Click here for
DISCLAIMER
Document starts on next page
-------
SEPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
Center for Environmental
Research Information
Cincinnati OH 45268
Technology Transfer
EPA/625/6-86/013
Handbook
Stream Sampling for
Waste Load Allocation
Applications
-------
EPA/625/6-86/013
September 19B6
Handbook
Stream Sampling for
Waste Load Allocation Applications
by
William B Mills, George L. Bowie, Thomas M Grieb, and Kay M Johnson
Tetra Tech, Incorporated
Lafayette, CA 94549
and
Raymond C Whittemore
NCASI
Medford, MA
Subcontract No 101
Eastern Research Group, Inc
Cambridge. MA
EPA Project Officer
H Douglas Williams
Center for Environmental Research Information
Cincinnati, OH 45268
U.S Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
Notice
This document has been reviewed in accordance with U.S Environmental
Protection Agency policy and approved for publication Mention of trade names
or commercial products does not constitute endorsement or recommendation
for use.
-------
Abstract
This report discusses sampling requirements in support of waste load allocation
studies in rivers and streams. Two approaches to waste load allocation are
addressed: the chemical-specific approach and the whole effluent approach.
Numerical or analytical toxicant fate models are used to implement the
chemical-specific approach. Modeling requirements and swrrmling guidelines
are delineated for this method.
For the whole effluent approach, the method is first summarized and then
instream dye study requirements are presented. The report concludes with
example applications of the chemical-specific approach for conventional and
toxic pollutants.
-------
Contents
Chapter Page
Abstract
-------
List of Figures
Number Page
1 -1 Summary of stream survey design consideration for water
quality modeling approach and whole effluent toxicity
approach to waste load allocation 1-3
2-1 Example stream network showing reaches and computational
elements 2-2
2-2 Physical representation of a stream by model segments 2-3
2-3 Effects of grid resolution on predicted dissolved oxygen
profiles 2-4
2-4 Derivation of cross-sectional area vs flow and velocity
relationships from stage-flow data 2-5
2-5 Recommended locations for a minimal sampling program 207
2-6 Recommended sampling locations at point sources 2-9
2-7 Allocation of sampling effort based on preliminary analyses 2-11
2-8 Processes affecting dissolved oxygen 2-13
2-9 Effect of pH and temperature on un-ionized ammonia 2-14
2-10 Major constituent interations in QUAL-II 2-16
2-11 Example computation of total BOD removal rate, K,, based on
BOD measurements 2-18
2-12 Procedure for estimating Ka and K, from BOD measurements ....2-19
2-13 BOD decay times for various decay rates 2-21
2-14 Example sampling network for a dissolved oxygen analysis 2-22
2-15 Results of a short-term intensive survey to establish the
dissolved oxygen profilte 2-24
2-16 Daily dissolved oxygen variation in two rivers 2-24
2-17 Typical concentration profiles of toxicants in rivers 2-31
2-18 Sampling locations for toxicants during low flow and
high flow periods 2-31
2-19 Typical suspended solids concentrations during (a) low
flow and |bi high flow periods 2-32
3-1 Overview of effluent toxicity testing procedures 3-2
3-2 Overview of ambient toxicity testing procedures 3-3
3-3 Distances below point source discharges required for
complete vertical and transverse mixing 3-4
3-4 Time required for a continuous release of dye to reach
steady-state concentrations at selected locations below
the point of discharge 3-5
-------
List of Figures fCont'd)
Number Page
3-5 Dye isopleths in wide and narrow rivers 3-6
3-6 Regions of observable toxicity in wide and narrow rivers 3-6
3-7 Example sampling locations in wide and narrow rivers 3-7
4-1 Eel River and environs showing summer of 1961 water
quality results 4-1
4-2 Location of sampling stationson Eel River 4-2
4-3 El Cahon River, Lake Chabot, and environs 4-3
4-4 Location of sampling stations on El Cahon River 4-4
-------
List of Tables
Number Page
1 -1 Waste Load Allocation Guidance Documents ................... 1-2
2-1 Data Requirements for Hand-Calculation Techniques Described
in WLA Guidance Documents and Screening Manual for
Analysis of Cor •<_•. ••*• ' .' . ' ts ............................ 2-15
2-2 Processes Simulated :>i UUA..-H ............................... 2-16
2-3 Non-Toxic Constituents Included in Stream Models ............. 2-17
2-4 Model Input Parameters for QUAL-II ........................... 2-17
2-5 Comparison of QUAL-II with Other Conventional Pollutant
Models Used in Waste Load Allocations ........................ 2-1 8
2-6 Methods for Determining Coefficient Values in Dissolved
Oxygen and Eutrophication Models ............................ 2-20
2-7 Summary of Data Requirements for Screening Approach for
Metals in Rivers .............................................. 2-26
2-8 Summary of Data Requirements for Screening Approach for
Organics in Rivers ............................................ 2-27
2-9 MICHRIV Model Data Requirements ........................... 2-28
2-10 Summary of Input Data Required for TOXIWASP ................ 2-29
2-1 1 Travel Times for Various C/Co Ratios Corresponding to
Different Toxicant Decay Rates ................................ 2-31
2-12 Summary of Sampling Guidelines for Toxicants ................. 2-33
4-1 Summary of Data to be Collected During Stream Survey for
Dissolved Oxygen Waste Load Allocation ....................... 4-2
4-2 Properties and Fate Processes for Pyrene ....................... 4-3
1 ~K|A*'Z
4-3 Range of '~e ......................................... 4-4
4-4 Summary of Data to be Collected During Stream Survey ......... 4-5
VIII
-------
A ckno wledgments
Many individuals contributed to the review of this handbook. Special recognition
is given to Steven Gherini of TetraTech; Timothy Stuart, Elizabeth Southerland,
and James Plafkin, Monitoring and Data Support Division, USEPA, Washington,
DC; Steven McCutcheon, USEPA Environmental Research Lab, Athens, GA; H.
Douglas Williams and Orville Macomber, USEPA Center for. Environmental
Research Information, Cincinnati, OH; William Richardson and Larry Fink,
'J3tr>A Large Lakes Research Station, Grosse He, Ml; Robert Bordner, William
Horning and Robert Safferman, USEPA Environmental Monitoring and Support
Lab, Cincinnati, OH; Edward Woo, USEPA Region I, Boston, MA; Noel Kohl,
USEPA Region V, Chicago, IL; Henry Holman, USEPA, Region VI, Dallas, TX; and
Bruce Zander, USEPA, Region VIII, Denver, CO.
Thanks is also expressed to Trudy Rokas, Susan Madson, and Gloria Sellers for
typing and preparing the report, and Marilyn Davies for providing graphics.
IX
-------
Chapter 1
Introduction
1.1 Background
The United States Environmental Protection Agency's
Monitoring and Data Support Division is presently
developing guidance manuals th i dusciibe ap-
proaches for allocating waste loa's «r rivers a-id
streams, lakes and impoundments, and estuaries.
The pollutants addressed in the manuals are bio-
chemical oxygen demand/dissolved oxygen, nutri-
ents, and toxic substances (ammonia, organics, and
metals). Other manuals in the series present related
topics, such as how to select the critical conditions for
the waste load allocation (WLA) (e.g., the appropriate
stream flow) Table 1 -1 summarizes the documents.
Water quality simulation models are often used for
WLA purposes. These models must adequately
predict water body responses to different waste loads
when large financial expenditures are at stake. Conse-
quently, where feasible, models should be calibrated
and verified prior to allocating waste loads. Sufficient
historical data to accomplish these objectives are
often lacking and of the wrong type, and additional
data should be collected. Water quality specialists.
therefore, have to decide what data are missing and
their importance, and then design surveys to gather
any required information. This handbook is intended
to guide specialists through these steps for waste
load allocations in rivers and streams. Both the
chemical specific and whole effluent approaches to
WLA are discussed.
This handbook can be used in conjunction with Book
II, Rivers and Streams, with Book VI, Design Condi-
tions, and with appropriate sections of Book VIII,
Screening Manual. Book V, The Technical Support
Document for Water Quality-based Toxics Control,
will be useful as well (See Table 1 -1).
Because the river water quality model QUAL-II (1,2)
and its followup QUAL-2E (3) is widely used for WLA
applications and is supported by the USEPA's Center
for Water Quality Modeling, example stream survey
designs for this particular model are included in this
handbook. Stream survey guidance for the toxicant
models TOXIWASP (4) and MICHRIV (5) are provided
as well. Users of other models will find much of the
guidance applicable to their models because of
similarities in model requirements
1.2 Purposes of Handbook
The primary purpose of this handbook is to help water
quality specialists design stream surveys to support
modeling applications for waste load allocations. The
planner is guided through the data collection process
so that models used for WLA can be calibrated,
verified, and applied to the critical design conditions.
Field sampling requirements of the whole effluent
approach to waste load allocation are also addressed.
This handbook does not discuss a number of facets of
stream sampling where significant reference mate-
rials already exist. These areas include:
« equipment requirements
• personnel requirements
• collection of samples
• determination of stream geometrical and flow
characteristics
• laboratory analytical techniques.
The Appendix summarizes the appropriate literature
in these categories. The references are primarily from
the U.S. Geological Survey's Water Resource Investi-
gation series, the U.S. Fish and Wildlife's Instream
Flow Information series, and from the U.S. Environ-
mental Protection Agency.
This handbook also recognizes that waterborne
viruses are pollutants which produce definite health
effects. However, these pathogens cannot be con-
sidered in the wasteload allocation process which are
involved only with parameters that have established
water quality criteria.
The second purpose of this handbook is to show how
models can be used to help design stream surveys.
Since the models will eventually be used to predict
the allowable waste loads, they can be set up and
applied before the stream surveys are finished. This
will assist planners in examining the available data,
allow preliminary sensitivity analyses to be made.
and thereby help identify the most needed data.
Stream surveys can then focus on the collection of
such data, and de-emphasize data that are less
important or previously well characterized.
The third purpose of this handbook is to educate field
personnel on the relationship between sampling
requirements and modeling requirements. Field
-------
Table 1-1
Waste Load Allocation Guidance Documents
Waste Load Allocation Guidance
Book I General Guidance
Book II Streams & Rivers
BOD/DO
Nutrients/Eutrophicatton
Toxic Substances
Simplified Methods for POTWs
Book III. Estuaries
6OD.DO
Nutnems/Euuophicatiori
Toxic Substances
Book IV. Lakes & Impoundments
BODTJO
Nutrients/Eutrophication
Toxic Substances
Book V -r^^.,.^1 a, .. n-^umgnt for
Mater uuaiitv-Based jx/cs Control
Book VI. ,Jtn>.ji: •• fu.wo., ~,tions
Design Flow
Design Temperature
Design pH
Design Effluent Flow
Design Rate Constants
Book VII. Permit Averaging
Book VIII Screening Manual
BOD/DO
Toxic Organic*
Toxic Metals
Nutrients/Eutrophication
Book IX. Innovative Waste Load Allocations
•Available from Monitoring and Data Support Division, USEPA
(WH553), Washington, D.C. 20460. See latest Monitoring and
Waste Load Allocation status report for completion dates for
these documents.
personnel may sometimes question why historical
data are not adequate, why specially designed
surveys are often required to generate the data, and
why certain sampling locations and parameters are
selected. By understanding the factors that go into
the selection process, field personnel are likely to
perform their tasks more effectively. When unfore-
seen field conditions dictate a change in sampling
strategy, there is a better basis for deciding how to
modify the sampling program design.
1.3 Overview of Approach
Figure 1 -1 summarizes the approach to stream survey
design discussed in this handbook. Two parallel
approaches are possible: the chemical specific ap-
proach and the whole effluent approach.
The chemical specific approach is selected if the
pollutants to be allocated are conventional pollutants,
or if toxic pollutants are to be allocated on a toxicant-
by-toxicant basis. For example, if BOD/00 and copper
are to be allocated, then QUAL-2E and MICHRIV
might be the water quality models selected for the
allocation.
