watershed, and
is not evapotranspired, runs off or becomes groundwater and eventually reaches
the lake. A topographical map enables one to locate the highest points of
land surrounding a lake. Topographical maps are printed by the United States
Geological Survey and must be ordered by quadrangle number or name at a
U.S.G.S. office. The highest points of land may be outlined and A, calcu-
lated by planimetry. Equation 3 req
may require adjustment of the units.
2
lated by planimetry. Equation 3 requires that A, be expressed as m which
Step IB: Estimation of r (annual unit runoff)
Average annual area! runoff has been mapped for many regions and again,
the U.S.G.S.. is a valuable source of information. Note that r must be ex-
3 2
pressed in m/yr (m /m /yr) , and note that it does not include ground water.
When A. and r are multiplied together an estimate of the average inflow
of water from surface runoff is obtained.
Step 1C: Estimation of AQ (area of lake)
The estimation of lake area requires the use of a good map or area! photo-
graph of a known scale. The most accurate method for calculating this area
2
is by planimetry. Note that A must also be expressed in m .
14
-------
Step ID Estimation of Pr (precipitation)
An estimation of the average annual net precipitation (taking into
account losses by evaporation) is also needed for Equation 3. This informa
tion can be obtained from the U.S.G S or the U.S. Weather Service. Note
that Pr must be expressed in m/yr.
The statistics required to solve Equation 3 and 4 for the Higgins Lake
example are presented in Table 3 Therefore, the necessary variables are
i) Total annual inflow volume of water to Higgins Lake*
Q = (Ad x r) + (A x Pr)
= 30.863 106m3/yr
11) The area! water loading-
= 0.804 m/yr
Step 2: Estimation of L (areal phosphorus loading)
Every watershed has a unique pattern of land use within its boundaries
and each use makes a unique contribution, by way of diffuse sources, to the
phosphorus loading of a lake. Technical, financial and practical constraints
prohibit most water quality endeavors from conducting "in situ" studies.
Therefore, many quantitative investigations rely on the application of phos-
phorus export coefficients derived from other studies. A compiled survey of
coefficients screened according to acceptable criteria (see Chapter 3 and the
Appendix) is located in Tables 6 through 12.
15
-------
Table 2: Minimum and maximum value for the data set used to develop the
phosphorus model (from Reckhow, 1979d)
Variable Minimum Maximum
p
L
q
.004 mg/1
2
.07 g/m -yr
0.75 m/vr
.135 mg/1
2
31 4 g/m -yr
187. m/yr
Table 3: Higgins Lake data necessary for the estimation of qt
Variable Estimate
Ad = Watershed area 87 41 106m2
r = Total annual unit runoff 0.2415 m/yr
/- f
AQ = Lake surface area 38.4 10 m'
Pr = Mean annual net precipitation .254 m/yr
Table 4: Land use areas in the Higgins Lake Watershed (Liebeskind et al., 1978)
ft 7
Land Use Area (hectares) Area (10m)
Agriculture 16 0.16
Forest 8347 83 47
Urban 378 3 78
16
-------
In practical applications it is recommended that high, most likely, and
low export coefficients be selected for some of the phosphorus source cate-
gories. This allows the calculation of high, most likely, and low total
loading estimates, which ultimately represent the uncertainty that the analyst
has in his/her estimates of phosphorus loading. The high and low loading
estimates represent the additional phosphorus loading error that must be
added to the model error for the calculation of total prediction uncertainty.
It is important that the high and low loadings represent only those char-
acteristics described in the steps below. This is because to a great extent,
the error in the phosphorus loading estimates is already contained in the
model error. Additional loading error for an application lake must be in-
cluded only when the loading is estimated (using the procedure herein) in a
different (and less precise) manner than it was estimated for the model
development data set. These differences are described in the following steps.
The selection of appropriate phosphorus export coefficients is a diffi-
cult task. Since a critical aspect of this modeling exercise is the estima-
tion of prediction errors, the analyst should realize that poor choice of
export values contributes to an increase in error. This contribution may be
explicit or implicit in the analysis, depending upon whether or not the
analyst is aware of all of the uncertainty introduced by his/her choice of
phosphorus export coefficients. Clearly, experience in the application of
this modeling approach is a valuable attribute.
The estimates of phosphorus loading error are based on high, most likely,
and low phosphorus export coefficients selected by the analyst. Loading un-
certainty may be caused by either variability or bias. Variability may re-
sult from natural fluctuations inherent in a characteristic (e.g., natural
variations in streamflow or stream phosphorus concentration), or from uncer-
tainty inherent in a statistic summarizing a set of data. Bias may result
17
-------
from a number of causes, all associated with the fact that the estimate may
not be representative of the characteristic that it was selected to estimate.
For example, some uncertainty due to possible bias is appropriate to situa-
tions where phosphorus export coefficients generated in one watershed are
applied in another watershed. As the analyst becomes more uneasy about a
selected export coefficient, he/she should express this uneasiness through
increased uncertainty and a greater range between high and low export co-
efficients.
At different points in the procedure presented below, the analyst is
alerted to possible sources of bias. These result from the difference be-
tween conditions in the model development data set lakes and application
lakes. When the characteristics of these two lake groups differ substantially,
it is possible that a procedure appropriate for analysis of one group is
inappropriate for analysis of the other This is the justification for Table
3, which presents the limitations on the basic model variables, as defined
by the model development data set. Now, when the allocation of phosphorus
loading sources differs between the two lake groups, there is probably no
need to restrict the use of the model, which is insensitive to the source of
phosphorus. However, the error analysis (but not the mean prediction) may
be affected because the loading estimation errors vary from source to source
Therefore, warnings of possible bias are stated herein when the application
lake phosphorus loading allocation differs substantially from that for the
model development data set. These warnings should be addressed, when appro-
priate, by the inclusion of a bias uncertainty addition to the high and/or
low loading estimates.
The total annual mass flow of phosphorus to a lake is estimated by
summing the annual phosphorus contribution from each of the nonpoint sources
18
-------
plus any additional point source input within the watershed. Total mass
loading (M) may be expressed as (in kg/yr)•
M = (Ecf x Areaf) + (Ec&g x Areaag) + (Ecu x Areau) + (Eca x AQ) +
(Ecst x # of capita-years x (1 - S R.)) + PSI (5)
where*
Ecf = Export coefficient for forest land (kg/ha/yr)
EC _ = Export coefficient for agricultural land (kg/ha/yr)
ay
EC = Export coefficient for urban area (kg/ha/yr)
EC., = Export coefficient for atmospheric input (kg/ha/yr)
a
EC . = Export coefficient to septic tank systems impacting
the lake (kg/(capita - yr) - yr)
Area,: = Area of forest land (ha)
Area _ = Area of agricultural land (ha)
ag
Area = Area of urban land (ha)
AQ = Area of lake (ha)
# of
capita- = # of capita-years in watershed serviced by septic
years tank/tile field systems impacting the lake
S R. = Soil retention coefficient (dimensionless)
PSI = Point source input (kg/year)
In order to facilitate the understanding and estimation of the variables
contained in Equation 5, Step 2 has been broken into 7 sub-steps.
°Step 2A: Estimation of Area,:, Area,.,, and Area (watershed areas).
T ag u
Recall that the watershed area was determined in Step 1. This area
must now be subdivided into agricultural, forest, and urban lands. The area
for each must be determined and expressed in hectares. Table 4 identifies
19
-------
the existing land uses and areas in the Higgins Lake watershed. When the
objective of this analysis is the projection of future conditions, high and
low area estimates are needed to reflect the uncertainty in the land use
projections.
Step 2B: Estimation of Ecf, EC . EC , EC, and EC (export coefficients)
T ag u a so
This substep requires that an export coefficient (Ec ) be chosen for
A
each of the phosphorus source categories found in the watershed. Candidate
export coefficients with a respective summary of watershed characteristics
(e.g., soil type, % impervious surface, etc ) are found in Tables 6 through 12
Values for atmospheric loading are located in Table 13. These coefficients
represent the expected annual phosphorus input to a lake or stream per unit
of source.
It is important to understand that EC . differs from the other export
coefficients in that it represents the expected annual amount of phosphorus
transported not to the lake, but from households to on-site septic systems,
A range of export coefficients that describes per capita export of phosphorus
from households to septic systems is presented in Table 14.
After the analyst estimates the amount of phosphorus received by septic
systems the next logical step is to determine how much of that phosphorus
is being retained in the tile field soils (i.e., how much is exported to the
lake). Phosphorus retention is addressed in substep 2C.
The high and low export coefficients selected should reflect the modeler's
confidence in the extrapolation of literature export values to the applica-
tion lake watershed. For example, in cases where the modeler knows that the
"most likely" export coefficient chosen was determined under a good sampling
program on a watershed quite similar to the application lake watershed, the
20
-------
high and low values should be selected to represent little uncertainty. A
single precipitation loading figure (the "most likely" value) is probably
adequate for most conditions. Only when the precipitation loading is deemed
substantial (perhaps 25% of the total loading) should it be necessary to
include possible precipitation loading error bias.
When the objective is future projection, the area estimates and export
coefficients should be combined according to* high area with high export,
most likely area with most likely export, and low area with low export This
calculation is to be made in Step 2F.
For the Higgins Lake example, various documents were consulted, includ-
ing a study conducted by the U.S. Environmental Protection Agency (USEPA,
1975), in order to familiarize the investigators with the watershed. Phos-
phorus flux reports from other area lakes were also surveyed. Despite the
existence of pertinent literature, the selection of application phosphorus
export coefficients is still an unavoidably subjective task. This, of
course, is the nature of the technique and consequently affirms the impor-
tance of che associated uncertainty analysis.
The forestland within the Higgins Lake watershed consists primarily of
coniferous species with some deciduous trees and constitutes the major land
use. Agriculture is rather limited and consists chiefly of grazing and
pasture. The urban areas are mainly residential/recreational and all units
are serviced by septic systems.
For demonstration purposes, high, most likely, and low export coeffi-
cients were chosen (based on the phosphorus export coefficients presented
in Chapter 3) to reflect phosphorus source conditions found in the Higgins
Lake watershed. The selected coefficients are presented in Table 5 . Note
that the high and low values selected for Higgins Lake are not as high or
21
-------
Table 5: High, most likely, and low export coefficients selected for Higgins lake
ro
ro
Source
Forest
Agriculture
(pasture and grazing land)
Urban
(residential)
Precipitation
High
.30 kg/ha/yr
1.30 kg/ha/yr
2.70 kg/ha/yr
.50 kg/ha/yr
Most Likely
.20 kg/ha/yr
.40 kg/ha/yr
.90 kg/ha/yr
30 kg/ha/yr
Low
.10 kg/ha/yr
.20 kg/ha/yr
.35 kg/ha/yr
.15 kg/ha/yr
Input to
septic tanks
1.0 kg/capita/yr 0.6 kg/capita/yr 0.3 kg/capita/yr
-------
as low as some of the candidate export coefficients presented in Chapter 3
This is because conditions in the Higgins Lake watershed were judged to be
not equivalent to the extreme conditions that are represented by the ranges
in candidate coefficients.
The EC . export coefficients were selected to take into account Michi-
gan's ban on the sale of phosphorus-based detergents. Thus, the selected
coefficients are on the lower side of the range exhibited in Table 14
Likewise, note that EC, is also on the lower side of the presented atmos-
a,
pheric export coefficient range This is because little agricultural and
industrial activity take place in the Higgins Lake area, which probably
results in small quantities of air-born phosphorus.
Step 2C- Estimation of S.R (soil retention coefficient)
On-site septic tank-tile field systems may or may not be effective in
trapping phosphorus and preventing it from entering a lake via groundwater
transport. The soil retention coefficient is an estimate of how well the
systems immobilize phosphorus. This coefficient may range from 0 to 1 0.
For example, if it is assumed that all phosphorus transported to septic
systems eventually reaches the lake, then a soil retention coefficient value
of 0 would be selected. If it is assumed that no phosphorus reaches the
lake, then S R =10.
Rodiek (1979) notes that effective tile drainage fields involve both
physical and chemical processes. Chemical fixation reactions require
effluent-to-soil contact of sufficient time length for chemical reactions
or adsorption to occur. There are four major aspects of watershed soils
(within the lake impacting zone) that influence contact duration time and
phosphorus immobilizing capabilities a-nd thus should be considered when
selecting S.R. These factors are. 1) phosphorus adsorption capacity,
23
-------
2) natural drainage, 3) permeability and, 4) slope. As one might expect,
all of these factors are closely related and of a dynamic nature. They are
discussed in depth in Chapter 3.
In addition to the above soil characteristics, there are four general
mechanisms of phosphate removal in soils. 1) rapid removal or adsorption,
2) slow mineralization and insolubilization, 3) plant uptake, and 4) bio-
logical immobilization (Tofflemire and Chen, 1977). The most important of
the phosphorus immobilization mechanisms in septic systems are the formation
of insoluble iron and aluminum phosphate compounds and the adsordtion of
phosphate ions onto clay lattice structures (Tilstra et al., 1972)
Assessment of the factors discussed above is useful in determining
S.R. However, because of the complexities involved, the modeler's estima-
tion of S.R. still must be based on his/her knowledge of the soil condi-
tions present in the application watershed, past experience with similar
watersheds and his/her professional intuition When the model is used to
predict future conditions, it is often sufficient to use a single estimated
soil retention coefficient;. Only when the estimated loading from septic
systems is thought to be substantial (perhaps 25% of the total loading),
should it be necessary to employ low, most likely, and high soil retention
coefficients. It is possible, however, that the error analysis may be biased
when the septic tank loading becomes a sizeable fraction of the total loading.
For the Higgins Lake example, it was found that sandy/gravel soils of
moraines and till plains predominate in the watershed, which tend to permit
rapid infiltration and transmission of water. Nearby Houghton Lake is sur-
rounded by various soils posessing moderate to poor phosphorus adsorbing
capacities (Ellis and Childs, 1973). Based on this evidence, soil retention
of phosphorus was estimated to be on the poor side A "most likely" S R.
coefficient of .25 a "low" coefficient of .50 and a "high" coefficient of
24
-------
.05 were selected to represent the soils surrounding Higgins Lake Since
evidence (USEPA, 1975) suggests that phosphorus loading from septic systems
may be a substantial fraction of the total loading, three (not one) soil
retention coefficients were chosen for Higgins Lake, as specified in the
instructions above.
Step 2D. Estimation of # of capita-years
The number of persons contributing to septic systems that impact a
lake must be estimated and expressed in capita-years. To ascertain this
figure, the analyst must first determine the size of the impact zone. Ofcen
this is a strip, perhaps 20-200 meters wide, surrounding the lake Sometimes
the analyst may include border strips along tributary streams when condi-
tions suggest that these remote areas may be important. Conditions that
dictate the size and location of the impact zone include drainage patterns,
water tables, and slopes
When the model is used to assess current conditions, population sur-
veys are quite useful for the estimation of the phosphorus loading from
septic tanks. When the goal is the prediction of future conditions, popula-
i
tion projections must be consulted For most lakes, the high and low loading
estimates for septic systems should then be based solely on the uncertainty
in the population projections (the source of possible bias). The total num-
ber of capita-years may be calculated by adding together permanent resident
capita-years and seasonal resident capita-years This is described in Equa-
tion 6 below.
25
-------
Permanent capita-year
Total # of _ average # of persons # days spent # of
capita-years ~ per living unit at unit per year living units
365
+ (6)
Seasonal capita-year
average # of persons # days spent # of
per living unit at unit per year living units
365
In this particular example, it was assumed that septic systems of only
lakeside dwellings impact Higgins Lake. According to the EPA National Eutro-
phication Survey (USEPA, 1975) there are an estimated 1,000 seasonal dwell-
ings on the Higgins Lake shoreline and all are served by septic systems A
facilities plan study estimated that each seasonal unit is occupied by an
average of 3.5 people who spend 60 days a year at their residence (Progres-
sive Engineering Consultants, 1976). This information may be inserted in
Equation 6 to estimate the number of capita-years impacting the lake.
Permanent Seasonal
Total # of = Q + o r x _60 x ,nno
capita-years ° + 3 5 x 365 x 100°
= 575.3
Step 2E: Estimation of PSI (point source input)
If the effluent from an industry, sewage treatment facility or other
point source is deposited within the watershed, the impact must be assessed
and expressed in kg phosphorus/yr.
At the present time there are no known point sources of phosphorus in
the Higgins Lake watershed. Thus, PSI = 0 kg/yr. However, if point sources
26
-------
exist In the application lake watershed, or are projected for the future, then
the uncertainty in the phosphorus loading from this source is represented by
the uncertainty in the size of the projected population to be served An ex-
ample of phosphorus loads from sewage treatment plants is presented in Table
15.
As a final note on the phosphorus loading sources, it is possible that~
the lake sediments may be a non-negligible source. Internal phosphorus load-
ing is most probable in shallow lakes that possess anoxic bottom waters. This
condition can promote an appreciable rate of phosphorus transport from the
sediment/water interface to overlying waters. In shallow lakes this sedi-
ment phosphorus may reach the photic zone and be used by the aquatic plants.
If this is thought to be so for an application lake, then high, most likely,
and low loading estimates should be used for the prediction and prediction
error (see Reckhow, 1979b for suggested values), to reflect this source of
possible bias in the uncertainty analysis.
Step 2F Calculation of M (total phosphorus mass loading)
When Steps 2A through 2E are complete, Equation 5 may be solved to yield
high, most likely, and low phosphorus mass loading estimates based on high,
most likely, and low phosphorus export and soil retention coefficients.
Thus, for the Higgins Lake example
27
-------
high
M(high) = (-30 x 834?) + 0-30 x 16) + (2.7 x 378) + (.50 x 3840) +
(1.0 x 575.3 x (1 - 0.05)) + 0
= 6012.04 kg/yr
most likely
M(ml) = (*20 x 8347^ + (*40 x
(0.6 x 575.3 x (1 - 0.25)) + 0
= 3426.9 kg/yr *
low
M(low) = t'1 x 8347^ + ('20 x 16^ + (>35 x 378) + ( 15 x 384°)
(0.3 x 575.3 x (1 - .50)) + 0
* 1632.5 kg/yr
Step 2G: Calculation of L (annual area! phosphorus loading)
In order to be used in this model, annual phosphorus input must be ex
pressed as a loading per unit lake surface area. This is accomplished by
dividing M by the lake surface area, A .
The units are then converted so that this area! phosphorus loading
term is expressed in grams per square meter of lake surface area per year.
A
Thus, for Higgins Lake:
28
-------
high
1 L(high) = """"" "^ = 157 x 10"U k9/mVyr = .157
most likely
, = 3426.9 kg/yr = 8g x 1Q-6 kg/m2/yr = >089 g/m2/yr
(m{) 38.4 x loV
low
5-= 40 X lO"6 '-'-2' rt"" ~'"2
38.4 x 10um'
4 low) = ' " 43 x ID" kg/m/yr - 043 g/m/yr
Uowj
'Step 3. Calculation of P (lake phosphorus concentration)
The model may now be solved for high, most likely, and low phosphorus
concentrations by substituting in values of q and L (high, ml, and low)
P = L (high)
'(high) 11.6 + 1.2qs
(ml)
11 6 + 1.2qs
L (low)
11.6 + 1.2qs
For Higgins Lake.
high
'(high) = 11.6 !'l?2 (0.804) = °'0125 mg/1
most likely
P
(ml) 11.6 + 1.2 (0.804) ~^'
low
p = ±
r(low) 11.6 -i
29
-------
Step 4: Estimation of Sj (prediction uncertainty)
In order to estimate the uncertainty associated with a prediction cal-
culated using the phosphorus model, estimates are needed for the error, or
uncertainty, in all terms in the model, and in the model itself However,
it has been shown by Reckhow (1979d) that for most applications of this
model, the error in the parameter v is small Further, error in q is pri-
^ s
marily a function of flow measurement error and hydro!ogic variability,
which also affect L. Since L and q are in the numerator and denominator,
respectively, in the model, the errors affecting both tend to cancel when
they are combined to yield the resultant error in P. In addition, hydro*
logic variability is unimportant in lakes with low flushing rates. There-
fore, it is assumed here that the prediction error is a function only of
model error and of aspects of phosphorus loading uncertainty that are
identified in Step 2. If the application lake flushes rapidly and is sub-
ject to great variations in year-to-year precipitation, then the modeler is
urged to include hydrologic variation in the error analysis using the error
propagation equation (see the Appendix for instructions)
The model error is represented by s , in the equations below and is
expressed in logarithmic units of phosphorus concentration error. The load-
ing error, s^, on the other hand, is expressed in untransformed units of
phosphorus loading error. Therefore, to combine these two values for an
estimate of total prediction uncertainty, some calculations are necessary.
The procedure presented below is based on first order error analysis
(Benjamin and Cornell, 1970). In this particular application, three assump-
tions are of some importance.
1. Model error, expressed in log-transformed concentration
units, is appropriately combined with variable error terms
after the transformation is removed.
30
-------
2. The "range" ("high" minus "low"), for phosphorus loading
error, is approximately two times the standard deviation.
This is based loosely on the characteristics of the
Chebyshev inequality identified below, where about 90%
of the distribution is contained within ± 2 standard de-
viations of the mean.
3. The individual error components are adequately described
by their variances (standard deviations).
In order to relax a previously imposed (Reckhow, 1979a) yet tenuous nor-
mality assumption, the confidence intervals constructed below are based on
a modification of the Chebyshev inequality (Benjamin and Cornell, 1970). Therefore,
it is no longer required that the total error term be normally distributed
Instead its distribution must only be unimodel and have "high order contact"
with the abscissa in the distribution tails. These are achievable assump-
tions under almost all conditions, and it is recommended (Reckhow and Chapra,
1979) that this type of nonparametric approach be adopted until the distri-
butions have been adequately studied and characterized
Step 4A: Calculation of log P/ ,\
Take the logarithm of the most likely phosphorus concentration, P(m-]\-
For Higgins Lake:
log P(ml) = log 0.0071
= -2.149
31
-------
Step 4B: Estimation of s + ("positive" model error)
The model error, (sm-joq)s was determined to be 0.128. Add s , to
log P/m-i\ and take the antilog of this value. Now calculate the difference
between this antilog value and P(m-M- Label this difference sm+, it re-
presents the "positive" model error.
= antilog [log P(ml) + s] -P (8)
For Higgins Lake:
s+ = antilog (-2.149 + 0.128) - 0.0071
m
= 0.0024 mg/1
Step 4C: Estimation of s - ("negative" model error)
Subtract s , from log P/m-i\ and take the antilog of this value. Now
calculate the difference between this antilog and P(m-M» and label this
difference s -.
-= antilog [log P(ml) - smlog] - P(ml) (9)
For Higgins Lake:
s - = antilog (-2.149 - 0.128) - 0.0068
= 0.0015 mg/1
32
-------
Step 4D Estimation of s,+ ("positive" loading error)
Now, one must convert the loading error estimate into units compatible
with the model error. Use the p(njon) concentration estimated in Step 3 and
calculate the difference between P(ulau) and P(m])» then divide this differ-
ence by 2. Label this value s,+; it represents the "positive" loading error
contribution.
- P(ml)
For Higgins Lake
SL+ = 0.0027 mg/1
Step 4E: Estimation of s,- ("negative" loading error)
Repeat Step 4D substituting the low concentration value P/-, \ for
P(hiah)' Labe^ the resultant value s,-, it represents the "negative" load
ing error contribution.
For Higgins Lake:
SL- = 0.0019 mg/1
Step 4F: Estimation of s^+ (total "positive" uncertainty)
Total positive prediction uncertainty is calculated using the equation:
33
-------
(SL+)2 (12)
or:
V =
For Higglns Lake:
= ^(0.0024)2 + (0.0027)2
ST+ = 0.0036 mg/1
Step 4G: Estimation of ST~ (total "negative" uncertainty)
Total negative prediction uncertainty is calculated using the equation
(sr>2 = (V )
or:
For Higgins Lake:
(0.0019)2
ST- = 0.0024 mg/1
Step 4H: Calculation of confidence limits
The prediction uncertainty may be expressed in terms of "confidence
limits" which represent the prediction plus or minus the prediction uncertainty.
34
-------
Confidence limits have a definite meaning in classical statistical inference,
they define a region in which the true value will lie a pre-specified per-
centage of the time.
Using the modification of the Chebyshev inequality CBenjamin and Cornell,
1970), the confidence limits may be written as:
Prob [(P(lnl) - hsr) s P , (P(ml) + hsT+)] , 1 - -J-y (16)
Equation 16 states that the probability that the true phosphorus concentration
lies within certain bounds* defined by a multiple, h, of the prediction
2
error, is greater than or equal to 1 - 1/2. 25h . (This relationship loses
its significance as h drops much below one.) Substituting values for h into
Equation 16 reveals that a value of one for h corresponds to a probability
of about 55% (.556 to be exact), and a value of two for h corresponds to a
probability of about 90% (.889 to be exact). Thus the 55% confidence limits
are
* -55
Substituting the Higgins Lake data this becomes
Prob [(0.0071 - 0.0024) < P < (0 0071+ 0.0036)] > 55
Prob [0.0047 mg/1 z P <; 0.0107 mg/1] * .55
Now that specific values for the prediction error have been inserted into
the confidence limits expression, its interpretation changes somewhat. It
is- "about 55% of the time (that confidence limits are estimated), one can
35
-------
expect that the actual average phosphorus concentration will lie tTithin the bounds
defined by the prediction plus or minus the prediction uncertainty " This
same interpretation format applies when the confidence limits are widened to
the 90% level (h = 2), and the Higgins Lake data are inserted:
Prob [{P(ml) " 2sT'} - P - (P(ml) + 2sT+)] * '90
Inserting the data yields:
Prob [(0.0071 - (2) (0.0024)) z P z (0.0071 + (2) (0.0036))] <; .90
Prob [0.0023 mg/1 <; P z 0.0143 mg/1] * .90
2. 4 Application Summary
The application of the technique and model to Higgins Lake resulted in
a "most likely" phosphorus concentration of 0.0071 mg/1 (Step 3), with 55%
confidence limits bounding the "true" phosphorus concentration between
0.0047 mg/1 and 0.0107 mg/1.
Relating back to the trophic classification in Table 2, Higgins Lake is
probably:
1) oligotrophic
2) clear, and suitable for water-based recreation and a cold water
fishery.
These predicted trophic conditions in fact describe the present observed
conditions in Higgins Lake. A median phosphorus concentration value of
.006 mg/1 was determined for Higgins Lake by the Environmental Protection
Agency's National Eutrophi cation Survey (USEPA, 1975).
36
-------
Chapter 3
NUTRIENT EXPORT COEFFICIENTS
3.1 Introduction
In Chapter 1, a distinction was made between descriptive and predictive
use of the modeling methodology Planning for proper lake quality manage-
ment necessitates the prediction of the impact of projected land use on lake
water quality. Measurement of a yet-to-be realized impact is clearly im-
possible. Instead, the planner must extrapolate impact assessments from
other, similar watersheds, possibly in the form of annual export coefficients
Thus nutrient loading estimates associated with watershed land uses are neces-
sary for lake trophic quality management planning.
In this chapter, annual nutrient export coefficients, identified in the
comprehensive literature survey by Beaulac (1980), are presented in tabular
and graphical form The criteria employed in the identification of these
"approved" export coefficients are described. To a great extent, these
criteria reflect the importance of good experimental design in the collec-
tion of nutrient flux data for the determination of export coefficients.
Also discussed in some detail are recommended criteria to be considered by
users of this methodology in the selection of export coefficients To
facilitate this selection process, the tabulated nutrient export coefficients
are presented along with data on related characteristics These include
watershed location, precipitation, soil type, and other site-specific fea-
tures that might affect nutrient runoff. The user may then match these
qualities with the characteristics of the application lake watershed so that
reliable export coefficients are chosen
37
-------
3.2 Criteria Employed in the Selection of Export Coefficients for this Manual
In screening nutrient export coefficients reported in the literature,
Beaulac (1980) established certain acceptance criteria. To a considerable
degree, these criteria are reflective of good experimental design employed
in watershed sbudies. So that the analyst may understand the approximate
reliability that can be attached to the tabulated export coefficients (pre-
sented below and in the Appendix), important screening criteria are dis-
cussed. For elaboration of this topic, see the Appendix
1. Accuracy In statistical terms, accuracy exists when the ex-
pected value for an estimator lies close to the true value
Inaccuracy in a study frequently results from faulty experi-
mental design. Therefore, to evaluate accuracy, one must
have a good conceptual knowledge of the characteristic of
concern. This includes an understanding of causal relation-
ships as well as temporal and spatial variability if rele-
vant. For nutrient export coefficients, accuracy is likely
if the researcher a) employed design controls for ex-
traneous variables not of immediate interest, b) incorpor-
ated into the analysis all causal factors not removed from
influence, and c) used good statistical sampling design
(see the Appendix) As an example, if the researcher is
interested in agricultural row crop runoff, he/she must
either exclude all other runoff sources from the test water-
shed, or include consideration of the additional sources
when analyzing the results. If it is known that row crop
runoff is quite dependent upon major storms and that parti-
culate material is important, then the design must reflect
these factors.
38
-------
2. Precision Like accuracy, precision is a statistical term.
An estimate is precise if it is estimated with low error
Good sampling design (see the Appendix) is a necessary but
not sufficient condition for precision Precision is also
affected by the number of samples taken and the amount of
useful information acquired per sample. Therefore, in
screening literature export coefficients, Beaulac (1980)
looked for accurate experimental designs including frequent
sampling (particularly if the observations are dependent).
3. Representativeness For the nutrient export coefficients
to contribute to reliable lake trophic management planning,
the analyst must carefully match the characteristics of
candidate export coefficient watersheds with the applica-
tion lake watershed (see section 3.3). This requires com-
prehensive information on the export coefficient watersheds.
Therefore, Beaulac (1980) looked for information on char-
acteristics such as geographic location, precipitation,
watershed size, soil type, fertilizer application (when
/
appropriate) and other important land use features This
is vital to the watershed matching process which dictates
export coefficient choice.
4. Temporal Extent of Sampling Nutrient budgets and input-
output lake models are generally based on yearly increments
so that the meterologically-induced annual variations in
nutrient export and lake quality are effectively removed
from analysis. Since weather and climate have a similar
effect on export coefficients, only yearly values were
accepted by Beaulac. The alternative, extending values
39
-------
reported for fractions of a year, must be rejected because
of methodological problems
5. Nutrient Flux Estimation There is no "best" method for com-
bining concentration and flow data to estimate mass flux.
However, the short discussion in the Appendix on sampling
design and flux estimation describes preferred techniques.
Only export coefficients estimated under documented methods
deemed relatively unbiased are reported herein.
6. Concentration and Flow Data Since flow is the prime deter-
minant of nutrient mass flux, as a rule only those export
coefficients estimated with continuous flow data are re-
ported. Most lake models and nutrient loading criteria
are based on total nutrient concentrations Therefore, pre-
ferred studies were those that reported total nutrient con-
centrations. Beyond that, since bioavailability (see the
Appendix) is an issue receiving increasing attention,
fractional forms of the nutrients are provided herein when
reported in the original study.
3.3 Export Coefficient Selection Criteria for the Modeler
Probably the most important task that the analyst performs in applying
the methodology in Chapter 2 is selecting the phosphorus export coefficients.
