EPA// 600/4-83-030 June .83
GUIDELINES FOR FIELD TESTING
AQUATIC FATE AND TRANSPORT MODELS:
FINAL REPORT -
by
Stephen C. Hern, George T. Flatman, Wesley L. Kinney,
and Frank P. Beck, Jr.
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Las Vegas, Nevada 89114
James E. Pollard
Department of Biological Sciences
University of Nevada, Las Vegas
Las Vegas, Nevada 89154
and
Alan B. Crockett
Earth and Life Sciences Branch
EG&G Idaho, Inc.
Idaho Falls, Idaho 83415
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
EPA// 600/4-83-030 June .83
GUIDELINES FOR FIELD TESTING
AQUATIC FATE AND TRANSPORT MODELS:
FINAL REPORT
by
Stephen C. Hern, George T. Flatman, Wesley L. Kinney,
and Frank P. Beck, Jr.
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Las Vegas, Nevada 89114
James E. Pollard
Department of Biological Sciences
University of Nevada, Las Vegas
Las Vegas, Nevada 89154
and
Alan B. Crockett
Earth and Life Sciences Branch
EG&G Idaho, Inc.
Idaho Falls, Idaho 83415
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
NOTICE
This paper has been reviewed In accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved
for presentation and publication.
11
-------
ABSTRACT
This guidance has been developed for the EPA's Office of Pesticides and
Toxic Substances (OPTS) to aid in field validation of aquatic fate and
transport models. The OPTS anticipates that several potentially useful
models will need to be validated for application to a variety of compartmental
aquatic fate and transport models, and, therefore, the guidance is general.
The guidelines are not a set of "cookbook" instructions but rather discussions
of the major steps in model validation, field plan development, and data
Interpretation. They also include overviews of environmental fate processes
and a discussion of specific field sampling methods.
Validation of a model is defined in this report as a comparison of
model results with numerical data derived from observations of the environment.
Complete model validation requires testing over the full range of conditions
for which predictions are intended. At a minimum, this requires a series of
validations in various aquatic environments (streams, lakes, estuaries),
with chemicals that typify the major fate and transport processes (biotrans-
forms t ion, hydrolysis, oxidation, photolysis, sorption, ionization, volatiliza-
tion, bioconcentration, and physical transport). Validated models are
useful in the regulatory process because they withstand scientific scrutiny and
are defensible in courts of law. However, models are ultimately judged by
their usefulness to the user rather than against a scientific standard. Thus,
by this definition, a model may be valid for one use but invalid for another.
Fate and transport models may be based upon either an empirical approach
or a theoretical approach that considers transport and fate processes.
Empirical models which are based on extensive field observation are usually
calibrated to specific existing sites and chemicals and provide no rational
basis for making predictions outside their range of prior observation. Thus,
models of this type are generally not suited for predicting the fate of new
chemicals and are not considered in these guidelines.
The theoretical approach is based upon an understanding of environmental
fate and transport processes. This type of model is considerably more
versatile since it is designed to predict environmental fate and pollutant
concentrations based upon degradation rate constants and relatively simple
chemical and environmental input data. Therefore, theoretical models can
be applied to chemicals which haven't yet been Introduced into the environment.
Such models are of considerable Interest to the U.S. Environmental Protection
Agency and in particular, the OPTS.
This guidance Is designed to be useful in validating a variety of models
but is largely based upon experience gained during field validations of
the EXAMS model. Validations of this model were conducted and this experience
used to modify and improve the document. Included in the document are
ill
-------
discussions of the major steps in validating models and sections on the
Individual fate and transport processes. For each process the following
information is provided: a general description of the process, a list and
discussion of environmental factors affecting the process, a list of the
priority pollutants for which the process is important, a list of model-specific
environmental inputs, and field methods for collecting these input data.
iv
-------
CONTENTS
Abstract lii
Figures vii
Tables viii
List of Contributors x
Introduction 1-1
Steps in the Field Validation of Aquatic Fate and Transport Models. . 1-4
Step 1: Identify Model User's Needs 1-4
Step 2: Develop Acceptance Criteria for Validations 1-4
Step 3: Examine the Model 1-5
Step 4: Evaluate the Feasibility of Field Validation 1-5
Step 5: Determine Field Validation Scenario 1-6
Step 6: Plan and Conduct Field Validation-Steps 1-9
a. Select Site and Compound(s). 1-9
Compound-Selection Factors 1-9
Site Selection Factors 1-11
Approaches to Selecting Compounds and Sites . . . 1-12
b. Collect Preliminary Data Sets and Conduct Sensitivity
Analyses .............. 1-13
Sensitivity 1-13
c. Develop a Field Study Design 1-16
Validation Scenario 1-18
Rate Constant Information 1-18
Model Assumptions 1-18
Type and Number of Samples 1-23
Compartmentalizatlon 1-23
Design Features to Eliminate Other Problems . . . 1-24
Sampling Location 1-24
Quality Assurance 1-29
Quality Assurance References 1-30
Evaluation of Literature and Unpublished Data . . 1-31
Supplemental Data Acquisition 1-31
Sampling 1-32
d. Conduct Field Study not addressed in this document
e. Analyze Samples. Selected references for methods are
presented in the Sample Collection, Handling and
Analysis section
f. Compare Model Performance with Acceptance Criteria . . 1-32
Data to be Compared 1-32
Graphical Comparison 1-33
Statistical Comparisons .... 1-36
-------
CONTENTS (Continued)
Page
Summary 1-41
References 1-42
Model Inputs and Outputs 2-1
References 2-6
Biotransformation 3-1
References 3-14
Hydrolysis 4-1
References 4-4
Photolysis 5-1
References 5-8
Oxidation . 6-1
References 6-4
lonization 7-1
References 7-6
Volatilization 8-1
References 8-9
Sorption 9-1
References 9-10
Bioconcentration 10-1
References. .......... 10-11
Physical Transport 11-1
References 11-38
Pollutant Loading Inputs .. ..... 12-1
References 12-8
Sample Collection, Handling and Analysis for Toxic Substances 13-1
References 13-3
Appendices
A. Time and Distance Calculation that the Compound of A-l
B. Bioconcentration Methods B-l
vi
-------
FIGURES
Number Page
1-1 Field validation matrix 1-8
1-2 Effluent sampling 1-21
1-3 Simulating compliance of steady state loading 1-21
1-4 Pollutant decay curve 1-27
1-5 Cross section of a river showing sections defined by depth
integrated samples 1-29
1-6 Regression lines and confidence bands for model and field data
with acceptance band 1-34
1-7 Graphic error analysis 1-35
8-1 Plot of solubility, vapor pressure, and Henry's Constant depicting
relative volatilization rates 8-4
vii
-------
TABLES
Number • Page
1-1 Steps in Field Validation of Aquatic Fate and Transport Models. . 1-2
1-2 EXAMS Concentration Output and Relative Sensitivity for Ranges
of pH Input for a Compound Which Degrades Primarily by
Hydrolysis 1-15
1-3 Checklist for Field Protocol Preparation 1-16
1-4 Assumptions Commonly Associated with Compartmentalized Aquatic
Fate and Transport Models ................... 1-19
1-5 Design Problems and Possible Solutions 1-25
2-1 Environmental Inputs by Process, to Aquatic Fate and Transport
Models 2-2
2-2 Pollutant Loading Inputs to Aquatic Fate and Transport Models . . 2-5
2-3 Pollutant Concentrations Predicted by Aquatic Fate and
Transport Models 2-5
3-1 Nutrients Required by Certain Microorganisms 3-3
3-2 Major Environmental Factors that Influence Degradation Rates. . . 3-5
3-3 Organic Priority Pollutants Whose Aquatic Fate is Significantly
Affected by Microbial Degradation 3-8
4-1 Priority Pollutants for Which Hydrolysis is Estimated to be an
Important Fate Process 4-3
5-1 Summary Table of Environmental Factors that Influence
Photolysis 5-3
5-2 Priority Organic Pollutants for Which Photolysis is
Considered an Important Fate Process 5-4
5-3 Comparison of Direct Measurement Techniques and Spectral
Radiometer Methods 5-5
6-1 Organic Priority Pollutants for Which Oxidation is Estimated
to be an Important Fate Process 6-2
viii
-------
TABLES (Continued)
Number Page
7-1 Environmental Factors that Influence lonization of Organic
Compounds in Water 7-4
8-1 Organic Priority Pollutants for Which Volatilization is
Estimated to be an Important Transport Process 8-5
8-2 Environmental Factors that Influence Volatilization Rates .... 8-5
9-1 Influence of Particle Size on Partition Coefficients 9-3
9-2 Environmental Factors and Sorbent Properties Which Most
Influence Sorption Rates 9-6
9-3 Organic Priority Pollutants for Which Sorption is Estimated to
be an Important Process 9-6
10-1 Organic Priority Pollutants Most Subject to Bioconcentration. . . 10-4
10-2 Principal Environmental Factors and Biological Components that
Influence Bioconcentration of Organic Compounds in Aquatic
Systems .............. 10-6
11-1 Outflow Concentration Divided By/Inflow Concentration at Steady
State as a Function of Number of Compartments and KT 11-10
11-2 Site Specific Environmental Inputs to Aquatic Fate and
Transport Models for Physical Transport 11-16
ix
-------
LIST OF CONTRIBUTORS
Jerald L. Schnoor, University of Iowa, Iowa City, Iowa
Frederick H. F. Au, EPA, EMSL-LV, Las Vegas, Nevada
Joseph V. Behar, EPA, EMSL-LV, Las Vegas, Nevada
William R. Mabey, SRI, Menlo Park, California
Clyde W. Frank, EG&G, Idaho Inc., Idaho Falls, Idaho
Chris S. Staley, EG&G, Idaho Inc., Idaho Falls, Idaho
Bryant C. Hess, Las Vegas Valley Water District Las Vegas, Nevada
-------
INTRODUCTION
Aquatic fate and transport models are being developed that predict the
fate and concentration of chemicals in natural waters. Such models may be
based upon either an empirical approach or a theoretical approach that
considers transport and fate processes. Empirical models are based on
extensive field observation, usually calibrated to specific existing sites
and chemicals, and provide no rational basis for making predictions outside
their range of prior observation. Thus, models of this type are generally
not suited for predicting the fate of new chemicals and are not considered
in these guidelines.
The theoretical approach is based upon an understanding of environmen-
tal fate and transport processes Including biotransformation, hydrolysis,
oxidation, photolysis, ionization, sorption, volatilization, bioconcentration,
and physical transport. This type of model is considerably more versatile
since it is designed to predict environmental fate and pollutant concentrations
based upon degradation rate constants and relatively simple chemical and
environmental input data. Therefore, theoretical models can be applied to
chemicals which haven't yet been Introduced into the environment. Such models
are of considerable interest to the U.S. Environmental Protection Agency and
in particular, the OPTS.
The guidance provided is organized into sections with the first addressing
steps in the validation process while subsequent sections cover the environ-
mental fate processes and field methods for collecting environmental input
and output data. The major steps in the validation process are outlined in
Table 1-1 and are discussed at length in the first section of this report.
Much of the Information presented relates to the identification of
potential problems associated with field validation of models. Validation
of a model is defined in this report as comparison of model results with
numerical data derived from observations of the environment. Complete
model validation requires testing over the full range of conditions for
which predictions are Intended. At a minimum, this requires a series of
validations in various aquatic environments (streams, lakes, estuaries)
with chemicals that typify the major fate and transport processes. Where
possible, suggested solutions or approaches have been presented but many
problems are specific to a particular site, compound, or model and will
have to be dealt with on a case-by-case basis. In addition, a model develop-
ment and subsequent validation is a dynamic process by its very nature.
Future research regarding fate and transport processes will refine fundamental
theory. This information can then be used to improve existing models. As
models are updated, the methods to measure model inputs and outputs must
be changed to reflect model Improvements.
1-1
-------
Sections describing fate and transport processes include the following
information: a general description of the process, a list and discussion
of environmental factors affecting the process, a list of the priority
organic pollutants for which the process is important, a list of model
specific environmental inputs and finally, field methods for collecting
these input data. The report also briefly covers the collection of input
loading data, field sampling for predicted model outputs, and quality assurance.
Experience gained during field validations of the EXAMS model (Pollard et
al., 1983, Hern et al., in press) was used to modify and improve this
document. The guidance provided by this document was constructed for simpli-
fied aquatic fate and transport models, e.g. EXAMS. However, the steps in
field validation and many of the environmental measurement techniques
would apply to all aquatic fate and transport models.
TABLE 1-1. STEPS IN FIELD VALIDATION OF AQUATIC FATE AND TRANSPORT MODELS
Step 1. Identify Model User's Need: The first step in field validation is to
obtain a clear understanding of the model user's need i.e., how will
the model be used.
Step 2. Develop Acceptance Criteria for Validations: The model user must
provide criteria against which the model is to be Judged.
Step 3. Examine the Model: This step involves a detailed examination of the
model to precisely define input data requirements, output predictions
and model assumptions.
Step A. Evaluate the Feasibility of Field Validation: Some models can not be
validated in the field and the validator should consider this
possibility.
Step 5. Determine Field Validation Scenario: Many different approaches to field
validation are possible. A scenario should be identified and
approved by the model user.
Step 6. Plan and Conduct Field Validations Which Should Include the
Following Steps:
Step 6a. Select a Site and Compound(s): There are many
important factors to consider in selecting a site and
compound(s).
Step 6b. Collect Preliminary Data and Performance of Sensitivity
Analysis:. Preliminary data are required to conduct a
sensitivity analysis and determine the most important
input variables.
(continued)
1-2
-------
TABLE 1-1. (Continued)
Step 6c. Develop a Field Study Design: Development of a detailed
field sampling plan for the specific model compound and
site.
Step 6d. Conduct Field Study: Implementation of the field plan
is not addressed in these guidelines.
Step 6e. Analyze Samples: Many analytical procedures are available
depending on the chemical and the matrix. Validated
methods should be used together with a sound quality
assurance program. Selected references for analytical
methods are presented in the Sample Collection, Handling
and Analysis section.
Step 6f. Compare Model Performance with Acceptance Criteria: A
comparison must be made between the model's performance
and the user's acceptance criteria.
1-3
-------
STEPS IN THE FIELD VALIDATION OF AQUATIC FATE AND TRANSPORT MODELS
STEP 1: IDENTIFY MODEL USER'S NEED
The first step in model validation is to obtain a detailed understanding
of the problem confronting a potential model user. This understanding should
be acquired through direct discussion with the user. The model validator
should elicit from the user how the model will be used, what model input data
will be available, how such data will be acquired, and what the expected model
outputs are (Donigian 1980).
This problem identification step is important because a valid model is
one which is capable of meeting the user's needs. Thus, it is essential to
have a thorough understanding of the problem to make an assessment regarding
the model's utility. If the model is to be used in several ways or for
several different purposes, each use or purpose needs to be defined at the
outset. Just because a model has been found valid for one use does not mean
it is valid for some other use. The validator may discover, after a detailed
review of the model, that the model cannot be used as proposed because of the
model's assumptions, the cost of obtaining the input data, etc., and thus save
the user considerable time and expense involved In attempting a field
validation.
STEP 2: DEVELOP ACCEPTANCE CRITERIA FOR VALIDATIONS
Prior to attempting to validate a model, the user should develop and
provide to the validator the criteria that will be used to accept or reject
the model. After identifying the problem and determining how the model may
assist in resolving the problem, the user should have a good idea as to the
accuracy and precision required of the model. In developing acceptance
criteria, the user should consider the predictive ability of alternate methods
available to solve the problem, e.g., a model ecosystem. The development of
such criteria, however, should not result in either automatic acceptance if the
criteria are met or automatic rejection if the criteria are not met (Davis
1980).
The acceptance criteria for a validation should be given in terms of
required accuracy, precision and confidence interval. An example of
acceptance criteria is as follows: a user wants to estimate the level of some
pollutant in sediment within + one order of magnitude and be correct 95
percent of the time. On the other hand, the criteria for the level of a
pollutant in water may be + 100 percent with the same confidence level since
water is of greater concern to the user.
Consideration should be given to the uncertainty associated with model
predictions. Knowledge of the confidence Interval associated with model
predictions "is of considerable practical and theoretical Importance,
permitting quantitative comparison of model ouput and validation data, clearly
indicating the resolution or precision of the model prediction, and placing
model predictions in the proper context" (O'Neill and Gardner 1979). Ideally,
1-4
-------
the user's desired confidence interval about the model's predictions will
fall within the accuracy acceptance interval about the field data.
Lack of acceptance criteria could result in attempts to validate a
model that could not possibly satisfy the user. It may be appropriate to
reject a model prior to any field testing after considering the effect of
input sampling errors on the model's predictions. For example, sensitivity
analyses might indicate that the input rate constants would have to be
determined an infeasible number of times to achieve satisfactory confidence
in the output estimates.
STEP 3: EXAMINE THE MODEL
This examination should concentrate on determining the required inputs to
the model (chemical, biological, and physical), the predictions made by the
model and the assumptions made in the construction of the model. A model
input should be determined only for each environmental fate process by which
the compound of interest degrades, e.g., photolysis, hydrolysis, etc.
In examining model Inputs, it is important to precisely define each input
and determine the units of measure. For example, "suspended particulate"
might be defined to include particulate organic matter or only inorganics and
the units could be mg/1 dry, wet, or ash free dry weight. Other considera-
tions include determining whether a model's input values should represent
an average for some compartment or box, an average from some river or lake
cross-section, etc., and whether these averages represent one point In
time or some length of time. It is essential to precisely identify all model
Inputs to ensure a fair test of the model.
Model outputs must also be carefully defined. A prediction of x ppm in
water could refer to whole water including organic and inorganic particulate
matter or it might mean filtered or centrifuged water. A pollutant
concentration in fish may be based upon whole body wet weight, whole body dry
weight, edible portion only, etc.
The final major considerations In model examination are the assumptions
upon which the model is based. Assumptions are frequently made to simplify a
model and this usually restricts the situations to which it is applicable.
Some model assumptions are considerably more important than others and
attempts are frequently made to verify a model under conditions that violate
some of its assumptions. However, serious violation of an assumption will
likely result in an invalid test of the model. A more detailed discussion of
model assumptions and designing around them is found in the section on
Development of a Field Plan.
STEP 4: EVALUATE THE FEASIBILITY OF FIELD VALIDATION
Field validation is probably the most credible test of a model but not
all models should be field tested. Some of the reasons that a model cannot be
validated relate to model assumptions, the inability to provide input data,
the inability to quantify model outputs in the field, and large sampling
errors associated with model inputs. The model validator has an obligation to
1-5
-------
determine whether field validation should be attempted. The results of such
inappropriate testing can have a significant impact upon the use and
credibility of a model even if the tests were poorly designed and implemented.
The foregoing is not intended to imply that field validation should not
be attempted, only that caution and judgment should be exercised. There are
means to design around some model assumptions, indirectly collecting some
input and output data and improving model precision. The following are
examples of Instances where the feasibility of field validation may be
questioned.
• A model that assumes steady state or a dynamic equilibrium may be
difficult to field validate fully. Steady state conditions do not
exist for long in the environment; temperature, flow rates,
pollutant loads, bacteria population, etc., are almost always
changing. Steady state may be approximated for brief periods in
rivers but it may be Impossible to achieve such conditions for
compounds that have long half-lives.
• In some models, input data may not be quantifiable. Examples of
such inputs are the active pollutant degrading fraction of the total
bacteria population and the molar concentration of oxidants in water
that are capable of causing induced photolysis. No standard or
routine procedure exists for directly acquiring these data although
it is sometimes possible to obtain such data indirectly.
• As an example of uncollectable output data, consider a model that
predicts the concentrations of organic pollutants on clay.
Currently, no methods exist for separating clay from sand and silt
without affecting the concentration of volatile or semi-volatile
organic compounds.
• An example where it would be infeasible to field test a model due to
a large input sampling error is as follows: assume a degradation
rate constant is measured several times and the results vary by
orders of magnitude. If this input is sensitive in the model (a
small change in the input results in a large change in the output)
the resultant prediction will also have a very large sampling error;
probably much greater than the user's acceptance criteria for
precision.
In summary, it is important to realize that field validation is not
always practical. Persons involved in possible field tests must make this
determination on a case-by-case basis.
STEP 5: DETERMINE FIELD VALIDATION SCENARIO
There are many possible field validation scenarios that could be
proposed. Below, three possible scenarios are considered and the various
1-6
-------
attributes and limitations of each are discussed. The selection of a scenario
for attempting to validate a model should be made by the model user. The
basis of the decision will involve consideration of the degree of credibility
required, the feasibility of field testing all environmental processes consid-
ered by the model, the length of time available for testing, environmental
impacts, and finally, costs. The scenarios to be discussed are: 1) laboratory
testing supported by field validations; 2) controlled field validation; and 3)
uncontrolled field validation. ;
Laboratory Testing Supported By Field Validations
This approach involves extensive testing of a model and each of its pro-
cesses in controlled environments followed by a few field tests using compounds
which undergo several environmental processes. The controlled environments
may be fairly simple laboratory microcosms or complex artificial streams such
as the ones at the EPA's Environmental Research Laboratory at Athens, Georgia.
Such systems offer a relatively high degree of environmental control with
varying degrees of complexity and "real world" simulation.
They also permit detailed study of processes at relatively low expense,
and can avoid some problems associated with model assumptions. The limitations
of the scenario are that the results of tests In artificial environments lack
credibility since artificial environments only approximate the "real world,"
and limited field testing may not detect significant model deficiencies result-
ing from extrapolation of laboratory studies to the field. A few field tests
on compounds with complex environmental fate do not adequately test each
process in the model but would add to the credibility of the laboratory tests.
*
Controlled Field Validation
Controlled field validation, as envisioned here, entails dosing natural
or seminatural aquatic systems with selected compounds to test each environ-
mental fate process. Each process in the model is actually a sub-model and
needs to be tested separately. It is entirely possible for the model of one
process to be in error while all other processes are acceptably accurate.
An environment versus process matrix (Figure 1-1) should be developed and
the required validations conducted. Many of the tests required to complete
such a matrix could be conducted in a single study. For example, a small pond
might be dosed with low levels of relatively non-toxic compounds with each
typifying one of the following environmental processes: oxidation, hydrolysis,
photolysis, blotransformatlons, ionlzation, volatilization, sorption, biocon-
centratlon, and physical transport. Thus, if all compounds were added to a
single pond, all the processes of the model pertaining to such an environment
could be simultaneously tested. Similar studies would, of course, have to be
conducted on other types of aquatic environments. A few such studies could
yield very credible tests in natural ecosystems at a relatively economical
cost compared with uncontrolled field validation. Controlled testing of this
type necessitates the addition of contaminants to aquatic systems. This may
be scientifically justifiable and environmentally acceptable, yet politically
1-7
-------
impossible. There are, however, several ways to avoid or minimize problems
associated with intentional dosing. Isolated locations might be used to
avoid exposure to the public or the effluent from a dosed pond could be
cleaned up. An example of the latter is the seminatural stream currently
being used at EPA's Monticello, Minnesota laboratory. Natural river water
is dosed, run through an artificial stream bed and the water is treated
before it is discharged (under permit) back to the river.
./
Turbid River
Stream
Pond
Oligotrohic Lake
Eutrophic Lake
Estuary
*
Figure'1-1. Field validation matrix.
Many of the environmental hazards associated with working with priority
pollutants can be avoided by using deuterium labeled natural water components
or relatively nontoxlc compounds, e.g., ethanol could be used to test
biotransformation instead of the priority pollutant, phenol. In summary,
the controlled field approach is a relatively inexpensive approach to
obtain credible field data but it may entail some political considerations.
Uncontrolled Field Validation
Uncontrolled field validation involves attempting to test each process of
an aquatic fate and transport model using natural water bodies and uncontrolled
(by the validator) contaminant loading. Such sites might be a river, lake or
pond receiving contaminant loads from direct Industrial discharge, non-point
sources (urban and agricultural runoff), atmospheric deposition, contaminated
ground water infiltration, municipal waste water discharges, other surface
water flows, etc. The main advantage of field validation using this scenario
is one of credibility. Having validated a model under "real world" conditions
is frequently regarded as the ultimate test of a model. The more complex
1-8
-------
the test with multiple sources of various types, the more credible is the
test. However, such increased complexity generally adds to the cost of
testing, increases the sources of errors, may result in compensating errors,
makes data interpretation more difficult, and yields less precise predictions.
This approach to model validation should be based on the same matrix of
validation tests as described in the previous scenario (Figure 1-1). Field
validation using this matrix consists of a series of tests using selected
sites and compounds that typify each environment and process. A model
that meets the user's acceptance criteria for each test in the matrix
would then be considered valid. The number of replications for each box in
the matrix depends upon the level of confidence desired by the user as well
as the resources available. This approach requires a large number of field
tests to fill in the matrix. Therefore, it would probably only be possible
to test a few processes at any one site. This method then becomes very
resource intensive.
Another consideration In this scenario is the selection of validation
sites. Most of the organic chemical discharge data pertains to the priority
pollutants. Because of EPA's interest in these pollutants and their potential
adverse effects, efforts are being made to minimize their release. This
results In low pollutant levels in effluents and frequently undetectable
levels In receiving waters.
STEP 6: PLAN AND CONDUCT FIELD VALIDATIONS
This section addresses the design of field studies to test a process type
aquatic fate and transport model. At this stage the model has been examined
and all required environmental and chemical inputs identified by process. The
outputs of the model are known and all the assumptions and design limitations
thoroughly understood. The model has been judged potentially capable of
meeting the user's needs, the user's acceptance criteria have been
established, and a preliminary decision has been made that field validation is
feasible. Finally, a validation strategy has been identified In consultation
with the model user.
With these Important preliminaries completed, the design of a field
validation study can begin. The steps in this process include site and
compound selection, collection of preliminary data sets, a sensitivity
evaluation, development of a field plan and its Implementation, analysis of
samples, and comparison of model performance with acceptance criteria.
STEP 6a: Select Site and Compound(s)
Site and compound selection must be considered together. It is most
likely that not all the selection criteria can be met at any one location and
the importance of each factor must be weighed by the investigator.
Compound-Selection Factors
Analytical Methods--
Methods must exist for quantifying the input loadings to the model and
1-9
-------
for determining the concentration of the compound of interest in the environ-
mental media (water, sediment, biota, etc.) considered by the model. The
EPA has developed standard methods for determining many organic compounds
in water but no similar EPA approved procedures exist for sediment and biota.
The precision, accuracy and limits of detection also need to be determined for
each medium. The detection limits should be reviewed after consideration of
the expected levels at a site. A final factor to consider may be the cost of
making the determinations.
Compound-Specific Model Inputs—
A well-designed field study is based, in part, upon a sensitivity
analysis of the model. Thus, preliminary model runs must be made and these
require compound specific Inputs such as aqueous solubility, octanol water
partitioning coefficients, degradation rate constants, etc. The availability
of such input data may be a factor in selecting compounds. Ultimately the
precision and accuracy of such Inputs are also required. Most rate constants
that are available are based on laboratory tests using the neutral species
of the test compound. Since little information is available for ionic
forms, usually rate constants are assumed to be the same for neutral and
ionic species. This assumption may not be valid, depending on the compound
under investigation. The validator should be aware of this potential
source of error when conductinng field studies.
Environmental Fate—
The environmental fate of the compound is important to the three vali-
dation scenarios presented. For the scenario in which a few field validations
are conducted to support laboratory tests, compounds that undergo multiple
processes should be sought. In the controlled and uncontrolled field valida-
tion scenarios, compounds should be selected to typify a single environmental
process to the extent possible. In all cases, the predicted half-life of the
compound by each process must be considered relative to the time which the
compound can be followed (tracking time). For example, a compound that
hydrolyzes with a half-life of 7 days would not adequately test a model if
that compound could only be followed for 2 days in a river. Ideally, the
compound should be tracked through several half-lives. The compound-specific
model Inputs and canonical environments can be run in the model to estimate
environmental fate.
Compound Toxlcity and Environmental Hazard—
This factor is usually only significant when a compound or compounds are
being added to a natural water body (scenario 2). If this approach is pursued
and a suitable site is located, the least toxic compounds representative of a
process should be selected. The priority pollutants are obviously undesirable
contaminants, but there are many readily degradable organic compounds that
could be added to a water body, traced through several half-lives and pose no
significant threat to the environment. For example, some dyes used as water
tracers are easily photolyzed and therefore might be used to test that
process. Other dyes such as rhodamine B might be used to examine sediment
transport and sorption.
Contaminant Source—
In controlled field validation, the selection of a pollutant source is
1-10
-------
obviously no problem since the Investigator actually doses the water body.
For the other two scenarios, the investigator is restricted to selecting
compounds that already exist in concentrations that can be tracked in aquatic
systems. Although many compounds have been detected in effluents, relatively
few have been found in easily detectable concentrations in the receiving
waters.
Site Selection Factors
Traceable Level of Compounds—
The key factor in site selection is that the compound of interest must be
present at sufficiently high levels so that it can be traced for a consider-
able distance or time period. This translates to being able to follow a
compound through several half-lives for processes involving degradation or
volatilization. For sorption, physical transport and bioconcentration, it means
detecting the compound of interest in various media at significant distance/
time from the source. Factors influencing traceability include the input
load, water body size, flow rate, transformation half-life of the compound,
mixing, and dilution by other water inputs such as merging streams.
Ability to Collect Site Specific Input Data—
The collection of most site specific model input data presents no unusual
problems; however, the required inputs should be reviewed to assure this.
Particular consideration should be given to mixing, flow, sediment transport,
ground water infiltration, and fish movement. Weather, season of the year,
size of water body, site access, and many other factors can also affect the
feasibility of collecting data.
Historical Data on Site Specific Model Inputs—
If historical data are available, they can be used to conduct preliminary
model runs and sensitivity analyses. Depending upon the amount and type of
data available, it may not be necessary to conduct a preliminary sampling
study.
Quantifiable Input Loadings—
The pollutant load to the aquatic system must be known to adequately test
the model. The accuracy of these data will depend upon the types and number
of sources, their relative loading, and their variability. A single, steady
state point source is probably the simplest situation while nonpoint sources,
contaminated ground water infiltration, and atmospheric deposition add signi-
ficant complexity, the additional complexity can result in a less definitive
test of a model. For example, if a nonpoint source (NFS) model is used to
predict loadings, the accuracy and precision of the model's estimate must
be considered. If the output of the NFS model is in error, the output of
the aquatic fate and transport model will also be in error. Input loading
models should be tested and validated separately from aquatic fate and
transport models.
It is recommended that sites with easily quantified inputs be used to
verify in-stream processes. Multiple sources may significantly increase the
expense of collecting data, Increase input errors, and complicate data
interpretation. Theoretically, if the model is capable of predicting the
1-11
-------
pollutant concentrations from one source with distance or time, it should be
capable of doing so for multiple sources.
Site Related Analytical Problems-
Chemicals found in Industrial effluents, or in the water body to be
studied, may interfere with analyzing for the compound(s) of interest.
Samples of effluents, water, and other media of interest should be analyzed
prior to any major field study. If interferences are found, sample cleanup
procedures must be developed prior to further sampling efforts.
Model Assumptions Relative to Site-
In selecting a site, it is necessary to review the assumptions made In
developing the model. Specific assumptions such as complete mixing, steady
state, lack of toxic effect, etc., are discussed in the section on development
of a field plan. Basically, conditions at the site should not seriously
violate assumptions of the model. Proper site and compound selection also
offer an opportunity to design around some model assumptions.
Water Body Type and Validation Scenario—
The site for the field validation must fit into the validation plan,
i.e., the matrix approach discussed under the controlled and uncontrolled
field validation scenarios. The selection of a water body type should be done
in conjunction with the compound selection. Compounds with short or long
half-lives can easily be studied in small, well-mixed ponds. Large lakes can
also be used, but defining horizontal and vertical mixing can pose problems
with point source loadings such as from rivers and effluents. Rivers and
streams are best suited for testing the degradation or volatilization process
of short-lived compounds. Generally long-lived compounds should only be
used to test physical transport and bloaccumulation processes. It is
usually impossible to track long-lived compounds through several half-lives
in rivers and streams.
Overall Site Simplicity—
Generally, the simpler the site in terms of the amount of data that must
be obtained, the more cost effective the validation effort. The site
should, of course, be chosen wisely and must result in a sufficient test of
the model. Multiple sources, many major dilutions, and significant
environmental changes from compartment to compartment (for box-type models)
add greatly to the costs and difficulty of field validating a model. Greater
complexity usually results in more sources of error which result in a wider
confidence interval about predicted values.
Approaches to Selecting Compounds and Sites
The approach to compound and site selection depends upon the validation
scenario selected (i.e., laboratory, controlled field, or uncontrolled field).
Some compounds degrade via multiple fate processes while others degrade by a
single process. Sites for field validations may cover the spectrum from
specially constructed environments and small research ponds, streams, and
lagoons to large natural rivers, lakes, and estuaries.
1-12
-------
The selection of a chemical for study depends greatly upon Its fate in
aquatic environments. Data on the "Water Related Environmental Fate of 129
Priority Pollutants" were assembled and published by EPA (Callahan et al.
1979). Additional Information on 114 of these has been assembled In a form
useful to the EXAMS model and released In a final draft report titled "Aquatic
Fate Process Data for Organic Priority Pollutants" (Mabey et al. 1982). In
addition, fairly complete data exist on the compounds which were studied
during the development of the SRI model (Smith et al. 1978): p-cresol,
benz(a)anthracene, benzo(a)pyrene, quinoline, benzo(f)quinoline, 9H-carbazole,
7H-dlbenzo(c,g)carbazole, benzo(b)thiophene, dibenzothiophene, methyl
parathion, and mirex. Probably the next most studied organic compounds are
the pesticides, but it is beyond the scope of these guidelines to provide
information or references.
Perhaps'the most direct and obvious method of finding a study site is to
consult with persons engaged in model development and testing. Other sources
of information are the Surveillance and Analysis Divisions of EPA's Regional
offices, and State and local water pollution control personnel. These people
may know of sites that meet some or most of the site selection criteria. In
locating sites where contaminants can be added to a water body, various
government owned reservations, government sponsored national laboratories,
private research contractors, and organizations equipped with semi-natural
streams, ponds, etc. should be contacted.
Several computerized data bases exist which may be helpful in site
selection. The STORE! (STOrage and RETrieval) data base (Taylor 1980)
contains information relating to the quality of waterways within and
contiguous to the United States. Data from a geographic area around a site
may be obtained or information on sites where specific compounds .have been
found can also be sought. For information and assistance, contact STORET User
Assistance in Washington, D.C., (202) A26-7792. The Industrial Facility
Discharge (IFD) file may also be useful. It contains data on individual
facilities and can be searched by type of industry. The Reach File is a data
system organized by hydrologic structure and is used for organizing water
resources and environmental data, including the waste water discharge
information and the IFD file. Final manuals on both the Reach File and the
IFD file are being prepared. For Information and assistance, contact the
Office of Water Regulations and Standards (WH-553) U.S. Environmental
Protection Agency, Washington, D.C. 20A60.
STEP 6b: Collect Preliminary Data and Conduct Sensitivity Analysis
Sensitivity
Sensitivity of a model to an input is the rate of change of the model's
output caused by a change in that input. If a change in an input causes a
large change in the output, the model is sensitive to that input; or if a
change in an input causes a small change in the output, the model is
Insensitive to that input. A sensitivity analysis identifies which inputs
have a large Influence on output and need more input accuracy and precision.
Using model sensitivities to plan the sampling design and to choose the number
of samples can aid in obtaining the appropriate accuracy and precision for
1-13
-------
the model users needs. The similarity of model sensitivities to the sensiti-
vity of the real system is tentatively assumed in designing the initial
sensitivity study and must be proven for validation.
There are stochastic techniques for determining the mathematical
sensitivity of input parameters which are based upon Monte Carlo simulation
(Hoffman and Gardner 1983). These techniques are relatively complex and
could be quite tedious to perform with a model such as EXAMS which often
requires numerous input parameters. Mathematical sensitivity analysis
becomes increasingly untenable as the number of input parameters to be
varied by the model user increases. Therefore, this type of sensitivity
analysis is generally inappropriate for the purposes of determining sampling
effort distribution unless the Monte Carlo simulation is built into the model.
Hoffman and Gardner (in press) present a technique for sensitivity screening
using a simple sensitivity index. This index can be easily calculated
with estimates of input parameter values, and input parameters variability.
Best estimates of input parameters and their associated variability (standard
deviations, ranges, etc.) can be obtained during preliminary surveys or
from the literature. The index is then calculated as follows:
S - 1 -
where: S - relative sensitivity
• low model output concentration based on one extreme input
parameter estimate
" high model output concentration based on the other extreme
input parameter estimate.
The index will vary between zero and one with zero representing unused
parameters and one representing hypersensitive parameters. Indices can be
calculated for all input parameters necessary for the model to simulate a
given compound's fate and transport. These indices can then be ranked
according to magnitude and decisions regarding distribution of sampling
effort made. These decisions should take into account the relative
sensitivity of input parameters, the accuracy and precision required by the
model user, and the funding available for the field validation.
The amount of information necessary for best estimates of input
parameters may be quite different depending upon the sensitivity of the
parameter. It is possible for a given input parameter to be hypersensitive
under one set of environmental conditions and insensitive under another
set of conditions. This can be best illustrated using a simple example of
a compound that degrades primarily by acid/base hydrolysis. For purposes
of this example, a hydrolysis-sensitive compound was entered into EXAMS
data base and model outputs generated by varying the pH between 5 and 9
while holding all other inputs constant. Sensitivity indices were calculated
for various ranges of pH using the methods presented above. EXAMS
concentrations from ten hours after dosing were used in the calculations
1-14
-------
since compound levels were simulated as being detectable for all pH levels
at this simulation time. Since pH is known to be an important input
parameter for hydrolytic degradation (Burns et al. 1982), relative
sensitivities should have reflected this by indicating pH as a hypersensitive
parameter. With our test compound example, this was not always the case. In
fact, acid/base hydrolytic sensitivities to pH varied from 0.0 to 0.99
depending on the range of pH entered into the calculation (Table 1-2).
TABLE 1-2. EXAMS CONCENTRATION OUTPUT AND RELATIVE SENSITIVITY OF pH INPUT
FOR A COMPOUND WHICH DEGRADES PRIMARILY VIA ACID/BASE HYDROLYSIS
Combinations of EXAMS Output Calculated
Parameter Ranges cmin/cmax Relative Sensitivity
pH
pH
pH
pH
pH
pH
pH
pH
PH
pH
5/9
5/8
5/7
5/6
6/9
6/8
6/7
7/9
7/8
8/9 .
27/27
27/1700
27/2470
27/1700
27/1700
1700/1700
1700/2470
27/2470
1700/2470
27/1700
0
0.98
0.99
0.98
0.98
0
0.31
0.99
0.31
0.98
This example points out a potential problem in use of this sensitivity
index. For parameters of this type it is necessary to have very good
estimates of the range of environmental conditions to be encountered
during a validation attempt. In this case an underestimate or overestimate
of the range of pH expected to be encountered during the study could have
provided a gross underestimate of the sensitivity of this parameter. This
would have, in turn, provided an erroneous basis for decisions regarding
distribution of sampling effort. Parameters of this sort can be Identified
by calculating sensitivities at various input parameter levels and checking
the results for anomalous values of the sensitivity index over-increased
ranges of the Input parameter values. Once hypersensitive parameters or
hypersensitive/insensitive parameters have been identified, appropriate
effort can be expended to define the range of expected environmental varia-
bility prior to the validation attempt.
Relative parameter sensitivity may become a very important consideration
for parameters which undergo significant degradation via other process path-
ways as environmental conditions change. For example, if the compound used
in the above example underwent moderate biotransformation within the pH
range of 5 to 9, but was only hydrolyzed rapidly above pH 8 or below pH 6,
it would be very important to know the precise value of the expected pH as
well as the range of pH fluctuations.
1-15
-------
In summary, calculation of simple sensitivity indices based on the
response of model output to changes in model input parameters can aid in
making decisions regarding the amount of sampling effort to be expended for
various model input parameters. In addition, input parameters whose
sensitivities change as environmental conditions change can be identified
using this simple index. Decisions regarding distribution of sampling
effort among various input parameters can then be facilitated based upon
the relative magnitudes of the calculated sensitivity indices.
STEP 6c: Develop a Field Study Design
In designing a field study one must consider the compound, site, required
input and output data, model assumptions, other possible design complications,
results of the sensitivity analysis, safety aspects, and the appropriate
statistical design. Once the compound and site are selected the sensitivity
analysis (see previous section) will help define the appropriate samples to be
collected, sample size, and the statistical design. Table 1-3 presents a list
of topics related to the selected compound and site which must be addressed or
considered in the preparation of a field protocol.
TABLE 1-3. CHECKLIST FOR FIELD PROTOCOL PREPARATION
A. DESIGN CONSIDERATIONS
1. Validation scenario satisfied - Selected site (type) and compound
(degradation routes) meet requirements of validation scenario
2. Rate constant information available
3. Model assumptions satisfied
4. Sample type required - Water, sediment, pH, etc.
5. Number of samples of each type required
6. Time and/or distance compound can be detected
7. Location where samples are collected - Time and/or space
8. Quality assurance - Field and laboratory replicates, knowns, splits,
blanks, etc.
9. Statistical design - Comparison of predicted and observed estimates with
validation criteria
(continued)
1-16
-------
TABLE 1-3. (Continued)
B. OTHER ELEMENTS OF THE FIELD PROTOCOL
1. Sample Collection
a. Sampling procedures .
b. Sample collection gear - Pumps, dredges, nets, etc.
c. Sample containers - Glass jars, teflon bottles, etc.
d. Sample container preparation - Soap and water wash, etc.
e. Sample preservation - H2S04, NaOH, HgC12, etc.
f. Sample coding
g. Shipment of samples to laboratory
2. Chemical Analysis - Methods for each compound and media
a. Detection limits
b. Proper sample size - Volume or mass
c.-Matrix problems'-Sediment, biological, etc.
3. Logistics
a. Schedules
b. Record keeping - Field notebooks, data forms, etc.
c. Vehicles - Cars, trucks, boats, etc.
d. Maps
e. Access - Permission to sample effluents, boat ramps, etc.
f. Notification of local, state, and federal authorities
g. Personnel - Competent team leader, reliable assistants
h. Data management - How data will be massaged, reduced, etc.
4. Safety Aspects - Proper handling procedures, emergency procedures
1-17
-------
Validation Scenario
The site and compound(s) selected should be an integral and supportive
component of any validation scenario. In the first scenario, a few
field tests would be conducted on compounds of complex environmental fate to
add credibility to laboratory tests. In contrast, the field test in validation
scenario three, would be conducted with a compound of simple environmental
fate and a specific environmental type. Therefore, it is important to ask
the question, does the site and compound(s) selected support the overall
objectives of the chosen validation scenario?
Rate Constant Information
It is essential to have reliable rate constant information to conduct a
field validation study. In most fate and transport models, rate constants
are coupled with site specific environmental parameters to produce site
specific rates. Conceptually the rate constants are the main determinant
of the fate of a compound which is only modified by environmental data.
Often, an investigator is faced with undesirable situations with regard to
rate constants:
1. Several rate constants are obtainable from the literature for a very
sensitive process but the rates vary over several orders of
magnitude. Since this input is sensitive in the model the resultant
prediction will also have a very large sampling error. If the
sampling error is much greater than the user's acceptance criteria
for precision it would be infeasible to field test the model.
2. Only one rate constant with no estimate of its variance can be
obtained from the literature for a given sensitive process. In this
case, one cannot determine confidence limits for the model's
predictions.
If rate constants do not exist, an investigator may be forced to
develop this information or select another compound. Methods to determine
rate constants do exist, e.g., SRI International has developed laboratory
protocols for determining aquatic rate constants for hydrolysis, photolysis,
oxidation, biodegradation, volatilization and sorption (Mill et al. 1982).
Rate constants are available from a variety of sources. Two excellent
sources available in the literature are Mabey et al. (1982) and Lyman et al.
(1982).
Model Assumptions
After the model examination step has been completed, the model's inputs,
outputs, and assumptions will be carefully detailed. The assumptions
associated with any fate and transport model must be met to conduct a totally
valid field test. Table 1-4 lists assumptions commonly used in aquatic fate and
transport models.
1-18
-------
Theoretically, it is pointless to conduct a field validation study if
any one of the model's assumptions is violated. Few, if any, field
validation studies could be conducted if an Investigator had to strictly
comply with all of a given model's assumptions. However, an investigator may
simulate compliance of a model's assumptions. For example, most models assume
that a compartment is a completely mixed (homogeneous) box where a single
parameter value represents conditions throughout the entire compartment. An
investigator may simulate compliance by using average values which are
representive of the true compartment mean values.
TABLE 1-4. ASSUMPTIONS COMMONLY ASSOCIATED WITH COMPARTMENTALIZED
AQUATIC FATE AND TRANSPORT MODELS
1. Steady state - Loading, water flows, bedload movement constant within a
given compartment over time.
2. Homogeneous compartments - A compartment is completely homogeneous
(completely mixed) with respect to all parameters. Therefore, only a
single value is required to represent a given compartment parameter.
3. Pollutant levels will not change pseudo-first order kinetics - (a) no
direct toxic effects on biota, and (b) no interaction between organic
pollutants, i.e., no synerglstlc or antagonistic reactions.
4. Sorption and ionization are essentially instantaneous processes within
each compartment, i.e., much faster than physical transport or
transformations.
5. Subroutines which use rate constants and site specific environmental
parameters actually account for the variation of transport and fate
phenomenon. Standard plate counts represent total bacterial populations
or organic carbon contents of sediments can be used to "normalize"
sorption coefficients for neutral hydrophobic organics.
6. Mass balance - The mass of the loadings equals the mass of the
degradation products and/or mass of the transport products.
7. Environmental variables exist within "normal" ranges - Extreme high
temperatures do not lower bacterial degradation rates, water never
freezes, etc.
Steady State--
The assumption of steady state (conditions are constant) is a difficult
assumption to meet. Many possible field testing situations will be eliminated
because of this assumption. Effluents, stream flows, environmental conditions,
and bedload movement do vary over time. If a true steady state situation was
discovered, a single set of samples, collected without regard for time (e.g.,
1-19
-------
samples collected on a single given day), would be sufficient to verify a
model.
Fortunately, conditions do not have to be constant over time ad
infinitum, they only need to be constant long enough for an equilibrium to be
established between the pollutant and its surrounding environment and long
enough to allow one to collect the necessary data on input and output
variables. This alone represents a strong rationale for the use of controlled
situations, i.e., the second validation scenario.
In addition, input and output parameters can be time-weighted averages
that can be used in model validation. However, this is not the ultimate
solution to all steady state problems. If equilibrium conditions are reached
in days, the time-weighted averages need only to reflect days; however, if
equilibrium conditions are not reached for weeks or months, then the time-
weighted averages must truly reflect these extended periods.
Steady State Loadings—Industrial effluents or other sources of pollutant
loads (model loadings) are rarely constant in pollutant concentration or flow
over extended periods of time. However, with an appropriate field plan one
may simulate compliance with the steady state loading assumption. For
example, an investigator may select a compound and site with the following
characteristics:
1. The compound does not sorb to any appreciable extent, (because a
pollutant normally reaches equilibrium with water faster than with
bottom sediments);
2. the compound obtains equilibrium with its surrounding environment in
a relatively short time, i.e., hours;
3. the receiving water, e.g., stream, demonstrates "plug" flow; and
A. the compound has a short half-life relative to tracking time.
Then the investigator can:
1. Establish a "window" of known concentration in the effluent for a
short period of time. This could be accomplished by collecting a
number of continuously pumped samples to estimate pollutant
concentration in the effluent (Figure 1-2). During each 30-minute
sampling period a sample would be collected using a peristaltic pump
equipped with chemically inert tubing.
2. Mark the known effluent "window" by adding rhodamine WT dye at the
mid-point of the effluent monitoring period. The amount of dye
needed to mark a certain section of stream and still be detectable
a given distance downstream can be determined by methods presented
in Kilpatrick (1970).
1-20
-------
Sampling Time (min.)
Elapsed Time
Dye Spike
30 I 30 30 I 30 I 30 I 30 30
I 30 I 30 I 30 1
30 60 90 120 150 180 210
Figure 1-2. Effluent sampling.
If no concentration change Is detected during the 3 1/2 hour effluent
sampling period, a known section of the stream will be contaminated with a
known amount of pollutant. Repeated cross-sectional sampling of the center of
this water mass (area of maximum dye concentration determined with a portable
fluorometer) would permit the pollutant decay curve to be determined. By
sampling within the marked water mass at the point of maximum dye concentration,
the investigator has simulated compliance with the steady state loading
assumption.
Also, another variation on the same theme can be employed by an
Investigator to simulate compliance with the assumption of steady state
loadings. In this case the investigator would mark a section of a stream with
rhodamine WT dye and establish a window of known pollutant concentration in
the stream Itself rather than the effluents to estimate pollutant load (Figure
1-3). Depth-integrated cross-sectional samples would be collected to establish
a "window" of known, and hopefully equal, pollutant concentration. This water
mass would then be marked at its mid-point with rhodamine WT dye by releasing
the dye at a constant rate as the investigator transverses the stream.
Repeated cross-sectional sampling of the water mass as it moves downstream
would allow the Investigator to determine the pollutant decay curve. Also,
the problems associated with multiple discharges can be circumvented in this
manner. Obviously, the criteria listed above for the site and compound would
still apply.
Pollutant Discharge
Pollutant
Discharge
Release of
Rhodamine WT Dye
Pollutant Discharge
Cross-Sectional
Depth-Integrated
Samples
Figure 1-3. Simulating compliance of steady state loading.
1-21
-------
Steady State Flows—An investigator may find the following suggestions
useful in simulating compliance with the assumption of steady state flows:
1. Select a site where flows are relatively stable, e.g., downstream
from a flow regulation dam,
2. Select time of year when flows are relatively stable—often summer,
3. Use controlled situations,
A. Use dye as discussed previously, or,
5. Use a pond or lake where short-term flow changes have little effect
on lake or pond volume.
Steady State Bedload Movement—Bedload movement is difficult to measure
and often erratic. A high percentage of the yearly bedload movement will
often occur in a short period of time in a given stream, e.g., after
rainstorms, floods, snow melt, etc. Therefore, a single set of sediment
samples often will not accurately reflect steady state bedload movement. To
aid an investigator in simulating compliance with the assumption of steady
state bedload movement the following suggestions are offered:
1. Select a site which reduces the probability of bedload movement -
lake, pond, or extremely sluggish river,
2. Select a stream which is flow regulated,
3. Use controlled situations - artificial stream, etc.,
4. Select a compound which does not sorb so that bedload movement is not
important, or,
5. Select a compound which sorbs but has a relatively short half life.
Also, conduct this study where there is a low probability of erratic
bedload movement (low flow in stream).