Sampling periods to collect data for model calibration
and verification are then selected Model calibration
refers to the process of adjusting model parameters
so predictions acceptably match field data Calibration
often requires that some of the input data, particularly
rate constants (eg. BOD decay rate) be adjusted
within realistic limits to provide better agreement
between observations and predictions. Model verifi-
cation is a comparison of model predictions against
an independent set of field data A model or model
component is verified when predictions and observa-
tions agree without having to arbitrarily adjust model
coefficients
Stream surveys used to calibrate and verify models
are typically intensive synoptic surveys These are
surveys that are.usually completed within a week or
so, and are intended to provide a definition of river
responses to a specific set of loadings
Since the models or calculation methods to be used in
the WLA process will eventually be adapted to the
river systems where sampling is to be conducted,
model adaptation to the system should be completed
prior to sampling The models are used to simulate
the parameters to be allocated and at the conditions
expected to be encountered during the surveys (based
on the best information available prior to sampling).
This will encourage the specialist to examine the
available data, determine what is missing, and to
estimate values of the missing data Then, by
performing sensitivity analyses (i.e., by varying
parameters and observing the effect on model
outputs), the specialist can establish which data are
more likely to influence model predictions and thus
establish sampling frequencies and location. Loca-
tions where water quality conditions change most
rapidly and where water quality standards are not
expected to be achieved are the critical areas to find
and sample
Stream survey design for model calibration and
verification can then be rationally executed with
informational needs fairly well defined Often, dye
studies are needed to accurately estimate pollutant
travel time through the river. Travel time reflects the
average velocity over distance, and can be quite
different from the velocity measured by a current
meter at a cross section, especially if the river cross
section changes from location to location. Normally,
travel time studies are conducted at more than one
stream flow so that travel times can be estimated at
the critical flow.
Sampling locations are established considering ac-
cessibility, historical locations, critical points of
maximum or minimum concentration, and other
locations where water quality standards are expected
to be violated Other considerations include intervals
between samples (smaller intervals are typically used
where stream response is most rapid) and point
source sampling. Sampling just below a point source
is risky because of the likelihood of obtaining un-
1-2
-------
Figure 1-1. Summary of stream survey design consideration for chemical (pacific approach and whole effluent approach to
waata load allocation.
Select Waata Load Allocation
Approach
Chemical Specific Approach
Criteria for Selection:
• Conventional Pollutants
Being Allocated
a Pollutant-By-Pollutant
Approach used for Toxicants
Summarize Information Needed
to Use Guidance
e River System Reaches
e Water Quality Parameters
to be Modeled
e Modeling Approach
e Critical Conditions
Select Periods for Stream
Surveys
e Model Calibration Period
e Model Verification Period
Set Up Model Prior to
Stream Surveys
e Develop Data Base to Run
Model from Available
Historical Data
Perform Model Simulations
e Identify Important Processes
e Selected Parameters for
Sensitivity Analysis
e Perform Simulations and
Analyze Results
Design Stream Surveys
e Summarize Informational Needs
e Design Dye Studies if Needed
e Establish Sampling Locations
e Determine When Samples are
Collected
e Determine Number and Types of
Samples
e Design Methods to Estimate
Rate Constants
e Estimate Duration of Surveys
Additional Considerations
e Finances
e Manpower
e Professional Judgment
and Site-Specific
Experience
Whole Effluent Approach
Criteria for Selection
e Complex Mixture of
Toxicants Discharged
e More Than One Discharger
in Close Proximity or NPS are a
Significant Component
« Comical Specific Evaluation
•« I r.practical
Summarize Information Needed
to Use Guidance
e River System Reaches
e Critical Conditions
e Discharge Locations
Select Periods for
Whole Effluent Toxicity
Stream Surveys
Perform Preliminary
Mixing Calculations
Design Stream Surveys
e Design Dye Studies
e Establish Sampling
Locations
e Determine Number of
Samples
e Estimate Duration of
Surveys
1-3
-------
representative samples (point source discharges may
not rapidly mix with stream flow), and often a mass-
balance calculation using the point source and a
location just upstream is preferable. Where the
pollutant sources cannot be adequately measured, a
downstream sample will be necessary to back
calculate the load from nonpomt source (NFS)
pollution, agricultural runoff, or point sources that
cannot be adequately measured
Special consideration is required for rate constant
determinations. Typically, rate constants, such as the
reaeration rate coefficient, are not directly measured
but are determined through a series of indirect
measurements, or are based on model calibrations
Field determinations of rate coefficients can be costly,
and the specialist should justify the need prior to
recommending this aspect of the field study.
Stream survey design and implementation must be
tempered by factors such as financial resources and
man-power limitations and should be conducted
during critical conditions if at all possible. The
judgement and experience of water quality specialists
who are not necessarily modelers but who have
considerable experience with the natural waters of
interest must also be weighed. The importance of the
data that are to be collected can help to guide and
prioritize sampling program activities. All environ-
mental monitoring tasks performed under EPA spon-
sorship must also be conducted under an approved
Q.A. project plan following guidance provided by the
EPA. Quality assurance is especially important when
sample number is limited due to other project
considerations.
The second approach to WLA of toxicants is called the
whole effluent approach. Streams that receive com-
plex or multiple effluent discharges may present a
complicated sampling problem. All potential pollu-
tants in complex wastes may not be identified nor
their interactions assessed In turn, pollutant bio-
availability may be difficult to measure. The EPA has
recently evaluated and validated this approach for
setting discharge limits based on effluent toxicity
objective of field sampling in support of the whole
effluent toxicity approach is to determine mixing
characteristics of the effluent in the stream or river
and to determine whether toxicity is decreasing due
to decay processes.
Every effort should be made to visit the proposed
sampling locations during a brief field reconnaissance
before executing the stream surveys for model
calibration/ verification. This wil I help to establish the
accessibility of the selected locations, or to decide if
for any other reason a sampling location change
should be made
For this approach, total toxicity in a river is treated
conservatively Under certain circumstances, an
effective decay rate can be estimated based on
toxicity decrease over distance below an outfall (6).
Traditional chemical-specific toxicant models are not
required for this approach.
The whole effluent approach may be used alone or in
many cases in conjunction with the chemical specific
approach to WLA. As pointed out in EPA policy, both
approaches will be needed in many cases. In this
manner it may be possible to develop a more complete
evaluation of instream effluent effects. The primary
1-4
-------
Chapter 2
Sampling Requirements for Waste Load Allocation Modeling
2.1 Model-Independent Considerations
Sampling requirements for water quality modeling
depend to some extent on the particular model or
calculation procedure being used. This in turn
depends on the type of problem being studied and the
level of detail required in the modeling analysis.
Models can range in detail from dilution models or
simple Streeter-Phelps type models of dissolved
oxygen to complicated models of stream ecosystems
which include many interacting processes and vari-
ables, for example, oxygen dynamics, nutrient cycles,
and algal and zooplankton dynamics The major
distinctions between different models are the specific
parameters and processes modeled, the equations
used to describe each process, the numerical tech-
niques used to solve the equations, and whether the
models are dynamic or steady-state.
However, in spite of these differences, all models
share many common features. As a result, many
sampling considerations are the same regardless of
the specific model or the particular WLA problem
being addressed. These model-independent consid-
erations are discussed in this section. Sampling
considerations specific to particular types of problems
and specific models are discussed in the following
sections.
2.1.1 Strtmm Geometry Dttt
All models require essentially the same types of
information to define the geometric characteristics of
the stream. Stream systems are divided into a series
of reaches for model analysis, with each reach
described by a specific set of channel geometry (i.e.,
cross-sectional dimensions) and flow characteristics
(i.e., flow rates, depths, and velocities or time of
travel). Reaches are defined between all major
tributary junctions and flow diversions, or whenever
stream geometry, hydraulic conditions, or biochem-
ical processes (i.e., sediment oxygen demand) within
the stream are expected to change significantly. The
models assume that these conditions are uniform
within each reach.
Each reach is in turn divided into a series of model
segments or computational elements in order to
provide spatial variation for the water quality analysis
(Figures 2-1 and 2-2). Each segment is represented
by a grid point in the model where all water quality
variables are computed. The number and size of the
segments depends on the spatial resolution desired.
Enough detail should belpr.video to characterize
anticipated spatial variation rn .va,.-. Ho uje to
different pollutant sources, dissolved oxygen sags,
and other significant processes within the stream. In
general, the model grid must have a much higher
resolution than the sampling network for computa-
tional reasons. For example. Figure 2-3 shows the
effects of varying the grid resolution on dissolved
oxygen predictions. The low resolution grid flattens
out the dissolved oxygen sag curve due to the effects
of numerical mixing in the model. Although 10
sampling locations are more than adequate to define
the dissolved oxygen profile in the field, the use of
only 10 computational nodes in the model results in
inaccurate predictions.
Channel geometry data are used to define the stream
configurations and segment characteristics, regard-
less of the particular model being used. This includes
both hand calculation methods and computerized
modeling techniques. Additional types of geometry
data may also be necessary depending on the
hydrologic algorithm used to route the flows through
the system. The basic types of data required for each
reach include.
1. segment or reach length
2. variation of channel width and cross-sectional
area with depth
3. bottom slope (or bed elevations)
4. variation of wetted perimeter or hydraulic radius
with depth
5. bottom roughness coefficient (Manning's n).
Variation of water depth with flow is also important,
but will be discussed later in the hydraulic data
section. All of the above parameters are typically
assumed constant for all model segments within a
defined reach.
Length and average slope over long distances can be
determined from topographic maps, while the other
variables usually require field surveys. The first two
2-1
-------
Figure 2-1. Example ttream network ehowing reaches end computational element* (1.2.3)
Mo«t Upstream
Point
Reach
Number
Computational
Element Number
data types, length and cross-sectional area, are
fundamental to any modeling study. The remaining
information may or may not be required, depending
on the type of hydraulic computations used in the
model. For example, if stage-flow relationships are
2-2
used to describe the hydraulics (e.g., QUAL-2E, and
SSAM IV (8). then only lengths and cross-sectional
areas are required to fully define transport through
the system. However, if Manning's equation (e.g.,
option in QUAL-2E) or the St. Venant equations (e.g..
-------
Figur* 2-2. Physical representation of • stream by mod*) segment* (adapted from (10)).
Control
Volume
V,
Segmented Stream Syttem
RECEIV-II (9), and WQRRS (10) are used to route the
flow, then the additional information (items 3 through
5 above) will be necessary for the hydraulic compu-
tations.
Many models internally compute the cross-sectional
area as a function of depth based on idealized
representations of the channel shape. For example, if
a trapezoidal channel is assumed, only the bottom
width and side slopes need to be specified. For a
rectangular channel, only the width is needed.
The level of detail required in describing the stream
geometry depends on the amount of variability in the
system. For streams which have uniform slopes and
cross-sections over the study area, only a few
transects will be necessary. However, in areas where
the channel geometry varies widely, the stream
should be divided into a series of representative
reaches, and enough transects measured along each
reach to adequately characterize the geometry. Three
to five cross-sections could be measured along each
reach, and the results could be averaged to define the
reach characteristics for the model. As a minimum,
one representative cross-section should be measured
in each reach. Some pool and riffle streams may
require dye studies and measurement of as many
cross sections as possible to obtain adequate stream
geometry.
2.1.2 Hydraulic Dmt»
Hydraulic data are needed to define the velocities.
flows, and water depths for the transport calculations
that are used to describe how pollutants move down
stream. Enough data are necessary to characterize
the hydraulic regime throughout the study area. This
includes the flows at the upstream boundaries of all
channels, as well as all significant tributary inflows,
lateral inflows (from groundwater or runoff), flow
diversions, return flows and stage at some locations.
In a general analysis, waste flows which represent a
significant portion (i.e., greater than 5 to 10 percent)
of the total stream flow should also be included in the
hydraulic analysis. Enough flow sources should be
characterized so that 90 to 95 percent of the total
stream flow is accounted for in the analysis.
While the upstream boundary flows, tributary flows,
and diversion flows can be measured directly, lateral
2-3
-------
Figure 2-3. Effects of grid resolution on predicted dissolved oxygen profile* using sn explicit finite difference solution scheme
9.0 H
8.0-
70
60.
50.
4.0-
o
u
c
•
2
o
i
5
3.0
Analytics I
. 10 Nodes 10,000'esch
40 Node* 2500 each
100 Mooes 1000 each
= 0.6/dav. k. = 2.3/d«v
10.00O 20.000 30.000 4O.OOO 5O.OOO
Distance Downstream, feet —
60.000
70.000
80.000
inflows from ground water or runoff must be esti-
mated from differences in measured flows at different
locations along the stream channel.
Many models allow the specification of stage-flow
relationships for each channel reach in the system.