Most of the other steps are quite explicit in the description of the task to be
conducted and in the associated uncertainty (if any). Export coefficient choice,
on the other hand, benefits from the experience of the analyst in the general
topic of land use-nutrient flux relationships. The problem associated with
faulty export coefficient selection is one of supplemental, or hidden, uncer-
tainty. This refers to prediction uncertainty that is unknown to the analyst
40
-------
and thus is not part of the uncertainty accounting process. Hidden uncertainty
results from inexperience, the analyst believes that a choice of high and low
export coefficients covers the true phosphorus loading when in fact it does not.
This leads to bias in the prediction and additional risk in planning. Knowledge
of the causal factors behind nutrient flux from land use activities and thought-
fulness in matching the application watershed with candidate export coefficient
watersheds can significantly reduce supplemental uncertainty.
There are two important issues to consider when selecting export coeffi-
cients. First, one must try to match the application lake watershed and
candidate export coefficient watersheds as closely as possible on the basis
of causal determinants of phosphorus loading. Second, the range of export
coefficients selected (i.e., the high and low values) should reflect the total
uncertainty for that characteristic. Factors important in watershed matching
are outlined in detail after a short discussion of phosphorus loading uncer-
tainty.
Uncertainty in phosphorus export may arise from 1) natural variability,
2) error and bias associated with the measurement and estimation of the export
coefficient, 3) error and bias associated with the representation of phosphorus
export in the application watershed by an export coefficient estimated at
another point in space and/or time, and 4) uncertainty in land use, population,
etc., projections. It was noted in Chapter 2 that the phosphorus model
standard error contains some phosphorus loading estimation error associated
with the model development data set. As a result, the description of export
coefficient selection—for the purpose of loading error estimation—is quite
specific. Thus not all uncertainty components identified above are necessarily
included in each set-of-three (high, most likely, and low) export coefficients
choice.
41
-------
One point should be made concerning the first item listed above, natural
variability. The nutrient export coefficient tables exhibit natural variability
(in addition to measurement and estimation error) among the export coefficients
presented. This is cross-sectional variability which in part represent various,
and different, conditions in the nutrient export coefficient watersheds.
This must be distinguished from natural longitudinal variability, which
reflects variability in export from a single watershed over time (see section
3.4 for histograms exhibiting cross-sectional and longitudinal variability).
It is likely that longitudinal variability is smaller in magnitude than cross-
sectional variability, since the causative factors for longitudinal variability
are relatively homogeneous (in comparison to the causative factors for cross-
sectional export coefficient variability). Since export coefficients are
chosen for single watersheds (over time), it is longitudinal export variability
that is important (in addition to extrapolation error and bias). Unfortunately,
there is little multi-year data on nutrient export in single watersheds, so
when needed, the estimation of longitudinal nutrient export variability is
necessarily subjective.
The second issue mentioned above for consideration in selecting export
coefficients is the process of matching the application lake watershed and
candidate export coefficient watersheds according to causal determinants of
nutrient export. To facilitate this matching process, an outline is presented
below listing important causative nutrient export factors according to land
use activity.
1. Forest Land Use
The range of phosphorus export coefficients is very narrow (.019 - .830
kg/ha/yr), and it is difficult to specify any one factor as the determinant
of loading in a particular watershed. Much of the variation among
42
-------
coefficients is probably within the range of experimental or sampling error.
a. Species Type
i. Pine-coniferous softwoods have demonstrated higher rainfall
interception capacity and evapotranspiration rates than have
hardwoods. A number of investigators report that annual stream-
flow was reduced about 20% below that expected for the hardwood
cover, 15 years after experimental watersheds in the Southern
Appalachians had been converted from a mature deciduous hardwood
cover to white pine (Swank and Douglass, 1977, 1974; Swank et
al., 1972; Swank and Miner, 1968). Therefore higher nutrient
loads could develop from tributaries draining hardwoods than
from tributaries draining softwoods.
11. Some hardwoods such as alder (Alnus sp.) are nitrogen fixers.
Brown et al. (1973) reported both higher nitrate concentrations
and higher nitrogen loads from alder watersheds than from those
streams that drained primarily douglas fir and western hemlock
(for streams in western Oregon).
b. Soil Type, Bedrock, and Parent Material
Dillon and Kirchner (1975) observed that forested watersheds with
sandy soils overlying granitic igneous formation had one-half the
phosphorus output than did forested watersheds with loam soils
overlying sedimentary formations. Loam soils are higher in nutrients
and more erodable than sands and gravels, sedimentary formations have
higher Teachability and erodability. Therefore soils and substrate
types such as these (loams and sedimentary formations) may cause shifts
toward the higher end of the phosphorus export range.
43
-------
c. Vegetation Age
Maturity is a function of species type (among other characteristics).
This factor is important only in very young, newly vegetated forests.
In this sense, "young" refers to trees of less than five years of age.
Woodlots of this age do not have a canopy developed enough to reduce
rainfall impact energy. Therefore soil in young forests are more dis-
rupted than soils in mature forests (see Disturbed Watershed, below).
The result is higher runoff and greater sediment phosphorus flux from
young forests.
d. Climate
This appears to be the major determinant of export of phosphorus from
forests. Areas of the country that exhibit warm climates with high
rainfall (such as the pacific northwest and southeastern piedmont
regions) are associated with high productivity, high runoff, and high
phosphorus export.
e. Disturbed Watersheds
i. Deforestation/timber harvest
Watersheds with ongoing timber harvest tend to have higher nutrient
export than do undisturbed systems. This is because deforestation:
1) blocks the nutrient uptake pathway; 2) raises forest floor
temperature; 3) increases the frequency of drying and wetting
(weathering); 4) increases microbial activity; and 5) increases the
nutrient pool by contribution of dead organic material (slash).
Therefore, nutrient output is increased. (The amount of increase
depends on the extent of the watershed under cultivation.)
n*. Forest fire
Nutrient export due to fires can increase over the normal range
44
-------
of export for undisturbed forests, but this will depend upon the
severity of the burn (% watershed burned) and the type of fire
(crown vs. brush (understory) fire) (Wells, 1971; Pntchett, 1979).
m. Fertilization
Nutrient export will increase only if fertilizers are applied directly
on the stream. This practice is currently not very common nationwide.
Increases in nutrient export will last only for a short time period
(one or two runoff periods), and will depend on the extent of area!
coverage of the fertilizer and fertilizer type (nitrogen or phosphorus)
(.Moore, 1970, 1975; Frednksen et al, 1975; Stay et al., 1978).
2. Agricultural Land Use. Crops
a. Soils
Because soils are exposed for long time periods (late fall, winter,
early spring), they will influence the magnitude of the phosphorus
load released from the watershed.
i. Sandy/gravel soils 1) do not erode easily, 2) have a low cation
content, and 3) cause a general downward flow of water to the
groundwater (high infiltration capacity). Thus phosphorus export
via runoff is low.
11. Clay soils (clay loams, silt loams etc.) have a 1) high cation
content (high phosphorus adsorption capacity), 2) high erodability,
and 3) low infiltration capacity. Therefore phosphorus export via
runoff is high.
11*1. Organic soils have 1) limited phosphorus retention capacity, 2)
low infiltration capacity, and 3) high nutrient content. As this
soil is used for cultivation, it decomposes rapidly. Therefore
phosphorus export via runoff is high.
45
-------
b. Fertilizer Type and Amount
i. The type of fertilizer is not as significant as the time of
application. Partially because of this, manure-type fertilizers
are thought to often cause high phosphorus export because manure
is frequently applied on frozen soils in winter or early spring.
When combined with snowmelt and high rainfall/runoff periods, the
result is very often high' export of phosphorus and nitrogen. If
application is followed with soil incorporation, phosphorus and
nitrogen export is substantially reduced (Minshall et al., 1970,
Klausner et al., 1976, Converse et al., 1976, Hensler et al., 1970).
ii. Heavy amounts of fertilizer (either manure or commercial grade)
applied above the recommended rate will cause increases in nutrient
export. The recommended rate is dependent upon the amount and
availability of nutrients in the soil (to growing crops).
c. Tillage Practices
i. Conventional tillage methods, in which the ground is left fallow
during non-growing periods and crop residues are removed at
harvest, are a prime cause of high amounts of nutrient export
(lead to high erosion of soils, etc.).
ii. Conservation tillage methods ideally have conservation of soil,
water and energy as the primary objective. These methods will
reduce the export of nutrients. Among the conservation tillage
methods are 1) nonmoldboard tillage, such as chisel plowing, that
does not use a moldboard plow and involves fewer tillage oper-
ations than conventional moldboard systems, and 2) "no-till,"
which involves planting directly into untilled soil (Pollard et al.,
1979).
46
-------
ill. Other techniques, which the above tillage methods may be combined
with, can reduce nutrient export further. These methods are 1)
contour planting, and 2) terracing (Alberts et a!., 1978).
d. Crop Types
i. Row crops (corn, soybeans etc.): Farmland planted with this type
of crop is subject to channelization and erosion. Export of
nutrients from watersheds consisting of row crops will be much
higher than export from non row crop watersheds.
n. Non row crops (wheat, millet, rye and other small grains): Growth
of these crops does not generally lead to channelization. Therefore
lower levels of nutrient export may be expected.
3. Agricultural Land Use- Pasture and Grazing Land
Nutrient export from these watersheds depends upon the method of
management of the cattle, sheep, etc. and not necessarily on the volume of
waste produced.
a. Rotational Grazing: Cattle are grazed on a particular piece of land
for a limited time period (e.g., summer only). This allows vegetation
to regrow which reduces runoff (including nutrient export).
b. Continuous Grazing: This results in 1) increases in soil compaction,
2) decreases in vegetation, and 3) increases in waste loads (manure).
Therefore, nutrient export is usually high (Menzel et al., 1978).
c. Fertilization: Pastured watersheds are often fertilized to increase
forage vegetation. This often increases the total amount of nutrient
export (01 ness et al., 1980).
47
-------
d. Animal Density: Studies indicate that the greater the number of animals
per unit area, the higher the amount of animal waste and the greater
the potential for high nutrient export (ChiChester et a!., 1979).
4. Agricultural Land Use: Feedlot and Manure Storage
a. Percent Impervious Surfaces: If the percent of paved surfaces is high,
the infiltration rate will be low and the runoff and nutrient export
will be high (Coote and Hore, 1978).
b. Animal Concentration: If the animal density is high, the nutrient
export can also be high (McCalla et al., 1972; Clarke et al., 1975).
c. Covered Feedlots: If the feedlot is inclosed with a roof, rainfall
impact energy will be reduced, and runoff and nutrient export will be
decreased (the higher the roof area/feedlot area ratio, the lower the
runoff) (Dornbush and Madden, 1973; Coote and Hore, 1978).
d. Detention Basin: If a detention basin is present, nutrient export will
be decreased (Coote and Hore, 1978).
5. Urban Land Use
Most urban runoff is channeled into storm drains, although not all
storm drains serve a single watershed. Therefore, the output from a storm
drain may consist of material from portions of one or more watersheds. It
is important to determine the extent of the drainage system (if storm drains
are used) in order to get an accurate estimate of area! nutrient loading.
Urban land uses consist of a number of sub uses, each with different
features.
48
-------
a. Characteristics of residential areas important to nutrient loading
include: 1) housing density, 2) grass and vegetation coverage; 3)
fertilizer applications, and 4) pet density, type (dogs, cats), etc.
These characteristics are important because grass and housing density
affect the infiltration/runoff ratio, while fertilizers and pets deposit
nutrients in the watershed. Decreases in grass cover and increases in
the other three increase nutrient export.
b. Public parks or park-like settings (campuses, research parks, etc.)
have more vegetation (lawns, trees, ponds, etc.) than do commercial
districts. Therefore, they can produce less runoff and nutrient export
than do commercial districts.
c. Commercial/business/industrial areas have considerable street/pedestrian
traffic. Thus there is more dust suspension, more contaminants from
v auto and industrial emissions, and more imprevious surfaces than in
residential areas. Therefore, higher nutrient storm runoff often results.
6. Atmosphere
Atmospheric inputs consist of two major components: 1) wind transported
material, commonly called dustfall, removed from the air by sedimentation or
impaction; and 2) soluble gases or salts which are scavenged by rainfall.
It is important when determining the magnitude of atmospheric loads to con-
sider both of these components. Estimates for the dryfall portion alone may
be as high as 70 - 90% of the total load (Heany and Sullivan, 1971, Likens
and Loucks, 1978; Swank and Henderson, 1976, Miklas et a!., 1977). In
addition, the size of the dryfall fraction is generally considered to be
independent of the amount of wetfall precipitation (Swank and Henderson, 1976;
Delumyea and Petel, 1977; Eisenreich et al.? 1977),
49
-------
The sources of air-borne nutrients are not necessarily limited by the
actual watershed boundary. That is, a lake's atmospheric input can be a
result of cultural practices occurring in neighboring watersheds. However,
the impact of atmospheric sources diminishes with distance.
a. In agricultural areas, increases in nutrient loads transported via
the atmosphere can be attributed to agrarian activities and associated
soil disturbances. These include:
i. ammonia volatilization from feedlots and fertilizers,
11. wind erosion of fertilized soils.
In addition, peak inputs of nutrients from the atmosphere tend to
occur in late spring and early fall, in a pattern that roughly
corresponds to fertilization and tilling periods (Andren et a!., 1977;
Delumyea and Petel, 1977; Eisenreich et al., 1977; Miklas et al., 1977;
Hoeft et al., 1972).
b. Urban atmospheric inputs of nutrients can also be higher than those
from forests. These increases can be attributed primarily to combustion
emissions, since:
i. Aviation and automotive fuels are known to contain organophosphorus
additives to reduce corrosion (Simpson and Hemens, 1979).,
ii. Fly ash from oil-fired boilers has been estimated to contain 0.9%
phosphorus as P205» and open-hearth furnaces have been found to
contain up to 0.3% phosphorus pentoxide (Delumyea and Pete!, 1977),
iii. Automotive emissions are believed to be the major source of NOV,
A
(Robinson and Robins, 1970), and
iv. Photo-oxidation and hydrolysis reactions in an atmosphere containing
hydrocarbons and oxides of nitrogen apparently are a major source
of nitrites, nitrates and nitric acid in precipitation (Likens, 1972;
Likens et al., 1977).
50
-------
7. Septic Tanks and Soil Adsorption Fields
a. 'On-site septic tank-tile filed systems are another "non-point" source
that must be considered because these systems are not always effective
in trapping nutrients and preventing them from entering a waterway
via groundwater transport. Three major waste fractions typically
compose the septic tank receiving water. These are: 1) garbage
disposal wastes; 2) toilet wastes, referred to collectively as sblack
water; and 3) sink, basin and appliance wastewater collectively referred
to as gray water (Siegnst et al., 1976, Bennett and Linstedt, 1975,
Ligman et al,, 1974; Olsson et al., 1968; Wallman and Cohen, 1974;
Laak, 1975, Siegrist, 1977). Each waste fraction contributes com-
parable amounts of nutrients.
i. Phosphorus in septic tank effluent originates from two main sources,
human excreta and phosphate detergents.
ii. Major sources of nitrogen (up to 80%) are feces and urine, with
the predominant forms occurring as NH^ and organic-N.
b. The mass loading of each nutrient to the septic system may depend on
a number of considerations, and per capita-year loading coefficients
presented in Table 14 should be chosen accordingly. These considerations
include:
i. Fraction of the year that the system is in use and the number
of people using this form of waste disposal (i.e., summer cottage
or year-round dwelling).
li. Amount of detergent used and the detergent phosphorus content.
Phosphorus detergent bans will substantially reduce the total
load since gray water phosphorus loads are high. Sawyer (1965)
estimated that detergent - based phosphorus accounts for approxi-
51
-------
mately 50 - 75% of the total phosphorus in domestic wastewater.
iii. Use of waste flow reduction methods. Siegnst (1977) estimates
that several devices and systems such as low volume flush toilets,
no-water toilets, wastewater recycle for toilet flushing, and
suds-saver clotheswashers should produce waste flow reductions of
up to 35%. This could significantly lower the concentration and/or
pollutant mass in the household wastewater stream.
c. The estimated nutrient loading from septic systems to lakes will depend
upon the location of the system with respect to the surface water body.
The hypothesized "impacting zone" should include those systems within
the watershed that contribute nutrients directly to a lake. For example,
Rodiek (1979) developed a phosphorus budget for Lobdell Lake in Michigan
and chose to use a 100 meter wide impact zone around the lake. Tributaries
to the lake should also be included if certain conditions exist. These
conditions include population distribution and other factors related to
soil retention such as:
i. Phosphorus adsorption capacity: Relative phosphorus adsorption
categories have been proposed by Schneider and Erickson (1972)
and are outlined below:
Rate classes Kilograms of phosphorus per hectare in top .9 meters
Very low Less than 1120 Kg per hectare
Low 1120 to 1460 Kg per hectare
Medium 1460 to 1800 Kg per hectare
High 1800 to 2240 Kg per hectare
Very high Over 2240 Kg per hectare
The percentage of phosphorus adsorbed is highly dependent on soil
type and pH. ToffTenure and Chen (1977) proposed a procedure for
52
-------
the evaluation of soil retention of phosphorus. They examined
several New York soils and found that, as a rule, for phosphorus
removal, acid soils are better than calcareous, tills are better
than outwashes, and clay soils are better than sandy soils.
ii. Soil drainage: Natural soil drainage is generally related to the
depth of the water table. For septic system suitability, it is
essential that a zone of aeration exist between the septic tile
field and the water table at all times of the year. This zone func-
tions as a chemical and physical filter for phosphorus (Ellis and
Childs, 1973). For nutrient retention, well-drained to moderately
well-drained soils are preferable because these conditions tend to
lower the risk of nutrient contamination of groundwater. In other
words, the greater the distance between the septic-tile field and
the water table, the greater the likelihood that phosphorus will be
immobilized and not transported to a surface water body via groundwater.
iii. Soil permeability: Permeability is the rate at which water is trans-
mitted through saturated soil. This transmission rate is generally
a function of soil texture and structure (i.e., the proportion of
sand, silt and clay). High rates of water transmission are usually
indicative of sandy soils, while low rates are usually associated
with clay soils, soils possessing a clay lens, or an impervious layer
at or near the ground surface. Relative classes of soil permeability
can be used to describe conditions of water transmission. Schneider
and Enckson (1972) propose the following classes of soil permeability.
53
-------
Rate classes Permeability rate in centimeters per hour
Very slow < .50
Slow .50 - 2.00
Moderate 2.00 - 6.40
Rapid 6.40 - 25.40
Very rapid > 25.40
Soils having a moderate rate of permeability are optimal in terms
of septic system operation since these rates are slow enough for
phosphorus adsorption reactions to occur, yet fast enough to
avoid system "back-up" that results in standing effluent.
iv. Groundwater movement: In addition to the above three factors,
both direction and flow rate for groundwater must be considered.
The presence of clay lenses or bedrock can substantially reduce
groundwater flows to a lake, and in some situations flow can be
redirected from the lake altogether (to subterranean reservoirs
or to another watershed).
v. Slope: Steep slopes and low permeability rates may cause erosion
problems and perhaps convert the septic tank effluent to overland
runoff. In soils of good drainage and high permeability, gravi-
tational forces hasten groundwater flow (and nutrient transport
to surface water bodies).
vi. System age: Soils have only a finite capacity for phosphorus
adsorption. Old systems may provide less soil retention of phos-
phorus than do new systems.
vii. Plant uptake: The presence of a "green strip" of vegetation con-
sisting of shrubs, bushes, trees, etc., between the septic tank-
54
-------
tile field and the waterbody can effectively reduce the amount of
phosphorus entering a lake. This will depend, in part, on vegetation
type and density.
viii. Season: High rainfall seasons, such as spring or late summer, keep
the soil saturated, thereby decreasing soil phosphorus adsorption
capacity.
ix. Other: Factors such as frequency of cleaning (of both septic tank
and drainfield) and the effluent-soil redox potential should also
be considered.
8. Sewage Treatment Plants
Wastewater treatment plants are another phosphorus source that have been
studied to some degree. As a result, data do exist for the estimation of
phosphorus loads and variability. Among the issues that should be considered
by the analyst attempting to use these loading coefficients (Table 15) are:
a. Type of plant: Plant type, or more appropriately, the type of treatment
the plant is using, will determine to a great extent how much phosphorus
will be contined in the effluent. Different treatment types provide
different levels of phosphorus reduction in the waste stream. Obviously,
those plants using phosphorus removal will have lower per capita phos-
phorus outputs than those that do not.
b. Separate or combined sewerage systems: Wastewater treatment facilities
have a finite capacity to treat sewage inputs. Under normal circumstances,
treatment capacity is closely related to the extent of the sewerage
system. If the system is composed of both storm and sewage drains,
treatment capacity if overtaxed during high rainfall events, and a
55
-------
portion of the combined inputs are short circuited through to the
outfall. Final mass loads of phosphorus can be appreciably higher
when this occurs.
c. Phosphate detergent ban: If phosphorus inputs are reduced, (i.e.,
through a ban on phosphate detergent additives), the final per capita
mass load will also be lower. From a study of 702 wastewater treat-
ment plants with a variety of treatments processes employed, Allum et
al., (1977), estimated a median phosphorus loading of 1.0 ± 0.04 kg/
capita/yr. However, for Indiana, with a full-year phosphate detergent
ban, this median figure was found to be 0.5 ±0.11 kg/capita/yr (25
plants). New York, with about a one-half year phosphate detergent ban,
fell between the two with a median phosphorus discharge of 0.7 ± 0.11
kg/capita/yr (42 plants).
In concluding this section on guidelines for export coefficient selection,
some comments on watershed size, proximity to the application lake, and bio-
availability are in order. It should be noted that small watersheds, such as
microplots (<0.5 hectares), provide less opportunity for redeposition of sus-
pended sediment (and nutrients) than do large watersheds. Even though the
"100-year" storm will scour considerable amounts of deposited nutrients from
streambeds—thus balancing any loading inequalities between large and small
basins—some investigators feel that in the short term, small runoff plots or
small watersheds tend to overestimate the mass of nutrients removed by surface
runoff.
In addition, Schuman et al., (1973) demonstrated that water samples for all
runoff events taken adjacent to the outflow of an agricultural watershed con-
tained considerably more inorganic phosphorus in solution than did samples taken
56
-------
70-230 meters downstream. This reduction in solution phosphorus was attributed
to the adsorption of phosphorus by the additional suspended soil material enter-
ing the stream from gully erosion. This decrease in solution phosphorus in
the runoff was accompanied by an increase in phosphorus on the sediment trans-
ported. Thus total phosphorus loss measured at the two sites agreed relatively
well. Studies by Meyer and Likens (1979) at Bear Brook (an undisturbed head-
water stream in the Hubbard Brook Experimental Forest, New Hampshire) indicate
that there was a net conversion of dissolved phosphorus and course particulate
phosphorus (leaves, organic fragments, etc.) to the fine particulate fraction,
which was the predominant form (62% of the total) exported downstream.
Therefore, small plots (from which many of the export coefficients presented
in Chapter 3 and in the Appendix are based) are likely to yield high export
values for certain situations. These values will consist of both high solution
fractions and high sediment fractions. These fractions will tend to be higher
than those reported for larger watersheds (several hectares in size). Thus,
small watershed export coefficients are most applicable to application lake
watershed sections adjacent to a surface water body (tributary streams or the
lake). This means, of course, that watershed size is an important watershed
matching criterion.
For large basins consisting of mixed agricultural activities, export
coefficients from the tables entitled "Mixed Agriculture" should be used.
Individual assignment of export coefficient according to each use may result
in an overly high total loading estimate due to the small watershed bias
mentioned above.
Bioavilability is another concern. In general, the more solution phosphorus
converted to sediment phosphorus, the lower the bioaviTable fraction. However,
57
-------
since the models are based on total phosphorus, bioavailability cannot be
incorporated into the Chapter 2 methodology. If it is thought that an unusually
large fraction of the phosphorus loading to a lake is not biologically avail-
able, then the analyst should note this and be aware of possible model prediction
bias.
3.4 The Phosphorus and Nitrogen Export Coefficients
1. Summary Tables - Text To facilitate the analyst's ability to use
the model and quantitative approach presented in this manual, nutrient
loading coefficients from overland runoff were identified in an ex-
tensive literature search (Beaulac, 1980). While the emphasis of
this report is on phosphorus management and modeling, for comparison
purposes, export coefficients are also given for nitrogen. Those
studies which conform to the sampling criteria discussed in earlier
sections of this chapter have been aggregated by land use and are
presented in tabular fashion in Tables 6 through 12. The major land
uses examined are undisturbed forests, agriculture and urban.
As previously discussed, the range of nutrient export from forest
land use is relatively narrow. Climate (i.e., precipitation and
runoff) and productivity appear to be the major criteria determining
nutrient export variability. The analyst is therefore urged to extrapo-
late only those coefficients originating from climatic conditions and
regions similar to the application watershed. For comparison, vege-
tation type, soil type, location, precipitation and runoff amount have
been tabulated along with the export coefficient and reference in
Table 6.
58
-------
Agricultural land uses consist of a number of different pertur-
bations, and sufficient studies exist in the literature to describe
these activities. Therefore, this land use was further subdivided
into row crops, non-row crops, pasture/grazing land, and manure storage/
animal feedlot. The loading coefficients are presented in Tables 7
through 10 respectively. For comparison purposes, and to allow for
estimation of nutrient export from highly mixed agricultural watersheds,
export coefficients were compiled for general (mixed) agricultural
activities in Table 11. In addition to the descriptive conditions
listed in the forest export tables, fertilization rate and crop type(s)
have also been included.
The coefficients describing urban land uses also exhibit a high
degree of variability depending primarily on the type of urban activity
(i.e., low density residential, heavy industrial) and the associated
percentage of impervious surface area. Unfortunately, sufficient data
do not currently exist in the literature to adequately compile summary
tables for each of these activities. Therefore, the analyst is urged
to pay particular attention to the accompanying descriptive criteria
listed in Table 12.
2. Summary Tables - Appendix To provide the reader with a more complete
record of the variability and magnitude of the chemical fractions com-
posing both phosphorus and nitrogen (i.e., sediment phosphorus, NOg-N),
a breakdown of these chemical fractions is included in the Appendix.
The tables in the Appendix include all "approved" nutrient runoff
coefficients presented in the text plus some information from studies
which did not focus on total nutrient loads. To reduce repetition,
59
-------
most of the watershed characteristics have been eliminated if the
particular study was adequately described in the text tables.
3. Histograms The effects of watershed characteristics and climatic
conditions on nutrient export can be observed from a study of the
loading coefficients in the above tables. However, this variability
can be more properly assessed through examination of the data in
frequency distributions or histograms. Accordingly, histograms des-
cribing nutrient export from the above land uses have been developed
and are presented in Figures 4 through 10
From the histograms it is apparent that the cross-sectional data
are highly skewed. Consequently, robust statistics such as the median
and interquartile range are generally less biased as summary statistics
than are the mean and standard deviation. These statistics accompany
the histograms for each land use.
The histograms allow the analyst to note the cross-sectional
variability resulting from different characteristics among watersheds
that determine nutrient export. As an example the reader is referred
to Figure 7a, representing phosphorus export from pastured and grazed
watersheds. The values on the left represent phosphorus export from
those watersheds grazed primarily in summer or on a rotational basis,
while those on the right represent export from watersheds with either
continuous grazing or forage fertilization. This cause-effect relation-
ship emphasizes the need for proper examination and selection of the
coefficients for extrapolation purposes.
Cross-sectional variability among watersheds must be distinguished
from longitudinal variability. Longitudinal, or time series, variability
60
-------
represents the variation in nutrient export in a single watershed over
time. To illustrate longitudinal variability, phosphorus exports from
two similar adjacent corn cropped watersheds, one with seven years of
identical fertilization rates and the other with five, were combined
to create the histogram in Figure 11. Since variation in precipitation
runoff is the probable key cause of longitudinal variablity, a histogram
of water runoff rates was also developed and presented in Figure 12
Note the high degree of similarity between the two distributions.
4. Box Plots A useful graphical technique for displaying batches of data
is the box plot. This technique is based on order statistics (ordering
the data points from low to high value) and the plot itself is constructed
from five values from the (ordered) data set. These values are: 1)
the median; 2) the minimum value; 3) the maximum value; 4) the 25 per-
centile value; and 5) the 75 percentile value (see Figure 2).
Visual comparisons of box plots may be enhanced by the incorporation
of the statistical significance of the median into the plot. This is
achieved by notching the box at a desired confidence level. For example,
if the 95% confidence level notches around two medians do not overlap
in the display, the medians are roughly significantly different at the
95% confidence level (see McGill et a!., 1978,and Reckhow, 1980 for
details on confidence limits and other aspects of box-plot construction).
In addition to the above information, the box plots can include
the following (Reckhow, 1980):
1. the interquartile range;
2. the sample range;
3. an indication of skew (from a comparison of the symmetry above
and below the median), and
4. the size of the data set.
61
-------
CO
UJ
i
i .i
<3!
LLJ
i
*
CD
Li_
O
I • t
LLJ
CO
CO
O
ZD
F1
O
^
f lax i mum
value
75%
value
i —
1 LJ
Median \ 7
value / \
/ \
25% 1
value
Minimum _
value
v
\
\
/ ""
/
/
—
—
Statistical
significance
of the median
Inter-
quartile
range
Group A
Group B
Fiqure 2: The Basic Configuration of a Box Plot and Comparison of
Two Plots Possessing Significantly Different Medians
62
-------
Note that the box plot medians for forested phosphorus and nitrogen
export (Figure 3) are significantly different from those of agri-
culture and urban land runoff (with the exception of pasture land).
5« Other Tables In addition to the thorough examinaton of the litera-
ture on nutrient runoff from forest, agriculture, and urban land use
activities, other non-point and point nutrient loading information was
compiled in tabular form for this document.
a. Atmospheric Inputs: Uttormark et a!., 1974, listed at least 40
factors influencing atmospheric nutrient contributions. The bulk
of these factors are related to local conditions. Therefore, a
literature review was conducted to collect data relating bulk
nutrient precipitation inputs to specific land uses. A major re-
quirement for data acceptability was that the nutrient inputs be
collected from one of three land uses: 1) undisturbed-forest;
2) agricultural-rural; and 3) urban-industrial. Studies dealing
with regional or cross-sectional watersheds were thereby disregarded.
In this respect, precipitation chemistry may more closely reflect
endemic situations. These nutrient coefficients are presented in
Table 13.
b. Septic Tank Inputs: Information was collected to define a range of
values for the nutrient load in household wastewater discharged into
septic tanks. Since the values expressed in Table 14 are not
quantified according to the number of sources contributing to the
total load (i.e., percent contributed by gray water vs. black water),
it is recommended that the reader examine the section dealing with
septic tanks in order to justify the selection of the export
coefficient.
63
-------
800-
186
623
5 On
40-
>-
a:
o
~c
a.
CO
o
2.0-
1.0-
0J
CO
LU
CC
O
79520
400-
300-
200-
100-
0J
Figure 3a: Box Plots of Phosphorus Export Coefficients from Various Land Uses.