Homogeneous Compartments—
Fate and transport models may assume that a compartment is a completely
mixed box (homogeneous) where a single parameter value represents conditions
throughout the entire compartment. To simulate compliance with this
assumption, an Investigator could reduce compartment sizes to reduce
compartment variability, and/or use average values obtained from either
composite or multiple samples.
Mass Balance—
Models equate the mass of the chemical loading to the mass of the
chemical in the output predictions (mass balance). The model's mass balance
equations account for removal from the modeled system by transport, changes
from one phase to another within the modeled system (e.g., sorption) and
losses within the modeled system (e.g., degradation). To account for the
chemical mass present within the system it may be necessary to collect
1-22
-------
suspended material, sediment, interstitial water, and biota, depending on
the chemical of interest.
Environmental Variables Exist Within Normal Ranges—
Most fate and transport models' subroutines are set so that extreme
environmental conditions (pH value of 1, temperature of 80°C, etc.) will not
change pseudo - first order kinetics. Temperature is one of the most
important environmental conditions determining degradation rates. The .
influence of temperature on physiological processes can be measured by a
rate called QIQ. This value is the rate at which processes increase
with a 10°C rise in temperature. For biological processes, QJQ
values generally range from 2 to 3. The principle operates only within
the range of tolerance of a given species. However, a model given an
ambient temperature of 80°C may increase physiological processes (bacterial
degradation rates, bioconcentratlon) according to the QIQ rule even though
most organisms would be eliminated.
Normally it would be inappropriate to field test models in extreme
environments since the user is generally concerned about the fate and
transport of compounds under "normal" environmental conditions. However, if
the intended use of the model Is for extreme environments (hot spring, acid
mine drainage system, etc.) the investigator must take into account the effect
of extreme environmental parameter values on the various subroutines.
Type and Number of Samples .
The type of samples to be collected is a function of the model's input
requirements, output statements, and the selected test compound and site.
Through a detailed examination of the model, an investigator will determine
input data requirements by process (oxidation, physical transport etc.),
output predictions and model assumptions. By completing the steps of this
report through the sensitivity analysis, an investigator will know by which
processes the compound of Interest will be degraded or be transported by and,
therefore, the type of samples required for the field validation study.
The sensitivity analysis step as previously described will determine the
relative number of samples of each type to be collected. The relative number
of samples is based on a parameter's sensitivity and variability. The
relative number of samples to be collected may have to be tempered by cost and
time considerations to give the actual sample size.
Compartmentalizatlon
Compartmentallzation refers to the segmentation of model ecosystems
Into various "completely-mixed" boxes of known volume and interchange.
Interchange betwen compartments is simulated via bulk dispersion or equal
counterflows between compartments. Compartmentalization is a popular
assumption in fate and transport modeling because the completely-mixed
assumption reduces the set of partial differential equations (in time and
space) to one of ordinary differential equations (in time only). Neverthe-
less, it Is possible to recover some coarse spatial information by introducing
1-23
-------
the appropriate number of interconnected compartments. To select the
appropriate number of compartments for a given field validation attempt,
the reader is referred to the Physical Transport section.
Design Features to Eliminate Other Problems
Besides problems which arise from compliance with model assumptions,
other problems can plague a validation study. These problems are often
related to the selected site, e.g., multiple sources of effluents. Proper
site selection can alleviate many problems associated with collecting all
of the necessary data for model validation. For example, Is It possible
to collect large enough quantities of biota for chemical analyses at the
selected site? If the model predicts a certain concentration of pollutant
in sediments, can representative sediment samples be readily obtained?
Other collection problems can arise, e.g., the model may predict
chemical concentration in suspended sediments and plankton. Mutually
exclusive and totally discrete samples of this kind are nearly impossible to
collect. In this case, an investigator may estimate concentrations by
analyzing unfiltered versus filtered water, analyzing mainly plankton samples
collected by plankton nets, and fine sediment samples collected from the upper
layers of the bottom sediment. Many problems can arise when hydrologlc
characteristics of the water body are not well defined prior to the initiation
of chemical monitoring. Aquatic dyes (e.g., rodamine WT) have been
successfully used to establish hydrologic characteristics of water bodies
including travel time, mixing zones downstream from effluents, dilution
effects, and pollutant dispersal characteristics over the entire length of
the water body anticipated to be monitored (Hubbard et al. 1982, Pendleton
1980, Yotsukura and Cobb 1972). In addition, dyes have been used to track
compounds, and to determine volume and discharge rates. Generalized techniques
for fluorometric procedures to measure dyes can be found in Wilson (1968).
Field validation of aquatic fate and transport models require the
collection and quantification of numerous chemical, physical, and biological
parameters over time for model input variables and model output variables.
Parameters which display large scale perturbations over time can generate
problems in quantification of model input data. For example field validation
studies which require standing crop estimates should not be attempted
in environments which exhibit large temporal biological perturbations (algal
blooms). Table 1-5 lists some common problems and suggests possible
solutions.
Sampling Location
Selection of sampling sites is a function of the characteristics of the
decay curve, i.e., the pollutant half-life, while the sensitivity analysis,
cost, and a knowledge of parameter variability dictate the number of samples
of a given type to be collected. The investigator needs to determine over
what distance and time the compound of Interest can be detected. This
will be accomplished through a knowledge of the pollutant loading, flow
1-24
-------
TABLE 1-5. DESIGN PROBLEMS AND POSSIBLE SOLUTIONS
Problem
Possible Solutions
Multiple sources of loadings
Dilution of the water of
interest
Conduct study downstream of multiple sources—
calculate loadings from cross-sectional
transects
or :
Measure all major sources but estimate minor
sources
or
Avoid multiple loading sites.
Account for differences between dilution effects
and degradation/volatilization/sorption effects
by using absolute dye concentrations, i.e., the
reduction of the concentration of dye due to di-
lution water is proportional to the amount of di-
lution. Therefore, one can estimate the reduc-
tion of the compound of interest due to dilution
alone.
or
Avoid, if possible.
Tracking long lived compounds Tracking long-lived compounds through several
in rivers - half-lives in rivers is impractical due to
dilution effects or multiple sources. There-
fore, compounds with long half-lives may require
1 lake or pond validation.
Movement of biota into and
out of the zone of polluted
water
Calculate water compartment
volume when dimension data
are not available
Defining lower boundary of
sediment compartments
Gaging data not available
Select territorial species (eunfish rather than
trout).
For flowing waters use dye travel time multi-
plied by the mean discharge rate to calculate
the volume of a water compartment. For small
pond or water bodies use dilution of a known
amount of dye to calculate volume.
Use depth of detection of compound of interest
or
Use depth of detection of synthetic compound.
Install weirs
or
Nitrogen or phosphorus as
limiting nutrients
Use time to fill buckets
or
Use dye dilution techniques.
Measure nitrogen and total phosphorus to detect
abnormally low levels.
1-25
-------
rates, decay rates, and the detection limit for the compound of interest.
See Appendix A for an example of how time and distance determinations were
calculated for phenol in a river study.
The goal of any field validation study is to collect sufficient data to
accurately describe the pollutant decay curve through several half-lives or
the steady state concentration of a pollutant and the environmental conditions
which influence the pollutant concentration over time. True steady state
conditions are rarely achieved and, therefore, difficult to find. It may be
possible to find certain holding ponds which closely approximate steady state
conditions. However, a more practical solution may be to conduct small lake
or pond studies by dosing the waterbody with a single input of a pollutant or
pollutants and describing the decay curve or curves by repeated sampling over
time. This type of study should be conducted in lakes or ponds which have
complete mixing.
Even models with steady state assumptions can be manipulated to give
downstream concentrations in pond tests by dummying in plug flow conditions
(no mixing of the water of interest with the dummy water volumes used to
create "flows"). Therefore, the model would predict various concentrations in
downstream compartments as a function of distance. Since distance and time
can be equated under "plug flow" conditions, e.g., one kilometer downstream
equals one hour of travel time, this method can be used for steady state
models to construct pollutant decay curves. In addition, computer codes of
steady-state loading models can be modified to accept non-steady loadings.
For example, a new version of EXAMS (EXAMS II or "Spike EXAMS") is
currently available. EXAMS II allows one to predict environmental pollutant
concentrations over time from single or multiple loading(s) of a specified
pollutant mass at specified time(s).
Pollutant decay curves are commonly described by exponential rather than
linear curves with the following general form:
CF - CQE-Kt
where: Cp - Final concentration
C0 * Initial concentration
K - Combined loss rate
t » Time elapsed between the initial and final concentrations.
To accurately describe any decay curve with a given number of points it
is optimal to have known points which represent equal changes in pollutant
concentration and encompass the entire curve (Figure 1-4). Therefore, if an
exponential curve is expected in a validation study, and if the physical
properties of a specific site do not dictate compartment size, it is advan-
tageous to have more sampling points in the initial phase than the tail
portion of a curve. Obviously, little information would be provided by a
number of sampling points in the tail portion of the curve where concentra-
tion does not vary to an appreciable extent over time or space. A geometric
spacing of compartments may be convenient to use. Additional sampling
1-26
-------
10-
9
8
7
o 6<
5
+*
I 6
u
o
4
«•
! 3
2
1
o
a
2 3 45
TIME OR DISTANCE
8
Figure 1-4. Pollutant decay curve.
points may be included after the initial sampling points are selected, in
order to verify that there are no anomalies in the decay curve.
Fate and transport models may assume that a compartment is a completely
mixed box (homogeneous) where a single parameter value represents conditions
throughout the entire compartment. An Investigator may simulate compliance by
using average values which are representative of true compartment mean values.
Calculation of simple mean values for parameters which vary over a large range
In a short period may cause substantial input data bias which would invalidate
the test of the model. For example, pH in a pond system may vary between 7 and
9 over a diurnal period. Compounds undergoing fast base but insignificant
neutral hydrolysis would be degraded at pH 9 but not degraded at pH 7. It
would be necessary to weight the pH value by the hydrolysis rate to reflect
the shifting hydrolysis rates In an average pH value. This could be done
by averaging the pseudo-first-order rate constants generated from model
simulations based on continual sampling of pH values over a diurnal
period, and adjusting the pH input value such that the model output would
equal this average rate constant. Also, an investigator can reduce compartment
size to reduce compartment variability so that an estimate of the true
compartment mean is easier to obtain.
1-27
-------
Although there are many possible designs to estimate pollutant means
within a given water compartment for river or pond studies the following are
offered to aid an investigator:
Well-Mixed Rivers
1. Collect single grab sample at the middle of the compartment. This
method assumes that the compartment Is homogeneous or that the
pollutant gradient is approximately linear from the upstream to the
downstream boundaries of the compartment.
2. Collect depth and flow integrated samples at equally spaced points
along a stream transect at the middle of a compartment. All samples
would then be composited. The number of points (depth flow and flow
integrated samples) is a function of stream width. This method
assumes vertical, horizontal, or flow gradients exist at the mid-
compartment transect. Depth and flow integrated samplers can be
obtained from the Federal Inter-Agency Sedimentation Project, St.
Anthony Falls Hydraulic Laboratory, 3rd Avenue and Hennepin Island,
Minneapolis, Minnesota, 55414.
3. Collect vertical, horizontal and flow composite samples, as above,
along transects at the upstream and downstream boundaries of a
compartment. The data obtained from the transects can be averaged
to obtain a mean parameter value. This method assumes a nonlinear
pollutant gradient through the compartment.
Poorly Mixed River
1. Collect discrete vertically Integrated samples at points along a
transect. Also, at the verticals collect mean discharge and total
depth data. The stream is then divided into sections (two verticals
represent the boundaries of a section) (Figure 1-5). No more than 10
percent of the total flow should pass through a given section. By
this method one can calculate the pollutant concentration weighted
by volume (PC) at a transect as follows:
n
1 VC
PC - *-* * *
n
I
1-1 V±
Where n • the number of sections
Vj - the volume of section 1
GI • pollutant concentration of section 1
(average of two verticals)
1-28
-------
The pollutant concentration calculated can then be compared to the
Bodel's predicted value. Such transects could be located at the
middle of a compartment or at the upstream and downstream boundaries
of a compartment as described for well mixed rivers. Methods to
measure stream discharge are presented in the Physical Transport
' section of this report.
2. Collect depth and flow Integrated samples at the verticals described
above. All depth and flow Integrated samples would be composited to
determine a pollutant concentration at the transect. Discharge data
may still be collected to assure that the verticals are properly
spaced and/or to satisfy physical transport input requirements.
•••» Depth Integrated Samples
n
LJ Sections
Cross section of a river showing sections defined by
depth integrated samples
Figure 1-5. Cross section of a river showing sections defined by
depth integrated samples.
2.
3.
Collect a single depth integrated sample. This assumes nearly
complete mixing but allows for some minor thermal stratification.
Collect depth integrated samples at points along two transects at
right angles to each other to account for horizontal gradients. The
depth Integrated samples may be averaged as composited.
Collect depth integrated samples at points on a grid system.
depth integrated samples can be averaged as composited.
The
Input parameter values may also be collected at the locations described
above. However, additional sampling points may be required dependent upon the
given parameter to represent its true compartment mean value.
Quality Assurance
The purpose of sampling in field validation of models is to provide
1-29
-------
quantitative environmental input data to the model and to collect field residue
data that can be compared to the levels predicted by the model. Such field
programs usually include the following operational steps:
Planning
Sample Collection
Sample Preservation
Sample Transport
Sample Storage
Sample Preparation
Sample Analysis
Data Acquisition
Data Manipulation
Data Interpretation
Reporting
To obtain valid data, an overall quality assurance (QA) program must
apply quality assurance to all pertinent operational steps. In addition to
the usual analytical and equipment QA procedures, a comprehensive QA program
should include details on the reliability of the sampling program. Sampling
schemes, data analysis strategies, and the objectives of the sampling program
must be well defined for a statistician to assist in the development of an
efficient collection program.
The importance of quality assurance is such that it is the EPA's policy
that all intramural and extramural EPA funded projects involving environmental
measurements must have an approved QA plan. Field validation of any aquatic
fate and transport model for EPA by contract, grant or inhouse effort must
conform to this requirement. The required format for project plans is given
in "Interim Guidelines and Specifications for Preparing Quality Assurance
Project Plans" from the Office of Monitoring Systems and Quality Assurance
(U.S. EPA 1980a). Special items that supplement the guidelines are addressed
below.
Quality Assurance References
EPA's "Handbook for Analytical Quality Control in Water and Wastewater
Laboratories", (U.S. EPA 1979a) addresses in detail the following areas: labora-
tory facilities, instruments, glassware requirements, and reagents. It further
deals with control of analytical performance, data handling and reporting.
Separate chapters are devoted to the special requirements for trace organic
analysis, water and wastewater sampling, microbiology, aquatic biology and
safety.
The "Quality Assurance Guidelines for Biological Testing" (U.S. EPA 1978)
discusses the following elements of quality assurance: QA policy and objectives,
design and analysis of experiments, sampling, precision and accuracy of tests,
physical environment of research, chemicals and reagents, control of performance,
and data handling and reports.
The "Handbook for Sampling and Preservation of Water and Wastewater"
(U.S. EPA 1982) was developed for guidance on field monitoring required under
1-30
-------
the National Primary Drinking Water Regulation, the National Pollutant Discharge
Elimination System, section 304(h) of the Clean Water Act, and the consent
decree for priority pollutants. The handbook addresses the sampling of waters
and wastewaters, such as industrial/ municipal wastewaters, agricultural run
off, surface waters, and sediments, as well as flow monitoring, handling and
preservation methods.
Detailed analytical procedures have recently been proposed by the U.S.
EPA (1979b) for determining the concentration of 113 organic toxic pollutants
in water. Methods 601-613 apply to the analysis of individual compounds or
groups of chemically similar compounds. Methods 624 and 625 are GC/MS proce-
dures for the analyses of the same compounds. The proposed methods cover
calibration of instruments, quality control, sample collection, preservation
and handling, sample extraction and analyses, and calculations. Revisions of
the above methods appear in "Test Methods, Methods for Organic Chemical Analysis
of Municipal and Industrial Wastewater" (Longbottom and Lichtenberg 1982).
The majority of the revisions were made for clarification or to add additional
flexibility for the analyst. Finalized procedures are to be published in the
Federal Register in 1984.
Another source of EPA QA procedures is the "Manual of Analytical Quality
Control for Pesticides and Related Compounds in Human and Environmental Samples"
(Sherma 1981). A general description of pesticide residue analytical procedures
is provided following a discussion on inter- and intra-laboratory quality
control. Also covered are gas chromatograph procedures and troubleshooting,
and procedures for analysis of samples including extraction, Isolation and
confirmation of pesticide residues. Although the manual deals primarily with
pesticides, many of the'procedures and recommendations apply to the analysis
of any organic chemical.
There are many other sources of quality assurance procedures. Other EPA
QA procedures are referenced in "The Quality Assurance Bibliography" (U.S. EPA
1980b). Many of these publications are available from the EPA, while the
others are available from the National Technical Information Services in
Springfield, Virginia.
Evaluation of Literature and Unpublished Data
Part of the data used in validation may have been recorded in the past by
different researchers in a variety of studies. While no rigid apriori criteria
for acceptance or rejection of these data can be imposed, it should be stressed
that they must be closely scrutinized. In many instances, data may be of
limited value or even useless because precision and accuracy were not reported,
or because of inadequate reporting of other parameters.
Supplemental Data Acquisition
Whenever samples are collected at field sites, a variety of parameters
important for interpretation and correlation of all data must be recorded.
These parameters may include: sampling site location (coordinates, site
number), sampling depth, flow data, date and time of day, meteorological
conditions (air temperature, wind speed and direction, percent cloud cover,
1-31
-------
precipitation, fog), water temperature, water quality parameters (to be deter-
mined in the lab or field from collected samples or with contact sensing
techniques: pH, conductivity, turbidity, color, dissolved oxygen, etc.)*
These supplemental data may help the researcher to recognize and explain data
trends.
Sampling
Sampling should be highly coordinated to collect a maximum of required
samples per sampling trip. It is Impossible to give directions covering all
conditions, and the choice of sampling technique must often be left to the
analyst's judgment. However, samples should be truly representative of exist-
ing conditions. This can sometimes be achieved, depending on circumstances,
by making composites of samples that have been collected over a period of
time, or at different sampling points. Composite samples should be replicated
so that an estimate of field variation is obtained for each sampling time.
All sample containers should be sealed and labeled before they are
shipped to the laboratory. Pertinent Information should be recorded on a
sample tag, e.g., sample, number, date and time taken, source of sample,
preservative, analyses to be performed, and name of sample collector.
EPA approved procedures should, wherever possible, be used in collecting,
preserving and analyzing samples. Many field sampling procedures, approved
and unapproved, will be discussed at length in the subsequent sections. The
references frequently provide information on the precision and accuracy of
sampling, sample preservation procedures, information on container selection,
etc., and are valuable sources of information on QA procedures. A complete
list of EPA approved and tentatively approved test procedures related to aquatic
sampling and analysis was published in the Federal Register (U.S. EPA 1979b).
Also included was a table on approved containers, preservation procedures and
holding times.
STEP 6f: Compare Model Performance With Acceptance Criteria
Data To Be Compared
The data to be compared are two sets of estimates of the amount or concen-
tration (called Y) of pollutant In a study area. One set of estimates is the
results of the field sampling (called c); the other set of estimates is the
output of the fate and transport model (called Y). The general output from
this type of model is a two-dimensional array of pollutant estimates, i.e., a
matrix {»ij}. the first index i runs over the media such as water, particulate
material, biota, dissolved materials, etc. The second index j runs over the
compartments (spatial grid) that describe the river or lake. If the pollutant
of interest uses several media in its transport and the river or lake requires
several compartments for adequate description, then there are many model
estimates, one for each medium within each compartment. The field sample
estimates must have the same dimensionality for comparison. If the pollutant
uses only one medium in its fate and transport over the compartments, or if
1-32
-------
the body of water is homogeneous enough to be represented by one compartment,
the matrix of estimates becomes a one-dimensional vector. In summary,
there are two estimates (model outputs and field sampling statistics) to
be compared and the estimate may be zero (scalar), one (vector), or two
(matrix) dimensional arrays.
Graphical Comparison
For a one or two-dimensional array, plot both the model estimate and
field statistics on one graph for comparison. Each medium would be plotted on
a separate graph and plots inspected for similarity of shape. Patterns of
similarity to be expected would include asymptotic behavior at the same extremes,
similar points of inflection, or relative maxima or minima at similar times or
distances. These curves may be die-away curves and could be made linear by
taking the natural log of the concentration. Linearization would simplify
comparison of the curves. Therefore, linearize, if necessary and least square
fit the data to a regression line. Confidence around the regression line can
be calculated as follows:
CL (Y) - c t S-., / 1 + 1 + (d-d)
where CL « confidence limits
Y - population parameter of amount or concentration of pollutant
c • regression line estimate of amount or concentration of
pollutant via field observations
t - t statistic for appropriate a and degrees of freedom
83.j • standard deviation of c for.fixed d
d • distance or time for die-away
The regression line is plotted as a solid line in Figure 1-6 and the
confidence curves plotted as dashed lines. The horizontal axis of Figure
1-6 Is time or distance and the vertical axis the natural log of concentration
or amount of pollutant. Time would be the dimension of die-away for a one
time Insult of pollutant to a pond while distance down stream would be the
dimension of die-away for a continuous insult to a river.
A confidence interval about the regressed line of model estimates (?ij)
comparable to the confidence Interval about the regressed line of observed
estimates (cij) is needed for comparison. However, replicate outputs from a
deterministic model run with the same input values would produce one constant
value and the standard deviation of the replicated output for a fixed set of
inputs (SY.x) would be identically zero and would generate a confidence band
of zero are!. O'Neill and Gardner (1979) suggest that the sample variation of
1-33
-------
the inputs be used to generate a multivalued replicate output. That is,
holding our distance or time input (d) constant at the point d, the model
inputs (X) have a range defined by their sampling errors (S). Then, by
rerunning the model for many combinations of the inputs randomly varied
(Monte Carlo technique, Hoffman and Gardner in press), over their ranges
would generate a multivalued sample of model outputs. The regression
fitting of these samples of model outputs gives a regression line and a
standard deviation S?«d for confidence curves about that regression line.
Calculating SY-d by this technique may be a measure of only the dispersion
of the intermediate Inputs and may be an underestimate of the dispersion
of the process being modeled; however, for a deterministic model it is the
best available method. The Monte Carlo technique is "easier said than
done;" however, some modelers have suggested building it into their models.
With a SY«d for the regression on model estimates, confidence curves can
be computed as was done for the regression line for the field observations.
The regression line of model estimates is plotted in Figure 1-6 as a broken
line (—) with dotted line confidence curves.
5
8
o
O
'->x-
Regression on Model Data
Confidence Band
•
Regression on Field Data
Confidence Bands
Acceptance Band
7 ' Distance
Figure 1-6. Regression lines and confidence bands for model and field
data with acceptance band.
1-34
-------
Lastly, the acceptance criteria discussed in the selection entitled Develop-
ment of Acceptance Criteria for Verification, describes an "acceptance Interval"
about the true pollution values Y. Since the true Y are unknown, the acceptance
Interval will be about the field sample estimates c. (It should be about the
confidence Interval of the field sample.) If the acceptance criteria were 1/2Y
to 2Y, then the acceptance Interval for a linearized model of c's would be
plotted as the — - — line in Figure 1-6.
Figure 1-6 will show whether the modeled and observed lines are unusably
different, or basically the same line with minor differences in intercept and
slope, or within acceptance criteria. In Figure 1-6 the modeled output lies
outside both confidence limits and acceptance criteria. One would conclude
from this that the model was invalid and unacceptable for this pollutant and
site. If the modeled and observed lines lie within each other's confidence
curves, statistical comparisons should be done.
Other less direct but more diagnostic tests are the plotting of residuals
(Eij • Yil - cij) or differences against time or distance (d), modeled concen-
tration (YiJ), and observed concentration (cij). The difference of the modeled
output minus the field observations plotted against distance or compartments
would show cumulative error or bias. The horizontal axis is distance or time
and the vertical axis would be differences or errors (Figure 1-7). There
would be a graph for each media. The differences of a good fit should be
random with a mean of zero. If the differences appear random but their average
is non-zero, there is a bias in the model or the field data. If there is an
increasing magnitude to the differences, there is a cumulative error in the
model, or a negative cumulative error in the field observations.
X
o
Bias
.8
Random
4ij
*i
Cumulative
Distance Axis
Distance Axis
Distance Axis
Figure 1-7. Graphic error analysis.
Next the differences should be plotted against media. Now the horizontal
axis Is media and there would be a plot for each compartment. This would show
bias or cumulative error over media. Again, the differences or errors could
be plotted against the observed estimate c or the model estimate Y to see if
the errors were observed concentration dependent or model concentration
1-35
-------
dependent. Random error is acceptable if unbiased and small variance, but
cumulative error is more restrictive. A model with a cumulative error might
be usable for a site with few compartments or a pollutant with rapid die-
away. A model may be usable for one site and chemical but not usable for
other sites or pollutants.
Finally, a simple but statistically definitive plot can be prepared
which compares observed points to predicted model output. This is accomplished
by computing confidence intervals (Sokal and Rohlf 1969) about the observed
points on a die-away curve, and plotting those points, including the confidence
about the points, on the same plot. The degree of observed and predicted
(model generated) overlap can be evaluated based upon the graphical overlap
of calculated confidence Intervals with the predicted die-away. To perform
this type of analysis it is necessary to replicate field sampling for the
pollutant for each point that is to be compared to the predicted die-away.
Statistical Comparisons
The t-test (Steel and Torrie 1960) for equality of means with unequal
variance for the point comparisons and Hotelling's T2-test (Morrison 1967)
for the vector comparisons would be statistically appropriate, but the needed
variances are not available because no ecological models output the variance
with their estimates. The variance of the model theoretically could be
estimated by the Monte Carlo method discussed in the Graphical Comparison
paragraphs, but in practice there would be problems of convergence and
sample size. The variance of the field study's sample statistics may be
estimatable from the observations.
Correlation coefficients can be calculated which will test for a
linear relationship between observed and model generated values. This test
will Indicate if the points are related by a scale factor and an offset
rather than absolute equality. The correlation coefficient must always be
tested for significant differences from zero. For small sample sizes,
apparently large correlations are not different from zero (i.e., an r of
0.8 Is not different from zero for an n of 5, Table A13 Steel and Torrie
1960). This test can only be done if the modeled points and field samples
points have the same abscissa (time or distance value). Strong correlations
indicate that differences amoung observations can be explained by the model,
while weak correlations indicate that differences among observations are
controlled by factors unaccounted for by the model (Hoffman and Gardner
1983).
For a field study, the amount of a pollutant in a medium within a
compartment is a continuous variable (U.S. EPA 1978). From the modeling point
of view, the amount of pollutant might be thought of as the number of units of
pollutant being distributed to various boxes (media within compartments, or
media within a compartment, or medium across compartments). Then the count of
pollution units per box would-be distributed as a multinomial model, and the
comparison of two estimates ("?ij, cij) of pollution units frequencies can
be done by the standard test for equality of multinomial frequencies (the chi-
square test). The chi-square test is easily used, generally known and does
not require estimates of variances (Steel and Torrie 1960). The chi-square
1-36
-------
loses power if there are fewer than four comparisons of observed (cij) and
modeled ("Yij) values (Cochran 1954). The experimental design planning must
provide adequate number of modeled and observed estimates for comparison (_>4).
Pollutants with few pathways in a homogeneous site may require more sampling
locations (more j) for concentration (cij, ^ij) comparison than for
precision of model inputs. The chl-square test cannot have expected values
(iij) less than one count (Cochran 1954). The combinations of media(i) and
compartments( j) that have zero expectation (Yij - 0) must be pooled or
omitted. Zero for the pollutant count would be the background level or the
detection limit whichever is larger. The degrees of freedom will be the
number of boxes with non-zero Yij minus .one.
The null hypothesis states that the two sets of counts of pollutant
amounts are the same and the apparent differences are not greater than could
be expected from chance. The acceptance of the null hypothesis sustains but
does not prove the validity of the model. The alternative hypothesis states
that the two sets of counts of pollutant amounts are different and the
differences are greater than could be expected from chance. The acceptance of
the alternative hypothesis would invalidate the model for this use. The test
statistic is mathematically stated as follows.
Let * • test statistic
- model estimate of counts of pollutant in medium i of
compartment J
• field test sample statistic of counts of pollutant in medium i
of compartment j
- chi-square table value for 1- probability with df degrees
of freedom.
compartments media -y
Note the double summation is for a two dimensional array or matrix and if
there were i media within j compartments the summations would be
« . Z I (c -V)2
J-l i-1 f-
The double summation can be for a one dimensional array or vectors for i media
within a compartment the summations would be
J-l
1-37
-------
or for medium across j compartments the summations would be
J 1
$ m I £
J-l 1-1
$ _<_ aXdf •*• no reason to reject H
* > «X2df -»• accept Hj
If the test statistic is less than or equal to the chi-square critical value
for a chosen probability and the appropriate degrees of freedom, the null
hypothesis is sustained and the model has passed the field test. However,
each field test is only a part of the complete validation process. If the
test statistic is more than the chi-square critical value, the alternative
hypothesis must be accepted and the model has failed this field test.
Regression Analysis—
If the model and observed lines laid within each others' confidence
curves as suggested for the first test of the Graphic Comparison section, an
appropriate statistical comparison is discussed on pages 173, 174 of Steel and
Torrle (1960). The test requires residual and regression sum of squares. These
are printed out by any regression program but three regressions must be run:
(1) model estimates regressed on time or space unit, (2) field estimates
c regressed on time or space unit, (3) pool of model and field estimates
regressed on time or space unit. The null hypothesis states there is no
significant difference between the regression equations, and the alternative
hypothesis states that the two regression equations are different. The
retention of the null hypothesis sustains but does not prove the validity of
the model. The acceptance of the alternative hypothesis would invalidate the
model for this use. The test statistic is mathematically stated as follows:
Let TS - test statistic
where SS-REG(y) - sum of squares regression for the regression of model
estimates on time or distance.
SS-RES(Y) - sum of squares residuals for the regression of model
estimates on time or distance.
SS-REG(c) » sum of squares regression for the regression of
observed estimates on time or distance.
SS-RES(c) • sum of squares residuals for the regression of
observed estimates on time or distance.
SS-REG(Pooled) • sum of squares regression for the regression of
modeled and observed estimates pooled together.
SS-RES(Pooled) • sum of squares residuals for the regression of
modeled and observed estimates pooled together.
1-38
-------
TS - SS-REG(Y) + SS-REG(c) - SS-REG(Pooled)
[SS-RES(?) + SS-RES(c)]/(n? + nc-4)
p
TS <^ a l,n + nc-A -»• no reason to reject HQ
TS > oFl,n •»• nc-4 •*• reject HO, accept Hj
If the test statistic (TS) is less than or equal to the F-critical value for a
chosen probability (a) and the appropriate degrees of freedom (n? + nc-A),
the null hypothesis is sustained and the model has passed the field test. If
the test statistic is greater than the F-critical value, the alternative
hypothesis must be accepted and the model has failed this field test.
Another regression analysis, first used by economic modelers (Cohen and
Cyert 1961), is useful for evaluating agreement of model output and observed
data. This test is appropriate for deterministic modeling but not for
stochastic modeling (Aigner 1972). The null hypothesis states that the
regression line of model estimates (Y) on field data (c) has zero intercept
and slope of one. The retention of the null hypothesis confirms the validity
of the model. The alternative hypothesis states that either the intercept is
not zero or the slope of the zero-intercept model is not one. This would
imply that the model failed the field test.
To test the Intercept and slope of a regression line, the number of
observations should be at least twenty or thirty and the coefficient of
determination (R^) should be high (2. .80). The tests are not independent
and require two regression models (r • bC and y « a + be). First, the
intercept (a) of the model Y - a + be should be tested for equality to
zero (a « 0).
If the test is omitted on your stat-package, fun both a zero-Intercept
and intercept regression and compute this test statistic:
Let TS - test statistic
F - model output
T" - average of model output
Y - r - r
C • field observations
C • average field observations
c - C - "C
TS - n 2 + SS-REG(Ytc) - SS-REG(r.C)
SS-RES(Y,c) / (n-2)
1-39
-------
where n
SS-REG(Y,c)
SS-REG(Y,C)
SS-RES(Y.c)
number of pairs of observations
the sum of squares regression on the intercept model
the sum of squares regression on the zero-intercept
model
the sum of squares residual on the Intercept model
TS <_ aFl,n-2 -»• a not different from zero
>aF!,n_2
If the calculated test statistic is less than or equal to the critical
F-value, the zero-intercept model is correct and the slope of the
zero-intercept model should be tested for equality to one. The test statistic
would be calculated as follows:
Let
TS » test statistic
b - slope of zero-intercept regression line ( r- bC)
TS
(b - 1.0)
(n-1) I
C2
SS-RES (F,C)
The critical value is a two-tail t-test for a/2 and n-1 degrees of freedom.
ITS I £ atn-l •* slope is not different from one
ITS| > Otn-l •* slope differs from one
If the absolute value of the test statistic is less than or equal to the
critical value of t for °/2 (two-tail) and n-1 degrees of freedom the null
hypothesis Is sustained and the field test confirms the model validity. If
the intercept was different from zero the slope may still be tested for
equality to,one for diagnostic reasons but SS-RES (Y,c) and (n-2) should be
used in an analogous test of the intercept model. If the intercept was
statistically different from zero or if the slope of the zero intercept
regression line was statistically different from one, the model estimates
disagree with the field observations.
In summary, the more obvious statistical tests (t and Hotelling T2)
require more information (s, {S}) than available, but with appropriate
qualifications the chi-square test or regression analysis are reasonably
appropriate. Either can accommodate the variable dimensionality of the data.
The regression approach will require a computer program with a zero-intercept
option to manipulate the data, while the chi-square can be done on a hand-held
calculator. The user should choose the method best suited to his resources.
1-40
-------
SUMMARY
Field validations of aquatic fate and transport models require the
collection and quantification of numerous chemical, physical, and biological
parameters over time for model input variables and model output variables.
Because of the complexities associated with field validation a staggered
approach (steps 1-6) has been suggested in this report (Table 1-1). ;
However, one does not have to sequentially complete one step before proceeding
to the next since many steps can be completed simultaneously. In addition,
preliminary planning should include provisions to repeat experiments when
performing validations in complex environments or with compounds that are
subject to multiple fates.
1-41
-------
REFERENCES
Aigner, D. T. 1972. A Note on Verification of Computer Simulation Models.
Management Science. Vol. 18, No. 11, July, 1972.
Burns, L. A., D. M. Cline and R. R. Lassiter. 1982. EXAMS: An Exposure
Analysis Modeling System: User Manual and System Documentation. EPA-
600/3-82-023. U.S. EPA, Environmental Research Laboratory, Athens,
Georgia.
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler,
J. R. Freed, P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri,
W. R. Mabey, B. R. Holt, and C. Gould. 1979. Water-Related
Environmental Fate of 129 Priority Pollutants. Vol. I and II.
EPA-440/4-79-029a and EPA-440/4-79-029b. U.S. Environmental Protection
Agency, Washington, D.C.
Cochran, W. G. 1954. Some Methods for Strengthening the Common X^ Tests.
Biometrics 10:417-451.
Cohen, K. T. and R. M. Cyert. 1961. Computer Models in Dynamic Economics.
Quarterly Journal of Economics. Vol. LXXV, No. 1.
Davis, T. T. (Chairman). 1980. State-of-the-Art Report of the Hazardous
Substances Committee. In: Workshop on Verification of Water Quality
Models. EPA-600/9-80-016, U.S. Environmental Protection Agency,
Washington, D.C.
Donigian, A. S. Jr. 1980. Recommendation to Improve the Use of Models in
Decision Making. In: Workshop on Verification of Water Quality Models.
T. T. Davis, ed. EPA-600/9-80-016, U.S. Environmental Protection Agency,
Washington, D.C.
Hern, S. C., J. E. Pollard, and F. P. Beck. In Press. A Field Test of the
EXAMS Model in an Industrial Waste Pond. U.S. Environmental Protection
Agency, Las Vegas, Nevada.
Hoffman, F. D., and R. H. Gardner. In Press. Evaluation of Uncertainties in
Radiological Assessment. J. E. Till and H. R. Meyer, Eds. Nuclear
Regulatory Commission.
Hubbard, E. F., F. A. Kilpatrick, L. A. Martens, and J. F. Wilson, Jr. 1982.
Techniques of Water-Resources Investigations of the United States
Geological Survey. Book 3, Applications of Hydraulics, Chapter A9,
Measurement of Time of Travel and Dispersion in Streams by Dye Tracing.
Washington, D. C.
1-42
-------
Kilpatrick, F. A. 1970. Dosage Requirements for Slug Injections of Rhodanine
BA and WT Dyes. In: Geological Survey Research 1970, Chapter B.
Geological Survey Professional Paper 700-B, United States Printing
Office, Washington, D.C.
Longbottom, J. E., and J. J. Lichtenberg eds. 1982. Test Methods, Methods
for Organic Chemical Analysis of Municipal and Industrial Wastewater.
EPA-600/4-82-057. U.S. Environmental Protection Agency, Cincinnati, Ohio.
Lyman, W. J., W. F. Reehl, and D. H. Rosenblatt eds. 1982. Handbook of
Chemical property estimation methods: Environmental Behavior of
Organic Compounds. McGraw-Hill, New York.
Mabey, W. R., J. H. Smith, R. T. Podoll, H. L. Johnson, T. Mill, T. W. Chou,
J. Gates, I. W. Partridge, and D. Vandenberg. 1982. Aquatic Fate Process
Data for Organic Priority Pollutants. EPA-440/4-81-014. SRI Inter-
national, Menlo Park, California.
Mill, T., W. A. Mabey, D. C. Bomberger, T. W. Chou, D. G. Hendry and
J. H. Smith. 1982. Laboratory Protocols for Evaluating the Fate of
Organic Chemicals in Air and Water. EPA-600/3-82-022. SRI International,
Menlo Park, California.
Morrison, D. F. 1967. Multivariate Statistical Methods. McGraw-Hill Co.
New York. .
i'
O'Neill/ R. V. and R. H. Gardner. 1979. Sources of Uncertainty in Ecological
Models. In: . Methodology in Systems Modeling and Simulation,
B. P. Zergler, M. S. Elzas, G. J. Klir and T. I. Oren, eds. North-
Holland, Amsterdam.
Pendleton, A. F. (Chairman). 1980. Work Group on Surface Waters. National
Handbook of Recommended Methods for Water-Data Acquisition. Chapter
1: Surface Water. Office of Water Data Coordination, U.S. Geological
Survey. Reston, Virginia.
Pollard, J. E., S. C. Hern, and A. B. Crockett. 1983. A Field Test of the
EXAMS Model: The Monongahela River Study. EPA-600/X-83-002. U.S.
Environmental Protection Agency, Las Vegas, Nevada.
Sherma, J. 1981. Manual of Analytical Quality Control for Pesticides and
Related Compounds in Human and Environmental Samples. EPA-600/2-81-
095. U.S. Environmental Protection Agency, Cincinnati, Ohio.
Smith, J. H., W. R. Mabey, N. Bohonos, B. R. Holt, S. S. Lee, T. W. Chou,
D. C. Bomberger, and T. Mill. 1978. Environmental Pathways of Selected
Chemicals in Freshwater Systems: Part II. Laboratory Studies.
EPA-600/7-78-074, U.S. Environmental Protection Agency, Environmental
Research Laboratory, Athens, Georgia.
1-43
-------
Sokal, R. R. and J. Rohlf. 1969. Biometry. W. H. Freeman Co., San
Francisco.
Steel, R. G. D. and J. H. Torrie. 1960. Principles and Procedures of
Statistics. McGraw-Hill Co., Inc., New York.
Taylor, P. L. 1980. STORET: A Data Base for Models. In: Workshop oil
Verification of Water Quality Models. EPA-600/9-80-016, U.S. Environ-
mental Protection Agency, Washington, D.C.
U.S. EPA. 1978. Quality Assurance Guidelines for Biological Testing.
EPA-600/4-79-043, U.S. Environmental Protection Agency, Las Vegas,
Nevada.
U.S. EPA. 1979a. Handbook for Analytical Quality Control in Water and
Wastewater Laboratories. EPA-600/A-79-019, U.S. Environmental Protection
Agency, Cincinnati, Ohio.
U.S. EPA. 1979b. Guidelines Establishing Test Procedures for the Analysis
of Pollutants. Federal Register, Washington, D. C. 44(233):69464-
69575.
U.S. EPA. 1980a. Interim Guidelines and Specifications for Preparing
Quality Assurance Project Plans. QAMS-005/80. U.S. Environmental
Protection Agency, Office of Monitoring Systems and Quality Assurance,
Washington D. C.
U.S. EPA. 1980b. The Quality Assurance Bibliography. EPA-600/4-80-009, U.S.
Environmental Protection Agency, Washington, D.C.
U.S. EPA. 1982. Handbook for Sampling and Preservation of Water and
Wastewater. EPA-600/4-82-029, U.S. Environmental Protection Agency,
Cincinnati, Ohio.
Wilson, J. F. Jr. 1968. Fluorometric Procedures for Dye Tracing: U.S.
Geological Survey Techniques of Water-resources Investigations. Book
3, Chapter A12.
Yotsukura, N. and E. D. Cobb. 1972. Transverse Diffusion of Solutes in
Natural Streams. U.S. Geological Survey Professional Paper 582-C.
1-44
-------
MODEL INPUTS AND OUTPUTS
This part of the report deals with collection of the site specific input
data required to run and evaluate models. Two types of model input data are
required that are needed to describe the fate and transport processes and
pollutant loading data. In addition, actual pollutant concentrations that
correspond to those predicted by the model are required.
Table 2-1 summarizes specific model Inputs for the EXAMS (Burns et al.
1982) and TOXICS (Schnoor and McAvoy 1981) model. Some of the PEST (Park
1980) model inputs are Included under bioconcentration. The table also
lists environmental inputs that are currently being developed for incorpora-
tion into models, and speculative parameters; those inputs known to be
important to a process but for which the relationships have not been defined.
The input loadings to aquatic fate and transport models are summarized in
Table 2-2. The predicted model outputs for the EXAMS, TOXICS, and PEST
models are provided in Table 2-3.
2-1
-------
TABLE 2-1.
ENVIRONMENTAL INPUTS BY PROCESS, TO AQUATIC FATE AND
TRANSPORT MODELS
Environmental Input and Units
Currently*
Used In
In
Development
Future?
Biotransformation
Temperature (°C)
Total Bacteria Pop. (Cells/ml) or
(Cells/100 g dry sed.)
Active Degrading Pop. (X of Total)
Nutrients C/N, P (mg/1)
Acclimation State
PH (pH Units)
Dissolved Oxygen (mg/1)
Hydrolysis
POH (pH Units)
PH (pH Units)
Temperature (°C)
General Acids/Bases
Oxidation
Temperature (°C)
Oxidant Concentration (moles/1)
Reaeration (cm/hr)
Suspended Particulate (mg/1)
Dissolved Oxygen (mg/1)
Dissolved Organic Carbon (mg/1)
Photolysis
Depth (m)
Chlorophyll (mg/1)
Latitude (degrees)
Cloudiness (tenths)
Dissolved Organic Carbon (mg/1)
Suspended Sediment (mg/1)
Spectral light intensity at surface
Altitude (m)
Temperature (°C)
Time of Day (24 hr. time)
Time of Year
E
ET
E
ET
E
E
E
ET
E
E
E
E
E
E
X
X
X
X
X
X
X
X
X
X
X
*See end of table, page 2-4 for footnotes.
(continued)
2-2
-------
TABLE 2-1. (Continued)
Environmental Input and Units
Currently*
Used In
In
Development Future?
lonizatlon
POH (pH Units) E
PH (pH Units) E
Temperature (°C) E
Total Dissolved Solids (mg/1)
Ionic Strength
Volatilization
Temperature (°C) E
Compartment Dimensions, area and E
volume
Reaeration Rate (cm/hr) E
Wind (m/s) E
Slope (m/m)
Water Velocity (m/s)
Fetch
Sediment Sorption
Organic Carbon Content (X of dry E
sediment)
Percent Water of Benthic Sediment ET
(100 Fresh Wt.)
1 Dry Wt. 5"
Bulk Density Benthic Sed. (g/cc) ET
Suspended Sediment (mg/1) ET
Compartment Dimensions & Areas ET
Cation Exch. Cap. (meg/100 g dry
sediment)
Anion Exch. Cap. (meg/100 g dry
sediment)
Particle Size (mm)
PH of Sediment (pH Units)
Bioconcentration
Total Biomass (mg/1 or g/m2) ET
Planktonic Biomass (fraction of total) ET
Fish (g/m3) PT
Water Bugs (g/m3) P
*See end of table, page 2-4 for footnote.
X
X
X
X
X
X
X
(continued)
2-3
-------
TABLE 2-1. (Continued)
Environmental Input and Units
Currently In
Used In Development Future?
Zooplankton (g/m3) p
Phytoplankton (g/ra3) p
Particulate Organic Matter (g/ra3) p
Floating Parttculates Organic Matter P
(g/*3)
Floating Macrophytes (g/m3) p
Dissolved Organic Matter (g/m3) p
Zoobenthos (g/m3) p
Chlorophyll a (mg/1)
Fish by Species (g/ra3)
Fish by Age or Size Class (g/ra3)
Periphyton (g/m3)
Zoobenthos by Functional Group
(g/ra3)
Temperature (°C)
Dissolved Oxygen (mg/1)
Macrophytes, Rooted (g/m3)
Physical Transport
Evaporation-(mm/month) ET
Interflow (m3/hr) ET
NPS Sediment Load (kg/hr) ET
NPS Water Load (m3/hr) ET
Percent Water of Bottom Sed. (100 x ET
fresh/dry WT. sed.)
Rainfall (mm/month) E
Suspended Sediment (mg/1) ET
Bulk Density Bottom Sed. (g/cc) ET
Stream Inflow (m3/hr) ET
Stream Borne Sediment Inflow (kg/hr) ET
Compartment Volume (m3) ET
Eddy Diffusivity (m2/hr) ET
Cross Section Area for Dispersive ET
Exchange (m2)
Surface Area (m2) ET
Dist. Between Compt. Centers (m) ET
Compartment Dimensions L, W, H (m) ET
Sediment Bed Load (kg/hr) ET
Planktonic Biomass (mg/1) E
Water Velocity (m/s) T
Dissolved Organic Carbon (mg/1)
Bed Load by Part. Size Classes (%)
Total Organic Carbon (mg/1)
X
X
X
X
X
X
X
X
X
X
E - EXAMS model suspended sediment settling velocity dispersive bed
scour/resuspenslon
T - TOXICS model
P - PEST model (incomplete)
2-4
-------
TABLE 2-2. POLLUTANT LOADING INPUTS TO AQUATIC FATE AND TRANSPORT MODELS
Currently In
Loading Input and Units Used In Development Future?
Aerial Drift Loading (kg/hr) E
Loading via Ground Water (kg/hr) E
Rainfall Loading (kg/hr) E
Loading via Stream Flow (kg/hr) ET
Loading via NPS Flow (kg/hr) E X
Loading via Particulate Organic Matter T X
(kg/hr)
Loading via Suspended Sed. (kg/hr) X
Loading via NPS Sed. (kg/hr) X
Loading via Bed Load (kg/hr) X
Loading via Plankton (kg/hr) X
Loading via Fish, Birds (kg/hr) X
Loading via Point Sources (kg/hr) E
E - EXAMS model (steady state)
T - TOXICS model (time variable)
TABLE 2-3. POLLUTANT CONCENTRATIONS PREDICTED BY AQUATIC FATE AND
TRANSPORT MODELS
Currently In
Model Output and Units Used In Development Future?
Water, Dissolved (mg/1) ET
Suspended Sediment (rag/kg) ET
Biota in Water Column (mg/1) E T
Sed. Pore Water (mg/1) ET
Benthic Sediment-Sorbed (mg/kg) ET
Benthic Biota (g/m2) E T
Water Column-Total (mg/1) ET
Benthic Sediment-Total (mg/kg) E T
Zooplankton (g/m3 of H20) P
Waterbugs (g/m3 of H20) P
Phytoplankton (g/m3 of 11,0) P
Macrophytes (g/m3 of H20) P
Particulate Organic Matter (g/m3 of P
H20)
Floating Organic Matter (g/mj of H-O) P
Clay (g/m3 of H20) P
Fish (g/m3 of H20) P T
E - EXAMS model
T - TOXICS model
P - PEST model
2-5
-------
REFERENCES
Burns, L. A., D. M. Cline, and R. R. Lassiter. 1982. Exposure Analysis
Monitoring System (EXAMS): User Manual and System Documentation. EPA
600/3-82-023 U.S. Environmental Protection Agency, Athens, Georgia.
Park, R. A., C. I. Connolly, J. R. Albanese, L. S. Clesceri, G. W. Heitzman,
H. H. Herbrandson, B. H. Indyke, J. R. Loche, S. Ross, D. D. Sharma,
and W. W. Shuster. 1980. Modeling Transport and Behavior of Pesticides
and Other Toxic Organic Materials in Aquatic Environments. Report No.
7 Center for Ecological Modeling, Rensselaer Polytechnic Institute,
Troy, N.Y.
Schnoor, J. L. and D. C. McAvoy. 1981. A Pesticides Transport and
Bioconcentration Model. Journal of the Environmental Engineering
Division, Proc. of ASCE Vol. 107-EE6:1229-1246.
2-6
-------
BIOTRANSFORMATION
PROCESS DESCRIPTION -
Microorgaaisras can adapt to a variety of environmental conditions and
biotransform and/or biodegrade a very large number of organic chemicals. Bio-
degradation Is defined as a biologically mediated reduction In complexity of
a particular chemical by splitting off one or more constitutent groups or
components. Whereas, biotransformation is an encompassing term which includes
biodegradatlon and biologically mediated additions to the parent chemical. It
has been stated that any compound which is synthesized biologically can be
degraded mlcrobiologically. However, many man-made compounds are strongly
resistant to microblal degradation (recalcitrant) such as certain pesticides,
hard detergents, petroleum hydrocarbons, and plastics (Alexander 1964).
Microbial transformation of a toxic substance may yield a product that Is
more or less toxic than the parent compound. In addition, transformation
products in conjunction with the parent compound can show synergistic
toxlcities.