This requires the simultaneous measurement of
water depth and flow for a series of flows over the
range of interest. While a minimum of two values are
necessary to construct a stage-flow rating curve.
three or more values are desired for more accurate
relationships. If possible, the flows measured should
cover the range of conditions to be addressed in the
WLA analysis. For a preliminary analysis it may be
possible to estimate the relationship between d, A. V
and Q at gauging stations. However, these stations
are rarely representative of long reaches because
these stations are located at control points in a stream
where a unique relationship exists between stage
and Q.
Stage-flow curves are constructed by plotting depth
versus flow on log log paper since depth and flow can
be related by an exponential equation of the form:
d— «. ^fc« /•} < |
= 81Q 1 (*-1}
where d = water depth
Q = flow
81 = coefficient of stage-flow relationship
bi = exponent of stage-flow relationship
The coefficient 81 and exponent b, of Equation (2-1)
are determined from the intercept and slope of the
log-log plot (Figure 2-4). Similar relationships can be
developed for cross-sectional area and velocity as
functions of flow.
a2QB2
a3QB3
(2-2)
(2-3)
where A
U
83 =
cross-sectional area
velocity
coefficient for cross-sectiorral area vs.
flow relationship
exponent for cross-sectional area vs. flow
relationship
coefficient for velocity vs flow relation-
ship
exponent for velocity vs flow relationship
2-4
-------
Cross-sectional area as a function of depth is obtained
from the channel geometry data, and velocity is
computed from the flow continuity equation (U. =
Q/A) These data are plotted against flow on log-log
paper to determine the values of the coefficients and
exponents in Equations (2-2) and (2-3) (Figure 2-4).
These parameters are required as input to certain
stream water quality models. There are considerable
geometry data available from: 1) USGS, especially
new gauging stations. 2) COE near reservoirs and
proposed reservoirs (also from Bureau of Reclama-
tion, TVA. Bonneville Power), 3) FEMA flood insur-
ance studies, 4) National Weather Service forecasting
centers. In areas where stream bed varies with time,
it is impordnt xo use the most recent geometry data.
Both the stream geometry and flow information are
critical to the transport calculations. When the stream
geometry varies widely within reaches and is difficult
to characterize in detail, or when lateral inflows are
not well defined (for example, because access
problems limit measurement), it is often useful to
supplement the hydrologic and geometric data with
travel time studies using tracer techniques, typically
with rhodamine WTdye. This information can be used
to adjust the geometry or flow data so that model
transport calculations match the results of the dye
study. This calibrates the transport portion of the
water quality model by defining the geometry or flow
data to produce the net transport observed in the field.
2.1.3 Mfttoroiogic*/ D»t»
Because temperature influences dissolved oxygen
saturation and the rates of almost all of the chemical
and biological processes occurring in streams, many
water quality models include options for simulating
temperature. Meteorological data are necessary to
perform the heat budget computations in these
models. Heat transfer at the air-water interface
depends on several processes including short-wave
solar radiation, long-wave atmospheric radiation.
Figure 2-4. Derivation of croM-eactional are* va. flow and velocity relationship* from atage-ftow data (from (11)}.
20.0
100-
• 50 H
Q
2.0-
1.0
Depth =0.312 Q"
2000
1000-
200
100
Are» = 19.5 Q
20 50 100 200
Flow | eft)
500
20 50 100 200
Flow (eft)
500
2.0
\
0.5
I 1
0.2
0.1
Velocity* 0.0613
20 50 100 200
Flow (eft)
500
2-5
-------
long-wave back radiation, convective heat exchange,
and evaporative heat loss. Most models compute
each of these processes separately and add them
together to give the net heat f I ux at the water surface.
The alternative method is the equilibrium tempera-
ture approach in which al I of the above processes are
combined into two parameters: the equilibrium
temperature and the surface heat exchange coef-
ficient, both of which vary dynamically with the
meteorological conditions. In fact, the equilibrium
temperature formulation is essentially equivalent to
the linearized version of the total heat budget which is
used in most water quality models, including
QUAL-2E.
The basic types of meteorological data required are
essentially the same regardless of the particular
model being used:
1. solar radiation
2. cloud cover
3. air temperature
4. relative humidity (or wet bulb temperature or
dew point temperature)
5. wind speed
6. atmospheric pressure.
Many models compute incident solar radiation inter-
nally in the model as a function of latitude, longitude,
day of the year, time of the day, and atmospheric
scattering and absorption of light due to dust. Cloud
cover data are then required to compute the amount
of solar radiation reaching the water surface. The
alternative approach used in other models is to input
measured solar radiation directly. In mountainous
areas, canyons, or in areas where riparian vegetation
is dense, additional reduction in solar radiation due to
topographic and vegetation shading should be in-
cluded in the model. This is handled by an additional
shading coefficient, by detailed formulations which
compute the shading dynamically (12), or by inputing
net solar radiation values which include these effects.
The five meteorological variables listed above can
often be obtained from nearby weather stations. Most
NOAA class A stations have monthly averages of
these variables available. Long term monthly aver-
ages of these parameters based on several years of
historical data are also available in the "Climatic
Atlas" published by the National Oceanographic and
Atmospheric Administration (13). More detailed
records of meteorological data at 3-hour intervals can
often be obtained on magnetic tape from the National
Climatic Data Center, Federal Building, Asheville,
North Carolina 28801 (phone (704) 259-0682).
Existing weather stations are usually adequate when
studying larger rivers, streams near weather stations.
or when water temperature and volatilization are not
critical components of the model study In other
cases, temporary weather stations can be set up. A
single station is generally sufficient However, in
areas where solar radiation, atmospheric pressure
and temperature varies over the length of a river
(greater than 1000 feet in elevation), it may be
desirable to set up two stations, one near the
upstream boundary and one near the downstream
boundary of the study reach.
2.1.4 Wattr Quality Data
Given the semi-empirical nature of water quality
models, water quality data are necessary to setup,
calibrate, and verify any water r"~'-*-' »-'»J~' *<~T-\\
data are needed for all param tar? whi«~h will r n
simulated. For models like QUML-2E that simulate
conventional pollutants, this may include tempera-
ture, dissolved oxygen, carbonaceous BOO, phos-
phorus, nitrogen (ammonia, nitrite, and nitrate),
coliforms, chlorophyll a or phytoplankton dry weight
biomass, and conservative constituents such as total
dissolved solids. Some models also include additional
constituents such as total inorganic carbon, alkalinity,
pH, inorganic suspended solids, suspended organic
detritus, periphyton, zooplankton, and benthic orga-
nisms. Toxic fate models require data for the specific
chemicals under investigation.
It is only necessary to collect data for the particular
constituents and processes which are being evalu-
ated, plus any other variables which significantly
affect these constituents. For example, if coliforms
are not of interest, there is no need to colled data for
them even though they may be included in the model
since they do not influence the other constituents.
Any arbitrary value could be assigned to coliforms. or
they could be set equal to zero when running the
model. Some models [e.g., QUAL-2E], DOSAG3 (14),
RECEIV-II, WQRRS (14), SSAM IV (8); HSPF (15)
include options which allow the user to "switch off"
many of the model constituents when running the
model. This bypasses the computations for
parameters which are of no importance in a particular
application.
This section discusses some of the water quality
sampling considerations which are basic to all WLA
studies. Sampling considerations which pertain to
particular types of problems (e.g.. DO/BOD analyses)
will be discussed later.
2.1.4.1 Sampling Locations
After an initial estimate of the constituents and
parameters which must be sampled in a modeling
study, it is necessary to determine where, when, and
how often the samples should be taken. The minimal
instream sampling effort should include the following
locations (Figure 2-5):
2-6
-------
Figure 2-6.
Recommended location* for • minimal sampling
program.
Point .2
Source
Key
t Upstream Boundary
2 Point Source
3 Upstream of Point Source
4 Mouth of Tributary
5 Upstream of Tributary
6 Upstream of Nonpomt Source
7 Downstream of Nonpomt Source
8 Downstream End of Study Area
1. Upstream end (headwater) of each stream reach
and tributary being modeled.
2. Mouths of all significant tributaries which are
not included in the model grid, just above their
entrances to the main stream.
3. Effluent samples of all significant point sources
before they enter the stream.
4. Upstream and downstream ends of stream
sections where nonpoint sources are expected
to be significant.
5. Downstream end of the study area.
Water quality data are needed at the upstream
extremities of all channels in the modeled stream
system in order to define the upstream boundary
conditions (i.e., flows and concentrations) for the
computations. The model starts with these boundary
conditions and routes the water along the channels,
simulating all of the chemical, biological, and physical
processes which act to change the concentrations of
the various constituents The water quality samples
at the upstream boundaries also define the back-
ground concentrations in the study area before
additional pollutant loads are added to the stream
For tributaries which are not explicitly included in the
model grid, water quality data are needed just above
their mouths in order to define tributary loading rates
for all constituents. However, if the flow contributions
and mass loadings of tributaries are insignificant in
comparison to the main channel flows and mass
fluxes (i.e., less than 5 percent), they can usually be
omitted from the analysis. Loadings due to tributaries'
which are included as part of the model grid are
computed internally in the model based on the
specified upstream boundary concentrations at the
head of the tributary and the simulated water quality
changes between the tributary headwater and con-
fluence with the main stream.
In addition to instream concentrations, effluent data
are needed to characterize pollutant loadings due to
all significant point source discharges. These data
can be obtained from the dischargers, NPOES permit
holders, and federal, state, and local government
regulatory agencies. However, it is most desirable to
collect point source data during the survey, as
historical data bases may not be indicative of survey
loads.
In areas where significant nonpoint source loadings
are known to exist, both the flow rate and constituent
concentrations should be measured in the stream
just above and below the area of the loading. If this
area is not so large that other water quality changes
are likely to occur during the travel time through the
area, it is reasonable to assume that the changes in
concentrations are due to the nonpoint sources and to
use these differences as a basis for estimating the
loads.
Water quality data should be collected at the down-
stream end of the study area for calibration and
verification. While a single downstream station is the
minimum requirement for short stream sections with
no major tributaries, additional sampling stations are
desirable to provide more spatial data for calibrating
and verifying the model. Logical locations for addi-
tional stations are biologically sensitive areas, areas
where water quality standards may be violated, areas
just above major tributaries or point source loadings,
and areas where stream changes may significantly
cause changes in kinetics. The latter locations allow
independent calibration of stream sections between
each tributary or discharge based primarily on
biochemical processes within the stream without the
complication of water quality changes associated
with major inflows or discharges. Water quality
below tributary junctions or waste discharges can be
2-7
-------
directly computed based on data above the junction or
discharge site and the tributary or point source
loading rates using simple flow weighted mixing
computations:
Ce =
_ Q. C. + Qi C,
(2-4)
Q.
where CB = average concentration below tributary or
discharge
C« = concentration in stream above the tribu-
tary or discharge
131 = concentration in tributary or discharge
Q« = stream flow above the tributary or dis-
QH«r~«»
Qi = t ibutary or dischar e flow rate
These values can then be used as upstream boundary
conditions to calibrate the next section of the stream.
Since most stream models are one-dimensional,
water quality is assumed to be well mixed and
uniform over each cross-section of the stream.
Therefore, samples taken immediately downstream
of a discharge or tributary would probably not match
conditions in the model unless they were taken far
enough downstream for complete cross-sectional
mixing to occur (Figure 2-6) (see Section 3.4 for a
method to estimate the distance for complete mixing).
If a stream branches into two separate channels
moving downstream, it is also useful to include a
sampling station at the head of the branch to define
upstream conditions in each reach.
In addition to the above sampling locations which are
based on the stream system configuration and waste
discharge locations, it is desirable to include more
stations where significant water quality gradients are
expected, for example, dissolved oxygen sags below
waste discharges. These stations provide data to
calibrate and verify the ability of the model to predict
important water quality variations. The appropriate
locations for these additional stations are often
difficult to determine in advance. Simplified screening
calculations or preliminary model runs can often be
useful in locating these stations. Guidance for
determining these locations will be discussed in later
sections where sampling considerations for specific
types of water quality problems are discussed.
2.1.4.2 Sampling Time and Frequency
If possible, water quality sampling should be con-
ducted during periods similar to the critical design
conditions which will be used in the WLA analyses.
These generally represent some type of "worst case"
situation, such as summer minimum flow and
maximum temperature conditions. The procedure for
determining these conditions is described in Book VI
(Design Conditions) of the waste load allocation
guidance. The selected design conditions will prob-
ably represent an extreme event such as a 7-day.