64
-------
3847
25i
796
20-
cc
>-
-------
c. Sewage Treatment Plant Inputs: The data set compiled by the EPA-
NES (1974) listing statewide sewage treatment plant phosphorus
loads was summarized according to treatment type in Table 15.
As previously discussed in earlier sections, a number of factors
can increase or reduce the coefficients presented. This may be
verified from an examination of the ranges given for each treat-
ment type. It is further stressed that the analyst examine the
actual conditions within the study watershed before the selection
of these coefficients is made.
66
-------
Table 6: Nutrient Export from Forested Watersheds
Land Use
75-100 year old
jack pine & black
spruce, with birch
& trembling aspen
(342 1 ha)
Climax hardwoods
maple, beech, red
oak, with yellow
birch and hemlock
(125 ha)
Jack pine -
black spruce
Jack pine -
black spruce
Mixed deciduous
forest
Mixed deciduous
forest
Nixed deciduous
forest ( 01 ha)
70% aspen
30% black spruce
and alder (10 ha)
Aspen - birch
forest (6 48 ha)
Location
Kenora Experi-
mental Watershed
Rawson Lake
Ontario, Canada
Clear Lake
Watershed
Haliburton
County, Ontario,
Canada
Northwest
Ontario, Canada
Northwest
Ontario, Canada
Southern
Ontario, Canada
Southern
Ontario, Canada
Lake Minnetonka
Watershed,
Minnesota
Marcel 1
Experimental
Forest,
Minnesota
Marcel 1
Experimental
Forest ,
Minnesota
Water Total Nitrogen
Soil Precipitation Runoff Export
Type/Texture cm/yr cm/yr kg/ha/yr
medium-fine 77 3a 26 55a 6 26a
silicate sand (70 1 - 96 7) (22 3 - 35 4) (5 69 - 7 32)
overlying
deeper deposits
containing some
clay fractions
126 3 68 00
sandy loam 2 37
sandy loam 1 38
sandy soils
overlying
granitic
igneous
formation
loam soils
overlying
sedimentary
formation
loam, silt 129 0 84 3
loam, clay
loam
70% loam, clay 17 70e 2 26e
& sands (15 5 - 19 2) (1 74 - 2 37)
30% organic
peats
loam, clay 79 48e 15 56e 2 46e
and sands (/5 51 - 82 10) (13 73 - 21 47) (1 92 - 3 29)
Total Phosphorus
Export
kg/ha/yr
309a
( 220 - 435)
090
060b
036b
047C
( 025 - 077)
107d
( 067 - 145)
090
157e
( 124 - 179)
280e
( 19 - 38)
Reference
Schindler et al ,
1976
Schindler and
Nighswander, 1970
Nicholson, 1977
Nicholson, 1977
Dillon and
Kirchner, 1975
Dillon and
Kirchner, 1975
Singer and Rust,
1975
Verry, 1979
Timmons et al ,
1977
-------
Table 6: (continued)
Land Use
Maple, birch and
beech (15 6 ha)
Deciduous hardwood
and pine (17 6 ha)
Oak-hickory
forest (97 5 ha)
CTi
00 Oak-hi ckory
forest (97 5 ha)
Oak maple yellow
poplar, black
cherry, beech
(34 ha)
Mixed pine and
hardwood (40 ha)
Mixed pine and
hardwood
99% mixed forest
1% developed
(6495 ha)
Loblolly and
slash pine
2 81 ha
1 93 ha
2 39 ha
1 64 ha
1 49 ha
Soil
Location Type/Texture
Watershed #6 sandy loam
Hubbard Brook
Experimental
Forest, New
Hampshire
Coshocton, Ohio silt loam
Walker Branch
Watershed,
Oak Ridge,
Tennessee
Walker Branch
Watershed,
Oak Ridge,
Tennessee
Fenrow Expen- silt loam
mental Forest,
Parsons, West
Virginia
Eatonton,
Georgia
Rhode River
Watershed,
Maryland
Woodlands, clays
Texas
Copperville, loess over
Mississippi sedimentary
deposits
Precipitation
cm/yr
132 2f
88 9e
(85 4 - 95 8)
1369
157 la
(128 2 - 187 5)
164 0
205 0
205 0
205 0
205 0
205 0
Water Total Nitrogen
Runoff Export
cm/yr kg/ha/yr
83 30f 4 01 f
32 00e 2 82e
(25 3 - 35 6) (1 37 - 3 16)
70 709 3 I9
94 65a 2 Og
(71 0 - 116 1) (17-2 2)
48 70
1 50
7 30
36 90
38 95
34 85
30 75
22 55
Total Phosphorus
Export
kg/ha/yr
01 9f
035e
( 0349 - 0722)
025a
( 010 - 030)
140e
( 040 - 180)
0 275
0 200
0 212
0 281
0 306
0 357
0 321
0 226
Reference
Likens et al ,
1977
Taylor et al ,
1971
Henderson and
Harris, 1973
Henderson et al ,
1977
Aubertin and
Patnc, 1974
Krebs and Golley,
1977
Correll et al ,
1977
Bedient et al ,
1978
Duffy et al , 1978
-------
Table 6: (continued)
CT)
Land Use
Douglas fir and
western hemlock
(47139 5 ha)
Douglas fir and
western hemlock
(32376 ha)
Douglas fir and
western hemlock
(10 1 ha)
Douglas fir and
western hemlock
Douglas fir and
western hemlock
Soil Precipitation
Locati on Type/Texture cm/yr
Yakima River,
Western Cascade
Range, Washington
Cedar River,
Western Cascade
Range, Washington
H J Andrews 215 0
Experimental
Forest, Western
Cascade Range,
Oregon
Fox Creek, silt & clay
Western Oregon loams
Coyote Creek, silt & clay
Western Oregon loams
Water Total Nitrogen Total Phosphorus
Runoff Export Export
cm/yr kg/ha/yr kg/ha/yr
3 32 0 830
0 360
135 0 0 520
158 0 0 180
76 0 0 680
Reference
Sylvester, 1960
Sylvester, 1960
Fredriksen, 1972
Fredriksen, 1979
Fredriksen, 1979
a Four year median
b Four year mean from twelve watersheds
c Two year median from twenty watersheds
d Two year median from four watersheds
e Three year median
f Twelve year mean
g Two year mean
-------
Table 7: Nutrient Export from Row Crops
Land Use
Corn ( 004 ha)
Corn, fresh
manure applied
in winter
( 004 ha)
Corn, fermented
manure applied
in spring
( 004 ha)
Corn, liquid
manure applied
in spring
( 004 ha)
Corn ( 004 ha)
Corn, fresh
manure applied
in winter
( 004 ha)
Corn, fermented
manure applied
in spring
( 004 ha)
Corn, liquid
manure applied
in spring
( 004 ha)
Corn ( 009 ha)
Corn ( 009 ha)
Fertilizer
Application
kg/ha/yr Soil
N P K Location Type/Texture
000 Lancaster, silt loam
Wisconsin
109 39 99 Lancaster silt loam
Winconsin
102 44 85 Lancaster, silt loam
Wisconsin
7833114 Lancaster,- silt loam
Wisconsin
000 Wisconsin silt loam
108 39 99 Wisconsin silt loam
108 34 99 Wisconsin silt loam
108 39 99 Wisconsin silt loam
112 29 Morris, loam
Minnesota
29 81 Morris, loam
Minnesota
Water
Precipitation Runoff
cm/yr cm/yr
77 Oa 10
(65 76 - 77 6) (8 51 -
77 Oa 12
(65 76 - 11 6) (5 97 -
77 Oa 11
(65 76 - 77 6) (5 59 -
77 Oa 12
(65 76 - 77 6) (5 61 -
11
(8 71 -
9
(7 11 -
8
(7 11 -
9
(8 10 -
62 6C 8
65 7d 10
7a
21 95)
26a
19 41)
51a
15 32)
45a
15 60)
52b
14 33)
32b
11 53)
81b
10 52)
45b
10 79)
6C
ld
Total Nitrogen
Export
kg/ha/yr
3 96a
(3 61 - 5 53)
*797a
(3 05 - 26 88)
3 38a
(3 35 - 5 32)
2 88a
(2 81 - 5 07)
433b
(4 08 - 4 58)
1525b
(4 44 . 26 06)
4 22b
(3 68 - 4 76)
3 88b
(3 70 - 4 07)
79 6C
44 2d
Total Phosphorus
Exoort
kg/ha/yr Reference
1 22a Minshall et al ,
(1 22 - 1 49) 1970
2 00a Minshall et al ,
(1 03 - 5 77) 1970
75a Minshall et al ,
( 68 - 96) 1970
95a Minshall et al ,
(76-1 18) 1970
1 30b Hensler et al ,
(1 00 - 1 60) 1970
3 40b Hensler et al ,
(1 13 - 5 66) 1970
81b Hensler et al ,
( 73 - 90) 1970
94b Hensler et al ,
( 91 - 97) 1970
18 6C Young and Holt,
1977
14 Od Young and Holt,
1977
-------
Table 7- (continued)
Land Use
Corn, surface
spread manure
( 009 ha)
Corn, plowdown
manure
( 009 ha)
Corn { 009 ha)
Corn ( 009 ha)
Corn, contour
planting (30 ha)
Corn, contour
planting (33 6 ha)
Corn, contour
planting (60 ha)
Corn (1 29 ha)
Corn ( 001 ha)
Soybeans, two
crops/yr, conven-
Fertilizer
Application
kg/ha/yr
N P K
29 81
plus
239 43
from nanure
29 81
plus
239 42
from manure
56 29
112 29
448 64
168 39
280 64
284 54
100 35 35
0 29 56
Location
Morris,
Minnesota
Morris,
Minnesota
Morn s ,
Minnesota
Morris,
Minnesota
Treynor, Iowa
Treynor, Iowa
Treynor, Iowa
Watkinville,
Georgia
Northern,
Alabama
Holly Springs,
Mississippi
Soil
Type/Texture
loam
loam
loam
loam
deep loess,
fine, silty
mixed mesics
deep loess,
fine, silty
mixed mesics
deep loess,
fine, silty
mixed mesics
sandy loam-
sandy clay
loam
silt loam
silt loam
Precipitation
cm/yr
65 7d
65 7d
57 2e
57 2e
79 79f
(63 07 - 105 95)
78 65f
(62 16 - 104 59)
73 76f
(52 8 - 102 5)
107 7
87 39
143 75b
Water
Runoff
cm/yr
38d
40d
457e
8 03e
547f
(1 37 - 12 57)
3 86f
(1 52 - 9 76)
175f
( 35 - 10 71)
13 0
55 75b
Total Nitrogen
Export
kg/ha/yr
27 9d
33 Od
14 24e
23 63e
8 69f
(2 2 - 72 47)
5 36f
(1 69 - 43 71)
21f
(67-26 7)
12 42
3 29
46 50b
Total Phosphorus
Export
kg/ha/yr
86d
98d
3 14e
555e
59f
(092-2 118)
35f
(083-1 288)
26f
( 024 - 613)
2 21
40
17 64b
Reference
Young and Holt,
1977
Young and Holt,
1977
Burwell et al ,
1975
Burwell et al ,
1975
Alberts et al ,
1978
Alberts et al ,
1978
Alberts et al ,
1978
Smith et al ,
1978
Bradford, 1974
McDowell et al
1978
tional tillage
( 01 ha)
-------
Table 7: (continued)
Land Use
Soybeans, two
crops/yr, no till
( 01 ha)
Cotton (17 9 ha)
Cotton (12 1 ha)
Soybeans - Corn
two crops/yr
no till ( 01 ha)
Corn - Soybeans
two crops/yr
no till ( 01 ha)
Tobacco and Corn
Fertilizer
Application
kg/ha/yr
M P K
0 29 56
33 25 24
33 25 24
0 29 56
136 20 37
85 40
Location
Holly Springs,
Mississippi
Chickasha,
Oklahoma
Chickasha,
Oklahoma
Northern,
Mississippi
Northern,
Mississippi
Rhode River
Watershed,
Maryland
Soil
Type/Texture
silt loam
silt loam
silt loam
silt loam
silt loam
fine sandy,
loam
Precipitation
cm/yr
143 75b
81 39
(72 7 - 97 3)
80 79
(72 9 - 96 3)
143 8
143 8
114 7
Water
Runoff
cm/yr
27 9b
13 I9
(8 8 - 24 1)
12 79
(8 0 - 24 8)
54 9
50 5
Total Nitrogen
Export
kg/ha/yr
51b
9319
(4 99 - 11 49)
11 169
(5 18 - 14 84)
23 0
19 3
3 7
Total Phosphorus
Export
kg/ha/yr
26b
4319
(2 38 - 11 52)
4589
(2 07 - 10 75)
7 2
3 7
1 4
Reference
McDowell et al ,
1978
Menzel et al ,
1978
Menzel et al ,
1978
McDowell et al ,
1978
McDowell et al ,
1978
Correll et al ,
1977
a Three year median
b Two year mean
c Ten year mean
d Three year mean
e Six year mean
f Seven year median
g Four year median
-------
Table 8. Nutrient Export from Non Row Corps
Land Use
Alfalfa ( 004 ha)
Alfalfa, fall
applied manure
( 004 ha)
Alfalfa, winter
applied manure
( 004 ha)
Alfalfa, spring
applied manure
( 004 ha)
Alfalfa and
Bromegrass
two plots
3 55 - 4 10 ha
Wheat (5 2 ha)
Wheat (b 3 ha)
Spring wheat
and summer stubble
Two year rotation
(4-5 ha)
Spring wheat
and summerf allow
(4-5 ha)
Spring wheat and
fall fertilized
summerfallow
(4-5 ha)
Fertilizer
Application
kg/ha/yr
N P K
000
121 24 100
121 24 100
121 24 100
45 7C
45 7C
000
000
50 54
Locati on
Madison,
Wisconsin
Madison,
Wisconsin
Madi son ,
Wisconsin
Madison,
Wisconsin
Eastern
South Dakota
Chickasha,
Oklahoma
Chickasha,
Oklahoma
Swift Current,
Saskatchewan,
Canada
Swift Current,
Saskatchewan,
Canada
Swift Current,
Saskatchewan,
Canada
Soil
Type/Texture
silt loan
silt loam
silt loam
silt loam
sandy clay
laom
silt loam
silt loam
loam
loam
loam
Precipitation
cm/yr
107 8a
(105 4 - 108 8)
107 8a
(105 4 - 108 8)
107 8a
(105 4 - 108 8)
107 8a
(105 4 - 108 8)
57 9b
(50 0 - 65 7)
80 4d
(72 9 - 96 5)
BO 5d
(72 9 - 96 6)
Water
Runoff
cm/yr
14 2a
(8 2 - 18 5)
78a
(52-90)
10 3a
(8 2 - 12 8)
10 la
(6 7 - 15 0)
2 69fa
8 75d
(5 5 - 20 8)
74d
(5 4 - 23 0)
35 Ob
(7 0 - 62 5)
58 5b
(19 0 - 98 0)
28 Ob
(7 0 - 49 0)
Total Nitrogen
Export
kg/ha/yr
6 28a
(5 66 - 14 67)
6 63a
(6 10 - 23 09)
7 82a
(5 88 - 38 22)
643a
(4 07 - 11 42)
97b
588d
(3 77 - 7 12)
6 53d
(2 89 - 8 95)
Total Phosphorus
Export
kg/ha/yr
76a
(75-2 40)
1 24a
(1 20 - 8 08)
64a
(58-6 09)
181a
(55-2 39)
10b
164d
{80-3 34)
156d
(59-4 29)
35b
(1-6)
135"
{4-23)
2 9b
(2-56)
Reference
Converse et al ,
1976
Converse et al ,
1976
Converse et al ,
1976
Converse et al ,
1976
Harms et al ,
1974
Menzel et al ,
1978
Menzel et al ,
1978
Nicholaichuk
and Read, 1978
Nicholaichuk
and Read, 1978
Nicholaichuk
and Read, 1978
-------
Table 8: (continued)
Land Use
Fertilizer
Application
kg/ha/yr
N P K
Location
Mater Total Nitrogen Total Phosphorus
Soil Precipitation Runoff Export Export
Type/Texture cm/yr cm/yr kg/ha/yr kg/ha/yr Reference
Millet ( 001 ha) 100 35 35 Northern
Alabama
Oats ( 009 ha)
Hay ( 009 ha)
18 30
Morris,
Minnesota
000 Morris,
Minnesota
silt loam
loam
loam
87 39
57 2e
57 2e
6 89e
14 2e
3 04
4 22e
4 09C
44
65C
64C
Bradford, 1974
Burwell et al ,
1975
Bur-well et al ,
1975
a Three year median
b Two year mean
c Eleven year mean
d Four year median
e Six year mean
-------
Table 9: Nutrient Export from Grazed and Pastured Watersheds
Fertilizer
Application
kg/ha/yr
Land Use N P K
Moderate dairy 37 16 8
grazing, blue-
grass cover
(1.88 ha)
Heavy dairy 149 64 12
grazing, blue-
grass cover
(1 48 ha)
Pasture (6 28 ha)
Winter grazed 56 0 0
and summer rota-
tional, orchardgrass
and bluegrass cover
(1 ha)
Summer grazed 56 0 0
(1 ha)
Rotation grazing 168 39
(42 9 ha)
Pasture for 000
brood cattle
(10 ha)
Continuous
grazing with some
supplementary winter
feeding, some hay
production
(351 2 ha)
Continuous 000
grazing, little
bluestem cover,
Active gullies
(11 1 ha)
Soil
Locati on Type/Texture
Wayneville,
North Carolina
Wayneville,
North Carolina
Eastern sandy clay
South Dakota loam
Coshocton, silt loam
Ohio
Coshocton, silt loam
Ohio
Treynor, silt loam
Iowa
Eatonton ,
Georgia
Rhode River well drained,
Watershed, sandy loams
Maryland
Cmckasha, silt loans
Oklahoma
Precipitation
crn/yr
106 la
(104 3 - 119 8)
106 la
(104 3 - 119 8)
58 4
108 0
108 0
75 44C
(73 3-77 83)
164 0
114 7
88 25d
(50 7 - 105)
Water
Runoff
cm/yr
21 3a
(12 3 - 24 6)
26 4a
(19 9 - 31 8)
4 44
12 94
2 92
3 86C
(94-4 39)
61 8
15 ld
(2 02 - 28 4)
»
Total Nitrogen
Export
kg/ha/yr
3 46a
(2 41 - 3 83)
10 99a
(8 31 - 18 05)
1 52
30 85
21 35
2 32C
(47-4 28)
13 0
6 13d
(1 33 - 9 23)
Total Phosphorus
Export
kg/ha/yr
T4a
( 12 - 16)
16a
( 11 - 70)
25
3 6
85b
0
251C
( 081 - 512)
1 35
3 8
146d
(27-3 86)
Reference
Kilmer et al ,
1974
Kilmer et al ,
1974
Harms et al ,
1974
Chi Chester et
al , 1979
Chichester et
al , 1979
Schuman et al ,
1973 a, b
Krebs and
Golley, 1977
Correll et al ,
1977
Menzel et al ,
1978
-------
Table 9: (continued)
CTi
Land Use
Rotation grazing
little bluestem
cover, good cover
(11 0 ha)
Conti nuous
grazing, little
bluestem cover
(7 8 ha)
Rotational
grazing, little
bluestem cover
(9 6 ha)
Conti nuous
grazing, little
bluestem cover
active gullies
(I1 1 ha)
Rotational
grazing, little
bluestem cover
(11 0 ha)
Fertilizer
Application
kg/ha/yr
N P K
000
83 72 0
87 76 0
000
000
Location
Chickasha,
Oklahoma
Chickasha,
Oklahoma
Chickasha,
Oklahoma
Chickasha,
Oklahoma
Chickasha,
Oklahoma
Soil Precipitation
Type/Texture cm/yr
silt loams 88 35
(52 4 - 109 1)
silt loam 76 5e
silt loam 78 2s
silt loam 76 5e
silt loam 78 2e
Hater
Runoff
cm/yr
595d
(35-17 8)
14 7
4 3
10 2
4 3
Total Nitrogen
Export
kg/ha/yr
148d
{ 15 - 2 3)
9 20
4 72
5 19
1 73
Total Phosphorus
Export
kg/ha/yr
25d
(02-1 44)
4 90
3 09
76
20
Reference
Henzel et al ,
1978
Olness et al ,
1980
Olness et al ,
1980
Olness et al ,
1980
Olness et al ,
1980
a Four year median, sediment phase not sufficiently examined
b Major contribution from underground spring
c Three year median
d Four year median
e Nine year mean
-------
Table 10: Nutrient Export from Animal Feedlots and Manure Storage
Land Use
Beef livestock
feedlot (4 76 ha)
Lamb feedlot
(21 32 ha)
Lamb feedlot
(12 63 ha)
Dairy confinement,
45 head of tattle
( 13 ha)
Beef and sheep
feedlot ( 603 ha)
Beef feedlot,
300 head of
cattle (1 6 ha)
Beef cattle
feedlot, 9 29
m2/cow ( 002 ha)
Beef cattle
feedlot, 18 6
m2/cow ( 002 ha)
Beef cattle
feedlot, 18 6
mz/cow ( 002 ha)
Beef cattle
feedlot, 500 - 600
cattle ( 25 ha)
Beef cattle
feedlot ( 17 ha)
Location
Brookings,
South Dakota
Brookings,
South Dakota
Brookings,
South Dakota
Brookings,
South Dakota
Brookings,
South Dakota
Brookings
South Dakota
Mead, Nebraska
Mead, Nebraska
Mead, Nebraska
Kent Co ,
Ontario, Canada
Waterloo Co ,
Ontario, Canada
Soil /Surf ace
Characteristics
1/2 concrete
1/2 grassed
includes
detention pond
includes
detention storage
culvert
concrete plus
roof runoff
concrete
surface
silty clay loam
overlying sand
silty clay loam
overlying sand
silty clay loam
overlying sand
concrete
paved and unpaved
Precipitation
cm/yr
60 71 a
(53 19 - 61 06)
54 48b
(49 96 - 59 0)
49 96
58 Ola
(48 16 - 62 53)
58 01 a
(48 16 - 62 53)
59 74b
(55 83 - 63 73)
70 7
78 6
Water iota! Nitrogen
Runoff Export
cm/yr kg/ha/yr
21 87a
(8 92 - 28 52)
2 07b
(1 96 - 2 18)
3 10
27 18a
(15 16 - 82 65)
15 24a
(14 40 - 30 35)
6 35b
(3 99 - 8 71)
15 87b 2923 2b
(14 68 - 17 07) (2016 0 - 3830 4)
17 93b 1344b
(16 59 - 19 28) (1254 4 - 1433 6)
24 94b 3584b
(24 59 - 25 3) (1388 8 - 2195 2)
33 2 3372 27
17 3 680 52
Total Phosphorus
Export
kg/ha/yr
523 Oa
(145 6 - 749 0)
26 88b
(20 16 - 35 84)
21 28
355 Oa
(301 3 - 521 9)
222 9a
(157 9 - 2635 4)
86 2b
(29 1 - 142 2)
795 2b
(291 2 - 1299 2)
347 2b
(224 0 - 470 4)
224b
(134 4 - 313 6)
425
170
Reference
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
McCalla et al ,
1972
McCalla et al ,
1972
Gilbertson et al ,
1975
Coote and Hore,
1978
Coote and Hore,
1978
-------
Table 10: (continued)
Land Use
Solid manure
storage agea
( 05 ha)
Manure storage
facility ( 05 ha)
Location
Elmira,
Ontario, Canada
Burlington,
Vermont
Soil /Surf ace
Characteristics
2/3 paves
1/3 unpaved
crushed
limestone
Precipitation
cm/yr
67 37
57 7
Hater
Runoff
cm/yr
20 9
33 5
Total Nitrogen
Export
kg/ha/yr
T891 07
7979 9
Total Phosphorus
Export
kg/ha/yr
172
539 9
Reference
Coote and Hore,
1978
Magdoff et al ,
1977
a Three year median
b Two year mean
c Derived from original values of kg/cow/yr with permission of authors
00
-------
Table 11 Nutrient Export from Mixed Agricultural Watersheds
Fertilizer
Application
kg/ha/yr
Land Use N P K
58% row crops 44
31% small grain
and pasture
6% woods
5% urban
(4950 ha)
63% row crops
26% small grain
and pasture
8% woods
3% urban (942 ha)
35% row crops
48% small grain
and pasture
5% woods
12% urban (714 ha)
50% pasture
25% rotation
cropland
25% hardwood
forest
(123 ha)
39% corn 134 46 120
46% legumes
and grass
9% small grain
2% idle
4% roads
(594 ha)
60% row crops 127 28
40% hay and
pasture
2 livestock
feedlots
(157 5 ha)
Location
Black Creek
watershed,
Marian,
Indiana
Smith-Fry
uram,
Harlan,
Indiana
Dreisbach
Drain,
Harlan,
Indiana
Coshocton,
Ohio
Ottowa ,
Ontario,
Canada
Macedonia,
Iowa
Soil Precipitation
Type/Texture cra/yr
silt loam, clay, 91 Oa
silty clay loam, (70 - 112)
silty clay
silt loam, clay, 91 Oa
silty clay loam, (70 - 112)
silty clay
silt loam, clay, 91 Oa
silty clay loam, (70 - 112)
silty clay
silt loam 88 8b
(77 7 - 92 7)
clay loam, 95 lc
sandy loam
silt loam 67 79
Water Total Nitrogen
Runoff Export
cm/yr kg/ha/yr
19 35a 28 65a
(11 2 - 27 5) (8 6 - 48 7)
20 75a 31 76a
(12 4 - 29 1) (10 3 - 53 2)
18 05a 25 85a
(10 1 - 26 0) (6 6 - 44 1)
33 35b 3 74b
(26 9 - 34 4) (1 67 - 10 61)
18 6C
(8 2 - 24 2)
10 74 9 64 '
Total Phosphorus
Export
kg/ha/yr Reference
3 15a Lake and
(11-52) Morrison,
3 25a Lake and
(11-54) Morrison,
3 00a Lake and
( 1 00 - 5 00) Morrison,
Taylor et
1971
60C Patni and
(01-08) 1978
1977
1977
1977
al ,
Hore
648 Burwell et
al , 1974
-------
Table 11: (continued)
Fertilizer
Application
kg/ha/yr
Land Use N P K
Three years, 343 67
pasture, two
years corn
(42 9 ha)
Intensive 120 33
Agriculture
crops and im-
proved pasture
(208 ha)
Active cropping
and pasture
CO
o
At least 80%
of watershed
devoted to
agricultural
activities
37 4% soybean
and whitebean
27 1% cereal
23% corn (5080 ha)
36 U woodland
25 0% cereal
22 2% tobacco
10 1% corn
3% pasture and
hay
(7913 ha)
Soil Precipitation
Location Type/Texture cro/yr
Treynor, Iowa silt loam 84 71
North Central, sand 96 5a
Florida (88 - 105)
South of
Washington,
D C
Southern
Ontario,
Canada
Thames River, lacustrine clay 72 9
Southern over till plain
Ontario, Canada over limestone
Big Creek, deep level
Southern deltaic sands
Ontario, Canada
Water Total Nitrogen Total Phosphorus
Runoff Export Export
cra/yr kg/ha/yr kg/ha/yr
17 65 14 11 27
16 7a 4 23a 1 1a
(12 1 - 21 3) (2 10 - 6 36) (86-1 34)
2 82 409
14 3d 1 29d
( 62 - 23 5) (05-2 30)
16 1 1 28
64 26
Reference
Burwell, et al,,
1977
Campbell, 1978
Gnzzard et al ,
1977
Avadhanula, 1979
Coote et al
(ed ), 1978
Coote et al
(ed ), 1978
31 3% corn
26 4SS cereal
17 9% pasture
and hay
12 1% soybean
and whitebean
7 5% woodland
(6200 ha)
AuSable River level clay
Southern till plain
Ontario, Canada over shale
86 0
41 5
91 Coote et al
(ed ), 1978
-------
Table 11 (continued)
CO
Fertilizer
Application
kg/ha/yr
Land Use N P K Location
37 2.% pasture
and hay
35 3% cereal
18 7% corn
6 9% woodland
(1860 ha)
42 3% corn
22 8% pasture
and hay
15 4% woodland
12 2% cereal
(3000 ha)
33 4% pasture
and hay
29 2% woodland
22 3% cereal
12 3% corn
37 4% woodland
28 5% pasture
and hay
10 7% cereal
10 4% corn
3 7% tobacco
(5645 ha)
44 2% pasture
and hay
18 4% cereal
17 8% woodland
16 2% corn
(3025 ha)
41 3% pasture
and hay
29 0% cereal
11 3% corn
7 5% woodland
(2383 ha)
Grand River,
Southern
Ontario, Canada
Middle Thames
River,
Southern
Ontario, Canada
Haiti and River,
Southern
Ontario, Canada
Shelter Valley
Creek, Southern
Ontario, Canada
Twenty Mile
Creek, Southern
Ontario, Canada
Humber River,
Southern
Ontario, Canada
Water Total Nitrogen Total Phosphorus
Soil Precipitation Runoff Export Export
Type/Texture cm/yr cm/yr kg/ha/yr kg/ha/yr Reference
silty clay 92 5
ground
moraine
calcareous 101 8
loamy till
drum! im zed 82 3
loam till
windblown 84 0
sand and silt
on scoping
sandy calcareous
till
lacustrine 77 9
and reworked
clay over
dolomite
stratified 73 7
clay over
shale and
limestone
20 3 1 00 Coote
(ed )
31 1 1 53 Coote
(ed )
14 3 16 Coote
(ed )
32 08 Coote
(ed )
15 5 1 53 Coote
(ed )
11 1 49 Coote
(ed )
et al
, 1978
et al
, 1978
et al
, 1978
et al
, 1978
et al
, 1978
et al
, 1978
-------
Table 11: (continued)
00
ro
Fertilizer
Application
kg/ha/yr
Land Use N P K
27 8% vegetables
22 8% corn
10 0% woodland
8 9% cereal
7 9% soybean and
whitebean
(1990 ha)
66 6% pasture
and hay
12 1% cereal
9 5% corn
9 4% woodland
(4504 ha)
Location
Hillman Creek,
Southern
Ontario, Canada
Saugeen River
Southern
Ontario, Canada
Soil Precipitation
Type/Texture cm/yr
shallow moraine 77 0
sand over clay
till plain over
limestone
reworked 92 4
lacustrine
clay over
clay till
Water Total Nitrogen
Runoff Export
cm/yr kg/ha/yr
25 2
9 4
Total Phosphorus
Export
kg/ha/yr
91
81
Reference
Coote
(ed )
Coote
(ed )
et al
, 1978
et al
, 1978
a Two year mean
b Four year median
c Three year median
d Estimates based on PLUARG Task C monitoring of selected sites in the Grand and Saugeen River basins
-------
Table 12. Nutrient Export from Urban Watersheds
Land Use
Residential (50 ha)
78% industrial
22% commercial
(49 ha)
Commercial
(15 8 ha)
Central business
district (9 3 ha)
go Industrial (8 1 ha)
Residential
(41 7 ha)
Low density
residential
subdivision,
Large lots with
complete grass
cover and trees
(46 82 ha)
Low density
residential,
Extensive grassed
areas, small lots,
(33 73 ha)
High density
residential
Location
Madison, Wis-
consin
Menominee,
Wisconsin
Appleton,
Wisconsin
Appleton,
Wisconsin
Appleton,
Wisconsin
Appleton,
Wisconsin
Okemos ,
Michigan
Holt, Michigan
East Lansing,
Michigan
Soil /Surf ace Precipitation
Characteristics cm/yr
27% impervious 69 93
surface
silt and clay
loams
clay loam /6 5
overlying
dolomite bedrock
clay loam 76 5
overlying
dolomite bedrock
clay loam 76 5
overlyi ng
dolomite bedrock
clay loam 76 5
overlying
dolomite bedrock
sandy loam, 77 19
sandy clay loam
sandy loam, 77 19
sandy clay loam
sandy loam, 77 19
sandy clay loam
Water Total Nitrogen
Runoff Export
cm/yr kg/ha/yr
10 49 50
24 03a
(7 88 - 40 18)
454b
38 47b
6 53b
367b
1 52C
69C
48C
Total Phosphorus
Export
kg/ha/yr
1 1
2 67a
(1 06 - 4 28)
88b
408b
75b
35b
0 19C
27C
1 lc
Reference
Kluesener and Lee,
1974
Konrad et al ,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Landon, 1977
Landon, 1977
Landon, 1977
townhouse complex,
limited open space
(7 ha)
-------
Table 12: (continued)
Land Use
High density
residential co-
operatives, large
amounts of open
grassed areas
(21 63 ha)
Commercial,
Shopping Center
(18 19 ha)
Commencal, light
industry and busi-
oo ness (4 19 ha)
•p.