Although degradation might be accomplished by any living organism, micro-
organisms play the dominant role because of their high numbers and catabolic
versatility, species diversity and metabolic rate per unit of weight. Algae,
actinomycetes, protozoans, and fungi can all participate in the degradation of
organics. Bacteria normally dominate, however, in the aquatic environment.
Fungi often are important decomposers under acidic sediment conditions. Micro-
bial decomposition of natural organics is normally accomplished for energy and
to produce degradation products which can be used as cellular constituents.
From 60 to 80 percent of the carbon in organic matter assimilated by bacteria
Is liberated as carbon dioxide and the remaining 20 to 40 percent is converted
Into cellular material. This material is the major source of bacterial biomass,
most of which is devoured by protozoa and higher life forms in the aquatic
environment (Tallon 1969).
Biological degradation of organic pollutants is often accomplished by a
number of species sequentially In an "assembly line" decomposition. The first
metabolic product then may serve as the substrate for another species and so
on until complete mineralization has been accomplished. However, intermediate
metabolic products are often produced which are resistant to further breakdown
or are only slowly degraded In the aquatic environment.
Alexander (1964) lists the following factors which influence the biotrans-
formation of organic materials:
1. Absence of essential growth factor - Growth would not occur in the
absence of nitrogen or phosphorus.
3-1
-------
2. Toxlcity of the environment - This may result from the pollutant,
biologically generated inhibitors (antibiotics) or toxins, high salt
concentrations, extremes in pH, temperature, light, or pressure.
3. Structural characteristics of the molecule - Molecular configurations
which do not permit the formation of an enzyme - substrate complex.
4. Inaccessibility of the substrate - The compound may exist in a
micro-environment which prevents or lessens microbial attack, e.g.,
sorbed onto clay minerals or embedded within a matrix.
5. Inactivation of requisite enzymes - Enzymes may lose activity by
sorption to clay minerals or may be inhibited by other substances
(metals) in the environment.
6. Inability of the microorganism community to metabolize the compound
because of some physiological inadequacy - An enzyme capable of
degrading the compound may not exist or the substance may not be
able to penetrate into the cells where appropriate enzymes do exist.
ENVIRONMENTAL FACTORS
Limiting Nutrients
Many microorganisms are adapted to sporadic occurrences of suitable energy
sources (e.g., autumn leaf-fall, death of large animals, etc.) and can grow
very rapidly when a suitable organic (energy-yielding) substrate is available.
The limiting nutrient concept, derived from Liebig's law of the minimum, is
that the nutrient that is least available relative to the growth of a given
organism imposes primary limitation on the growth of that organism. Table 3-1
lists the nutrients required by certain microorganisms. The availability of
carbon, nitrogen, and phosphorus frequently inhibit microbial growth.
Dissolved Oxygen
The quantity of dissolved oxygen present in water influences the kinds of
organisms and the rate of decomposition. Four groups of microorganisms may be
distinguished according to their oxygen requirements (Rheinheimer 1974);
1. Obligate aerobes - which grow only in the presence of oxygen,
2. Microaerophiles - which grow optimally at low oxygen concentrations,
3. Facultative aerobes - which grow in presence or absence of oxygen,
and
4. Obligate anaerobes - which grow only in the absence of oxygen.
The majority of aquatic bacteria are facultative aerobes. Molecular
oxygen is vital for obligatory aerobes as it is necessary as a terminal
hydrogen acceptor in respiration. Fluctuations of ambient oxygen
concentrations normally do not affect obligatory aerobes. For example, the
3-2
-------
TABLE 3-1. NUTRIENTS REQUIRED BY CERTAIN MICROORGANISMS
1. Energy Source
2. Electron acceptor
3. Carbon source
A. Growth factors
a. Amino acids
b. Vitamins
Organic compounds
Inorganic compounds
Sunlight
Organic compounds
N03-,N02-,N20,S04-,C02,Fe
C02,HC03 -
Organic compounds
Alanine, aspartic acid, glutamlc. acid, etc.
Thiamine, blotin, pyridoxlne, riboflavin,
nicotinic acid, pantothenic acid,
-aminobenzolc acid, folic acid, thioctic
acid, Bi2» etc.
c. Others
Purine bases, pyrlmidine bases, choline,
inositol, peptides, etc.
N,P,K,Mg,S,Fe,Ca,Mn,Zn,Cu,Co,Mo,Cl,Si,
B,Na,V,I
Modified from Alexander (1961),
nitrate bacteria Nitrobacteria wlnogradski maintains its oxidation rate at
dissolved oxygen levels from saturation down to 2 ppm. Only a further
decrease of oxygen makes the oxidation rate drop. In contrast, obligatory
anaerobes are Impaired at very low oxygen levels (Rheinheimer 1974).
Anaerobic decomposition of organic matter is normally much slower than
aerobic decomposition. There is a lower cell yield and degradation is
incomplete under anaerobic conditions. Generally the end products of aerobic
decomposition are carbon dioxide and water while anaerobic respiration often
yields organic acids (Mitchell 1974).
Concentration of Toxic Organic
The concentration of a toxic organic pollutant is an important factor
influencing its. degradation rate. At high concentrations an organic pollutant
may demonstrate bactericidal or bacterlostatic properties whereas at low
concentrations the organic pollutant may be readily degraded. Phenol is
commonly used as a disinfectant at high concentrations (2 to 5 percent);
however, it is readily used by bacteria as a substrate at low concentrations
(Pelczar and Reid 1972). Any biological system can be overloaded with a toxic
organic pollutant. For example, chlorodiphenyl oxide has a slower degradation
rate at 9.2 mg/1 than at lower concentrations (Branson 1978).
3-3
-------
Besides, bactericidal and bacterlostatic affects, biotransformation
rate constants may not be independent of chemical concentration (usually
assumed in fate models). The biotransformation rate may change at low
contaminant concentrations found in most natural environments. Unfortunately,
roost laboratory tests to determine biotransformation rates are conducted
with concentrations of test chemicals far higher than those found in rivers,
lakes and marine waters. Little attention is given to how the rate of
transformation In artificial microcosms approximates the rate under natural
conditions (Boethling and Alexander 1979).
Temperature
Temperature is one of the most important environmental conditions
determining degradation rates. The temperature coefficient of biological
processes is sometimes expressed in terms of QIQ* This value is the ratio
of the velocity constant of a process at a given temperature to the velocity
constant at a temperature 10°C higher. Optimal temperatures for degradation
and the response of degradation to different temperature can be reasonably
predicted (Nesbitt and Watson 1980a).
Three groups of bacteria may be distinguished by their temperature
tolerance. Most organisms are mesophiles (15° to 45°C) with an optimum range
of 25° to 35°C. Psychrophiles grow best at temperatures below 20°C while
thermophiles grow readily from A5° to 65°C (Alexander 1961). High
concentrations of mesophiles and psychrophiles can be detected in many
aquatic systems, while thermophiles are basically restricted to thermal
springs. An increase in temperature results in an increased rate of
biologically mediated decomposition, e.g., pesticide degradation (Lichtenstein
1972) and oil degradation (Westlake et al. 1978) by bacteria are increased at
elevated temperatures.
The growth and reproduction of microorganisms are affected by the
hydrogen ion concentration of the medium. Most bacteria can only grow within
the pH range of 4 to 9 while the optimum for most aquatic bacteria is between
pH 6.5 to 8.5. This corresponds to the pH range of most water bodies. There
are more acidophilic fungi than bacteria; therefore, the proportion of fungi
in the microflora Increase In acid. waters or sediments.
Acclimation
There can be a considerable lag between the exposure of a new chemical
to the organism and the initiation of degradation by microorganisms. This
lag time exists because organisms must acclimatize to a new substrate In
any of several ways. Cells may adapt by synthesizing new enzymes in response
to new substrate added to the environment (enzyme induction). Second, a
change in the environment may allow for the selection of mutant types. Third,
a change in the environment may increase the population of existing organisms
that grow on the added substrate (Brock 1966).
3-4
-------
In nature, acclimation time may have a great influence on the rate of
degradation since microorganisms may not have an opportunity to acclimate to a
new chemical (Howard et al. 1978). This is particularly important in flowing
waters that receive periodic discharges. Under these circumstances, attached
organisms will only be exposed to a pollutant "slug" for a short time while
planktonlc forms traveling the "slug" will have a greater opportunity to
acclimate. However, dilution effects may result in such low concentrations of
the pollutant that the acclimation response may not be fully realized. Since
exposure times to intermittent discharges are greater in lake or pond environ-
ments than in fast flowing streams, the probability for full acclimation is
enhanced. However, continuous discharges provide the greater opportunity for
mlcroblal acclimation, regardless of the environment. Experiments to determine
microbial degradation rates Involving long time periods may allow development
of a microbial population acclimated to the compound present. This can lead
to apparent high rate constants. Perhaps even more significant, rate constants
developed from short-term experiments may lead to apparent low rate constants
(Branson 1978).
Many organic contaminants only occur at low concentrations (< 100 ug/1) in
natural aquatic systems. Dlckson et al. (1981) found no lag phase for
napthalene and lindane biotransformatlon at concentrations ranging from 100 to
500 ug/1. However, experiments with phenol using the same concentration range
demonstrated approximately a 3-day lag phase. Therefore, acclimation state may
not be an important consideration at low contaminant concentrations for certain
chemicals.
Major Environmental Factors
Major factors governing microbial decomposition include temperature, pH,
acclimation state, organic matter levels and sufficient levels of nitrogen and
phosphorus. As a general rule, environmental modifications that favor
microbial proliferation Increase site specific degradation rates. Thus, a
rise in temperature, increased available carbon and neutral pH, all tend to
Increase site specific degradation rates. Table 3-2 lists the major
environmental factors that influence microbial degradation rates.
TABLE 3-2. MAJOR ENVIRONMENTAL FACTORS THAT
INFLUENCE DEGRADATION RATES
Nutrients - Carbon, nitrogen, phosphorus
Temperature
pH
Dissolved oxygen concentration
Size and composition of the active microbial
community
Acclimation state
3-5
-------
COMPOUND-RATE RELATIONSHIPS
Obviously, the specific organic chemical of concern governs the
transformation rate under optimal environmental conditions for the given
chemical. Half-lives vary from hours for certain compounds such as sugars,
starches, etc., to months or years for other compounds such as certain
chlorinated hydrocarbons or trisubstituted phenoxyaliphatic acids. In recent
years considerable effort has been devoted to relating the chemical structure
of a molecule to its oicrobial biodegradability. Although these investiga-
tions have been restricted to a limited number of compounds, a few
generalizations can be made:
1. Short-chain aliphatic hydrocarbons are not as readily degraded as
those of high molecular weight (Tallon 1969).
2. Straight-chain alkanes are degraded more rapidly than branched
alkanes (Mitchell 1974).
3. Unsaturated aliphatics are more readily attacked than the
corresponding saturated hydro-carbon (Tallon 1969).
4. The branched-chain alkyl benzene sulfonate (ABS) detergents are
poorly degraded compared to linear alkyl sulfonates (LAS) detergents
(Mitchell 1974).
5. A molecule is more resistant when a carbon atom in a chain is
replaced by oxygen, sulfur, nitrogen, halogens, nitrile, or
phosphorus (Tallon 1969).
6. Many aromatic hydrocarbons are strongly recalcitrant and frequently
accumulate in the sediments; e.g., no organisms have been Isolated
that will grow with polycyclic hydrocarbons that contain more than
three aromatic rings (Gibson 1978).
7. Comparable resistance to biotransformation of pesticides as a function
of structure is presented below. The compounds are ranked from the
most to the least resistant to microbial breakdown (Mitchell 1974).
Chlorinated hydrocarbons
Trisubstituted phenoxyaliphatic. acids
Dlsubstituted phenoxyaliphatic acids
Monosubstituted phenoxyaliphatic acids
Long-chain phenoxyaliphatic acids
Organophosphates
Aliphatic acids
BIOTRANSFORMATION RATE CONSTANT
Biotransformation rate constants have been expressed as first order and
as second order rate constants (pseudo-first order). First order rates
need only chemical concentration to describe biodegradation although the
3-6
-------
importance of compounds, the character of the microbial population, and
nonlinear responses over ranges of concentration are recognized (Boethling
annd Alexander 1980, Larson 1980, Johnson 1980). Second order or pseudo-
first order rate constants to describe biodegradation rates require bacterial
population numbers and chemical concentration (Baughman et al. 1980).
Apparently, heterotrophic microbial biomass is a major factor in biotrans-
formation of chemicals in aquatic systems. However, others have found .
little or no relationship between degradation rates and estimates of biomass
(Nesbitt and Watson 1980b, Dickson et al. 1981). Organic priority pollutants
for which microbial degradation is a significant fate process are listed
in Table 3-3.
MICROBIAL METHODS
Microbial Degradation Rate Constant
The microbial degradation rate constant is usually determined in the
laboratory and, for the purposes of these guidelines, is considered a chemical
input parameter. The rate constants which were developed under laboratory
conditions which most closely approximate conditions at the validation site
should be chosen. Factors to be considered in the selection are chemical
species and concentration, microbial population and acclimation state.
For some toxic organics, such as phenol, this rate constant is extremely
critical, i.e., the model outputs are very sensitive to a change in the
rate constants. Three rate constants may be important depending upon the
model, pollutant, biomass, etc. They are the degradation rates of the
pollutant in the soluble, biosorbed, and sediment sorbed forms. The
importance of the rate constants can be determined during the sensitivity
analyses.
If rate constants are expressed as percent disappearance of a pollutant
per unit of biomass (or cell count) per unit of time, both field and
laboratory measurements for the rate constants of biomass or cell count
must be Identical or equivalent. The usual procedure has been to use standard
plate counts (SPC) to estimate the number of cells per volume. Other means of
measuring cell counts or biomass include direct count methods such as
epifluorescence, the measurement of adenosine triphosphate (ATP), and the
determination of bound llpopolysaccharide (LPS).
There are advantages and disadvantages to each of the above techniques of
determining microbial populations. The most appropriate technique depends
upon how well the measurement of biomass or cell count correlates with
microbial degradation at the selected site. In the development of a
biodegradation rate constant, biological lag time should also be determined.
Acclimation State
An investigator should select the appropriate degradation rate constant
which reflects the acclimation state of the bacteria at the selected site and
3-7
-------
TABLE 3-3. ORGANIC PRIORITY POLLUTANTS WHOSE AQUATIC FATE IS
SIGNIFICANTLY AFFECTED BY MICROBIAL DEGRADATION
Pesticides
Acrolein
Aldrin
DDD
Endosulfan and Endosulfan sulfate
Hexachlorocyclohexane
Y-Hexachlorocyclohexane (Llndane)
TCDO
2-Chloronaphthalene
Monocyclic Aromatics
Nitrobenzene
2,4-Dinitrotoluene
2,6-Dinltrotoluene
Phenol
2,4-Dichlorophenol
Phthalate Esters
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Bis (2-Ethylhexyl) phthalate
Butyl benzyl phthalate
Polycyclic Aromatic Hydrocarbons
Acenaphthalene
Acenaphthylene :
Anthracene
Benzo [a] anthracene
Benzo [b] fluorantheae
Benzo [e] fluoranthene
Benzo [ghi] perylene
Benzo [a] pyrene
Chrysene
Dibenzo [a,h] anthracene
Fluoranthene
Fluorene
Indeno [1,2,3-cd] pyrene
Naphthalene
Phenanthrene
Pyrene
Source: Callahan et al. 1979.
the concentration of the chemical. Therefore, if a site only receives
periodic discharges of the compound of interest, one would utilize a rate
constant which was developed without regard for acclimation. Whereas, if
the selected site receives continuous discharge, the rate constant should
be developed with organisms which have been acclimated.
Methods to Estimate Total Microbial Populations
Water—
The standard plate count method can be used to estimate the viable
aerobic and facultative anaerobic bacteria in an aquatic environment (Bordner
and Winter 1978 and APHA 1980). An aliquot of the water sample or its
dilution is pipetted into a sterile petri dish and a liquified, tempered
agar medium added. The petri dish is then rotated to evenly distribute
the bacteria. Theoretically, each bacterium present multiplies into a
3-8
-------
visible colony which is counted to estimate the bacterial population, How-
ever, the procedure does not yield the true number of bacteria since not
all viable bacterial cells in the water sample can reproduce under a single
set of culture conditions imposed in the test. Underestimates of bacterial
density also occur because, 1) some microorganisms will not develop within
the specified incubation period, 2) the procedure does not allow fastidious
aerobes or obligate anaerobes to develop, and 3) clumps of organisms will
appear as a single colony. Another disadvantage to the SPC method is the
relatively short holding time. Sample analysis should be initiated as
soon as possible to minimize changes in the bacterial population. If the
sample cannot be analyzed at once, it should be refrigerated and analyzed
within six hours (Bordner and Winter 1978). Incubation temperature and
duration must be identical to those used to determine the laboratory based
rate constants. This simple technique is a useful tool for determining
bacterial density for fate and transport studies if, 1) the total bacterial
population is capable of degrading the pollutant, or 2) the percent of
the total population detected by the SPC is relatively constant between
field sampling and rate determination tests.
Prescott et al. (1946) reported that the standard deviation of individual
counts from plates with 30-300 colonies will vary from 0-30 percent. The
plating error was 10 percent for plate counts within the 100-300 range. In
addition to plating error, a dilution error of about 3 percent is expected for
each dilution required. Large variations can be expected from high density
samples, e.g..sewage. Laboratory personnel should be able to duplicate their
plate count values on the same plate within 5 percent and the count of others
within 10 percent.
Direct microscope counting methods can be used to estimate bacterial
numbers in aquatic environments. The advantage of direct counts is that the
total population is enumerated rather than some percentage of the total
population as with the SPC method. Therefore, direct counts are generally
much higher (1 to 3 orders of magnitude) than plate counts. The eplfluores-
cence counting method is suggested (Watson et al. 1977). An eptfluoresrence
microscope is employed to count bacteria that have been concentrated on
membrane filters and stained with acridine orange. This method reduces
errors of improper identification of detrltal material versus bacterial
cells commonly associated with previous direct count methods. Another
advantage with this method is that water samples may be preserved for 2
weeks prior to examination. Also, the method seems to yield more precise
information (less variance) than SPC methods. The method is relatively
inexpensive (about 30 dollars per sample) and only requires about 1 hour to
perform. Disadvantages include: specialized equipment is required (i.e.,
Zeiss standard microscope equipped with an epifluorescent illumination system
containing a 100-W halogen lamp, a BG12 excitation filter, a LP510 barrier
filter and an FT beam splitter), and the investigator cannot distinguish
between viable and non-viable cells. This procedure would eliminate one of
the difficulties associated with the SPC, i.e., detecting only a percentage of
viable total population.
Biomass determinations eliminate the problems associated with SPC and
direct counts but they possess other disadvantages. The biomass of microbial
3-9
-------
populations can be estimated by measuring the amount of cellular bound
lipopolysaccharide (LPS), adenosine triphosphate (ATP), or limulus amebocyte
lysatc (LAL).
Measuring ATP levels provides an estimate of the total viable planktonic
biomass. ATP occurs in all living plant and animal cells and is not
associated with non-living particulate material. The ratio of ATP to biomass
varies from species to species and with the physiological state of a cell.
However, the method can be used to estimate biomass (APHA 1980).
The ATP content of a water sample is determined by concentrating the
organisms on a membrane filter. The ATP content of the organisms is extracted
by boiling the filter in a buffer solution. The ATP thus extracted is reacted
with luciferase in the presence of oxygen and magnesium to produce water,
luciferin, adenosine monophosphate (AMP), inorganic polyphosphate (PPi), and
light. The amount of light emmitted is proportional to the ATP content of the
water sample (APHA 1980). The method is simple and relatively inexpensive,
and the instrument is stable and reliable. The disadvantages of the procedure
are, 1) it does not estimate the active degrading biomass, and 2) it does not
distinguish between living forms, e.g., bacteria, zooplankton, algae, fungi or
yeast.
The biomass of bacteria in various waters can also be estimated by
measuring the amount of cellular bound LPS. With this technique, LPS, which
is a component of the cell walls of all gram-negative bacteria including
cyanophytes (blue-green algae), is quantified.
Limulus amebocyte lysate (an aqueous extract of the blood cells of the
horseshoe crab) will clot in the presence of LPS. Experimental conditions
can be controlled so that LAL will become turbid rather than clot. Turbidity
is then proportional to the concentration of LPS. Both bound and free LPS
occur in aquatic samples. The cellular bound LPS must be quantified to
estimate bacterial biomass. Centrifugation is used to separate the free from
the bound LPS (Watson and Hobbie 1979). The procedure is relatively simple,
inexpensive and yields reproducible results.
The test can be used to estimate bacterial biomass since most bacteria in
the aquatic environment are gram-negative. However, the procedure does not
detect gram-positive bacteria which could be an Important population component
in certain aquatic environments. In addition, the method estimates blue-green
algae biomass which can be an important planktonic component in lakes,
reservoirs, and slow moving streams.
Microbial biomass may also be estimated using the epifluorescence
technique. However, it is extremely time consuming to make enough
measurements to calculate the bacterial biomass from cell numbers and size.
Sediments
Microbial populations in sediments can be estimated using modification of
the plate count method and the epifluorescence technique. The same advantages
and disadvantages of each method apply to their sediment counterparts as
3-10
-------
previously described. Plate counts can be used to measure bacterial numbers
by transferring a known quantity of sediment (10 grams) into a dilution bottle
containing glass beads. The bottle is then shaken for 10 minutes In a
reciprocating shaker. Dilutions are made and plated out as described for the
SPG method (Clark 1965).
The following modification can be used to estimate bacterial numbers In
sediments using the epif luorescence technique: The sediment sample is poured
into a blender and 100 ml of pref lltered water Is added. Prefiltered water Is
water which passed through a filter of 0.2 m pore size no more than 6-3 hours
before use. The contents are mixed at high speed for 1 minute. An aliquot of
0.5 ml of the mixture is removed and directly counted.
Active Degrading Population
As previously stated, all the above procedures are used as indices or
measures of total biomass or population size. A possible procedure for
measuring the active degrading population Is to use standard plating
procedures with the organic compound of interest as the sole carbon source.
Though the procedure has many disadvantages as discussed below, there does not
appear to be any simple practical means of determining the size of the
microbial population actually doing the degrading.
one disadvantage of the plate count or other methods that use
the pollutant of Interest as the sole carbon source is a gross underestlmatloa
of the active population because:
1. The pollutant may be toxic at concentrations needed to produce
growth.
2. An acclimation period which may be required is not adequately
provided.
3. Often pollutants are degraded bacteriologically in an "assembly
line," where essential prerequisites (extracellular enzymes,
vitamins, cofactors, etc.) may not be produced in the test
environment.
4. The environmental conditions (pH, temperature, etc.) of the test
environment may not be suitable for growth of the active population.
In addition, there is a series of reactions where a substance is only
oxidized in the presence of other organic substances which would not be
provided in a sole source carbon media. However, If the above series of
significant problems can be eliminated, any method which could accurately
measure the active population would be vastly superior as it would measure
only the population of interest.
It is often assumed that biological degradation is essentially bacterial
degradation. However, some compounds may be significantly degraded by algae,
fungi, etc., and If their population is large enough, they may have to be
3-11
-------
considered. If the fate and transport model being evaluated can accept such
biological degradation rate constants, these data will also have to be
acquired. The procedure for field measurements of blomass, cells, etc., must
be the same as was used in determining the rate constants.
The measurement of algae, macroinvertebrates, zooplankton, and fish
biomass are discussed in the Bioconcentration section of this report.
Whether such measures of biomass are necessary for validating a model depends
upon whether the model has provisions for such inputs and whether the
compound being studied undergoes significant degradation by the communities.
Methods to measure pH are presented in the Hydrolysis section of the report.
Temperature
Biological degradation of organic compounds is simply biologic*!
mediation of standard chemical reactions. These reactions usually result in a
release of energy which is then used by the organism. The Qio'8 °f each
biological degradation rate constant should be determined. This is a
laboratory procedure and, since it is considered a chemical input parameter,
procedures for developing Q}Q rates are not provided here.
Environmental temperature Is, however, a measurement which must be made
in the field during the testing of models. Hie temperature of interest is
that of the biota during the degradation process. Since most aquatic life
forms have no thermoregulatory mechanisms, they are considered to be at the
same temperature as the surrounding water.
Temperature measurement methods are rather simple and may involve the use
of thermometers or thermocouples. Instruments must be accurately calibrated
and procedures for their use are found in APHA (1980). The sampling design
for determining temperature depends upon the model and the aquatic environment
being sampled. If a compartments! type of model is being used, the temperature
input should be representative of the biomass temperature of that compartment.
When the range of temperatures detected within a compartment is great,
consideration should be given to reducing compartment size. Temperature
profiles of lakes and streams are useful in determining the size of
compartments since the temperature of well mixed water Is constant.
NUTRIENTS
Satisfactory methods to analyze water samples for total phosphorus, ortho
phosphorus, ammonia, nitrlte-nltrate-nltrogen, and Kjeldahl-nltrogen can be
found in methods for chemical analysis of water and wastes (U.S. EPA 1979)
or (APHA 1980).
DISSOLVED OXYGEN
Both membrane electrode and the Winkler methods and its modifications can
be used to measure dissolved oxygen (U.S. EPA 1979, APHA 1980). The Winkler
method is a tltrlmetric procedure based on the oxidizing property of dissolved
3-12
-------
oxygen, while the membrane electrode procedure Is based on the rate of
diffusion of molecular oxygen across a membrane.
3-13
-------
REFERENCES
Alexander, M. 1961. Introduction to Soil Microbiology. John Wiley and Sons,
Inc. New York, New York.
Alexander, M. 1964. Microbiology of Pesticides and Related Hydrocarbons.
In: Principles and Applications in Aquatic Microbiology. H. Huekelekian
and N. C. Dondero (eds.). John Wiley and Sons, Inc. New York, New York.
APHA. 1980. Standard Methods for the Examination of Water and Wastewater.
Fifteenth edition. APHA/AWWA/WPCF. Washington, D.C.
Baughman, G. L., D. F. Paris, and W. C. Steen. 1980. Quantitative Expression
of Biotransformation Rate. In: Biotransformation and Fate of Chemicals
in the Aquatic Environment. A. W. Maki, K. L. Dlchson and J. Caairns,
Jr. eds. Washington, D. C. Amer. Soc. for Microbiol.
Boethling, R. S., and M. Alexander, 1979. Effect of Concentration of Organic
Chemicals on their Biodegradation by Natural Microbial Communities.
Applied and Environ. Microbial. 37(6):1211-1216.
Bordner, R. H. and J. A. Winter (eds.). 1978. Microbiological Methods for
Monitoring the Environment. EPA-600/8-78-017, U.S. Environmental
Protection Agency, EMSL-Cincinnati, Cincinnati, Ohio.
Branson, D. R. 1978. Predicting the Fate of Chemicals in the Aquatic Environ-
ment from Laboratory Data. In: Estimating the Hazard of Chemical
Substances to Aquatic Life. ASTM STP 657. J. Cairns Jr., K. L. Dickson,
and A. W. Maki, eds. American Society for Testing and Materials,
Philadelphia, Pennsylvania.
Brock, T. D. 1966. Principles of Microbial Ecology. Prentice-Hall, Inc.
Englewood Cliffs, New Jersey.
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler, J. R. Freed,
P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri, W. R. Mabey,
B. R. Holt, and C. Gould. 1979. Water-Related Environmental Fate of
129 Priority Pollutants. Vol. I and II. EPA-440/4-79-029a and
EPA-440/4-79-029b. U.S. Environmental Protection Agency, Washington, D.C.
Clark, F. E. 1965. Agar-Plate Method for Total Microbial Count. In: Methods
of Soil Analysis. C. A. Black, ed. American Society of Agronomy.
Madison, Wisconsin.
Dickson, K. L., J. H. Rogers, Jr., and F. Y. Saleh. 1981. Measuring Rate
Constants for Chemicals in Simple Model Aquatic Laboratory Systems.
Final Report. Chemical Manufacturers Association, Project No. ENV-8-W.
Gibson, D. T. 1978. Microbial Transformations of Aromatic Pollutants. In:
Aquatic Pollutants: Transformation and Biological Effects, 0. Hunt-
zinger, J. H. Van Lelyveld, and B. C. J. Zoeteman, eds. Pergamon Press,
New York, New York.
3-14
-------
Howard, P. H., J. Saxena, and H. Sikka. 1978. Determining the Fate of
Chemicals. Env. Sci. Tech. 12:398-407.
Johnson, B. T. 1980. Approaches to Estimating Microbial Degradation of
Chemical and Contaminants in Freshwater Ecosystems, In: Biotransformation
and Fate of Chemicals in the Aquatic Environment. A. W. Maki, K. L.
Dickson and J. Cairns, Jr. eds. Washington D. C. Amer. Soc. for
Microbiol.
Larson, R. J. 1980. Role of Biodegradation Kinetics in Predicting
Environmental Fate, In: Biotransformation and Fate of Chemicals in
the Aquatic Environment. A. W. Maki, K. L. Dickson and J. Cairns, Jr.
eds. Washington D. C. Amer. Soc. for Microbiol.
Lichtenstein, E. P. 1972. Environmental Factors Affecting Fate of
Pesticides. In: Degradation of Synthetic Organic Molecules in the
Biosphere. National Academy of Sciences. Washington, D.C.
Mitchell, R. 197A. Introduction to Environmental Microbiology. Prentice
Hall, Inc. Englewood Cliffs, New Jersey.
Nesbitt, J. J. and J. R. Watson. 1980a. Degradation of the Herbicide 2,4-
D in River Water-II. The Role of Suspended Sediment, Nutrients and
Water Temperature. Water Research 14:1689-1694.
Nesbitt, J. J. and J. R. Watson. 1980b. Degradation of the Herbicide 2,4-D
.' in River Water-I.< Description of the Study Area and Survey Rate
Determining Factors. Water Research 14:1683-1688.
Pelczar, M. J. Jr. and R. D. Reid. 1972. Microbiology. McGraw-Hill Book Co.
New York, New York.
Rheinheimer, G. 1974. Aquatic Microbiology. John Wiley and Sons, Inc. New
York, New York.
Prescott, S. C., C. E. A. Winslow, and M. H. McGrady. 1946. Water Bacter-
iology (6th ed.). John Wiley and Sons, Inc.
Tallon, G. R. 1969. Microbes and Microorganisms in Water. In: Microorganic
Matter in Water. ASTM Special Publication No. 448.
U.S. EPA. 1979. Methods for Chemical Analysis of Water and Wastes.
EPA-600/4-79-020, U.S. Environmental Protection Agency, Cincinnati, Ohio.
Watson, S. W. and J. E. Hobbie. 1979. Measurement of Bacterial Biomass as
Lipopolisaccharide. American Society for Testing and Materials. Special
Technical Publication 695.
Watson, S. W., T. J. Novitsky, H. L. Quinby, and F. W. Valois. 1977.
Determination of Bacterial Number and Biomass in the Marine Environment.
Appl. Environ. Microbiol. 33:940-946.
3-15
-------
Westlake, D. W. S., F. D. Cook, and A. M. Jabson. 1978. Microbial Degrada-
tion of Petroleum Hydrocarbons. EPA-600/17-78-148. U.S. Environmental
Protection Agency, Washington, D.C.
3-16
-------
HYDROLYSIS
As a result of a significant amount of fundamental research that has been
applied to evaluations of hydrolysis of chemicals in the aquatic environment,
the hydrolysis of organic chemicals in solution is probably the best
understood environmental process.
Hydrolytic reactions involve the introduction of a hydroxyl group (OH-)
into a substrate molecule with loss of a leaving group (x) (Mabey and Mill
1976, Wolfe et al. 1976) (cf: Eq. 1).
R - X + H20-»R - OH + H - X ( 1 )
where R is usually an organic moiety, and X is a leaving group (e.g.
alcaholate, halide, etc.). Hydrolysis usually results in products which are
more water soluble than the parent compound; whether or not the hazard
presented by the parent compound is moderated by hydrolysis is, of course,
dependent on the properties, reactivity and toxicology of the products
themselves. To assess the reactivity of compounds under hydrolytic conditions
similar to those of Equation 1, the environmental factors of pH and
temperature must be addressed (Wolfe et al. 1976) as well as the unique
characteristics of the compound itself.
The rates of hydrolysis in streams will depend heavily on the pH of the
stream. The kinetics of hydrolysis have been described by Mabey and Mill
(1976), where the rate of hydrolysis (Rhd) mav ^e expressed as:
-dS j.
- - R , • k [S] - kB IDS'] [S] + kA (S) (H+)-* (HO) (S) (2)
dt hyd obs ° A N 2
where (S) - concentration of hydrolyzable pollutant, kg and kA are the
rate constants of the base [OH~] and (H+) acid promoted hydrolyses
respectively, and k^ Is the rate constant for the reaction of the chemical
with water. The half-life of the pollutant is calculated from:
(3)
4-1
-------
Since in aquatic systems the hydrolysis of an organic chemical will not
affect the concentration of water (55.5 M), the term kN [l^O] is
essentially constant and can be represented as a first order constant, kN.
From the autoprotolysis equilibrium expression, the concentration of hydroxyl
ion [OH~], is:
[OH~] - ^/(H) (A)
Equation 2 can then be rewritten:
kobs - kA IHJ + kN + kB
—
l« J
Thus, the hydrolysis rate constant of a chemical is dependent on the relative
values of k^, kg and k^, and on the hydrogen ion concentration [H*].
The rate of substitution of the hydroxyl group (OH-) for the leaving
group (-X) on a molecule will also vary over several orders of magnitude (pH
and temperature held constant) for different classes of compounds (Mabey and
Mill 1976) as well as for compounds in the same class but with different
peripheral structures. Wolfe et al. (1978) found this to be the case with
carbamate 'pesticides. The hydrolytic half-life varied dramatically with
structure.
Three types of hydrolytic reactions have been outlined by Mabey and Mill
(1976) as important when considering the hydrolysis of different classes of
compounds. The relative importance of the three processes will vary for
different classes of compounds. For instance, the kN term is often
negligible for hydroysis of some esters, and hydrolysis rates may then be
strongly pH dependent with a minimum hydrolysis rate at around pH 5 (Mabey and
Mill 1976). Other esters and carbamates (Wolfe et al. 1978) show a pH
independent behavior at moderate pH values due to a dominant k^ process, but
are acid or base catalyzed at pH values <4 or >8, respectively. Simple
non-halogenated alkanes, however, appear to have pH independent hydrolysis
rates up to pH MO, which is usually beyond environmental relevance.
Hence, knowledge of the hydrolytic capabilities of an aquatic ecosystem
requires a knowledge of the characteristics of the compound being hydrolyzed,
the pH of the water body and the temperature. Organic priority pollutants for
which hydrolysis is a significant fate process are listed in Table 4-1.
METHODS
The two environmental factors that unequivocally affect the rate of
hydrolysis of organic chemicals are pH and temperature. Although some authors
have suggested that metal ions or humic acids may catalyze hydrolysis
reactions, sufficient evidence for the environmental relevance of these
4-2
-------
TABLE 4-1. PRIORITY POLLUTANTS FOR WHICH HYDROLYSIS IS ESTIMATED TO BE
AN IMPORTANT FATE PROCESS
Pesticides
Endosulfan and Endosulfan sulfate
Heptachlor
Halogenated Aliphatics
Bromomethane
Bis (chloromethyl) ether
Bis (2-chloroethoxy) methane
Source: Callahan et al. 1979.
processes is not available, and therefore will be omitted at this time. These
factors, pH and temperature, may need to be quantified in the field if the
compound being investigated undergoes hydrolysis in natural aquatic systems.
The pH of a solution is its hydrogen ion activity expressed as the
logarithm of the reciprocal of the hydrogen ion activity in moles per liter
at a given temperature. The pH of a solution should be measured
electrometrically since colorimetric methods suffer from interferences due
to color, turbidity, salinity, colloidal matter and various oxidants and
reductants (U.S. EPA 1979). To obtain an average compartment pH value,
when more than one pH value is measured per compartment, an investigator
may convert the pH values to hydrogen ion (H+) activity. Next, he would
average the (H+) activity and reconvert the average H+ value to an average
compartment value.
It is also important that pH be monitored over a 24-48 hr period to
confirm stability. If pH proves to be unstable then it would be necessary
to weight the pH value by the hydrolysis rate to reflect the shifting
hydrolysis rates in the average pH value. This could be done by averaging
the pseudo-first-order rate constants generated from model simulations
based on continual sampling of pH values over a diurnal period, and adjusting
the pH input value such that the model output would equal this average
rate constant.
pOH
The pOH of. a solution can be determined by subtracting the pH value
from 14. Mean pOH values should be determined in an anlogous manner as pH
values described above.
Temperature
Methods to describe temperature are outlined in the Biotransformation
section. In addition, temperature changes over time may have to be considered
as were pH changes.
4-3
-------
REFERENCES
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler, J. R.
Freed, P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri, W. R.
Mabey, B. R. Holt, and C. Gould. 1979. Water-Related Environmental Fate
of 129 Priority Pollutants. Vol. I and II. EPA-440/4-79-029a and
EPA-440/4-79-029b. U.S. Environmental Protection Agency, Washington,
D. C.
Mabey, W. R. and T. Mill. 1976. Kinetics of Hydrolysis and Oxidation of
Organic Pollutants in the Aquatic Environment. In: Symposium on
Nonbiological Transport and Transformation of Pollutants on Land and
Water: Processes and Critical data required for predictive description.
L. H. Gevantman, ed. National Bureau of Standards, U.S. Dept. of
Commerce, Washington, D. C.
U.S. EPA. 1979. Methods for Chemical Analysis of Water and Wastes.
EPA-600/4-79-020. U.S. Environmental Protection Agency, Cincinnatti, Ohio.
Wolfe, N. L., R. G. Zepp, G. L. Baughman, R. C. Fincher and J. A. Gordon.
1976. Chemical and Photochemical Transformation of Selected Pesticides
in Aquatic Systems. U.S. EPA-600/3-76-067, U.S. Environmental Protection
Agency, Athens, GA.
Wolfe, N. L., R. G. Zepp and D. F. Paris. 1978. Use of Structure -
Reactivity Relationships to Estimate Hydrolyic Persistence of Carbamate
Pesticides. Water Res. 12:563.
4-4
-------
PHOTOLYSIS
Photolysis is the process whereby light energy is used to alter organic
and inorganic chemicals. This process is influenced by several factors. Two
important factors are the spectral qualities and intensity of the light. The
spectral characteristics indicate the range of energies available in the photons.
The intensity measures the number of photons available at a given energy level
(wavelength).
When the photon interacts with the molecule, the molecule may absorb the
energy available. If the wavelength is greater than i450 nm the molecule will
not generally have sufficient energy to cause bond rupture (Stern and Walker
1978). Wavelengths in the 300-450 nm range will (if absorbed) provide
sufficient energy for the photolytic process (^ 95 K cal/mole) (Sundstrom and
Ruzo 1978). In water, the shorter wavelengths <300 nm are not available because
of absorbtlon by atmospheric ozone. Once excited by a photon, a molecule may
lose the energy through fluorescence or phosphorescence or it may undergo
oxidation, reduction, displacement, isomerization, elimination or substitution
(Crosby 1976). The molecule may also transfer its energy to another molecule.
Several other factors may influence the photolytic process in aqueous
media. These are pH, temperature, the concentration of the chemical being
studied, other chemicals available to interact with the chemical being studied,
and turbidity of the water.
Temperature has always been a controlling factor in chemical reactions
and it may be important in photochemical reactions. Turbidity has both been
called a limiting factor (Herbes et al. 1979) and an enhancing factor depending
on the study reported (Miller and Zepp 1979). One researcher showed that it
reduced the rate (Herbes et al. 1979) and another showed that although the
longer wavelengths were absorbed by the water, the scattering effect on the
shorter wavelengths «400 nm) actually caused a greater number of photons to
be available (Miller and Zepp 1979).
Many environmental measurements can be made that give an indication of
the wavelengths and intensity of light available for photolysis. The latitude,
longitude, altitude, cloud cover, ozone concentration, time of year (season),
and time of day affect the spectrum, intensity and angle at which the light
strikes the water (Baughman and Lassiter 1978, Stern and Walker 1978, and Zepp
and Cline 1977). This affects the shorter wavelengths more than the longer :
ones since the shorter wavelengths are affected by ozone, atmospheric absorp-
tion, and scattering (Sundstrom and Ruzo 1978). The angle at which the light
hits the water also influences the amount of light reflected by the surface.
This in turn affects the amount of light available in the water to promote
5-1
-------
photolysis (Baughman and Lassiter 1978, Stern and Walker 1978, and Zepp and
Cline 1977).
The surface water conditions (rough versus smooth) as well as suspended
particulates, biota, pigments, DOC and other turbidity factors determine light
penetration and what wavelengths are available to act on the chemical. Eddy
diffusivity also determines how much of the dissolved substrate is present in
the photic ozone. If the chemical is not brought to the surface it may never
be exposed to the photolytic process. Water pH and the chemical's condition
(sorbed to soil particle, dissolved, ionized, etc.) also affect the rate of
photolysis.
If one is unable to measure (throughout the day) the actual spectral
characteristics of the site with a submersible spectral radiometer, then it is
essential that not only the above conditions be noted, but that an estimate of
cloud cover and any other factors (e.g., tree cover) that would reduce the
total light available be obtained so that an estimate of available light
energy can be made from data for clear day sunlight data. Most of these para-
meters (date, latitude, longitude, time of day, altitude, temperature, percent
cloudiness, general ozone concentrations) are available for most areas of the
country from NOAA, USGS, and EPA (air monitoring group).
Because spectral characteristics change throughout the year, they must be
measured at the time of evaluation. The spectral radiometer will measure how
much energy is available at each wavelength at the surface of the water, or at
any depth of interest, light flux may also be evaluated using actinometers.
The chemical characteristics that are used to indicate susceptibility to
change (pH, chemical concentration, etc.) can be measured in a laboratory
using samples taken in the field. Table 5-1 lists major environmental factors
which influence photolysis while Table 5-2 lists priority organic pollutants
for which photolysis is estimated to be an important fate process. For more
information on photolysis the following references may be of use: Mabey et
al. 1982, Tinsley 1979, Smith et al. 1977, Haque 1980, Stumm and Morgan 1981,
and Mill et al. 1980.
METHODS
Latitude (degrees)
Using a USGS 7-1/2' or 15' map of the area in which you are working, list
the latitude of your site.
Longitude (degrees)
Using the same maps as used for latitude and altitude, find and record
the longitude.
5-2
-------
TABLE 5-1. SUMMARY TABLE OF ENVIRONMENTAL FACTORS THAT
INFLUENCE PHOTOLYSIS
1. Spectral Characteristics and Intensity
a) Determined by
1) Latitude
2) Longitude
3) Altitude
A) Angle of Sun
a) Time of Day
b) Time of Year
5) Cloud Cover
2. Light Intensity as Measured by Spectral Radiometer/Actinometer
3. Absorption Characteristics of Water and Solutes (dissolved organic carbon)
4. Turbidity
a) Suspended sediment
b) Microflora and Fauna (partially indicated by chlorophyll
levels)
5. Eddy diffusivity
6. Depth
7. pH
Altitude (m)
Using the USGS map (7-1/21 or 15') of your site, find the elevation above
(or below) sea level. Multiply that number (in feet) by 0.305 to get meters,
and report the data in meters.
Time of Day (2A-hour time)
List time of day using military time (24-hour clock).
Cloud Cover (tenths)
Use of a pyranograph to quantify the available solar radiation is suggested
when photolysis is a significant fate process. Pyranographs measure direct
and incident sunlight in Cal/cm^. Relative radiation on a daily basis can
be estimated from comparison of daily pyranograph curves. Estimates of total
dally radiation in Cal/cm^ can be determined by cutting out and weighing the
area under the daily pyranograph curves. To convert from Cal/cm^ to percent
cloud cover one must determine total radiation on a cloudless day (maximum
value on cloudless day - 0% cloud cover) and on a day with extremely heavy
cloud cover (minimum value or 100% cloud cover). The daily baseline values
5-3
-------
TABLE 5-2. PRIORITY ORGANIC POLLUTANTS FOR WHICH PHOTOLYSIS IS CONSIDERED
AN IMPORTANT FATE PROCESS
Pesticides
DDE
Dieldrin
Isophorone
Monocyclic Aromatics
Nitrobenzene
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Phenol
Pentachlorophenol
2-Nitrophenol
4-Nitrophenol
2,4-Dinitrophenol
2,4-Dimethylphenol (2,A-Xylenol)
p-Chloro-m-cresol
4,6-Dinitro-o-cresol
Nitrosamines
Dimethylnitrosaoine
Diphenylnitrosamine
Di-n-Propylnitrosamine
Polycyclic Aromatic Hydrocarbons
Acenaphthene
Acenaphthylene
Anthracene
Benzo [a] Anthracene
Benzo [b] Fluoranthene
Benzo [k] Fluoranthene
Benzo [g,h,i] Phylene
Benzo [a] Pyrene
Chrysene
Dibenzo [a,h] Anthracene
Fluoranthene
Fluorene
Indeno [1,2,3-cd] Pyrene
Naphthalene
Phenanthrene
Pyrene
Source: Callahan et al. 1979.
can be established by subtracting the minimum value from all other observations.
By proportioning the daily baseline values to the maximum value, daily percent
cloud cover can be determined.
Additional research is needed to establish representative minimum and
maximum values. Research is also needed to determine if a linear function
is appropriate between extremes. Even with the above constraints, this method
is felt to be superior to the subjective estimates based on visual observations
of an untrained observer. If a NOAA weather station is near the study site,
and weather patterns for the site and the NOAA facility are similar, NOAA
percent cloud cover estimates could be used.
Light Intensity Measurements
Light intensity measurements may be made in several ways, and will depend
on equipment or analytical instruments available as well as the time period of
5-4
-------
interest for the validation effort. For short term validations (few hours) a
spectral radiometer or an actinometer system is suitable; for efforts of
several days, actinometer systems are more practical. The use of actinometers
is often preferable because the measurements take into account sunlight varia-
tions over a period of time (hours and days), while the radiometer method
provides instantaneous light intensity readings at specific wavelengths.
These methods are presently being further evaluated, and their relative capa-
bilities are illustrated in Table 5-3.
TABLE 5-3. COMPARISON OF DIRECT MEASUREMENT TECHNIQUES AND
SPECTRAL RADIOMETER METHODS
Spectral Radiometer Method
Direct Measurement of Kj
Moderately difficult, requires
equipment (>. $10K), calibration,
expertise, power source at
site, and probably two persons
Applicable to all compounds if
uv spectra data available
Practical limits on depth due
to probe length in some units,
other units are submersible
Obtains K values at discrete X
Use of method more difficult
over long time periods
(time used to collect data and
calculate K)
Fundamentally sound method
Easy to do; requires chemicals
and standard analytical
equipment
Strictly applicable to chemicals
of similar spectra
Usable for all depths
Obtains integrated K value over
X and light Intensity variations
No problem with time period if
chemical half-life is relevant
to study8
Experimental development needed
to assess limits
aUse of a meter gives Instantaneous light reading, which must then be
integrated over time, but not wavelength.
Source: Mabey (Personnel communication)
5-5
-------
Spectral Radiometer Measurements—
Use a spectral radiometer to measure light intensity at specific wave-
lengths, X, at which absorption coefficients of chemical and natural water are
measured; wavelengths described by Zepp and Cllne (1977) or Mabey et al.
(1982) are suitable. The amount of data obtained in the field measurement
effort Is in part determined by the intended use of the data and the model
being validated.
EXAMS, for example, will accept the spectral radiometer readings taken
when the sun is at its highest point or will accept a rate constant developed
in the field or from its SOLAR program. At the present time, so many new
electronic devices for measuring solar energy are available or in develop-
ment, that anyone preparing to validate a model using a photolytically active
chemical should carefully investigate the new products available.
Actionometer Measurements—
An actinometer should be a photo-reactive chemical whose chemistry is
well understood. Rate constants as a function of seasons, latitude, and time
of day should be known. The present state-of-knowledge allows such values to
be estimated from absorption coefficients and reaction quantum yields for an
actinometer at clear, cloudless sky conditions.
For its simplest use in a field study, the actinometer is exposed to
sunlight under field conditions. The actinometer should be used in an area
free of shadows, and with a non-reflecting background. If possible the photo-
lysis rate of the actinometer should be chosen to approximate that of the
chemical under study. The first-order rate constant for photolysis of actino-
meter [A] is the calculated from a plot of In A vs T, and is designated k a
(measured). The ratio of k_a (measured) to the rate constant estimated for
photolysis of actinometer under clear sky conditions then gives a correction
factor that can be used to adjust any photolysis (or possibly abiotic oxidation
rate constant) from clear-sky conditions to field sunlight conditions. One
actinometer solution suitable for field use has been described by Mill et al.
(1981), and its use is presently under final development.
Dissolved Organic Carbon (mg/1)
In order to measure for dissolved organic carbon (DOC) in a water sample,
it Is first necessary to filter the sample using glass-fiber filter discs, to
remove nonfilterable residue. Methods and materials for this filtering
procedure are outlined in APHA (1980). The filtered sample is then analyzed
using an infrared analyzer for total organic carbon (APHA 1980).
Water samples for DOC analyses should be collected and stored in glass
bottles, preferably brown, and sealed with a teflon-lined cap. Samples should
be analyzed promptly, but if delay is unavoidable, store samples at ice
temperature with minimal exposure to light and atmosphere and acidify with
hydrochloric acid to a pH not over 2 (APHA 1980).
5-6
-------
Suspended Sediment (mg/1)
See methods section in Physical Transport.
Chlorophyll (mg/1)
The attenuation of irradiance in natural waters is a result from
absorption of light by the water itself, green plants, dissolved organic
matter, and suspended sediments. EXAMS relates absorption due to green plants
to chlorophyll £, chlorophyll ]>, chlorophyll £, and phaeophytin £. Methods to
determine the above pigments can be found in APHA (1980), and Weber (1973).
Eddy Diffusivity (m2/hr)
See methods section in Physical Transport.
Depth (meters)
See methods section of Physical Transport.
pH (in pH units)
See methods section in Hydrolysis.
5-7
-------
REFERENCES
APHA. 1980. Standard Methods for the Examination of Water and Wastewater.
Fiftenth Edition. APHA/AWWA/WPCF. Washington, D.C.
Baughman, G. L. and R. R. Lassiter. 1978. Prediction of Environmental :
Pollutant Concentration. In: Estimating the Hazard of Chemical
Substances to Aquatic Life, ASTM STP 657. John Cairns Jr., K. L. Dickson
and A. W. Maki, eds. American Society for Testing and Materials,
Philadelphia, Pennsylvania.