10-year low flow (rQio) which occurs on the average
every 10 years. Direct sampling of such conditions
may not be possible and although similar flows may
occur averaging shorter periods (e.g., 1 or 2 days)
each year the sampling period may not be sufficiently
long to accomplish all of the program goals. Therefore,
the sampling program should be conducted at times
most likely to approach these conditions when the
same water quality processes are important.
In addition to the calibration data, another set of
water quality samples should be collected under
different flow and water quality conditions for pur-
poses of model validation. If several flow or water
quality conditions will be evaluated in the WLA
analyses, the calibration and validation samples
should be collected at times which will bracket the
conditions of the analyses.
The duration and frequency of water quality sampling
depends to a large extent on whether a steady-state
or a dynamic model will be used. Because they are
easier to apply and require less data, steady-state
models are generally used in WLA analyses. Steady-
state models compute water quality conditions
assuming everything remains constant through time.
This includes:
• flows and stream geometry (depths, widths, etc.)
• meteorological conditions
• temperature and quality of the water entering the
up-stream boundary of the reach being modeled
(upstream boundary conditions)
• temperature, quality, and flow rates of all tribu-
taries
• temperature, quality, and flow rates of all nonpoint
and point source loadings
• rates of all physical, chemical, and biological
processes occurring in the stream.
Steady-state models simulate spatial (downstream)
variations in the above factors, but not temporal
changes. These models are appropriate for predicting
water quality conditions at different locations in the
stream when the above conditions do not change
significantly with time. Since the travel time through
an impacted stream reach is generally on the order of
days, it is reasonable to assume that hydrologic and
meteorologic conditions can remain fairly constant
over this period and to apply a steady-state model.
Seasonal variations can be analyzed by repeatedly
running the model for different scenarios, for example
monthly average conditions, monthly extreme condi-
tions, etc. The major limitations of steady-state
models are that they do not account for continuous
flow variations or transient events such as storms or
toxic spills, and that they do not directly simulate the
2-8
-------
Ftgun 2-6. Recommended aampaling location* at point tourcei.
Sampling Locations
fa)
Aerial View of River
Sampling Locations
Diacharge
Vertical
Mixing
Zone
Side View of Rivar
diurnal dynamics of temperature and oxygen. Some
quasi steady-state models are available which simu-
late these latter effects (e.g., 16,17,3).
Since steady-state models assume conditions remain
constant with time, it is important to conduct the
sampling program during a period when this assump-
tion is valid Synoptic surveys (e.g., sampling all
stations over 1 to 2 days) should be conducted to the
extent possible so that water quality conditions at
different locations are not affected significantly by
changes in the weather or variations in the waste
discharges. However, since temperature varies dt-
urna lly and temperature influences the process rates
of most biological and chemical reactions, some
variability will be inevitable in the sampling results. If
diurnal variations are important (for example, in some
dissolved oxygen problems where algal activity is
significant), then a 24-hour survey should be con-
ducted for at least one station (preferably the station
at the 00 sag) and usually more. Short-term intensive
surveys (diurnal measurements over 2 or 3 days) are
recommended in all water quality studies, since this
will provide enough data for sample variability (e.g.,
variances, confidence intervals, etc.) to be estimated.
The alternative approach to steady-state modeling is
dynamic modeling. Dynamic models simulate streams
in the same basic manner as steady-state models
(i.e., they route water downstream and compute the
physical, chemical, and biological processes occur-
ring in the stream and the resulting changes in the
water quality parameters). However, in addition.
dynamic models compute the continuous changes
which occur over time due to variations in stream
flows, upstream water quality and temperature.
2-9
-------
tributary inflows, nonpomt and point source loadings,
meteorology, and processes occurring within the
stream. In dynamic modeling, all of the factors which
are assumed constant for a steady-state analysis are
free to vary continuously with time. This allows an
analysis of diurnal variations in temperature and
water quality, as well as continuous prediction of
daily variations or even seasonal variations in water
quality.
Dynamic model studies generally require much more
detailed sampling programs than steady-state stud-
ies. Enough data must be collected to define the
temporal variations in water quality throughout the
simulation period at the upstream ends of all stream
channels and the major pollutant loadings so that the
model boundary conditions can be specified. Since
dynamic models are used to study transient events
such as combined sewer overflows during storms,
toxic spills, and diurnal variations in temperature or
dissolved oxygen, the duration and frequency of the
sampling should be commensurate with the duration
of the event plus the travel time through the study
area. For toxic spills, one travel time plus the time for
the trailing edge to pass is necessary to track the
toxicant through the system. For storm runoff prob-
lems, the duration of the storm runoff should also be
added to the sampling period since the pollutant
loadings and stream hydrologic response will vary
throughout the storm runoff period. For diurnal
studies of temperature or dissolved oxygen, sampling
at specified intervals (1 or 2 hours, for example)
should be conducted over at least 24 hours.
important to locate the stations in places that will
provide the most information Preliminary model
calculations can be used to determine the best
locations for sampling, as well as the critical times for
sampling if dynamic analyses are being performed
For example, when analyzing dissolved oxygen
problems in streams with several discharges, more of
the sampling effort should be allocated to areas
where water quality standards are most likely to be
violated (Figure 2-7). Also, areas where large water
quality gradients exist should be sampled more
thoroughly. These areas can be determined with
Streeter-Phelps type calculations or simplified com-
puter r. —.—.,., .u—- 9S or waste loadings which
are ni J. ...r._... _.ii . ten be omitted from these
preliminary analyses(particularly if hand calculations
are being used). Simple mixing calculations can be
used to help determine which waste sources are
significant. Mixing zone calculations can also be
made to estimate the distance required for complete
mixing of the waste water with the stream, and to
estimate concentrations within the mixing zone. Rate
coefficients and model parameters can be estimated
from literature values (18,19.20) before site specific
measurements are available. For important param-
eters such as the BOO decay rate (K«), sensitivity
analyses can be performed to evaluate the effects of
different Kd values on the location of the DO sag.
These analyses should provide enough information
so that sampling stations can be located on the critical
portion of the sag curve.
Long-term dynamic simulations of seasonal varia-
tions in stream water quality may be impractical.
Where seasonal variation is of interest, the general
practice is to run a steady-state model or a dynamic
model (with short term simulations) several times for
different sets of conditions that represent the full
spectrum of conditions expected over the period of
interest. Enough data should be collected to char-
acterize the seasonal variations, and to provide
adequate data for calibrating and verifying the model.
If possible, enough data should be collected to cover
the full range of conditions of the model analysis. As a
minimum, this should include conditions at both
extremes of the seasonal range, as well as a few
intermediate conditions (e.g., monthly averages).
2.1.4.3 Use of Models in Designing Sampling
Programs
Models can be very effective tools in the design of
sampling programs. This includes both computerized
models and simple hand calculation techniques.
Since sampling resources are generally limited, it is
2.1.5 Plug Flow Sampling
Plug flow sampling is a type of instream sampling
where a particular parcel of water is followed as it
moves downstream, and samples of water quality are
taken from the same parcel of water at different
locations. Typically a dye (e.g., rhodamine WT) is
injected into the river and is used to determine when
to sample at selected downstream locations. Passage
of the peak dye concentration indicates when to
sample. Centroid-to-centroid measurements (rather
than peak-to-peak measurements) are not used
because centroids are not readily determinable in the
field.
Plug flow sampling is particularly useful for rate
constant determinations. Suppose, for example, that
waste loading to a river segment is highly time
variable. By sampling a particular slug of water as it
moves downstream, the effects of time variability in
waste loading can be eliminated.
Accessibility to the stream or river at multiple
locations is necessary to implement plug flow samp-
ling. For larger rivers, a boat may be appropriate to
2-10
-------
Figure 2-7 Allocation of sampling effort based on preliminary analyses.
Point
Source
Point
Source
Point
Source
1
Less Intensive Sampling Here
Key
DO Saturation
00 Profile
Stream Standard
More Intensive
Sampling Here
54
River Mile. Above Mouth
move from location to location, while for smaller
streams, land transportation may be easier.
Before samples are collected, the dye should be well-
mixed across the river. Section 3.2 provides guide-
lines on the distance required. For large rivers,
complete mixing can take many miles, and plug flow
sampling would be inappropriate.
It is not always necessary, or even desirable, for the
dye to be injected at the upstream boundary of the
segment under investigation. The dye can be injected
at some distance further upstream so that when the
dye reaches the segment boundary, it has attained its
one-dimensional profile. If the upstream boundary is
a wastewater treatment plant, the effluent from the
plant is sampled at the time the peak dye concentra-
tion passes the plant, and subsequent samples are
taken at selected downstream locations when the
peak dye concentration passes those locations. Travel
times between locations are calculated to determine
stream velocities. Two methods for Lagrangian
sampling are: 1) find peak every 2 hours or so by
moving boat, or 2) await arrival of peak at predesig-
nated sites.
2.2 Sampling Requirements for
Conventional Pollutants
2.2.1 G»n*r»l Modeling Approaches
Most waste load allocations in streams focus on
dissolved oxygen. Dissolved oxygen dynamics depend
on the interactions of several constituents and
processes. The constituents include dissolved oxygen,
carbonaceous BOD, nitrogenous BOD (ammonia and
nitrite), temperature, and in some cases phytoplank-
2-11
-------
ton. penphyton, and aquatic plants. The major
processes include (Figure 2-8):
Reaeration
CBOD decay
CBOO settling
Sediment oxygen demand
Nitrification
Photosynthesis
Respiration
These constituents and processes are typically
modeled by a set of coupled mass balance equations
Dissolved Oxygen
~ =K.(0»,-Oa)-KaL-
NOj ^ (o3 ^ - cu r) A
(2-5)
Carbonaceous BOD
dL
= - *„ L - K. L = - K, L
dt
Nitrogen forms
NH3
dNQ3
dt
= KwaNOj- tnv A
(2-6)
(2-7)
(2-8)
(2-9)
Algae
«» A
_,__-* A
(2-10)
For identification of coefficients for these equations,
see p. 2-34.
other equations may be coupled to the dissolved
oxygen equation in an indirect way Michaelis-
Menten type saturation kinetics are typically used to
compute nutrient limitation effects on algal growth,
and often light limitation as well. Other saturation
relationships (21) are also used for light limitation.
Periphyton and aquatic plants are rarely included in
water quality models because of the difficulty in
predicting these parameters. When they are, they are
modeled by equations analogous to those used for
algae (Equation (2-10)), except that the settling term
is replaced by a sloughing or nonpredatory mortality
term.
The above equations give the general framework
which forms the basis of a II dissolved oxygen models.
However, many models use a simplified framework
which ignores or combines some of the processes.
For example, in systems where photosynthesis and
respiration are not important, the corresponding
terms and equations can be left out of the analysis
(e.g., DOSAG1 [22], and SNSIM [23]) Simple models
and hand calculation techniques often lump the
nitrogen cycle into a single nitrogenous BOD equation
analogous to Equation (2-6) (e.g., DOSAG1, SNSIM),
or else combine the nitrogenous and carbonaceous
BOD into a single constituent representing total BOD
(24). In the latter case, only the first three terms of
Equation (2-5) and a total BOD equation analogous to
Equation (2-6) are left in the model.
Even when the nitrogen cycle is not lumped into a
BOD equation, models differ in the number of stages
included in the cycle. The complete sequence should
include hydrolysis of organic nitrogen to ammonia
and oxidation of ammonia to nitrite and nitrite to
nitrate. However, most models do not even include
organic nitrogen as a separate constituent (e.g.,
QUAL-II. DOSAG3, WQRRS). However, QUAL-2E
does have organic nitrogen and organic phosphorus
capability. Many models also leave out nitrite so that
ammonia is oxidized directly to nitrate in the model
equations (e.g., SSAM IV). As a result, some of the
constituents and process rates may take on a different
meaning since they represent two or more consti-
tuents and corresponding decay processes combined.
In addition to dissolved oxygen analyses, other
conventional pollutant problems such as ammonia
toxicity and eutrophication are sometimes important
in waste load allocations.
The above equations are simplified in that they do not
include the pollutant loading or transport (advection
and dispersion) terms. All of the process rates are
temperature dependent. In addition, algal growth
depends on light, phosphorus, and other nutrients, so
Ammonia toxicity is due to the un-ionized form of
ammonia. The un-ionized fraction of total ammonia
increases with pH and temperature. Figure 2-9 shows
this relationship. Most currently available water
quality models do not simulate un-ionized ammonia
2-12
-------
Figure 2-8. Proc«(M( affecting dissolved oxygen.