64% residential
13% recreational
12% comnencal
6% transportation
1% industrial
(958 ha)
Residential
and light commer-
cial (11 ha)
At least 60% of
watershed devoted
to urban land use
Industrial and
residential (414 ha)
Commercial (212 ha)
Location
Lansing,
Michigan
Meridian Twp
Ingham Co
Michigan
Lansing,
Mi chi gan
Montgomery
Creek, Kitchner
Ontario, Canada
Cincinnati,
Ohio
Southern
Ontario, Canada
Third Creek
Watershed,
Knoxville,
Tennessee
Fourth Creek
Watershed,
Knoxville
Tennessee
Soil /Surf ace Precipitation
Characteristics crn/yr
sandy loam, 77 19
sandy clay loam
sandy loam 77 19
sandy clay loam
sandy loam, 77 19
sandy clay loam
37% impervious 76 2
surface
carbonatic 150 0
bedrock with
shales, 28% imper-
vious surfaces
soluble 155 0
dolomitic car-
bonate rock, 45%
impervious sur-
faces
Hater Total Nitrogen Total Phosphorus
Runoff Export Export
cm/yr kg/ha/yr kg/ha/yr Reference
5 5C 56C Landon, 1977
20 5C 1 7C Landon, 1977
4 Oc 66C Landon, 1977
757 O'Neill, 1979
28 19 9 97 Weibel et al ,
1964
9 48d 1 63d Avadhanula, 1979
(6 65 - 10 2) (73-2 05)
84 3 14 95 4 17 Betson, 1978
41 1 12 78 4 85 Betson, 1978
-------
Table 12' (continued)
Land Use
Suburban (62 ha)
60% residential
19% commercial and
industrial
12% institutional
10% unused
(432 54 ha)
60% residential
Co 19% commercial and
tri industrial
12% institutional
10% unused
(432 54 ha)
20% urbanized,
large scale resi-
dential (47900 ha)
Single family resi-
dential (19 2 ha)
67% residential
13% conmercial
12% woodland
8% agriculture
(792 ha)
Residential (6 8 ha)
Location
Plantation Hills
Residential Area
Knoxville,
Tennessee
Durham, North
Carolina
Durham, North
Carolina
Bull Run Basin,
Occoquan
Watershed,
Virginia
Broward Co ,
Florida
Tallahasee,
Florida
Durban,
South Africa
Soil/Surface Precipitation
Characteristics cm/yr
soluble dolomitic 153 0
carbonate rock
23% impervious
surfaces
29% impervious 108 2
surfaces
29% impervious
surfaces
sedimentary
sandstones and
shales
quartz sand, 125 6
39% impervious
surface
well drained 249 0
loamy soils
quartz sand 113 06
with some clay
Water Total Nitrogen Total Phosphorus
Runoff Export Export
cm/yr kg/ha/yr kg/ha/yr Reference
94 1 56 43 Betson, 1978
16 26 1 23 Bryan, 1970
24 64 5 26 Colston, 1974
33 76e 1 91 2e Gnzzard et al ,
1978
9 42 1 48 0 21 Mattraw and
Sherwood, 1977
48 3 6 23 Burton et al ,
1977
18 99 40 06 Simpson and
Hemens, 1978
content, 20% imper-
vious surface
a two year mean
b Estimates based on annual streamflow measurements and nine monitored runoff events during 8 month water quality sampling period
c Estimates based on annual streamflow measurements and five months water quality sampling
d Estimates based on PLUARG Task C monitoring of selected sites in the Saugeen and Grand River basins
e Suspected of having nonurban influences
-------
CO
TABLE 13a: Forest Atmospheric Inputs
LOCATION
PHOSPHORUS (kg/ha/yr)
Dissolved-P Total-P
NITROGEN (kg/ha/yr)
N03-N NH3-N Orgamc-N Total-N
REFERENCE
Rawson Lake, Ontario,
Canada
Clear Lake, Ontario,
Canada
White Mountains,
New Hampshire
Hubbard Brook Exp Forest
New Hampshire
Walker Branch Watershed
Tennessee
Coweeta Experimental
Watershed, N Carolina
North Carolina
Duke Forest, N Carolina
N East, Minnesota
H J Andrews Exp Forest,
Western, Oregon
N Central, Minnesota
Mississippi
Northern Mississippi
Northern Mississippi
New Mexico
Sapelo Is , Georgia
Watersmeet, Michigan
Beaver Island, Mich
Beaver Island, Mich
Rock Island St Pk , Wis
Finger Lakes Area, NY
327
26
035
54
19
21
28
14
27
48
3
07
41
19
036 216
032
039
181
7
6 0
4 3
3 9
2 88
C
—3
1 46
135
2 25
3 12
2 64
1 255
5.37
1 CO
1 \J\J
2 8
2 24
2 0
52
54
74
85
2 74
5 73
1 74
95
3 37
6 27 Schindler et al., 1976
Schindler et al , 1970
Martin, 1979
Likens et al , 1977
8 7 Henderson, 1977
Swank & Henderson, 1976
Wells & Jorgensen, 1975
1 33 3 53 Wells et al , 1972
Wright, 1976
99 Fredriksen, 1972
2 32 7 32 Verry & Timmons, 1977
11 3 Switzer & Nelson, 1972
Schreiber et al , 1976
Duffy et al, 1978
2 39 6 77 Gosz, 1978
633 2 84* Haines et al , 1976
Eisenreich et al , 1977
Eisenreich et al , 1977
Murphy & Dos key, 1976
Murphy & Doskey, 1976
Likens, 1972
*wetfa11 only
-------
CO
TABLE 13tr Agricultural-Rural Atmospheric Inputs
LOCATION PHOSPHORUS (kg/ha/yr)
Dissolved-P Total-P
Silver Lake St Pk ,
Michigan
Wisconsin
Wisconsin
Great Britain
Eatonton, Georgia
NITROGEN (kg/ha/yr)
N03-N NH3-N Organic-N Total-N
REFERENCE
Treynor, Iowa
Rhode River Watershed
Edgewater, Maryland
Coshocton, Ohio
Morris, Minnesota
Southern Ontario, Canada
Pellston, Michigan .20
Houghton Lake,
Michigan 29
82
.20
125
97
25
31
, ... 7 9fi «... 9rhitman ft RnvinfoT 1 1Q7/1
4 71 5 66 10 49 Miklas et al , 1977
p o rhirho"tr>r rt nl 1070
2 45 5 09 Burwell et al , 1975
38 0 Sanderson, 1977
4 85 3 09 Richardson & Merva, 1976
3 21 2.09 Richardson & Merva, 1976
086
74
192
3 51 12 22 14 43 30 16
2 73 2 86 6 54 13 13
13 1
Murphy et al , 1976
Hoeft et al , 1972
Hoeft et al , 1972
Frissel 1978
Krebs & Golley, 1977
-------
CO
CO
TABLE 13c: Urban-Industrial Atmospheric Inputs
LOCATION PHOSPHORUS (kg/ha/yr)
Dissolved-P Total-P
NITROGEN (kg/ha/yr)
N03-N NH3-N Orgamc-N Total-N
REFERENCE
Washington, D C
Knoxville, Tennessee
San Francisco, Cal.
Wisconsin
Madison, Wisconsin
Madison, Wisconsin
Milwaukee, Wise
Grand Haven, Mich
Saginaw Bay, Mich
Chicago, Illinois
Chicago, Illinois
Halifax, Nova Scotia
Durban, South Africa
Munich, Germany
Hamburg, Germany
Stockholm, Sweden
London, England
Pans, France
2 58**
2 17 3 67
26
99
1 02
.372
415
1 12 1 21
327
084 558
.56
27 .52
80
2 0
1 6
2 1
1 6
4.4 3.4 24 8
23 76
3 73 3.61 6 19
4 73
1 21
3 95 4 22 14 74
8 26 36
3 3 20.2
8 2 1
25 17 3
23 14 8
9.09**
24.8
13 53
24 0
23 0
22 91
23 5
7 4
19 8
17 1
Randall et al , 1978
Betson et al , 1978
McDoll et al , 1978
Hoeft et al , 1972
Likens & Loucks, 1978
Kluesener, 1972
Eisenreich et al , 1977
Eisenreich et al , 1977
Richardson et al , 1976
Murphy et al , 1976
Eisenreich et al , 1977
Hart & Ogden, 1977
Simpson & Hemens, 1979
Goettle, 1978
Fnssel, 1978
Frissel, 1978
Frissel, 1978
Frissel, 1978
**dustfall only
-------
Table 14: Nutrient Loads for Household Wastewater
Discharged into Septic Tanks, (kg/capita/yr).
Total P Total N Reference
1.49 6.45 Ligman et al, 1974
1.43 5.99 Laak, 1975
2.65 Bennet and Linstedt, 1975
.74 4.61 Chan et al, 1978
1.59 Ellis and Childs, 1973
1.49 2.15 Siegrist et al, 1976
3.00 Bernhard, 1975
.80 Otis et al, 1975
8.20 Walker et al, 1973
1.28 3.20 EPA-NES, 1974
89
-------
Table 15: Magnitude and Variability for Phosphorus Loading from Waste Water
Treatment Plant Effluent*
Treatment Type
Activated Sludge
Trickling Filter
Phosphorus Removal
Primary Settling and Digestion
Oxidation Pond
Sand Filter
Median Loading
(kg/capita/yr)
.89
1.10
.57
.82
1.07
2.86
Range
(kg/capita/yr)
.32
.39
.23
.27
.36
.77
- 4.99
- 5.44
- 1.81
- 3.18
- 3.63
- 6.11
Sample
Size
183
158
16
53
52
11
*Ten to fourteen samples taken per year. Adapted from EPA-NES Working Paper Number 22
(U.S.E.P.A., 1974)
-------
lO-i
PHOSPHORUS EXPORT FROM
oo
56
LU
Z3
O
UJ
£ 4
no
0
i lyui
FORESTED WATERSHEDS
MEDIAN ,206
MEAN ,236
INTERQUARTILE RANGE ,218
STANDARD DEVIATION .200
RANGE ,019 - ,830
SAMPLE 26
PHOSPHORUS EXPORT (KG/HA/YR)
-------
6n
«,
NITROGEN EXPORT FROM
FORESTED LAND USE
rv>
5-
4-
O
O
LU
I -
0
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
2,46
2,86
1,19
1,37
1,38 - 6,26
11
NITROGEN EXPORT (K6/HA/YR)
-------
vo
Figure 5a: PHOSPHORUS EXPORT FROM ROW CROPS
10-
8-
o
UJ
=D
O
UJ
tr.
4-
2-
0
0
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
2,24
4,46
4,37
5,24
,26 - 18,6
26
6 8 10 12 14 16
PHOSPHORUS (KG/HA/YR)
18 20
-------
10
14
12-
10-
8-
UJ
ID
O
UJ
o:
0
0
NITROGEN EXPORT FROM ROW CROPS
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
16,09
18,05
18,06
2,1 - 79,6
26
10 20
30 40 50 60
NITROGEN (KG/HA /YR)
70 80 90
-------
1C
cn
Figure
PHOSPHORUS EXPORT FROM
NONROW CROPS
6-
5-
>•
I4-
o
UJ -2
QC J
U_
2-
\ -
n -
MEDIAN ,76
MEAN 1,08
INTERQUARTILE RANGE ,92
STANDARD DEVIATION ,77
RANGE ,10 - 2,90
SAMPLE SIZE 13
0 .5 1.0 1.5 2.0 2.5 30
PHOSPHORUS (KG/HA/YR)
3.5
-------
6b: NITROGEN EXPORT FROM
NONROW CROPS
O1
O
LU
O
LU
0
0
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
2 4 6 8 10
NITROGEN (KG/HA/YR)
12
6,08
5,19
2,38
2,07
,97 - 7,82
10
-------
ID
--si
4-
o
z
LU
O
UJ
tr
2-
0
a. PHOSPHORUS EXPORT FROM
GRAZED AND PASTURED WATERSHEDS
MEDIAN ,81
MEAN 1,50
INTERQUARTILE RANGE 2,43
STANDARD DEVIATION 1,64
RANGE ,14 - 4,90
SAMPLE 14
1.0 1.5 20 2.5 3.0 3.5 4.0
PHOSPHORUS (KG/HA/YR)
4,5 5.0 5.5
-------
Figur.71,: NITROGEN EXPORT FROM
GRAZED AND PASTURED WATERSHEDS
-------
UD
Figure 8a: PHOSPHORUS EXPORT FROM
ANIMAL FEEDLOT AND MANURE STORAGE
MEDIAN 224,0
MEAN 300,7
INTERQUARTILE RANGE 255
STANDARD DEVIATION 226,64
t -
o
LU
0
LU
CC
U_
2-
n.
RANGE
SAMPLE S
21,28 - 795,2
IZE 13
i i
PHOSPHORUS (KG/HA/YR)
-------
o
O
Figure 8b: NITROGEN EXPORT
FROM
ANIMAL FEEDLOT AND MANURE STORAGE
4-
LU
O
LU
2-
0
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
2923,2
3110,7
1860,6
2400,4
680,5 - 7979,9
7
NITROGEN (KG/HA/YR)
-------
7-
"6-
5-
«
54-
LU
13-
U.
2-
\ -
n .
r iyu
ic ya. i i i w v-/ 1 i i vy i » *k/ \J l_/\i \yi\i I i\v/i*i
MIXED AGRICULTURAL WATERSHEDS
MEDIAN ,91
MEAN 1 , 134
INTERQUARTILE RANGE ,88
STANDARD DEVIATION ,964
RANGE ,08-3,25
SAMPLE SIZE 20
0 .5 1.00 1.5 2.0 2,5 3.0 3.5
PHOSPHORUS (KG/HA/YR)
4.0 4.5
-------
Figure 9b: NITROGEN EXPORT
o
ro
6-
_
FREQUENCY
D f>0 -P*
I II I I
I-KUM
MIXED AGRICULTURAL WATERSf
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
i
0 10 20 30 40 50
NITROGEN (KG/HA/YR)
60
11,30
16,53
15,80
10,77
2,82 - 41,50
21
-------
10
8-
6-
LU
O
LU
4-
0
0
PHOSPHORUS EXPORT FROM
URBAN WATERSHEDS
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
1,1
1,91
2.07
1,79
,19
23
-6,23
1234567
PHOSPHORUS EXPORT (K6/HA/YR)
8
-------
8
Figure 10b: NITROGEN EXPORT FROM
URBAN WATERSHEDS
6-
o
UJ
o
UJ
cc
u.
5-
4-
3-
2-
i -
0
MEDIAN
MEAN
INTERQUARTILE RANGE
STANDARD DEVIATION
RANGE
SAMPLE SIZE
5,50
9,97
7,36
10,16
1,43 - 38,47
19
0
10 15 20 25 30
NITROGEN (KG/HA/YR)
35 40 45
-------
Figure 11•
o
en
FREQUENCY
o
TOTAL PHOSPHORUS EXPORT FROM TWO CORN CROPPED
WATERSHEDS ILLUSTRATING VARIABILITY OVER TIME
(FROM ALBERTS ET AL,, 1973)
0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
PHOSPHORUS EXPORT (KG/HA/YR)
-------
Figure 12: SURFACE WATER RUNOFF FROM TWO CORN CROPPED
WATERSHEDS ILLUSTRATING VARIABILITY OVER TIME
(FROM ALBERTS ET AL,, 1978
6 -
o
en
o
LU
ZD
O
LU
CC
2-
o
SURFACE WATER RUNOFF (cm/YR)
-------
Chapter 4
CONCLUDING COMMENTS
The procedure described in Chapter 2 should be useful for lake trophic
management planning because of the inclusion of uncertainty analysis and the
carefully screened tables of export coefficients Its value will be en-
hanced, however, if the analyst is mindful of the limitations of the methodo-
logy. Listed below are several items reflecting these limitations as well
as guidelines for the interpretation and communication of the modeling/
uncertainty analysis results
1. The prediction of quantitative water quality impacts associated
with changes in land use necessitates the use of a mathematical
model Projected or anticipated land use changes cannot be
measured so information must be extrapolated from other points
in space and/or time Both the application of the mathematical
model and the extrapolation of information imply prediction
error. This error is therefore unavoidable, but when quanti-
fied, prediction uncertainty can be extremely useful in the
planning process.
2. Prediction uncertainty is a measure of the information value
contained in a prediction. If the uncertainty is small, the
prediction is precise, and the predictive information is
valuable Alternatively, if the uncertainty is large, the
prediction is imprecise, and the predictive information is
less valuable. Prediction uncertainty is caused by natural
107
-------
process variability, and bias and error in sampling, measure-
ment, and modeling. Prediction uncertainty can be useful to
the planner as long as it is reliably estimated. However, it
is possible that unquantified supplemental uncertainty (Hos-
teller and Tukey, 1977) exists. This uncertainty term generally
results from errors that are unknown to the analyst. For
example, supplemental error may be introduced, but unquanti-
fied, because of poor choice of export coefficients. This
hidden error may increase planning risks because the error is
not included in the error analysis. Therefore the analyst
must exercise care in the selection of the export coefficients
and in the conduct of the modeling process.
3. The notion of supplemental uncertainty and hidden planning
risks underscores the importance of selecting representative
nutrient export coefficients. The watershed matching process
described in Chapter 3 is central to this concern. The analyst
must be aware of those watershed characteristics that are the
major determinants of nutrient export. Then the appropriate
export coefficients are selected according to a match between
application lake watershed and export coefficient watershed,
on the basis of these causal characteristics. This match
leads to representative and reliable coefficients and dimin-
ishes supplemental uncertainty.
4. The discussion in Chapter 2 identifies the major limitations
on the modeling/uncertainty analysis methodology In funda-
mental terms, the limitations are generally associated with
the fact that the model development data set for any parti-
cular model represents a subpopulation of lakes. Application
108
-------
lakes that differ substantially from the model development
subpopulation may not be modeled well (i.e , results may be
biased). Any limnologic characteristic that is a causal
determinant of lake phosphorus concentration is a candidate
as a limiting, or constraint, variable. These include con-
straints on the model variables (e.g., all model development
data set lakes have P < .135 mg/1), constraints on hydrology
(e.g , there are no closed lakes in the model development
data set), or constraints on climate (e.g., the model develop-
ment data set contains only north temperate lakes).
5. The methodology described in Chapter 2 can be used to quantify
the relationship between watershed land use and lake phos-
phorus concentration. Yet phosphorus by itself is not an ob-
jectionable water quality characteristic. The real quality
variable of concern (i.e , the charactenstic(s) that lend(s)
value or human benefit to the water body, abbreviated "qvc")
may be algal biomass, water clarity, dissolved oxygen levels,
or fish populations (see Figure 1). Therefore the modeling
methodology and the error analysis do not include all of the
calculations necessary to link control variables (land use)
with the qvc. This means that the relevant prediction error
(on the qvc) is underestimated by the phosphorus model pre-
diction error, and planning and management risks are inade-
quately specified. More useful methodologies are needed that
quantitatively link control variables with the qvc for a
particular application.
6. The error analysis procedure presented in Chapter 2 should
provide a reasonable estimate of prediction uncertainty.
109
-------
However, there are still problems in interpretation and appli-
cation. For instance, the model error component was estimated
from a least squares analysis on a multi-lake (cross-sectional)
data set. This error is then applied to a single lake in a
longitudinal sense. Thus, much of the model error term actu-
ally results from multi-lake variability, whereas when the
model is applied to a single lake, the model error term should
consist primarily of lack-of-fit bias and single lake vari-
ability. On the basis of present knowledge, it is not clear
how a multi-lake-derived error relates to a single lake analy-
sis.
7. A second issue associated with the error analysis concerns the
subjective determinations of phosphorus loading and hence,
loading estimation error. Statisticians and modelers generally
prefer objective measures of uncertainty, such as calculated
variability in a set of data However both limited available
data and the obviously unmeasurable nature of future impacts
favor (or necessitate) subjective estimates Given this sub-
jectivity, and the inexperience of most planners and analysts
with phosphorus loading estimation, there may be uncertainty
in the uncertainty estimates. This is exacerbated by the po-
tential for loading error "double counting" (see Reckhow,
1979d), although the procedure described in Chapter 2 is de-
signed to reduce error double counting. It is likely that as
analysts gain experience in loading and error estimation,
this problem will be of less concern.
8. A third uncertainty analysis issue concerns the precise de-
scription of error terms presented in Chapter 2 to minimize
110
-------
error double counting. It was noted in Chapter 2 that some
variable error is already incorporated into the model standard
error. The error analysis procedure proposed is designed to
require additional application lake error only for those fac-
tors not already included in the model error. Therefore the
analyst is urged to closely follow the guidelines in Chapter 2
for export coefficient selection and error estimation The al-
ternative may be a well-intentioned but inaccurate estimate
of prediction uncertainty.
9. The simplicity of this technique necessarily limits its adapt-
ability to certain situations that may occur within a water-
shed. This procedure may be flexible enough to accommodate
some of these situations, but others may require more inten-
sive study (than the procedure provides for). Therefore, it
must be left to the judgment of the analyst as to whether or
not this method is appropriate. Examples of events or char-
acteristics that would alter the effectiveness of this pro-
cedure are:
a) the input of phorphorus from sources not considered
in the method presented. These sources might include
a large number of resident water fowl in and around
the lake or fertilizers applied to shoreline lawns,
b) the trapping of phosphorus by mechanisms not con-
sidered These phosphorus traps might include aquatic
plants or an upstream lake within the watershed,
c) the occurrence of an unnatural phenomenon that alters
the lake ecosystem. These phenomena might include
dredging, filling, and chemical treatment,
m
-------
d) lake types not modeled well with this black box
nutrient model. These types include closed lakes
(lakes without well-defined outlets) and lakes
with strong internal concentration gradients (i e ,
lakes with significant local quality variations)
10. Water quality management planning and modeling incur a cost
that is presumably justified in terms of the value of the in-
formation provided. The actual achievement of a water quality
level often requires management and pollutant abatement costs
but also carries with it various benefits. The analyst must
be cognizant of the fundamental economic nature of environ-
mental management, planning, and decision making The acquisi-
tion of additional data or the conduct of additional modeling
and planning studies should be justified in terms of informa-
tion return for improved decision making
11. Finally, the planner or analyst conducting a lake modeling
study has as his/her primary goal the effective communica-
tion of the work carried out. This does not simply mean doc-
umentation of the calculations and presentation of the pre-
diction and prediction uncertainty. Rather, effective commun-
ication requires consideration of the knowledge and concerns
of the likely audience. The analyst must then describe his/
her study so that the audience can comprehend the results,
can understand the study's limitations, and can act (if
necessary) in an informed manner As a rule, this means
that the analyst should complecely describe procedural
limitations and assumptions made in conducting the study
112
-------
Beyond that, the analyst should explain how the limitations
and assumptions affect the interpretation of the results for
planning. As a related issue, the analyst should justify
his/her choice of export coefficients. A comprehensive dis-
cussion of the application of the modeling/uncertainty
analysis methodology that meets the needs of the intended
audience facilitates good water quality management planning.
113
-------
REFERENCES
1. Alberts, E. E., Schuman, 6. E., and Burwell, R. E. 1978. Seasonal Run-
off Losses of Nitrogen and Phosphorus from Missouri Valley Loess Watersheds.
J. Environ. Qua!. 7(2):203-208.
2, Allum, M. 0., Glessner, R. E., and Gakstatter, J. H. 1977. An Evaluation
of the National Eutrophication Survey Data Working Paper No. 900. Corvallis
Environmental Research Laboratory, Con/all is, Oregon, National Eutrophication
Survey, Office of Research and Development, U.S. Environmental Protection
Agency, GPO 699-440.
3. Andren, A. W., Eisenreich, S. J., Elder, F. C., Murphy, T. J., Sanderson,
M., and Vet, R. J. 1977. Atmospheric Loading to the Great Lakes: A
Technical Note Prepared for the International Reference Group on Pollution
of the Great Lakes from Land Use Activities - International Joint Commission.
4. Armstrong, D. E., Perry, J. R., and Flatness, D. 1979. Availability of
Pollutants Associated with Suspended or Settled River Sediments Which Gain
Access to the Great Lakes. Final Report on EPA Contract No. 68-01-4479.
Water Chemistry Laboratory, University of Wisconsin, Madison.
5. Aubertin, G. M., and Patric, J. H. 1974. Water Quality After Clearcutting
a Small Watershed in West Virginia. J. Environ. Quality 3(3)243-349.
6. Avadhanula, M. R. 1979. Pollution from Rural, Transportation, Extractive,
and Undisturbed Land Uses in the Ground and Saugeen Watersheds. PLUARG
Technical Report Series, 60 pp.
7. AVCO Economic Systems Corp. 1970. Storm Water Pollution from Urban Land
Activity. FWQA, WPC Series 11034, FKL 07/70.
8. Beaulac, M. N. 1980. Sampling Design and Nutrient Export Coefficients:
An Examination of Variability Within Differing Land Uses. Master of
Science Thesis, Dept. of Resource Development, Michigan State Univ.,
East Lansing.
114
-------
9. Bedient, P. G., Harned, D. A., and Characklis, VI. G. 1978. Stormwater
Analysis and Prediction in Houston. J. Env. Eng. Div. ASCE 104:1087-
1100.
10. Benjamin, J. R., and Cornell, C. A. 1970. Probability Statistics, and
Decision for Civil Engineers. McGraw-Hill, New York.
11. Bennett, E. R., and Linstedt, E. K. 1975. Individual Home Wastewater
Characterization and Treatment. Colorado State Univ., Env. Resources
Ctr., Report #66.
12. Bent, P. C. 1971. Influence of Surface Glacial Deposits on Streamflow
Characteristics. United States Department of the Interior, Geological
Survey, Water Resources Division.
13. Bernhard, A. P. 1975. Return of Effluent Nutrients to the Natural Cycle
Through Evapotranspiration and Subsoil - Infiltration of Domestic Waste-
water. In: Proc. National Home Sewage Disposal Symposium, ASAE.
175-181 pp.
14. Betson, R. P. 1978. Bulk Precipitation and Streamflow Quality Relation-
ships in an Urban Area. Water Resources Research, 14(6):1165-1169.
15. Bormann, F. H., and Likens, G. E. 1967. Nutrient Cycling. In: Science,
155:424-429.
16. Bouldin, D. R., Johnson, A. H., and Lauer, D. A. 1975. The Influence
of Human Activity on the Export of Phosphorus and Nitrate from Fall Creek.
In: Nitrogen and Phosphorus - Food Production, Waste, and the Environ-
ment, K. S. Porter, ed., Ann Arbor Science, Ann Arbor, Michigan.
17. Bradford, R. R. 1974. Nitrogen - Phosphorus Losses from Agronomy Plots
in North Alabama. U.S. Environmental Protection Agency, EPA-660/2-74-033.
115
-------
18. Brown, G. W.5 Gahler, A. R., and Marston, R. B. 1973. Nutrient Losses •
After Clear-cut Logging and Slash Burning in the Oregon Coast Range.
Water Resources Research 9(5)1450-1453.
19. Bryan, E. H. 1970. Quality of Stormwater Drainage from Urban Land
Areas in North Carolina. Univ. N. Carolina, WRRI Report #37.
20. Burton, T. M., Turner, R. R. and Harriss, R. C. 1977. Nutrient Export
from Three North Florida Watersheds in Contrasting Land Use. In:
Watershed Research in Eastern North America, A Workshop to Compare
Results. February 25 - March 3, 1977.
21. Burwell, Schuman, Heinemann and Spomer. 1977. Nitrogen and Phosphorus
Movement from Agricultural Watersheds. J. Soil and Water Cons.,
32(5):226-230.
22. Burwell, R. E., Schuman, G. E., Piest, R. F., Spomer, R. G., and McCalla,
T. M. 1974. Quality of Water Discharged from Two Agricultural Water-
sheds in Southwestern Iowa. Water Resources Research 10(2):359-365.
23. Burwell, R. E., Timmons, D. R., and Holt, R. F. 1975. Nutrient Transport
in Surface Runoff as Influenced by Soil Cover and Seasonal Periods.
Soil. Sci. Soc. Amer. Proc. 39(3):523-528.
24. Campbell, K. L. 1978. Pollution in Runoff from Nonpoint Sources. Univ.
of Florida WRRC Report #42.
25. Chan, H. T. 1978. Contamination of the Great Lakes by Private Wastes.
PLUARG Technical Report Series, 269 pp.
26. Chapra, S. C. 1977. Total Phosphorus Model for the Great Lakes. J. Env.
Eng. Div., ASCE 103(2):147-161.
27. Chapra, S. C., and Reckhow, K.-H. 1979. Expressing the Phosphorus
Loading Concept in Probabilistic Terms. J. Fish. Res. Board of Canada.
36(2):225-229.
116
-------
28. Chichester, F. W., Van Keurens R. W.s and McGuinness, J. C. 1979.
Hydrology and Chemical Quality of Flow from Small Pastured Watersheds:
II Chemical Quality. J. Environ. Qua!. 8(2):167-171.
29. Clarke, R. N., Gilbertson, C. B., and Duke, H. R. 1975. Quantity and
Quality of Beef Feedyard Runoff in the Great Plains. In: Managing
Livestock Wastes, ASAE Proc. 275, St. Joseph, Michigan. 429-431 pp.
30. Colston, N. V. 1974. Characterization and Treatment of Urban Land
Runoff. U.S. Environmental Protection Agency, EPA 670/2-74-096.
31. Converse, J. C., Bubenzer, G. W., and Paulson, W. H. 1976. Nutrient
Losses in Surface Runoff from Winter Spread Manure. Trans. ASAE,
19:517-519.