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler, J. R. Freed,
P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri, W. R. Mabey, B. R.
Holt, and C. Gould. 1979. Water-Related Environmental Fate of 129 Priority
Pollutants. Vol. I and II. EPA-440/4-79-029a and EPA-440/4-79-029b.
U.S. Environmental Protection Agency, Washington, D.C.
Crosby, D. G. 1976. The Photochemistry of Xenobiotlcs. In: Symposium on
Non-biological Transport and Transformation of Pollutants on Land and
Water: Processes and Critical data required to predictive description,
L. H. Gevantman, ed. National Bureau of Standards, U.S. Department of
Commerce, Washington, D.C.
Haque, R. 1980. Dynamics, Exposure and Hazard Assessment of Toxic Chemicals.
Ann Arbor Science Publishers, Inc. Ann Arbor, Michigan.
Herbes, S. £., G. G. Southworth, D. L. Shaefferm, W. H. Griest and M. P.
Maskariwec. 1979. Critical Pathways of Polycyclic Aromatic Hydrocarbons
in Aquatic Environments. Oak Ridge National Laboratory, Oak Ridge,
Tennessee.
Mabey, W. R., J. H. Smith, R. T. Podoll, H. L. Johnson, T. Mill, T. W. Chou,
J. Gates, J. Waight Partridge and D. Vandenberg. 1982. Aquatic Fate
Process Data for Organic Priority Pollutants. EPA 400/4-81-014 U.S.
Environmental Protection Agency, Washington, D.C.
Mill, T., D. Dulin, and J. Davenport. 1981. Development and Application of
Chemical Actinometer for Solar Gradients. American Chemical Society
preprint. Paper No. 82. 181st National ACS Meeting at Atlanta, Georgia.
Mill, T., D. G., Hendry, and H. Richardson. 1980. Free-radical oxidants in
natural waters. Science 207:886-7.
Miller, G. C., amd R. G. Zepp. 1979. Effects of Suspended Sediments on
Photolysis Rates of Dissolved Pollutants. Water Res. 13:453-459.
Smith, J. H., W. R. Mabey, N. Bohonos, B. R. Holt, S. S. Lee, T. W. Chou,
D. C. Bomberger, and T. Mill. 1977. Environmental Pathways of Selected
Chemicals in Freshwater Systems Part I: Background and Experimental
procedures. EPA 600/7-77-113. U.S. Environmental Protection Agency,
Athens, Georgia.
5-8
-------
Stern, A. M. and C. R. Walker. 1978. Hazard Assessment of Toxic Substances:
Environmental Fate Testing of Organic Chemicals and Ecological Effects
Testing. In: Estimating the Hazard of Chemical Substances to Aquatic
Life, ASTM. STP 657, John Cairns Jr., K. L. Dickson, and A. W. Maki, eds.
Am. Soc. for Testing and Materials. Philadelphia, Pennsylvania.
Stumm, W., and J. J. Morgan. 1981. Aquatic Chemistry. An Introduction
Emphasizing Chemical Equilibria in Natural Waters. John Wiley and Sons,
Inc. New York, New York.
Sundstrom, G. and L. 0. Ruzo. 1978. Photochemical Transformation of
Pollutants in Water. In: Aquatic Pollutants: Transformation and
Biological Effects, 0. Huntzinger, J. H. Van Lelyveld and B. C. J.
Zoeteman, eds. Pergamon Press, New York, New York.
Tinsley, I. J. 1979. Chemical Concepts in Pollutant Behavior. John Wiley
and Sons, Inc. New York, New York.
Weber, C. I. (Ed.). 1973. Biological Field and Laboratory Methods for
Measuring the Quality of Surface Waters and Effluents. Environmental
Monitoring Series. EPA-670/4-73-001. U.S. Environmental Protection
Agency. Cincinnatti, Ohio.
Zepp, R. G. and D. M. Cline. 1977. Rate of Direct Photolysis in Aquatic
Environments. Env. Sci. and Technol. 11(4):359-366.
5-9
-------
OXIDATION
Three different mechanisms can be identified for the oxidation of organic
contaminants in water:
1. Photooxidatlon
2. Chemical oxidation
3. Microbial oxidation
Actually, oxidation is the common end point of all the above processes.
They differ only in the initiation mechanisms. Both photooxidation and micro-
bial oxidation are discussed in other sections and will only be mentioned
here with respect to their relationship to oxidation.
Photooxidation is an oxidation process which begins with the absorption
of a photon of light energy by a molecule, raising it to an excited electronic
state. It is this excited electronic state which subsequently gives rise
either directly or indirectly to the oxidized species. Photooxidation is the
major source of oxidant production.
Oxidation processes are dependent on several factors. Some of these
factors appear to be the rate that oxygen dissolves, the concentration of
oxygen, and the homogeneity of the dissolved oxygen in the water. These factors
depend variously on meteorological conditions and surface and subsurface
characteristics of the water body under investigation. Factors such as wind
speed and surface turbulence play an important role in the dissolution rate of
oxygen (Rathbun 1977; Holley 1978).
The temperature of the river will also have an effect on the rate of
oxidation of an organic contaminant. This is true not only of chemical oxida-
tion but also when metabolic function associated with microblal oxidation is
considered. Microbial oxidation is Initiated by enzymatic or metabolic
processes which essentially predispose the pollutant molecule to oxidation.
As the temperature of the water decreases, microblal activity decreases until
it eventually becomes negligible. Therefore, season of the year as well as
geographical location must be considered.
The production of oxidants by means of photooxidation is one mode of
oxidation of organic contaminants. Photo production of oxidants such as singlet
oxygen, alkyl peroxyl, or hydroxyl radicals are dependent upon such factors as
the intensity and distribution of Incident radiation, humic substances, and
water temperature. Organic priority pollutants for which oxidation is estimated
to be an important fate process are listed in Table 6-1. More specific infor-
mation on oxidation is available in Mabey et al. 1982, Tinsley 1979, Smith et
al. 1977, Haque 1980, and Stumm and Morgan 1981, Mill et al. 1980.
6-1
-------
TABLE 6-1. ORGANIC PRIORITY POLLUTANTS FOR WHICH OXIDATION IS EXPECTED
TO BE AN IMPORTANT FATE PROCESS
Monocyclic Aromatics
Phenol
Nitrosamines
Benzidine
3,3' -Dichlorobenzidine
Source: Callahan et al. 1979.
METHODS
Dissolved Organic Carbon
Methods to measure dissolved organic carbon are described in the
Photolysis section.
Free Radical Oxidants
The determination of the average steady-state free radical oxldant
concentration in the natural water would be conducted in a manner similar to
the singlet oxygen, except that cumene would be used as the probe chemical
instead of DMF. Cumene is the suggested probe since reaction with free
radical oxidants is the most important abiotic transformation process, and has
been used as the probe chemical for such oxidations in aquatic systems (Mill
et al. 1980).
Oxidant Concentration
Standard methods for the determination of oxidant concentration in
aqueous media have not been developed. The state-of-knowledge for oxidation
process leaves no alternative except to employ research methods which directly
estimate free radical oxidants or singlet oxygen present in natural waters.
These measurements of oxidant concentration need only be performed when the
rate of oxidation becomes a significant factor in chemical breakdown.
Reaeration Rate.
Methods to measure reaeration rates are described in the Volatilization
section.
6-2
-------
Singlet Oxygen
Singlet oxygen concentrations in natural waters can be determined by
reactions of a chemical known to only undergo reaction with singlet oxygen;
Zepp et al. (1978) have used dimethyl furan (DMF), as a probe chemical. In
the proposed method, DMF would be dissolved in a sterilized solution of
natural water, and photolyzed in sunlight during the field study. The
solution would be monitored for loss of DMF, and the concentration versus time
data would be used to determine the first order rate constant for DMF loss.
This rate constant would be divided by the known rate constant value for
reaction of DMF with singlet oxygen. This ratio is then the average steady-
state concentration of singlet oxygen for the duration of the experiment,
and could be used in the computer model.
In addition, Zepp et al. (1981) suggest an approach for correlating
singlet oxygen reaction rate constants with total organic carbon
concentrations in natural waters.
Suspended Particulate
Methods to measure suspended particulates are found in the Physical
Transport section.
Temperature and Dissolved Oxygen
Methods for temperature and dissolved oxygen are described In the
Biotransformation section.
6-3
-------
REFERENCES
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler, J. R.
Freed, P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri, W. R.
Mabey, B. R. Holt, and C. Gould. 1979. Water-Related Environmental Fate
of 129 Priority Pollutants. Vol. I and II. EPA-440/4-79-029a and
EPA-440/4-79-029b. U.S. Environmental Protection Agency, Washington,
D. C.
Haque, R., ed. 1980. Dynamics, Exposure and Hazard Assessment of Toxic
Chemicals. Ann Arbor Science Publishers Inc. Ann Arbor, Michigan.
Holley, E. R. 1978. Oxygen Transfer at the Air Water Interface. In:
Transport Processes in Lakes and Oceans. R. J. Gibbs, ed. Plenum Press.
Mabey, W. R., J. H. Smith, R. T. Podoll, H. L. Johnson, T. Mill, T. W. Chou,
J. Gates, J. Waight Partridge and D. Vandenberg. 1982. Aquatic Fate
Process Data for Organic Priority Pollutants. EPA 440/4-81-014. U.S.
Environmental Protection Agency, Washington, D. C.
Mill, T., D. G. Hendry, and H. Richardson. 1980. Free-radical oxidants in
natural waters. Science 207:886-7.
Rathbun, R. E. 1977. Reaeration Coefficient of Streams - State of the Art.
J. Hydraulics Divn. ASCE 103, No. HY 4 409.
Smith, J. H., W. R. Mabey, N. Bohonos, B. R. Holt, S. S. Lee, T. W. Chou,
D. C. Bomberger, and T. Mill. 1977. Environmental Pathways of Selected
Chemicals in Freshwater Systems Part I: Background and Experimental
Procedures. EPA-600/7-77-113. U.S. Environmental Protection Agency,
Athens, Georgia.
Stumm, W. and J. J. Morgan. 1981. Aquatic Chemistry. An Introduction
Emphasizing Chemical Equilibria in Natural Waters. John Wiley and Sons,
Inc. New York, New York.
Tinsley, I. J. 1979. Chemical Concepts in Pollutant Behavior. John Wiley
and Sons, Inc. New York, New York.
Zepp, R. G., G. L. Baughman and P. F. Schlotzhauer. 1981. Comparison of
Photochemical Behavior of Various Humic Substances in Water. II.
Photosentized Oxygenations. Chemosphere. 10:119-126.
Zepp, R. G., N. L. Wolfe, G. L. Baughman, and R. C. Hollis. 1978. Singlet
Oxygen in Natural Water. Nature 267:421-423.
6-4
-------
IONIZATION
The type of ion population characteristic of a lotic water body may
directly influence the state of an added chemical contaminant. The extent to
which the monovalent (C+,A~) and divalent (CT^A™) ion population may
ionize neighboring atoms and molecules (the ionization potential, H) depends
on the amount of energy required to displace an electron on a neighboring atom
or molecule (Morrison and Boyd 1973) and may be expressed by the following
equation la.
A $ A+ + E- - H (la)
Similarly, dissociation may be expressed by equation Ib.
A - B f A+ + B- (Ib)
The equilibrium between the ionized and un-ionized forms of a chemical
species may be represented by the ionization constant (K^), where (referring
to equation la):
Increases in entropy resulting from temperature increases, increases or
decreases in pH, substitution of divalent for monovalent ions and vice versa
(increase or decrease in electrostatic forces) or an addition of high-
molecular-weight compounds, may influence the ionization potential and
hence the equilibrium (Stumm and Morgan 1970).
However, not all molecules are affected similarly with respect to the
effect of temperature on ionization. The influence of temperature upon
Ionization of water, carbon dioxide, and acetic acid reported by Stumm and
Morgan (1970) showed large Increases, slow Increases, and a Maxwell-Boltzmann
distribution of increase, respectively.
The pH of a solution may also characterize the ionization potential of a
solution. For every acid or base there exists an acid dissociation (Ka)
or base dissociation (K^) constant determined from the following equilibria:
7-1
-------
HA + H20ir H30+ + A~ (3)
K -
a (HA)
(Note: In equation 4, (H20) is incorporated in Ka).
B -f H20 5. HB+ + OH~ (5)
(Note: In equation 6, (H20) is incorporated in
Since Ka and K^ (equations 4 and 6) represent the ratio of ionized to
un-ionized species, the larger Kg (or K^,), the greater the ionization
potential of the system (Morrison and Boyd 1973). Temperature increases or
decreases will affect this relationship according to the resultant entropy
change.
In addition to the influence of pH and temperature on ionization, the
total dissolved solids will give an estimate of both the un-ionized and
ionized species and their concentration in the stream (Reid 1961). Specific
conductance (conductivity) is commonly used to estimate the total dissolved
solids (Reid 1961).
Modern procedures may include a species specific analysis. Some of the
more common cations (calcium, magnesium, sodium, and potassium) and anions
(sulfate, chloride, fluoride, nitrate, carbonate, and bicarbonate) found in
streams characterize the stream's ion population. Other common ion species
may include arsenic, barium, boron, bromine, chromium, copper, iodine, iron,
lead, manganese, phosphate, selenium, silicon, strontium, and zinc (Hem
1970).
Water also self-protonates according to the following:
H20 + H20 * H30+ + OH~ (7)
This must be considered (especially in dilute solutions) as a process
which may alter the dissociation of any acid or base with a pKa comparable
to that of water (Stumm and Morgan 1970). Equilibria of ionized and un-ionized
7-2
-------
species may also depend on the particular characteristics of the ionized form
of a chemical. A selectivity coefficient Q of the form:
Xv R (NA+)
Q (Na R * KR) - -£ (8)
2 XNaR(K+)
or for second order reactions:
Xr_ R (Na"1")2
Q(NaR+CaR) - -_E2 __ (9)
2 X2Na R (Ca ++)
can tell us whether sodium tends to bind with either calcium or potassium.
These coefficients represent the equilibria for an Interchange of ions but may
not be constant (Stumm and Morgan 1970).
A compound may not reside in the water column but may precipitate into
the sediment. If this happens, the cation anion composition may further deter-
mine the fate of the compound (Stumm and Morgan 1970). The substitution of
less hydrated Ca"1"1" or Mg** by strongly hydrated Na+ will result in swelling
of the sediment and a reduced permeability (Stumm and Morgan 1970). As a
result, transport through the Interstitial waters will depend upon the hydra-
tion tendency of Ionic species in the sediment, the elasticity of interlayer
forces (which allow different interlayer spacing to develop), and factors
influencing osmotic control, electrostatic force, and van der Waals forces.
The process of ionlzation, however, is not limited to specific ion Interaction
or selfprotolysls, but augments its potential as an environmental process
through the formation of colloidal suspensions.
Two classes of colloids exist: those which are hydrophobic and those
which are hydrophllic (Stumm and Morgan 1970). The metal sols, such as silver
halldes and nonhydrated metal oxides are typical hydrophobic sols, whereas
gelatin, starch, proteins, hydrated compounds and biocollolds (viruses-bacteria)
are hydrophilic (Stumm and Morgan 1970). The ion exchange structure of these
colloidal suspensions depends on interlayer attractions and lattice forces.
The elasticity of attraction increases with organic content (Stumm and Morgan
1970). Addition of an ion with unmatched charge to a parent colloid may result
in alterations of the parent colloid's surface chemical behavior. The flucta-
tlon of ionized species is primarily related to the charge and concentration
of the ions in solution. The stability of the colloid depends exponentially
on the relative magnitude of its energy barrier and can be represented by the
equation:
_J
W - ±— e Vmax/kT (10)
2ae
7-3
-------
where: W is an indicator of stability; k~Va is the ratio of the double
layer thickness (the distance characterizing the repulsive interaction) to
particle radius, k (the Boltzman constant); and T (temperature) represents the
energy of attraction (Stumm and Morgan 1970).
If Vmax (the energy barrier) exceeds a value less than 15 kT, relatively
stable colloids will be found. For example if Vmax _> 15 kT only 1 out of 106
collisions will be successful (Stumm and Morgan 1970).
Colloidal destabilization can be accomplished by reducing k~* or increasing
the concentration of electrolytes (unless aggregate size increases). This
physical theory (a combination of van der Waals attractional forces and the
repulsion of the Gouy-Chapman double-layers) has provided an analytical basis
for the valency rule of Shulze and Hardy (Stumm and Morgan 1970). This rule
simply states that the critical coagulation concentration (CCC) varies for
mono-, di-, and trivalent compounds proportional to the ratio of (6/z)6 (z
represents ion valence) or 100:1.6:0.13. The rule is not applicable to metal
oxide sols. Although the influence of aggregate size, and ion valency contri-
bution, on stability may be complex; as a rule increasing stability is observed
with increasing aggregate particle radius (Stumm & Morgan 1970).
Seasonal variation in runoff or the geochemical characteristics will
determine the relative concentration, composition and longitudinal dispersal
of the total dissolved solid present in the stream (Reid 1961). The physiography
related to arid or humid plains, mountains, and valleys may determine the
ionized and un-ionized species present. Reid (1961) characterizes those rivers
with their headwaters in semi-arid plains as high in sodium, sulfate, and
chloride but low in calcium and carbonate. Other examples Include: rivers
draining extensive plains in humid or temperate regions are characteristically
high in sulfate and carbonate concentrations, but valley streams briginating
in granitic mountain regions show increased total dissolved solids. Youthful
granitic mountainous region streams show high silica content and relatively
low total dissolved solids (Reid 1961).
In summary, the ionized and un-ionized species in streams will contribute
to the process of ionization if one or more of the environmental factors listed
in Table 7-1 is changed. Whether the pH, temperature, total dissolved solids,
or ionic strength vary may depend on geochemical characteristics inherent to
the stream or the ionization capability of an anthropogenic chemical.
TABLE 7-1. ENVIRONMENTAL FACTORS THAT INFLUENCE IONIZATION OF ORGANIC
COMPOUNDS IN WATER
Temperature
pH
Ionic Strength
Total Dissolved Solids
7-4
-------
METHODS
Temperature
Methods for temperature are described in the Biotransformation
section.
£H
Method for pH determinations are described in the Hydrolysis section.
Ionic Strength
No standard method exists for direct determination of ionic strength;
however, it may be indirectly determined via conductivity measurements (APHA
1980).
Total Dissolved Solids
Methods for total dissolved solids may be found in Standard Methods
(APHA 1980).
7-5
-------
REFERENCES
APHA. 1980. Standard Methods for the Examination of Water and Wastewater.
Fifteenth Edition. APHA/AWWA/WPCF, Washington, D.C.
Hem, J. D. 1970. Study and Interpretation of the Chemical Characteristics of
Natural Water. 2nd ed. USGS Water-Supply Paper 1473, U.S. Government
Printing Office, Washington, D.C.
Morrison, R. T. and R. W. Boyd. 1973. Organic Chemistry. 3rd ed. Allyn
and Bacon, Inc.
Reid, G. K. 1961. Ecology of Inland Waters and Estuaries. Van Nostrand
Reinhold Company, New York, New York.
Stumm, W. and J. J. Morgan. 1970. Aquatic Chemistry: An Introduction
Emphasizing Chemical Equilibria in Natural Waters. Wiley - Interscience,
New York, New York.
7-6
-------
VOLATILIZATION
Volatilization of a chemical that is dissolved in water is the transport
of the chemical from the water body to the atmosphere. The volatilization of
organic substancees from aquatic environments to the atmosphere can be a signi-
ficant environmental pathway, by which compounds are transported over wide
areas. Volatilization rates are dependent on the vapor pressure of the chemical
and the environmental factors that influence the movement to and from the
evaporative surface. The pressure exerted by equilibrium vapor is known as
the vapor pressure and is temperature dependent. Vapor pressures provide
relative indications of the tendency of solutes to vaporize under unperturbed
conditions (laboratory conditions). However, volatilization from a water
solution is Influenced by many environmental factors that tend to reduce the
effective vapor pressure from the vapor pressure of a solute in distilled
water under equilibrium conditions.
In moving from the bulk of the water to the atmosphere, a contaminant
experiences three diffusive (transport) processes, each with a resistance.
Normally, one of the three resistances will dominate. The resistance approach
has been utilized to quantify contaminant movement from the water to the air
(Mackay and Leinonen 1975, Mackay and Yuen 1980). The three diffusive
resistances are:
1. rj » Resistance to diffusion from the bulk of the water to the
interface,
2. *2 • Resistance to diffusion through the near-surface liquid to the
interface,
3. r3 - Resistance to diffusion through the atmosphere in the layer
close to the water surface.
The first resistance (r^) is to that process needed to replenish the
volatile compound in the top layer of water or to maintain an equal concentra-
tion of the pollutant throughout the water column. The process is normally
sufficiently active, so that no concentration gradient of the pollutant will
exist to retard the volatilization process. This resistance is thought to be
Important only in deep sluggish rivers or lakes and ponds. A knowledge of
vertical mixing (eddy diffusivity) as a function of depth and time is required
to understand the magnitude of this resistance in quiescent aquatic environments.
The second resistance (r£) is referred to as liquid-phase resistance and
represents the resistance to a molecule migrating from the near-surface water
to the air-water Interface. It is now generally accepted that turbulence in
the near-surface region in the water phase controls the volatilization rate
of highly volatile compounds in most rivers. Oxygen reaeration rates can be
8-1
-------
used as a measure of turbulence. The oxygen reaeratlon rate is the rate at
which oxygen in the atmosphere dissolves in water that has a concentration of
dissolved oxygen lower than the equilibrium concentration with the atmosphere
(Mill et al. 1982). Oxygen reaeration rates of rivers have been correlated
with river velocity, depth, and slope (Rathbun 1977, Mackay et al. 1982, Holley
1978). In general, turbulence increases with increasing velocity and slope
and decreases with increasing depth.
Wind can also generate turbulence in near-surface layers. This effect,
while generally minimal in lotic environments, is important in lentic environ-
ments. Cohen et al. (1978) demonstrated the importance of wind speed in
laboratory wind-wave tests. They suggest that the prediction of liquid-phase
resistance is best achieved by considering three wind-velocity ranges measured
at 10 cm above the water column. At velocities below 3 m/s the effect of wind
is negligible and liquid-phase resistance is strongly influenced by mixing
originating from within the water body. In the second range, 3 to 10 m/s,
liquid-phase resistance is basically controlled by wind-induced waves. Waves
also Increase the interfacial area, but since their height to length ratios do
not normally exceed 0.143, this cannot account for more than a 4 percent increase
in the transfer rate. Most environments have wind speeds in this region. In
the third region, above 10 m/s, wave breaking may commence. A further decrease
in liquid-phase resistance occurs in this region due to an increase in wave
surface area, water spray, bubble entrainment and the disintegration of wave
crests.
Solar radiation adsorption, direct heat transfer, and water evaporation
may affect r2> When surface waters are cooler than subsurface water layers,
density gradients will tend to promote mixing and decrease r£. Conversely,
warm surface waters increase r^ as they tend to stagnate and not mix with
subsurface water layers (Mackay 1978).
The third resistance (r3) is against diffusion through the surface layer
to the atmosphere. Henry's constant, H, (the equilibrium partition coefficient
expressed as vapor pressure divided by water solubility) and temperature
basically control the process (Mackay and Yuen 1980). This resistance is
relatively insensitive to temperature variations under normal environmental
conditions since aqueous solubilities and mass transfer coefficients do not
dramatically change. The magnitude of this resistance is inversely proportional
to H and, therefore, compound dependent. For ionic compounds H is zero and no
volatilization occurs (Mackay and Yuen 1980). Since aqueous solubility is one
of the components of H, any environmental factor, e.g., pH and temperature,
which can greatly influence solubility becomes important.
The total resistance rj is:
rT - ri + r2 + r3
Most fate and transport models use the two-resistance (T£ and T$) or
"two-film" concept to compute the export of chemicals across the air-water
interface. The concept was originally developed by Whitman (1923) and has been
successfully adapted to environmental problems by Liss (1973), Lin and Slater
8-2
-------
(1974), and Mackay (1978). Most often, the resistance to mass transfer lies
in these films, the other resistance (r^) is often not considered in fate
and transport models, e.g., chemical movement through a thermocline. Strong
diel shifts in transport resistances in each phase of the air/water interface
are possible. For example, volatilization of a nominally liquidphase resis-
tance dominated compound can be retarded under still atmospheric
conditions (Burns 1982).
Figure 8-1 is a logarithmic plot of vapor pressure versus aqueous solu-
bility, with a number of compounds of environmental concern noted. Since H is
vapor pressure divided by solubility, H is represented by solid diagonal lines.
By locating the properties of a given compound an estimate can be made of its
relative potential for volatilization. It is interesting to note that compounds
of very low volatility such as DDT and PCB's will volatilize appreciably because
of their low solubility. Hydrocarbons and halogenated hydrocarbons have a
high volatilization potential while the more soluble organics, such as phenols,
are less likely to volatilize (Mackay and Yuen 1980). The priority pollutants
for which volatilization is thought to be an important process are listed in
Table 8-1.
Henry's constants (H) are often determined in the laboratory using dis-
tilled water and the compound of concern. Predictions of volatilization based
on "distilled water" chemical data are often high when applied to natural
water where the pollutant Is in low concentrations because of sorption of the
compound to organic or mineral materials (Mackay 1978).
Table 8-2 lists the major environmental factors that Influence volatiliza-
tion rates. In addition, any process which tends to sequester, transform, or
degrade the parent compound such as lonization, sorption, photolysis, microbial
degradation/transformation will obviously retard the volatilization rate of
the parent compound.
Surface films impede transfer processes at the interface. This may be
due either to a "blocking" effect or to a reduction of near-surface turbulence.
When visible scums due to algal blooms or higher aquatic plants or slicks are
present, the rate of volatilization will be reduced. Also, volatilization
will be severely reduced or eliminated when a continuous ice cover is present
(MacKay and Yuen 1980).
For additional information, the reader is referred to "Volatilization of
Organic Pollutants from Water" by Mackay et al. (1982). The report elucidates
factors which control volatilization, and develops methods for calculating
rates. Also, it contains theoretical and experimental studies and a comprehen-
sive review of the equilibrium physical chemistry and thermodynamics of systems
involving hydrophobic organic solutes and water.
METHODS
The input parameters necessary for volatilization include pH and tempera-
ture (°K), compartment dimensions, area and volume, and mixing. Parameters
necessary as input to mixing are eddy diffusivity, the reaeration rate of the
8-3
-------
00
I
10-
10"
3
CO
£10
K
n.
-------
TABLE 8-1. ORGANIC PRIORITY POLLUTANTS FOR WHICH VOLATILIZATION IS ESTIMATED
TO BE AN IMPORTANT TRANSPORT PROCESS
Halogenated Aliphatics
Chloromethane (methyl chloride)
Dichloromethane (methylene
chloride)
Trichloromethane (chloroform)
Tetrachloromethane (carbon
tetrachloride)
Chloroethane (ethyl chloride)
1,1-Dichloroethane (ethylidine
chloride)
1,2-Dichloroethane (ethylene
dichloride)
1,1,1-Trichloroethane (methyl
chloroform
1,1,2-Trichloroethane
1,1,2,2,-Tetrachloroethane
Chloroethane (Vinyl chloride)
1,1-Dichloroethane (Vinylidine
chloride)
1,2-trans-Dichloroethane
Trichloroethane
1,3-Dichloropropane
Bromomethane
Dichlorodifluoromethane
Trichlorofluoromethane
Ethers
2-Chloroethyl vinyl ether
Monocyclic Aromatics
Benzene
Chlorobenzene
1,2-Dichlorobenzene (0-Dichlorobenzene)
1,3-Dichlorobenzene (m-Dichlorobenzene)
1,4-Dichlorobenzene (p-Dichlorobenzene)
1,2,4-Trichlorobenzene
Ethyl benzene
Toluene
Pesticides
DDD
DDE
DDT
Dieldrin
Source: Callahan et al. 1979.
TABLE 8-2. ENVIRONMENTAL FACTORS THAT INFLUENCE
VOLATILIZATION RATES
pH
Temperature
Mixing
Eddy Diffusivity
Reaeration Rate
Wind Speed
Slope
Water Velocity
8-5
-------
compound, wind speed near the water surface, slope of the river, and water
velocity.
The measurement of pH is described in the methods section of Hydrolysis.
Temperature
Procedures for measuring water temperature are described in the methods
section of Biotransformatlon.
Compartment Dimensions, Area and Volume
Procedures for determining compartment dimensions, area and volume are
described in the methods section of Physical Transport.
Mixing
Parameters necessary as input to mixing are eddy dlffuslvity, reaeration
rate, wind (speed), slope (of river surface) and water velocity.
1. Eddy Diffusivity. For determining eddy diffusivity refer to the
methods section of Physical Transport.
2. Reaeration Rate. The reaeration rate can be measured directly but
this is not often possible because of the specialized equipment and
time required. Rathbun and Grant (1978) have developed a technique
using ethylene or propane as indicators of volatilization and
Rhodamine WT dye as an Indicator of dispersion and dilution. This
method provides a direct measure of volatilization of compounds with
high H (Henry's constant) values. In addition, reaeration rates can
be estimated in the field by measuring the rate of oxygen release
from the water into a nitrogen-purged dome (Copeland and Duffer 1964,
Hall 1970).
Where direct measure is not possible, the reaeration rate for a
specific site can be computed based on a knowledge of the physical
characteristics of the site. McKay and Yuen (1980) presented
several equations for calculating the reaeration rate for oxygen
in streams:
Tsivoglou and Wallace (1972) K2 - 638VS
Parkhurst and Pomeroy (1972) K2 - 1.08 (1.0 + 0.17F2)(VS)0.375Z-1
Churchill et al. (1962) K2 - 0.00102V2'695Z~3*085 s~°'823
where S is the river slope (m/m), Z average depth (m), V velocity
8-6
-------
(m/s), and F the Froude Number V/(gZ)0»5 (g»gravitational force)
If no slope data are available the following equations can be used:
Isaacs and Gaudy (1968) K2 - 0.223VZ"1'5
Langbein and Durum (1967) K2 - 0.241VZ~1<33
Mackay and Yuen (1980) state that, for a given river it is likely
that these equations would give an estimate of K2 with an average
error between 10 and 50 percent. They suggest that one approach may
be to apply all the equations, discard the outlying results and take
an average of the remainder.
Reaeration may be primarily determined by local winds in lakes and
ponds. The effect of windspeed on reaeration rates can be separated
into two distinct zones as described by Banks (1975) and Banks and
Herrera (1977). At windspeeds (at 10 m height) greater than about
5.5 m/s the exchange constant Increases as the square of the wind
velocity. At windspeeds of less than about 5.5 m/s exchange constants
correlate with the square root of windspeed. Banks (1975) presents
the following oxygen exchange equations:
KL - 3.2 x 10~7 U2 (U > 5.5 m/s)
KL - 4.19 x 10"6 /U*(U < 5.5 m/s)
where
KL - oxygen exchange constant (m/s)
U - windspeed (m/s) at 10 m above the water surface.
Finally, Smith et al. (1977) have presented the following depth-
dependent ranges and "suggested values" for oxygen reaeration rates
for various water bodies:
Range
(hour"1) (hour"1)
Pond 0.0046-0.0096 0.008
River 0.0042-0.39 0.04
Lake 0.004 -0.013 0.01
3. Wind.. The Influence of wind upon mixing is greatest for lakes and
ponds and less Important for rivers. In many cases, weather station
data may suffice. Wind speeds recorded at a given height can be
converted to EXAMS input values (wind velocity at 10 cm above the
water surface) by the assumption of a logarithmic wind profile
8-7
-------
(Israelsen and Hansen 1962). The following equation describes the
relationship:
V!/V2 - log z!/Z0 / log
where :
Vj - Windspeed in m/s at height Zj in
V2 * Windspeed in m/s at height Z2 in
ZQ - Effective roughness height in mm. The roughness
height is generally on the order of one millimeter
over water bodies.
When weather stations are not sufficiently close to the water body of
interest, or if an investigator deems that greater accuracy is required,
direct field measurements may be taken as follows. Wind speed should
be measured in close proximity (shoreline) to the waterbody of interest
or above the water surface. Units should be measured in or converted
to m/sec. The anemometer chosen should be capable of measuring wind
speed below 3 m/sec, as this is an important value with respect to
liquid phase diffusion. For information on anemometer types, consult
U.S.G.S. (1977).
A. Slope. The slope (of water surface) of a particular stream section
can be determined using elevations from U.S.G.S. quad sheets. Units
should be converted and reported in meters change in elevation/meters
distance along stream.
5. Water Velocity. Procedures for measuring water velocity are described
in the methods section of Physical Transport.
8-8
-------
REFERENCES
Banks, R. B. 1975. Some Features on Wind Action on Shallow Lakes. J.
Environ. Eng. Div., Proc. ASCE 101 (EE5):813~827.
Banks, R. B. and F. F. Herrera. 1977. Effect of Wind and Rain on Surface
Reaeration. J. Environ. Eng. Div. Proc. ASCE 103 (EE3):489-504.
Burns, L. A. 1982. Identification and Evaluation of Fundamental Transport
and Transformation Process Models. In: Modeling the Fate of Chemicals
in the Aquatic Environment. K. L. Dickson, A. W. Maki, J. Cairns,
Jr. Eds. Ann Arbor Science Publishers.
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler,
J. R. Freed, P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri,
W. R. Mabey, B. R. Holt, and C. Gould. 1979. Water-Related
Environmental Fate of 129 Priority Pollutants. Vol. I and II.
EPA-440/4-79-029a and EPA-440/4-79-029b. U.S. Environmental Protection
Agency, Washington, D.C.
Churchill, M. A., H. L. Elmore, and R. A. Buckingham. 1962. The Prediction
of Stream Reaeration Rates. J. Sanlt. Eng. Divn. ASCE 88, No. SA4.
Paper 3199.
Cohen, Y., W. Cocchio and D. Mackay. 1978. Laboratory Study of Liquid - Phase
Controlled Volatilization Rates in Presence of Wind Waves. Environ. Sci.
Technol. 12:553-558.
Copeland, B. J. and W. R. Duffer, 1964. Use of a Clear Plastic Dome to
Measure Gaseous Diffusion Rates in Natural Waters. Limnol. Oceanogr.
9:494-499.
Hall, C. A. S. 1970. Migration and Metabolism in a Stream Ecosystem.
Ph.D. Thesis. University of North Carolina at Chapel Hill.
Holley, E. R. 1978. Oxygen Transfer at the Air-Water Interface. In:
Transport Processes in Lakes and Oceans. R. J. Gibbs ed. Plenum Press.
Isaacs, W. P. and A. F. Gaudy. 1968. Atmospheric Oxygenation in a Simulated
Stream. J. Sanit. Divn. ASCE 94 No. SA2. Paper 5905 319.
Israelsen, 0. W. and V. E. Hansen. 1962. Irrigation Principles and
Practices. John Wiley and Sons, Inc., New York.
Langbein, W. B. and W. H. Durum. 1967. The Aeration Capacity of Streams.
U.S. Geol. Survey Clrc. No. 542, Reston, Virginia.
Liss, P. S. 1973. Processes of Gas Exchange Across an Air-Water Interface.
Deep Sea Res. 20:221-225.
Liss, P. S. and P. G. Slater. 1974. Flux of Gases Across the Air-Sea
Interface. Nature. 247:181-184.
8-9
-------
Mackay, D. 1978. Volatilization of Pollutants from Water. In: Aquatic
Pollutants: Transformation and Biological Effects, 0. Huntzinger,
J. H. Van Lelyveld and B. C. J. Zoeteman eds. Pergamon Press, New York,
New York.
Mackay, D. and P. J. Leinonen. 1975. Rates of Evaporation of Low Solubility
Contaminants from Water Bodies to Atmosphere. Environ. Sci. Technol.
9:1178-1180.
Mackay, D. and T. K. Yuen. 1980. Volatilization Rates of Organic
Contaminants from Rivers. Water Poll. Res. J. Canada 15(2):83-98.
Mackay, D., W. Y. Shiu, A. Bobra, J. Billington, E. Chau, A. Yeun, C. Ng,
and F. Szeto. 1982. Volatilization of Organic Pollutants from Water.
EPA-600/53-82-019. U. S. Environmental Protection Agency, Athens,
Georgia.
Mill, T., W. A. Mabey, D. C. Bomberger, T. W. Chou, D. G. Hendry and
J. H. Smith. 1982. Laboratory Protocols for Evaluating the Fate of
Organic Chemicals in Air and Water. EPA-600/3-82-022 SRI International,
Menlo Park, California.
Parkhurst, J. D. and R. D. Pomeroy. 1972. Oxygen Adsorption in Streams. J.
Sanit. Eng. Divn. ASCE 98 No. 5A1. Paper 8701 101.
Rathbun, R. E. 1977. Reaeration Coefficient of Streams - State-of-the-Art.
J. Hydraulics Div. ASCE. 103 No. HY4. Paper 409.
Rathbun, R. E. and R. S. Grant. 1978. Comparison of the Radioactive and
Modified Techniques for Measurement of Stream Reaeration Coefficients.
U.S. Geol. Survey Water Res. Investigations 78-68.
Smith, J. H., W. R. Mabey, N. Bohonos, B. R. Holt, S. S. Lee, T. W. Chou,
D. C. Bomberger, and T. Mill. 1977. Environmental Pathways of Selected
Chemicals in Freshwater Systems, Part I: Background and Experimental
Procedures. EPA 600/7-77-113. U.S. Environmental Protection
Agency, Washington, D. C.
Tsivoglou, E. C..and J. R. Wallace. 1972. Characterzation of Stream
Reaeration Capacity. EPA-R3-72-012. U.S. Environmental Protection
Agency, Washington, D.C.
U.S.G.S. 1977. National Handbook of Recommended Methods for Water-Data
Acquisition. Reston, Virginia.
Whitman, R. G. 1923. A Preliminary Experimental Confirmation of the Two-film
Theory of Gas Adsorption. Chem. Metallurg. Eng. 29:146-148.
8-10
-------
SORPTION
PROCESS DESCRIPTION
The partitioning of compounds between water, sediments, and biota in
aquatic systems can be an important process affecting pollutant fate. Sediments
can act as a sink for sorbed chemicals, as they are removed from the water
column or as a source at a later time when the chemicals are released (desorbed)
from the sediments. When a solution containing a dissolved chemical contacts
a solid surface sorption may occur. Normally when the total amount of chemical
is increased in a system, the amount of chemical which is sorbed is also
increased (Mill et al. 1982).
Sorption is defined as the binding process of chemicals to surfaces of
solids through chemical or physical interactions or a combination of both.
The process of adsorption implies only physical interaction of molecules of
gases, dissolved substances or liquids to the surface of solid bodies. The
tern sorption will be used in this report so as not to exclude chemical
mechanisms in the binding process.
A sofption-desorption phenomenon depends on the physical and chemical
characteristics of the sorbate (chemical pollutant), sorbent (material that
the chemical adheres to), and environmental conditions.
CHARACTERISTICS OF THE SORBATE
The properties of the sorbate are key factors determining the sorption or
partitioning of a chemical in the environment. Poinke and Chesters (1973)
state that the basic distribution (partitioning) depends more on the physico-
chemical characteristics of the chemical than on the characteristics of the
sorbent or environmental conditions. An excellent review of the properties
that determine the role of the sorbate in sorption and desorption is provided
by Bailey and White (1970). They conclude that the following properties of
the sorbate molecule are important: 1) chemical character, shape, and configu-
ration; 2) acidity or basicity; 3) water solubility; 4) charge distribution
for organic cations; 5) polarity; 6) molecular size; and 7) polarizabillty.
Since the purpose of the report is to review the Influence of environmental
factors on fate and transport processes, the Influence of the sorbate properties
will not be presented in detail. However, the following represent important
generalizations which were derived from the literature:
1) Organic molecules with a negative charge (anions) are unlikely to be
sorbed to an appreciable extent (Knight et al. 1970, Alexander 1961);
2) Cationic compounds are sorbed to a greater extent than neutral molecules.
9-1
-------
Their sorption can occur by cation exchange (Faust 1977, Baughman and
Lassiter 1978);
3) Larger cations are sorbed more readily than smaller ones (Faust 1977);
A) With the exception of sorption by cation exchange, the most insoluble
compounds (neutral) are the most strongly sorbed (Baughman and Lassiter
1978);
5) Neutral, hydrophobic compounds with high octanol/water partition coeffi-
cients are most likely to be sorbed (Karickhoff et al. 1979);
6) The sorption of hydrophobic compounds will be positively correlated to
the organic carbon content of the sorbent (Karickhoff et al. 1979, Faust
1977, Polnke and Chesters 1973);
7) Desorption of many pesticides occurs at a much slower rate than sorption,
e.g., simazlne, atrazine, linuron, monuron, and aldrin (Faust 1977,
Poinke and Chester 1973).
CHARACTER OF THE SORBENT
The properties of the sorbent that influence sorption are primarily related
to the area and configuration of the surface, and to the magnitude, distribution
and intensity of the electrical field at the surface (Bailey and White 1970).
In the aquatic environment a chemical can adhere to the surface of soil
particles, clay minerals, colloids, (such as oxides and hydroxides of iron,
aluminum and manganese), biotic organic matter such as phytoplankton, periphyton,
higher aquatic plants, fish, insects, zooplankton, bacteria, and fungi, and
other abiotic organic matter.
The sorption of most hydrophobic compounds is positively correlated with
organic carbon content of the sorbent and less well with mineral content (Poinke
and Chesters 1973). This results, in part, from 1) the high charge and large
specific surface area associated with highly-decomposed organic matter, and 2)
the fact that clay minerals form organo-clay complexes which can reduce the
number of sorption sites available.
Another factor in the sorption of cation and hydrophobic compounds to
sediments is the correlation of sorption to decreasing particle size (Baughman
and Lassiter 1978, Karickhoff et al. 1979). Karickhoff et al. (1979) were able
to show that the sorption behavior of hydrophobic pollutants can be estimated
within a factor of two from a knowledge of particle size distribution and
organic content of the sediment, and the octanol/water partition coefficient
of the pollutant.
Sediment sorption of cation and hydrophobic organic compounds is strongest
with fine particles because of the higher surface to volume ratio (Baughman
and Lassiter 1978). Table 9-1 shows the effect of particle size on partition
coefficients for hydrophobic chemicals (methoxychlor and pyrene) and cation
exchange (paraquat).
9-2
-------
TABLE 9-1. INFLUENCE OF PARTICLE SIZE ON PARTITION COEFFICIENTS
Size Fraction
Sand
Coarse silt
Medium silt
Fine silt
Clay
KOC*
Methoxychlor
22,000
73,000
88,000
93,000
81,000
Pyrene
19,000
92,000
128,000
122,000
109,000
v*
Paraquat
400
25,000
35,000
250,000
750,000
Source: Baughman and Lassiter (1978)
*KQC The amount of chemical sorbed per unit of organic carbon divided by
the concentration of chemical in the water.
**Kp The concentration of a pollutant in the sorbent divided by concentration
of pollutant in the water at equilibrium.
While the sorption of nonpolar and polar neutral organic molecules is
usually positively correlated with the organic carbon content of the sorbing
solid, charged organic molecules often behave differently with their sorption
more related to the clay mineral content. Two important properties of the
mineral sorbent that control sorption are cation exchange capacity and surface
area. Sorption of charged organic molecules will generally increase with an
increase in the surface area and in the cation exchange capacity of the
sorbent.
Biota play an important role in the distribution of organic compounds in
the aquatic environment (Faust 1977). Poinke and Chesters (1973) point out
that aquatic weeds can sorb significant quantities of organophosphorus and
organochlorine pesticides in relation to the bottom sediments. For example,
in one study the most rapid sorption and greatest concentration of p,p'-DDT
was associated with vegetation and not the sediments (Bridges et al. 1963).
Also, it should be noted that considerable quantities of organics can be
sorbed by fish, insects, and zooplankton (Hamelink et al. 1977).
Sorption of a chemical to the sediments is basically limited to the upper
exposed layers. Important physical factors Influencing sediment sorption
are the actual amount of sorbent (dry weight) per volume of sediment or
conversely, the percent water content and bulk density of the sediment.
If the sediment contains a high percentage of water and low bulk density,
less material per volume of sediment will be available for sorption than
if the sediment contained a low percentage of water and high bulk density
per volume of sediment.
9-3
-------
ENVIRONMENTAL FACTORS INFLUENCING SORPTION
Besides the characteristics of the sorbate and sorbent, environmental
conditions affect sorption phenomena. The major environmental factors that
influence sorption are pH, temperature, biota, and the presence of other
dissolved organics.
Since the ionization state of a chemical can greatly influence sorption,
a knowledge of the pH of the water and sediment is important. The strength of
sorbtion will depend on the degree of ionization particularly where the
sorption mechanism is ion exchange.
Neutral compounds or strong acids or bases will not be affected by changes
in pH, since their form is not dependent on the pH of their environment (Mill
et al. 1982). Depending on the pH of the environment, weak acids or bases may
be present in two different forms. Under low pH conditions, weak bases will
be in the cationic form and weak acids in the free acid form. Since most
sediments have negatively charged sites the free acid and cationic forms will
be more highly sorbed than their respective anionic acid and free base counter-
parts. Therefore, as the pH of a sorption system decreases the sorption of
weak acids and bases increases.
Since sorption processes are normally exothermic and desorption processes
are endothermlc, an Increase in temperature would normally tend to reduce
sorption and favor desorption (Bailey and White 1970). With Increased tempera-
tures attractive forces between solute and the solid surface are weakened,
with a concomitant increase in solute concentration. Cationic pesticides that
are sorbed to highly organic soils are normally not significantly influenced
by temperature variations (Foinke and Chesters 1973).
Organic chemicals are generally sorbed less when other indigenous dissolved
organic chemicals are present. This decrease in sorption is apparently due to
two processes. First, the increased solubility of the organic chemical of
interest in the presence of other organic materials, since sorption has been
found to correlate inversely with solubility. The second process that results
in decreased sorption is the competition for sorption sites. Naturally occuring
organics will also be sorbed on sediments and any sorption will decrease the
total number of sorption sites available to the chemical of interest (Mill et
al. 1982).
Sorption of a chemical to bottom sediments Is often limited to the upper
exposed layers. However, bioturbatlon can increase the depth of the distribu-
tion of the chemical in the bottom sediments (Aller 1978, Rippey and Jewson
1982). Bioturbation may result In physical disturbance by demersal fishes,
Irrigation of sediments by tube-dwelling macrofauna, and mixing of contaminated
layers by macrofauna populations after the contaminant has sorbed to the sedi-
ment surface. Benthic organisms may also affect partical size distribution
and seasonal layering of sediment deposits by ingesting sediments In sub-surface
layers and depositing fecal material at the surface (Davis 1974).
9-4
-------
Table 9-2 lists the environmental factors, Including the properties of the
sorbent, which most influence sorption rates. Organic priority pollutants for
which sorption is thought to be an important process in aquatic systems are
listed in Table 9-3.
METHODS
General Considerations
Sorption is defined as the binding process of chemicals to surfaces of
sediments through chemical and/or physical interactions. In the aquatic envi-
ronment a chemical generally adheres to sediments, suspended particles, and
biota. Sorption of nonpolar organic chemicals onto sediments is a function of
the partition coefficient of the compound, the organic carbon and water content
and bulk density of the sediment. Also, under certain conditions particle
size may greatly influence sorption.
The partition coefficient of a chemical is measured by equilibrating an
aqueous solution of the pollutant in an environmentally realistic concentration
with a known concentration of sediment. After reaching equilibrium the distri-
bution of chemical between the water and sediment phases is measured. The
partition coefficient for a particular nonpolar compound can be normalized to
the organic content of aquatic sediments. The normalized sorption constant
(KQC) is described by the following equation at equilibrium conditions:
Chemical sorbed to sediment/dry weight of sediment
K • Water concentration of chemical x 100
oc ' Percent organic carbon in sediment
KQC can be a stable, system-independent measure of an intrinsic property
of nonpolar organic compounds. For a compartmentalized fate and transport
model, Koc can be used to compute a partition coefficient for sediment phases
of an aquatic system as a function of the organic carbon content of each sedi-
ment compartment. Koc is strongly correlated with the octanol-vater partition
coefficient (K^) (Karickhoff et al. 1979). By definition, KQW expresses
the equilibrium concentration ratio of an organic chemical partitioned between
octanol and water in a dilute solution. Koc can be estimated as 41 percent
of Kow for whole sediments.
Sorbtion of a chemical to the sediments is basically limited to the upper
exposed layers. An important factor influencing sediment sorption is the
actual amount of sorbent (dry weight) per volume of sediment or conversely the
percent water content and bulk density of the sediment. Therefore, if an
investigator is given a K0C or K^ value as part of the chemical input
data to a model, he would have only to describe the average organic carbon
content, bulk density, and water content of each sediment compartment to
estimate sediment sorption.
However, it is possible that in addition to organic carbon, sorption can
be correlated with total amount of surface area of the sorbing material
9-5
-------
TABLE 9-2. ENVIRONMENTAL FACTORS AND SORBENT PROPERTIES WHICH MOST
INFLUENCE SORPTION RATES
Environmental Factors
Sorbent Properties
pH
Temperature
Dissolved organic carbon
Biota
Percent organic matter
Particle size/surface area
Percent water
Bulk density
Cation exchange capacity
TABLE 9-3. ORGANIC PRIORITY POLLUTANTS FOR WHICH SORPTION IS ESTIMATED
TO BE AN IMPORTANT PROCESS
Pesticides
Aldrin
ODD
DDE
DDT
Dieldrin
Endosulfan and Endosulfan sulfate
Heptachlor epoxide
Hexachlorocyclohexane
TCDD
Toxaphene
Polychlorinated biphenyls
Polycyclic Aromatic Hydrocarbons
Acenaphthene
Acenaphthylene
Anthracene
Benzo [a] Anthracene
Benzo [b] fluoranthene
Benzo [k] fluoranthene
Benzo [g,h,i] perylene
Benzo [a] pyrene
Chrysene
Dibenzo [a,h] anthracene
Fluoranthene
Fluorene
Indeno [1,2,3-cd] pyrene
Naphthalene
Phenanthrene
Pyrene
Monocyclic Aromatics
Hexachlorobenzene
2,A-Dinitrotoluene
2,6-Dinitrotoluene
Pentachlorophenol
2-Nitrophenol
4-Nitrophenol
2,4-Dinitrophenol
4,6-Dinitro-o-cresol
Phthalate Esters
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Bis (2-ethylhexyl) phthalate
Butyl benzyl phthalate
Halogenated Aliphatics
Hexachlorobutadiene
Nitrogen-containing Compounds
Benzidine
3,3-'Dichlorobenzidine
1,2-Diphenylhydrazine
Source: Callahan et al. 1979.