Phyioplsnkron
Periphyton
Aquatic Plants
NH4*
NO]
NOj
or pH. Therefore, waste load allocations which involve
ammonia toxicity must usually be based on totat
ammonia simulations using equations such as (2-7)
through (2-10) above in combination with field
measurements of pH and temperature. Un-ionized
ammonia concentration can be calculated from
model-projected total ammonia and a relationship
such as shown in Figure 2-9.
Eutrophication analyses require models which simu-
late nutrient and algal dynamics. Phosphorus and
nitrogen are generally the only nutrients considered.
The major processes include algal uptake, algal
excretion, sediment release, and nitrification. The
mass balance equations for the nitrogen cycle and
algae were given above in Equations (2-7) through
(2-10). The only additional equation required is a
mass balance for orthophosphate, which is typically
expressed as:
dPQ4
dt
(2-11)
2.2.2 Modfl Data R*qufr»m»ntt
This section summarizes the data requirements for
the different types of models used to allocate
conventional pollutants. The modeling approaches
range from simple hand-calculation techniques to
complex computer models. Dissolved oxygen anal-
yses using Streeter-Phelps type hand-calculations
are probably the most commonly used techniques in
waste load allocation analyses. Simplified methods
are limited for eutrophication analyses since several
constituents with complex interactions are involved.
However, a few hand-calculation techniques for
predicting algal concentrations and their effects on
dissolved oxygen are described in Chapter 2 of Book II
of the WLA guidance documents (see Table 1-1).
Table 2-1 summarizes the data requirements for the
various hand-calculation methods available.
The models QUAL-II, NCASI (26). and QUAL-2E are
probably the most widely used computer model for
predicting the effects of conventional pollutants in
2-13
-------
Figure 2-9 Effect of pH and temperature on un-ionized
ammonia (From (26)).
10 15 20 25
Temperature (°C)
30 35
streams. The data requirements for QUAL-2E are, in
general, the same as most other stream models,
except models such as DOSAG1 which are restricted
to simple dissolved oxygen analyses and therefore
require less data. QUAL-2E simulates the following
constitutents:
Dissolved oxygen
Biochemical oxygen demand
Temperature
Algae as chlorophyll a
Organic nitrogen
Ammonia
Nitrite
Nitrate
Organic phosphorus
Dissolved phosphorus
Coliforms
Arbitrary nonconservative constituent
Three conservative constituents.
The model equations and process formulations are in
general identical to those discussed in Equations (2-
5} to (2-11) for dissolved oxygen, nutrients, and
phytoplankton. Figure 2-10 shows the interactions of
the various constituents, and Table 2-2 lists the
processes which are simulated for each constituent
Table 2-3 compares QUAL-2E with other models
commonly used in WLA analyses with respect to the
constituents simulated
Table 2-4 summarizes the input data requirements
for QUAL-2E. Note that many of the process rates can
vary with each reach. This feature is useful since
waste characteristics may vary between different
discharges, resulting indifferences in the BOD decay
rates and nitrification rates at different locations in
the stream. Other process rates such as sediment
•xygen demand, phytoplankton settling rates, and
,eaeration rates may also vary with distance, since
these are affected by the hydraulic characteristics of
the stream.
QUAL-2E is capable of running in either a steady-
state or a quasi-dynamic mode. The dynamic option is
used primarily for simulating diurnal variations in
dissolved oxygen and temperature since (he stream
flows, point source loadings, and nonpoint source
loadings cannot be varied during the simulation. Only
the constituent concentrations at the upstream
boundaries, the meteorological conditions, and the
resulting water quality response are free to change
Table 2-5 compares the general features of QUAL-2E
with other computer models used in waste load
allocation analyses. DOSAG1 and SNSIM are limited
to steady-state DO/BOD analyses, while QUAL-2E
and RECEIV-II can be used for eutrophication anal-
yses as well as dissolved oxygen analyses QUAL-2E
and RECEIV-II both simulate the effects of photo-
synthesis, respiration, and temperature on diurnal
variations of dissolved oxygen. RECEIV-II is truly
dynamic since it simulates continuous temporal
variations in stream hydraulics and waste loadings.
QUAL-2E assumes these features remain constant,
but allows the meteorology and water quality condi-
tions downstream of the upstream boundaries to
vary.
2.2.3 Sampling Guidatinas
2.2.3.1 Constituents Sampled
The specific constituents which must be sampled, as
well as the sampling frequency, depend to some
extent on the particular modeling framework which
will be used in the waste load allocation analysis. The
selected model should include all of the processes
which are significant in the stream being analyzed,
without the unnecessary complexity of processes
which are insignificant. A few preliminary measure-
ments may be useful to define which processes are
important
2-14
-------
Table 2-1. Data Requirements tor Hand-Calculation Techniques Described in WLA Guidance Documents and Screening Manual
(27) For Analytic of Conventional Pollutants
AKpl Predictions Algal Predictions
Streeter-Phelp* NH3 ToxkHty Without With Algal Effects on Algal Effects
Data Requirements PO Analyses* Calculations'' Nutrient Limitation1 Nutrient limitation* Dairy Average PCX on Diurnal DO
Hydreu/ic and Geometry DM
Ftowratef X
Velocity X
Depth X
Croas-eectional are* X
Reach length X
Constitutent Concentration**
DO X
CBOO. NBOO X
MM,
Temperat' - X
Inorganic '
Inorganic i
Chlorophyll t
pH
DO/BOD Parameters
Reaeration rate coefficient X
Sediment oxygen demand X
CBOD decay rate X
CBOD removal rate X
NBOD decay rate X
NHj oxidation rate
Oxygen per unit chlorophyll a
Algal oxygen production rate X
Algal oxygen respiration rate X
Pnyrop/anAron Parameters
Maximum growth rate
Respiration rate
Settling velocity
Saturating light intensity
Phosphorus half-saturation constant
Nitrogen hatf-saturation constant
Phosphorus to chlorophyll ratio
Nitrogen to chlorophyll ratio
Light Parameters
Dairy solar radiation
Photoperiod
Liprii «yiinctK>n coefficient
X
X
X
X
X
X
X
X
X
X
X
X
X
x«
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
'Streeter-Phelps DO calculations are described in Chapter 1 of Book II of the WLA guidance documents (Table 1-1) and the Screening
Manual 127).
"Ammonia toxicrty calculations are described in Chapter 1 of Book II of the WLA guidance documents.
'Algal predictions and their effects on DO are discussed in Chapter 2 of Book II of the WLA guidance documents.
dFlow rates are needed for the river and all point sources at various points to define nonpoint flow.
•Constituent concentrations are needed at the upstream boundary and all point sources.
'Chlorophyll a concentrations are also needed at the downstream end of the reach to estimate net growth rates.
The absolute minimum sampling requirements for all
dissolved oxygen studies should include dissolved
oxygen, temperature, carbonaceous BOD, and total
Kjeldahl nitrogen (measure of nitrogenous BOD),
since these are fundamental to any dissolved oxygen
analysis. BOD is typically measured as 5-day BOD
(BODs). However, a few measurements of long-term
or ultimate BOD (BOD*>) are also necessary to
establish the BODzo/BODg ratio since ultimate BOD
is simulated in the models. If a model which considers
only a total BOD component is selected, the analyst
should be aware that nitrogenous BOD and carbon-
aceous BOD decay at different rates, which will cause
both the composition of the remaining BOD and the
net decay rate to change as the waste moves
downstream. Therefore, the total BOD approach
should only be used in situations where the nitro-
genous components of the waste sources are known
to be unimportant (e.g., less than 10 percent of the
total BOD).
In addition to total Kjeldahl nitrogen (TKN), ammonia
and nitrate (or nitrite plus nitrate) should be measured
in both dissolved oxygen and eutrophication studies
for models which simulate the nitrogen cycle. Even if
they are not modeled, ammonia, nitrate, and nitrite
2-15
-------
Figure 2 10 Major constituent interaction* in QUAL-2E (3)
Dis^nlvM
1
SOD
r/'f'n
data are useful for estimating the nitrogenous BOD
decay rate or ammonia oxidation rate as discussed
below in Section 2.2.3.2. Ammonia, pH, and temper-
ature must be measured in all studies involving
ammonia toxicity. In streams where algae activity is
significant, diurnal variations in pH as much as 1.5
units per day may occur. The potential effect of pH
variation on ammonia toxicity should be taken into
effect when designing a sampling program.
For models which simulate algae, concentrations of
algal dry weight biomass or chlorophyll a should be
measured. Orthophosphate concentrations and light
extinction coefficients
-------
T»bU 2-3.
Non-Toxic Constituents Included In Stream Models
W«l»' OmlitY V
CBOO or
To«l
Tot Org
Model Hunt *)««*f*oc« DO BOO NBOO SOO Temp P P PO»
WQAV (271 XXX
OOSAG1 (221 XXX
DOSAG3 1141 X X X*
SNSIM 1231 XXX
QUA1-II OJt XXX*
DUAL-IM (31 XXX*
R£Cfrv.ii Ml X X X*
WASP (281 XXX*
AESOP (291 XXX*
WSPf 1151 XXX*
HAR03 (301 X X
FEDBAK03 (311 X X
MCTONM XXX*
EXPLORE 1 (32) XXX*
WQRRS (10) X X X*
*NBOD s»mui«t«d as nitrification of •rtmonw
* *T»mp«r»iur« specified by mod** u*«rt
Table 2-4 Model Input Parameters
Input Parameter
Disso/ved Oxygen Parameters
Reaeration rate coefficients
0; consumption per unit NH3 oxidation
0; consumption par unit N02 oxidation
Oj production per unit photosynthesis
Oj consumption per unit respiration
Sediment oxygen demand
Carbonaceous BOO Parameters
CBOO decay rate
CBOO settling rate
Oroen/c Nitrogen
Hydroliie to ammonia
Ammonia Parameters
Ammonia oxidation rate
Benthic source rate
Nitrite Parameters
Nitrite oxidation rate
A/rfrafe Parameters
None
'Oryanic Priospnorus
Transformed to diss. p
Pnospnare Parameters
Benthic source rate
Phytop*ari«ton Parameters
Maximum growth rate
Respiration rate
Settling rite
Nitrogen half-saturation constant
Phosphorus half-saturation constant
Light hall-saturation constant
Light extinction coefficient
Ratio of chlorophyll a to algal biomass
Nitrogen fraction of algal biomass
Phosphorus fraction of algal biomass
ColHorm Parameters
Die-off rate
XXX X
X"
X X" X
X X"
XX X
XX X
X X" X X
XX" XX
XX" XX
XX X
X"
X"
XX" XX
XX X
forOual-2E
Variable Variable
by Reach with Time
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Toi Org AJg««r CM-« pUxitton pH Alt
X
XXX X
XXX X
XXX X
X X X X X X
XX XXX X
XX XXX X
X X X X X
X X X X X X
X X X X X X X
X X X X X X
Input Parameter
A/onconsaYva'ive Corxfrfuwr)' Parameters
Decay rate
Meteoro/oorc*/ Dtlt
Solar radiation
Cloud cover
Dry bulb temperature
Wet bulb temperature
Wind speed
Barometric pressure
Elevation
Oust attenuation coefficient
Evaporation coefficients
Srreem Geometry Oatt
Corss-sectionsl area vs. depth
Reach lengths
Hydnultc Oft* (Stage-Flow Curve Option)
Coefficient tor stage-flow equation
Exponent for stage-flow equation
Coefficient for velocity-flow equation
Exponent for vetocfty-ftow equation
Hydrtulic Dfti (Manning's Equation Option)
Manning's n
Bottom width of channel
Side stopes of channel
Channel slope
flowDMa
Upstream boundaries
Tributary inflows
Point sources
Nonpoint sources
Diversions
Consfrruenr Conc*ntrttion*
Initial conditions
Upstream boundaries
Tributary inflows
Point sources
Nonpoint sources
X X
X X
X X
Variable
by Reach
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Colrlorm
IDS B*C!*ri«
X X
X X
X X
X X
X X
X
X
X X
X X
Variable
with Time
Yes
Yes
Yes
Yes
Yes
Yes
Yes
dissolved oxygen and eutrophication analyses. Some
dissolved oxygen models include the effects of algal
photosynthesis and respiration without actually
simulating algae (e.g., 16,17). This can be done in
many cases by using photosynthesis and respiration
fluxes obtained from light-dark bottle measurements,
or by measuring the diurnal DO variations and
superimposing them on the daily average concentra-
tions predicted in the model, typically assuming a
sinusoidal relationship.