32. Coote, D. R., and More, F. R. 1978. Pollution Potential of Cattle
Feedlots and Manure Storages in the Canadian Great Lakes Basin. PLUARG
Technical Report Series, 89 pp.
33. Coote, D. R., MacDonald, E. M., and Dickinson, W. T., eds. 1978.
Agricultural Watershed Studies in the Canadian Great Lakes Drainage
Basin; Final Summary Report. PLUARG Technical Report Series, 78 pp.
34. Correll, D. L., Wu, T. L., Fnebele, E. S., and Miklas, J. 1978.
Nutrient Discharge from Rhode River Watersheds and Their Relationships
to Land Use Patterns. In: Watershed Research in Eastern North America.
A workshop to compare results. Vol. 1, Feb. 28 - Mar. 3, 1977.
35. Cowen, W. F., and Lee, G. F. 1976a. Phosphorus Availability in Parti-
culate Materials Transported by Urban Runoff. J. Water Poll. Control
Fed. 48:580-591.
36. Cowen, W. F., and Lee, G. F. 1976b. Algal Nutrient Availability and
Limitation in Lake Ontario during IFYGL, Part I, U.S. Environmental
Protection Agency, EPA 600/3-76-094a.
117
-------
37. Cowen, W. F. 1974. Algal Nutrient Availability and Limitation in Lake f
Ontario during IFYGL. Ph.D. Thesis, University of Wisconsin, Madison.
38. Delumyea, R. G., and Pete!, R. L. 1977. Atmospheric Inputs of Phosophorus
to Southern Lake Huron. April-October, 1975. U.S. Environmental Pro-
tection Agency. EPA-600/3-77-038.
39. Dillon, P. J., and Kirchner, W. B. 1975. The Effects of Geology and
Land Use on the Export of Phosphorus from Watersheds. Water Research
9:135-148.
40. Dillon, P. J., and Reid, R. A. 1980. The Input of Biologically Avail-
able Phosphorus by Precipitation to Precambrian Lakes. (Unpublished
manuscript).
41. Dillon, P. J., and Rigler, F. H. 1974. A Test of a Simple Nutrient Budget
Model Predicting the Phosphorus Concentration in Lake Water. J. Fish Res.
Board Can., 31(11).1711-1778.
42. Dillon, P. J., and Rigler, F. H. 1975. A Simple Method for Predicting
the Capacity of a Lake for Development Based on Lake Trophic Status.
J. Fish Res. Board of Canada. 32(9):1519-1531.
43. Dornbush, J. N., and Madden, J. M. 1973. Pollution Potential of Runoff
from Production Livestock Feeding Operations in South Dakota. S. Dakota
State Univ. WRRI Report #A-025-SDAK.
44. Duffy, P. D., Schreiber, J. D., McClurkin, D. C., and McDowell, L. L.
1978. Aqueous- and Sediment-phase Phosphorus Yields from Five Southern
Pine Watersheds. J. Environ. Qual. 7(1):45-50.
45. Edwards, W. M., Simpson, E. C., and Frere, M. H. 1972. Nutrient Content
of Barnlot Runoff Water. J. Environ. Qual. 1(4):401-405.
118
-------
46. Eisenreich, S. J., Emmling, P. J., and Beeton, A. M. 1977. Atmospheric
Loading of Phosphorus and other Chemicals to Lake Michigan. Internat.
Assoc. Great Lakes Res. 3(3-4):291-304.
47. Ellis, B. G., and Childs, K. E. 1973. Nutrient Movement from Septic
Tanks and Lawn Fertilization. Michigan Dept. of Natural Resources,
Technical Bulletin No. 73-5. Lansing, Michigan.
48. Fredricksen, R. L. 1972. Nutrient Budget of a Douglas-Fir Forest on
an Experimental Watershed in Western Oregon. In: J. F. Franklin, C.
J. Dempster, and R. H. Waring (eds.) Proc.-Research on Coniferous Forest
Ecosystems - A Symposium. Bellingham, Washington. Forest Service, U.S.
Dept. Agriculture, Portland, Oregon.
49. Fredriksen, R. L. 1979. Unpublished data, Pacific Northwest Forest and
Range Experiment Station, Portland, Oregon. (Typewritten).
50. Fredriksen, R. L., et al., 1975. The Impact of Timber Harvest, Fertili-
zation, and Herbicide Treatment on Streamwater Quality in Western Oregon
and Washington. In: Forest Soils and Land Mangement: Proc. 4th North
American Forest Soils Conference held at Laval University, Quebec City,
August, 1973.
51. Frissel, M. J., (ed.). 1978. Cycling of Mineral Nutrients in Agri-
cultural Ecosystems. Elsevier Scientific Publishing Co., New York,
New York.
52. Gilbertson, C. B., Ellis, J. R., Nieneber, J. A., McCalla, T. M., and
Klopfenstein, T. J. 1975. Physical and Chemical Properties of Outdoor
Beef Cattle Feedlot Runoff. Univ. Nebr. Agricultural Exp. Station
Bulletin #271.
53. Goettle, A. 1978. Atmospheric Contaminants, Fallout and their Effects
on Storm Water Quality. Prog. Wat. Tech. 10(5):455-467.
119
-------
54. Gosz, J. R. 1978. Nitrogen Inputs to Stream Water from Forests along
an Elevational Gradient in New Mexico. Water Research 12:725-734.
55. Gregory, K. J., and Walling, D. E. 1973. Drainage Basin Form and
Process. John Wiley and Sons, New York. 465 pp.
56. Griffin, D. W., Grizzard, T. J., Randall, C. W., and Hartigan, J. P.
1978. An Examination of Nonpoint Pollution Export from Various Land
Use Types. Presented at: International Symposium on Urban Stormwater
Management, Univ. Kentucky, Lexington, Kentucky, July 24-27, 1978.
57. Grizzard, T. J., Hartigan, J. P., Randall, C. W., Kim, J. I., Smullen,
J. T., and Derewianka, M. 1977. Assessing Runoff Pollution Loading
for 208 Planning Programs. Presented at: ASCE National Environmental
Engineering Conference, Nashville, Tenn. July 13-15, 1977.
58. Grizzard, T. J., Randall, C. W., Hoehn, R. C., and Saunders, K. G.
1978. The Significance of Plant Nutrient Yields in Runoff from a
Mixed Land Use Watershed. Prog. Wat. Tech. 10(5/6)-577-596.
59. Haines, E. B. 1976. Nitrogen Content Acidity of Rain on the Georgia
Coast. Water Resources Bulletin. 12(6):1223-1231.
60. Harms, L. L., Dornbush, J. N., and Andersen, J. R. 1974. Physical and
Chemical Quality of Agricultural Land Runoff. Journ. Water Poll. Contr.
Fed. 46(11):2460-2470.
61. Hart, W. C., and Ogden, J. C. 1977. Precipitation and Nutrient Export
from a Small Coastal Ecosystem in Nova Scotia. In: Watershed Research
in Eastern N. America, A Workshop to Compare Results. Vol. 1. Feb. 28 -
March 3, 1977.
62. Heaney, J. P., and Sullivan, R. H. 1971. Source Control of Urban Water
Pollution. J. Water Poll. Contr. Fed. 43(4):571-578.
120
-------
63. Henderson, G. S., and Harris, W. F. 1975. An Ecosystem Approach to
Characterization of the Nitrogen Cycle in a Deciduous Forest Watershed.
In: B. Bernier and C. H. Wingett (eds.). Forest Soils and Forest
Land Management. Les Presses de 1'Universite Laval, Quebec, Canada.
179-193 pp.
64. Henderson, G. S., Hunley, A., and Selvidge, W. 1977. Nutrient Discharge
from Walker Branch Watershed. In: Watershed Research in Eastern North
America. A workshop to compare results. Vol. I, Feb. 28 - Mar. 3, 1977.
65. Hensler, R. F., Olsen, R. J., Witzel, S. A., Attoe, 0. 0., Paulson, W. H.,
and Johannes, R. F. 1970. Effect of Method of Manure Handling on Crop
Yields, Nutrient Recovery and Runoff Losses. Trans. ASAE 13:726-731.
66. Hetling, L. J., Carlson, G. A., and Bloomfield, J. A. 1976. Estimation
of the Optimal Sampling Interval in Assessing Water Quality of Streams.
In: Proceedings of the Conference on Environmental Modeling and Simula-
tion, W. R. Ott, ed., U.S. Environmental Protection Agency, Washington,
D.C., EPA 600/9-76-016.
67. Hoeft, R. G., Keeney, D. R., and Walsh, L. M. 1972. Nitrogen and Sulfur
in Precipitation and Sulfur Dioxide in the Atmosphere in Wisconsin.
J. Environ. Qua!. 1(2):203-208.
68. Holland, M. E. 1969. Runoff from Forest and Agricultural Watersheds.
Colorado State University, Natural Resources Center Completion Report
Series #4.
69. Hollis, G. E. 1975. The Effect of Urbanization and Floods of Different
Recurrence Internal. Water Resources Research, 11(6):431.
70. Ikuse, T., Mimura, A., Takeuchi, S., and Matsuchita, J. 1975. Effects
of Urbanization on Run-off Characteristics. In: Publication n° 117 de
1'Association Internationale des Sciences. Hydrologiques Symposium de
Tokyo. December, 1975.
121
-------
71. Kilmer, V. J., Gillian, J. W., Lutz, J. F., Joyce, R. T., and Eklund,
C. D. 1974. Nutrient Losses from Fertilized Grussed Watersheds in
Western North Carolina. J. Environ. Quality 3(3):214-219.
72. King, D. L. 1979. Lake Measurements. In: Lake Restoration: Pro-
ceedings of a National Conference. U.S. Environmental Protection Agency,
EPA-440/5-79-001.
73. Kissel, D. E., Richardson, C. W. and Burnett, E. 1976. Losses of
Nitrogen in Surface Runoff in the Black!and Prairie of Texas. J.
Environ. Qua!. 5(3):288-292.
74. Klausner, S. D., Swerman, P. J., and Ellis, D. F. 1974. Surface
Runoff Losses of Soluble Nitrogen and Phosphorus Under Two Systems of
Soil Management. J. Environ. Qua!. 3(l):42-46.
75. Klausner, S. D., Zwernian, P. J., and Ellis, D. F. 1976. Nitrogen and
Phosphorus Losses from Winter Disposal of Dairy Manure. J. Environ.
Qua!. 5(l):47-49.
76. Kluesener, J. W. 1972. Nutrient Transport and Transformation in Lake
Wingra, Wisconsin. Ph.D. Thesis, Water Chemistry Dept., Univ. of
Wisconsin, Madison, Wis.
77. Kluesener, J. W.9 and Lee, G. F. 1974. Nutrient Loading from a Separate
Storm Sewer in Madison, Wisconsin. J. Water Poll. Contr. Fed. 46(5):
920-936.
78. Konrad, J. G., Chesters, G., and Bauer, K. W. 1978. Menomonee River
Basin, Wisconsin; Summary Pilot Watershed Report. PLUARG Technical
Report Series, 77 pp.
122
-------
79. Krebs, J. E., and Golley, F. B. 1977. Budget of Selected Mineral Nutrients
for Two Watershed Ecosystems in the Southeastern Piedmont. NTIS PB 272
286.
80. Laak, R. 1975. Relative Pollution Strengths of Undiluted Waste Materials
Discharged in Households and the Dilution Waters Used for Each. Manual
of Grey Water Treatment Practice - Part II, Monogram Inudstries, Inc.,
Santa Monica, California.
81. Lake, J., and Morrison, J. 1977. Environmental Impact of Land Use on
Water Quality. U.S. Environmental Protection Agency, EPA 905/9-77/007-B.
82. Landon, R. J. 1977. Characterization of Urban Stormwater Runoff in
the Tri-County Region. 208 Water Quality Management Plan. Michigan Tri-
County Regional Planning Commission. 136 pp.
83. Larsen, D. P., and Mercier, H. T. 1976. Phosphorus Retention Capacity
of Lakes. J. Fish Res. Board Can., 33(8):1742-1750.
84. Lee, G. F., Jones, R. A., and Rast, W. 1979. Availability of Phosphorus
to Phytoplankton and its Implications for Phosphorus Management Strategies.
Proc. IOC/Cornell University Conference on Phosphorus Management Strategies
for the Great Lakes, 1979.
85. Leopold, L. B. 1968. Hydrology for Urban Land Planning - A Guidebook on
the Hydrologic Effects of Urban Land Use. U.S. Geol. Survey Circular
#554.
86. Liebeskind, A., Welch, D., Moore, L., Lawrence, T., Parson, R., and Leach,
J. 1978. A Study of the Phosphorus Flux of Higgins Lake and the Effect
on the Water Quality. Unpublished manuscript. Dept. of Resource Develop-
ment, Michigan State University, East Lansing.
123
-------
87. Ligman, K., Hutzler, N.5 and Boyle, W. C. 1974. Household Wastewater
Characterization. J. Env. Eng. Div.s ASCE, 150(EEI):201-213.
88. Likens, 6. E. 1972. The Chemistry of Precipitation in the Central
Finger Lakes Region. Cornell Umv. Water Resources and Marine Sciences
Center Tech. Dept. Report #50.
89. Likens, 6. E., Bormann, F. H., Pierce, R. S., Eaton, J. S., and Johnson,
N. M. 1977. Bio-Geo-Chemistry of a Forested Ecosystem. Springer-Verlag,
Inc., New York, 146 pp.
90. Likens, 6. E.s and Loucks, 0. L. 1978. Analysis of Five North American
Lake Ecosystems. Ill Sources, Loading and Fate of Nitrogen and Phosphorus.
Verh. Internat. Verein. Limnol. 20:568-573.
91. Lindh, G. 1972. Urbanization: A Hydrological Headache. AMBIO 1(16)-
185-201.
92. Logan, T. J., Verhoff, F. H., and DePinto, J. V. 1979. Biological
Availability of Total Phosphorus. Technical Series Report, Lake Erie
Wastewater Management Study, U.S. Army Corps of Engineers, Buffalo, New
York.
93. Long, F. L. 1979. Runoff Water Quality as Affected by Surface-applied
Dairy Cattle Manure. J. Environ. Qua!., 8(2):215-218.
94. Magdoff, F. R., Amadon, J. F., Goldberg, S. P., and Wells, G. D. 1977.
Runoff from a Low-Cost Manure Storage Facility. TRANS. ASAE 20(4):658-665.
95. Marsalek, J. 1975. Sampling Techniques in Urban Runoff Quality Studies,
Water Quality Parameters. ASTM STP 573. American Society for Testing
and Materials, pp. 526-542.
124
-------
96. Martin, C. W. 1979. Precipitation and Streamwater Chemistry in an
Undisturbed Forested Watershed in New Hampshire. Ecology, 60(1):36-42.
97. Mattraw, H. C., and Sherwood, C. B. 1977. Quality of Storm-water Runoff
from a Residential Area, Broward County, Florida. Jour. Research U.S.
Geol. Survey, 5(6):823-834.
98. McCalla, T. M., Ellis, J. R., and Gilbertson, C. B. 1972. Chemical
Studies of Solids, Runoff, Soil Profile and Groundwater from Beef Cattle
Feedlot at Mead, Nebraska. Proc., 1972 Cornell Waste Management Conference,
Syracuse, New York, 211-223 pp.
99. McColl, J. G., and Bush, D. S. 1978. Precipitation and Throughfall
Chemistry in the San Francisco Bay Area. J. Environ. Qua!. 7(3):352-357.
100. McDowell, L. L., Ryan, M. E., McGregor, K. C., and Greer, J. D. Nitrogen
and Phosphorus Losses in Runoff from No-till Soybeans. Presented Paper
#78-2508. 1978 Winter Meeting ASAE, Palmer House Hotel, Chicago, Illinois,
Dec. 18-20, 1978.
101. McGill, R. J., Tukey, J. W., and Larsen, W. A. 1978. Variations of Box
Plots. Am. Stat. 32:12-16.
102. Menzel, R. G., Rhoades, E. D., Olness, A. E., and Smith, S. J. 1978.
Variability of Annual Nutrient and Sediment Discharges in Runoff from
Oklahoma Cropland and Rangeland. J. Environ. Qua!. 7(3):401-406.
103. Meyer, J. L., and Likens, G. E. 1979. Transport and Transformation of
Phosphorus in a Forest Stream Ecosystem. Ecology 60(6):1255-1269.
104. Miklas, J., Wu, T. L., Hiatt, A., and Correll, D. L. 1977. Nutrient
Loading of the Rhode River Watershed via Land Use Practice and Pre-
cipitation. In: Watershed Research in Eastern N. America, A Workshop
to Compare Results. Feb. 28 - Mar. 3, 1977.
125
-------
105. Minshall, N. E., Witzel, S. A., and Nichols, M. S. 1970. Stream Enrich-
ment from Farm Operations. J. San Engr. Div., Amer. Soc. Chem. Engineers
96(SA2):513-524.
106. Moore, D. 6. 1970. Forest Fertilization and Water Quality in the Pacific
Northwest. Am. Soc. Agron. Abstr. 1970:160-161.
107. Moore, D. G. 1975. Effects of Forest Fertilization with Urea on Stream
Quality - Quilcene Ranger District, Washington. USDA For. Serv. Res.
Note PNW-241. 9 pp.
108. Moore, W. L., and Morgan, C. W. 1969. Effects of Watershed Changes on
Streamflow. Univ. Texas Press, Austin, Texas, 289 pp.
109. Mosteller, F., and Tukey, J. W. 1977. Data Analysis and Regression:
A Second Course in Statistics. Addison-Wesley Publishing Co., Reading,
Mass.
110. Much, R. R., and Kemp, 6. 1978. Characterization of Non-Point Waste
Sources, Report #6, Summary of Non-Point Waste Loads. Fox Valley Water
Quality Planning Agency, Wisconsin. 135 pp.
111. Murphy, T. J., and Doskey, P. V. 1976. Inputs of Phosphorus from Pre-
cipitation to Lake Michigan. J. Great Lakes Research 2(1):60-70.
112. Murphy, T. J., and Doskey, P. V. 1975. Inputs of Phosphorus from Pre-
cipitation to Lake Michigan. U.S. Environmental Protection Agency,
EPA 600/3-75-005.
113. Nelson, D. W., Monke, E. J., Bottcher, A. D., and Sommers, L. E. 1978.
Sediment and Nutrient Contributions to the Maumee River from an Agri-
cultural Watershed. In: Best Management Practices for Agriculture and
Silviculture. Proceedings of the 1978 Cornell Agricultural Waste Manage-
ment Conference, 491-505 pp.
126
-------
114. Nicholaichuk, W., and Read, D. W. 1978. Nutrient Runoff from Fertilized
and Unfertilized Fields in Western Canada. J. Environ. Qua!., 7(4):542-544.
115. Nicholson, J. A. 1977. Forested Watershed Studies, Summary Technical
Report. PLUAR6 Technical Report Series, 23 pp.
116. Okuda, A. 1975. Change in Runoff Patterns Due to Urbanization of River
Basin. In: Publication n° 117 de 1'Association Luternationale de Sciences
Hydrologiques Symposium de Tokyo (Dec. 1975).
117. Olness, A., Rhoades, E. D., Smith, S. V., and Menzel, R. 6. 1980.
Fertilizer Nutrient Losses from Range!and Watersheds in Central Oklahoma.
J. Environ. Qual., 9(l):81-85.
118. Olsson, E., Karlgren, L., and Tullander, V. 1968. Household Wastewater.
The National Swedish Inst. for Bldg. Research, Box 26 163-102 52 Stockholm
27, Sweden.
119. O'Neill, J. E. 1979. Pollution from Urban Land Use in the Grand and
Saugeen Watersheds. PLUAR6 Technical Report Series, 55 pp.
120. Otis, R. J., Boyle, W. C., and Sauer, D. K. 1975. The Performance of
Household Wastewater Treatment Units Under Field Conditions. Proc.
National Home Sewage Disposal Symposium, ASAE. pp~ T91-201.
^
121. Patni, N. K., and Hore, F. R. 1978. Pollutant Transport to Subsurface
and Surface Waters in an Integrated Farm Operation. PLUARG Technical
Report Series, 79 pp.
122. Pollard, R. W., Sharp, M. H., and Madison, F. W. 1979. Farmers'
Experience with Conservation Tillage: A Wisconsin Survey. Journ. Soil
and Water Conserv. 34(5):215-219.
127
-------
123. Pritchett, W. C. 1979. Properties and Management of Forest Soils.
John Wiley and Sons, New York, 500 pp.
124. Progressive Engineering Consultants. 1976. Facilities Plan for Lyon,
Markey and Garrfsh Townships. Grand Rapids, Michigan.
125. Randall, C. W., Helsel, D. R., Gnzzard, T. J., and Hoehn, R. C. 1978.
The Impact of Atmoshperic Contaminants on Storm Water Quality in an Urban
Area. Prog. Water Tech. 10(5):417-431.
126. Reckhow, K. H. 1979a. Empirical Lake Models for Phosphorus: Development,
Applications, Limitations, and Uncertainty, in Perspectives on Lake Eco-
system Modeling, edited by D. Scavia and A. Robertson, Ann Arbor Science
Publishers, Ann Arbor, Mich.
127. Reckhow, K. H. 1979b. Sampling Design for Lake Phosphorus Budgets. In:
Proceedings of the American Water Resources Assocation Symposium on the
Establishment of Water Quality Monitoring Programs, San Francisco, Calif.
128. Reckhow, K. H. 1979c. Quantitative Techniques for the Assessment of
Lake Quality. U.S. Environmental Protection Agency, EPA-440/5-79-015.
129. Reckhow, K. H. 1979d. Uncertainty Analysis Applied to Vollenweider's
Phosphorus Loading Criterion. J. Water Pollut. Control Fed., 51(8):
2123-2128.
130. Reckhow, K. H. 1980. Techniques for Exploring and Presenting Data Applied
to Lake Phosphorus Concentration. Can J. Fish. Aquat. Sci., 37:290-294.
131. Reckhow, K. H., and Chapra, S. C. 1980. Engineering Appraoches for Lake
Management: Data Analysis and Modeling. Ann Arbor Science, Ann Arbor,
Mich. (In press).
128
-------
132. Reckhow, K. H.s and Chapra, S. C. 1979. The Need for Simple Approaches
for the Estimation of Lake Model Prediction Uncertainty. Paper presented
at the International Institute for Applied Systems Analysis Task Force
Meeting on "Uncertainty and Forecasting of Water Quality." Laxenburg,
Austria.
133. Reckhow, K. H., and Simpson, J. T. 1980. A Procedure Using Modeling and
Error Analysis for the Prediction of Lake Phosphorus Concentration from
Land Use Information. Can J. Fish. Aquat. Sci. (In press).
134. Richardson, C. J., and Merva, G. E. 1976. The Chemical Composition of
Atmospheric Precipitation from Selected Stations in Michigan. Water,
Air and Soil Pollution, 6:373-385.
135. Robinson, E., and Robbins, R. C. 1970. Gaseous Nitrogen Compound Pol-
lutants from Urban and Natural Sources. Journ. Air Poll. Contr. Ass'n.,
20(5):303-306.
136. Rodiek, R. K. 1979. Some Watershed Analysis Tools for Lake Management.
In: Lake Restoration* Proceedings of a National Conference. U.S.X
Environmental Protection Agency, EPA-440/5-79-001.
137. Rogers, J. S., Stewart, E. H., Calvert, D. V., and Mansell, R. S. 1976.
Water Quality from a Subsurface Drained Citrus Grove. Proc. 3rd National
Drainage Symposium, ASAE, Palmer House, Chicago, 111., Dec. 13-14, 1976.
pp. 99-103.
?
138. Sanderson, M. 1977. Agricultural Watershed Studies in the Canadian Great
Lakes Drainage Basin, Precipitation - Quantity and Quality. PLUARG Tech-
nical Report Series, 137 pp.
139. Sawyer, C. N. 1965. Problems of Phosphorus in Water Supplies. J. AWWA,
57:1431.
129
-------
140. Schindler, D. W., Newbury, R. W., Beaty, K. 6., and Campbell, P. 1976.
Natural Water and Chemical Budgets for a Small Precambnan Lake Basin
in Central Canada. J. Fish. Res. Board Can., 33:2526-2543.
141. Schindler, D. W., and Nighswander, J. E. 1970. Nutrient Supply and
Primary Production in Clear Lake, Eastern Ontario. J. Fish. Res. Bd.
Canada, 27:2009-2036.
142. Schneider, I. F., and Erickson, A. E. 1972. Soil Limitations for Disposal
of Municipal Waste Waters. Research Report #195, Farm Science Series,
Michigan State University Agricultural Experiment Station.
143. Schreiber, J. D., Duffy, P. D., and McClurkin, D. C. 1976. Dissolved
Nutrient Losses in Storm Runoff from Five Southern Pine Watersheds. J.
Environ. Qua!., 5(2):201-205.
144. Schuman, 6. E., Spomer, R. 6., and Piest, R. F. 1973. Phosphorus Losses
from Four Agricultural Watersheds on Missouri Valley Loess. Soil Sci.
Soc. Amer. Proc., 37:424-427.
145. Schuman, 6. E. and Burwell, R. E. 1974. Precipitation Nitrogen Contri-
bution Relative to Surface Runoff Discharges. J, Environ. Qua!., 3(4):
366-369.
146. Schuman, 6. E., Burwell, R. E., Piest, R. F., and Spomer, R. 6. 1973.
Nitrogen Losses in Surface Runoff from Agricultural Watersheds on Missouri
Valley Loess. J. Environ. Qua!., 2(2):299-302.
147. Sherwani, J. K., and Moreau, D. H. 1975. Strategies for Water Quality
Monitoring. Univ. North Carolina WRRI Report #398.
148. Siegrist, R. L. 1977. Waste Segregation to Facilitate Onsite Wastewater
Disposal Alternatives. Proceedings of the Second National Home Sewage
Treatment Symposium, Chicago, December 12-13.
130
-------
149. Siegrist, R. L., Witt, M., and Boyle, W. C. 1976. Characteristics of
Rural Household Wastewater. J. Env. Eng. Div., ASCE, 102(EE3):533-548.
150. Simpson, D. E., and Hemens, J. 1978. Nutrient Budget for a Residential
Stormwater Catchment in Durham, South Africa. Prog. Wat. Tech. 10(5):631-
643.
151. Singer, M. J., and Rust, R. H. 1975. Phosphorus in Surface Runoff from
a Deciduous Forest. J. Environ. Quality, 4(3)-307-311.
152. Smith, C. N., Leonard, R. A., Langdale, 6. W., and Bailey, 6. W. 1978.
Transport of Agricultural Chemicals from Small Upland Piedmont Watersheds.
U.S. Environmental Protection Agency, EPA-600/3-78-056.
153. Smith, R. V., and Steward, D. A. 1977. Statistical Models of River Loadings
of Nitrogen and Phosphorus in the Lough Neagh System, Water Research 11:631-
636.
154. Snedecor, G. W., and Cochran, W. G. 1973. Statistical Methods. Iowa
State University Press, Ames, Iowa. 593 pp.
155. Sonzogni, W. C., and Chapra, S. C. 1980. Phosphorus Availability - Signifi-
cance to Modeling and Management. Presented paper. Great Lakes-80, 23rd
Conference on Great Lakes Research. Queen's University, Kingston, Ontario.
May 19-22, 1980.
156. Stay, F. S., Malueg, K. W., Austin, R. E., Crouse, M. R., Katko, A., and
Dominquez, S. E. 1978. Ecological Effects of Forest Fertilization with
Urea on Small Western Cascade Streams of Oregon, U.S.A. Verh. Internat.
Limnol., 20:1347-1358.
157. Swank, W. T., and Douglass, J. E. 1974. Streamflow Greatly Reduced by
Converting Deciduous Hardwood Stands to Pine. In: Science, 85:857-859.
131
-------
158. Swank, W. T., and Douglass, J. E. 1977. Nutrient Budgets for Undisturbed
and Manipulated Hardwood Forest Ecosystems in the Mountains of North
Carolina. In: Watershed Research in Eastern North America - A Workshop
to compare Results,February 28 - March 3.
159. Swank, W. T., Soebel, N. B., and Helvey, J. D. 1972. Interception Loss
in Loblolly Pine Stands of the South Carolina Piedmont. J. Soil Water
Conserv., 26:160-163.
160. Swank, W. T., and Henderson, 6. S. 1976. Atmospheric Input of Some Cations
and Anions to Forest Ecosystems in North Carolina and Tennessee. Water
Resources Research 12(3):541-546.
161. Swank, W. T., and Miner, N. H. 1968. Conversion of Hardwood - Covered
Watersheds to White Pine Reduces Water Yield. Water Resources Research
4(5):947-954.
162. Switzer, 6. L., and Nelson, L. E. 1972. Nutrient Accumulation and Cycling
in Loblolly Pine Plantation Ecosystems: The First Twenty Years. Soil
Sci. Soc. Am. Proc., 36:143-147.
163. Sylvester, R. 0. 1960. Nutrient Content of Drainage Water from Forested,
Urban and Agricultural Areas. Tech. Rep. W61-3, Trans. 1960 Seminar on
Algae and Metropolitan Wastes, Robert A. Taft San. Engr. Ctr., Cincinnati,
Ohio. 80-87 pp.
164. Taylor, A. W., Edwards, W. M., and Simpson, E. C. 1971. Nutrients in
Streams Draining Woodland and Farmland Near Coshocton, Ohio. Water
Resources Research, 7(l):81-89.
165. Thomas, N. A., Robertson, A., and Sonzogni, W. C. 1979. Review of Control
Objectives: New Target Loads and Input Controls. Proc. IJC/Cornell
University Conference on Phosphorus Management Strateaies for the Great
Lakes.
132
-------
166. Tilstra, J. R., Malueg, K. W., Larson, W. C. 1972. Removal of Phosphorus
and Nitrogen from Wastewater Effluent by Induced Soil Percolation. J.
Water Poll. Cont. Fed., 44(5):796-805.
167. Timmons, D. R., Verry, E. S., Burwell, R. E., and Holt, R. F. 1977.
Nutrient Transport in Surface Runoff and Interflow from an Aspen-Birch
Forest. J. Environ. Qua!., 6(2):188-192.
168. Tofflemire, T. J., and Chen, M. 1977. Phosphate Removal by Sands and
Soils. Ground Water. 15(5):377-387.
169. Treunert, E., Wilhelms, A., and Bernhardt, H. 1974. Einfluss der
Probenetnahme-Haufigkeit auf die Ermittlung der Jahres-Phosphor-
Frachwerte mittlerer Bache. Hydrochem. Hydrogeol. Mitt. 1:175-198.
170. Turner, R. R., Burton, T. M., and Harriss, R. C. 1977. Descriptive
Hydrology of Three North Florida Watersheds in Contrasting Land Use.
In: Watershed Research in Eastern North America, A Workshop to Compare
Results. February 28 - March 3, 1977.
;
171. Unger, U. 1970. Berechung von Stofffrachten in Flussen durch wenige
Einzelanalysen im Vergleich zu kontinuierlichen einjahren chemischen
Untersuchungen, gezeigt am Beispiel des Bodenseezuflusses Argen (1967/68).
Schweiz. Z. Hydro!. 32:453-473.