9-6
-------
(particle size). Increased sediment sorption generally correlates to
decreasing particle size (Baughman and Lassiter 1978). Particle size nay be
particularly important for sediments with an organic content so low that
sorption to organic surfaces is Insignificant compared to sorption by
inorganic surfaces. . In these instances particle size data may be. required to
accurately estimate sediment sorption. Further information on this topic is
provided by Polnke and DeAngelis (1979).
Polar organic molecules also sorb to sediments, suspended particles, and
biota. Sorption of polar molecules (ions) also correlates with organic
carbon, but not as closely as do nonpolar ones. In general polar compounds
often behave differently, and some property or combination of properties other
than organic carbon may have to be used in the development of a normalized
sorption constant for polar compounds. Possible factors that could be used in
the development of a normalized sorption constant for polar compounds include
cation exchange capacity, anion exchange capacity, particle size, and pH. The
complexity of this process has hindered the development of robust, system-
independent analogs of the Koc for nonpolar organlcs.
If the only chemical input data for sediment sorption is a partition
coefficient, an Investigator can apply it equally to all sediments or develop
a partition coefficient which is specific for the chemical and site of
interest.
As noted earlier, chemicals also adhere to suspended sediments and biota.
To estimate sorption to suspended sediments, Koc or Kow constants can be
applied If one knows the mass of suspended sediments and their percent organic
carbon.
The actual quantities of a chemical captured by the biomass are often
relatively small, compared to the amounts sorbed by sediments or dissolved in
the water column. Therefore, biomass is often relatively insignificant as a
transport or capture medium affecting the fate of an organic pollutant.
However, biomass accumulation of pollutants is extremely important because of
the possibility of direct toxic effects on aquatic life or the possibility of
toxic effects on man through food-chain vectors. Also, biota may be an
important factor in distribution of the contaminant via bioturbation.
A biomass partition coefficient (bioconcentration factor) may be provided
as part of the chemical input data to a given model. This factor may apply to
specific populations, e.g., fish, benthos, etc., or to functional groups
(carnivores and herbivores) or the entire population. With this factor an
investigator would need to measure the biomass of the appropriate population.
Methods to measure biomass are described in the Bioconcentration section
of this report.
Sediment Sampling
To utilize normalized sorption coefficients in a compartmentalized fate
and transport model, representative sediment samples are required. For
example, the average percent organic carbon content of each sediment
9-7
-------
compartment would be a required Input value if a K^c or KQW were provided.
In addition, the volume of each sediment compartment would be needed. Length
and width measurements of sediment compartments can be readily obtained from
naps. However, what Is an appropriate depth of a sediment compartment?
Obviously, chemicals must be exposed to sediments before sorption can occur.
Often, only the upper surface layers of sediment can participate in sorption
phenomena. One possible solution to this problem is to take sediment core
samples. The depth at which the chemical of interest can be detected can
be used as the depth of the sediment compartment.
Sediments are largely eroded soils that have been subjected to
redlspersal and particle size fractionation. In the aquatic environment,
dispersion of suspended particles follows sedimentation patterns.
Sedimentation patterns are a function of particle size and density, and
current velocities. Sediments in a given aquatic environment may be fairly
uniform. For example, sediments in a river system may be distributed as
follows: 1) clay suspended in the water column, 2) sand in the middle bottom
sediments, and 3) silt at the edge of the stream. In a reservoir or lake
environment coarser grained sediments are generally deposited near the
headwaters of a reservoir/lake with the bed sediments near the center composed
of fine grained materials. For sediment samples to be representative of a
given compartment, sample allocation needs to be proportional to the
geographical distribution of the various size fractions of the sediments.
Methods for sediment collection are described in the Physical Transport
section of this report.
Sediment Analysis
An investigator may have to measure the following sediment properties for
inputs to fate and transport models: total organic carbon content expressed
as percent, anion exchange capacity, cation exchange capacity, pH, percent
water content, bulk density, and particle size. In addition,one may need
to measure suspended sediments and their percent organic carbon. Methods
for determination of suspended sediments, bulk density, percent water,
particle size, and total organic carbon content of bottom sediments are
given in the Physical Transport section.
pH of Sediments—
Sediment pH can be measured by a procedure described by Peech (1965).
One part of a sediment is added to two parts of .01M CaC12 to adjust the
ionic strength of the suspension. The suspension Is stirred several times
over a 30-minute period and allowed to stand for 30 minutes to settle before
measurement with a glass electrode. The glass electrode is immersed into the
suspension, but the reference electrode remains In the clear supernatant above
the suspension to minimize liquid junction potential effects.
Cation Exchange Capacity—
Cation exchange in sediments is a reversible chemical reaction. Cations
held on the surface and within sediments can be replaced with cations of salt
solutions and acids. Cation exchange capacity (CEC) is defined as the sum of
the exchangable cations of a sample (usually expressed as mllliequivalents per
100 g of sample).
9-8
-------
Many methods have been and continue to be proposed to determine CEC.
Kelley (1968) and Peech (1965) describe in detail the pitfalls of
various methods. The ammonium acetate and sodium acetate methods are often
used as the "standards" for the comparison of new methods. The "standard"
methods do not- accurately reflect field conditions since the pH and ionic
strength of the test medium (important factors affecting CEC) are altered in
the analytical procedures. The sodium and ammonia methods are described
by Black et al. (1965).
Anion Exchange Capacity—
The anion exchange reaction of sediments may be described as the substi-
tution of one anion by another which is available in solution in a greater
concentration or possesses a stronger tendency to hold its position on the
sediment. The anion exchange capacity (AEC) of a sediment is measured by the
amount of an anion sorbed and then displaced from a sediment. Phosphate or
arsenate anions are frequently used to measure AEC. Dean and Rubins (19A7)
described methods to measure AEC.
9-9
-------
REFERENCES
Alexander, M. 1961. Introduction to Soil Microbiology. John Wiley and Sons,
Inc., New York, New York.
Aller, R. C. 1978. Experimental Studies of Changes Produced by Deposit
Feeders on Pore Water, Sediment, and Overlaying Water Chemistry. AM. J.
Sci. 278:1185-1234.
Bailey, B. W. and J. L. White. 1970. Factors Influencing the Adsorption,
Desorption, and Movement of Pesticides in Soil. Residue Review 32:29-82.
Baughman, G. L. and R. R. Lassiter. 1978. Prediction of Environmental
Pollutant Concentration. In: Estimating the Hazard of Chemical
Substances to Aquatic Life, ASTM STP 657 John Cairns Jr., K. L. Dickson
and A. W. Maki, eds. American Society for Testing and Materials,
Philadelphia, Pennsylvania.
Black, C. A., D. D. Evans, J. L. White, L. E. Ensminger, and F. E. Clark
(eds.). 1965. Methods of Soil Analysis. Volumes 1 and 2. Amer. Soc.
Agron. Madison, Wisconsin.
Bridges, W. R., B. J. Kallman, and A. K. Andrews. 1963. Persistence of DDT
and its Metabolites in a Farm Pond. Trans. Amer. Fish Soc. 92:421-423.
Burns, L. A., D. M. Cline, and R. R. Lassiter. 1982. Exposure Analysis
Modeling System (EXAMS): User Manual and System Documentation. EPA-
600/3-82-023. U.S. Environmental Protection Agency, Athens, Georgia.
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler,
J. R. Freed, P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri,
W. R. Mabey, B. R. Holt, and C, Gould. 1979. Water-Related
Environmental Fate of 129 Priority Pollutants. Vol. I and II.
EPA-440/4-79-029a and EPA-440/4-79-029b. U.S. Environmental Protection
Agency, Washington, D.C.
Davis, R. B. 1974. Stratigraphic Effects of Tubificids in Profundal Lake
Sediments. Limnol. Oceanogr. 19(3):466-488.
Dean, L. A. and E. J. Rubins. 1947. Anion Exchange in Soils: I. Exchange-
able Phosphorus and the Anion-Exchange Capacity. Soil Science.
63:377-387.
Faust, S. D. 1977. Chemical Mechanisms Affecting the Fate of Organic
Pollutants in Natural Aquatic Environments. In: I. H. Suffet, ed.
Advances in Environmental Science and Technology Vol. 8:317-365, Fate of
Pollutants in the Air and Water Environments. Part 2. Chemical and
Biological Fate of Pollutants in the Environment. Symposium at the 165th
National American Chemical Society Meeting. Philadelphia, Pennsylvania.
April 1977. John Wiley and Sons, Inc. New York.
9-10
-------
Hamelink, J. L., R. C. Waybrant, and P. R. Yant. 1977. Mechanisms of
Concentration of Mercury and Chlorinated Hydrocarbon Pesticides by Fish
in Lentic Ecosystems. In: I. H. Suffet, ed. Advances in Environmental
Science and Technology, Vol. 8:261-282. Fate of Pollutants .in the Air
and Water Environments. Part 2. Chemical and Biological Fate of
Pollutants in the Environment. Symposium at the 165th National American
Chemical Society Meeting. Philadelphia, Pennsylvania. April 1975. John
Wiley and Sons, Inc. New York.
Karickhoff, S. W., D. S. Brown, and T. A. Scott. 1979. Sorption of
Hydrophobic Pollutants on Natural Sediments. Water Res. 13:241.
Kelley, W. P. 1968. Cation Exchange in Soils. Reinhold, New York.
Knight, B. A., G. J. Counts, and T. E. Tomllnson. 1970. Sorption of Ionized
Pesticides by Soil. In: Sorption and Transport Processes in Soils.
J. G. Gregory and C. J. Rawlins, eds. Staples Printers Limited.
Rochester, Kent, England.
Mill, T., W. R. Mabey, D. C. Bomberger, T. W. Chou, D. G. Hendry, and J. H.
Smith. 1982. Laboratory Protocols for Evaluating the Fate of Organic
Chemicals In Air and Water. Draft Document Prepared by SRI
International, Menlo Park, California, for Environmental Research
Laboratory, Athens, Georgia.
Peech, M. 1965. Hydrogen-ion Activity. Agronomy 9(2):914-926.
Poinke, H. B. and G. Chesters. 1973. Pesticide - Sediment - Water
Interactions. J. Environ. Quality 2:29-44.
Poinke, H. B. and R. J. DeAngelis. 1979. To be published in: "System for
Evaluating Nonpoint Source Pollution on a Field Scale," USDA. Scientific
Education Administration Series.
Rippey, B. and D. H. Jewson. 1982. The rate of Sediment-water Exchange of
Oxygen and Sediment Bioturbatlon in Lough Neagh, Northern Ireland.
Hydrobiologia. 92:377-382.
9-11
-------
BIOCONCENTRATION
INTRODUCTION
The characteristic of living organisms to act as compositors and
integrators of toxic substances from the environment is well known. Aquatic
organisms concentrate and accumulate substances directly from the water and
through the food chain (Rudd 1964; Woodwell et al. 1967). Uptake of toxic
substances is not restricted to aquatic organisms directly in contact with the
water. This is illustrated by thinning of egg shells and reproductive failure
of birds that prey upon fish contaminated with DDT and its derivatives (Hickey
and Anderson 1968; Anderson and Hickey 1970), and reproductive difficulties of
commercially raised mink that were fed Lake Michigan coho salmon (40 percent
of their diet) contaminated with PCBs (Metcalf 1977).
Three processes are involved in biological uptake of toxic substances:
bioconcentration, bloaccumulation, and biomagnification. The terminology of
the three processes is clarified by Brungs and Mount (1978) as follows:
"Bioconcentration is usually considered to be that process by which toxic sub-
stances enter aquatic organisms, by gill or epithelial tissue, from the water.
Bioaccumulatlon is a broader term in the sense that it usually includes not
only bioconcentration but also any uptake of toxic substances through consump-
tion of one organism. Biomagnification refers to the resultant total process
including bioconcentration and bioaccumlation by which tissue concentration of
bloaccumulated toxic substances increases as this material passes up through
two or more trophic levels."
UPTAKE RATES AND MECHANISMS
A number of factors influence uptake of organic constituents from the
water and through the food chain, but there is a lack of agreement among
investigators regarding the significance of various factors. Hamelink et
al. (1977) list the primary factors governing the quantity of residues
available directly from the water as the concentration of the compound In
the water, the volume of water passed over the gills and the efficiency
with which fish extract the compound from the water. Metcalf (1977) points
out that absorption of lipid soluble substances from water is a function
of the concentration of compound times the exposure time. Several investiga-
tors have noted direct relationships between residue concentration in
tissues of fish and the age or size of the animal. Hamelink et al. (1977)
suggested that the rate at which DDE and mercury are acquired from both
food and water by fish versus their growth rate may largely determine
residue concentrations. Metcalf (1977) observed that bioconcentration of
DDT and dleldrin by fish from coldwater lakes is essentially a linear
function of the age of the fish. Youngs et al. (1972) reported that
concentrations of DDT residues and its metabolites increased with the age of
lake trout (Salvellnus namaycush) from Cayuga Lake, New York. Relnert (1970)
reported that concentrations of DDT and dleldrln increased steadily with the
size of Lake Michigan lake trout. Residue concentrations may also vary
among species of fish (Reinert 1970), suggesting that feeding habits,
10-1
-------
distribution patterns, physiological characteristics, life cycles, fat content
and other species-specific factors influence bioaccumulation rates.
DeFoe et al. (1978) noted that female fathead minnows accumulated about
twice as much PCBs as males exposed to similar concentrations in water, pre-
sumably because of the greater amount of lipid in the females. Veith et al.
(1979) reported sizable differences in bioconcentration factors of warm water
and cold water fish exposed to similar concentration of PCBs, hexachlorobenzene
and 1,2,4-trichlorobenzene at various temperatures. Warm water species showed
a large temperature dependence on uptake up to approximately 15°C, whereas
trout were less affected by temperature variation.
Other investigators point out that there is substantial evidence that
effects of species, size, age, temperature and chemical concentrations can be
accounted for, or are insignificant, or negligible, relative to bioconcentration
(Dickson et al. 1981). Veith et al. (1979) also contend that the age and
source of test fish as well as the presence of multiple pollutants in water
have little effect on bioconcentration.
The bioconcentration of organic compounds by fish has been shown to be
directly related to the octanol-water partition coefficient (Neely et al.
1974, Veith et al. 1979) and inversely correlated with water solubility (Metcalf
1977, Lu and Metcalf 1975, and Kapoor et al. 1973) for a variety of compounds.
The basic concept is based on the tendency of lipid soluble organic chemicals
to partition out of the water to compartments containing non-polar cellular
components (e.g., lipids in fatty tissues of living organisms) and cell
membranes. The partition of chemicals between lipids and water determines the
entry of lipid soluble substances, either directly from water through cuticle,
or from water to blood via the gills and from blood to tissue (Metcalf 1977).
Metcalf (1977) suggests that the octanol/water partition coefficient for any
organic compound can be used as a first approximation of its relative tendency
to bioconcentrate in living organisms. However, he further states that the
end result is clearly a function of the stability of the compound in water and
the rate at which more highly partitioned degradation products are formed in
the organisms (Metcalf 1977).
Norstrom, et al. (1976) developed a pollutant accumulation model to predict
the concentration of PCBs and methyl mercury in tissues of yellow perch from
the Ottawa River, Canada. The model relates the uptake of pollutants from
food—for a given fish species—to both the energy requirements (growth and
maintenance) of that species and the concentration of pollutants in its food
source. The assimilation efficiency of pollutants from specific food sources
is also incorporated as an element in the model. Pollutant uptake from water
is based on the flow of water past the gills, the concentration of the pollutant
in the water and the efficiency of removal by the gills. The metabolic rate
expression in the model includes a term for estimating the effects of seasonal
and annual growth in each age class. Also given is a pollutant clearance
function related to body weight, but independent of metabolic rate.
Neely (1979) discussed the rates of uptake and clearance of organic sub-
stances by fish and proposed a method for estimating rate constants based on
10-2
-------
the physiology of the fish and properties of the chemical. His method expands
on the work of Norstrom et al. (1976) relative to the bioenergetics of chemical
uptake by fish and the work of Neely et al. (1974) showing the relationship
between the octanol-water partition coefficient and bioconcentration in fish.
. Neely (1979) points out that the efficiency with which a fish extracts a
chemical from the water flowing past its gills is a function of the efficiency
with which a chemical moves across the gill membrane, the oxygen concentration
in the water and the amount of oxygen required to maintain the respiration
rate of the fish. The efficiency with which a chemical moves across gill
membranes (efficiency transfer coefficient) increases directly with the log of
the octanol water partitioning coefficients (Kow) and the rate constant for
uptake (kj) of the various compounds.
ENVIRONMENTAL FACTORS INFLUENCING BIOCONCENTRATION
Temperature and dissolved oxygen content of the water are frequently men-
tioned environmental factors that influence bioconcentration of organochlorine
pesticides in aquatic organisms. Both water temperature and the dissolved
oxygen content influence the metabolic activity of the organism. Dickson et
al. (1981), however, mention that effects of environmental factors, such as
temperature, on bioconcentration and uptake depuration rates have not
been adequately studied.
It has been fairly well established that those compounds with a high
lipid/water partition rate and low rates of degradation and elimination may
continue to concentrate in organisms as long as the organisms are exposed.
That is, equilibrium between water and tissue lipids may not be achieved with
stable compounds, because the time required .for this to occur often exceeds
the longevity of the organism (Metcalf 1977 and Hamelink et al. 1977). Conver-
sely, a state of equilibrium may be rapidly achieved with less persistent
compounds such as lindane (Gakstatter and Weiss 1967, Dickson et al. 1981) or
methoxychlor which is rapidly cleared by organisms (Metcalf 1977). Table 10-1
presents a list of organic priority pollutants most subject to bioconcentration,
arranged by classes of compounds.
The bioconcentration potential of any given toxic organic substance in a
particular stream segment depends upon, 1) the type and quantity of biota
present in the stream reach; 2) the availability of the compound for
biological involvement (concentration in the water, period of exposure to
receptors); and 3) the chemical nature of the compound in terms of its
solubility in water, its tendency to partition to lipids and its degradation
characteristics both in water and within organisms.
A number of factors Influence the availability of a compound for biocon-
centration, including the quantity and rate of delivery of the substance, to
the system, the dilution effects of receiving waters (considering volume of
flow and vertical and horizontal mixing characteristics of the stream); and .
the nature and quantity of suspended and bedload sediment in the stream reach
(including rates of transport and depostion of sediments and sorbed pollutants
within the stream segment).
10-3
-------
TABLE 10-1. ORGANIC PRIORITY POLLUTANTS MOST SUBJECT TO BIOLOGICAL UPTAKE
Pesticides
Aldrin
Chlordane
DDD
DDE
DDT
Dieldrin
Endrin and Endrin Aldehyde
Heptachlor Epoxide
TCDD
Polychlorinated Biphenyls
Nitrosamines
1,2-Diphenylhydrazine
Acrylonitrile
Monocyclic Aromatics
1,3-Dichlorobenzene (m-Dichlorobenzene)
1,4-Dichlorobenzene (p-Dichlorobenzene)
1,2,4-Trichlorobenzene
Hexachlorobenzene
Phthalate Esters
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Bis (2-ethylhexyl) phthalate
Butyl benzyl phthalate
Source: Callahan et al. 1979.
Obviously, a number of compound specific processes such as tendencies to
undergo sorption, volatilization, photolysis, oxidation, ionization,
hydrolysis and/or microbial transformation are key factors governing the
availability of a compound for biological uptake within a given system. These
processes, and the environmental factors that influence the nature and rates
of each, are of secondary importance In discussions of bioconcentration per
se, and are addressed in other portions of this report under discussions of
specific processes. For example, physical transport of a substance through a
system is dependent not only upon the residence time of a volume of water in
the stream segment, but the form of the substance (dissolved, colloidal,
sorbed onto sediments), nature of the flow, and sediment transport, deposition,
resuspension, etc.
APPLICATION OF BIOLOGICAL DATA TO FATE AND TRANSPORT MODELS
Many fate and transport models that attempt to predict the ultimate
disposition of chemicals introduced to waterways incorporate only general
provisions for biological uptake, storage, transformation and transport of
substances in streams. For example, input data for particular models may
require gross estimates of total biomass within a particular stream segment or
compartment, with no consideration of the percentage contribution of various
community components to the total biomass (e.g., EXAMS). Other models (e.g.,
PEST) accept biomass Input data separated by major categories of communities
based on distribution, ecological niches or behavior patterns; e.g.,
stationary (zoo benthos, rooted macrophytes) vs. mobile or drifting forms
(fish, phytoplankton, zooplankton, and floating macrophytes). Some models
emphasize bioaccumulation, but ignore chemical process (Thomann 1978). Models
10-4
-------
such as PEST incorporate provisions to accept site specific and organism
specific input data reflecting current knowledge of the differential response
of communities—and components within communities—to exposure, in terms of
their "efficiency" as compositors, their ability to degrade or transform
substances, and their potential for exporting substances across compartment
or ecosystem boundaries (e.g., sediment to water, one stream segment to
another, riffles to pools, streams to lakes, etc.). These features represent
significant advances in model sophistication for purposes of examining the
behavior of toxic substances in the context of an entire aquatic ecosystem.
As experience is gained in the use of these models and as refinements are
made, capabilities to predict fate and transport of toxic organic substances
in natural aquatic systems can only be improved.
In using models such as EXAMS, which are designed for use as quick
screening tools, only gross estimates of the quantities of particular
substances that will partition to the biota in a given stream segment
under steady-state conditions can be derived as output data. Even if
octanol-water partition coefficients for a particular introduced chemical
are known, the generated output data cannot accurately predict the quantities
of a substance stored, removed or transferred by the biota unless the
total biota of the stream segment consist of a known blomass for which
uptake potential has been established. This is because the lipid content
and the bioconcentration potential of the biota vary between community
types, e.g., vascular plants, periphyton, predaceous fish, insects, mollusks,
crustaceans, etc. Also, the life cycles, habits and distribution patterns
of these receptors are sufficiently different that duration of exposure
varies considerably.
It must be assumed that model sensitivity will continue to be improved
for acceptance of input data reflecting selectivity of the various biological
components as compositors of toxic organic contaminants. Information is
available in the literature on uptake potential of a number of community types
for selected contaminants (primarily pesticides). However, continued
biological testing of other priority pollutants is essential if the role of
the biota in the fate and transport of contaminants is to be adequately
defined.
Table 10-2 presents a list of those biological and environmental factors
that Influence biological uptake of toxic organic substances in aquatic systems.
Proper sampling of the biological components listed in conjunction with
measurements of the key environmental factors (dissolved oxygen and water
temperatures) will provide sufficient information for most types of aquatic
fate and transport models that include biological uptake as a fate process.
This table should not be taken as an exhaustive list of input variables for
any and all aquatic fate and transport models for toxic organic substances,
but rather as a minimum list of environmental parameters and community compo-
nents to be measured or sampled in the field for model validation purposes.
In designing a field sampling program, the investigator must be well
aware of the manner in which the data will be utilized in the model and tailor
his sampling program accordingly. For example, if total biomass per compart-
ment is identified as an input variable, the Investigator must determine the
10-5
-------
dimensions of the compartment in question through appropriate field measurements
and calculations and plan his sampling to adequately describe the standing
crop within the compartment boundaries. He should also be well aware of any
special measurement, preparation, or sample handling requirements for particular
samples.
TABLE 10-2. PRINCIPAL ENVIRONMENTAL FACTORS AND BIOLOGICAL COMPONENTS THAT
INFLUENCE BIOCONCENTRATION OF ORGANIC COMPOUNDS IN AQUATIC SYSTEMS
Factor
Dissolved oxygen content
Water temperature
Fish
Total biomass per compartment
Size and age classes
Species distribution
Food Habits - percentage herbivores, predators
Resident and migratory species
Macroinvertebrates
Total biomass per compartment
Percentages by functional groups (e.g., scrapers, shredders, collectors,
etc.)
Periphyton
Total biomass per compartment
Total diatom biomass
Total nondiatom biomass
Aquatic Vascular Plants (Rooted, Floating)
Total biomass per compartment
Phytoplankton
Total biomass per compartment
Total diatom biomass
Total nondiatom biomass
Zooplankton and Invertebrate Drift
Total biomass per compartment
Particulate organic matter (Seston)
Total mass per compartment
METHODS
Problems of assessing biological influence on transport and fate of toxic
organic substances in flowing systems are compounded by sampling problems
associated with obtaining reliable quantitative biomass estimates of the various
communities. This portion discusses stream communities and identifies key
references to field sampling and measurement methods for gathering biological
data that may be required for model validation purposes. Because each aquatic
10-6
-------
community or taxonomic group presents special sampling problems and requires
the use of special equipment and methods, the various ecosystem components are
addressed separately. Field procedures for measuring water temperature and
dissolved oxygen content were described in the Biotransformatlon section of the
report. The reader is referred to that discussion for measurement of those
parameters.
BENTHIC MACROINVERTEBRATES
Stream Habitats and Standing Crop Estimates
Stream biologists commonly recognize three major types of stream habitats:
riffles, runs, and pools. Each habitat has its own benthic community charac-
terized by particular species associations. Biomass also normally differs
greatly among the three habitats. To obtain reliable biomass estimates within
a given stream reach, the proportion of each habitat must be estimated through
mapping, and the biomass of each habitat estimated through appropriate sampling
methods.
Because each habitat represents a discrete compartment, characterization of
the communities requires use of sampling procedures and equipment tailored to
the particular environment. Detailed descriptions of samplers and their opera-
tion are provided by Usinger (1956); Weber (1973); Merrit and Cummins (1978);
Southwood (1978); and Resh (1979). The samplers described are known as absolute
(or "quantitative") samplers because they are designed to estimate density (or
biomass) per unit area of habitat (Southwood 1978). Although these are many
"Quantitative" sampling methods available for detemming Biomass, the investi-
gator should be aware that sampling devices do not always provide an absolute
measure of biomass. For example it has been shown that different benthic sampling
devices that collect from the same area of stream bottom provide different
estimates of benthic biomass (Kroeger 1972, Pollard and Kinney 1979). This
should be taken into consideration when deciding which sampling device to use
to obtain biomass estimates for model input. Appendix B-l describes the various
habitats and discusses sampling procedures and field processing techniques
appropriate for each.
ALGAE
Biomass Estimates
Biomass measurements provide important information on algal abundance and
growth. Algal biomass is the amount of algal material present, i.e., the
standing crop of these primary producers. Many models, predicting the conse-
quences of nutrients, turbidity, toxicants, hydrological modifications, etc.,
require accurate measurements of algal biomass. Algal biomass may occur as
phytoplankton—the algae suspended within the water column—or periphyton—
attached algae and other algae associated with them. In lakes, especially
those with poorly developed shorelines, the plankton are usually the most
Important primary producers. In contrast, lotic systems seldom allow for
plankton development and periphyton may virtually be the only primary producers
in this ecosystem, therefore occupying a very Important role (Grzenda and
Brehmer 1960).
10-7
-------
Detailed discussions of the materials and methods for collection, preserv-
ation, and analysis of algal samples are found in APHA (1980), Lund and Tailing
(1957), Sladeckova (1962), Sournia (1978), Vollenweider (1969), Weitzel (1979),
and Weber (1973). Appendix B-2 provides a discussion of methods appropriate
for sampling and processing algal communities.
FISH :
Significance as Concentrators of TOS
Fish may represent the aquatic community of principal interest in any
investigation designed to assess or predict the impact of Introduced contam-
inants on aquatic ecosystems. This is particularly true in Investigations
involving biocumulative toxic substances owing to the position of fish at the
upper levels of the aquatic food web and the direct link they provide in the
transfer of toxic biocumulative substance from the aquatic environment to man.
Because fish function as compositors of toxic substances, tissue levels fre-
quently exceed concentrations of ambient waters and of the lower trophic level
food organisms by several orders of magnitude. In some cases, under conditions
approaching steady-state, equilibrium may be reached between concentrations in
tissue and in the surrounding water for short lived compounds.
Sampling and Inventory Methods
Appendix B-3 examines fish sampling and inventory approaches and methods
for estimating populations in a range of situations likely to be encountered
in stream surveys. The advantages, disadvantages, limitations and uses of
each of the various techniques and sampling devices are addressed.
Fish collection methods are frequently discussed under two categories:
active and passive. Active collection methods include the use of electro-
shockers, seines, trawls, poisons, and angling gear. Passive methods involve
the use of gill, Fyke, trammel, hoop and pound nets, and D-traps, and
purchasing fish from fisherman. Other inventory methods, which normally do
not Involve collecting specimens, include visual observations and the use of
remote sensing devices.
In some instances techniques involving diversion of entire streams have
been used to make population estimates. This technique was used by Embody
(1929) who diverted a trout stream and counted the fish remaining In the old
channel. Needham et al. (19A5) periodically made population estimates in a
trout stream by diverting the water and pumping the remaining pools. Such
techniques are not generally useful except in very small streams and under
unusual circumstances, and will not be further discussed.
Detailed description of the various fish sampling and inventory approaches,
methods, and equipment are available in the literature. Three especially
useful sources are: Bagenal (1978), Weber (1973), and U.S. EPA (1979). These
manuals should be studied and maintained in the possession of anyone contem-
plating undertaking fish surveys for any purpose.
10-8
-------
The Intent of this section is to summarize the important factors of the
various approaches and techniques to enable the investigator to select the
method most suited to his purpose and to the particular conditions of the
stream under investigation. A combination of methods may be required to obtain
the necessary information. However, the investigator will have to make that
determination on a case-by-case basis.
MACROPHYTES
!
Significance as Concentrators of TOS
Macrophytes include all aquatic plants possessing a multicellular structure
with cells differentiated into specialized tissue (Weber 1973). They include
the mosses, liverworts, and flowering plants and range in size from nearly
microscopic water meal to massive cypress trees. The most common and conspicuous
stream macrophytes in most systems are the rooted vascular plants.
Macrophyte abundance may vary dramatically from one watershed to the next
or even within a particular stream. Physical factors such as depth, current
speed, turbidity, water level stability, and the volume of the substrate are
primary factors governing the areal extent of coverage and abundance of macro-
phytes in most flowing waters. Obviously chemical and biological factors are
also important, as nutrient deficiencies, herbicide contamination or excessive
grazing by herbivores tend to reduce or eliminate macrophytes In habitats with
physical characteristics suitable for abundant growths.
Macrophytes may constitute the bulk or at least a substantial portion of
the total biomass in particular stream segments or habitats. Consequently,
vascular plants may represent the major aquatic community component influencing
fate and transport of introduced toxic organic compounds. Macrophytes are
generally not considered to be as efficient concentrators of toxic organic
substances as are secondary consumers, for example, but, by virtue of their
sheer mass alone, the quantities of substances temporarily removed from the
water column could be considerable. Although some transformation and degrada-
tion of organic compounds occur as a result of uptake by macrophytes, they
undoubtedly function primarily as temporary sinks. Because macrophytes exhibit
pronounced cyclic, seasonally-dependent growth patterns, the significance of
their involvement in removal, transformation and transport of aquatic toxic
organic substances varies considerably throughout the year. Secondly, dif-
ferential rates of uptake occur among communities and species owing to physio-
logical and metabolic differences as well as to distribution patterns and
growth habits that influence duration of exposure.
Survey Methods and Procedures
Quantitative sampling of aquatic macrophytes for purposes of obtaining
absolute standing crop or biomass estimates are rarely undertaken in stream
surveys. For the purposes of most surveys, estimations of the areal extent
and relative abundance (e.g., dense, moderate, sparse) of the most conspicuous
plant communities based on visual inspection are adequate. If several species
are present, percentages of the individual taxa are often estimated and
qualitative samples collected for Identification purposes.
10-9
-------
Qualitative and quantitative aquatic macrophyte sampling techniques are
discussed in Slack et al. (1973), Weber (1973), and Vollenweider (1969).
(Also see Appendix B-4.)
ZOOPLANTON AND DRIFT FAUNA
Occurrence in Flowing Systems
The invertebrate fauna in the water column of flowing systems consists
both of true zooplankters and benthic forms dislodged from their normal
habitats. Much of the zooplankton community of streams represents contributions
from slack water areas. Hynes (1970) points out that there are no planktonic
animals that are known to be restricted to flowing systems. True reproducing
zooplankton occur in large, sluggish rivers, but such communities are generally
very unstable. Rich zooplankton communities frequently are found downstream
from reservoirs as a result of reservoir export. Such export drastically
affects the trophic relationships in these reaches, with the gross effect
being a great increase in total invertebrate biomass.
Drift of bottom dwelling organisms is a normal occurrence in all streams,
and represents a mechanism of redistribution of populations and recolonlzation
of denuded areas. The total biomass in drift during periods of peak movement
frequently exceeds the total standing crop of the established benthos. Drift
organisms are important forage items for stream fish and periods of peak feeding
activity frequently coincide with periods of maximum benthic drift.
Stream biological surveys designed to investigate invertebrate density
and distribution for purposes of determining fish food availability, water
quality relationships or similar purposes should include measurements of drift
fauna as one component of the stream benthos. The drift community represents
components of benthic organisms associated directly with the stream bottom as
well as those animals associated with littoral vegetation that through
behavioral or catastrophic (e.g., flooding) activity have dislodged from the
substrate. Drift represents a mechanism of transport of considerable inverte-
brate biomass. Much of the community associated with behavioral drift will
become reestablished in downstream reaches—although predation takes a heavy
toll—representing a mechanism of redistribution of organisms and associated
bloconcentrated substances. Organisms dislodged from their habitats as a
result of heavy flooding or disturbances to the stream bed or littoral vegeta-
tion are frequently badly mutilated and are unable to recolonlze downstream
reaches. Some of these animals are eaten by fish while others are deposited
along with debris in slack water areas including lakes and reservoirs with
habitats unsuitable for colonization.
Procedures for sampling stream drift organisms are described in Weber
(1973), Elliott (1970), Waters (1962, 1969) and others. Appendix B-5 provides
a discussion of field methods appropriate for obtaining drift biomass estimates.
10-10
-------
REFERENCES
Anderson, D. W. and J. J. Hickey. 1970. Oological Data on Egg and Breeding
Characteristics of Brown Pelicans. Wilson Bull. 82:14-28.
APHA. 1980. Standard Methods for the Examination of Water and Wastewater.
Fifteenth Edition. APHA/AWWA//WPCF. Washington, D.C.
Bagenal, T. B. (Ed.). 1978. Methods for Assessment of Fish Production in
Fresh Waters. IBP Handbook No. 3, 3rd edition. Blackwell Scientific
Publication, Oxford.
Brungs, W. D. and D. I Mount. 1978. Introduction to a Discussion of the Use
of Aquatic Toxicity Tests for Evaluation of the Effects of Toxic
Substances. In: Cairns, J., K. L. Dickson and A. W. Maki, eds.
Estimating the Hazard of Chemical Substances to Aquatic Life. ASTM
STP 657, pp. 15-26.
Callahan, M. A., M. W. Slimak, N. W. Gabel, I. P. May, C. F. Fowler,
J. R. Freed, P. Jennings, R. L. Durfee, F. C. Whitmore, B. Maestri,
W. R. Mabey, B. R. Holt, and C. Gould. 1979. Water-Related
Environmental Fate of 129 Priority Pollutants. Vol. I and II.
EPA-440/4-79-029a and EPA-440/4-79-029b. U.S. Environmental Protection
Agency, Washington, D.C.
DeFoe, D. L., G. D. Veith, and R. W. Carlson. 1978. Effect of Aroclor
1248 and 1260 on the Fathead Minnow (Pimepholes promelas). J. Fish.
Res. Bd. Con. 35:1060-1066.
Dickson, K. L., J. H. Rodgers, and F. Y. Solek. 1981. Measuring Rate
Constants for Chemicals in Single Model Laboratory Studies. Final
Report. Chemical Manufacturing Association Project MO. ENV-8-W.
Section 7, pp. 1-27.
Elliott, J. M. 1970. Methods of Sampling Invertebrate Drift in Running
Waters. Ann. Limnol. 6:133-159.
Embody, G. C. 1929. Stocking Policy for the Streams, Lakes and Ponds of the
Erie-Niagara Watershed, Exclusive of Lake Erie: A Biological Survey of
the Erie-Niagara System. Suppl. 18th Ann. Rept. (1928) N.Y. State
Conserv. Dept.: pp. 23-24.
Gakstatter, J. L. and C. M. Weiss. 1967. The Elimination of DDT-C14,
Dieldrin-C14 and Lindane-Cl4 from Fish Following a Single Sublethal
Exposure in Aquaria. Trans. Amer. Fish. Soc. 96:301-307.
Grzenda, A. R. and M. L. Brehmer. 1960. A Quantitative Method for the
Collection and Measurement of Stream Periphyton. Limnol. Oceanogr.
5:190-194.
10-11
-------
Hatnelink, J. L., R. C. Waybrant, and P. R. Yant. 1977. Mechanisms of
Bioconcentration of Mercury and Chlorinated Hydrocarbon Pesticides by
Fish in Lentlc Ecosystems. In: Advances in Environmental Science and
Technology. 1. H. Suffet, ed. Vol. 8:261-282. Fate of Pollutants
in the Air and Water Environments. Part 2. Chemical and Biological Fate
of Pollutants in the Environment. Symposium at the 165th National
American Chemical Society Meeting. Philadelphia, Pennsylvania. April 1,
1975. John Wiley and Sons, Inc. New York.
Rickey, J. J. and D. W. Anderson. 1968. Chlorinated Hydrocarbons and
Eggshell Changes in Raptorial and Fish-eating Birds. Science
162:271-273.
Hynes, H. B. N. 1970. The Ecology of Running Waters. University of Toronto
Press. Toronto.
Kapoor, I. P., R. L. Metcalf, A. S. Hirwe, J. R. Coats, and M. S. Khalsa.
1973. Structure Activity Correlations of Biodegradability of DDT
Analogs. J. Agri. Food Chem. 21:310-315.
Kroeger, L. 1972. Underestimation of Standing Crop by the Surber Sampler.
Limnol. Oceanogr. 17:(3) 475-479.
Lu, Po-Yung and R. L. Metcalf. 1975. Environmental Fate and Biodegradability
of Benzene Derivatives as Studied in a Model Aquatic Ecosystem.
Environ. Health Perspect. 10:269-284.
Lund, J. W. G. and J. F. Tailing. 1957. Botanical Limnological Methods with
Special Reference to the Algae. Bot. Rev. 23:489-583.
Merritt, R. W. and K. W. Cummins. 1978. An Introduction to the Aquatic
Insects of North America. Kendall/Hunt Publishing, Dubuque, Iowa.
Metcalf, R. L. 1977. Biological Fate and Transformation of Pollutants in
Water. In: Advances in Environmental Science and Technology.
I. H. Suffet, ed. Vol. 8:195-221. Fate of Pollutants in the Air and
Water Environments. Part 2. Chemical and Biologial Fate of Pollutants
in the Environment. Symposium at the 165th National American Chemical
Society Meeting. Philadelphia, Pennsylvania. April 1975. John Wiley
and Sons, Inc. New York.
Needham, P. R., J. W. Moffett and D. W. Slater. 1945. Fluctuations in Wild
Brown Trout Populations in Convict Creek, California. J. Wildl. Mgt.
9:9-25.
Neely, W. B. 1979. Estimating Rate Constants for the Uptake Clearance of
Chemicals by Fish. Environ. Sci. and Technol. 13:1506-1510.
Neely, W. B., D. R. Branson, and G. E. Blau. 1974. Partition Coefficient to
Measure Bioconcentration Potential of Organic Chemicals in Fish.
Environ. Sci. and Technol. 8:1113-1115.
10-12
-------
Norstrom, R. J., A. E. McKinnon, and A. S. W. deFreitas. 1976. A
Bioenergetics-Based Model for Pollutant Accumulation by Fish. Simulation
of PCB and Methylmercury Residue Levels in Ottawa River Yellow Perch
(Perca flavescens). J. Fish. Res. Bd. Can. 33:248-267.
Pollard, J. E. and W. L. Kinney. 1979. Assessment of Macroinvertebrates
Monitoring Techniques in an Energy Development Area. A Test of the
Efficiency of Three Macrobenthic Sampling Methods in the White River.
EPA-600/7-79-163. U.S. Environmental Protection Agency, Las Vegas,
Nevada.
Reinert, R. E. 1970. Pesticide Concentrations in Great Lakes Fish. Pest.
Mont. Jour. 4:233-240.
Resh, V. H. 1979. Sampling Variability and Life History Features: Basic
Considerations in the Design of Aquatic Insect Studies. J. Fish Res.
Bd. Can. 36:290-311.
Rudd, R. L. 1964. Pesticides and the Living Landscape. Univ. of Wise.
Press, Madison, Wisconsin.
Slack, K. V., R. C. Averett, P. E. Greeson and R. G. Lipscomb. 1973. Methods
for Collection and Analyses of Aquatic Biological and Microbiological
Samples. In: Techniques of Water-Resources Investigation of the United
States Geological Survey. Book 5, Chapter A4. U.S. Government Printing
Office, Washington, D.C.
Sladeckova, A. 1962. Limnological Investigation Methods for the Periphyton
("Aufwuchs") Community. Botanical Review 28:286-350.
Sournia, A. (Ed.). 1978. Phytoplankton Manual. Monograph on
Oceanographic Methodology 6. United Nations Education, Scientific, and
Cultural Organization, Paris.
Southwood, T. R. E. 1978. Ecological Methods. John Wiley and Sons, New
York.
Thomann, R. V. 1978. Size Dependent Model of Hazardous Substances in Aquatic
Food Chain. Final Report on Contract No. R803680030, EPA-600/3-78-036.
U.S. Environmental Protection Agency, Duluth, Minnesota.
Usinger, R. L. 1956. Aquatic Insects of California. University of
California Press, Los Angeles.
U.S. EPA. 1979. Sampling Protocols for Analysis of Toxic Pollutants in
Ambient Water, Bed Sediment and Fish. EPA Contract No. 68-01-3852. U.S.
Environmental Protection Agency, Washington, D.C.
Velth, G. D., D. L. DeFoe, and B. V. Bergstedt. 1979. Measuring and
Estimating the Bloconcentration Factor of Chemicals in Fish. J. Fish
Res. Board Can. 36:1040-1048.
10-13
-------
Vollenweider, R. A. (Ed.)- 1969. A Manual on Methods for Measuring Primary
Production in Aquatic Environments. The IBP Handbook No. 12. 2nd ed.
Blackwell Scientific Publication, Oxford.
Waters, T. F. 1962. Diurnal Periodicity in Drift of Stream Invertebrates.
Ecology 43:316-320.
Waters, T. F. 1969. Invertebrate Drift - Ecology and Significance to Stream
Fishes. In: Symposium on Salmon and Trout in Streams. T. G. Noalhiate,
ed. N. R. MacMillon Lectures in Fisheries. Univ. British Columbia,
Vancouver, pp. 121-134.
Weber, C. I. (Ed.). 1973. Biological Field and Laboratory Methods for
Measuring the Quality of Surface Water and Effluents. EPA-670/4-73-001.
U.S. Environmental Protection Agency, Cincinnati, Ohio.
Weitzel, R. L. (Ed.). 1979. Methods and Measurements of Perlphyton
Communities: A Review. ASTM Special Technical Publication 690.
American Society for Testing and Materials. Philadelphia.
Woodwell, C. M., C. F. Wurster, and P. A. Isaacson. 1967. DDT Residues in an
East Coast Estuary: A Case of Biological Concentration of a Persistent
Insecticide. Science 156:821-824.
Youngs, W. D., W. R. Gutenmann, and D. J. Lisk. 1972. Residues of DDT in
Lake Trout as a Function of Age. Environ. Sci. Technol. 6:451-452.
10-14
-------
PHYSICAL TRANSPORT
TRANSPORT
Toxic organic chemicals, at low concentrations in natural waters, exist
in a dissolved phase and a sorbed phase. Dissolved substances are trans-
ported by water movement with little or no "slip" relative to the water.
They are entirely entrained in the current and move at the water velocity.
Likewise, organics which are sorbed to colloidal material or fine suspended
solids are essentially entrained in the current, but they may undergo addi-
tional transport processes such as sedimentation and deposition or scour and
resuspension. These processes may serve to retard the movement of the sorbed
substances relative to the water movement. Thus in order to determine the
fate and transport of toxic organic substances, we oust know both the water
movement and sediment movement.
Water Budget
The importance of a good water budget should not be ignored. Physical
transport of water in a clearly defined ecosystem is accounted for in a water
balance.
Accumulation Direct
of H20 - Inflows + Precipitation - Outflows - Evaporation
(+ Infiltration - Exfiltration + Overland Runoff)
Water can be stored within lakes or rivers by a change in elevation or stage.
Inflows and outflows should be gaged or measured over the period of investiga-
tion. Precipitation gages and evaporation pans can be utilized with sufficient
accuracy. In the best of situations, It Is possible to achieve an annual
water balance within five percent (total inflows are within 5 percent of total
outflows). Confounding factors include infiltration, exfiltration, and
overland runoff.
Transport of Chemicals In Water
The transport of toxic chemicals in water principally depends on two
phenomena: advectlon and dispersion. Advectlon refers to movement of
dissolved or fine particulate material at the current velocity in any of :
three directions (longitudinal, lateral or transverse, and vertical).
Dispersion refers to the process by which these substances are mixed within
the water column. Dispersion can also occur in three directions. There are
a number of processes v.'hich contribute to the mixing (dispersion) process:
11-1
-------
1. Molecular diffusion. This is the mixing of dissolved chemicals due
to the random walk of molecules within the fluid. It is caused by
kinetic energies of molecular vibrational, rotational, and
translational motion. In essence, molecular diffusion corresponds
to an increase in entropy whereby dissolved substances move from
regions of high concentration to regions of low concentration
according to Pick's laws of diffusion. It is an exceedingly slow
phenomena, such that it would take on the order of 10 days for 1
mg/1 of dissolved substance to diffuse through a 10 cm water column
from a concentration of 10 mg/1. It is generally not an important
process in the transport of dissolved substances in natural waters
except relating to transport through thin and stagnant films at
the air-water interface or transport through sediment pore water.
2. Turbulent Diffusion. Turbulent or eddy diffusion refers to mixing
of dissolved and fine particulate substances due to micro-scale
turbulence. It is an advective process at the microscale level
caused by eddy fluctuations in current velocity. Shear forces
within the body of water are sufficient to cause this form of
mixing. It is several orders of magnitude larger than molecular
diffusion and is a contributing factor to dispersion. Turbulent
diffusion can occur in all three directions, but is often aniso-
trophic (i.e., there exist preferential directions for turbulent
mixing due to the direction and magnitude of shear stresses).
3. Dispersion. The interaction of turbulent diffusion with velocity
profiles in the water body causes a still greater degree of mixing
known as dispersion. Transport of toxic substances in streams and
rivers is predominantly by advection, but transport in lakes and
estuaries is often dispersion controlled. Velocity gradients are
caused by shear forces at the boundaries of the water body, such as,
vertical profiles due to wind shear at the air-water interface,
and vertical and lateral profiles due to shear stresses at the sedi-
ment-water and bank-water interfaces. Also velocity gradients can
develop within the water body due to channel morphology, the sinuousity
and meandering of streams, and thermal or density stratification
and instabilities in lakes and estuaries. Morphological causes of
dispersive mixing in rivers include dead spots, side channels, and
pools where back-mixing occurs. When turbulent diffusion
causes a parcel of fluid containing dissolved substances to change
position, that parcel of fluid becomes entrained in the water body
at a new velocity, either faster or slower. This causes the parcel
of fluid and the toxic substance to mix forward or backward relative
to its neighbors. The mixing process is called dispersion and
results in a mass flux of toxic substances from areas of high
concentration to areas of low concentration, analagous to molecular
diffusion, but at a much more rapid rate. The mass flux rate can
be described by Pick's first law of diffusion:
t
dc
J - - K A — (1)
dx
11-2
-------
J
K
A
dc
dx
max flux rate, M/T
diffusion or dispersion coefficient,
cross sectional area through which diffusion occurs
— » concentration gradient, M/L^-L
Advective-Dispersive Equation
The basic equation describing advection and dispersion of dissolved
natter is based on the principle of.conservation of mass and Pick's law. For
a conservative substance, the principle of conservation of mass can be stated:
Rate of change
of mass in
control volume
at
where C
t
'Rate of change of"
mass in control
volume due to
advection
Rate of change of
mass in control
volume due to
diffusion
dC
-u
Transformation"
Reaction Rates
(Degradation)
(2)
R -
concentration, M/L.3
time, T
average velocity in the i'th direction, L/T
distance in the i'th direction, L
reaction transformation rate, M/L3-T
E! is the diffusion coefficient in the i'th direction. For laminar
flow, t± • CM, the coefficient of molecular diffusion. For turbulent flow,
CJL - ej + e^ where CT is the coefficient of turbulent diffusion. In Fickian
diffusion theory, it is assumed that dispersion resulting from turbulent open-
channel flow is exactly analogous to molecular diffusion. The dispersion
coefficients in the x, y, and z directions are assumed to be constants,
given by Kx, Kv and Kz. The resulting equation, expressed in Cartesian
coordinates is:
ac ac ac ac
— + u — + u — + u — -K
at x ax y ay z az x
+ K
+ K
- R
(3)
the solution of equation (3) depends on the values of Kx, Kv and Kz. Various
authors have arrived at equations to approximate the values of the dispersion
coefficients (K) in the longitudinal (x), lateral (y), and vertical (z)
directions.
Longitudinal Dispersion Coefficient in Rivers
Liu (1977) used the work of Fischer (1967) to develop an expression for
11-3
-------
the longitudinal dispersion coefficient in rivers and streams (K^, which has
units of length squared per time):
2 3 2
uv B QR
*x - P -Z - P -5—^ (A)
UA D3
where Liu (1978) defined,
D » mean depth, L
B « mean width, L
U - bed shear velocity, L/T
ux - mean stream velocity, L/T
A • cross sectional area, L*
QB » river discharge, L3/T
P does not depend on stream morphometry but on the dimensionless bottom rough
ness. Based on existing data for KX in streams, the value of Kx can be pre-
dicted to within a factor of six by equation (A). The bed shear velocity is
related empirically to the bed friction factor and mean stream velocity:
U -JI° - Jlu2 (5)
*
in which To » bed shear stress, M/L-T2
f • friction factor =0.02 for natural, fully turbulent flow
p » density of water, M/L3
Lateral Dispersion Coefficient in Rivers
Elder (1959) proposed the equation for predicting the lateral dispersion
coefficient, Ky:
K - * D U^ (6)
where is equal to 0.23. The value of $ - 0.23 was obtained by experiments
In long, wide laboratory flumes.