2-17
-------
Table 2-5. Comparison of QuaMI With Other Conventional Pollutant Models U»ad in Waste Load Allocation* (Adapted from (11)
Variable
Temporal Variability i n^.ny
Model Water Quality Hydraulics Rates
DOSAG-I Steady-state Steady-state No
SNSIM Steady-state Steady-state No
OUAL-II Steady-state Steady-state No
or Dynamic
RECEIV-II Dynamic Dynamic Yes
Types of
Loads
multiple
point
sources
multiple
point
sources &
nonpoint
sources
multiple
point
sources &
nonpoint
sources
multiple
point
sources
Spatial Water
Dimensions Body
1 -0 stream
network
1-0 stream
network
1-D straan
network
1 -D stream
or network or
2-D well-
mixed
estuary
Water Quality
Parameters
Modeled
DO. CBOD. NBOD.
conservative
DO. CBOD. NBOD, conserve
tive
DO, CBOO. temperature, am-
monia, nitrate, nitrite, algae.
phosphate, coliforms, non-
conservative substances,
three conservative sub-
stances
DO, CBOO. ammonia, nitrate.
nitrite, total nitrogen, phos-
phate, coHtorms. elgae.
salinity, one metal ion
Processes Simulated
Chemica (.'Biological
1st -order decay ol
NBOD. CBOD. cou
pled DO
1st -order decay o1
NBOD. CBOO, cou
pled DO. benthic
demand (»l. photo-
synthesis (s)
1st -order decay of
NBOD, CBOO. cou
pled DO. benthic
demand It), CBOD
settling (s).
nutnent-elgel cycle
1st -order decay of
CBOO. coupled DO.
benthic demand
Is). CBOD settling
Is), nuthent-algal
cycle
Phystcal
dilution.
sdvectton.
reaeration
dilution
advection.
reaeration
dilution.
edvection.
reaerauon.
heat
balance
dilution.
advection,
reaeration
(st-
2.2.3.2 FMd Data Used to Estimate Model
Coefficient*
Besides sampling for the constituents to be simu-
lated, additional measurements may be necessary to
help quantify the various coefficients and parameters
included in the model equations. Coefficient values
can be obtained in four ways, 1) direct measurement,
2) estimation from field data, 3) literature values, and
4) model calibration. Model calibration is usually
required regardless of the selected approach. How-
ever, coefficients which tend to be site specific or
which can take on a wide range of values should
either be measured directly or estimated from field
samples. This could include the following param-
eters:
• Carbonaceous BOD decay rate
• Carbonaceous BOD settling rate
• Ammonia oxidation rate (nitrogenous BOD decay
rate)
• Sediment oxygen demand.
Carbonaceous BOD decay and settling rates can be
estimated from field data by plotting CBOD meas-
urements versus travel time on semi-log paper. The
decay rates are estimated from the slopes of the lines
(Figure 2-11). Slope calculations should be limited to
reaches where tributaries are negligible. In situations
where CBOD settling is important, a two-stage curve
usually results, with a steep slope on the first part and
a more gradual slope on the second pan (Figure 2-12).
The first part of the curve gives the total removal rate
when both settling and decay are significant (K,).
while the second part generally represents CBOD
decay after most of the settling has taken place (K0)
The settling rate (K.) can then be estimated from the
difference between K, and Ka. Carbonaceous BOD
decay rates can also be measured in the laboratory
using nitrogen inhibited tests or calculated by other
techniques (33), but the above approach is generally
preferred since it reflects the actual conditions
measured in the field.
Figure 2-11. Example computation of total BOD removal
rate. K,, based on BOO measurements
V = 4 mi/day
K,= -Slope XV
= 23(06)
- 1 4/day
. 14
Point Source
2-18
-------
Figure 2-12 Procedure for estimating Ktf and Kr from BOD
measurements
30-
20
a
o
m
1 0
09
08
05'
,.— K, -- 1 25 day
Kd = 02/day
a 05 10 15
Time 01 Travel. Days
20
25
Ammonia oxidation (or nitrogenous BOD decay) rate
coefficients can also be estimated from field data
using the same graphical technique. Total Kjeldahl
nitrogen (TKN) rather than ammonia is generally
plotted, since TKN includes both hydrolyzable organic
nitrogen and ammonia, both of which will ultimately
be oxidized to nitrate. Unfortunately, ammonia con-
centrations are influenced by algal uptake and
respiration in addition to oxidation, so these pro-
cesses may affect the slope of the curve. Similar plots
of nitrate versus travel time can be used to provide a
second estimate of the ammonia oxidation rate, but
nitrate is also influenced by algal uptake. Unlike
CBOO, most water quality models do not include
separate components for settleable and non-settle-
able NBOD If the model does include separate
formulations for NBOD settling and decay, the settling
rate can be estimated using the procedure above.
Sediment oxygen demand (SOD) should be measured
in situ in situations where it is a significant com-
ponent of the oxygen budget. This is most likely to
occur in shallow streams where the organic content
of the sediments is high The significance of SOD can
be evaluated by comparing it to the carbonaceous
BOD and nitrogenous BOD fluxes. For example, SOD
can be neglected if:
and
KSOD
Ksoo
«KdL
«KNiNH3
(2-12)
(2-13)
Measurements should be taken both upstream and
downstream of the waste discharges, since the
background SOD will probably be lower than the SOD
in the area impacted by the discharge
In addition to the above model parameters which are
determined primarily from the results of field samp-
ling surveys, several other rate coefficients can be
measured in the field. For example, stream reaeration
rates can be measured using tracer techniques
(34,35,36,37,38). However, the usual procedure is to
select an appropriate reaeration rate formula e.g.,
(39,40,41) and compute the reaeration rate as a
function of the hydraulic characteristics of the stream.
Most computer model&provide several options for the
reaeration rate equation, si nee many of the equations.
are »!'P;; *.***'? o«ty over certain ranges of depth and
velocity n.q . QUA" -2E, DOSAG3, RECEIV-II, and
WQRRS
As mentioned above, algal photosynthesis and respi-
ration rates can be measured using light-dark bottle
techniques. However, it is usually more convenient to
estimate these rates by model calibration using field
measurements of diurnal variations in dissolved
oxygen and spatial and temporal variations in algal
concentrations.
Table 2-6 summarizes the methods typically used to
determine the values of each model parameter in
Equations (2-5) through (2-11).
2.2.3.3 Sampling Locations
The general model-independent sampling locations
discussed in Section 2.1.4.1 (i.e., upstream bound*
aries, tributaries, point and nonpoint sources, etc.)
are the minimum sampling requirements for all
conventional pollutant studies. In addition, enough
•nations should be sampled to characterize the shape
of the dissolved oxygen profile below each major
waste source in dissolved oxygen studies. A minimum
.of five or six stations are necessary to define the
•hape of a typical DO sag, assuming the location and
extent of the sag curve are known in advance. Since
this is rarely the case, a few more stations (2 or 3)
should generally be included so that at least one
station is near the dissolved oxygen minimum. It is
important to sample this region since this will be the
area where water quality standards are likely to be
violated. Where violations exist, more intensive
•ampling should be conducted in the sag region to
determine the extent of the violations (Figure 2-7). In
general, more sampling effort should be allocated to
those waste discharges which have the most impact
on the stream. These can be determined by comparing
the mass fluxes of CBOD and NBOD with the cor-
responding ambient fluxes in the stream. In many
cases, the DO sags from different discharges will
overlap, reducing the total number of sampling
locations required.
2-19
-------
Table 2-4.
*a-j|. n .1- £__
MetnOQS IOC
Model Parameter
Coefficient Vetoes In Dissolved Oxygen and EutropMcetion Models
Symbol Method of Determination
Dissolved Oxygen Parameters
Reaeration rate coefficients
Oj coneumption per unit NH) oxidation
O) consumption per unit NO) oxidation
Oj production per unit photusyiiuiesis
O) consumption per unit respiration
Sediment oxygen demand
Carbonaceous BOO Parameters
CBOO decay rate
CBOD settling rat*
Ammonia Parameters
Ammonia oxidation rate
Benthic source rate
rVrffffe Parameters
Nitrite oxidation rate
Phospnate Parameters
Benthic source rate
PnyroptonAton Parameters
Growth rate
Respiration rate
Settling rate
Nitrogen fraction of algal biomass
V,
a*, a* o7
Phosphorus fraction of algal biomass a*, a.
Half-saturation constants for nutrients K«. K.
Saturating light intensity or half- I, or K^
saturation constant for light
Compute as • function of depth and velocity using an appropriate for-
mula, or measure in field using tracer techniques.
Constant fixed by biochemical stoichiometry.
Constant fixed by chemical stoichiometry.
Literature veluea. model calibration and measurement by light to dark
bottles and chambers.
Literature values and model calibration.
In »*tu measurement and model calibration.
Plot CBOO measurements on semi-log peper or measure in-laboratory.
Plot CBOO mea«'"»^em« r»- «*o»Mog paper and estimate from steep
part of curve.
Plot TKN measurements and NO3 + NO2 measurements on semi-log
paper.
Model calibration.
Use literature values and calibration, since this rate is much faster than
the ammonia oxidation rate.
Model calibration.
Literature values and model calibration, or measure in field using light-
dark bottle techniques,
Literature values and model calibration, or measure in field using light-
dark bottle techniques.
Literature values and model calibration.
Literature values and model calibration or laboratory determinations
from field samples.
Literature values end model calibration or laboratory determinations
from field samples.
Literature values and model calibration.
Literature values end model calibration.
Note: Literature values for model coefficients ere available in ref. (18. 19. 20)
As discussed in previous sections, simplified
Streeter-Phelps calculations or preliminary model
runs may be useful estimating the location of the DO
sag prior to sampling. The total length of the sag
region can also be estimated by considering the BOD
decay rate along with the travel times and stream
velocities below the discharge (Figure 2-13). since
BOD decay is generally the major process removing
oxygen from the stream. However, this latter infor-
mation does not provide any information on the
location of the DO minimum.
All of the major water quality parameters of interest
(DO, CBOD, TKN, NHj. NO*. PO* temperature. pH.
etc.) should be measured at each station in the
sampling network. Some constituents, however, may
be unimportant at certain locations, i.e., BOD in the
area of recovery (past the sag point). Algal measure-
ments may not be necessary at all stations in areas
where the sampling grid is close, for example, if
intensive sampling is conducted in the critical region
of the DO sag. since algal concentrations should not
change significantly over small distances. Sediment-
related processes such as SOD only need to be
measured at a few locations. These locations do not
have to coincide with the locations of the other water
quality samples. Rather, they should be located in
areas which will characterize the differences in
sediment characteristics throughout the study area.
Typical locations would be downstream of a major
discharge to define the SOD in areas where signif-
icant settling of BOD occurs, and a site away from the
2-20
-------
Figure 2-13. BOD dsciy tim«i for various d»c«y ratas (From (11))
100
SI
0 B
Q C
i!
£ 40
20
influence of the discharge to characterize the back-
ground SOD. Figure 2-14 shows an example of the
sampling locations for a typical dissolved oxygen
analysis.
2.2.3.3.b Sediment Oxygen Demand
Measurement Strategy
The selection of sampling locations for SOD meas-
urement is not usually quantitatively addressed by
water quality modelers. The first step is to determine
which stream reaches should be selected for SOD
measurement, and then outline a strategy for deter-
mining the measurement frequency for the selected
reaches.
In developing a strategy for SOD measurement, it is
logical to assume that those factors important in
establishing model reaches or segments are also
relevant to selecting SOD measurement sites. The
more important of these factors are:
• Geometric—depth, width
• Hydraulic—velocity, slope, flow, bottom roughness
• Water Quality—location of: point sources, non-
point sources runoff, abrupt changes (large gradi-
ents) in DO/BOD concentrations, tributaries, dams
and impoundments.
The most important factor for SOD is likely to be the
location of abrupt changes in DO/BOD concentra-
tions.
In the absence of historical water quality information,
it is best to assume for planning purposes that SOD
should be measured in each model reach. This
recommendation is particularly important in rivers
and streams where significant DO deficits occur. In
these cases, it is important that the modeler obtain
sufficient data to independently account for the
effects of point sources and SOD on water quality.