172. U.S. Environmental Protection Agency. 1974. Nitrogen and Phosphorus in
Wastewater Effluents, Working Paper No. 22. National Eutrophication
Survey, Pacific Northwest Environmental Research Laboratory, Con/all is,
Oregon, GPO 697-032.
173. U.S. Environmental Protection Agency. 1975. National Eutrophication
Survey Working Paper on Higgins Lake. U.S.E.P.A. Laboratory, Corvallis,
Oregon.
133
-------
174. U.S. Environmental Protection Agency. 1975. National Eutrophi cation
Survey Methods 1973-1976, Working Paper No. 175. Office of Research
and Development, GPO 699-440.
175. Uttomark, P. D., Chapin, J. D., and Green, K. M. 1974. Estimating
Nutrient Loading of Lakes from Nonpoint Sources. United States
Environmental Protection Agency. EPA-660/1 3-74-020.
176. Vallentyne, J. R. 1974. The Algal Bowl. Dept. of Env. Fisheries and
Marine Service, Ottawa, Canada.
177. Verhoff, H., Yaksich, S. M., and Melfi, D. A. 1980. River Nutrient
and Chemical Transport Estimation. J. Env. Eng. Div., ASCE. 106(3) :591-
608.
178. Verry, E. S. 1979. Nutrient Yields from Forested Areas, (Typewritten).
179. Verry, E. S., and Timmons, D. R. 1977. Precipitation Nutrients in the
Open and Under Two Forests in Minnesota Can. Journal Forest Research
180. Vitousek, P. M. 1977. The Regulation of Element Concentrations in
Mountain Streams in the Northeastern United States. Ecological Mono-
graphs 47:65-87.
181. Vitousek, P. M., and Reiners, W. A., 1975. Ecosystem Succession and
Nutrient Retention: A Hypothesis. Biosci. 25-376-381.
182. Vollenweider, R. A. 1968. The Scientific Basis of Lake and Stream
Eutrophi cati on, with Particular Reference to Phosphorus and Nitrogen
as Eutrophication Factors. OECD Tech., Report, Paris, DAS/CSI/68.21.
183. Vollenweider, R. A. 1975. Input-output Models with Special Reference
to the Phosphorus Loading Concept in Limnology. Schweiz. Z. Hydro!.,
37:53-84.
134
-------
184. Walker, W. C., Bouma, J.5 Keeney, D. R.s and Olcott, P. 6. 1973.
Nitrogen Transformations During Subsurface Disposal of Septic Tank
Effluent in Sands: II. Ground Water Quality. J. Environ. Qua!.
2:521-525.
185. Walker, W. W., Jr. 1977. Some Analytical Methods Applied to Lake
Water Quality Problems. Ph.D. dissertation, Harvard University,
Cambridge, Mass. 528 pp.
186. Wallman, H., and Cohen, S. 1974. Demonstration of Waste Flow Reduction
from Households. U.S. Environmental Protection Agency. EPA-670/2-77-071.
187. Weibel, S. R., Anderson, R. J., and Woodward, R. L. 1964. Urban Land
Runoff as a Factor in Stream Pollution. J. Water Poll. Con. Fed.
36:914-924.
188. Wells, C. G. 1971. Effects of Prescribed Burning on Soil Chemical
Properties and Nutrient Availability. In: Proc. Prescribed Burning
Symp. USDA For. Serv., SE For. Expt. Sta., Asheville. 86-99 pp.
189. Wells, C. G., and Jorgensen, J. R. 1975. Nutrient Cycling in Loblolly
Pine Plantations. In: B. Bermer and C. H. Winget (eds.) Forest
Soils and Forest Land Management. Les Presses de Tuniversite Laval,
Quebec. 137-158 pp.
190. Wells, C., Whigham, D. and Lieth, H. 1972. Investigation of Mineral
Nutrient Cycling in Upland Piedmont Forest J. Elisha Mitchell Sci.
Soc. 88:66-78.
191. Whipple, W., Hunter, J. V., Ahlert, R. C., and Yu, S. L. 1978.
Estimating Runoff Pollution from Large Urban Areas - the Delaware
Estuary. Rutgers University WRRI Report.
135
-------
192. Wright, R. F. 1976. The Impact of Forest Fire on the Nutrient Influences
to Small Lakes in Northeast Minnesota. Ecology 57(4):649-663.
193. Yoshino, F. 1975. Runoff Characteristics of Small Urbanized Areas.
In: Publication No. 117 de I1Association Internationale des Sciences
»
Hydrologiques Symposium de Tokyo (Dec. 1975).
194. Young, R. A., and Holt, R. F. 1977. Winter-applied Manure: Effects
on Annual Runoff, Erosion, and Nutrient Movement. J. Soil and Water
Cons. 32(5):219-222.
195. Young, T. C., Depinto, J. V., Flint, S. E., Switzenbaum, M. S., and
Edzwald, J. K. 1980. Bioavailability of Phosphorus in Municipal Waste-
waters. Presented paper. Great Lakes - 80, 23rd Conference on Great
Lakes Research. Queen's University, Kingston, Ontario. May 19-22, 1980.
136
-------
APPENDIX
A. 1 Research Methodology for the Assessment of Nutrient Runoff in Field
Studies
During his literature survey of nutrient export coefficients, Beaulac
(1980) found considerable variability in the methods employed by researchers
for experimental design, sampling design, mass flux estimation, and result
reporting. Unfortunately the lack of a standard methodology resulted in the
rejection of some reported values for this manual. Since it is likely that
some readers of this manual will at some time be involved in studies designed
to directly measure nutrient mass transport to surface water bodies, research
and reporting methods are discussed below This is particularly important
since a premise supporting this manual is that export coefficients are trans-
ferable among selected watersheds. Researchers are urged to adopt certain
standard procedures so that their results may be added to the literature on
nutrient export coefficients.
A. 1. 1 Watershed Designs
Of the criteria necessary for a nonpoint source monitoring program, the
sampling location, or more importantly, the watershed design, is crucial for
accurate estimation of nutrient yields. To facilitate the sampling site/
design selection process, two key interrelated factors are involved, the
specific objective of the network design and the representativeness of the
sample to be collected. To accommodate these factors, two basic approaches
to diffuse load assessment are, in turn, available.
137
-------
The first approach involves sample collection from relatively large
streams draining large watersheds. If storm and seasonal hydro!ogic response
are routinely sampled throughout the year, an accurate representation of
total annual nutrient flux from particular drainage basins can be obtained.
This approach has been extensively used to obtain estimates of Great Lakes
tributary loads by the Pollution From Land Use Activities Reference Group
(PLUARG) associated with the International Joint Commission.
A number of disadvantages to this approach have been noted (Whipple et
a!., 1978). First, many large streams, particularly in urban areas, include
inputs from industrial and municipal point sources, so that total loading
does not relate directly to pollution from storm water runoff. Second, sub-
traction of known point loads from total yield can result in a biased diffuse
load estimate. This occurs because the magnitude of reactions such as sedi-
ment attenuation, nutrient uptake and degradation by bioseston are not accu-
rately accounted for at the downstream sampling site. Since point sources
determined at their end-of-pipe source do not undergo these transformation
processes, their subtraction from total loads may result in an underestima-
tion of diffuse source contributions. (Alternatively if there is no net
accumulation of material in the stream, over a sufficiently long time period
all phosphorus discharged will reach the lake. In the steady state, this
suggests no bias from point source subtraction.)
Third, the land use of large watersheds is very often mixed, in propor-
tions which vary from one tributary to the next. This makes it difficult if
not impossible to determine the percent loading contribution from each land
use, and application of the results to other watersheds for prediction purposes
remains questionable.
If the objective of the sampling design is to describe runoff loads from
specific perturbations, representativeness will depend on a comprehensive
138
-------
approach. This second approach is more specific and is based on the examina-
tion of drainage from catchment basins which define a particular land use
In order to maintain homogeneity, the monitored watersheds are relatively
small (except for some forested systems).
The advantages to this approach are essentially two-fold First, land
use - water quality relationships are more carefully defined allowing for
contrasts between natural and manipulated ecosystems. By comparison this
can provide information about the functional efficiency and "health" of a
particular land use For instance, is a particular land use conservative
of nutrient inputs (forests) or is the assimilation capacity limited (pasture)
or exceeded (feedlots)7 Second, the results can be used in conjunction with
other similar studies to predict future water quality changes corresponding
to projected land alterations *
Because of the identified advantages, a large percentage of nonpoint
source water quality investigations have utilized this latter approach with
forest, agricultural and urban activities as the major land use categories
studied. The remainder of this subsection contains a discussion on how diffuse
runoff is monitored from each of these land use types.
forest land use
In order to provide hydrologic and nutrient flux information from
natural (undisturbed) ecosystems, a number of experimental forested water-
sheds have been established across a wide range of climates, geology and bio-
logical structure. Some of the well-known watersheds are Hubbard Brook Experi-
mental Forest in New Hampshire, Walker Branch Watershed in Tennessee, H. J.
Andrews Experimental Forest in Oregon and Coweeta Hydrologic Laboratory
in North Carolina.
139
-------
Although biological (species type and age) and geological character-
istics (bedrock and soil) are often substantially different among watersheds,
the watershed design is usually quite similar. Each drainage basin has to
some degree vertical and horizontal borders, demarcated by ridges and func-
tionally defined by biological activity and the drainage of water (Bormann
and Likens, 1967).
Accurate monitoring of total hydrologic flux can pose problems. Since
forest cover and litter layer dissipate much of the energy from precipitation
events, infiltration is high and the opportunity for overland flow is slight
The runoff that does occur is usually associated with snowmelt events To
register the greater percentage of subsurface flow, v-notch weirs or flumes
are often anchored to the bedrock at the base of each watershed
As the size of the forested area increases, flow measurement methods
change. Drainage basins covering hundreds or even thousands of hectares use
gauging staffs or other flow measuring devices to determine the proportionately
greater flow volumes. While automatic sampling devices facilitate collection
in the smaller basins, manual methods often still persist in the larger water-
sheds because of the relative uniformity of forest flow and chemical concen-
tration.
agricultural land use
Water quality monitoring in agricultural settings is often conducted in
a manner similar to that for forested systems Areas of agrarian activity
are defined and the resulting runoff is examined separate from the influence
of other land activities. Numerous studies are available which give repre-
sentative loading estimates from general agricultural land use (Avadhanula,
1979; Campbell, 1978; Burton et al., 1977, Lake and Morrison, 1977, Grizzard
et al., 1977, Nelson et al., 1978, Burwell et al., 1974, Taylor et al , 1971).
140
-------
In contrast to forested systems nutrient export from agricultural areas
demonstrates wide variability. Practices are highly diversified and an agri-
cultural basin can consist of a mosaic of different uses such as pasture,
feedlots, row and nonrow crops. Each type of perturbation creates different
hydro!ogic responses, and depending on the percent composition of the basin,
the effect of one activity can influence the final nutrient load In order
to further delineate these effects, individual activities should be, and often
are, separately monitored.
Separation of the various agrarian activities into discrete hydrologic
units is conducted through two basic approaches, and the differences between
approaches are based primarily on the size of the basin under study. The
first approach relies on relatively large hydrologic units ranging from 5-500
hectares in size. In spite of these dimensions, the entire catchment basin
contains a single activity such as row crop or pasture (Alberts et al., 1978,
Chi Chester et al., 1979).
The second technique employs several small runoff plots, usually much
less than a hectare in area. Separated by raised metal, wood or concrete
borders, the individual plots are 2-5 meters wide and 10-25 meters long.
Runoff studies using these plots may include 1 to 20 individual plots. At
the base of each plot is the flow/sampling device often consisting of a
collecting tank which relies heavily on the "batch" collection methods.
Because of the low area and labor requirements, this particular design
has increased in use by university agricultural experiment stations and other
research agencies. Small size permits close proximity to research facilities
and personnel, which has allowed for both close monitoring and manipulation
of environmental conditions such as soil, slope, fertilizer, tillage methods
and crops.
141
-------
urban land uses
Sampling site selection for urban runoff monitoring potentially poses
the greatest difficulty of the three land uses. Since it is not economically
feasible to re-create urban settings using small runoff plots, available con-
ditions must be utilized. These conditions simultaneously impose an expand-
ing set of limitations on data transferability.
Urban runoff is often channeled into storm sewers which later discharge
into nearby tributaries. In order to derive an area! loading rabe, however,
it is first necessary to ascertain that the network of storm sewers is re-
stricted to the boundaries of the watershed and does not contribute runoff
from other basins.
Many cities have combined storm and municipal sewers During high run-
off events, domestic sewage often overflows and mixes with effluent within
the sewer system. While providing valuable information about a particular
site, the results are difficult to apply to other areas because of the in-
ability to separate the proportion of point source contributions from total
flow.
If the above spatial uncertainties can be accounted for, the "flashy"
nature of the individual runoff event must be suitably monitored To accu-
rately assess these transient events, flow must be continuously monitored
(To reduce monitoring costs, it is often necessary to locate the study site
in close proximity to established stream gauges such as those used by USGS )
Similarly, water quality samples are (or should be) collected with automatic
samp!ers.
Similar to agricultural lands, urban areas consist of a number of differ-
ent land activities. These activities include industrial complexes, business
and commercial districts, parking lots, residential areas, parks and playgrounds
142
-------
Because of differing surface characteristics, the hydro!ogic and water quality
responses from city parks or even large heavily vegetated residential lots
are often quite different from the response from the essentially sealed surface
of shopping malls or industrial complexes. Separation of these discrete types
of activities into distinct drainage basins is not always possible because of
the lack of conformity with topographical boundaries.
A study by AVCO (1970) indicated that aside from these problems, the
following factors also influence site selection for urban runoff studies.
1. Minumum area requirements for the acquisition of a measurable
sample
2. Security of the sampling equipment from vandalism
3. Accessibility of the sampling site
A. 1. 2 Sampling Design and Flux Estimation
The estimation of phosphorus export from watersheds requires good experi-
mental and sampling design. Design considerations include the methods of
acquisition of the concentration samples and flow values, the extent of temporal
sampling, and the method of combination of concentration and flow data for flux
estimation. Use of an inadequate methodology for any of the tasks mentioned ,can
bias the resultant export coefficients. The discussion presented below on these
issues is probably most appropriate for watersheds of moderate to large size,
although the concepts discussed are generally applicable to all watersheds.
Systematic temporal sampling (not including storm sampling) throughout
the year has been examined in the literature for stream quality assessments.
Allum et al. (1977) reported on intensive sampling of tributary phosphorus
discussed in three papers (Treunert et al., 1974; Unger, 1970; Hetling et al.,
1976). In all three studies, the sampling was quite frequent (twice weekly or
143
-------
daily); samples taken from the data set at a reduced frequency, on a systematic
basis, could then indicate the effectiveness of less frequent sampling. In
general, these three studies found that, at a concentration sampling interval
of between 14 and 28 days, the standard error of the annual phosphorus flux
varied between 10% and 20% of the "true" flux.
In addition, Walker (1977) evaluated the effect of serial correlation of
the phosphorus concentration measurements on the equivalent sample size.
Treating the time sequence of samples as a first-order autoregressive (Markov)
process, and assuming a phosphorus concentration serial correlation coefficient
of 0.75 to 0.90 (one day lag), the effective number of samples and the actual
number of samples are essentially equivalent at sampling intervals of 14 to 28
days (or longer). At more frequent sampling, the effective number/actual number
ratio drops below one,indicating that less information is being acquired per sample.
Therefore, a sampling interval of about 14 to 28 days may be a general
guideline for phosphorus concentration. This must be considered in light of the
following comments, however.
1. More frequent sampling will still reduce uncertainty in the phosphorus
concentration, but at a reduced efficiency.
2. Less frequent sampling can still be used to estimate phosphorus
concentration, but at a greater risk of significant error (see data
presented in Hetling et a!., 1976).
3. Sampling should not be systematic with respect to time (e.g., every
two weeks). A better approach is to establish sampling as systematic
with respect to flow, with a random start. This means that the year
should be divided into n equal flow periods, for the purpose of taking
n concentration samples per year.
-------
Sampling should also occur during storm events, as storms may be the major
transporter of phosphorus from the land to surface water bodies in certain
situations. During storms, the method of acquisition of the concentration
samples is important, because of significant concentration variability. Con-
centration sampling should preferably be a composite on a flow-weighted or
mass-flow-weighted basis, not on an equal time basis (Marsalek, 1975). Alter-
natively, grab samples could be collected, at perhaps five to ten minute intervals
during a storm, but this would lead to a higher sample processing cost. One way
grab sampling may be acceptable is through stratification of the sampling with
respect to time, assuming a model of first flush followed by an exponential decay
with time. This can be thwarted, however, by storms that, due to fluctuating
intensity, produce several runoff peaks. Remote automatic sampling units may be
necessary because human response may be too late for the important first flush.
Flow estimation can basically be undertaken in three ways. Continuous flow
measurement is clearly preferable, but it is costly and often not feasible.
An acceptable alternative is an annual flow regression equation developed by the
USGS. These should be available for each state (e.g., Bent, 1971), and they
provide an estimate of the annual flow and the standard error of the flow estimate.
A third alternative, which must be considered unacceptable here because it does
not yield an estimate of precision, is to simply measure instantaneous flow at
the time of concentration sampling.
Finally, flux estimation can follow several approaches, each of which can be
most appropriate under certain conditions. These include techniques dependent
upon a:
1. regression of mass flux versus watershed characteristics,
2. flow-weighted concentration,
3. regression of concentration versus flow, and
4. regression of flux versus flow.
145
-------
The following comments outline the approaches taken. Walker (1977) looked at
several flux estimation approaches and concluded that flow-weighted concentration
times average flow is the best (determined by bias, variance, and calculation
effort) estimator when concentration does not vary greatly with flow. The
EPA-NES (1975) developed a concentration versus flow regression from data taken
at 250 sampling sites. Their equation indicates that a 1% change in flow results
in a -0.11% change in phosphorus concentration and a -0.06% change in nitrogen
concentration. The magnitude and direction of these changes must be considered
with the fact that the EPA-NES data included watersheds containing major point
sources. Bouldin et al. (1975) developed a regression equation for phosphorus
concentration as a function of flow and the rate of change of flow. Smith and
Stewart (1977) looked at eight different approaches for the estimation of annual
nutrient flux. Included among these approaches were flow-weighted concentration
times mean flow and concentration/flow polynomials. They selected a regression
of log flux on log flow because of both good results and mathematical simplicity.
Finally, Verhoff et al. (1980) found that a flow interval method relating
phosphorus flux to streamflow provides the best fit to Lake Erie tributary data.
In conclusion, the estimation technique used should probably depend upon
the:
1. intended use, (A regression on watershed characteristics and land uses
may be useful for future predictions.)
2. fit of the data to the equations, and
3. simplicity of the mathematics.
A. 1. 3 Standardization of Results Reported in the Literature
In addition to the need for statistical considerations in sampling designs,
there is also a necessity for uniformity in the presentation of results. Nutrient
146
-------
contributions from overland drainage have been and continue to be reported
in a variety of forms - usually expressed as either concentration (mass/volume)
or loading (mass/unit area-time). Because of difficulties in interpretation,
however, these results must sometimes be analyzed and compared carefully. Cross-
sectional comparisions of concentrations are particularly risky.
Streamwater concentrations alone can suffice for total output comparison
provided several important assumptions are satisfied. If the watersheds to be
compared have similar values for precipitation, precipitation chemistry, evapo-
transpiration and chemical response characteristics (or if the differences in
these properties among watersheds can be measured), then Streamwater chemistry
is a sufficiently accurate measure of total elemental losses (Vitousek, 1977;
Vitousek and Reiners, 1975).
However, a better unit for comparison is an area yield rate such as loading.
This is the product of flow volume and concentration over time divided by water-
shed area. This unit incorporates runoff duration and catchment area directly,
as well as rainfall intensity and catchment character indirectly (Betson, 1978;
Griffin et al., 1978). Not only are comparisions between watersheds and land
uses possible, but relationships between certain inputs (i.e., precipitation) and
outputs are more definitive. Therefore, investigators conducting studies of
nutrient runoff from land use activities are urged to report unit area! loading
or export in addition to concentration.
A. 2 Issues Important in the Determination of Phosphorus Loading to Lakes
A. 2. 1 Phosphorus Fractions and Availability
The transport of contaminants, especially those emanating from dxffuse
147
-------
sources, is intimately connected with the hydrologic cycle. Nutrient flux
to streams and lakes is generally positively correlated with rainfall, runoff,
and sediment inputs. While linked with one common transport vector,
the forms of these contaminants are source-dependent Groundwater inputs
are primarily in the dissolved phase, while precipitation, stormwater runoff
and point source effluents consist of both dissolved and particulate species.
The form of particular nutrients has become increasingly important in
terms of biological availability. Until recently, eutrophication control
programs have been based largely on the regulation of any fraction of phos-
phorus that was amenable to management, irrespective of whether the phosphorus
was in an available fraction which could support algal growth. This has raised
some serious questions concerning what fractions should be collected and/or
measured.
It is generally agreed that the soluble inorganic forms of phosphorus
are readily available biologically. This included forms such as the soluble
orthophosphates and condensed phosphates. There is a high degree of uncer-
tainty, however, concerning what fractions of particulate inorganic and
organic forms are available. Complicating matters is the presence of dynamic
and complex sets of physical, chemical and biological processes which deter-
mine this availability in the aquatic system. For example, sediment-attached
phosphorus that is not available under certain chemical conditions at one
point in time, may become available under the same or different chemical
conditions at another point in time. This is in sharp contrast to the static
and controlled nature of the laboratory conditions where a variety of techni-
ques are used to correlate algal uptake with actual and highly variable
"in situ" conditions. Consequently, any estimates of bioavailability must
148
-------
be viewed with a high degree of uncertainty and as only "ball park" approxi-
mations.
One of the more comprehensive studies concerned with assessing algal-
avaiTable phosphorus was conducted by Cowen and Lee (1976a, b) and Cowen
(1974). From both urban runoff samples collected in Madison, Wisconsin and
agricultural runoff samples obtained in New York State, these investigators
determined that in the absence of site-specific data, an upper bound estimate
could be made of the available phosphorus in tributary waters:
available P = SRP + .2 PPT (A-l)
where.
SRP = soluble reactive phosphorus
PP.,. = total parti cul ate phosphorus
Lee et al. (1979) later made the following recommendation for the
available phosphorus load from urban stormwater drainage and normal-tillage
agricultural runoff. If the runoff enters a lake directly, or encounters a
limited distance of tributary travel between source and lake, then the avail-
able phosphorus loading may be estimated as:
available P = SPQ + 0.2 P?T (A-2)
where:
SP = soluble orthophosphorus
149
-------
Additional studies have demonstrated comparable, albeit variable, results.
Based on independent, but limited, studies of rivers in the Great Lakes
basin, 40% or less of the suspended sediment phosphorus was estimated to be
in a biologically available form. Overall, probably no more than about
50-60% of the tributary total phosphorus (including soluble P) is likely
to be biologically available (Logan et al., 1979; Armstrong et al., 1979,
Songzoni and Chapra, 1980; Thomas et al., 1979).
The issue of phosphorus availability has also been directed towards
other inputs such as precipitation and point sources. For precipitation,
Dillon and Reid (1980) estimated that up to 28% of the total bulk loads and
40% of the total P in wet-only precipitation was available. Studies by
Murphy and Doskey (1975) speculated that 50% of the total phosphorus in
bulk loads was ultimately available.
The availability of point source phosphorus is variables, depending
upon whether phosphorus removal is practiced (i e., iron, aluminum, or cal-
cium hydroxide precipitation), or depending upon factors such as limita-
tions on phosphorus detergents. It is generally believed, however, that
the major fraction of wastewater phosphorus is available (Lee et al., 1979).
Studies by Ynung et al. (1980) indicate that up to 72% of total phosphorus,
55% of the total particulate phosphorus and 82% of total soluble phosphorus
are available
It should be stressed that availability usually applies to the phos-
phorus fraction that is utilized within one growing season. Depending on
conditions, there is, however, a potential for at least some (if not all)
of the remaining fraction of particulate phosphorus to be utilized at a
later date (due to sudden equilibrium changes). Regardless of what percent
of the total is initially utilized, or what fraction of the remainder has
150
-------
future potential availability, it is imperative that sampling be undertaken
for both soluble and particulate forms. This is especially important since
particulate phosphorus can be an order of magnitude greater in quantity than
the reported dissolved fraction.
Proper assessment of the particulate fraction requires a greater emphasis
on sampling during storm events since the bulk of this fraction is carried
with stormwater runoff. The cumulative effect of many storm events is not
only considerable enough to degrade water quality but often sufficient to
negate the positive aspects of local point source pollution abatement pro-
grams. Many studies have demonstrated that just a few storms during a given
year were responsible for the bulk of the total annual nutrient load (Alberts
et al., 1978; Kissel et al., 1976; Schuman et al., 1973).
Both dissolved and particulate fractions respond to storm events differ-
ently. Although variation exists, their response relative to the storm
hydrograph can be discussed in somewhat general terms (see Figure A-I).
The initial increase in streamflow is often associated with a decrease
in the dissolved nutrient fraction. This decrease is attributed to the dilution
effect of the greater runoff volume, resulting in the lowest dissolved con-
centration at the peak of the hydrograph. As flow rates decrease, the dissolved
component tends to gradually increase to concentrations approaching that of the
pre-storm baseflow conditions.
For the particulate (or sediment) fraction, a different response is
evident. During the initial rapid rise of the hydrograph, the particulate
component increases dramatically, often reaching a maximum concentration
preceding peak flow. This phenomenon, often referred to as "first flush", is
151
-------
en
ro
Flow
Dissolved
Fraction
Particulafe
/ Fraction
TIME (hours)
Figure A-l Dissolved and Participate Nutrient Response to the Storm Hydrograph.
-------
the result of the dislodging of particulate matter from the land surface
during the initial stages of runoff, leaving little material for transport
at later periods. Regardless of where the particulates "peak out" relative
to the hydrograph peak, a decrease in flow is accompanied or preceded by a
decrease in particulate concentration.
A. 2. 2 Variability, Precision, and Accuracy
Variation in nutrient flux through time has been intimately linked to
changes in flow. To adequately account for these variabilities, and to
reduce the amount of uncertainty irv the phosphorus loading estimate, the
sampling frequency should be dictated by the hydro!ogic response. Many
previous sampling studies have failed to address this issue but have instead
made broad but untested assumptions concerning watershed hydrology and load-
ing responses. Sampling intervals have ranged from once per week to irregular
periods during the year, resulting in many of the more sporadic storm
events being missed.
Hydrologic response (and sampling frequency) differs according to drain-
age basin characteristics. As land use progresses toward urbanization,
channels are straightened or paved, small tributaries are filled and the
watershed surface generally becomes smoother and more conducive to sheet
runoff. Therefore, as land use is intensified (i.e., rural to urban) the
effect on drainage basin hydrology is to:
1. increase the storm peak discharge,
2. increase the storm runoff volume while reducing baseflow,
3. decrease response time,
4. increase annual runoff and reduce groundwater recharge, and
5. increase the number of days of no (baseflow) discharge.
153
-------
(Turner et al., 1977, Ikuse et al., 1975, Okuda, 1975, Yoshino, 1975;
Hollis, 1975; Gregory and Walling, 1973; Lindh, 1972; Moore and Morgan,
1969; Holland, 1969; Leopold, 1968).
The result of the first three of these effects is visually interpreted
in Figure A-2
Since peak discharge and flow volume are higher in urban areas, urban
nutrient storm loads are often substantial. In a comparison between urban
and rural watersheds, Burton et al. (1977) reported that up to 98% of the
total phosphorus load was exported in storm flow on an urban watershed while
storm events accounted for slightly more than half this amount on the rural
basin. Conversely, overland runoff from forested basins is a rare event
with an extended response time resulting from slow discharge after precipi-
tation. Hence, sampling frequency need not be as rigorous as in "flashy"
urban watersheds.
To sufficiently describe the nutrient export from differing land uses,
Sherwani and Moreau (1975) describe the desired frequency of measurement
as a function of the following considerations:
1. the response time of the system,
2. expected variability of the parameters,
3. half-life and response time of constituents,
4. seasonal fluctuations and random effects,
5. representativeness under different flow conditions,
6. short terra pollution events,
7. the magnitude of response, and
8. variability of the inputs.
154
-------
Ol
on
cc
5,
o
Q
Urban
Agricultural
Agricultural-
Forest
Forest
1 1 1 1 U
TIME (hours)
Figure A-2 Hydrographic Response of Varying Land Uses to a Storm Event.
-------
Simply stated, there is no single best sampling frequency for all con-
ditions.
To reduce loading uncertainty, a greater degree of accuracy and precision
may be gained by maintaining complete flow records while obtaining enough
concentration samples to adequately characterize the flow variability. While
accumulation of flow records is fairly straight forward (using USGS stream
gauging stations, for example), the concentration sample collection process
can often be made reasonably efficient if stratified random sampling is
employed (Reckhow, 1979b). Under this sampling scheme, the population is
divided into homogeneous sub-populations (strata) that are separately
*
sampled according to the degree of variability which they exhibit (Snedecor
and Cochran, 1973). The underlying assumption is that the population can be
more accurately represented as the sum of sub-populations, therefore reducing
the sample variance.
In the context of hydro!ogic data collection, two temporal strata are
evident:
1. high flow events produced by rainfall runoff and snowmelt, and
2. baseflow produced by groundwater flux.
To expect a gain in precision over simple random sampling, more fre-
quent measurements should be applied to the stratum represented by high flow
events. If the sample size is increased in this stratum and the final con-
centration properly weighted, a more precise and accurate estimate of the
population average will be obtained.
The studies selected for inclusion in the export coefficient tables
employed a wide variety of sampling techniques, but nearly all were based
156
-------
upon complete flow records. While stowm runoff was not sampled ac every
event, it was felt that a sufficient number of events were examined to allow
for realistic estimates of the total nutrient load for a particular land
use.
*
A. 2. 3 Temporal Extent of Sampling
Climate determines local weather conditions which in turn influence
the quantity and duration of baseflow and the number and periodicity of
storm events. While some areas of the country exhibit relatively uniform
climates (e.g , pacific northwest) evenly distributed periods of precipita-
tion are usually not the norm. Winter thaws and spring/simmer rains often
create seasonal cycles of high and low runoff.
Intimately associated with climatic periodicity is the modifying impact
land use has on hydrologic response. The relatively uniform annual flow
patterns of many undisturbed forests is in sharp contrast to the highly
variable flows eminating from urbanized and agricultural basins. As vegeta-
tive cover is artificially reduced and the basin is increasingly developed,
groundwater recharge and flux are reduced Baseflow and nutrient export
are often either inconsequential or absent during dry summer or winter periods
Consequently, a greater percentage of nutrient export occurs during wet
periods of the year for disturbed watersheds than for undisturbed watersheds.
As a result of this seasonal variability, high runoff seasons exhibit
greater variance in nutrient concentrations and total nutrient loads than do
low runoff or baseflow periods. For a given confidence level (precision)
and a margin of error (accuracy), the temporal extent of sampling must in-
clude these high and low runoff periods (especially for the more disturbed
watersheds). If sampling duration focuses exclusively on one season (e g ,
spring), the nutrient flux estimate may sufficiently describe that time
157
-------
period but may not be indicative of other unsampled periods For this
reason, the reader is warned against extrapolating seasonally reported re-
sults toward more extended time frames. This will bias the nutrient flux
estimate toward whatever season in which the sampling was performed To?
better account for this seasonal variability and to allow for a more stan-
dardized unit of measure for comparison purposes, a more informative approach
is to sample and report the data in yearly increments.