Many authors have since Investigated the value of $ in both laboratory
flumes and natural streams. Sayre and Chang (1968) reported $ - 0.17 in a
11-4
-------
straight laboratory flume. Yotsukura and Cobb (1972) report values of $ for
natural streams and irrigation canals varying from 0.22 to 0.65, with most
values being near 0.3. Other reported values of $ range from 0.17 to 0.72.
The higher values for $ are all for very fast rivers. The conclusions drawn
are that; 1) the form of equation (6) is correct predicting Ky, but $ may
vary, and 2) application of Flckian theory to lateral dispersion is correct
as long as there are no appreciable lateral currents in the stream.
Okoye (1970) refined the determination of $ somewhat by use of the aspect
ratio, X * D/B, the ratio of the stream depth to stream width. It was found
that 4> decreased from 0.24 to 0.093 as X increased from 0.015 to 0.200.
The effect of bends in the channel of Ky is significant. Yotsukura and
Sayre (1976) reported that 4> varies from 0.1 to 0.2 for straight channels,
ranging in size from laboratory flumes to medium size irrigation channels; $
varies from 0.6 to 10 in the Missouri River, and $ varies from 0.5 to 2.5 in
curved laboratory flumes. Fischer (1968) reports that higher values of $ are
also found near the banks of rivers.
Vertical Dispersion Coefficient in Rivers
Very little experimental work has been done on the vertical dispersion
coefficient, Kz. Jobson and Sayre (1970) reported a value for marked fluid
particles of:
K2 -
-------
Vertical eddy diffusivities can be calculated from temperature data by
solving the vertical heat balance or by the simplified estimations of Edinger
and Geyer (1965). Schnoor and Fruh (1979) have demonstrated that the
mineralization and release of dissolved substances from anaerobic sediment
can be used to calculate average hypolimnetic eddy diffusivities. This
approach avoids the problem of assuming that heat (temperature) and mass
(dissolved substances) will mix with the same rate constant, i.e., that the
eddy diffusivity must equal the eddy conductivity. A summmary of dispersion
coefficients and their order of magnitude appears below.
Dispersion Coefficient, cm2/sec
Molecular Diffusion 10~5
Compacted Sediment 10~7 - 10~5
Bioturbated Sediment 10~5 - 10~*
Lakes - Vertically 10~2 - 101
Large Rivers - Lateral 10~2 - 103
Large River - Longitudinal 10^ - 105-5
Estuaries - Longitudinal 106 - 107
Choosing a Transport Model
It is possible to estimate the relative importance of advection compared
to dispersion with the Peclet number:
Pe - uL/K (9)
in which Pe • Peclet number, dimensionless
u - mean velocity, L/T
L • segment length, L
K - dispersion coefficient, L2/T
If the Peclet number is significantly greater than 1.0, advection predominates;
if it is much less than 1.0, dispersion predominates the transport of dissolved,
conservative substances.
If there is a significant transformation rate, the reaction number can
be helpful:
kK
Rxn No. - — (10)
u
where k - the first order reaction rate constant, T"1. If the reaction number
is less than 0.1, then advection predominates and a model approaching plug
flow is appropriate. If the reaction number is greater than 10, then
dispersion controls the transport and the system is essentially completely
mixed. Otherwise a number of compartments in series will best simulate the
prototype water body.
11-6
-------
Compartmentalization
Compartmentaliration refers to the segmentation of model ecosystems into
various "completely-mixed" boxes of known volume and interchange. Interchange
between compartments is simulated via bulk dispersion or equal counterflows
between compartments. Compartmentalization Is a popular assumption in fate
and transport modeling because the completely-mixed assumption reduces the
set of partial differential equation (in time and space) to one of ordinary
differential equations (in time only). Nevertheless it is possible to recover
some coarse spatial information by introducing a number of interconnected
compartments.
A completely-mixed flow-through (CMF) compartment contains an ideal
mixing of fluid in which turbulence is so large that no concentration gradients
can exist within the compartment. This corresponds to the assumption that
Kx „ 2 » oo. Equation (3) reduces to:
p -i ("Mass 1 ["Dispersive"! T Mass "|
I Accumulation of Mass! - llnflowsl -f I Inflows I - (Outflows! -
|_w/in Compartment jj |_ to 3 J L to J J |_ from j J
[Dispersive! ^Transformation"!
Outflows - I Reactions I
from J J L w/in J J
dCj n n n n
VJ 7" " A QJ*kCk + J, QvkCk ' A *•& " A Qk-JcJ '
dt k-1 k-1 J k-1 k-1 k
o
in which Vj » volume of j compartment, L
C. - concentration within j compartment , M/L^
t - time, T
n • number of adjacent compartments to j
QJ k - inflow from compartment k to compartment j , L /T
Cj^ » concentration in compartment k, M/L^
Q' • dispersive (interchange) flow from k to j ,
j'k ,
Q^ j » outflow from J to k, LJ/T
Q1 • dispersive (interchange) flow from j to k,
fc.J
k - pseudo-first order rate constant for transformation,
and Q' - Q'j, 4 for a symmetric matrix with zero diagonal
J.fc 'J
11-7
-------
Equation (11) can be rewritten in terms of bulk dispersion coefficients:
dCj n n n
dt k-1 " k-1 ' k-1
where K' - bulk dispersion coefficient,
*
A. , - interfacial area between compartments j and k, 1*
j t*-
- distance between midpoints of compartments, L
There is one mass balance equation (e.g. equation 12) for each of n compart-
ments. This set of ordinary differential equations is solved simultaneously
by numerical computer methods.
Bulk dispersion coefficients between compartments are dependent on the
scale chosen for the compartments. They are not equivalent to measured
dispersion coefficients from dye studies (these are usually derived from the
continuous partial differential equations). The very nature of the
compartmentalized system introduces considerable mixing into the model. Such
mixing or numerical dispersion is in addition to the bulk dispersion specified
by the bulk dispersion coefficient.
Streams and swift-flowing rivers may approach a 1-D plug flow system
(i.e. the water is completely mixed in the lateral and vertical dimensions,
but there is no mixing in the longitudinal dimension). In an ideal plug flow
system, the longitudinal dispersion coefficient is equal to zero since there
is no forward or backward mixing. For this case an infinite number of
compartments (of infinitestlmal length in in the longitudinal direction)
would be required in order to produce zero longitudinal mixing. Since it is
impossible to specify an infinite number of compartments, one chooses a finite
number of compartments and accepts the artificial dispersion which accompanies
that choice. One method of estimating the artificial or numerical dispersion
of a compartmentalized model for an ideal, plug flow system is given by
equation (13).
uAx uAt
E (1 ) (13)
x 2 Ax
O
where EX - artificial numerical dispersion coefficient, I//T
u • mean longitudinal velocity, L/T
Ax • longitudinal length of equal spaced compartments, L
At * time step for numerical computation, T
One approach would be to set the artificial dispersion coefficient equal
to the measured or estimated dispersion coefficient from equation (4). With
this approach It is not necessary to use bulk dispersion coefficients; rather,
one allows the artificial dispersion of the model to account for the actual
dispersion of the prototype.
11-8
-------
In general most river simulations require many compartments due to their
nearly plug flow nature, as indicated by their large Peclet number (equation
9). The greater the number of compartments, the greater the tendency towards
plug flow conditions. It is a poor choice to simulate a riverine environment
with one completely-mixed compartment.
Lakes, reservoirs, and embayments may require a number of compartments
if one desires some spatial detail, such as concentration profiles, these
compartments should be chosen to relate to the physical and chemical realities
of the prototype. For example, a logical choice for a deep lake is to have
two compartments: an epillminlon and hypolimnion. Mixing between compart-
ments can be accomplished by interchanging flows:
J » QexDepi - Qexchypo
where J . » net mass flux from epilimnion to hypolimnion due to vertical
mixing, M/T
Qex - exchange flow, L /T
C • concentration of chemical, M/L^
The magnitude of the interchange flow, Qex, can be determined from tracer
studies or from temperature profiles and simulations. Bulk dispersion
coefficients can then be calculated based on the interchange flow as
Sometimes only coarse Information is required for a given use of the
model. There are many lucid examples in the literature of modeling efforts
based on very simple transport models. The U.S. Great Lakes have often been
simulated as single compartment, completely-mixed lakes in series (Chapra,
1977). Toxic chemical screening methodologies are usually based on organic
chemical properties that are known only within an order-of -magnitude. In
such cases it may not be necessary to simulate transport with great accuracy.
There exists a distinct trade-off between errors in transport formulations
and errors in reaction rate constants as shown in Table 11-1. If the sum of
the pseudo-first order reaction rate constant is accurately determined in the
field or laboratory, then an accurate model simulation will require a
realistic transport formulation.
For example, consider a hypothetical lake whose steady state outlet
concentration of a toxic chemical is determined to be 0.01 times the inflow
concentration. Suppose the hydraulic detection time, t, of the lake is 10
days, and the transformation reaction rate constant is determined to be
1.0/day (kt»10). The lake is behaving like three compartments-in-series
according to Table 11-1. However the model calibration would have required a
reaction rate constant of 10.0/day in order to obtain the observed result of
C/Co • 0.01 if only the completely-mixed compartment had been assumed.
11-9
-------
TABLE 11-1. OUTFLOW CONCENTRATION DIVIDED BY/INFLOW CONCENTRATION AT
STEADY STATE AS A FUNCTION OF NUMBER OF COMPARTMENTS AND kx
C/Co VALUES
Rate Constant x Detention Time
kt-O.Ol kx=0.1 kx-1.0 kt-10 kt-100
CMF*
(1-compart.)
0.99
* C/Co - l/(k-c+l)
+ C/Co - l/([kT/n
t C/Co - exp(-kt)
0.91
0.50
0.09
0.01
3-compart ."*"
10-compart."*"
PFt
(» compart.)
0.
0.
0.
99
99
99
0.
0.
0.
91
91
90
0.
0.
0.
42
39
37
0
1 X
5 x
.01
10-3
10-5
2
4
4
x
x
X
10-5
10-"
io-«*
where T • total hydraulic detection time
k » first order reaction rate constant
where n - number of compartments
In summary, the first approach should be to choose compartments that
correspond to the morphometric features of the aquatic ecosystem prototype.
If better than order of magnitude accuracy is required in the model, one can
estimate dispersion coefficients from semi-empirical equations, e.g. equations
(4) through (8). If still greater accuracy is desired, one should utilize
tracer studies and simulations. Temperature and velocity are thermal and
momentum "tracers" that can be used to calibrate the mixing characteristics
of the water body. Salinity can often serve as a conservative tracer in
estuarine studies, but dye or radioisotope tracers generally yield the best
quality information.
SEDIMENT TRANSPORT
Chemicals can be associated with sediments by adsorption (electrostatic
surface forces), chemisorption (chemical bonding between the chemical and
surface), or ion exchange (surface exchange of ionic species). Suspended
solids (also termed "suspended sediment") can undergo transport in the
horizontal and vertical dimensions.
Bed load transport includes those particles which saltate, bounce, roll,
or flow within a few particle diameters of the bottom sediment. In rivers,
downstream transport of suspended sediment includes particles which settle to
the bottom and very fine particles which do not settle substantially and are
termed "wash load". Total sediment transport consists of suspended load
plus bed load. It comprises sediments of autochthonous (internally produced
by primary and secondary aquatic production) and allochthonous (externally
produced from land and air) origin. Suspended and bed load sediments consist
of a distribution of particle types and sizes, including those classified as
sand, silt, and clay.
11-10
-------
Partitioning
A chemical is partitioned into a dissolved and particulate adsorbed
phase based on its sediment-to-water partition coefficient, Kp (Karickhoff
et al., 1979). The dimensionless ratio of the dissolved to the particulate
concentration is the product of the partition coefficient and the concentra-
tion of suspended solids, assuming local equilibrium.
Cp/C - KpM (15)
where Cp - particulate chemical concentration, ug/A
C - dissolved chemical concentration, ug/A
Kp - sediment/water partition coefficient, i/kg
M • suspended solids concentration, kg/A
The particulate and dissolved concentrations can be calculated from knowledge
of the total concentration, CT, as stated in equations (16) and (17).
KpM
Cp - CT (16)
1 4 KpM X
1
CT (17)
1 + KpM l
These concentrations can be calculated for the water column or the bed
sediment, by using the concentration of suspended solids in the water (M) or
in the bed (Mj,), respectively.
Suspended Load
The suspended load of solids in a river or stream is defined as a flow
rate times the concentration of suspended solids, e.g. kg/day or tons/day.
The mean suspended load of a river is greatly weighted by peak flows. Peak
flows cause large inputs of allochthonous material from errosion and runoff
as well as increases in scour and resuspenslon of bed and bank sediment.
The average suspended load is not equal to the average flow times the
average concentration, as stated in equations (18) and (19).
Q x C + QC (18)
QC - QxC + ITC"7 (19)
The mean fluctuation of mass, Q'C', is usually greater than the first term of
equation (19) and contributes greatly to the average suspended load. These
equations hold true for the mass of suspended solids as well as the mass of
adsorbed chemical.
11-11
-------
Bed Load
Several formulas- have been reported to calculate the rate of sediment
movement very near the bottom. These were developed for rivers and non-
cohesive sediments, i.e. fine-to-coarse sands and gravel. It is important
to note that it is not sands, but rather silts and clays, which transport
most sorbed chemicals. Therefore these equations are of limited predictive
value in environmental exposure assessments. Generally bed load transport
Is a small fraction of total sediment transport (suspended load plus bed
load). However, in estuaries, bed load transport of fine silts and clays may
be an important contribution to the fate of chemical contaminants. Un-
fortunately, predictive equations have not been developed for bed load
transport in such applications. For a detailed discussion of available
techniques the reader is referred to the methods section presented below.
Sedimentation
Suspended sediment particles and adsorbed chemicals are transported down-
stream at nearly the mean current velocity. In addition, they are transported
vertically downward by their mean sedimentation velocity. Generally the
particles settle, according to Stoke's Law, In proportion to the square root
of the particle diameter and the difference between sediment and water
densities. One modification of Stoke's Law that is frequently used to
calculate particle fall velocities is the Rubey equation (1933):
w - Fj I gds (20)
in which
w » particle fall velocity, ft/sec
Ys • specific weight of sediment particle, lbs/ft3
•y • specific weight of water, lbs/ft^
g - gravitational constant, ft/sec^
ds - sediment particle diameter, ft
" •,
A.
1/3 ' .
36v2
3, . ..
(21)
v • kinematic viscosity of water, ft^/sec
Generally it is the washload (fine silt and clay-size particles) which
carry most of the adsorbed chemical. These materials have very small fall
11-12
-------
velocities, on the order of 0.3 m/day for clays of 2-4 urn nominal diameter
and 3 in/day for silts of 10-20 \m nominal diameter.
Once a particle reaches the bed, there exists a certain probability that
it can be scoured from the bed sediment and resuspended. The difference
between sedimentation and resuspension represents net sedimentation. Often
It is possible to utilize a net sedimentation rate constant in the fate and
transport model to account for both processes. In many ecosystems where the
bed is aggrading, sedimentation is much larger than resuspension, Schnoor
and McAvoy (1981). The net sedimentation rate constant can be calculated as
follows:
w w
- - - ku - -
H H
(22)
where ks - net sedimentation rate constant, 1/T
w • mean particle fall velocity, L/T
H • mean depth, L
1^ - scour/resuspension rate constant, 1/T
Scour/Resuspension
Quantitative relationships to predict scour and resuspension of cohensive
sediments are difficult to develop due to the number of variables Involved.
Sayre and Chang (1968) reported on the vertical scour and dispersion of silt
particles in flumes. DiToro et al. (1982) recommended a resuspension velocity
(Wr8) of about 1-30 mn/yr based on model calibration studies. The turbulent
vertical eddy diffusivity for sediment (cs) is also related to the scour
coefficient and/or a resuspension velocity.
Under steady state, the sedimentation of suspended sediment must equal
the scour and resuspension of sediment.
w C +
oC
es —
(23)
where
w
6 •
C •
- sedimentation velocity, L/T
suspended sediment vertical eddy diffusivity, L/T
concentration of suspended sediment,
However under time-varying conditions, the boundary condition at the bed-
water interface is more complex. According to Onlshi and Wise (1979) the
following equation applies, based on the work of Krone (1962) and Partheniades
(1965).
&C
(1-p) w C + es — - SD - SR
(24)
11-13
-------
where p » probability that descending particle will "stick"
SD - rate of bed deposition, M/L2-T
SD -
o
SR • rate of bed scour, M/L-T
M. • erodibility coefficient, M/L2-T
r\
TC£ * critical bed sheaar required for resuspension, M/L-T
TCD « critical bed shear stress which prevents deposition, M/L-T
h • ratio of depth of water to depth of active bed layer
Equation (24) shows that the bed can either be aggrading or degrading at any
time or location depending on the relationship between SD and SR.
Desorption/Dif fusion
In addition to sedimentation and scour /resuspension, an adsorbed chemical
can desorb from the bed sediment. Likewise dissolved chemical can adsorb
from the water to the bed. Both pathways can be presented by a diffusion
coefficient (K^) and a concentraation gradient or difference between water
and sediment concentrations.
SUMMARY
Sediment mass balances must Include terms for advection, sedimentation,
scour /resuspension, and possibly vertical or longitudinal dispersion. At
the bottom, bed load movement may be included. Processes which affect the
fate of dissolved substances Include desorption from the bed (or adsorption
from the water column), advection, dispersion, and transformation reactions.
Adsorbed particulate chemical is removed from the water column by sedimentation
and returned to the water column by scour. Models which are used to evaluate
fate and transport should Include these processes.
Often it is possible to neglect the kinetics of adsorption and desorption
In favor of a local equilibrium assumption. Over the time scales of interest,
this may be a good assumption. Also bed load movement is sometimes small
relative to wash load movement and can be neglected. Under steady state
conditions, net sedimentation rates are often used to simplify the transport
of sedimentation and scour. All of these assumptions have their applications
but should be carefully considered in each model application.
11-14
-------
METHODS
In designing a program to field validate aquatic fate and transport
models for toxic organics substances, special consideration must be given to
selection of a study area. The investigator must be familiar with the input
requirements of the particular model in question, and select candidate study
sites that minimize complexing influences. That is, the least complex study
area that meets the minimal requirements for the model in question should
be selected whenever possible. Consideration must be given to such factors
as: stability of the flow regime; uniformity of the stream channel; sources
of pollutants and inflow within the study area; flow diversions and obstruc-
tions; availability of historical data; accessibility to the stream; and
relative ease in obtaining flow data and required measurements. It is parti-
cularly critical that pollution and dilution influences that are not easily
quantified (e.g., non point sources, ground water inflows, intermittent
inflows, discharges, etc.) are avoided whenever possible. It is also highly
desirable to base parameter field inputs on measured values rather than
calculated values whenever possible. For example, non point source loading
is not easily measured in most situations, but a number of models are avail-
able that predict NFS loading functions and rates. The difficulties asso-
ciated with incorporation of such calculated input values are obvious; they
introduce a potential soure of error in input values of unknown magnitude,
thereby reducing the level of confidence in derived output values.
Table 11-2 presents a partial list of site specific environmental input
values currently required or anticipated as future requirements for aquatic
fate and transport modes such as EXAMS, TOXICS, PEST, etc. The following
addresses field methodologies most appropriate for obtaining the required
field data for each parameter.
COMPARTMENT DIMENSIONS
Length (M)
The length of a stream segment (compartment) can be determined by two
principal methods: 1) measurement from maps of suitable scale and accuracy
(e.g., USGS 15 min. quadrangle maps; or US Coast Guard Navigation Charts); or
2) direct field measurements. Measurements from maps can be accomplished
using map measurer devices or with a set of calibrated map dividers (Welch
1948). Field measurements can be determined through use of steel tapes or tag
lines, by chaining along the bank or with the use of calibrated range
finders.
Width (M)
Stream width measurements can sometimes be determined from maps of
appropriate scale and accuracy following procedures described by Welch (1948)
and Wetzel and Likens (1979). More frequently, however, width measurements
are made in the field from cableways or bridges, from a boat or by wading
following techniques described by Buchanan and Somers (1969). Cableways and
bridges can be marked with paint at measured intervals to indicate distances.
Steel tapes or tag lines marked with solder beads or tags at regular Intervals
11-15
-------
TABLE 11-2. SITE SPECIFIC ENVIRONMENTAL INPUTS TO AQUATIC FATE AND TRANSPORT
MODELS FOR PHYSICAL TRANSPORT
Compartment Dimensions
Length (M)
Width (M)
Height (Depth) (M)
Volume (M3)
Surface Area (M2)
Cross Section Area (M2)
Distance Between Compartment Centers (M)
Water, Sediments, Biota, and Organic Particulate Matter Entering Compartments
Rainfall (mm/h)
Groundwater (m3/h)
Interflow (m^/h)
Nonpoint Source Water volume (M3/h)
Tributary Water inflow (M3/h)
Streamflow (M3/s)
Nonpoint Source Sediments (kg/h)
Tributary Sediment inflow (kg/h)
Suspended Sediments (kg/h)
Bedload Sediments (kg/h)
Biota
Particulate Organic matter
Measurements Within Compartments
Suspended Sediments
Concentration determination (mg/1)
Discharge calculation (kg/D)
Size distribution (%)
Settling Velocity (%)
Organic Matter (%)
Bedload Sediments and Bed Material
Discharge (kg/D)
Size distribution (%)
Organic matter (%)
Bulk density (c/cm3)
Water content (%)
Biota
Water Column
Water Temperature (C)
Dissolved oxygen (mg/1)
Dissolved organic carbon (mg/1)
Total organic carbon (mg/1)
Biota
Eddy diffusivity (M2/hr)
(continued)
11-16
-------
TABLE 11-2. (Continued)
Losses from Compartment
Suspended sediments
Bedload sediments
Water volume
Biota
Particulate organic matter
Evaporation
can be used from boats or when wading to measure distences. The important
factor to keep in mind when making stream width measurements is that stream
width may vary dramatically with water level and longitudinal location in the
channel. In natural stream channels a number of measurements are required to
determine the mean width for the segment or compartment of Interest (Buchanan
and Somers 1969).
Height (Depth) (M)
The height of a given water compartment usually coincides with the depth
of the stream or the designated stratum. Depth measurements are generally
made along several transects within a compartment (segment) using conventional
techniques described by Welch (1948), Buchanan and Somers (1969), and Betson
(1978). In large streams a recording depth finder mounted on a boat simplifies
depth measurements and an analog strip chart recorder provides a permanent
record of depth data.
Volume (M3)
The volume of water within a compartment at any given moment can be
computed from the previously described measurements (L x H x W).
Surface Area (M2)
The surface area of a compartment is determined from the product of length
and width measurements previously discussed. This applies both to water and
benthlc compartments.
Cross Section Area (M2)
Cross sectional area is restricted to water compartment applications and
represents the product of width and mean depth.
Distance Between Compartment Centers (M)
Linear distances between compartment centers are determined from maps of
appropriate scale and accuracy or measured directly in the field following
techniques described for determining stream length.
11-17
-------
WATER, SEDIMENTS AND BIOTA ENTERING COMPARTMENTS
Rainfall (mm/h)
A variety of methods and Instruments are available for measuring rainfall
and other forms of precipitation. Basically, the ordinary rain gage consists
of a funnel that leads into a collector with some sort of measuring device
that is operated manually or automatically. Pogerman (1980) presents detailed
discussions of methodologies for rainfall (and other forms of precipitation)
measurements.
Groundwater (M^/h)
Quantification of groundwater contribution to the total volume of surface
water compartment is generally not attempted in exercises designed for purposes
of validating aquatic fate and transport models. Costs associated with obtain-
ing quantitative groundwater flow data for such purposes are normally prohibi-
tive. However, if considerable historical data are readily available for the
study site or if the means of gathering current data exist (e.g., wells suitable
for sampling are already present) and the model in question can accept such
data, groundwater data could be Included as an additional dimension to the
model validation program.
Darcy's law can be used for determining rates and quantities of water
moving in a large sand and gravel aquifer as follows:
Q - KA £il
dt
dh
^ — J^ra
where: Q • Discharge
K • The coefficient of permeability
A » The cross sectional area of the aquifer
Q. - The hydraulic gradient (Todd 1959).
dt
K is derived mathematically from data collected from pumping a well while
using observation wells to define the drawdown curve.
The feasibility and necessity of attempting to obtain groundwater data
should be carefully evaluated before the decision is made to incorporate
these data into aquatic fate and transport models. If, after all factors are
considered, such efforts can be justified, the investigator should consult
Appel (1980), Stettman (1971), and Reed (1980) and references cited therein
for procedural guidance.
Interflow (M3/h)
Water that infiltrates the soil surface and moves laterally through the
upper soil horizon and enters a stream channel without reaching the water
11-18
-------
table is known as interflow or subsurface flow (Linsley et al. 1949). The
significance of interflow contribution to the total volume of streamflow varies
widely from area to area depending upon soil types and rainfall rates. Consi-
derable differences of opinion exist among hydrologists regarding the qualita-
tive breakdown between groundwater and interflow. Because of the poorly defined
nature of interflow and problem associated with the measurement, interflow and
groundwater are sometimes considered as a single component of the hydrograph,
while surface water is considered as the other component (Linsley et al. 1949).
For purposes of validating aquatic fate and transport models for toxic
organic substances, interflow is best considered as a component of ground-
water. Unless extensive data are available for a particular study reach where
testing is to be conducted, Interflow along with groundwater and nonpolnt
source surface runoff are best treated as a composite gain to streamflow between
upstream and downstream boundaries.
Nonpoint Source Water Volumes (M^/h)
The volume of water and associated pollutants contributed to a water
compartment via nonpoint sources are difficult to quantify owing to nonspecific
(diffuse) nature of nonpoint source contributions. In many watersheds, contri-
butions to waterways from nonpoint sources are negligible except during periods
of rainfall when surface runoff may be substantial. To minimize factors
associated with episodic nonpoint source contributions, selection of the study
area is of vital importance. Potential contributing sources should be avoided
to the extent possible in the event of a significant precipitation event.
Such sources would include stormwater drains, runoff from roadways and paved
surfaces such as parking lots, drainage ditches, and culverts. If it proves
impossible to avoid all sources of nonpoint contributions to a compartment,
the volumes of nonpoint source contributions plus groundwater and interflow
can be computed by determining differences in discharges between upstream and
downstream compartment boundaries, and factoring out point source and tributary
stream inflow contributions. Assuming no substantial loss from the compartment
occurs as a result of diversion, withdrawals, etc., the increases in volume at
the downstream boundary of the compartment can be attributed to the collector
Influences of nonpoint sources groundwater and subsurface flows (interflow).
Tributary Inflow (M3/h)
The volume of water entering a water compartment via tributary streams
can be calculated or measured following procedures described in the Streamflow
section below.
Streamflow Entering Compartment (M^/s)
Streamflow is defined as the discharge in a natural channel (Langbein and
Iseri 1960). Discharge is the volume of water that passes a given point within
a unit of time. Stream discharge values are normally reported in units of
cubic meters per second (M^/S - SI units) or in cubic feet per second (CFS -
English units).
11-19
-------
A value for stream discharge at a site can be obtained: 1) from estab-
lished stream gaging stations; 2) by installing a staff gage and establishing
a rating curve for the reach; or 3) by on-site measurement at the time of
sample collection.
Established Gaging Stations
The most desirable approach is to locate the sampling site near an esta-
blished and operational stream gaging station. Most such stations within the
United States are operated by the United States Geological Survey (USGS).
Although, in general, only mean daily discharge values are published, Instan-
taneous or at least mean hourly discharge values with a known error band can
be obtained by contacting the servicing USGS office. Many gaging stations are
also used as USGS water quality monitoring sites and thus an historic water
quality record may exist. Information on site selection, development and
operation of USGS gaging stations is given by Carter and Davidlan (1968),
Buchanan and Somers (1968), and Buchanan and Somers (1969). Location data and
flow records are published on an annual basis for the Individual states by the
USGS in cooperation with each state. These documents have the general title,
"Water Resources Data for (state name) - Water Year 19 ."
Installation of Staff Gages
If a stream gaging station is not available or the sample collection site
cannot be located at a gaging station for other reasons, it may be desirable
to install a simple stage measurement (staff) gage and develop a rating curve
or stage-discharge relationship for the site. Due to the effort involved,
this approach should be considered only if a long term (years) sampling program
is to be maintained and a very stable channel with good channel control exists
at the site.
The methodology for developing a stage-discharge relationship is described
briefly in the following paragraphs. It is in essence the same approach as
used to establish a stream gaging station except that automatic water level
recording equipment would be incorporated in a permanent gaging Installation.
Detailed descriptions of these procedures are included in Carter and Davidian
(1968) and Buchanan and Somers (1968).
A stage-discharge relationship is an expression of the correlation
(mathematical or graphical) between $age height (stage or depth of water) and
discharge at a point on the stream channel. Development of a reliable stage-
discharge relationship requires careful selection of the channel reach. The
reach should be relatively straight and stable; that is, not prone to signifi-
cant channel shifts or cutting and filling under different flow regimes. In
addition, a good channel control should exist at the outflow end of the reach.
An example of natural control would be a solid rock channel bottom or constric-
tion; artificial control would be provided by concrete bridge abutments, broad
crested weirs, etc. The purpose of channel control is to assure that the
stage-discharge relationship will not be significantly altered over time due
to major channel downcutting, aggradation or by bank erosion causing consequent
channel shifts or size changes.
11-20
-------
Once the channel segment is selected, a simple water level or staff gage
is installed near the stream bank. This allows visual determination of the
elevation of the water surface above some arbitrary datum. Discharge measure-
ments (described below) and corresponding gage readings are then taken under a
variety of flow conditions until enough data pairs are available to define a
graphical relationship. Subsequently, stream discharge values can be obtained
for the site by simply reading the gage height and consulting the constructed
graph.
An advantage of this approach is that in a long-term sample collection
program a stream discharge measurement is not required at the time of each
water quality sample collection. Disadvantages are: 1) the level of "upfront"
effort involved, 2) the low probability that a stable channel and channel
control will be available at a desired water quality sampling site; and 3) the
inherent deficiencies of stage-discharge relationships at flow extremes which
result in significant errors in estimated discharge values under very low or
high flow conditions.
On Site Measurements
In the majority of cases where permanent gaging station data are not
available it will be necessary to physically measure stream discharge at the
time of sample collection. This can be accomplished by current meter measure-
ments on the majority of streams, or by installation of portable weirs or
flumes or by volumetric measurements on very small streams.
The.basic instruments for current or stream velocity measurements are the
Price and pygmy meters. Both of these current meters are basically structured
as cup-type bucket wheels moving on vertical shafts with bearings that operate
in air pockets. The Price meter is appropriate for deep and swiftly flowing
streams, being accurate at velocities up to 6 meters per second (20 fps). The
smaller pygmy meter is accurate at much slower velocities and will function in
streams as shallow as 0.06 meters (2 to 3 inches). Standard procedures for
the operation and maintenance of both meters are given in Buchanan and Somers
(1969) and Smoot and Novak (1968).
The procedures followed in making current measurements are similar for
both types of meters. For very deep and swiftly moving streams the measure-
ments with a Price meter must be conducted from a secured boat or from bridges
or other structures spanning the stream. In streams that can be safely
traversed by wading, the meter (Price or pygmy) is attached to a hand-held
wading rod. A great variety of ancillary equipment is necessary for meter
operation, especially in the case of non-wading measurements. A description
of the various equipment options and functions is beyond the scope of this
discussion but is adequately covered in Buchanan and Somers (1969).
The discharge measurement itself is a summation of the products of par-
tial areas of stream cross sections and their respective average velocities or
Q - (av)
11-21
-------
where:
Q - Total discharge
a - An individual partial cross section area
v » The mean velocity of flow for the area as measured by the current
meter.
To conduct a discharge measurement, a relatively straight reach of stream
should be identified which is free of debris and other flow interferences.
Near the middle of this reach a cross section is selected and by use of a
steel tape or other measurement device is divided into segments of equal
width. The USGS recommends at least 20 segments if feasible; very small
streams would be the obvious exception. At the midpoint of each segment the
water depth is measured and the flow velocity is determined with the current
meter. When all segments in the cross section have been measured they are
summed to produce a value for total stream discharge. Again, this has been
a very simplified description addressed more to the concept than to actual
procedures. A comprehensive description of equipment and standard USGS
procedures for making discharge measurements is given in Buchanan and Somers
(1969).
Discharge of small streams (<0.1 m^/s), which are too shallow or slow
moving for reliable current meter operation can be obtained relatively easily
and quickly by use of portable weirs or flumes or by volumetric measurement.
Small 90 V-notch weirs can be easily fashioned from 10- to 16-gauge sheet
metal. If properly used, such a weir with a 1-foot deep notch will measure
flows from 0.006 to 0.07 m3/s (0.002 to 2.5 CFS) within 3% accuracy. Small
modified Parshall flumes made from lightweight aluminum are available com-
mercially. They are easier to install than weirs and provide comparable
accuracy over a flow range of 0.00003 to 0.014 m3/s (0.001 to 0.5 CFS).
Both devices are Installed by placing them within the stream channel, and
carefully leveling and damming the remaining channel so that all flow is
diverted through them. Measurements are taken by determining the head or
depth of water passing through the rated control section (weir notch or
throat of flume). Discharge values are then obtained from published rating
tables. In very small streams, where the total flow can be diverted into a
bucket or other container of known capacity, the discharge rate can be deter-
mined volumetrically. This "bucket and stopwatch" approach is fairly self-
explanatory. Detailed descriptions of portable weir and flume installation,
operation, and measurement and of the volumetric method are provided in
Buchanan and Somers (1969) and U.S. Bureau of Reclamation (1967).
Nonpoint Source Sediments Entering Compartment (kg/h)
Quantification of NFS sediment loading to stream compartments is at least
as difficult as determining the volumes of water contributed by NFS. By far,
the preferred method of dealing with NFS sediment contribution for purposes of
validating specific fate and transport models is to avoid stream reaches and
watersheds with a high potential for NFS sediment contribution. Sediments
11-22
-------
entering streams, regardless of their origin, are Incorporated into, and trans-
ferred between both sediment and water compartments. Naturally, the finer
sediments with slower settling rates are more likely to remain as components
of the water compartment than the coarser, heavier materials.
Potential nonpoint sources of stream sediments are prevalent in most
watersheds, thus, it is not always possible to select study reaches that are
immune from sediment influences. Careful surveillance of possible study sites
with an eye toward minimizing effects of sediment input, particularly during
periods of heavy rainfall and high discharge, can greatly reduce the compli-
cating effects introduced by sediments. Major contributing sediment sources
should be avoided if possible; these include: 1) stream reaches with large
expanses of easily erodable, exposed, stream banks comprised largely of uncon-
solidated soils; 2) recently tilled agricultural lands especially those planted
to row crops with little or no buffer zone (green belt) along the stream; 3)
areas of disturbed landscape such as ongoing housing development projects,
logging operations, road construction, mining activities Including unstabilized
spoil areas, roadways and parking lots; and A) in western states, particularly,
areas of denuded landscape resulting from overgrazing and the generally sparse
nature of vegetation. If the study reach selected is subject to nonpoint
source sediment influences, it may be possible to estimate the quantities of
sediment Introduced to the stream. Subsequent sampling within the stream may
yield sufficient data to permit computation of the quantities of sediments in
the bedload, in bed material, and in suspension in the water column.
Nonpoint sediment yield can be estimated using equations developed for
specific application. The U.S. EPA (1973) and McElroy et al. (1976) discuss
sediment prediction methods for various agricultural sources. Also, a number
of individual papers using predictive methods are compiled in the proceedings
of the third Federal Interagency Sedimentation Conference of the Water Resource
Council (1976).
Although the majority of sediment loading functions have been developed
primarily for application to cropland, some, e.g., the Universal Soil Loss
Equation (USLE), have application to non-cropland and to some extent to silvi-
cultural, construction, and mining activities (McElroy et al. 1976). The U.S.
EPA (1973) and McElroy et al. (1976) provide sufficient information to decide
the feasibility computing NPS sediment loading functions for a particular
situation, and to select the most appropriate model. The investigator should
be well aware of the potential for introducing errors as a result of incor-
porating data derived through mathematical computations into aquatic fate and
transport models.
Tributary Sediment Inflow (Kg/")
The amount of sediment entering stream compartments per unit time via
tributary inflow can be determined following procedures described for sediment
sampling within compartments. The sampling site is established near the mouth
of the tributary stream in question and the total quantity of suspended and :
bedload materials are computed. It is assumed that all materials in transport
at the mouth are discharged to the receiving stream. Owing to differences in
11-23
-------
flow velocities, total volume of flow, gradient, etc. between tributary
streams and the main receiving stream, the percentage distribution of con-
tributed suspended and bedload sediments will not necessarily remain con-
sistent from tributary stream to the main channel. That is, a large per-
centage of sediment in transport as bedload in the tributary stream may
become suspended components of the water column upon entering the main
channel. Consequently, extrapolation of percentage distribution of sedi-
ment data from tributary to mainstream is not necessarily valid. Sediment
measurements should be made within the stream compartment to determine the
distribution of total sediments between the water column, bedload and stream
bed materials.
Suspended Sediments
The amount of suspended sediment entering a stream sampling compartment
or reach is determined by the same procedures as described for suspended-
sediment discharge. The sample site will, by definition, be located at the
upstream boundary of the sampling compartment. To produce a value more
appropriate for comparison with other water quality variables measured in
the compartment, the suspended sediment discharge may be computed in units
of kg/hour (tons/hour) or kg/min (tons/min). This requires mean hourly or
instantaneous stream discharge values for the rate calculations.
Bedload Sediment
The amount of bedload sediment entering a stream water quality sampling
compartment or reach is best determined by the same computational procedures
as described for bedload sediment discharge. The stream cross section for
data collection will, by definition, be located at the upstream boundary of
the sampling compartment. The bedload sediment discharge values may be com-
puted in units of kg/hour (tons/hour) or kg/min (tons/min) to allow for
meaningful comparison with instantaneous measures of water quality that may
be obtained within the stream compartment.
Biota
Living organisms, both aquatic and non-aquatic, are continuously entering
and leaving water and sediment compartments. The biota are integral components
of aquatic ecosystems and their role in contributing, dispersing, transforming,
and removing organic pollutants to, within, and from ecosystem compartments
is substantial. Procedures for obtaining biomass measurements In aquatic
system are discussed at length in Appendix B.
PARTICULATE ORGANIC MATTER
Like the biota, nonliving particulate organic matter is a component of
all natural aquatic systems serving as an attachment surface and transport
mechanism for pollutants with sorptive tendencies. Particulate organic matter
In water and streambed sediment is typically measured by weight loss through
11-24
-------
ignition at 550°C of previously dried and weighed samples. The organic con-
tribution to the total sample is expressed either as a percentage of the
total dried sample weight or in absolute weight per volume (e.g., mg/1 water)
or wt per wt (e.g., mg/kg sediment) values. Procedures for handling and
analyzing water samples for measurement of particulate organic matter (volatile
residues) are provided in APHA (1980) and U.S. EPA (1974). Methods for col-
lecting and analyzing sediment samples for determining the organic factor are
provided in this report under procedures for measuring percent organic matter
in bed material.
MEASUREMENTS WITHIN COMPARTMENTS
Suspended Sediments
Suspended sediment is that portion of the total stream sediment load that
stays in suspension for appreciable lengths of time. Suspended-sediment dis-
charge is the quantity of suspended sediment passing through a given stream
cross section per unit of time. This discharge is usually expressed in units
of kilograms per day (Kg/D-SI units) or in tons per day (T/D-English units).
The procedure for determining suspended-sediment discharge can be considered
in the following segments: 1) site selection; 2) sample collection; 3) deter-
mination of concentrations; and A) discharge calculations. Each of these com-
ponents is discussed individually in the following paragraphs.
Site Selection
Stream discharge is required to calculate the suspended-sediment discharge;
therefore, the site would ideally be located at or near a gaging station. If
this is infeasible, then an actual measurement of stream discharge will be
required concurrent with collection of the suspended-sediment sample. Other
site criteria are:
a. uniform channel with no major flow restrictions;
b. stable banks which are not being actively eroded;
c. lack of significant eddies or backwater effects;
d. lack of significant tributary inflows immediately upstream; and
e. existence of a bridge or other base for sampling if the stream is
not wadable. This latter criterion often conflicts with the
desirability of not having major flow restrictions in the channel.
Sample Collection
Collection of suspended-sediment data requires a sampling device that
will obtain a sample representative of the water-sediment mixture moving in
the stream in the vicinity of the sampler. In addition, the sampler must be
capable of collecting a composited vertical profile of streamflow. Vertical
profiling is necessary because of the gravity effect on suspended sediment
which causes fines to predominate in the upper levels of streamflow and coarser
material near the stream bottom.
Several types of sediment sampling devices which meet the above criteria
have been developed. For the purposes considered In this report, only standar-
dized suspended-sediment samplers used by the USGS should be employed. Two
11-25
-------
major types are available: depth-integrating and point-integrating. Depth-
integrating samplers collect and accumulate a sample as they are lowered to
the bottom of the stream and raised back to the surface. The sampler must be
moved at a uniform rate in a given direction but not necessarily at equal
rates in both directions. Sampling depth is limited to approximately 4.6
meters (15 feet). The point-integrating sampler can be operated to obtain a
depth-integrated sample in deep streams by manually opening and closing "a
sampling valve to integrate the stream depth in parts.
Both types of samplers are basically a streamlined metal casting with a
hollowed interior for the sample container. An opening for interchangeable
nozzles is located at the front of the samplers and leads to the internal
sample container. Openings for air-exhaust from the sample container are also
provided. The standard sample container is a 473 ml (1-pint) glass "milk"
bottle. For certain sampling conditions a 946 ml (1-quart) sample bottle is
used. The samplers are available in a variety of sizes and weights, the use
of which is dependent upon stream conditions. The small and lightweight models
are appropriate for sampling in wadable streams. The larger models, which
weigh up to 91 kg (200 pounds), are used for cable-suspended sampling from
bridges of very deep and/or fast moving streams.
A detailed description of suspended-sediment samplers and sampling proce-
dures is contained in Guy and Norman (1970) and Vanoni (1975). The major
concern in sampler operation is to move the sampler at such a rate that the
sample container becomes filled, or nearly so, just at the end of one or more
vertical sampling trips. The major deficiency of existing samplers is their
inability to collect a sample of water-sediment mixture near the streambed. A
zone within 9 to 15 cm (3.5 to 6 inches) of the stream bottom is not sampled
because of the need to keep the sample nozzle out of the bottom sediments.
Coarse grained sediment is prevalent in this zone. Therefore, the concentra-
tion of suspended-sediment samples is usually lower than the true suspended-
sediment concentration. For this reason the suspended-sediment discharge
computed from samples is called measured suspended-sediment discharge.
Reliable determination of suspended-sediment discharge requires that the
sample(s) be representative of the total stream cross section. The sampling
methodologies to achieve this can range from simple to fairly complex. For
streams with a stable channel and a uniform lateral suspended-sediment concen-
tration, sampling at a single vertical will usually be adequate. However, the
location of that "mean" vertical must be determined by trial multi-vertical
sampling. For streams with unstable channels or with highly variable sediment
loads, multi-vertical sampling (commonly 4 to 20 sites in a cross section)
should be conducted at all times. Two general approaches for multi-vertical
sampling are used by the USGS: 1) method of centroids-of-equal discharge
increments (EDI) across the stream; and 2) method of equally spaced verticals
across the stream with an equal sampling transit rate (ETR) at all verticals.
A detailed explanation of these methods is contained in Guy and Norman (1970).
Both methods produce results of comparable accuracy. The ETR method has an
advantage in that the Individual samples can be composited for laboratory
analysis of suspended-sediment concentration. In addition, It is a more intui-
tive method which can be quickly and easily implemented by field personnel.
11-26
-------
Concentrations Determination
After field collection, samples are shipped to an appropriate laboratory
for determination of suspended-sediment concentration. No special procedures
for sample handling are necessary. The most important requirements are that
the bottles be adequately sealed and labeled with detailed site and collection
procedure information. •
The common unit for expressing suspended-sediment concentration is milli-
grams per liter (mg/1). This is computed as one million times the ratio of
the dry weight of sediment in grams to the volume of water-sediment mixture in
cubic centimeters. Because it is more convenient to obtain the weight of the
water-sediment mixture (the sample) than its volume, the following formula
involving parts per million (ppm) is generally used.
. c weight of sediment x 1,000,000
weight of water-sediment mixture
where: C is a variable adjustment factor that accounts for the fact that one
mg/1 is not equal to one ppm at concentration levels greater than
16,000 ppm.
Given the purpose of sampling considered in this report, only the mean
suspended-sediment concentration for each stream cross section will be required.
Therefore, a single composite sample can be analyzed rather than individual
samples from all verticals. Samples collected by the EDI method can be
composited only If individual sample volumes are equal; samples collected by the
ETR method must be composited, either physically or arithmetically.
The two most common and accepted methods for determining the weight of
sediment in samples are evaporation and filtration. The filtration method is
faster but is not amenable to samples with high suspended-sediment concentration
because of filter clogging. The evaporation method requires an adjustment for
dissolved solids if the dissolved load is high In comparison to the sediment
load. A detailed description of laboratory procedures for both methods is
given in Guy (1969).
Discharge Calculations
The basic method for calculating the average suspended-sediment discharge
at a stream cross section is
Qs - <*» cs k
where :
Qs - The sediment discharge in kg/day (tons/day)
0 - The stream discharge in M3/S (CFS)
11-27
-------
Cs » The mean concentration of suspended sediment in the stream cross
section in mg/1 (ppm)
k » A coefficient to adjust units.
More complex and presumably accurate methods are available to compute suspended-
sediment discharge. All methods are described in Porterfield (1972).
Size Distribution
Size distribution of suspended sediment refers to the breakdown of a
suspended-sediment sample into various size classes. The results are usually
expressed as percentages or proportions by weight.
Suspended-sediment samples for particle size analysis are collected with
the same equipment and in the same manner as samples for determination of
suspended-sediment concentration and discharge. The same criteria, for site
selection and single vertical vs. multi-vertical sampling, hold. Therefore,
collection of duplicate bottles at each sampling vertical is recommended if
both discharge and size distribution of suspended sediment are to be determined.
Also, because the primary focus is on average particle size distribution for
the entire stream cross section, multi-vertical samples collected in a cross
section should be composited, when possible, prior to laboratory analysis.
There are two generally accepted methods for separation of suspended-
sediment samples into size fractions—separation by sieving and separation
according to fall velocity in still water. The size range of material that
can be analyzed by the methods is different. Therefore, the anticipated size
range and the sampling objectives must be considered to determine which indivi-
dual or combination of methods is appropriate for a given sample.
Sieve analysis can be performed on either wet or dried sediment samples.
The wet-sieve method is preferred. Sieving is limited to size classes coarser
than a 0.0625 mm, which Is material retained on a 250 mesh sieve. If the
distribution of finer particles is desired, the residual from the 250 mesh
screen must be analyzed by one of the settling velocity techniques described
below.
Although sieves are available which will classify material up to 16-32 mm
(coarse gravel), a standard sieve analysis will result In determination of the
following size classes.
Class Name Size Range (mm)
Very fine sand 0.052 - 0.125
Fine sand 0.125 - 0.250
Medium sand 0.250 - 0.500
Coarse sand 0.500 - 1.000
Very coarse sand 1.000 - 2.000
Results from a sieve analysis should be reported in percent to the nearest
whole number. A detailed description of sample preparation is given in Guy
11-28
-------
(1969). Standard ASTM procedures for conducting a sieve analysis should be
followed and are described In ASTM (1938).
Settling Velocity
A number of methods to determine particle size distribution based on
differences in particle fall velocities have been developed. The methods
recommended by the USGS are: 1) pipet; 2) visual-accumulation tube (VA tube);
and 3) bottom-withdrawal tube (BW tube). Concentration of sediment in samples
has a considerable effect on the accuracy of results. Prior to analysis,
samples should be split or diluted to meet the following criteria.
Method Concentration in Tube (mg/1)
Pipet 2,000 - 5,000
BW tube 1,000 - 10,000
VA tube total sediment not to exceed 12 cm
of tube height
The VA tube method is valid only for material in the size range 0.062 to
2.0 mm and thus is comparable to sieve analysis. The pipet and BW tube methods
are valid only for sediments finer than a 0.062 mm. The following size classes
will normally be determined by these methods.
. . . Class Name Size Class (mm)
Coarse clay 0.002 - 0.004
Very fine silt 0.004 - 0.998
Fine silt 0.008 - 0.016
Medium silt 0:016 - 0.031
Coarse slit 0.031 - 0.062
Results from all particle fall velocity methods should be reported in percent
to the nearest whole number. A description of laboratory procedures for these
methods Is given in Guy (1969).
Organic Matter
Organic matter in suspended sediment ranges from macroscopic plant
material to microscopic colloidal humus. In the majority of streams, organic
material will constitute a small percentage of total suspended sediment load.
However, for certain water quality applications it may be Important to know
the actual amounts present. In addition, quantitative determination of organic
content of suspended sediment is recommended for samples analyzed for particle
size distribution by fall velocity methods. This is because of the effect the
lower specific gravity organic material has on settling velocities. This
issue is further discussed in Guy (1969).
Measurement of percent organic matter should be performed on additional
duplicate bottles collected during sampling for determination of suspended-
11-29
-------
sediment discharge and particle-size distribution. The recommended methodology
for quantifying organic matter content involves weighing of a dry sediment
sample followed by removal of the organic matter and re-weighing of the remain-
ing sediment. The weight percentage of organic matter is then determined by
difference. Three techniques are available for removal of organic matter
from a sediment sample: 1) hydrogen peroxide oxidation; 2) specific gravity
separation; and 3) combustion. Laboratory procedures for each of the tech-
niques are described in Guy (1969). ;
BEDLOAD SEDIMENTS
Discharge
For a variety of reasons, physical measurement of true bed load is diffi-
cult. First, many mechanical sampling devices, placed on or within the stream-
bed, will decrease the velocity at the sampler and thus decrease the rate of
bedload movement. Another major difficulty is that the velocity of water near
the streambed varies greatly both in time and space. This results in bedload
particles moving intermittently and at an average velocity which is much less
than that of the stream flow. This behavior makes it very improbable that any
given sample will be representative of long term conditions at the sample
point or of conditions at other points in a stream cross section. More
detailed explanations of the problems associated with bed load sampling are
given in Guy and Norman (1970) and Hubbell (1964).