Lumping instream BOD decay for example, with SOD
is not good modeling practice and should be avoided
in models used for waste load allocation.
As a practical matter, however, this recommendation
is difficult to implement completely due to a number
of financial constraints imposed during modeling
studies. For these situations, it is recommended that
the partially calibrated model be used to determine
which stream reaches (segments} are critical in terms
of DO concentration. SOD measurements can then
be concentrated in these areas.
The modeler should also be aware that the sensitivity
of DO to SOD (and other model mechanisms) can
change significantly when forecasts are made at the
7Q10 flow (or other worst case conditions). Typically,
the model is calibrated with water quality data
collected at flows higher than the 7Q10 flow. It is
possible in this case that the DO sensitivity to SOD is
low, and the modeler might elect to reduce the
number of SOD measurements accordingly. This
action becomes inappropriate if the stream DO
sensitivity to SOD increases at the longer travel times
usually associated with the rQio flow.
227
-------
Figure 2-14 Example tamplmg network for • dusolved oxygen «n»ly»i»
Key
•»• Waste Discharge
O Water Quality Sample
• SOD Measurement
(ai Aerial View Of River
8-
7-
c 6
u
S ^
01
>
"5
»
6
3-
2 -
(b) Measured DO Profile
10
15
20
25
30
l
35
40
River Distance. Miles
Once critical reaches are defined, several considera-
tions that should be addressed by the modeler include:
measurement technique, measurement precision,
and measurement frequency.
Although it is not the purpose of this handbook to
review SOD measurement, it is important to note that
there is a controversy regarding the accuracy of
appropriate laboratory based procedures when com-
pared to preferred in-situ methods (18,42). There is
growing evidence, however, that laboratory based
procedures can be used as a reliable surrogate for
in-situ measurements (18). These data would further
suggest that basin specific correlations of laboratory
SODs with in-situ SODs are credible alternatives to
extensive in-situ measurements. In this case, the
modeler would collect the data to develop the
correlation at one location and then rely upon the less
2-22
-------
expensive laboratory technique for remaining stream
segments This option would be especially advan-
tageous for large basins with similar sediment
characteristics throughout.
There is some data available for SOD measurement
precision for both in-situ and laboratory methods
(42,43,44,45) To a large degree, the precision is a
function of the experience of the field crew or
laboratory analyst. For in-situ work with an experi-
enced crew, a precision defined by the coefficients of
variation of multiple measurements could be as high
as ±40% With additional field experience, the
precision should consistently improve to the ±20%
range. Laboratory precision is usually better than in-
situ precision and as a rule in the range ±10% to
±20%. li is recommended that measurement crews
pay close attention to measurement precision in
model studies used for wasteload allocation. Five
duplicate measurements at one location prior to
extensive SOD work throughout the basin are ad-
visable to define this potentially important factor for
model calibration Reference (16) presents guidance
for both laboratory and in-situ SOD methods.
Since most water quality models require that SOD be
specified as a single value per reach, the input value
must be an average for the entire reach. For large,
slow moving rivers a minimum of 2 to 3 measure-
ments per reach is recommended and should include
both mid-channel and shallower stream bank areas.
One measurement per reach may be appropriate for
small, shallow streams if bottom conditions are
consistent within each cross-section. Visual obser-
vations of the streambed should provide the modeler
with a basis for this judgment. For all streams,
however, duplication of at least 10% of all SOD
measurements is recommended for quality assurance
purposes.
The final point to consider is that SOD may vary with
season. This observation is particularly relevant to
some estuarine and impoundments dominated by
algal activity and/or oxidation of organic and in-
organic nutrients by benthat microorganisms, both of
which may occur seasonally The modeler should
thus be aware of this potential concern and structure
the SOD measurement times accordingly.
2.2.3.4 Sampling Time and Frequency
The general model-independent sampling concerns
discussed in Section 2.1.4.2 are directly applicable in
conventional pollutant studies. Most WLA analyses
use steady-state models, except in some cases where
diurnal variations in oxygen are important or when
long term eutrophication analyses are necessary. The
analyses are typically conducted for a low flow
condition with a high summer temperature since
dissolved oxygen problems are usually most severe
under these conditions. Procedures for selecting the
appropriate flow and temperature conditions are
described in Book VI ("Design Conditions") of the
WLA guidance documents (see Table 1-1) The
sampling program should be conducted during the
time of the year that most closely approaches the
conditions to be used in the analysis. Samples should
be collected during a period when weather, waste
loading, and stream flows are expected to remain
approximately constant If possible, a short-term
intensive survey should be conducted in which
several samples are collected at each station at
different times of the day over a period of 2 or 3 days.
This approach provides enough data to accurately
define the average DO profile, as well as the variabitity
in the profile (Figure 2-15).
Diurnal variations in dissolved oxygen can be i o-
portant in streams when phytoplankton, periphytun.
or aquatic plant densities are high, or in streams
which have large diurnal variations in temperature
(5°C or more). In the first case, the DO variations are
due to photosynthesis and respiration, while in the
second case the variations are due primarily to the
effects of temperature on DO saturation. Photosyn-
thesis and respiration produce maximum DO concen-
trations in late afternoon and minimum concentra-
tions in early morning. Temperature variations result
in essentially the opposite effects, minimum DO
levels in mid afternoon and maximum levels at dawn.
Figure 2-16 shows examples of two rivers in which
these effects cause diurnal variations of about 2 or 3
mg/l. These could be significant if the background
levels were close to the water quality standards.
If diurnal variations are important, a dynamic model
or a quasi steady-state model which simulates these
effects should be used in the WLA analysis Pre-
liminary sampling over a 24-hour period at a few
stations should first be conducted to determine if
diurnal effects are significant The significance of the
variations depends on the context of the problem. For
example, if the daily average DO concentration is
around 5 mg/l or less, then a diurnal variation of less
than 1 mg/l could be very important with respect to
meeting water quality standards, while if the average
DO concentration is around 10 mg/l, then diurnal
variations of 2 or 3 mg/l may not matter. However,
these latter variations would be important if future
projected waste loads were being analyzed since
these loads could lower the ambient DO levels in the
stream to a point where a 2 to 3 mg/l diurnal
fluctuation could violate standards. If preliminary
sampling indicates diurnal variations are important,
then the sampling program should include 24-hour
sampling for dissolved oxygen and temperature at all
of the key stations. As a minimum, these would
include the upstream boundary, all major tributaries,
and a few stations near the low points of the major DO
sags. If there is reason to suspect that the significant
diurnal variations in characteristics of the waste
2-23
-------
Figure 2-16. Results of • short-term intensive survey to establish the dissolved oxygen profile (modified from Clarence J Velz.
Appktd Stettm Stnitttion. copright c 1984 by John Wiley & Sons. Inc.) |47|
' ' I ' I ' I ' I '
III!
I I I I
Stations
Dissolved Oxygen Sample
Observed Mean Value
90S. Confidence Range .
^""-
I _
-^. *
^-&
i i ii i i +i i
i i \k -
20 19 18 17 16 15 14 13 12 T1 10 9
Miles Above Mouth
Figure 2-16. Deity dissolved oxygen variation in two
streams (From (27)).
12
1,1
c
1
610
•g
1 9
eft
o
8
7
Photosynthesis
And Respiration
Effect ^.A
. ^
;J>N
.' \^
"\— .-----'
Temperature
Effect
i i i i
Wyman Creek. CA
August 6. 1 962
Average 10 1 mg/l
tf
\ ,-'--"
x
•*' NI^~*~'"~*-'^_
River Ivel. England
May 31 1959
Average 104 mg/l
L L 1
0600 09OO 1200 N 15OO
1800 2100
Hours
2400 0300 06OO
discharges will occur, than the discharges should be
sampled. These locations satisfy the minimum re-
quirements of defining the boundary and loading
conditions plus a few calibration stations in the
critical portions of the DO sags. However, additional
stations would also be desirable, for example, up-
stream of the tributaries and waste discharges, and at
several locations along the major DO sags. As with
the other data, two sets of sampling data are required,
one for calibration and one for verification. The
diurnal sampling should be conducted at the same
time or as close as possible to the rest of the water
quality sampling.
2.3 Sampling Requirements for Toxic
Pollutants
2.3.1 Introduction
Thousands of toxic pollutants are discharged into
rivers across the United States. The toxicants can
arbitrarily be grouped in many different ways (e.g., by
use, by quantity produced, by volatility, or by mole-
cular structure). For design of stream surveys, the
following categorization is convenient:
Toxicants
/ \
Organic*
Metals
Strong
Adsorption
To Sediments
\
\
Moderate To
No Adsorption
To Sediments
Strong
Adsorption
To Sediments
Moderate To
No Adsorption
To Sediments
Sampling requirements are generally more intensive
when toxicants adsorb to suspended and bottom
sediments because data are needed to quantify such
interactions. Since not all toxicants adsorb to sedi:
2-24
-------
ments, however, the assumption should not auto-
matically be made that sediment-toxicant interactions
must be quantified in stream surveys. Such an
assumption can lead to needless expenditures. For
example, many metals can be transported largely as
dissolved species if river water pH is low (e.g., 6.0 to
6.5), and if suspended solids concentrations are also
low (e.g., 0 to 25 mg/l). These conditions pertain in
many rivers in the Northeast and Southeast during
moderate to low flow periods.
Also, many of the organic toxicants are transported
predominantly in dissolved form at low suspended
solids concentrations. Adsorption of organics can be
evalu"""J ".s'Tt •*••» Blowing expression:
C/Cr =
1
1 +KP-S-KT*
(2-14)
where C/Cr = fraction of organic toxicant in dissolved
form, dimensionless
KP= partition coefficient, 1/kg = 0.6.fc.Ko«
[See (27)] for details and values of KM
S = suspended solids concentration, mg/l
fc = fraction by weight of organic carbon on
suspended sediments (typically 0.01 -
0.10)
kow = octanol-water partition coefficient
For conditions when C/CT approaches unity (e.g.,
20.9), adsorption is unimportant, and pollutant-sedi-
ment interactions can be neglected. For example,
suppose fc = 0.03, Kow = 50, and S = 25 mg/l. Then
C/CT =
1
0.6 • 0.03 • 50 • 25 -10'
= 0.9999
and adsorption is negligible.
Equation (2-14) has been used with limited success
for metals as well as organics. However, the partition
coefficient Kp is usually taken as site-specific for
metals and local data for pH, suspended sediments
and other WQ paramters are needed to reliably use
this approach.
The fate of organic toxicants can be controlled by
processes in addition to adsorption such as photol-
ysis, biodegradation, hydrolysis and volatilization.
Surprisingly, however, the fate of many organic
toxicants are often dominated by a single process. For
example, the following organics are commonly dis-
charged into rivers, and are also commonly found at
Superfund sites:
Trichloroethylene (TCE)
Toluene
Benzene
PCBs
Chloroform
Tetrachloroethylene.
Volatilization probably controls the fate of five of the
toxicants, while adsorption is most important for the
remaining one(PCBs). The volatilization rate constant
as shown in (27) can be found from:
k, =
(2-15)
where k, = volatilization rate
ks = reaeration rate of dissolved oxygen
MW = molecular weight of the organic compound
that is volatilizing
Consider, for example, TCE (MW = 131| -n -> -••o:—
where the atmospheric reaeration rate is
The volatilization rate is:
:.o day
k, =
2 = 1.4/day
Other processes (hydrolysis, photolysis, biodegrada-
tion) are insignificant compared to the volatilization
rate. Further TCE has a low Kow so that adsorption can
also be neglected. Consequently, the atmospheric
reaeration rate is the major process that must be
quantified to predict the fate of TCE in streams. This
example illustrates that simple approaches can be
used to allocate waste loads for some toxicants, and
that instream data requirements may not be prohib-
itively expensive.
When multiple processes mutually influence the fate
of toxicants, stream surveys cannot always be easily
designed to segregate out the significance of each
process. However, the composite rate constant can
be found in the same manner as for the BOD decay
rate by plotting toxicant concentration versus dis-
tance (See Section 2.2). Transformation rates for
toxicants are usually determined from theoreticaf
relationships, or in the laboratory, and the sum of the
rate constants can then be compared to the instream
composite rate. Only under special circumstances
can the individual transformation processes be found
from a stream survey (see the example problem in
Section 3.2, for example) or by resorting to more
elaborate approaches (Hern at at.. 1983).