While the bulk of studies included in the export tables are the result
of intensive sampling and annual flow data, many investigators have refined
the sampling period within the water-year time frame. According to Likens
et al. (1977), the ideal water-year is that successive twelve-month period
that most consistently, year after year, gives the highest correlation be-
tween precipitation and streamflow.
Examination of precipitation-streamflow data at Hubbard Brook resulted
in a water-year beginning June 1 and ending May 31 Since the beginning of
this water-year corresponds with the appearance of foliage, it allows for a
separation of the vegetation growth and dormancy periods. This concept has
been effectively applied by other investigators working with agricultural
land uses (Alberts et al., 1978, Burwell et al., 1975)
A. 3. Prediction Uncertainty Estimation for Area! Water Loading (q ) Error
The methodology presented in Chapter 2 is based on the assumption that
model variable error is contributed only by uncertainty in phosphorus load-
ing (L). Under some conditions and in some lakes, uncertainty in area! water
loading (q ) may also be significant. For example, since uncertainty in-
o
eludes natural variability, lakes with highly variable flushing rates may be
158
-------
candidates for q -error analysis. In addition, since measurement error is
also a part of total uncertainty, lakes for which flushing rates are poorly
characterized might also be analyzed for q -uncertainty.
The procedure presented below is designed to interface with the steps
in the Chapter 2 methodology. It is assumed that the uncertainty may orig-
inally be estimated in terms of Q (the annual volumetric water flow through
a lake), but that for analysis purposes it is re-expressed as qs = Q/AQ
(where A = the lake surface area (a constant)).
The contribution to total prediction uncertainty from uncertainty in
; i
1970).
q is calculated using the error propagation equation (Benjamin and Cornell,
s(P)
n
-
1/2
(A-3)
where:
s(P) = contribution to total uncertainty in the model (P), due
to uncertainty in variables x and x ,
' J
x ,x = model parameters or independent variables,
s(x1) = uncertainty (standard error) in x , and
P(X ,x ) = correlation between x and x .
* J I J
The phosphorus lake model is
P = 11.6+L1.2qs (A-4)
159
-------
Therefore, using the error propagation equation, the additional prediction
uncertainty in total phosphorus concentration due to uncertainty in q is
s
s
1.44 L2 2. x 2.4 L
S (> - s(q)s(L)p(Ljq)
(11.6 + 1.2q_r (11 6 + 1 2q_)
3
C * \iiw » »_ ^ _ y
1/2
(A-5)
The confusing array of symbols necessitates interpretation. In conjunc-
tion with their interplay with the steps in Chapter 2, the symbols are
1. p(L, q ) is the correlation between L and q_. Since both
o o
are primarily determined by Q, this correlation should be
positive, which diminishes the importance of the q -uncertainty
contribution. Ideally this correlation should reflect a
time series of data for an application lake In the absence
of this site-specific information, cross-sectional studies
suggest a correlation coefficient between L and q of + 5 to
+.8.
2. s(q ) is the estimate of uncertainty in q determined by
the analyst. It is different from s which is defined
below.
3. s(L) is the estimate of uncertainty in L It has positive
and negative components. In Step 2G, the high, most likely,
and low phosphorus loading terms are calculated The re-
sultant uncertainties in loading are
s(L)+ = (A-6)
S(L)- „ (A_7)
160
-------
4. s0 is the contribution to the total phosphorus concentra-
Hs
tion prediction uncertainty due to uncertainty in qg It,
too, has positive and negative components (resulting from
the positive and negative components in s(L)). Thus
a. Sqs+ is found using p(L, qg), s(qg) and s(L) in
Equation A-5.
b sn - is found using p(L, q_), s(q ) and s(L)" in
Me o o
Equation A-5.
Then:
a. Sq + is squared and added to the right side of Equation
12 in Step 4F.
b. Sq - is squared and added to the right side of Equation
14 in Step 4G
This modification results in positive and negative error in-
tervals reflecting all known uncertainties (including uncer-
tainty in qs)
161
-------
Table Ala: Phosphorus Export from Forested Watersheds
Land Use
75-100 year old
jack pine -
black spruce
(34 ha)
Climax hardwoods
(125 ha)
Jack pine -
black spruce
Jack pine -
black spruce
70% aspen
30% black spruce
and alder
(10 ha)
Aspen - birch
(6 48 ha)
Precip-
itation
cm/yr
96 7
80 3
70 1
74 3
126 3
82 1
79 48
75 51
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Particulate/Sednment
cm/yr PO.-P Total P Phosphorus
29.7
35 4
22.3
23 4
68 0
032 028
024 012
17 7
19 2
15 5
21 47 05
15 56 20
13 73 16
Total
Phosphorus
329
435
289
220
090
060
036
124
179
157
19
38
28
Reference
Schindler et al , 197
Schindler and
Nighswander, 1970
Nicholson, 1977
Nicholson, 1977
Verry, 1979
Timmons et al , 1977
Maple, birch,
beech (15 6 ha)
132 2
83 3
007
012
019
Likens et al , 1977
-------
Table Ala: (continued)
CT>
co
Land Use
Deciduous
hardwood and
pine (17 6 ha)
Mixed deciduous
forests, sandy
soils - igneous
formation
Mixed deciduous
forests, loam
soils, sedimen-
tary formation
Precip- Water Phosphorus Export (kg/ha/yr)
itation Runoff Dissolved Phosphorus Parti cul ate/Sediment Total
cm/yr cm/yr PO.-P Total P Phosphorus Phosphorus Reference
854 253 .035 035 Taylor et al, , 1971
88.9 35 6 072 072
92 8 32 0 035 035
070 Dillon and Kirchner,
047 1975
067
075
046
050
037
025
060
072
030
052
025
035
037
077
048
038
027
041
145 Dillon and Kirchner,
092 1975
122
-067
Mixed deciduous
forest ( 01 ha)
129 0
84 3
090
Singer and Rust, 1975
-------
Table Ala: (continued)
Precip-
itation
Land Use cm/yr
Oak hickory 139 5
forest (97.5 ha) 128 2
187 5
174 7
Oak, maple,
yellow poplar,
black cherry,
beech (34 ha)
Mixed pine and 164 0
hardwood
(40 ha)
Mixed mature
hardwoods,
Coweeta hydro-
logic lab,
North Carolina
(12 1 - 61 1 ha)
Mixed pine and
hardwood
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Particulate/Sediment
cm/yr PO.-P Total P Phosphorus
74 5 01
71 0 02
114 8 03
116 1 03
48 7 265 010
€2
02
02
03
02
02
03
Total
Phosphorus
01
02
03
03
18
14
08
275
20
Reference
Henderson et al , 1977
Aubertin and Pa trie,
1974
Krebs and Golley, 1977
Swank and Douglas,
1977
Correll et al , 1977
99% mixed forest
Y/o developed
(6495 ha)
7 3
212
Bedient et al , 1978
-------
Table Ala' (continued)
01
Land Use
Loblolly and
slash pine
Mississippi
(2 81 ha)
(1 93 ha)
(2 39 ha)
(1 64 ha)
(1 49 ha)
Loblolly and
slash pine
(2 81 ha)
(1 93 ha)
(2 39 ha)
(1 64 ha)
(1 49 ha)
Douglas fir
and western
hemlock
(47139 5 ha)
(32376 0 ha)
Douglas fir
and western
hemlock
(10 1 ha)
Precip-
itation
cm/yr
189 08
189 08
189 08
189 08
189 08
205 0
205 0
205 0
205 0
205 0
215 0
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Parti cul ate/Sediment
cm/yr PO.-P Total P Phosphorus
39 48 04
46 40 05
37 88 04
30 26 04
39 63 05
36 90 094 187
38 95 110 196
34 85 097 260
30 75 083 238
32 55 055 171
135 0
Total
Phosphorus Reference
Schreiber et al ,
1976
281 Duffy et al , 1978
306
357
321
226
830 Sylvester, 1960
360
520 Fredriksen, 1972
Douglas fir and
western hemlock
158 0
08
180
Fredriksen, 1979
-------
Table Ala: (continued)
Land Use
Precip-
itation
cm/yr
Water
Runoff
cm/yr
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus
P04-PTotal P
Particulate/Sediment
Phosphorus
Total
Phosphorus
Reference
Douglas fir and
western hemlock
76 0
47
680
Frednksen, 1979
en
-------
Table Alb* Nitrogen Export from Forested Watersheds
Land Use
75-100 year old
jack pine-
black spruce
(34 ha)
Climax hard-
woods (125 ha)
Jack pine -
black spruce
Jack pine -
black spruce
70% aspen
30% black spruce
and alder
(10 ha)
Aspen-birch
(6 48 ha)
Precipitation
cm/yr
96 7
80 3
70 1
74 3
126 3
82 1
79 48
75 51
Water
Runoff
cm/yr
29 7
35 4
22 3
23 4
68 0
17 7
19 2
15 5
21 47
15 56
13 73
N03-N
1 26
108
171
20
05
33
17
19
09
Nitrogen Export (kg/ha/yr)
Dissolved Nitrogen Particulate/Sediment Nitrogen
NH4-N
126
037
23
10
37
16
47
19
TKN-N ORG-N Total-N NOg-N NH.-N TKN-N ORG-N Total-N
2 028 342
1 22 164
1 83
2 21
1 34
2 13
2 63
1 64
Total
Nitrogen
6 45
7 32
6 07
5 69
2 37
1 384
2 26
2 37
1 74
2 46
3 29
1 J2
Reference
Schindler et al ,
1976
Schindler and
Nighswander, 1970
Nicholson, 1977
Nicholson, 1977
Verry, 1979
Timmons et al ,
1977
Sugar maple, yellow
birch, beech, red
spruce, balsam fir
and paper birch
New Hampshire
(607 ha)
6 6
04
Martin, 1978
-------
Table Alb: (continued)
00
1 and Use
Maple, birch,
beech (15 6 ha)
Deciduous hardwood
and pine (17 6 ha)
Oak-hickory
forest (97 5 ha)
Oak hickory
forest (97 5 ha)
Oak, maple, yellow
poplar, black
cherry, beech
(34 ha)
Precipitation
cm/yr
132 2
84 4
88 9
92 8
136 0
189 5
174 7
Hater
Runoff
cn/yr
83 3
25 3
35 6
32 0
70 7
114 8
116 1
Mixed mature hard-
woods, Coweeta hydro-
logic lab , North
Carolina
(12 1 - 61 1 ha)
Mixed pine and
hardwood
N03-N
80
1 60
70
40
1
2
45
60
86
03
06
05
35
05
05
15
Nitrogen Export (kg/ha/yr)
Dissolved Nitrogen Harticu late/Sediment Nitrogen
NH^-N TKN-N ORG-N Total-N NOj-N NH^-N TKN-N ORG-N Total-N
3 90 11
1 10 1 60 3 10
321 2,2
215 17
1 52
84
86
03
05
05
10
07
06
07
Total
Nitrogen
4 01
1 37
3 16
2 82
3 10
2 2
1 7
1 50
Reference
Likens et al ,
1977
Taylor et al ,
1971
Henderson and
Harris, 1973
Henderson et al ,
1977
Aubertiti and
Patric, 1974
Swank and Douglas
1977
Correll et al ,
1977
-------
Table Alb (continued)
I ami Use
Precipitation
cm/yr
Water
Runoff
cm/yr
Dissolved Nitrogen
Nitrogen Export (kg/ha/yr)
N03-N
TKN-NORG-N Total-N
Particu1 ate/Sediment Nitrogen
,-N NH.-N TkN-N ORG-N"tStaf-
o H
Total
Nitrogen Reference
99% mixed forest
~\% developed
(6495 ha)
7 3
286
Bedient et al ,
1978
Loblolly and
slash pine,
Mississippi
2 81 ha)
1 93 ha)
2 39 ha)
(1 64 ha)
(1 49 ha)
189 08
189 08
189 08
189 08
189 08
39 48
46 40
37 88
30 26
39 63
33
39
30
24
33
4 40
Schreiber et al
1976
CTi
Douglas fir
and western
hemlock
(47139 5 ha)
3 32 Sylvester, 1960
Alpine forest,
New Mexico
Gosz, 1978
91 4% pine
8 6% pinion-
juniper (116 ha)
004
03
06
56% mixed conifer
44% spruce-fir
(180 ha)
06
13
23
-------
Table Alb: (continued)
Hater Nitrogen Export (kg/ha/yr)
Precipitation Runoff Dissolved Nitrogen _ Particulate/Sedlnient Nitrogen Total
I and Use cm/yr cni/yr N03-N NH.-N TKN-N ORG-N Total-N NOj-N NH.-N TKN-N ORG-N Total-N Nitrogen Reference
68 3* spruce-fir 05 12 Gosz, 1978
14 02 aspen (continued)
11 0% mixed conifer
6 7% pine
(164 ha)
64% spruce-fir 25 40
23% subalpine
grassland
13% aspen (100 ha)
Aspen (3 4 ha) 14 32
48 9% aspen 13 28 82
39 0% subalpine
grassland
11 1% spruce-fir
1 0% alpine tundra
(415 ha)
84 4% spruce-fir 08 25
15 6% aspen (122 ha)
75 5% spruce-fir 55 43 99
24 5% alpine tundra
(163 ha)
Douglas fir and 251 0 170 0 58 Frednksen, 1972
western hemlock 215 0 135 0 38
(10 1 ha)
-------
Table Alb: (continued)
Precipitation
1 ami Use cm/yr
Douglas fir and
western hemlock
Douglas fir and
western hemlock
Alder and
douglas fir
Western Oregon
6835 alder
32% douglas fir
(203 14 ha)
68% alder
32% douglas fir
25% patch cut
(303 32 ha)
Water Nitrogen Export (kg/ha/yr)
Runoff Dissolved Nitrogen Particulate/Sediment Nltroqen Total
cm/yr N03-N NH4-H TKN-N ORG-N Total-N N03~N NH^-N TKN-N ORG-N Total-N Nitrogen
158 0 07 70 77
76 0 02 71 73
35 04
37 40
28 45
24 95
31 46
25 40
28 42
24 54
Reference
Fredriksen, 1979
Fredriksen, 1979
Brown et al ,
1973
-------
Table A2a: Phosphorus Export from Row Crops
ro
Land Use
Corn
continuous plant-
ing ( 004 ha)
Corn
continuous plant-
ing
Precip-
itation
cm/yr
77 6
77 0
65 76
77 6
77 0
65 76
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus
cm/yr PfVP Total P
10 7
8 51
21 95
12 26
5 97
19 41
Particulate/Sediment Total
Phosphorus Phosphorus
1 22
1 49
1 22
5 77
1 03
2 00
Reference
Minshall et al ,
1970
f
Minshall et al ,
1970
fresh manure
winter applied
( 004 ha)
Corn
continuous plant-
ing
fermented manure
spring applied
( 004 ha)
77 6
77 0
65 76
11 51
5 59
15 32
96
75
68
Minshall et al ,
1970
Corn 77 6 12 45
continuous plant- 77 0 5 61
ing, liquid manure 65 76 15 60
spring applied
( 004 ha)
1 18
95
76
Minshall et al ,
1970
Corn
continuous plant-
ing, no manure
( 004 ha)
8 71
14 33
1 00 Hensler et al ,
1 60 1970
-------
Table A2a* (continued)
oo
Precip-
itation
Land Use cm/yr
Corn
continuous plant-
ing, fresh manure
winter applied
( 004 ha)
Corn
continuous plant-
ing, fermented
manure
spring applied
( 004 ha)
Corn
continuous plant-
ing, liquid
manure, spring
applied ( 004 ha)
Corn 62 6
continuous
( 009 ha)
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Parti cul ate/Sediment
cm/yr PO^-P Total P Phosphorus
7 11
11 53
7 11
10 52
8,10
10 79
8.6 34 18 2
Total
Phosphorus
5 66
1 13
73
90
91
97
18 6
Reference
Hensler et al ,
1970
Hensler et at ,
1970
Hensler et al ,
1970
Young and Holt,
1977
Corn ( 009 ha)
65 7
10 1
13 7
14 0 Young and Holt,
1977
Corn
surface spread
manure (.009 ha)
65 7
3 8
8.1
8 6 Young and Holt,
1977
-------
Table A2a: (continued)
Land Use
Precip-
itation
cm/yr
Water
Runoff
cm/yr
Dissolved Phosphorus
P04-PTotal P
Phosphorus Export (kg/ha/yr)
Particulate/Sediment
Phosphorus
Total
Phosphorus Reference
Corn
plowdown
manure ( 009 ha)
65 7
4 0
9 4
9 8 Young and Holt,
1977
Corn
rotation
planting
( 009 ha)
57 2
4 57
11
17
2 97
3 14 Burwell et al ,
1975
Corn
continuous
planting
( 009 ha)
57 2
8 03
18
33
5 22
5 55 Burwell et al ,
1975
Corn
continuous
contour planting
(30 - ha)
79 79
80 04
73 8
86 2
105 95
63 07
78 25
41
47
12 57
86
64
37
63
19
085
237
04
175
019
043
306
948
881
554
104
073
244
496
1 033
2 118
594
279
092
287
Alberts et al ,
1978
Corn
continuous
contour planting
(33 6 - ha)
80 11
78 29
74 08
86 45
104 59
62 16
78 65
93
86
9 76
81
5
52
2 11
094
046
189
028
205
026
052
163
477
1 099
426
048
057
301
257
523
1 288
454
253
083
353
Alberts et al ,
1978
-------
Table A2a (continued)
Land Use
Corn
continuous
terraced
(60 - ha)
Precip-
itation
cm/yr
52 8
73 12
76 41
95 24
102 46
53 81
73 76
"Water
Runoff
cm/yr
70
35
1 75
10 71
8 49
66
2 9
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus
P04-P Total P
081
009
059
119
238
018
128
Parti cul ate/Sediment
Phosphorus
009
015
228
494
161
032
131
Total
Phosphorus
09
024
287
613
399
050
259
Reference
Alberts et al , 1978
en
Corn 107 7
continuous plant-
ing (1 29 ha)
13 0
25
54
1 67
2 21
Smith et al , 1978
Corn
6 replications
( 001 ha)
87 39
40
Bradford, 1974
Soybeans
two crops/yr
conventional till
( 01 ha)
118 0
169 2
28
83
025
25
17 5
17 75
McDowell et al , 1978
Soybeans
two crops/yr
no till
( 01 ha)
118 3
169 2
13 0
42 8
1 2
1 8
1 1
2 9
McDowell et al , 1978
-------
Table A2a: (continued)
Land Use
Cotton,
continuous plant-
ing (17 9 ha)
Cotton ,
continuous plant-
ing (12 1 ha)
Soybeans - corn
two crops/yr
no till ( 01 ha)
Corn - soybeans
two crops/yr
no till ( 01 ha)
Corn
silt loam soils
Aurora, New York
( 32 ha)
Citrus grove
surface tillage,
Precip-
itation
cm/yr
97 3
72 7
88.1
74 4
96 3
73 1
88 2
72 9
118 3
169 2
118 3
169 2
98 1
163 5
146 1
Water
Runoff
cm/yr
24 1
8 8
12 6
13 6
24 8
8 0
11 9
13 5
21 5
88 2
66 2
50 5
8 9
17
23
Phosphorus
Export (kg/ha/yr)
Dissolved Phosphorus Parti cul ate/Sediment
PO.-P Total P Phosphorus
2 18
68
86
1 06
1 67
51
70
98
1 3
5
0 8
2 2
21
01
01
9 34
1 70
2 68
4 01
9 08
1 56
2 80
4 68
6 3
2 2
Total
Phosphorus
11 52
2 38
3 54
5 07
10 75
2 07
3 5
5 66
6 8
4 4
Reference
Menzel et al , 1978
Menzel et al , 1978
McDowell et al , 1978
McDowell et al , 1978
Klausner et al , 1974
Rogers et al , 1976
sand soil,
Gainesville, FL
(9 ha)
-------
Table A2a (continued)
Land Use
Citrus grove
surface tillage,
sand soil ,
Gainesville, FL
(3 ha)
Citrus grove
surface tillage,
Precip-
itation
cm/yr
163 5
146 1
163 5
146 1
Water
Runoff
cm/yr
9 68
4 43
8 89
6 60
Dissolved
P04-P
37
15
32
23
Phosphorus Export (kg/ha/yr)
Phosphorus Parti cul ate/Sediment Total
Total P Phosphorus Phosphorus
Reference
Rogers et al , 1976
Rogers et al , 1976
sand soil,
heavy lime
application
Gainesville, FL
(9 ha)
-------
Table A2b: Nitrogen Export from Row Crops
CO
Precipitation
1 .ind Use cni/yr
Corn, continuous
planting ( 004 ha)
Corn, continuous
planting, fresh
manure, winter
applied ( 004 ha)
Corn, continuous
planting, fermented
manure, spring
applied ( 004 ha)
Corn, continuous
planting, liguid
manure, spring
applied ( 004 ha)
Corn, continuous
planting, no
manure ( 004 ha)
Co--n, continuous
ol anting, fresh
"arJ'-e, winter
•i.". led ' 004 ha)
77 6
77 0
65 76
77 6
77 0
65 76
77 6
77 0
65 76
77 6
77 0
65 76
Hater Nitrogen Export (kg/ha/yr)
Runoff - Dissolved Nitrogen
on/yr N03-N NH^-N TKN-N ORG-N Total-N
10 7
8 51
21 95
12 26
5 95
19 41
11 51
5 59
15 32
12 45
5 61
15 60
8 71
14 33
7 11
11 53
Partlculate/Sedlment Nitrogen Total
NOj-N NH.-N TKN-N ORG-N Total-N Nitrogen
5 53
3 61
3.96
26 88
3 05
7 97
5 32
3 35
3 38
2 81
2 88
5 07
4 08
4 58
26 06
4 44
Reference
Minshall, et al ,
1970
Minshall et al ,
1970
Minshall et al ,
1970
Hinshall et al ,
1970
Hensler et al ,
1970
Hensler et al ,
1970
-------
Table A2b (continued)
Water
Nitrogen Export (kg/ha/yr)
1 arid Use
Precipitation
cm/yr
Corn, continuous
planting, fermented
manure, spring
applied ( 004 ha)
Corn, continuous
planting, liquid
manure, spring
applied ( 004 ha)
Corn, continuous
( 009 ha)
-J
10 Corn ( 009 ha)
Corn, surface
spread manure
( 009 ha)
Corn, plowdown
manure ( 009 ha)
Corn, rotation
planting ( 009 ha)
62 6
65 7
65 7
65 7
57 2
Runoff Dissolved Nitrogen
cm/yr N03-N NH4-N TKN-N ORG-N Total-N
7 11
10 52
8 10
10 79
8624 40
10 1 2 7 48
386 32
4012 25
4 57 44 18 33
Particulate/Sediment Nitrogen Total
N03-N NH4-N TkN-N ORG-N Total-N Nitrogen
3 68
4 76
4 07
3 70
0 75 6 79 6
0 39 4 44 2
0 24 7 27 9
0 30 5 33 0
21 13 08 14 24
Reference
Hensler et al ,
1970
Hensler et al ,
1970
Young and Holt,
1977
Young and Holt,
1977
Young and Holt,
1977
Young and Holt,
1977
Burwell et al ,
1975
-------
Table A2b: (continued)
00
o
Precipitation
land Use cm/yr
Corn, continuous
planting ( 009 ha)
Corn, continuous
contour planting
(30 ha)
Corn, continuous
contour planting
(33 6 ha)
Corn, continuous
terraced (60 ha)
57
79
80
73
86
105
63
78
80
78
74
86
104
62
78
52
73
76
95
102
53
73
2
79
04
80
20
95
07
25
11
29
08
45
59
16
65
8
12
41
24
46
81
76
Water
Runoff
cm/yr
8
6
5
12
3
6
1
2
5
3
9
3
7
1
2
1
10
8
2
03
41
47
57
86
64
37
63
93
86
76
81
5
52
11
70
35
75
71
49
66
9
Nitrogen Export (kg/ha/yr)
Dissolved mtroqen •
PartlcuJate/Sedlment Nitrogen
Total
N03-N NH^-N TKN-N ORG-N Total-N NOj-N NH.-N TKN-N OR6-N Total-N Nitrogen Reference
1
2
1
1
2
1
1
3
2
11
3
45
31
65
03
04
53
45
53
95
49
60
31
32
24
14
16
26
56
54
80
37 72
54
42
2 10
29
33
08
08
95
34
1 46
14
22
70
03
12
03
59
36
33
01
19
36
5
34
69
23
5
1
4
2
25
41
27
*2
3
7
23
4
1
21 18
85
73
06
42
19
08
43
96
15
30
22
02
68
99
31
52
03
08
19
55
11
23
8
36
7?
/ &
1
c
9
5
26
43
27
2
1
4
7
26
7
1
2
63 Burwell et al ,
1975
69 Alberts et al ,
60 1978
ftf
*T/
OC
OQ
OA
U*T
36 Alberts et al ,
02 1978
71
85
84
69
34
67 Alberts et al ,
69 1978
78
70
08
10
10
-------
Table A2tr (continued)
Co
Precipitation
land Use cm/yr
Corn, continuous
planting (1 29 ha)
Corn, 6 replica-
tions ( 001 ha)
Soybeans, two
crops/yr ,
conventional till
( 01 ha)
Soybeans, two
crops/yr ,
no till ( 01 ha)
Cotton, continuous
planting (17 9 ha)
Cotton, continuous
planting (12 1 ha)
Soybeans - corn
two crops/yr ,
no till ( 01 ha)
107 7
87 39
118 0
169 2
118 3
169 2
97 3
72 7
88 1
74 4
96 3
73 1
88 2
72 9
118 3
169 2
Water
Nitrogen Export (kg/ha/yr)
Runoff Dissolved Nitrogen Particulate/Sediroent Nitrogen Total
cm/yr N03-N NH^-N TKN-N ORG-N Total-N NOj-N NH.-N TKN-N ORG-N Total-N Nitrogen
13 0 86 1 48 5 90
078 55
28 3 70 1 5
83 2 1028
13 0 1221
42 8 0616
24 1
8 6
12 6
13 6
24 8
8 0
11 9
13 5
21 5 3022
88 2 3038
85 5 66 12 42
3 29
42 7 46 5
23 45
11 49
4 99
9 79
8 82
14 84
5 18
10 03
12 19
17 0 23 8
Reference
Smith et al ,
1978
Bradford, 1974
McDowell et al ,
1978
McDowell et al ,
1978
Menzel et al ,
1978
Menzel et al ,
1978 •
McDowell et al
1978
-------
Table A2b: (continued)
Precipitation
Land Use cm/yr
Corn - soybeans
two crops/yr ,
no till ( 01 ha)
Corn, silt loam
soils, Aurora,
New York ( 32 ha)
Citrus grove
surface tillage,
sand soil,
Gainesville, FL
_, (9 ha)
00
ro
Citrus grove
surface tillage,
sand soil,
Gainesville, FL
(9 ha)
Citrus grove
surface tillage,
sand soil,
heavy lime applica-
tion, Gainesville,
FL (9 ha)
118 3
169 2
98 1
163 5
146 1
163 5
146 1
163 5
146 1
Hater
Runoff
cm/yr
66 2
50 2
8 9
17
23
9 68
4 43
8 89
6 60
Nitrogen Export (kg/ha/yr)
Dissolved Nltroqen
N03-N HH.-N TKN-N ORG-N Total -N
81 31
89 67
1 16 42
.