Despite these inherent problems, there is a long history of attempts to
develop bed load sampling devices that would be reliable yet simple enough for
field application. Many of the designs tested are described in Hubbell (196A).
One of the most frequently used devices is the Helley - Smith (1971) pressure-
difference sampler. Research is continuing, but to date no bed load sampling
devices have been developed which are routinely used or recommended for use by
the USGS.
In lieu of direct measurement, standard USGS procedures are to compute
bed load movement or discharge through use of one of several mathematical
formulas. The formulas in general use are based primarily on semi-empirical
considerations of stream hydraulics and particle behavior. As such, a compre-
hensive description of the individual formulas is beyond the scope of this
report. The methods are described in detail in Einstein (1950), Colby and
Hembree (1955), and Colby and Hubbell (1961) and Brooks (1965).
The Einstein procedure (Einstein 1950) was developed in 1950 as a method
for computing the total discharge of sediment of the sizes found in appreciable
quantities in the streambed. The method is based on both theoretical considera-
tions and experimental findings and requires hydraulic data from cross-sections
within a reach and particle size data. A modified Einstein procedure was sub-
sequently developed (Colby and Hembree 1955) which computes the total discharge
charge of all particle sizes (the total sediment load) in the stream and
requires a single depth-integrated suspended load sample. The formulas in the
modified procedures are based on such easily measurable quantities as the
concentration and particle size distribution of suspended sediment, the particle
size distribution of streambed material, and the mean velocity of streamflow.
11-30
-------
The basic equation in the modified procedure for computing the rate of
bed load discharge (in tons/day) of particles of a given size range (igQa) is
IgQg - ibW(43.2)(l,200)D3/2 $/2
where: 1^ « the fraction by weight of the bed material in the size range
U • the width of the stream, in feet
D - the geometric mean size (the square root of the product of the
upper and lower sizes of the range), in feet
$/2 » the Intensity of bedload transport
As suggested by the equation, computations may be performed for several
ranges of particle sizes and the results summed to achieve a value for total
bed load discharge. To reduce this computational requirement, another version
of the modified Einstein procedure has been developed which replaces many of
the numeric calculations with graphical methods (Colby and Hubbell, 1961).
An earlier but widely utilized method for determining total bed load
discharge per unit of stream width is summarized by the formula (DuBoys, 1879).
where qg • bedload discharge, N/m-sec or lbf/ft-sec
•& - 0.17/d3/^ » bed-material discharge coefficient, m3/N-sec or
ft3/lbf-sec
2 2
TO » yHS • shear stress at the water-bed boundary, N/m or Ibj/ft
i\ f\
TC • critical shear stress, N/m or Ibj/ft
Y • specific weight of the fluid, lbf/ft3
d » median particle diameter, mm
H * mean depth, m or ft
S » energy slope, m/m or ft/ft
This procedure requires the following field data inputs: 1) the flow
depth and slope; 2) characteristics of bed material, specifically the median
particle diameter 3) the critical shear stress of the bed material; and 4) the
specific weight, of the transporting water.
Total Load
The total sediment load includes bed load plus suspended sediment load.
Many formulas exist, including a modification of the Einstein method (Colby
and Hembree, 1955), three formulas which have been calibrated with flume and
11-31
-------
field measurements (Engelund-Hansen, 1966, 1967; Inglls-Lacey, 1968; and
Toffaletti, 1968, 1969).
The Engelund-Hansen formula requires field measurements on the specific
weight of bed sediment, the mean stream velocity, the median diameter of bed
sediment, the specific weight of the water, and the bed shear stress. It
yields the total sediment discharge per unit width of stream.
*
q,. - 0.05
" " ' ' " \(Y.-Y>d/
where qt » total sediment discharge, Ib/ft-sec
Ys • specific weight of sediment, Ib/ft
Y « specific weight of water, Ib/ft^
u • mean stream velocity, ft/sec
d - median bed particle diameter, ft
TO - bed shear stress, Ib/ft
g - gravitational constant - 32.2 ft/sec2
This equation was calibrated with data from a very large flume (8 ft wide x
150 ft long) with sediments larger than 0.19 mm median diameter. The authors
do not recommend it for use on fine bed materials, less than 0.15 mm.
The Inglis-Lacey formula relates total sediment discharge to the mean
velocity of the stream to the fifth power.
(ug)l/3 U2 .^3
qt - 0.562
w gH g
where qt - total sediment discharge per unit width, Ib/ft-sec
v • kinematic viscosity of fluid, ft 2/sec
u - mean stream velocity, ft/sec
g - gravitational constant - ft/sec2
H - flow depth, ft
Y " specific weight of water, lbf/ft
W - fall velocity of median size bed particle, ft/sec
This equation was calibrated with field data from large, stable irrigation
canals.
Toffaletti (1968, 1969) divided the stream into four depth zones for the
purpose of calculating a total sediment discharge. The bed sediment is divided
into standard size fractions. Hydraulic quantities which are required Include
the temperature and viscosity of the water, mean stream velocity, hydraulic
radius, energy slope, stream width, and bed shear velocity. The formula was
based on extensive data from seven rivers and flume data from four investiga-
tors. Most of the transport was due to sand in these instances. To date,
there is not a verified formula which describes total sediment transport of
silts and clays.
11-32
-------
Size Distribution (X)
Size distribution of bedload sediment would be defined as the percentage
breakdown of a bedload sediment sample into various particle size classes.
However, as discussed in the section of this report on determining bedload
discharge, obtaining representative samples of true bedload is improbable at
best. For this reason, size distribution analysis of streambed sediments are
normally performed on samples of what is termed "bed material." Bed material
is defined as the mixture of streambed sediment that is exposed to the errosive
force of flow and which can, or may move during any relatively short period of
time. This material is located from the surface of the streambed to a depth
of approximately 3 to 20 cm (1 to 8 inches). Bed material and bedload are not
synonymous. Bed material will normally be coarser grained than true bedload
and the exclusion of the finer grained sediments that may bounce or roll along
the bed.
Three basic types of bed-material samplers are recommended and used by
the USGS: 1) vertical; 2) rotating bucket and scoop; and 3) surface samplers.
Details of design and operation of these samplers are contained in Guy and
Norman (1970).
Vertical samplers consist of a hand-held cylindrical tube which is forced
into the streambed and retrieved with the sample core retained by suction.
Collection of a sample up to 15 cm (6 inches) in depth is possible in fine
grained bottom sediments. However, only the top 25 mm (1 inch) of sample is
normally retained for particle size analysis. This procedure is followed to
make the sample more closely representative of true bedload.
The rotating bucket scoop sampler is available in two different size-
weight configurations for varied stream depth and/or velocity conditions.
Both models are designed for cable-suspended sampling of non-wadable streams.
They consist of a streamlined metal bomb with a spring loaded rotating scoop
built into the bottom surface. When the sampler encounters the streambed, the
scoop is released and rotates around and upwards into the body of the sampler.
Both models sample to a depth of about 50 mm (2 inches). All of the sample
must be retained for particle-size analysis because of the mixing which occurs
during the scooping process.
The streambed surface sampler used by the USGS consists of a small circular
disk filled with petroleum jelly mounted on the bottom of a cable-suspended
sounding weight of appropriate size for stream flow and depth conditions.
Upon contact with the streambed, the surface particles adhere to the Jelly.
The sample is then retrieved and sent in a sealed container to a laboratory
where the sediment is separated from the jelly.
It should be noted that all of the described samplers are designed to
recover particles in the silt to coarse sand size range. Effective sampling
of larger particles (gravel, cobbles, and boulders) is difficult and no reliable
mechanical sampler exists. The need to consider these larger particles should
be evaluated for each individual site in view of sampling program objectives.
If a sample is required, gravel and larger material may be collected by scoop
11-33
-------
sampling or manual picking within a measured area large enough to provide a
representative sample. There is no acceptable methodology for sampling coarse
particles in non-wadable streams.
Standard containers for storage and transport of samples from all of the
described mechanical or manual sampling methods are wide-mouthed plastic
bottles or waxed cardboard containers. No special sample handling procedures
are necessary beyond assuring that the containers do not leak or spill during
transit.
Procedures for bed-material sampling for particle size analysis are com-
parable to sampling procedures for suspended sediment. Basically, a sufficient
number of individual verticals or sites must be sampled on a cross section to
adequately represent the distribution of particle sizes over the entire channel.
A large number of individual samples (5 to 15) may be appropriate in an unstable
mixed silt, sand, and gravel channel. Depending upon channel width, 3 to 5
individual samples may be adequate for streams with fairly stable and homoge-
neous bottoms. The depth of samples taken at all individual sites within a
cross section should be constant. Following sample collection, individual
samples may be composited or analyzed for size distribution separately, depend-
ing upon the objectives of the data collection program.
Organic Matter in Bed Material
Measurement of the amount of organic matter in stream bed material should
be performed on a set of duplicate samples collected concurrently with those
for determination of particle size distribution. Equipment and procedures for
collection of these samples are outlined in the preceding section of this
report.
The recommended methodology for quantifying organic matter content involves
weighing of a dry bed-material sample followed by removal of the organic matter
and re-weighing of the remaining sediment. The weight percentage of organic
matter is then determined by difference. Three techniques are available for
removal of organic matter from the sediment sample: 1) hydrogen peroxide
oxidation; 2) specific gravity separation; and 3) combustion. Laboratory
procedures for the listed techniques are described in Guy (1969).
Bulk Density (g/cm3)
Bulk density of soil (or sediments) is the ratio of the mass to the bulk
or macroscopic volume of soil particles plus pore space in a sample (Black et
al. 1965). The mass is determined after drying to constant weight at 105°C,
and the volume is that of the sample as collected in the field. For sediment
samples in which the void spaces are saturated with water, W, density, D, is
expressed as the ratio of the mass (both solids and water) for a unit volume,
V, of sediment (Guy 1978).
Thus, D -
Where D - Density
11-34
-------
Ws • Mass of dried soil or sediment (g)
Wu - Mass of water (g)
V - Total volume (cm^).
Water Content
Moisture content of bottom sediment is determined by weighing a fresh
sediment sample, drying the sample to a constant mass at 105eC-110°C and
reweighing the dried sample (Guy 1978). Thus, percent moisture on a dry
weight basis (Md) is computed as follows:
Md - ws ~ ds x 100,
wds
where: Wws - Mass of wet sediment
W^g • Mass of dry sediment.
Collection of samples for determination of moisture content follows
procedures described for sampling the bedload and bed materials. Since the
measure is based on loss of moisture through evaporation, the sediment sample
must be stored in air tight container to prevent moisture loss.
Biota
The biota associated with the streambed (benthic macroinvertebrates,
periphyton, rooted aquatics) are important components of these ecosystem.
Procedures for obtaining biomass measurements of these components of the
community are described in the section on Bioconcentratlon and in Appendix B.
WATER COLUMN
Water Temperature (°C)
Water temperature measurements are determined following procedures
described in the Biotransformation section.
Dissolved Oxygen (mg/1)
Procedures for collecting, handling, and analyzing water samples for
dissolved oxygen determination are described in the Biotransformation section.
Dissolved Organic Carbon (mg/1)
Procedures for collecting, handling, and analyzing water samples for
dissolved organic carbon determination are described in the Photolysis section
of this report.
11-35
-------
Total Organic Carbon
Total Organic Carbon (TOC) is determined using procedures and materials
outlined in APHA (1980), pp. 471-474. An infrared analyzer detects carbon
concentrations of from 1 to 150 mg/1. Greater or lesser concentrations can be
measured by appropriate dilution or concentration of samples. Because the
carbon analyzer measures all carbon present In the sample, steps must be taken
to separate the inorganic carbon. This is accomplished either by acid decom-
position and volatilization of the inorganic carbonates prior to analysis, or,
if the analyzer is equipped to measure carbonate and bicarbonate carbon, by
direct measurement.
Water samples for TOC analysis should be collected and stored in glass
bottles, preferably brown, and sealed with Teflon-lined caps. Samples should
be analyzed promptly, but if delay is unavoidable, store samples at ice tem-
perature with minimal exposure to light and atmosphere or acidify with
hydrochloric acid to a pH not over 2 (APHA 1980).
Biota
Procedures for obtaining quantitative estimates of the biomass associated
with the water column (e.g., phytoplankton, fish, zooplankton, drift organism,
floating and rooted macrophytes, etc.) are discussed in depth in the Biocon-
centration section and Appendix B.
Eddy Diffusivity (M2/h)
Eddy diffusivity (turbulent diffusion) refers to mixing of dissolved and
fine particulate substances due to micro-scale turbulence. Determination
of eddy diffusivity is discussed in Berner (I960), Fisher (1979), Yotsukura and
Cobb (1972), Yotsukura and Sayre (1976), Edinger and Geyer (1965), and Schnoor
and Fruh (1979).
LOSSES FROM COMPARTMENT SEDIMENTS
Suspended Sediments
The amount of suspended sediment leaving a stream sampling compartment or
reach is determined by the same procedures as described for suspended-sediment
discharge. The sample site will, by definition, be located at the downstream
boundary of the sampling compartment. If significant tributary inflows do not
occur within the compartment and if the stream channel is stable, it may be
assumed that the suspended-sediment discharge will be equal to that entering
the compartment. If these conditions are not met, sampling at both upstream
and downstream boundaries of the compartment may be appropriate.
Bedload Sediments
The amount of bedload sediment leaving a stream water quality sampling
compartment or reach is best determined by the same computational procedures
as described for bedload sediment discharge. The stream cross section for
data collection will, by definition, be located at the downstream boundary of
11-36
-------
the sampling compartment. If significant tributary inflows do not occur within
the compartment and if the stream channel is stable it may be assumed that the
outflow bedload sediment discharge will be equal to that entering the compart-
ment. If these conditions are not met, separate determinations of bedload
sediment discharge at both upstream and downstream boundaries of the compart-
ment may be appropriate.
Water Volume (Streamflow M^/sec)
Procedures for determining the volume of water leaving a stream compart-"
ment are described under the Streamflow section. This determination is a total
discharge measurement and is synonomous with the volume of water entering a
compartment, except discharge is computed at the downstream compartment
boundary.
Biota
Living organisms are continuously in transit into, within and between
compartments. Procedures for estimating the biomass lost through voluntary
and involuntary movement are discussed in the Bioconcentration section and
Appendix B.
Particulate Organic Matter
Particulate organic matter associated with the stream and suspended in
the water column Is subject to bedload and Streamflow movement through the
compartment. Procedures for sampling and measuring particulate organic matter
in transit are described in previous sections.
Evaporation
Loss of water to the atmosphere may represent a substantial portion of
the total water budget, particularly in waters of the southwestern U.S. Perez
et al. (1974) point out that evaporation losses from surface water result in:
1) changes in concentrations of channel constituent due to changes in water
volume; 2) losses of non ionic substances (e.g., ammonia); and 3) changes in
the heat content which affects chemical and biological reaction rates.
Procedures for obtaining evaporation estimates in the field involve three
phases (Perez et al. 1974): 1) Pan evaporation losses are measured and
recorded; 2) the basic wide evaporation losses are estimated from measured
values of all other variables in the water budget; and 3) a pan coefficient
(proportionality factor) is found that will best relate the two results
(Thornthweite and Mather 1955, Eagleson 1970, and Kohler 1954).
11-37
-------
REFERENCES
APHA. 1980. Standard Methods for the Examination of Water and Wastewater.
Fifteenth Edition. APHA/AWWA/WPCF. Washington, D.C.
Appel, C. A. (Chairman). 1980. Work Group on Ground Water. National Hand-
book of Recommended Methods for Water-Data Acquisition. Chapter 2:
Groundwater. Office of Water Data Coordination, U.S. Geological Survey.
Reston, Virginia.
ASTM. 1938. Designation C136-38T, Tentative Method of Test for Sieve Analysis
of fine and Coarse Aggregates: Proc., V. 38-1, p. 808-810.
Black, C. A., D. D. Evans, J. L. White, L. E. Ensminger, and F. E. Clark eds.
1965. Methods of Soil Analysis. Volumes 1 and 2. Amer. Soc. Agron.
Madison, Wisconsin.
Berner, R. A. 1980. Early Diagenesis: A Theoretical Approach. Princeton
University Press.
Betson, R. P. (Chairman). 1978. Work Group on Physical Basin Characteristics
for Hydrologic Analyses. National Handbook of Recommended Methods for
Water-Data Acquisition. Chapter 7: Physical Basin Characteristics for
Hydrologic Analyses. Office of Water Data Coordination, U.S. Geological
Survey, Reston, Virginia.
Brooks, N. H. 1965. Calculation of Suspended Load Discharge from Velocity,
and Concentration Parameters. Proceedings of the Federal Interagency
Sedimentation Conference, Miscellaneous Publication No. 970. Agricultural
Research Service, U.S.D.A., Washington D.C.
Buchanan, T. J. and W. P. Somers. 1968. Stage Measurement at Gaging Stations:
Techniques of Water-Resources Investigations of the United States Geo-
logical Survey. Book 3, Chapter A7. U.S. Geological Survey, Arlington,
Virgina.
Buchanan, T. J., and W. P. Somers. 1969. Discharge Measurements at Gaging
Stations: Techniqus of Water-Resources Investigations of the United States.
Book 3, Chapter A8. U.S. Geological Survey, Alexandria, Virginia.
Carter, R. W., and J. Davldian. 1968. General Procedure for Gaging Streams:
Techniques of Water-Resources Investigations of the United States Geolo-
gical Survey. Book 3, Chapter A6. U.S. Geological Survey, Arlington,
Virginia.
Chapra, S. C. 1977. Total Phosphorus Model for the Great Lakes. Journal of
the Environmental Engineering Division, American Society of Civil
Engineers, 103:147-161.
Colby, B. R., and C. H. Hembree. 1955. Computations of Total Sediment Dis-
charge Niobrara River near Cody, Nebraska: U.S. Geological Survey
Water-Supply Paper 1357.
11-38
-------
Colby, B. R., and D. W. Hubbell. 1961. Simplified Methods for Computing
Total Sediment Discharge with the Modified Einstein Procedure: U.S.
Geological Survey Water-Supply Paper 1593.
DiToro, D. M., D. J. O'Connor, R. V. Thomann, and J. P. St. John. 1982.
Simplified Model of the Fate of Partitioning Chemicals in Lakes and
Streams. In: Modeling the Fate of Chemicals in the Aquatic Environment,
K. L. Dickson, A. W. Maki, J. Cairns, Jr., Eds. Ann Arbor Science
Publishers, Inc.
Duboys, P. 1879. LeRohne et les Rivieres a Lit Affouillable. Annales des
Fonts et Chaussees, Series 5, 18: 141-195.
Eagleson, P. J. 1970. Dynamic Hydrology. McGraw Hill. New York.
Edinger, J. E. and J. C. Geyer. 1965. Heat Exchange in the Environment.
Cooling Water Studies for the Edison Electric Institute. Johns Hopkins
University, Baltimore, MD.
Einstein, H. A. 1950. The Bed-Load Function for Sediment Transportation in
Open Channel Flows: U.S. Dept. of Agriculture Tech. Bulletin No. 1026.
Elder, J. W. 1959. The Dispersion of Marked Fluid in Turbulent Shear Flow.
Journal of Fluid Mechanics, 5(4):544-560.
Engelund, F. and E. Hansen. 1966. Investigations of Flow in Alluvial Streams.
Acta Polytech. Scand. Ci-35.
Engelund, F. and E. Hansen. 1967. A Monograph on Sediment Transport in
Alluvial Streams. Teknisk Forlag, Copenhazen, Denmark.
Fischer, H. B. 1967. The Mechanics of Dispersion in Natural Streams.
Journal of the Hydraulics Division, American Society of Civil
Engineers, HY6, 93:187-216.
Fischer, H. B. 1968. Methods for Predicting Dispersion Coefficients in
Natural Streams, with Applications to Lower Reaches of the Green and
Duwamish Rivers Washington. U.S. Geological Survey Professional
Paper 582-A.
Fischer, H. G., E. J. List, R. c. Y. Koh, J. Imberger, and N. H. Brooks. 1979.
Mixing in Inland and Coastal Waters. Academic Press.
Guy, H. P., and V. W. Norman. 1970. Field Methods for Measurement of Fluvial
Sediment: Techniques of Water-Resources Investigations of the United
States Geological Survey. Book 3, Chapter C2. U.S. Geological Survey,
Arlington, Virginia.
Guy, H. P. (Chairman). 1978. Work Group on Sediment. National Handbook of
Recommended Methods for Water-Data Acquisition. Chapter 3: Sediment.
Office of Water Data Coordination, U.S. Geological Survey, U.S. Department
of the Interior, Reston, Virginia.
11-39
-------
Guy, H. P., and V. W. Norman. 1970. field Methods for Measurement of fluvial
Sediment: Techniques of Water-Resources Investigations of the United
States Geological Survey. Book 3, Chapter C2. U.S. Geological Survey,
Arlington, Virginia.
Helley, E. J. and Smith, W. 1971. Development and Calibration of a Pressure-
Difference Bed Load Sampler. United States Geological Survey, Water
Resources Division, Open File Report, Menlo Park, California.
Hubbell, D. W. 1964. Apparatus and Techniques for Measuring Bedload: U.S.
Geological Survey Water-Supply Paper 1748.
Inglis, C. C. 1968. Discussion of: Systematic Evaluation of River Regime,
by C. R. Neill and V. J. Galey. Journal of Waterways and Harbors
Division, ASCE, 94, WW1: 109-114.
Jabson, H. E. and W. W. Sayre. 1970. Vertical Transfer in Open Channel Flow.
Journal of the Hydraulics Division, American Society of Civil Engineers,
HY3, 96:703-724.
Karickhoff, S. W., D. S. Brown, and T. A. Scott. 1979. Sorption of Hydro-
phobic Pollutants on Natural Sediments. Water Research, 13:421-428.
Kohler, M. A. 1954. Lake and Pan Evaporation. In: Water Loss Investigations
- Lake Heper Studies. U.S. Geological Survey Prof. Paper No. 268.
Krone, R. B. 1962. Flume Studies of the Transport of Sediment in Estuarial
Shoaling Processes. Hydr. Engr. Lab., University of California,
Berkeley, California.
Langbein, W. B., and K. T. Iseri. 1960. General Introduction and Hydrologic
Definitions: U.S. Geological Survey Water-Supply Paper 1541-A.
Linsley, R. K., M. A. Kohler. and J. L. H. Paulhus. 1949. Applied Hydrology.
McGraw-Hill Book company, Inc., New York.
McElroy, A. D., S. Y. Chiu, J. W. Nebgen, A. Aleti, and F. W. Bennett. 1976.
Loading Functions for Assessment of Water Pollution from Nonpoint sources.
EPA600/2-76-151. U.S. Environmental Protection Agency. Midwest Research
Institute, Kansas City, Missouri, Office of Research and Development,
Washington, D.C.
Mortimer, C. H. 1941. The Exchange of Dissolved Substances Between Mud and
Water in Lakes. Journal of Ecology, 29:280-239.
Okoye, J. D. 1970. Characteristics of Transverse Mixing in Open-Channel
Flows. Report No. KH-R-23. W. M. Kech Laboratory of Hydraulics and
Water Resources, California Institute of Technology, Pasadena, California.
11-40
-------
Onishi, Y., and S. V. Wise. 1979. Mathematical Model, SERATRA, for Sediment
and Pesticide Transport In Rivers and its Application to Pesticide
Transport in Four Mile and Wolf Creeks in Iowa. Prepared for the U.S. EPA
under contract 23111-03243. Battelle Pacific Northwest Laboratories,
Richland, Washington.
Partheniades, F. 1965. Errosion and Deposition of Cohensive Soils. Journal
of the Hydraulics Division, ASCE, 91, HY1:105-139.
Perez, A. I., W. C. Huber, J. P. Heaney, and E. E. Pyatt. 1974. A Water
Quality Model for a Conjunctive Surface-Groundwater System. EPA-600/5-74-
013. Dept. Environ, engineering Sciences, University of Florida, Gaines-
ville. U.S. Environmental Protection Agency, Office of Research and
Development, Washington, D.C.
Pogerman, W. I. (Chairman). 1980. Work Group on Hydrometerological Observa-
tions. National Handbook of Recommended Methods for Water-Data Acquisi-
tion. Chapter 10: Hydrometeorological Observations. Office of Water
Data coordination, U.S. Geological Survey, Reston, Virginia.
Porterfield, G. 1972. Computation of Fluvial-Sediment Discharge: Techniques
of Water-Resources Investigations of the United States Geological Survey.
Book 3, Chapter C3. U.S. Geological Survey, Arlington, Virginia.
Reed, J. E. 1980. Type Curves for Selected Problems of Flow to Wells in
Confined Aquifers. Techniques of Water-Resources Investigations of the
United States Geological Survey. Book 3, Chapter 63. U.S. Geological
Survey, Arlington, Virginia.
Rubey, W. W. 1933. Settling Velocities of Gravel, Sand and Silt Particles.
American Journal of Science, 25, 148:325-338.
Sayre, W. W. and F. M. Chang. 1968. A Laboratory Investigation of Open-
Channel Dispersion Processes for Dissolved, Suspended and Floating
Dispersants. U.S. Geological Survey Professional Paper 433-E.
Schnoor, J. L. and E. G. Fruh. 1979. Dissolved Oxygen Model of Detention
Time Reservoir with Anaerobic Hypolimnion. Water Resources Bulletin,
15(2):506-518.
Schnoor, J. L. and D. C. McAvoy. 1981. A Pesticide Transport and Bio-
concentration Model. Journal of the Environmental Engineering Division,
American Society of Civil Engineers, EE6, 107:1229-1246.
Smoot, G. F., and C. E. Novak. 1968. Calibration and Maintenance of Vertical-
Axis Type Current Meters: Techniques of Water-Resources Investigations of
the United States Geological Survey. Book 8, Chapter B2. U.S. Geological
Survey, Arlington, Virginia.
Stettman, R. W. 1971. Aquifer-test Design, Observation, and Data Analyses:
Techniques of Water-Resources Investigations of the United States Geological
Survey. Book 3, Chapter B-l. U.S. Geological Survey, Arlington, Virginia.
11-41
-------
Tennekes, H. and J. L. Lumley. 1972. A First Course in Turbulence The
MIT Press, Cambridge, Massachusetts and London, England.
Thornthweite, C. W., and J. R. Mather. 1955. The Water Balance. Laboratory
of Climatology Dept., Drexel Inst. of Technology. Centerlon, New Jersey.
Toffaleti, F. B. 1968. A Procedure for Computation of the Total River Sand
Discharge and Detailed Distribution, Bed to Surface. Technical Report
No. 5. Corps of Engineers, U.S. Army, Vicksburg, Mississippi.
Toffaleti, F. B. 1969. Definitive Computations of Sand Discharge in Rivers.
Journal of the Hydraulics Division, ASCE, 95, HY1.-225-248.
U.S. Bureau of Reclamation. 1967. Water Measurement Manual: Second Edition,
U.S. Department of the Interior.
U.S. EPA. 1973. Methods for Identifying and Evaluating the Nature and Extent
of Nonpoint sources of Polutants. EPA-430/9-73-014. U.S. Environmental
Protection Agency, Washington, D.C.
U.S. EPA. 1974. Methods for Chemical Analysis of Water and Waste. EPA-625/6-
74-003. U.S. Environmental Protection Agency, Cincinnati, Ohio.
Vanoni, V. A. (Ed.). 1975. Sedimentation Engineering, American Society of
Civil Engineers, New York, N. Y.
Water Resources Council. 1976. Proceedings of the Third Federal Interagency
Sedimentation Conference, Denver, Colorado.
Welch, P. S. 1948. Limnological Methods. McGraw-Hill Book company, Inc.,
New York.
Wetzel, R. G., and G. E. Likens. 1979. Limnological Analyses. W. B. Saunders
Co., Philadelphia.
Yotsukura, N. and E. D. Cobb. 1972. Transverse Diffusion of Solutes in
Natural Streams. U.S. Geological Survey Professional Paper 582-C.
Yotsukura, N. and W. W. Sayre. 1976. Transverse Mixing in Natural Channels.
Water Resources Research, 21:695-704.
11-42
-------
POLLUTANT LOADING INPUTS
Pollutant loading Inputs to aquatic fate and transport models are
summarized in Table 2-2.
POINT SOURCES
The most conspicuous and easily measured sources of pollutant loads to
aquatic systems are point source discharges confined to pipes that lead
directly to the receiving water body. Computation of loading rates for such
discharges is generally straightforward since the volume of flow in a
discharge pipe and the concentrations of pollutants are relatively easily
measured. Generally, considerable diurnal variability in both the volume of
flow and pollutant concentrations is evident in municipal and industrial
discharges owing to man's daily schedule of peak daytime activity interrupted
by periods of reduced activity during the night. Consequently, it is common
practice to calculate an average loading rate based on flow and pollutant
concentration measurements taken over a 24 hour period. Such loading rates
are generally expressed in pollutant mass per unit time (e.g., kg/h or kg/d).
Obviously, loading computations based on a single instantaneous measurement
will.generally differ considerably from those based on daily mean values.
For purposes of aquatic fate and transport model validation studies, it
is essential that a distinction be made between loading values based on
instantaneous measurements and those based on dally, weekly or other mean
values. This holds whether 1) the discharge is the primary source of the
toxic organic pollutant whose fate and transport is being assessed in the
model, 2) the discharge is a minor contributing source of the pollutant,
or 3) the discharge is contributing an agent that tends to alter the activity,
availability, concentration or other chemical characteristic of the pollutannt
in question.
It is also important to observe the manner in which a discharge is
assimilated by the receiving stream-and to note any variability in dispersion
rates and patterns that may occur through time. Considerable variation in
mixing rates and boundaries of mixing zones is common in receiving waters
whose flow is regulated by dams, diversion, withdrawal and return flows, or
which experience considerable diurnal variability in the temperature regime as
a result of periodic thermal or hypollmnetic discharges or natural temperature
fluctuations. Also, the mixing patterns of the effluent containing the
pollutant of study, may vary considerably depending upon the volume of the
discharge, the chemical composition and temperature of the effluent.
These factors must all be considered in selecting a study area and
designing an aquatic fate and transport model validation study for toxic
organic substances. For further information on sampling program design and
12-1
-------
delineation of mixing zones for streams receiving regional wastes, the reader
is referred to National Academy of Sciences (1973) and Fetterolf (1973).
STREAMFLOW
Procedures for computing pollutant loads entering a designated reach or
stream compartment are straightforward if the system is well mixed and the
channel is relatively uniform and straight. The total discharge (Q) entering
the upstream compartment boundary is calculated following procedures described
in the Physical Transport section under the heading of Streamflow Entering
Compartment. Sufficient samples are collected and analyzed to provide a mean
concentration value (C) for the pollutant in question. Thus,
Y - KCQ
where: Y » The pollutant load (kg/h)
C - The mean pollutant concentration at the upstream compartment
boundary (mg/1)
Q - The discharge (m^/s)
K - A coefficient to adjust the units.
SUSPENDED AND BEDLOAD SEDIMENTS
Procedures for computing the loading rates of sorbed pollutants to stream
compartments via suspended and bedload sediments involve the same concepts as
Streamflow loading determinations. The total quality of both suspended and
bedload sediments entering a stream compartment are determined following
procedures described in the Physical Transport section. Mean concentrations
of sorbed pollutants associated with the respective sediment functions are
determined and loading rates computed for each as follows:
For suspended sediments: Ygp • QgCgpK
where: Yep • The total pollutant load associated with suspended sediment
(kg/h)
QS » The mean suspended Sediment discharge (kg/d)
CSP • The concentration of pollutant associated with the suspended
sediment (mg/kg)
K - A coefficient to adjust units.
For bedload sediment: Ybp - qtC^pK
where: Yvp • The total pollutant load associated with the bedload
discharge (kg/hr)
qt - The bedload discharge (kg/d)
12-2
-------
Cfep - The concentration of pollutant associated with the bedload
(og/kg)
K » A coefficient to adjust units.
BIOTA
Loading of pollutants associated with and transported by the biota' is
computed by: 1) sampling the particular community component of Interest
following procedures described in the Biological Uptake section and Appendix
B-l, 2) determining the total biomass of the appropriate community component
in transit, 3) determining mean pollutant levels in tissues of each community
of interest, and A) computing the total pollutant load (Yg) delivered to a
compartment via the biota or biological component. Thus, Yg • CgBKL;
where: Yg • The total pollutant load transported to a particular
compartment by the biota (kg/hr)
CB - The mean concentration of pollutant in biological tissue
(mg/kg)
B - The total biomass delivered to a compartment (kg/h)
K - A coefficient to adjust units.
PARTICULATE ORGANIC MATTER
Measuring the amount of particulate organic matter in suspended and
bedload sediments for purposes of determining concentrations of associated
pollutants must be accomplished through the specific gravity separation
process following procedures described by Guy (1969). Computations of loading
rates of sorbed pollutants associated with particulate organic matter are
subsequently determined as follows:
YPOM ' CPOM°.POMK
where: Ypo^ • The total pollutant load associated with particulate
organic matter (mg/hr)
mean concentration of pollutant associated with
particulate organic matter (mg/kg)
quantity of particulate organic matter the compartment
(kg/m)
K - A coefficient to adjust units.
GROUNDWATER
Some aquatic fate and transport models are designed to accept data on
pollutant loading to surface waters via ground water flow. Typical pollutant
sources that would contribute such loadings are hazardous waste dumps, holding
12-3
-------
ponds, and soil contaminants that have percolated or leached into ground-
water. Theoretically, such sites can be used as pollutant sources for
verification of models, but from a practical aspect, direct quantification of
such sources is infeasible.
Darcy's law is quite simple and can be used for determining rates and
quantities of water moving in large aquifers composed of sand and gravel.
However, application of the formula for an area of relatively small size, such
as at a waste disposal site, is very difficult. The formula is:
Q - KA QL
dt
where Q « The flow rate
K - The coefficient of permeability
A - The cross-sectional area
dh/dt - The hydraulic gradient (Todd 1959).
K is mathematically determined from data collected by drilling and
pumping a well while using a couple of observation wells to define the draw
down curve. The cross-sectional area is also difficult to determine because
height of the aquifer contributing to surface water recharge is unknown. It
is possible for groundwater to flow upward into a streambed. The hydraulic
gradient is more easily determined and only requires drilling about 3 wells to
measure the change in height of the water table over distance. These measure-
ments and Darcy's formula can be used to estimate groundwater flow to a
surface water body.
To quantify the pollutant loading to surface water, however, it is also
necessary to determine the average pollutant concentration in the ground-
water. This can be done by drilling and sampling a series of wells in the
contaminated aquifer but determining the depth to drill is again a problem.
The foregoing is not intended to indicate that waste dumps and ponds
cannot be used as pollutant sources for model verifications. Such sources can
be used but the suggested approach (for flowing waters) is to sample the
cross-section of the river at a point downstream from the source. Depending
upon the mixing in the river, it may be necessary to take a series of depth
and flow integrated samples to obtain a precise loading estimate. Using this
approach, the model loading is the pollutant load of the river, just upstream
from the first compartment.
The use of groundwater flow as a source of contamination has one major
advantage over many other sources. Groundwater loading rates are relatively
steady compared to loads from effluents, atmospheric deposition and non-point
source runoff. This steady type of input probably approximates steady state
better than most uncontrolled sources.
12-4
-------
ATMOSPHERIC DEPOSITION
An important source term or input loading factor for particulate matter
In a body of water is atmospheric deposition. Two primary mechanisms may
contribute to the total load: dry deposition and wet deposition Including
rainout and washout. We can begin by defining the deposition velocity Vg
Vg - - J (1)
8 C
where Vg has units of centimeters per second (cm/sec). J is the flux of
material to a surface in nanograms per square centimeter per second
(mg/cm^/sec), and C is the atmospheric concentration in nanograms per cubic
centimeter (mg/cm-*). The flux is assumed to be constant with height above
the surface although C and Vg are functions of height.
Flux and concentration are related by:
J - (D + E)l£ - VsC (2)
DZ
where D and E are the Brownian and Eddy diffusivlties, respectively, in square
centimeters per second (cm^/sec), Z is the height above the surface in
centimeters (cm) and Vs is the sedimentation velocity in centimeters per
second (cm/sec). The sedimentation velocity of a particle is reached when
the aerodynamic drag force on a falling particle exactly balances the
gravitational force. .The.minus signs are necessary since the flux downward to
a surface is a negative quantity by convention and the deposition velocity is
defined as positive.
Equation (2) applies to both small and large particles. For very small
particles, however, the sedimentation term is negligible. It is also assumed
that D is very much less than E, except very close to the surface or when
there is no wind. Therefore, for small particles, let E • ku*Z, and
substituting in equation (2):
j . - EdC . _ ku*zl£ - - ku* dc (3)
dZ dZ C(hrZ)
where k Is von Karman's constant equal to 0.4, u* is the friction velocity in
cm/sec, and hrZ is the natural logarithm of height.
For large particles, the diffuslvitles are negligible hence equation (2)
reduces to:
J - - VsC (4)
Comparison of equation (1) and (4) demonstrates that for large particles, the
deposition velocity equals the sedimentation velocity. It should be mentioned
that although equation (2) simplifies to equation (3) or (4) for small or
12-5
-------
large particles respectively, the general case has been examined by some
researchers. For example equation (2) has been integrated by Chamberlain
(1966) and shown in greater detail by Sehmel (1970), based on special
assumptions concerning the form of E near the surface.
Several workers (Chamberlain 1966; Islitzer and Dumbauld, 1963; Sehmel
1972; Sehmel et al. 1973) have demonstrated that deposition velocities over
rough surfaces are generally greater than sedimentation velocities for
similarly sized particles. This effect is especially evident at high wind
speeds. Impaction, turbulent deposition, and eddy diffusional deposition on
the roughness elements are responsible for the high deposition rate when
compared with sedimentation.
For a relatively smooth surface, Sehmel (1973) has shown that
sedimentation may be controlling. In particular, Vg is shown to be equal to
sedimentation velocity (Vs) for a wind speed of 2.2 meters per second (m/sec)
for particles larger than 0.3 urn in equivalent diameter. A smoother brass
surface was used for this study.
Wet deposition processes include "rainout" or removal of pollutants from
within clouds during precipitation events, and "washout" or below cloud
scavenging. Some researchers believe that wet deposition is more efficient
than dry deposition (Frantisak et al. 1980). However, the precipitation
quality resulting from scavenging processes will vary from region to region
and from storm to storm depending on the composition of the atmosphere at the
time precipitation occurs.
Dry deposition is determined by sampling fallout on exposed surfaces of
stainless steel or teflon for varying periods of time. Friedlander (1977)
and Chamberlain (i960) provide standard methods for sampling both dry and wet
deposition.
NONPOINT SOURCE FLOW AND SEDIMENTS
As was pointed out in the discussion of nonpoint source flow in the
Physical Transport section, the optimal means of dealing with nonpoint source
contributions to water courses is to select study areas with minimal potential
for nonpoint source Influences. In the event nonpoint source flow, sediments
and pollutant contribution cannot be avoided for whatever reason, the most
cost-effective means of estimating the loading inputs is by treating all
unaccountable increases in pollutant loads as nonpoint contributions, or as
the collective contributions of nonpoint source, groundwater and interflow
depending on the availability of data for the latter sources. The most
probable cause of nonpoint contributions would be a precipitation event
while the study is ongoing. The potential for contribution of toxic organic
pollutants to the water body, depends on the nature of the watershed in
terms of soil types, degree of slope, vegetative cover and, of course, its
uses and quantities of mobile pollutants contained therein. Agricultural
lands, for example, especially those dedicated to row crops, represent a
potential source of organic pesticides that may be easily transported via
overload runoff to a water course during an intense rainfall event. Similarly,
12-6
-------
mine tailings, dump sites, process waste piles and holding ponds from
numerous industrial activities represent potential sources of a myriad of
organic pollutant that could be mobilized by intense rainfall.
The investigator must be alert to the possibilities of such inputs both
during the early surveillance stages of site selection as well as during the
course of the study. Generally, visual observation (including aerial surveys)
is the most efficient means of judging whether or not nonpoint contributions
are of significant magnitude to warrant concern.
The use of loading models to estimate nonpoint contributions is generally
not encouraged because it introduces another source of error to the
total program. However, under some circumstances, this approach may be
Justified. For example, a number of watersheds have been intensively studied
and modeled to examine the rates of pesticide movement from the field to water
courses. In doing so, the hydrology, soil types, pesticide application rates,
etc. have been sufficiently well described to permit good estimates of roles
and routes of transport to receiving waters.
U.S. EPA reports by Crawford and Donigian (1973), McElroy et al. (1976),
Haith and Loehr (1979), and Evans and Duseja (1973) will prove to be useful
sources of information for persons interested in past efforts to evaluate non-
point source loading contributions to surface water.
Procedures for determining through instream measurements, the nonpoint
source contribution of pollutants that reach the water body compartments,
whether transported by the waterflow or as sorbed components of the sediment,
are essentially the same as those described previously for streamflow and
suspended and bedload sediments; i.e., the quantities of pollutants in the
respective compartment are determined at the upstream and downstream
boundaries, and unaccountable increases in loading within the compartment is
attributed to nonpoint pollutants.
12-7
-------
REFERENCES
Chamberlain, A. C. 1960. Aspects of Deposition of Radioactive and Other
Gases and Particles. In Aerodynamic Capture of Particles, E. G.
Richardson, ed. Pergamon Press, New York.
Chamberlain, A. C. 1966. Transport of Lycopodium spores and other small
particles to rough surfaces. Proc. Roy. Soc. 296:45-70.
Crawford, N. H. and A. S. Donigian, Jr. 1973. Pesticide Transport and
Runoff Model for Agricultural Lands. EPA-660/2-74-013. U.S.
Environmental Protection Agency, Office of Research and Development,
Washington, D. C.
Evans, J. 0. and D. R. Duseja. 1973. Herbicide Contamination of Surface
Runoff Waters. EPA-R2-73-266. Utah State University, Logan. U.S.
Environmental Protection Agency, Office of Research and Development,
Washington, D. C.
Fetterolf, C. M., Jr. 1973. Mixing zone concepts. Biological Methods for
the Assessment of Water Quality, "ASTM" STP 528, American Society for
Testing and Materials.
Frantisak, F., G. H. Pelletier, M. Bdard, and G. Gastonguay. 1980. Precipi-
tation Quality in Northeast Quebec-Canada. Vol. 4, report #80-52.5 of
the 73rd Annual Meeting of the APCA, Proceedings. June 22-27, 1980.
Friedlander, S. K. 1977. Smoke, Dust, and Haze Fundamentals of Aerosol
Behavior. John Wiley and Sons, Inc., New York.
Guy, H. P. 1969. Laboratory Theory and Methods for Sediment Analyses:
Techniques of Water-Resources Investigations of the United States
Geological Survey. Book 5, Chapter Cl. U.S. Geological Survey,
Arlington, Virginia.
Haith, D. A. and R. C. Loehr. 1979. Effectiveness of Soil and Water
Conservation Practices for Pollution Control. EPA-600/3-79-106. College
of Agriculture and Life Sciences, Cornell University Ithaca, New York.
U.S. Environmental Protection Agency, Office of Research and Development,
Washington, D. C.
Islitzer, N. F. and R. K. Dumbauld. 1963. Atmospheric diffusion -
Deposition Studies Over Flat Terrain. Int. J. Air Wat. Poll. 7:999-1022.
McElroy, A. D., S. Y. Chiu, J. W. Nebgen, A. Aleti, and F. W. Bennett. 1976.
Loading Functions for Assessment of Water Pollution from Nonpoint
Sources. EPA-600/2-76-151. Midwest Research Institute, Kansas City,
Missouri. U.S. Environmental Protection Agency, Office of Research and
Development, Washington, D. C.
12-8
-------
National Academy of Sciences (NAS). 1973. Water Quality Criteria 1972.
EPA-R3-73-033. U.S. Environmental Protection Agency. Ecological
Research Series.
Sehmel, G. A. 1970. Particle Deposition from Turbulent Air Flow. J.
Geophysical Res. 75(9):1766-1781.
Sehmel, G. A. 1972. Turbulent Deposition of Monodisperse Particles on
Simulated Grass. In: Assessment of Airborne Particles, Fundamentals,
Applications and Implications to Inhalation Toxicity. T. T. Mercer,
P. E. Morrow, and U. Stober, eds. Charles C. Thomas, Springfield,
Illinois.
Sehmel, G. A. 1973. Particle Eddy Dlffusivities and Deposition Velocities
for Isothermal Flow and Smooth Surfaces. J. Aerosol Sci. 4:125-138.
Sehmel, G. A., S. L. Sutter, and M. T. Dana. 1973. Dry Deposition
Processes. In: Pacific Northwest Laboratory Annual Report for 1972
to the USAEC Division of Biomedical and Environmental Research. Vol. II:
Physical Sciences. Part I: Atmospheric Sciences. C. L. Simpson, ed.
Battelle Northwest Laboratories, Rlchland, Washington.
Todd, D. K. 1959. Ground Water Hydrology. John Wiley and Sons, Inc., New
York, New York.
12-9
-------
SAMPLE COLLECTION, HANDLING, AND ANALYSIS FOR TOXIC SUBSTANCES
The previous sections on sampling have been primarily concerned with the
collection of environmental Input data that may be required for aquatic fate
and transport models. Field validation of models also requires that the pre-
dicted outputs of a model (Table 2-1) be compared with actual field data. For
these guidelines, it is assumed that the prime output of models being evaluated
are concentrations of organic compounds in various media, i.e., fish, water,
sediment, zooplankton, etc.
It is, however, beyond the scope of these guidelines to provide sample
collection methods for all model outputs and then to provide sample preservation
and analytical procedures for all organic chemicals that could be used in
model evaluation. Most sampling procedures for water, sediment and biota have
been described in previous sections (Bioconcentration and Physical Transport);
however, special precautions are required during collection to avoid contamina-
tion or loss of organic pollutants. The "Handbook for Sampling and Preservation
of Water and Wastewater" (U.S. EPA 1982) was developed for guidance on field
monitoring required under the National Primary Drinking Water Regulation, The
National Pollutant Discharge Elimination System, Section 304 (h) of the Clean
Water Act, and the consent decree for priority pollutants. The handbook
addresses the sampling of water and wastewater such as industrial/municipal
wastewater, agricultural runoff, surface water, and sediments, as well as flow
monitoring, handling and preservation methods. Extra attention is required to
preserve the sample until the compound(s) of interest is extracted and analyzed.
Individual investigators will have to identify the most appropriate sampling
procedure based upon the projections of the model, the compound whose fate is
being studied, and specific environmental factors at the study site. Prior to
field sampling it is essential to precisely define the model's outputs.
Sampling procedures specific to priority pollutants are contained in
"Sampling Protocols for Collecting Surface Water, Bed Sediment, Bivalves, and
Fish for Priority Pollutant Analysis" (VERSAR Inc. 1982). This manual is
currently in draft form but final publication is scheduled for 1984. Included
in the document are sections on sampling ambient water, bed sediment and fish.
All sections contain discussions of site selection, sampling equipment and
use, and sample preservation and shipping. Container selection and cleaning
are covered In the water and fish sections.
REFERENCES TO ANALYTICAL METHODS FOR ORGANIC COMPOUNDS
Detailed analytical procedures have recently been proposed by the U.S. EPA
(1979) for determining the concentration of 113 organic toxic pollutants in
water. Methods 601-613 apply to the analysis of individual compounds or groups
of chemically similar compounds. Methods 624 and 625 are GC/MS procedures for
13-1
-------
the analyses of the same compounds. The proposed methods cover calibration of
instruments, quality control, sample collection, preservation and handling,
sample extraction and analyses, and calculations. Revisions of the above
methods appear in "Test Methods, Methods for Organic Chemical Analysis of
Municipal and Industrial Wastewater" (Longbottom and Lichtenberg 1982). The
majority of the revisions were made for clarification or to add additional
flexibility for the analyst. Finalized procedures are to be published in the
Federal Register in 1984.
Currently there are no EPA approved methods for determining priority pol-
lutants in sediments or fish although the Environmental Monitoring and Support
Laboratory in Cincinnati has prepared interim guidelines (U.S. EPA 1981). These
procedures were originally prepared in 1977 and were revised in 1980 and 1981.
Final guidelines will not be available until 1984.
Another source of analytical procedures is the "Analysis of Pesticide
Residues in Human and Environmental Samples: A Compilation of Methods Selected
for Use in Pesticide Monitoring Programs" (Sherman and Beroza 1980). Brief
comments are addressed to the collection, preservation and storage of samples
but the cleaning of glassware is covered in detail. A separate section is
devoted to sampling analysis of water for pesticides with specific reference
to free acid herbicides. Gas-liquid chromatographic procedures for the analysis
of chlorinated hydrocarbon and organophosphorus pesticides in water are also
covered.
These are but a few of the many references available for the determination
of priority pollutants. Where possible, established methods should be used
but this is not always possible or practical. Whatever method is selected, it
must be tested and verified by the analyzing laboratory. For any EPA funded
project, the laboratory must also prepare and follow a quality assurance plan.
13-2
-------
REFERENCES
Longbottom, J. E. and J. J. Lichtenberg. Eds. 1982. Test Methods, Methods
for Organic Chemical Analysis of Municipal and Industrial Wastewater.
EPA-600/4-82-057. U.S. Environmental Protection Agency, Cincinnati,
Ohio.
Sherman, J. and M. Beroza. 1980. Analysis of Pesticide Residues in Human and
Environmental Samples: A Compilation of Methods Selected for Use in
Pesticide Monitoring Programs. EPA-600/8-80-038. U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
U.S. EPA. 1979. Guidelines Establishing Test Procedures for the Analysis of
Pollutants. Federal Register, Washington, D.C. 44(233):69464-69575.
U.S. EPA. 1981. Interim Methods for the Sampling and Analysis of Priority
Pollutants in Sediments and Fish Tissue. EPA-600/4-81-055. U.S.
Environmental Protection Agency, Cincinnati, Ohio.
U.S. EPA. 1982. Handbook for Sampling and Sample Preservation of Water
and Wastewater. EPA-600/4-82-029. U.S. Environmental Protection
Agency, Environmental Research Center, Cincinnati, Ohio.
VERSAR, Inc. 1982. Sampling Protocols for Collecting Surface Water, Bed
Sediment, Bivalves, and Fish for Priority Pollutant Analysis. Final
Draft Report. U.S. Environmental Protection Agency, Office of Water
Regulations and Standards, Washington, D.C.