01
02
69
25
82
44
Partlculate/Sedlment Nitrogen Total
N03-N NH.-N TKN-N ORG-N Total-N Nitrogen Reference
McDowell et al ,
59 21 3 1978
Klausner et al ,
1974
Rogers et al ,
1976
Rogers et al ,
1976
Rogers et al ,
1976
-------
Table A3a- Phosphorus Export from Non-Row Crops
00
co
Land Use
Alfalfa
no fertiliza-
tion ( 004 ha)
Alfalfa
fall applied
manure ( 004 ha)
Alfalfa
winter applied
manure ( 004 ha)
Alfalfa
spring applied
manure ( 004 ha)
Precip-
itation
cm/yr
105 4
107 8
108 8
105 4
107 8
108 8
105 4
107 8
108 8
105 4
107 8
108 8
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Parti cul ate/Sediment Total
cm/yr PO/rP Total P Phosphorus
8 2
14 2
18 5
5 2
7 8
9 0
8 2
10 3
12 8
6 7
10 1
15 0
Phosphorus
75
76
2 40
1 24
1 20
8 09
64
58
6 09
2 39
55
1 81
Reference
Converse et al ,
1976
Converse et al ,
1976
Converse et al ,
1976
Converse et al ,
1976
Alfalfa and
bromegrass
two plots
(3 55 - 4 10 ha)
57 91
2 69
24
73
97
Harms et al , 1974
Wheat
continuous
planting
(5 2 ha)
96 5
72 9
87 7
73 1
20 8
7 0
10 5
5 5
61
19
36
13
2 73
51
60
2 19
3 34
80
96
2 32
Menzel et al , 1978
-------
Table A3a: (continued)
Land Use
Wheat
continuous
planting
(5 3 ha)
Spring wheat
and summer
Precip-
itation
cm/yr
96 6
73.1
87 9
72 9
Water
Runoff
cm/yr
23 0
5.5
9 3
5 4
62 5
7 0
Phosphorus Export (kg/ha/yr)
Dissolved
P04-P
0 3
0 1
Phosphorus
Total P
.52
09
.26
11
Parti cul ate/Sediment
Phosphorus
3 77
.50
.53
2 21
Total
Phosphorus
4 29
59
79
2.32
0 6
0 1
Reference
Menzel et al
Nicholaichuk
Read, 1978
, 1978
and
00
stubble
two year rota-
tion (4-5 ha)
Spring wheat
and summer
fallow
two year rota-
tion (4-5 ha)
98 0
19 0
0 9
0 1
2 3 Nicholaichuk and
0 4 Read, 1978
Spring wheat
and fall
fertilized
summer fallow
(4-5 ha)
49 0
7 0
2 3
0 2
5 6 Nicholaichuk and
0 2 Read, 1978
Millet
six replications
( 001 ha)
87 39
0 44
Bradford, 1974
-------
Table A3a: (continued)
Precip- Water Phosphorus Export (kg/ha/yr)
Itation Runoff Dissolved Phosphorus Parti cul ate/Sediment
Land Use cm/yr cm/yr PO^-P Total P Phosphorus
Oats 57 2 6 89 09 22 ,43
Rotation Plant-
ing ( 009 ha)
Total
Phosphorus Reference
65 Bur-well et al ,
1975
Hay
Rotation Plant-
ing ( 009 ha)
57 2
14 2
31
60
04
64 Burwell et al ,
1975
oo
01
Wheat
Aurora, New York
( 32 ha)
98 1
10 7
.20
Klausner et al , 1974
-------
Table A3b: Nitrogen Export from Non-Row Crops
CO
Precipitation
land Use on/yr
Alfalfa
no fertilization
{ 004 ha)
Alfalfa
fall applied
manure ( 004 ha)
Alfalfa
winter applied
manure ( 004 ha)
Alfalfa
spring app^ed
manure ( 004 ha)
Alfalfa and
bromegrass
two plots
(3 55 - 4 10 ha)
Wheat
continuous plant-
ing (5 2 ha)
105 4
107 8
108 8
105 4
107 8
108 8
105 4
107 8
108 8
105 4
107 8
108 8
57 91
96 5
72 9
87 7
73 1
Hater
Runoff
on/yr
8 2
14 2
18 5
5 2
7 8
9 0
8 2
10 3
12 8
6 7
10 1
15 0
2 69
20 8
7 0
10 5
5 5
Nitrogen Export (kg/ha/yr)
NU3-N
1 36
1 14
2 02
1 24
78
1 50
1 51
79
1 88
2 69
98
1 73
Dissolved Nitrogen
tm/ftt TKN-N ORG-N Total -N
1 27
98
3 86
2 63
3 41
8 92
1 71
1 17
13 12
3 58
85
2 75
Partlculate/Sedlraent Nitrogen Total
H03-N NH4-N TKN-N ORG-N Total-N Nitrogen
6 28
5 66
14 67
6 10
6 63
23 09
7 82
5 88
38 22
6 43
4 07
11 42
10
6 12
3 77
5 63
7 12
Reference
Converse et al ,
1976
Converse et al ,
1976
Converse et al ,
1976
Converse et al ,
1976
Harms et al ,
1974
Henzel et al ,
1978
-------
Table A3b: (continued)
Precipitation
Land Use ctn/yr
Wheat 96 6
continuous plant- 73 1
ing (5 3 ha) 87 9
72 9
Spring wheat and
summer stubble,
two year rotation
(4 - 5 ha)
Spring wheat
and summer fallow,
— ' two year rotation
S3 (4-5 ha)
Spring wheat and
fall fertilized
summer fallow,
(4-5 ha)
Millet 87 39
six replications
( 001 ha)
Oats, rotation 57 2
planting ( 009 ha)
Hay, rotation 57 2
planting ( 009 ha)
Hater Nitrogen Export (kg/ha/yr)
Runoff Dissolved Nitrogen Parti culate/Sedfment Nitrogen
cm/yr N03-N NH4-N TKN-N OR6-N Total -N NOj-N NH4-N TKN-N ORG-N Total -N
23 0
5 5
9 3
5 4
62 5 02 ^11
70 01 01
98 0 07 65
19 0 06 09
49 0 04 77
70 02 05
0 10 0 50
6 89 1 57 29 71 03 1 89
14 2 63 1 41 1 67 01 17
Total
Nitrogen
8 95
2 89
4 31
8 74
3 04
4 22
4 09
Reference
Menzel et al ,
1978
Nicholaichuk and
Read, 1978
Nicholaichuk and
Read, 1978
Nicholaichuk and
Read, 1978
Bradford, 1974
Burwell et al ,
1975
Burwell et al ,
1975
-------
Table A3b: (continued)
00
Precipitation
1 and Use cm/yr
Wheat ( 32 ha)
Aurora, New York
coastal
bermuda grass
light manure
fertilization,
sandy loam soil,
Alabama ( 04 ha)
Coastal
burmuda grass,
heavy manure
fertilization,
sandy loam soil,
Alabama ( 04 ha)
98 1
134 8
108 9
191 2
134 8
108 9
191 2
Hater
Runoff
cm/yr
10 7
42 0
10 8
38 6
41 9
10 0
33 6
N03-N
79
3 6
1 0
2 1
26 0
4 6
3 4
Nitrogen Export (kg/ha/yr)
Dissolved Nltroqen Partlculate/Sedlment Nitrogen Total
NH^-N TKN-N OR6-N Total-N NO-j-N NH.-N TKN-N ORG-N Total-fi Nitrogen Reference
56 Klausner et al ,
1974
1 0 Long, 1979
8
1 2
3 2 Long, 1979
1 2
0 8
-------
Table A4a- Phosphorus Export from Pastured and Grazed Watersheds
03
10
Land Use
Moderate dairy
grazing, blue-
grass cover
(1 88 ha)
Heavy dairy
grazing, blue-
grass cover
(1 88 ha)
Pasture
(6 28 ha)
Winter grazed
and summer
rotation (1 ha)
Summer grazed
(1 ha)
Rotation grazing
(42 9 ha)
Pasture for
brood cattle
(10 ha)
Precip-
itation
cm/yr
106 8
104.3
105 5
119.8
106 8
104.3
105 5
119 8
58 4
108 0
108 0
77 83
'73 30
75 40
164 0
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Parti cul ate/Sediment Total
cm/yr PO^-P Total P Phosphorus
20 2
22 5
12 3
24 6
26.0
26 9
19.9
31 8
4 44
12 94 30 15
2 92 40 0
4 39 193 058
, 94 064 017
3 86 386 126
61 8 1 269 076
Phosphorus
,15
16
13
12
70
18
11
12
25
3 6
85
251
081
512
1 345
Reference
Kilmer et al
Kilmer et al.
Harms et al ,
Chi Chester et
1979
Chi Chester et
1979
Schuman et al
1973
, 1974
, 1974
1974
al ,
al ,
>
Krebs and Go Hey,
1977
-------
Table A4a: (continued)
Land Use
Continuous grazing
with some
supplementary
winter feeding
some hay pro-
duction
(351 2 ha)
Continuous grazing
little bluestem
cover, active
gullies (11 1 ha)
Rotation grazing
little bluestem
cover, good
cover (11 0 ha)
Precip-
itation
cm/yr
114 7
105 0
77 8
98 7
50 7
109 1
77 3
99 4
52 4
Water
Runoff
cm/yr
28 4
12 6
17 6
2 02
17 8
4 2
7 7
35
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus
P04-P Total P
14
07
03
01
10
02
02
00
Parti cul ate/Sediment
Phosphorus
3 72
99
1 83
26
1 34
22
25
02
Total
Phosphorus Reference
3 8 Con-ell et al ,
1977
3 86 Menzel et al , 1978
1 06
1 86
26
1 44 Menzel et al ,
24 1978
27
02
Continuous grazing
little bluestem
cover (7 8 ha)
76 5
14 7
3 27
1 63
4 90
Olness et al , 1980
Rotational graz-
ing, little
bluestem cover
(9 6 ha)
78 2
4 3
2 43
0 66
3 09 Olness et al , 1980
-------
Table A4a: (continued)
Land Use
Precip-
itation
cm/yr
Water
Runoff
cm/yr
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus
P04-PTotal P
Parti oil ate/Sediment
Phosphorus
Total
Phosphorus
Reference
Continuous graz-
ing, little
bluestem cover
active gullies
(11.1 ha)
76.5
10.2
01
75
76
01 ness et al , 1980
Rotational graz-
ing, little
bluestem cover
(11 0 ha)
78.2
4.3
02
18
20
01 ness et al , 1980
-------
Table A4b: Nitrogen Export fron Pastured and Grazed 1/atersheds
to
Precipitation
Land Use cm/yr
Moderate dairy
grazing,
gluegrass cover
(1 88 ha)
Heavy dairy grazing,
bluegrass cover
(1 88 ha)
Pasture (6 28 ha)
Winter grazed and
summer rotational
(1 ha)
Summer grazed (1 ha)
Rotation grazing
(42 9 ha)
Continuous grazing,
little bluestem cover,
active gullies
(11 1 ha)
106 8
104 3
105 5
119 8
106 8
104 3
105 5
119 8
58 4
108 0
108 0
77 83
73 30
75 40
105 0
77 8
98 7
50 7
"ater
Runoff
ot/yr
20 2
22 5
12 3
24 6
26 0
26 9
19 9
31 8
4 44
12 94
2 92
4 39
94
3 86
28 4
12 6
17 6
2 02
Nitrogen Export (kg/ha/yr)
HU3-N
2 97
2 63
1 68
2 14
16 10
10 79
7 20
7 28
40
5 75
0 5
1 14
17
96
Dissolved Nltroqen
NH^-N TKN-N OR6-N Total -N
47
44 77
18 55
21 1 15
1 95
61 1 32
19 99
32 1 46
1 12
7 8
0 6
66
09
43
Partlculate/Sedlment Nltroqen Total
N03-N NH.-N TKN-N ORG-N Total-N Nitrogen
3 44
3 83
2 41
3 47
18 05
12 71
8 31
9 26
1 52
8 25 30 85
0 21 85
52 2 32
21 47
2 89 4 28
6 84
5 43
9 23
1 33
Reference
Kilmer et al ,
1974
Kilmer et al ,
19/4
Harms et al , 1974
Chi Chester et al ,
1979
Chichester et al ,
1979
Schuman et al ,
1973
Menzel et al ,
1978
-------
Table A4b: (continued)
<£>
CO
1 and Use
Rotation grazing,
little bluestem
cover, good cover
(11 0 ha)
Continuous grazing,
little bluestem
cover (7 8 ha)
Rotational grazing,
little bluestem
cover (9 6 ha)
Continuous grazing,
little bluestem
cover, active
gullies (11 1 ha)
Rational grazing,
little bluestem
cover (11 0 ha)
Precipitation
cm/yr
109 1
77 3
99 4
52 4
76 5
78 2
76 5
78 2
Water
Runoff
cm/yr
17 8
4 2
7 2
35
14 7
4 3
10 2
4 3
Nitrogen Export (kg/ha/yr)
Dissolved Nltroqen Particulate/Sediment Nitrogen Total
N03-N NH4-N TKN-N ORG-N Total-N NOg-N NH^-N TKN-N ORG-N Total-N Nitrogen
2 02
95
2 30
15
68a 4 18a 8 52a 9 20
31 a 3 67a 4 41a 4 72
34a 16a 4 85a 5 19
20a 12a 1 53a 1 73
Reference
Menzel et al ,
1978
01 ness et al ,
1980
Olness et al ,
1980
Olness et al ,
1980
Olness et al ,
1980
Consists of both soluble and non-soluble fractions
-------
Table A5a: Phosphorus Export from Animal Feedlot and Manure Storage
Land Use
Beef livestock
feed lot
(4 76 ha)
Lamb feed lot
(21 32 ha)
Lamb feedloc
(12 63 ha)
Dairy confine-
ment, 45 head
of cattle
( 13 ha)
Beef and sheep
feed lot
( 603 ha)
Beef feeding
(1 6 ha)
Beef cattle
feed lot
9 29 m2/cow
( 0019 ha)
Precip-
itation
cm/yr
61 06
60 71
53 19
59 0
49 96
49 96
62 53
58 01
48 16
62 53
58 01
48 16
63 73
55.83
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Parti cul ate/Sediment
cm/yr PO^-P Total P Phosphorus
28.52
8 92
21 87
1 96
2 18
3 10
15 16
82 65
27 18
15 24
30 35
14 40
3 99
8 71
17 07
14 68
Total
Phosphorus
523 0
145 6
749 3
35 84
20 16
21 28
521 9
355 0
301 3
2635 4
222 9
157 9
29 1
142 2
1299 2
291 2
Reference
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
McCalla et al ,
1972
-------
Table A5a: (continued)
ID
Land Use
Beef cattle
feedlot,
18 6 m2/cow
( 0019 ha)
Beef cattle
feedlot,
18 6 nr/cow
( 0019 ha)
Beef cattle
feedlot,
500-600 cattle
( 245 ha)
Beef cattle
feedlot
( 165 ha)
Solid Manure
storage area
( 05 ha)
Precip- Water
itation Runoff
cm/yr cm/yr
16 59
19 28
24 59
25 30
70 7 33 2
78 6 173
67 37 20 9
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus Parti cul ate/Sediment Total
PO.-P Total P Phosphorus Phosphorus
470 4
224 0
313 6
134 4
163 0 425 0
73 0 170 0
86 0 172 0
Reference
McCalla et al ,
1972
Gilbertson et al ,
1975
Coote and More, 1978
Coote and More, 1978
Coote and More, 1978
Manure storage 57 7
facility ( 047 ha)
33 5
539 9 Magdoff et al ,
1977
Barnlot runoff,
370 cows/ha,
Ohio ( 17 ha)
31 9
27 9
14 11
19 98
Edwards et al ,
1972
-------
Table A5b: Nitrogen Export from Animal Feedlot and Manure Storage
Hater
Nitrogen Export (kg/ha/yr)
IO
Precipitation
i.ind Use cni/yr
Beef livestock
feedlot (4 76 ha)
Laub feedlot
(21 32 ha)
Lamb feedlot
(12 63 ha)
Dairy confinement,
45 head of cattle
( 13 ha)
Beef and sheep
feedlot
( 603 ha)
Beef feeding
(1 6 ha)
Beef cattle feedlot,
9 29 m2/cow
61 06
60 71
53 19
59 0
49 96
49 96
62 53
58 01
48 16
62 53
58 01
48 16
63 73
55 83
Runoff Dissolved Nitrogen
on/yr NQ3-N NII^-N TKH-N ORG-N Total -N
28 52
8 92
21 87
1 96
2 18
3 10
15 16
82 65
27 18
15 24
30 35
14 40
3 99
8 71
17 07
14 68
Partlculate/Sediment Nitrogen
N03-N NH.-N TKN-N ORG-N TotaJ-N
1332 8a
577 9a
1196 2a
32 48a
53 76a
64 96a
705 6a
1561 28a
1154 70a
973 3a
433 4a
287 84a
99 68a
975 40a
Total
Nitrogen Reference
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
Dornbush and
Madden, 1973
3830 4 McCalla et al ,
2016 0 1972
( 0019 ha)
-------
Table A5b: (continued)
I and Use
Water
Precipitation Runoff
Nitrogen Export (kg/ha/yr)
DissolvedNltrogen Partlculate/Sedtment Nitrogen Total
cm/yr cm/yr N03-N NH4-N TKN-N ORG-N Total-N N03-N NH4-N 1KN-N OR6-N Total-N Nitrogen Reference
Beef cattle feedlot
18 6 m2/cow
( 0019 ha)
16 59
19 28
1254 4
1433 6
McCalla et al ,
1972
Beef cattle feedlot,
18 6 mz/cow
( 0019 ha)
24 59
25 30
1388 8
2195 2
Gilbertson et al
1975
Beef cattle feedlot,
500-600 cattle
( 245 ha)
70 7
33 2
6 27 862 0 2504 0
3372 27 Coote and Hore,
1978
VD
Beef cattle feedlot
( 165 ha)
78 6
17 3
1 42 138 0 541 0
680 52 Coote and Hore,
1978
Solid manure
storage area
( 05 ha)
67 37
20 9
2 07 776 0 1112 0
1891 07 Coote and Hore,
1978
Manure storage
facility ( 047 ha)
57 7
33 5
5831 0°
7979 9 Magdoff et al ,
1977
Barnlot runoff
370 cows/ha
Ohio
( 17 ha)
31 9
27 9
6 45
3 04
66 86
107 09
Edwards et al ,
1972
a Consists of both dissolved and particulate fractions
-------
Table A6a: Phosphorus Export from Mixed Agricultural Watersheds
Land Use
58% row crops
31% small grain
Precip-
itation
cm/yr
112 0
70 0
Water
Runoff
cm/yr
27 5
11 2
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus
P04-P Total P
Parti cul ate/Sediment
Phosphorus
Total
Phosphorus
5 2
1 1
Reference
Lake and Morrison,
1977
and pasture
6% woods
5% urban (4950 ha)
10
CO
63% row crops
26% small grain
and pasture
8% woods
3% urban (942 ha)
112 0
70 0
29 1
12 4
14
06
24
09
5 20
98
5 4 Lake and Morrison,
1 1 1977
35% row crops
48% small grain
and pasture
5% woods
12% urban (714 ha)
112 0
70 0
26 0
10 1
34
18
46
22
4 5
73
5 0 Lake and Morrison,
1 0 1977
50% pasture 77 7 26 9
25% rotation 88 6 34 4
cropland 88 9 33 9
25% hardwood 92 7 32 8
forest (123 ha)
031
080
067
077
Taylor et al , 1971
39% corn 99 0
46% legumes and 93 5
grass, 9% small 92 7
grain, 2 6% idle
4% roads (594 ha)
0 10
0 80
0 60
Patni and More,
1978
-------
Table A6a- (continued)
Land Use
60% row crops
40% hay and
pasture
two livestock
feedlots
(157 5 ha)
Three years
pasture and
two years
corn (42 9 ha)
Intensive
agricultural
Precip-
itation
cm/yr
67 79
84 71
105 0
88 0
Water
Runoff
cm/yr
10 74
17 65
21 3
12 1
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus Parti cul ate/Sediment
P04-P Total P Phosphorus
319 329
19 08
1 21
63
Total
Phosphorus
648
27
1 34
86
Reference
Burwell et al ,
1974
Burwell et al ,
1977
Campbell, 1978
crops and
improved pasture
(202 ha)
Active cropping
and pasture
409 Gnzzard et al ,
1977
At least 80% of
watershed devoted
to agricultural
activities
233
1 29
Avadhanula, 1979
-------
Table A6a: (continued)
Land Use
37 4% soybean
and whitebean
27 1% cereal
23% corn
(5080 ha)
Precip-
itation
cm/yr
72 9
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus
cm/yr P04-P Total P
21
Parti cul ate/Sediment Total
Phosphorus Phosphorus
1 28
Reference
Coote et al , 1978
no
o
o
36 1% woodland
25 0% cereal
22 2% tobacco
10 1% corn
3% pasture
and hay
(7913 ha)
06
26
Coote et al , 1978
31 3% corn
26.4% cereal
17.9% pasture
and hay
12 1% soybean
and whitebean
7 5% woodland
(6200 ha)
37 2% pasture
and hay
35 3% cereal
18 7% corn
6 9% woodland
(1860 ha)
86 0
50
91
Coote et al , 1978
92 5
33
1 00
Coote et al , 1978
-------
Table A6a (continued)
Land Use
Precip-
itation
cm/yr
Water
Runoff
cm/yr
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus
P04-P Total P
Parti cul ate/Sediment
Phosphorus
Total
Phosphorus
Reference
42 3% corn
22 8% pasture
and hay
15 4% woodland
12 2% cereal
(3000 ha)
101 8
43
1 53
Coote et al , 1978
ro
o
33 4% pasture
and hay
29 2% woodland
22 3% cereal
12 3% corn
(5472 ha)
82 3
07
16
Coote et al , 1978
37 4% woodland
28.5% pasture
and hay
10 7% cereal
10.4% corn
3 7% tobacco
(5645 ha)
840
03
.08
Coote et al , 1978
44 2% pasture
and hay
18.4% cereal
17 8 Woodland
16 2% corn
(3025 ha)
77 9
.51
1 53
Coote et al , 1978
-------
Table A6a: (continued)
Land Use
41.3% pasture
and hay
29.0% cereal
11 3% corn
7 5% woodland
(2383 ha)
Precip-
itation
cm/yr
73.7
Water
Runoff
cm/yr
Phosphorus
Dissolved Phosphorus PC
P04-P Total P
20
Export (kg/ha/yr)
jrtlcul ate/Sediment
Phosphorus
Total
Phosphorus
49
Reference
Coote et al , 1978
ro
o
ro
27 8% vegetables
22 8% corn
10 0% woodland
8 9% cereal
7 9% soybean
and whitebean
(1990 ha)
77 0
36
91
Coote et al , 1978
66 6% pasture
and hay
12 1% cereal
9 5% corn
9 4% woodland
(4504 ha)
92
36
81
Coote et al , 1978
-------
Table A6b: Nitrogen Export from Mixed Agricultural Watersheds
fVJ
o
co
Precipitation
land Use cm/yr
58% row crops
31% small grain
and pasture
6S» woods
5% urban (4950 ha)
63% row crops
26% small grain
and pasture
8% woods
3% urban
(942 ha)
35% row crops
48% small grain
and pasture
5% woods
12% urban (714 ha)
50% pasture
25% rotation crop-
land,
25% hardwood forest
(123 ha)
39% corn
46% legumes and grass
9% small grain
112 0
70 0
112 0
70 0
112 0
70 0
77 7
88 6
88 9
92 7
49 0
93 5
92 7
Water Nitrogen Export (kg/ha/yr)
Runoff Dissolved Nitrogen Particulate/Sedjment Nitroqen
cm/yr N03-N NH4-N TKN-N ORG-N Total-N NOj-N NH^-N TKN-N ORG-N Total-N
27 5
11 2
29 1
12 4
26 0
10 1
26 9
34 4
33 9
32 8
12 3a 6 3a
8 3a 16 Oa
5 6a 2 6a
Total
Nitrogen
48 7
8 6
53 2
10 3
44 1
6 6
1 67
3 11
10 61
4 38
18 60
24 20
8 20
Reference
Lake and Morrison,
1977
Lake and Morrison,
1977
Lake and Morrison,
1977
Taylor et al ,
1971
Patm and More,
1978
2 6% idle
4% roads (594 ha)
a Consists of both dissolved and particulate fractions
-------
Table A6b: (continued)
land Use
Precipitation
cm/yr
Hater
Runoff
cw/yr
Nitrogen Export (kg/ha/yr)
Dissolved nitrogen
N03-N NH^-fl TKH-N ORG-fl Total-N
Partlculate/Sedlwent Nitrogen iota]
Ji03-H NH.-H TKN-N ORG-N Total-N Nitrogen
Reference
60% row crops
40% Jiay and pasture
two livestock
feedlots
(157 5 ha)
67 79
10 74 1 19 1 37
7 08
9 64 Burwell et al ,
1974
Three years
pasture and two
years corn
(42 9 ha)
84 71
17 65 11 68 40
2 03 14 11 Bur-well et al ,
1977
ro
o
Intensive
agriculture crops
and improved
pasture (208 ha)
105 0
88 0
21 3 37 68
12 1 09 09
S 3
1 92
6 36 Campbell, 1978
2 10
Active cropping
and pasture
2 83 Gnzzard et al ,
1977
At least 803! of
watershed devoted
to agricultural
activities
14 3 Avadhanula, 1979
37 4% soybean and
wTiitebean
27 U cereal
23% corn (5080 ha)
72 9
10 7
5 3
16 1
Coote et al
1978
-------
Table A6b (continued)
Land Use
36 1% woodland
25 0% cereal
22 2% tobacco
10 1% corn
3% pasture and
hay (7913 ha)
Water
Precipitation Runoff
cm/yr cm/yr
Nitrogen Export (kg/ha/yr)
Dissolved Nitrogen Parti dilate/Sediment Nitrogen
N03-N NH4-N TKN-N ORG-N Total-N N03-N NH4-N tKN-N ORG-N tbtal-N
43 22
Total
Ni trogen
6 4
Reference
Coote et al ,
1978
o
en
31 3% corn
26 4% cereal
17 9% pasture
and hay
12 1% soybean
and whitebean
7 5% woodland
(6200 ha)
37 2% pasture
and hay
35 3% cereal
18 1% corn
6 9% woodland
(1860 ha)
92 5
37 4
14 9
4 2
5 4
41 5
Coote et al
1978
20 3
Coote et al
1978
42 3% corn
22 6% pasture and
hay
15 4% woodland
12 2% cereal
(3000 ha)
101 8
24 1
7 1
31 1
Coote et al
1978
-------
Table A6b: (continued)
1 and Use
33 4% pasture
and hay
29 2% woodland
22 3% cereal
12 3% corn
(5472 ha)
Precipitation
cm/yr
82 3
Hater
Runoff
on/yr
Nitrogen Export (kg/ha/yr)
Dissolved ((Kronen Particulate/Sedlraent Nitrogen
N03-H NH^-N TKN-N ORG-H Total-N NOg-N NH.-N TKN-N ORG-N Total-N
11 3 29
Total
Nitrogen
14 3
Reference
Coote et al ,
1978
ro
o
37 4% woodland 84 0
28 5% pasture
and hay, 10 4% corn
10 7% cereal
3 7% tobacco
(5645 ha)
44 2% pasture 77 9
and hay
18 4% cereal
17 8% woodland
16 2% corn
(3025 ha)
2 1
7 0
1 1
8 5
3 2
15 5
Coote et al
1978
Coote et al
1978
41 3% pasture
and hay
29 0% cereal
11 3% corn
7 5% woodland
(2383 ha)
73 7
8 3
2 0
11 1
Coote et al
1978
-------
Table A6b: (continued)
Precipitation
land Use cm/yr
27 8% vegetables 77 0
22 83! corn
10 0% woodland
8 9% cereal
7 9% soybean
and whitebean
(1990 ha)
66 6% pasture 92 4
and hay
12 1% cereal
9 55! corn
9 4% woodland
^o (4504 ha)
o
Water Nitrogen Export
(kg/ha/yr)
Runoff Dissolved Nitrogen Particulate/Sediment Nitrogen Total
cm/yr N03-N NH4-N TKN-N ORG-M Total -N NOj-N
21 0
5 4
NH4-N TKN-N OR6-N Total-N Nitrogen Reference
42 25 2 Coote et al ,
1978
41 94 Coote et al ,
1978
-------
Table A7a: Phosphorus Export from Urban Hatersheds
o
CO
Land Use
Residential
(50 ha)
78% industrial
22% commercial
Commercial
(15 8 ha)
Central business
district (9 3 ha)
Industrial
(8 1 ha)
Residential
(41 7 ha)
Low density
residential
(46 82 ha)
Low density
residential
(33 73 ha)
Precip-
itation
cm/yr
69 93
76 5
76 5
76 5
76 5
77 19
77 19
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Particulate/Sediment Total
cm/yr PO.-P Total P Phosphorus Phosphorus
10 49 64 1 10
7 88 1 06
40 18 4 28
64 88
3 58 4 08
62 75
27 35
0 19
2 7
Reference
Kluesener and Lee,
1974
Konrad et al ,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Landon, 1977
Landon, 1977
-------
Table A7a: (continued)
IN}
o
Land Use
High density
residential
(7 ha)
High density
residential
(21 63 ha)
Commercial
(T8 T9 ha)
Commercial
(4 19 ha)
Precip- Water
itation Runoff
cm/yr cm/yr
77 19
77 19
77 19
77 19
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus Parti cul ate/Sediment Total
PO.-P Total P Phosphorus Phosphorus
1 1
56
1 7
.66
Reference
Landon, 1977
Landon, 1977
Landon, 1977
Landon, 1577
64% residential
13% recreational
12% commercial
6% transportation
1% industrial
(958 ha)
,757 O'NefH, 1979
Residential and
light commercial
(11 ha)
76 2
28.19
.90
Weibel et al ,
1964
-------
Table A7a: (continued)
Land Use
At least 60%
of watershed
devoted to
urban land use
Industrial and
residential
(414 ha)
Commerci al
(212 ha)
Suburban
(62 ha)
Precip-
itation
cm/yr
150 0
155 0
153 0
Water Phosphorus Export (kg/ha/yr)
Runoff Dissolved Phosphorus Parti cul ate/Sediment
cm/yr P
-------
Table A7a: (continued)
Land Use
Single family
residential
(19.2 ha)
67% residential
13% commercial
Precip-
itation
cm/yr
125 6
249 0
/
Hater
Runoff
cm/yr
9 42
48 3
Phosphorus Export (kg/ha/yr)
Dissolved Phosphorus
P04-P Total P
18 22
Parti cul ate/Sediment
Phosphorus
Total
Phosphorus
0 21
6 23
Reference
Mattraw and Sherwood,
1977
Burton et al ,
1977
ro
12% woodland
8% agricultural
(792 ha)
Residential
(6 8 ha)
113 06
18 99
17
22
60 Simpson and Hemens,
1978
74 7% residential
12 6% institu-
tional
7 4% industrial
5 3% commercial
(2261 ha)
Tulsa, Oklahoma
94 61
15.14
92
AVCO, 1970
-------
Table A7b: Nitrogen Export from Urban Uatersheds
1 and Use
Residential (50 ha)
Commercial
(15 8 ha)
Central Business
district (9 3 ha)
Industrial
(8 1 ha)
i Residential
ro (41 7 ha)
Low density
residential
(46 82 ha)
Low density
residential
(33 73 ha)
High density
residential
(7 ha)
Precipitation
on/yr
69 93
76 5
/6 5
76 5
76 5
77 19
77 19
77 19
Hater Nitrogen Export (kg/ha/yr)
Runoff Dissolved Nitrogen Partkulate/Sediipent Nitrogen
cro/yr N03-H NH^-N TKN-N ORG-N Total -N NOj-N NH.-N TKN-N ORG-N Total-N
10 49 67 50
1 74a 59a 2 80a
10 36a 6 98a 28 lla
2 04a 2 36a 4 49a
1 19a 72a 2.48a
82a 70a
2 9a 4 Oa
2 Oa 2 8a
Total
Nitrogen
5 00
4.54
38 47
6 53
3 67
1 52
6 9
4 8
Reference
Kluesener and Lee,
1974
Much and Kemp,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Much and Kemp,
1978
Landon, 1977
Landon, 1977
Landon, 1977
-------
Table A7b (continued)
land Use
Hater
Precipitation Runoff
Dissolved Nitrogen
cm/yr cra/yr N03-N NH4-N TKN-N ORG-N Total -N NOj
Nitrogen Export (kg/ha/yr)
Parti culate/Sedlment Nitrogen"Total
,-N NH.-N TKN-N
-N ORG-N Total-N Nitrogen Reference
High density
residential
(21 63 ha)
77 19
2 2"
2 3a
5 5 Landon, 1977
Commercial
(18 19 ha)
77 19
12 0°
8 5°
20 5 Landon, 1977
Commercial
(4 19 ha)
77 19
1 8°
2 2°
4 0 Landon, 1977
ro
__i
co
Residential and
light commercial
(11 ha)
76 2 28 19
9 97 Weibel et al ,
1964
At least 60% of
watershed de-
voted to urban
land use
3 05
9 48 Avadhanula, 1979
Industrial and
residential
(414 ha)
150 0 84 3
5 62a 83a
8 5"
14 95 Betson, 1978
Commercial
(212 ha)
155 0 41 1
3 14a 44a 9
12 78 Betson, 1978
Suburban
(62 ha)
153 0
9 4
47
98°
1 56 Betson, 1978
-------
Table A7b: (continued)
ro
Precipitation
1 and Use ctn/yr
60% residential
19% commercial and
industrial
12% institutional
10% unused (432 54 ha)
20% urbanized
large scale
residential
(47900 ha)
Single family 125 6
residential
(19 2 ha)
67% residential 249 0
13% commercial
12% woodland
8% agricultural
(792 ha)
Residential (6 8 ha) 113 06
Hater Nitrogen Export (kg/ha/yr)
Runoff Dissolved Nltroqen Particulate/Sedlroent Nltroqen Total
cm/yr N03-N NH^-N TKN-N ORG-N Total-N NOg-N NH.-N TKN-N ORG-N Total-H Nitrogen Reference
24 64 6 83 Colston, 1974
33 76 Gnzzard et al ,
1978
9 42 1 48 Mattraw and
Sherwood, 1977
48 3 24 17 Burton et al ,
1977
18 99 40 Simpson and
Hemens, 1978
a
o
74 7% residential 94 61 15 14
12 6% institutional
7 4% industrial
5 3% commercial
(2261 ha) Tulsa,
Oklahoma
a Consists of both dissolved and particulate fractions
2 16
AVCO, 1970
------- |