13-3
-------
APPENDIX A
TIME AND DISTANCE CALCULATION
-------
APPENDIX A
EXAMPLE OF TIME AND DISTANCE CALCULATION THAT THE COMPOUND
OF INTEREST CAN BE DETECTED
SAMPLE DESIGN
A sensitivity analysis of the EXAMS model using phenol in the Monongahela
River predicts that phenol will only be detectable in water samples. Major
factors to be considered in designing the water sampling program are:
1. Distance and time of water travel between phenol dischargers.
Phenolic wastes are discharged at river mile (RM) 39.4 and no
other significant phenol discharges occur until downstream of RM
23.8. The time of travel in the Monongahela River has been
estimated by the U. S. Corps of Engineers (USCE). The water
transit time between RM 39.5 to Lock 3 (RM 23.8) can be estimated
by the following equation:
V - aQb
where: V • Velocity in feet per second
a - correction coefficient (0.000196)
b • correction coefficient (0.92)
Q - River flow in cubic feet per second (CFS)
The Monongahela River is regulated to maintain a minimum navigable
flow (1,000 cfs) and climatic conditions control discharges above
this level. The river flow on September 15 was approximately A,000
cfs and was used as the best estimate of flow conditions for early
October.
Therefore:
a) Velocity - V - aQb - .000196(4000).92 - 0.04378 ft/sec •=
0.275 miles/hr .
b) Distance between the effluent discharge point and Lock
3 - 15.7 miles
c) Time of travel • 57 hours
2* Concentration of phenol in Monongahela River below effluent.
An estimate of 22.679 kg/day of phenol loading was obtained from
the plant manager.
A-l
-------
Loading
22.679 kg/day
(22.679 x 109 wg/day)/(9.79 x/109/day) - 2.3 ug/1
River Flow
4000 CFS x 86AOO sec/day x 7.48 gal/CF x 3.78 1/gal - 9.79 x 109 I/day
3. Half-life of phenol in the Monongahela River.
Phenol degrades through bacterial breakdown and chemical
oxidation.
Combined bacterial and chemical oxidation degradation rates have
been estimated to be:
1.0 x 10-7 ml/cell/hr*
Average » 6.5 x 10~& ml/cell/hr
3.0 x 10-8 ml/cell/hr*
Concentration of Bacteria in Monongahela River:
1.0 x 105
1.0 x 106 Average - 1.0 x 106 Cells/ml
1.0 x 107
T 1/2 - .693/K
where:
T 1/2 - half-life
K - site specific rate
0.693 - natural log of 2
T 1/2 - 0.6S3/C6.5 x 1Q-8) (1.0 x 10$) - 0.693/0.065 - 10.66
hours
4. Time and distance before phenol falls below minimum detection limit:
The following formula can be used to estimate time elapsed
before phenol can be no longer detected:
c/co » e~kT,
A-2
-------
where: c - minimum detection limit (0.5 ug/1)
c0 - concentration at outfall (2.3 ug/1)
k • estimated rate of 6.5 x 10-2 per hour
c/c0 - 0.5/2.3 -e-°-065 T
T - 23.5 hours
Therefore phenol can theoretically be tracked for about 24
hours or 6.5 miles based upon a flow of 4,000 cfs and a minimum
detection limit of 0.5 Pg/1.
^Normalize to the number of cells per ml of water estimated by standard
plate count methods.
A-3
-------
APPENDIX B
BIOCONCENTRATION METHODS
-------
CONTENTS
B-l Description of Stream Habitats and Appropriate Macroinvertebrate
Sampling and Processing Procedures B-l-1
Pools B-l-1
Ekman grab sampler B-l-1
Core samplers. ............ B-l-1
Riffles B-l-1
Enclosed net samplers B-l-2
Colonization trap B-l-2
Runs B-l-2
Grab samplers B-l-3
Sampling Considerations B-l-3
Processing Benthic Samples B-l-4
References B-l-6
B-2 Algal Sampling and Processing Procedures for Stream
Environments B-2-1
Periphyton B-2-1
Phytoplankton B-2-1
Sample Preservation B-2-2
Estimating Algal Biomass B-2-2
Microscopic methods B-2-2
Numerical abundance B-2-2
Cell volume (biovolume) B-2-3
Cell surface area B-2-3
Non-microscopic methods for plant pigment measurements . . .
References B-2-5
B-3 Fish Sampling Methody B-3-1
Population Estimates B-3-5
References B-3-4
B-A Macrophyte Sampling Methods B-4-1
Qualitative sampling B-A-1
B-i
-------
Quantitative Sampling B-4-1
References B-4-3
B-5 Invertebrate Drift Sampling Methods B-5-1
References B-5-2
B-6 Zooplankton Sampling Methods B-6-1
References B-6-2
B-ii
-------
Appendix B-l. Description of Stream Habitats and Appropriate
Macroinvertebrate Sampling and Processing Procedures.
Pools—
Pools, or quiet waters, are usually found behind obstructions, in side
eddies, or immediately downstream from riffles. The finest sediments (silts
and finer) are deposited in pools, often along with considerable amounts of
organic debris. Although the smallest proportion of the total substrate of a
stream is pool-type substrate, the highest biomass per unit area is often
found here. This is due to the stability of the substrate and the large
amount of organic matter entrapped in the pools. The variety (diversity) of
organisms found in pools, however, is normally much lower than is found
elsewhere. In many cases, the greatest part of the biomass is represented by
biting and non-biting midges and segmented worms.
If export of biomass through emergence is Included in the model, it would
be very useful and simple to separate the midges (which emerge) from the worms
(which do not) while extracting the organisms from the samples. The most
common sampling devices for use in pools include the Ekman grab sampler
and coring devices described below.
Ekman Grab Sampler—This is probably the most widespread sampler used to
sample soft (fine-particle) substrates. The sampler is lowered onto the
substrate with jaws open and is triggered to close by a messenger if lowered
by line, or through a specially built pole used for hand operation. As the
sampler closes it grabs a fixed area of bottom substrate. The Ekman grab
collects a fairly shallow sample and will thus miss organisms living several
inches deep in the substrate. Weber (1973) lists a coefficient of variation
mode for this sampler of from 41 to 50 percent.
Core Samplers—These sample soft stream substrates in the same manner as
soil samples are collected. Although they reach deeper into the substrate
than the Ekman grab, special techniques are often required to hold the sample
in place during retrieval.
Riffles —
The dominant substrate of riffles is rocks, ranging from pebble to
boulder size. This habitat occurs where the stream gradient is steep, and
ranges from gentle, shallow riffles occurring only occasionally along lowland,
large rivers, to rapids and cascades occurring over entire reaches of mountain
streams. Owing to the heterogeneity of habitat—the various size rocks will
also entrap "microhabitats" of debris, gravel, sand and silt—and generally
high oxygen content, the greatest diversity of organisms in a stream reach
usually occurs in the riffle areas. The total biomass of invertebrates is
generally also quite high In riffles, particularly when the water is rich in
nutrients and the substrates are stable. The high heterogeneity of riffle
habitats also increases the natural patchy distribution of the invertebrates
over the habitat, thus Increasing sampling variability.
B-l-1
-------
Again, if "export" is relevant, it would probably be useful to separate
the non-insect invertebrates and the riffle beetles from the remainder of the
insects which "export" their biomass from the streams at emergence. Because
mechanical grab samplers and corers are ineffective in rocky riffle sub-
strates, sampling devices for riffles generally consist of some type of
enclosed net sampler or colonization trap.
Enclosed Net Samplers—Absolute techniques involve enclosing a unit-area
(usually 0.093 m^) of substrate with a frame or box of which the upstream
end is either open (Surber sampler) or consists of a fine-meshed netting
(Portable Invertebrate Box Sampler (PIBS)] that allows the stream current to
pass through the sampler. The downstream end of the sampler is always
equipped with a fine-meshed net that collects the debris and organisms that
have been dislodged through scrubbing of the substrate material of the
enclosed area by hand. The Surber sampler is less adequate than the
completely enclosed PIBS sampler because of loss of organisms through
backwash. Disadvantages of these samplers are that they can be operated only
in shallow waters (arm's length deep) and, like the Ekman, they do not sample
deep into the substrate where a large proportion of the invertebrate biomass
may be present (Hynes 1970). Weber (1973) reports a mode of coefficients of
variability for the Surber sampler of from 41 to 50 percent. Efficiency of
the enclosed net samplers varies greatly depending on ease of sampling,
heterogeneity of habitat, and richness of biota. A cautionary note here:
There may be a tendency to sample very close to the bank or where the water is
shallowest because this may be the only area available for effective net
operation. In streams that exhibit pronounced temporary fluctuation in
discharge in accord with storms and water usage patterns, such shallow areas
may yield unreasonably low estimates of biomass because they may have only
recently been inundated.
Colonization Trap—Where the water is too deep for enclosed net samplers,
a bag, tray, or box of known size is filled with cleaned substrate material
from the riffle and allowed to recolonize over a period of about 6 weeks. The
sampler is carefully retrieved so as to prevent escape of the fast-moving
invertebrates (such as larger stoneflies which may make up a large proportion
of the actual biomass.) The use of SCUBA techniques may be particularly
effective. Artificial substrate samplers, such as a basket filled with
ceramic spheres or a multiple-plate sampler, although effective for
comparative studies, do not accurately duplicate the actual stream conditions
and may give misleading biomass estimates.
Runs—
Except for mountain streams with very steep gradients, most stream reaches
consist of alternating riffles and runs, with runs generally making up an
increasingly greater proportion of the habitat as the stream gradient
decreases. Unlike pools, water flows over runs, but often at a slow rate.
The larger rivers and streams of lowlands consist almost entirely of runs.
The substrate of runs ranges from coarse silt to sand to gravel. Because of
the high instability of this substrate In flowing water and the abrasive
nature of the particles, the invertebrate fauna of this habitat is usually
B-l-2
-------
extremely sparse. It is not uncommon to sample several square meters of
substrate without finding more than a dozen organisms (normally midges,
burrowing mayflies and dragonflies), and it is easy to see how misleading a
biomass estimate would be if pool or riffle samples were used to estimate
total stream biomass when much of the stream bottom consists of the "run"
habitat. The most common sampling devices for runs are heavy, jawed grab
samplers, such as the PONAR. ;
Grab Samplers—PONAR grab samplers operate similarly to the Ekman
grab, but are designed to work effectively in sand or gravel. They are
heavy samplers that are lowered on a cable from a bridge or boat with jaws
that are held open and released at impact to "bite" the substrate. Because of
the extremely low density of organisms found in sandy substrates, the patchy
nature of invertebrate distribution can result in very high replicate-to-
replicate variability of data. Therefore, extra caution is needed when
estimating the invertebrate biomass of runs.
Sampling Considerations
Chapter 12 of Hynes (1970) should be required reading for all
investigators attempting to obtain stream macroinvertebrate biomass or
standing crop estimates. The inherently heterogeneous nature of streams and
the patchy distribution of the organisms over the substrate result in highly
variable data. However, if the model in question accepts inherent natural
variability of benthic invertebrate distribution and requires an allowance
for error of about 50 or 100 percent, then three to five replicates should
normally be adequate. Because the collection of benthic samples takes
only a fraction of the time and effort involved in processing these samples,
a safeguard strategy to assure that even this error (50 to 100 percent) is
reached would be to collect a large number of samples (e.g. 5 to 10), and
then process and analyze the replicates only until the desired level of
error is obtained. This method is similar to the sequential sampling
technique recently developed by pest entomologists (Resh 1979) and assures
the desired level of error is reached with the minimum amount of processing
effort. For additional information on required sample size for a given
level of precision see the biometrics chapter in Weber (1973).
Consideration should be given to the variability of stream invertebrate
biomass over time. Life cycle features of the dominant insect species present
in a reach of stream can greatly influence the amount of biomass in the stream
at any given time, as described by Hynes (1970). For example, in a stream
dominated by spring mayflies, invertebrate biomass can fall sharply during
late spring as many of the mature individuals emerge. Streams dominated by
summer species of black flies and midges or by amphipods (which breed and grow
during the summer months) may reach their peak in biomass during late summer.
Thus, a "transport-fate" model that requires invertebrate biomass as an input
should take these large seasonal fluctuations into account. This requires
that benthic samples be taken from the stream reach in question on a seasonal :
basis to determine the amplitude and duration of the seasonal variations.
B-l-3
-------
In addition to seasonal changes of biomass, less predictable but
substantial changes occur on an annual basis. Hynes (1970) presents
invertebrate data taken from a river station during the same season over a
9-year period that show that the total number or organisms changed as much as
threefold from year to year. These changes are probably in part due to annual
differences in intensity and timing of climatic events and features. Periods
of high water often have an especially drastic effect on biomass, reducing it
by several-fold. In the semiarid western United States, localized summer
thunderstorms can produce extremely severe floods and substrate scouring,
reducing the invertebrate populations from thousands per square meter to close
to zero. In short, even with the most carefully designed sampling metho-
dology, extreme caution is necessary when evaluating invertebrate biomass,
and consideration must be given to the above factors.
Processing Benthic Samples
Processing riffle and run Invertebrate samples for biomass estimates
will consist chiefly of extracting the invertebrates from the gravel and
debris. Separation of debris and Invertebrates from gravel is usually
quick and simple. The sample is poured into a pan with water and swirled
about with the lighter weight debris and organisms which collect at the top
retained (gold-panning in reverse). This can be done in the field or in the
laboratory. Before discarding the gravel, it is checked for snails, clams and
stone-cased caddisflies. If present, they will have to be hand-sorted from
the gravel component. Pool samples will consist of mud and some debris. The
mud can be rinsed through a large fine-meshed sieve or a bucket especially
equipped with a fine-meshed bottom.
Extraction of invertebrates from the remaining debris (sorting) is
usually by far the most time-consuming and tedious task involved with
invertebrate analysis and Is normally done in the laboratory. Hynes (1970)
reviews techniques which have worked with varying success to expedite the
sorting. Weber (1973) also describes processing procedures for macroinverte-
brate samples.
Subsampling, as suggested by Weber (1973), should also be considered
as a means of greatly reducing processing effort. For purposes of estimating
biomass, the entire sample should first be searched for the large, conspicuous
organisms (dragonfly and stonefly nymphs, horse fly larvae, large caddisfly
larvae, etc.). These will be quickly sorted and will often consist of the
great majority of the biomass. The smaller organisms (of which there may be
hundreds or thousands) can then be sorted from a fraction of the sample (e.g.
1/4 or 1/8) and the resulting biomass from this fraction back-multiplied by
the reciprocal of the portion sorted and added to the biomass of the larger
organisms which were sorted from the entire sample.
Naturally, the replicate samples must be kept discrete throughout the
processing, so that the biomass values from these separate replicates are
available for use in evaluating the variability among the replicate samples
and, hence, the precision of the biomass estimate yielded by the sampling
methodology employed. As noted earlier, it may also be important to separate
invertebrates that emerge from the stream from those that do not, if biomass
B-l-4
-------
"export" is an input in the model. Also, recording the size distribution of
the more important species will assist in determining whether biomass was
increasing or decreasing at time of sampling.
B-l-5
-------
REFERENCES
Hynes, H. B. N. 1970. The Ecology of Running Waters. University of Toronto
Press, Toronto.
Resh, V. H. 1979. Sampling Variability and Life History Features: Basic
Considerations in the Design of Aquatic Insect Studies. J. Fish Res.
Bd. Can. 36:290-311.
Weber, C. I. (Ed.). 1973. Biological Field and Laboratory Methods for
Measuring the Quality of Surface Waters and Effluents. Environmental
Monitoring Series. EPA-670/A-73-001. U. S. Environmental Protection
Agency. Cincinnati, Ohio.
B-l-6
-------
Appendix B-2. Algal Sampling and Processing Procedures for Stream
Environments.
Periphyton
The three basic stream habitats (pools, riffles, and runs) can have a
variety of substrate types (sand, pebbles, large rocks, aquatic plants,'
sticks, etc.) each with a unique periphyton community. Difficulties in making
accurate estimates of periphyton biomass in streams are largely due to the
variability of substrate types and current velocities. In some streams light
can also be a very important controlling factor.
Periphyton on natural substrates is highly variable due to different
textures and types of substrate, and the irregular surface is difficult to
quantify. For these reasons artificial substrates have come into use as they
seem to reflect the natural community clearly and are now accepted by most
researchers (Weitzel et al. 1979). Sladeckova (1962) describes numerous
natural and artificial substrates and gives detailed methods for their use.
On natural substrates, mason jar lids or pieces of sheet rubber with
circular holes punched in them have been used to describe the area for
sampling (Pers. Comm. with William D. Taylor, UNLV). The sheet rubber
conforms to the shape of many substrates thereby increasing the accuracy of
the sample. The substrate should be held over a pan while the periphyton are
scraped or brushed from it. The area is then rinsed into the pan to ensure
that all of the periphyton is included in the sample.
When using artificial substrates particularly good success with unglazed
ceramic tiles in streams and fiber glass plates in lakes has been reported
(William D. Taylor, UNLV Pers. Comm.). The ceramic tiles can be attached to
concrete blocks or bricks with silicone rubber and are easily cut free after
appropriate exposure. Lowe and Gale (1980) tested substrates for monitoring
periphyton in deep rivers. They recommended frosted acrylic for future
studies employing benthic artificial periphyton substrates. Weber (1973)
recommends the standard (plain 25 x 75 mm) glass microscope slide as the
most suitable artificial substrate for quantitative sampling.
Phytoplankton
Phytoplankton can be collected with several types of sampling devices.
Kemmerer and Van Dorn samplers collect known volumes of water (1, 2, and 3
liters are standard) from discrete depths. A tube sampler is designed to
sample the entire water column to a depth equal to the length of the tube.
This sampler integrates algae in the water column so that vertical layering
information is lost, but it has the advantage of collecting all of the algae
in the vertical segment of water. Phytoplankton nets (64 urn mesh) have been
used for years to concentrate plankton algae from large volumes of water,
but they cannot be considered quantitative samples suitable for obtaining
biomass estimates. Unfortunately the small nanoplankton, which often
constitute the bulk of the algal biomass, readily pass through the net and
are lost from the sample.
B-2-1
-------
All sampling for phytoplankton should take into account the vertical and
horizontal patchiness of the plankton community to properly characterize the
water body. A single sample taken from one depth at a station may totally
misrepresent the actual overall conditions of the aquatic ecosystem being
examined.
Sample Preservation ;
A variety of solutions are available for the fixation and preservation of
algal samples (Vollenweider 1969 and Weber 1973). It is the experience of
Taylor et al. (1979) and that of others (Vollenweider 1969) that acid Lugol's
solution is the fixative and preservative of choice for freshwater algae.
Acid Lugol's solution is prepared by mixing 10 g iodine and 20 g potassium
iodide in 200 ml of distilled water. To this, 20 ml of glacial acetic acid Is
added a few days prior to using. The solution should be stored in a dark
bottle to reduce oxidation of the iodine. This solution is added to a sample
at a 1:100 ratio. If large amounts of materials are collected in a sample,
enough acid Lugol's solution should be added to maintain a tea color in the
water. If samples are to be stored for more than 1 year, the preferred
preservative is formalin (40 percent formaldehyde - 100 percent formalin),
which has been neutralized with sodium tetraborate (pH 7.0 to 7.3). Five
ml of neutralized formalin are added for each 100 ml of sample (Weber
1973).
Estimating Algal Biomass
Various microscopic, gravimetric and chemical techniques are used to
measure the quantity of algal biomass (APHA 1980). The following quantitative
measurements can be made using microscopic techniques: numerical.abundance,
cell volume, cell surface area, and plasma volume. Common gravimetric and
chemical procedures for measuring biomass include the following parameters:
dry weight, ash-free dry weight, carbon, phosphorus, nitrogen, and chlorophyll
a.
Direct microscopic examination provides the most useful kind of informa-
tion (Fogg 1965) and has three basic advantages over other methods. The first
is that the algae are observed each time a count is made so that any changes
in appearance, size, shape, or aggregation of cells can be recorded. The
second is that dead and living cells may be differentiated. Finally, exact
information on algal species composition and size distribution is obtainable.
Non-microscopic determinations of algal biomass may be impaired by the
presence of detrital material, particulate organic matter, zooplankton and
bacteria, but are much less time consuming than microscopic counting methods.
Except for pigment determination, non-microscopic methods are generally
Inferior for obtaining reliable biomass estimates and are not further
discussed.
Microscopic Methods—
Numerical abundance—The use of numerical abundance (cells/ml or cells/
cm2) is of limited value as a measurement of biomass. This is attributable
B-2-2
-------
to the variation in cell size within individuals of a species and between dif-
ferent species of algae. Cell counts do not express these differences since
equal numerical value is assigned to each algal cell, regardless of size.
Paasche (1960) reported that cell numbers tend to be biased towards the
smaller, usually more numerous species in the community. Munawar et al.
(1974) reported that cell numbers do not give information about phytoplankton
biomass nor can they be correlated with primary production, particularly where
algal populations are variable in size. However, cell numbers have been
correlated with chlorophyll a. Taylor et al. (1979) found a rank correlation
between cell numbers and chlorophyll a_ for 44 eastern and southeastern
U.S. lakes to be 0.72 (p < 0.01). Munawar et al. (1974), also reported a
significant correlation between cell abundance and chlorophyll a_ (r » 0.59,
p < 0.01).
Cell volume (Biovolume)— Determination of cell volume (vo3 per
individual or colony) provides a measure of the algal biomass (mg fresh
weight/m3, assuming that the specific weight of algae is approximately
unity. This measurement of standing crop is widely accepted in quantita-
tive surveys (Rodhe et al. 1958, Nauwerck 1963, and Munawar and Nauwerck
1971). The appropriate dimensions of at least 25 randomly selected cells
are measured, and the volume of each of the measured cells calculated,
from which the mean cell volume is derived (Smayda 1965). The mean cell
volume should not be calculated from the average linear dimensions of the
individual cells. Cell volumes are usually reported in urn3/!, or urn3.
Simple geometric formulae may be used to compute the cell volumes although
some algal cells may have to be subdivided into several shapes because of
their complex geometric configurations. Cell volumes are computed by
simply integrating the volumes calculated for each form. Standard volumes
from published sources should be used with great care in these calculations
since differences in cell dimensions vary considerably from one place to
another and even seasonally at the same place.
Results of phytoplankton surveys expressed in terms of bio volumes may
tend to overemphasize the importance of the larger forms as producers (Paasche
1960). The small nanoplankton generally assimilate much more carbon per unit
of biomass than the larger forms (Findenegg 1965).
Cell volumes generally provide good correlations with other biomass
parameters. Munawar et al. (1974) reported that cell volume was better
correlated to chlorophyll a_ than cell surface area and numerical abundance.
Taylor et al. (1979), however, reported better correlations with cell
numbers and chlorophyll £ than biovolumes and chlorophyll a_ (rs - 0.72 and
0.66, respectively). Cell volumes can be converted to biomass using the
following formula from Strickland (1960) and Lund and Tailing (1957).
Biomass (mg) - vtotal ( m3) x 1Q-9 2E3 x S,
m3
where: Vtotaj » sum of the volume for each species, and S is 1.0, the
approximate specific gravity of algae.
B-2-3
-------
The results of most surveys are expressed in terms of fresh weight (mg) where
the approximate specific gravity of algae is assumbed to be 1.0 (Rhode et
al. 1958).
Cell surface area—Cell surface area (pm^) provides a better method of
estimating standing crop than numerical abundance; however, it is not as
widely used or as quantitative as cell volume (Munawar et al. 1974). Cell
surface area is important since it represents the assimilative area for
pollutants or nutrients. The area computation is similar to the method used
in computing cell volumes.
Nonmicroscopic Methods for Plant Pigment Measurements—
Chlorophyll £ is the predominant chlorophyll pigment in algae and
assumes considerable importance in standing crop estimates. The speed and
simplicity of chlorophyll £ analysis are the two main reasons that this
method is the most popular for estimating standing crop (Strickland 1960).
Results are usually reported in g/1 or g/cm2. The analysis is far less
time consuming than the microscopic "counting" methods; however, it does
not furnish information on algal species and size composition. This method of
estimating biomass is also faced with certain problems, i.e., pigment
extraction is not always complete, chlorophyll content varies with the age and
light or shade adaptation of the population, relative pigment composition of
various algae groups is not always constant, and degradation products may be
present that are extracted along with active chlorophyll by ordinary extraction
processes. Kalff and Knoechel (1978) provide the following formula for con-
verting algae fresh weight to chlorophyll a:
Chlorophyll £ (mg) - F x fresh weight (mg),
where:
F - 0.006 to 0.008 when the community is dominated by Cryptophyta,
and
F - 0.003 when the community is dominated by diatoms.
Chlorophyll a_ data are especially informative when used in conjunction
with other biomass parameters (Fruh et al. 1966). Weber (1973) and APHA
(1980) provide descriptions of appropriate procedures for chlorophyll £, b,
c and phaeophyton determinations using fluorometric annd spectrophotometric
methods.
B-2-4
-------
REFERENCES
APHA. 1980. Standard Methods for the Examination of Water and Wastewater.
Fifteenth Edition. APHA/AWWA/WPCF. Washington, D.C.
Findenegg, I. 1965. Relationship between standing crop and primary
productivity. In: Primary Productivity in Aquatic Environments. .C. R.
Goldman (ed.). Mem. 1st. Ital. Idrobiol. 18 Suppl., University of
California Press, Berkeley, pp. 273-289.
Fogg, G. E. 1965. Algal cultures and phytoplankton ecology. University of
Wisconsin press, Madison, Wisconsin.
Fruh, E. G., H. M. Stewart, G. F. Lee, and G. A. Rohlich. 1966. Measures of
eutrophication and trends. J. Water Pollut. Control Fed.
38:1237-1258.
Kalff, J. and R. Knoechel. 1978. Phytoplankton and Their Dynamics in
Oligotrophic and Eutrophic Lakes. Ann. Rev. Ecol. Syst»--;;9»-475-495.
Lowe, R. L. and W. F. Gale. 1980. Monitoring River Periphyton with
Artificial Benthic Substrates. Hydrobiologia 69:235-244.
Lund, J. W. G. and J. F. Tailing. 1957. Botanical Limnological Methods With
Special Reference to the Algae. Bot. Rev. 23:489-583.
Munawar, M. and A. Nauwerck. 1971. The composition and horizontal distribu-
tion of phytoplankton in Lake Ontario during the year 1970. In; Proc.
14th Conf. Great Lakes Res., Int. Assoc. Great Lakes Res. pp. 69-78.
Munawar, M., P. Stadelman, and I. F. Munawar. 1974. Phytoplankton biomass,
species composition and primary production at a nearshore and midlake
station of Lake Ontario during IFYGL. Proc. 17th Conf. Great Lake Res.
Internat. Assoc. Great Lakes Res. pp. 629-652.
Nauwerck, A. 1963. The Relationships Between Zooplankton and Phytoplankton
in Lake Erken. Sumb. Bot. Uppsal. 17:1-163.
Paasche, E. 1960. On the relationship between primary production and
standing stock of phytoplankton. J. Cons. Int. Explor. Mer. 26:33-48.
Rodhe, W., R. A. Vollenweider, and A. Nauwerck. 1958. The Primary Production
and Standing Crop of Phytoplankton. In; Perspectives in Marine Biology.
A. A. Buzzati - Traverse, ed. University of California Press,
Berkeley, pp. 299-322.
Sladeckova, A. 1962. Limnological Investigation Methods for the Periphyton
("Aufwuchs") Community. Botanical Review 28:286-350.
B-2-5
-------
Smayda, T. J. 1965. A quantitative analysis of the phytoplankton of the Gulf
of Panama II: On the relationship between C14 assimilation and diatom
standing crop. Inter-American Tropical Tuna Commission Bulletin
9:467-531.
Strickland, J. D. H. 1960. Measuring the Production of Marine Phytoplankton.
Bulletin No. 122. Fish Res. Bd. of Canada, Queens Printer, Ottawa,
Canada.
Taylor, W. D., L. R. Williams, S. C. Hern, and V. W. Lambou. 1979. Phyto-
plankton Water Quality Relationship in U.S. Lakes. Part VII. Comparison
of some new and old indices and measurements of trophic state.
EPA-600/3-79-079. U.S. Environmental Protection Agency, Las Vegas,
Nevada.
Vollenweider, R. A. (ed.). 1969. A Manual on Methods for Measuring Primary
Production in Aquatic Environments. IBP Handbook No. 12. Blackwell
Scientific Publications, Oxford.
Weber, C. I. (ed.). 1973. Biological Field and Laboratory Methods for
Measuring the Quality of Surface Waters and Effluents. EPA-670/4-73-001.
U.S. Environmental Protection Agency, Cincinnati, Ohio.
Weitzel, R. L., S. L. Sanocki, and H. Holecek. 1979. Sample Replication of
Periphyton Collected from Artificial Substrates. In: Methods and
Measurements of Periphyton Communities: A Review, ASTM STP 690, R. L.
Weitzel, ed. American Society for Testing and Materials.
B-2-6
-------
Appendix B-3. Fish Sampling Methods
Population Estimates
Total population estimates, including species and size class distribution
data, would be required to provide a precise estimate of the amount of a given
toxic organic substance potentially bound in fish tissue in a particular
stream segment at any given moment. Such precision is unnecessary given the
present state of predictive fate and transport models because they are not
sufficiently refined to accept this level of input data. However, in
anticipation of further refinement of predictive models it behooves the
biologist to obtain the best population estimate possible that is consistent
with economic effort and model sensitivity and resolution.
Inasmuch as it is generally impractical to attempt to obtain complete
counts or absolute biomass measures by size class and/or species in most
fisheries surveys, the biologist must attempt to obtain reasonable estimates
of these parameters based on samples of the population. Such estimates are
especially difficult in large rivers as pointed out by Cleary and Greenbank
(1954).
"On the whole, a large river probably contains a less uniformly
distributed fish population and a less stable one, in any given locality,
than does either a lake or a small stream. Thus, It is extremely
difficult to make a quantitative population estimate of the entire river,
or even of a sizeable stretch of it, by sampling random localities at
random times, unless a formidably large number of samples is taken. The
task is made all the more difficult by the fact that most known sampling
devices do not take a non-selective sample even in the immediate locality
in which they are operated.
Finally, it must be taken into consideration that the total
population changes from day to day, season to season, and year to year.
Recruitment comes about by the hatching of young, ingress, and by
artificial stocking. Removal is by angling, commercial fishing,
predation, death by natural and unnatural causes, and exodus into
connecting waters."
Commenting on the adequacy of "complete" counts in large stream, i.e.,
actual counts of all or nearly all individuals in a population, Cleary and
Greenbank (1954) noted the following:
"The physical nature of large streams usually precludes bringing to
hand the entire fish population, except In special instances, such as the
counting of fish at a fishway during spawning migration. However, the
idea of a total count may be applied to a given portion of a stream.
This portion then becomes a whole, in a sense, and may even be considered
a "body of water." For instance, a section of a stream is blocked off
and a population count made with an electric shocker. Probably this
count should more properly be considered a complete enumeration of the
B-3-1
-------
population of that stream section rather than a random sample of the
entire stream's population."
For purposes of obtaining total population or biomass estimates for large
segments of streams; extrapolation of "complete" count data from small repre-
sentative sections to large reaches warrants consideration. The advantages
and disadvantages of this approach will have to be examined on a case by.
case basis. If small stream segments that are representative of larger
reaches can be clearly delineated and adequately sampled, this technique
could prove very effective.
To obtain "complete" counts, the portion of the stream to be sampled
would have to be completely blocked to prevent ingress and egress of fish in
the area. The most common technique for blocking a section invloves the
placement of seines at each end of the segment to be sampled. Obviously,
small, slow moving, streams can be more effectively blocked in this manner
than large, swift streams. Drifting objects, boat traffic, etc., seriously
interfere with this technique in large systems.
A number of options are available for actually collecting the fish in a
complete count survey. Many of the sampling previously discussed by Weber
(1973) and Bagenal (1978) could be used singly or In combination with
others for such purposes. In very small streams with relatively uniform
channels and little rooted vascular plant growth, seining techniques are
effective provided obstructions that would interfere with the operation of
the seine or provide refuge for the fish are absent or are removed prior
to the operation. Drag seines have been used for years for obtaining
population estimates as well as for qualitative work. Hoover (1938) blocked
off portions of small trout streams and a large crew attempted complete
removal inside the blocked off areas. Seining efficiency, as indicated by
the mark and recapture method plus dynamiting, ranged from 70 percent to
100 percent depending on the number of obstructions in the area. Gerking
(1949) reported a seining efficiency of 88 percent based on his work with
a drag seine in a small warm-water stream in Indiana.
Despite these efficiencies with seines, Cleary and Greenbank (1954)
regard shocking as the most generally effective method of obtaining entire
population counts of sections of a stream or entire small streams. Poisons
and narcotizing agents are also effective for removing fish from entire
sections, but an obvious disadvantage is the hazard posed to fish downstream
from the experimental area. Even if attempts to completely remove fish from
stream segments are successful, extrapolation of the data to large stream
reaches is a potential source of considerable error in biomass estimates. If
the segments sampled are truly representative of the entire reach under
Investigation, a direct linear extrapolation of the biomass data is easily
accomplished through application of a proportional factor. This is the
simplest and most direct approach and probably is adequate for most survey
purposes. This assumes that the area (or volume) of each habitat within the
experimental section or sections sampled, is in direct spatial proportion to
the actual occurrence of each habitat throughout the entire section for where
biomass estimates are sought. If individual habitats, e.g., pools, runs,
riffles, weed beds, etc., are blocked off and sampled, the percentage of
B-3-2
-------
stream reach comprised by each habitat will have to be determined and the
biomass data applied proportionally. The investigator will have to make the
decision as to whether biomass data are reported for each species, for groups
of species and size classes occupying similar habitats, or by some other
system. Typically in streams supporting diverse fisheries different habitats
will be occupied by different species with fair consistency. Length and
weight measurements and size class grouping of each species should provide the
level of information required in practically all surveys.
Complete fish counts for a stream section or individual habitat may often
be impossible to obtain, particularly in large, deep, and/or fast moving
streams. For this reason, the method of "mark and recapture" is usually
employed for estimating populations. This procedure involves marking (e.g.,
by fin clip, tag, or dye) all fish in a catch and releasing them. In a
subsequent catch the number of fish with marks (i.e., "recaptures") and
without marks are noted and by simple proportion, a population estimate
achieved. For a complete description of this method, Including the formula
for computing standard error of the population estimate, biases inherent in
the procedure, and methods for minimizing these biases, refer to Chapter 6
"Estimation of Population Number and Mortality Rates" in Bagenal (1978).
B-3-3
-------
REFERENCES
Bagenal, T. B. (ed). 1978. Methods for Assessment of Fish Production in
Fresh Waters; IBP Handbook No. 3, 3rd ed. Blackwell Scientific
Publication, Oxford.
Cleary, R. E. and J. Greenbank. 1954. An Analysis of Techniques Used in
Estimating Fish Population in Streams, With Particular Reference to Large
Non-Trout Streams. J. Wildl. Mgt. 18:461-477.
Gerking, S. D. 1949. Characteristics of Stream Fish Populations. Invest.
Indiana Lakes and Str. 3:283-309.
Hoover, E. E. 1938. Fish Populations of Primitive Brook Trout Streams of
Northern New Hampshire. Trans. N. Amer. Wildl. Conf. 3:486-496.
Weber, C. I. 1973. Biological Field and Laboratory Methods for Measuring
the Quality of Surface Water and Effluents. EPA-670/4-73-001. U.S.
Environmental Protection Agency, Cincinnati, Ohio.
B-3-4
-------
Appendix B-4. Macrophyte Sampling Methods.
Qualitative Sampling—
Samples of macrophytes are gathered by hand, with grappling devices,
rakes or dredges. Plants collected for analyses of cumulative substances
should be removed in their entirety including flowers, seed pods, roots,;
rhizomes and tubers if possible. Representatives of all major communities
should be collected and placed in appropriate containers and frozen for
storage and shipment. Generally, identification of most common species can be
made in the field by an experienced aquatic biologist with the aid of
appropriate taxonomic keys. Rare or questionable specimens can usually be
identified by a macrophyte taxonomist if the specimens are handled properly
during and following collection. Two common references that provide taxonomic
treatment of aquatic vascular plants are Fassett (1968) and Muenscher (1944).
Quantitative Sampling—
Quantitative sampling of macrophytes for purposes of obtaining biomass
estimates for an entire stream reach or section thereof is a laborious task.
A critical aspect of such an undertaking is the accurate mapping of the stream
reach for purposes of delineating vegetative zones or plant communities. The
investigator can exercise considerable judgement and personal preference in
the delineation of plant zones regarding the number and sizes of discrete
zones. The important consideration is that the community boundaries be
clearly defined and plotted on scale maps as accurately as possible.
Communities may be delineated based on growth patterns (e.g., submersed,
floating, emersed), density of stand (abundance), diversity or "purity" of
stands or a combination of these characters. "Limnological Methods" (Welch
1948) is a good reference for conducting stream surveys, and for determining
areal extent of designated vegetation zones through the use of planimetry.
The sampling design should be such that the total biomass within each
designated zone can be estimated and expressed on a wet weight per unit area
basis. Sampling procedures Involve the establishment of transects or grids
following procedures adapted from terrestrial plant survey techniques. The
following references will be useful for this purpose: Edwards and Owens
(1960), Forsberg (1959), Boyde (1969), and Westlake (1966).
The actual sampling may involve wading, use of boats or diving. Plant
material may be removed by hand or by a sampling apparatus, depending upon the
nature of the substratum, water depth, and the type of plant communities being
sampled.
Determination should be made prior to initiation of sampling as to what
portion of the plant material is considered within the quadrant, e.g., all
plant material rooted within the quadrant, all plant material within the
sampled projections of the quadrant, etc. The sampling protocol should also
specify whether the entire plant will be removed intact or whether portions
above the substrate will be removed and the roots removed separately.
B-4-1
-------
All sampling techniques and equipment have their limitations. Grabs such
as the Ekman or PONAR can be used in relatively deep waters. They sample
small areas, but they are inconsistent with respect to the amount of root
material removed, and it is very difficult to restrict the sampling to the
confines of a quadrant. Corers are efficient for sampling roots in mud, sand
and small gravel, but they are not effective samplers of leafy plant material.
SCUBA techniques are very useful in many streams enabling the
investigator to remove all materials within the confines of a quadrant with
direct "hands on" contact. Turbidity may impede visibility of the divers even
though sufficient light penetrates to permit plant growth.
Plant materials retrieved are typically accompanied by large quantities
of soil, epiphytes and attached animals. Removal of soil requires that the
plants be washed over a sieve in a stream of water, but this also removes most
of the epiphytes and associated fauna. Estimates based on the work of Odum
(1957) and Edwards and Owens (1965) indicate that the weight of epiphytes
alone may be of the same order as the weight of macrophytes to which they are
attached.
Advances in remote sensing techniques are encouraging in terms of their
potential for providing biomass estimates of plant materials. Techniques
range from the use of aerial photography to define the spatial distribution of
plant beds to the use of laser fluorescence devices for estimating the
quantities of plant pigment in a given volume of water. Pigment
concentrations can then be related to standing crop or biomass. Currently
remote sensing techniques are in the developmental stages for purposes of
providing quantitative standing crop estimates on an aerial unit basis.
B-4-2
-------
REFERENCES
Boyde, C. E. 1969. Production, Mineral Nutrient Absorption and Biochemical
Assimilation by Justicia americana and Alternantheia philoxeroides.
Archiv. Hydrobiol. 66:139-160.
Edwards, R. W. and M. Owens. 1960. The Effects of Plants on River
Conditions. I. Summer Crops and Estimates of Net Productivity of
Macrophytes in a Chalk Stream. J. Ecol. 48:151-160.
Edwards, R. W. and M. Owens. 1965. The Oxygen Balance in Streams. In:
Ecology and the Industrial Society. Goodman, G. T.; R. W. Edwards and J.
M. Lambert, eds. Synp. Brit. Ecol. Soc. 6:149-172.
Fassett, N. C. 1968. A Manual of Aquatic Plants. Univ. Wise. Press.
Madison, Wisconsin.
Forsberg, C. 1959. Quantitative Sampling of the Sub-Aquatic Vegetation.
OIKOS 10:238-240.
Muenscher, W. C. 1944. Aquatic Plants of the United States. Cornell Univ.
Press. Ithaca, New York.
Odum, N. T. 1957. Trophic Structure and Productivity of Silver Springs,
Florida. Ecol. Monogr. 27:55-112.
Welch, P. S. 1948. Limnological Methods. McGraw-Hill, New York.
Westlake, D. F. 1966. The Biomass and Productivity of Gylrecia maxima. I.
Seasonal Changes in Biomass. J. Ecol. 54:745-753.
B-4-3
-------
Appendix B-5. Invertebrate Drift Sampling Methods.
Typically, sampling drift organisms involves the use of specially
designed nets anchored just above the stream bottom (but below the surface in
shallow streams). Drift nets are left in place for a few minutes up to 24
hours depending upon the density of organisms and amount of floating debris in
the water. Results are expressed in terms of biomass or number of animals
captured per unit time, and, if desired, composition of the fauna can be
determined. For further discussions of sampling procedures for drift organisms
see Weber (1973), Elliott (1970), Waters (1962, 1969) and Edmondson and
Winberg (1971).
B-5-1
-------
REFERENCES
Edmondson, W. T. and G. G. Winberg (eds.). 1971. A Manual for the Assessment
of Secondary Productivity in Fresh Waters. Blackwell Scientific, Oxford.
Elliott, J. M. 1970. Methods of Sampling Invertebrate Drift in Running
Waters. Ann. Limnol. 6:133-159.
Waters, T. F. 1962. Diurnal Periodicity in Drift of Stream Invertebrates.
Ecology 43:316-320.
Waters, T. F. 1969. Invertebrate Drift - Ecology and Significance to
Stream Fishes. In: Symposium on Salmonn and Trout in Streams. T. G.
Noalhiate, ed. N. R. MacMillon Lectures in Fisheries. Univ. British
Columbia, Vancouver, pp. 121-134.
Weber, C.I. (ed). 1973. Biological Field and Laboratory Methods for
Measuring the Quality of Surface Water and Effluents. EPA-670/4-73-001.
U.S. Environmental Protection Agency, Cincinnati, Ohio.
B-5-2
-------
B-6. Zooplankton Sampling Methods.
Zooplankton biomass in the water column of deeper streams can be
estimated by using appropriate mesh netting [No. 10 (0.158 mm) for micro-
crustacea and No. 20 (0.076 mm) for rotifers and nauplii]. Of course, the
finer mesh netting which clogs very rapidly in many streams is best used at
some distance above the substrate, whereas much coarser macroinvertebrate
sampling net material (0.5 mm) is better suited to use near the stream
bottom where most of the coarser suspended material is transported.
Conventional Zooplankton sampling techniques, both quantitative and
qualitative, may be appropriate for very large streams. These procedures
involve the use of metered tow nets deployed from boats and drawn through
water either horizontally or vertically, as well as filtering water col-
lected in containers or pumped to the surface. Weber (1973), Welch (1948),
and Edmondson and Winberg (1971) provide descriptions of Zooplankton sampl-
ing apparatus and their uses for a variety of sampling situations and
study objectives.
B-6-1
-------
REFERENCES
Edmondson, W. T., and G. G. Winberg. 1971. A Manual on Methods for the
Assessment of Secondary Production in Fresh Waters. Blackwell Scientific,
Oxford.
Weber, C.I. (ed). 1973. Biological Field and Laboratory Methods for
Measuring the Quality of Surface Water and Effluents. EPA-670/4-73-001.
U.S. Environmental Protection Agency, Cincinnati, Ohio.
Welch, P. S. 1948. Limnological Methods. McGraw-Hill, New York.
B-6-2
-------
INDEX
Acceptance criteria, 1-4
comparison of model performance
with, 1-32
Acclimation, of microorganisms to
new substrate, 3-4, 3-7
Actinometer measurements, 5-5
Algae
biomass
estimating, 10-7
gravimetric and chemical
methods, B-2-2
microscopic methods, B-2-2
preservation of samples, B-2-2
sampling and processing, B-2-1
Altitude, 5-3
Assumptions, model, 1-12, 1-18, 1-19
ATP method, for estimating viable
biomass, 3-10
Bacteria
active degrading population,
3-11
determination in sediment,
3-10
determination in water, 3-8
Biological uptake
environmental factors affecting,
10-3, 10-6
Biotransformation, 3-1
factors Influencing, 3-2
rate constants, chemical
properties influencing,
3-6
determination in laboratory,
3-11
Chlorophyll
measurement, 5-7
method for estimating algal
biomass, 10-7, B-2-2
Cloud Cover
measurement, 5-3
Flow
measurement, 11-19
Groundwater, 11-18
Comparisons, of model performance
with acceptance criteria
graphical, 35-38
statistical, 38-43
Compartments
designation, 133-134
dimensions, 134-137
Compound
selection factors, 10-14
toxicity, 11
Decay curve, 30
Degradation rate, microbial
compound properties affecting,
3-6
environmental factors
affecting, 3-2
rate constants, 3-6
Direct counting method
for bacteria in water, 3-9
Dissolved organic carbon
measuring, 5-6
Dissolved oxygen
measurement, 65
Eddy diffusivity, 11-5, 11-36
Epifluorescence, method for deter-
mining microbial biomass, 3-9
Evaporation, measurement, 11-37
Field validation
def., 1-1
feasibility, 1-5
scenarios, 1-6
steps, 1-2
Field verification
design problems/solutions, 1-25
Fish
as concentrators of TOS,
population estimates,
sampling methods, B-3-1
chemical poisons, B-3-2
electroshocking, B-3-2
Model outputs, 2-5
Model user needs, 1-4
Non-point source water, 11-19
1-1
-------
Guaging stations, 11-20
Henry's constant, 8-3
Inputs to models
see Model Inputs
Interflow, 11-18
Invertebrate drift
general,
sampling methods, B-5-1
lonization, 7-1
factors influencing, 7-4
Latitude, 5-2
Longitude, 5-2
Lipopolysaccharide (LPS)
method for determining microbial
biomass, 3-7, 3-10
Macroinvertebrates, benthic, 10-7, B-l
pOH methods, 4-3
Macrophytes
as concentrators of TOS, 10-9
sampling methods, B-4-1
Microorganisms
active degrading population,
3-11
estimating populations, 3-8
nutrients required, 3-3
Mixing, measurement of, 8-6
Model inputs
environmental, by process, 2-2
pollutant loading, 2-5
Bioconcentration
Priority pollutants, organic, which
Biotransformation, 3-8
undergo:
Biotransformation, 3-8
hydrolysis, 4-3
microbial degradation, 3-8
oxidation, 6-2
photolysis, 5-4
sorbtion, 9-6
volatilization, 8-5
Protocol, preparation checklist,
1-16
Q10, 3-4, 3-12
Quality assurance, 1-24
references, 1-30
Rainfall, measurement, 11-18
Rate constants, 1-18
Reaeration rate, 8-6
Nutrients, required by microorgan-
isms, 3-3
measurement, 3-12
Octanol-water partition coefficient,
9-5
Organic compounds
analytical methods, references,
Outputs of models,
see Model Outputs
Oxidation
mechanisms, 6-1
methods
free radical oxidants, 6-2
singlet oxygen, 6-3
Particulate organic matter, 11-24
pH
hydrolysis effect on biodegradation rates,
3-4
measurement, 4-3
Phosphorus,
Photolysis
factors influencing,
Physical transport
factors influencing, 5-3
Pollutant loading sources
atmosphere, 12-5
biota, 12-3
groundwater, 12-3
non-point, 12-6
particulate organic matter, 12-3
point, 12-1,
sediments, 12-2
streamflow, 11-19
discharge calculations,
11-27
organic matter, 11-24
settling velocity, 11-29
size distribution, 11-28, 11-33
tributary inflow of, 11-19, 11-23
transport,
Sensitivity, 1-13
Sieving, for sediment separation,
11-28
Site selection, 1-11
Sorbtion
environmental factors which
1-2
-------
Samplers
colonization trap,
core, B-l-1
Ekman, B-l-1
enclosed net, B-l-2
grab, B-l-3
Sampling
for toxic substances, 13-1
Sediment
bedload, 11-12
anion exchange, 9-9
bulk density, 11-34
cation exchange, 9-8
discharge measurement,
11-30
leaving a compartment, H-36
organic matter, 11-34
pH, 9-8
size distribution, 11-33
total load, 11-31
water content, 11-35
non-point source, 11-22
suspended, 11-24, 11-25, 11-36
concentration determination,
11-27
Validation, field: see field
validation
Volatilization, 8-1
factors influencing, 8-5
affect, 9-6
Spectral energy, measurement, 5-4
Spectral Radiometer methods, 5-6
Standard plate count method, 3-8
Streams, habitats and macroinverte-
brate sampling, B-l-1
Streamflow, 11-19
measurement, 11-19
Temperature
measurement, 3-12
Time and distance,
detection of compound, example,
A-l
Total organic carbon (TOC),
analysis, 11-36
Wind
measurement, 8-7
Zooplankton
occurrence In flowing streams,
10-10
sampling methods, B-6-1
1-3
-------
TECHNICAL REPORT DATA
(ntattnad /nttructiont on the rtvene btfort complerint)
1. REPORT NO.
a.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
GUIDELINES FOR FIELD TESTING AQUATIC FATE
AND TRANSPORT MODELS:
6. REPORT DATE
Aoril 1983
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
S. C. Hern, G. T. Flatman, W. L. Kinney,
F. P. Beck. Jr.. J. E. Pollard, and Alan B. Crockett
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Monitoring Systems Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, NV 89114
10. PROGRAM ELEMENT NO.
3ACR60JOAB
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency—Las Vegas, NV
Office of Research and Development
Environmental Monitoring Systems Laboratory
Las Vegas. NV 8911A
13. TVPE OF REPORT AND PERIOD COVERtO
14. SPONSORING AGENCY CODE
JPA/600/07
16. SUPPLEMENTARY NOTES
16. ABSTRACT
This guidance has been developed for those attempting to field validate aquatic
fate and transport models. Included are discussions of the major steps in validating
models and sections on the individual fate and transport processes: biodegradation,
oxidation, hydrolysis, photolysis, ionization, sorption, bioconcentration,
volatilization, and physical transport. For each process the following Information is
provided: a general description of the process, a list and discussion of environmental
factors affecting the process, a list of the priority pollutants for which the process
is important, a list of model-specific environmental inputs, and field methods for
collecting these input data.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATl Field Croup
Mathematical models
Aquatic fate and
transport models
EXAMS model
Model validation
07 C
06 F,M
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (Tliii Report)
UNCLASSIFIED
21. NO. OF PAGES
230
30. SECURITY CLASS (Thit page/
UNCLASSIFIED
22. PRICE
IPA Pwa 3220-1 («•». 4-77) FMivious IOITION is OXOLKTE
------- |