2.3.2 Model Dtte Requirement*
This section summarizes data requirements for
methods that can be used to to determine the amount
of toxicants that can be assimilated. The methods
range from simple to complex:
• screening techniques for organic toxicants and
metals
• the MICHRIV model, a steady-state computer
model for metals and organic toxicants
• the TOXIWASP model, a dynamic computer model
for toxic organics
2-25
-------
Tables 2-7 through 2-10 summarize the data re-
quirements for each approach. The first two tables
show data requirements for screening techniques.
The requirements in those two tables have been
further subdivided as shown below:
Metals
Organics
• dilution only is
considered
• dilution and
adsorption are
considered
• dilution only is
considered
• dilution and
adsorption are
considered
dilution, adsorption, •
and interactions with
streambed are
considered
dilution, adsorption,
interactions with
streambed, and
speciation are
considered
dilution, adsorption,
and decay are
considered {often
volatilization, dilution,
and adsorption are the
most important)
Depending on the specific situation and resources
available, the analyst can select an appropriate level
of complexity, and collect data accordingly. The
screening methods are most applicable when one or
two sources of toxicants are present, when hydraulics
are simple, and when fate processes are easily
quantified.
The data requirements of MICHRIV (Table 2-9) are
similar to the most complex level of screening
analysis. However because MICHRIV is a computer
model, multiple waste sources and spatially variable
parameters are more easily accommodated. The data
required for metals and organics are indicated
separately, and those associated with adsorption are
shown with an asterisk (*).
The data requirements for TOXI..*.;>i- ;idot« i-io)
are presented in a very summarized format. Data
requirements are greater than for the previous
approaches. The analyst should consult the TOXI-
WASP user manual (4) for specific details.
TOXIWASP is designed explicitly for organic toxicants
(and not metals) and requires more technical exper-
Table 2-7. Summary of Data I
> for Serening Approach for Metate In Rivers (27)
Data
Calculation
Methodology Where
Data are Used*
Remarks
Hydraulic Data
1. Rivers:
• River flow rate, Q D, R, S. L
• Cross-sectional area, A D, R. S
• Water depth, h D, R, S. L
• Reach lengths, x R, S
• Stream velocity, U R, S
2, Lakes:
• Hydraulic residence time, T L
• Mean depth, H L
Source data
1. Background
• Metal concentrations, CT D, R, S, L
• Boundary flow rates. 0^ D, R, S, L
• Boundary suspended solids, D, R. S, L
• Silt, clay friction of sus- L
pended solids
• Locations D, R, S, L
2. Point Sources
• Locations D, R, S, L
• Flow rate, Q,, D, R, S. L
• Metal concentration, CT- D, R, S. L
• Suspended solids, Sw D, R, S, L
An accurate estimation of flow rate is very important because of
dilution considerations. Measure or obtain from USGS gage.
The average water depth is cross-sectional area divided by surface
width.
The required velocity is distance divided by travel time. It can be
approximated by Q/A only when A is representative of the reach
being studied.
Hydraulic residence times of lakes can vary seasonally as the flow
rates through the takes change.
Lake residence times and depths are used to predict settling of ab-
sorbed metals in lakes.
Background concentrations should generally not be. set to zero
without justification.
One important reason for determining suspended solids concentra-
tions is to determine the disserved concentration, C, of metals,
based on CT, S, and Kp. However, If C is known along with CT
•nd S, this information can be used to find Kp.
2-26
-------
Table 2-7. (Continued)
Data
Calculation
Methodology Where
Data are Used*
Remarks
Bed Data
• Depth of contamination
• Porosity of sediments, n
• Density of solids in sediments
(e.g., 2.7 for sand). n,
• Metal concentration in bed
during prolonged scour pe-
riod, CTJ
Derived Parameters
• Partition coefficient, Kp
• Settling velocity, w,
• Resuspension velocity, wn
Equilibrium Modeling
Water quality characterisation
of river:
• PH
• Suspended solids
• Conductivity
• Temperature
• Hardness
• Total organic carbon
• Other major cations and anions
For the screening analysis, the depth of contamination is most use-
ful during a period of prolonged scour when metal is being input
into the water column from the bed.
All
S. L
R
The partition coefficient is a very important parameter. Site-specific
determination is preferable.
This parameter is derived based on suspended solids vs. distance
profile.
This parameter is derived baaed on suspended solids vs. distance
profile.
Equilibrium modeling is required only if predominant metal spe-
cie* end estimated solubility controls are needed.
Water quality criteria for many metals are keyed to hardness, and
allowable concentrations increase with increasing hardness.
*D - dilution (includes total dissolved and adsorbed phase concentration predictions)
R - dilution and resuspension
S - dilution and settling
L - lake
E - equilibrium modeling
Table 2-9.
Summery of Data fleujuhemenU for Ocreent
Methodology Where
i for Toade Orfentee In Mvers (27)
Data
Data are Used*
Remarks
River Hydraulic Data
• Flow rate. Q D, DA. DAK
• Cross-sectional area, A D, DA, DAK
• Water depth, h DAK
• Reach lengths, x DAK
• Stream velocity, U DAK
Source Data
1. Background
• Toxicant concentrations D, DA. DAK
• Boundary flow rates D, DA, DAK
• Boundary suspended solids DA, DAK
2. Point Sources
• locations D, DA, DAK
• Flow rates, Qw D, DA, DAK
• Total toxicant concentration,
CT D, DA, DAK
• Suspended solids, Sw DA, DAK
An accurate estimate of flow rate la very imporent beceuee of dilu-
tion, which for many organic* is the most important process that
influences their fate. Measure or obtain from USGS gage.
ch es volatilization and
Water depth on influence rate proceeei
photolysis.
U • Q/A should be used only where A is representative of the
reach being analyzed. Otherwise dye tracers, measured from
centroid to centroid of the dispersing dye is a better method of
finding velocity (indirectly es distance divided by travel time).
Concentrations of organic toxicants may be negligible in areas not
influenced by man.
Suspended solids era used to help determine the dissolved end
adsorbed phaee concentrations.
2-27
-------
Table 2-t. (Continued!
Data
Methodology Where
Data are Used*
Remarks
Partition Coefficient and Rate
Constant Data
Solid-liquid partition
coefficients. K
Acid-base speciation
Volatilization rate
DA. DAK
DA. DAK
DAK
Stream surveys can not always be easily designed to calculate rate
constants or partition coefficients for toxic organic*. A step-by-step
procedure for calculating each rate constant and partition coeffi-
cient discussed here can be found in Mills et al. (in press). Input
data needed to calculate rate constants and partition coefficients
are identified here, and ranges of values lor the data are found in
(27. 49)
Data required:
• KM, octanol-water partition coefficient (use literature, e.g., Leo er
al.. 1971)
• Xfc. organic carbon fraction of sand in suspension (typically 0.00-
0.05)
• XJK, organic carbon fraction of silt-day in suspension (typically
0.03-0.10).
Data required:
• pH of water
• K, or K«, the association constant for the organic acid or base
(from literature, e.g., (51))
Data required:
• Henry's Law Constant (from) literature, e.g., (52)
• Stream depth
• Reaeration rate for dissolved oxygen
• Wind speed (only for toxicants with small Henry's Constant, e.g.,
Biodegradation rate
Hydrolysis rate
Photolysis rate
DAK
DAK
DAK
Typically, only an approximation of biodegradation rate is obtain-
able due to factors such as adaptability to stream environment
Data required:
• pH of river
• Acid or base catalyzed hydrolysis rate constants (from literature.
e.g.. (53)1
• Neutral hydrolysis rate (from literature e.g., (53))
Data required:
• Solar radiation
• Water depth
• Concentrations of light-attenuating substances (chlorophyll a,
DOC, SS)
*D - dilution only (total organic in water column, sum of dissolved and adsorbed phases)
DA - dilution plus adsorption (to predict dissolved and adsorbed phases)
DAK • adsorption and rate processes both considered.
Table 2-9 Mfchriv Model Data Requirement* (5)
Variable Pollutant Category'
Remarks/Qualifications
Channel Data
• River flow, Q
• Velocity. U
Cross section area, A
Reach length, x
Depth of water, h
M, 0 Measure or obtain from USGS gage.
M. O Measure directly with time-of-passage dye study, (Ref. (54)) or
compute from area and flow: U - Q/A.
M, O Compute from measured width and depth, or compute from
velocity and flow.
M. O Reaches determined by significant morphometric changes, trib-
utaries, or point sources; measure from charts, confirm in
field.
M, O Measure directly or compute from cross section area and mea-
sured width.
2-28
-------
Table 2-9
(Continued)
Variable
Pollutant Category1
Remarks/Qualifications
Loading Data
1. Upstream "Boundary' Concentration
• Toxicant, Cu
• Suspended solids. Sw
2. Point Sources
• Flow, Qw
• Concentration toxicant, Cw
• Concentration-suspended solids
Bed and Paniculate Data
• Thickness of Active Sediment. H2
Solids concentration in bed, m2
Porosity, n
Solids type
Size distribution
Settling Velocity, w,
• Resuspension velocity. wn
• Partition coefficient, Kp
• Sediment diffusion, K,
Hate Constants and Related Data
M. O
M*. 0*
M. 0
M, 0
M*, O*
M*. O*
M«. O"
M*, O*
M', O*
M*, O*
M, O
Direct measurement of loading data is preferable for WLA model-
ing.
Estimate from core samples, measuring vertical distribution of
contaminants; or use typical published values. This parameter
has no effect on steady state results unless significant decay
O- OOifc ". .'..*• ut»u.
Measure or estimate: m2 «= (particle density) (1-n)
Estimate from particle-size distribution and stream turbulence
coupled with published data or Stokes formula. Measure with
sedeiment traps or in lab. Adjust by calibration.
Calibrate to suspended solids data; estimate from theory.
Calibrate from dissolved and paniculate data. Otherwise, use
literature values (5)
Use literature values (5)
• Volatilization coefficient, fc»
Reaeration coefficient, kj
Solubility, S
Vapor pressure, P
• Photolysis rate, kp
Chlorophyll a
Diss. organic carbon
Suspended solids
Solar/UV radiation
Near surface rate
• Biolysis rale, k^
Cell count
Chlorophyll a
Hydrolysis rate pH
Ancillary data: temperature
0
O
O
O
0
0
0
O
0
0
0
0
0
0
0
Calculate from theory.
Use published data (49)
Use published data (49)
Calculate from theory (27) or by Actinometer.
Meausred by Actinometer at water surface, or in 'laboratory (49)
Laboratory experiment at different pH values or from pub-
lished data (53)
'M = Metals; 0 = Orgenics. The asterisk (•) indicates the data are required only if adsorption to sediments is important.
TaM* 2-10. Summary of Input Dcta Required for TOXIWASP
Category
Data
Mass Exchange
Volumes
Flows
• exchange coefficients
• interfacial cross-sectional area
• river segment lengths
• volumes of segments
• flow between segments
• flow routing information
• piecewise linear approximation of time variable flows
2-29
-------
Table 2-10. (Continued)
Category
Data
Boundary Condition
Forcing Function*
Parameters
Constants
Miscellaneous Time Function*
Initial Conditions
Stability and Accuracy
Criteria
boundary'
nitrations
piecewise linear approximation of time variable flows
loading rates
piscewiee linear approximation of time variable function
temeprature v» time function
deptn of compartments
water velocity
wind speed
bacterial population
biomaes
reeeration rate
molar concentration of environmental oxidants
organic carbon content of fine sediments
percent' . • . • -• irtments
PH
settling iwte» ot t,.ie
66 constants required
• initial concentrations of concentrations, water temperature, etc.
tise and resources than the approaches discussed
previously. TOXIWASP may be an appropriate model
to use for WLA when:
e waste loadings (and other boundary conditions)
are highly variable over time
• the flow field is highly dynamic (e.g., during a
storm)
e other significant parameters (e.g., water tempera-
ture) are time variable
• detailed sediment—water column interactions are
required (for toxicants that adsorb strongly).
2.3.3 Stmpling Quid»Hn»t
Profiles of toxicants in rivers, under most hydrologic
conditions, often approach gradually curved lines, as
shown by Figure 2-17. Exceptions occur in the vicinity
of point sources where an abrupt increase or decrease
in toxicant concentrations may occur. Because toxi-
cant profiles do not exhibit "sag points" as do
dissolved oxygen profiles, sampling stations can
usually be more evenly distributed downstream of the
source.
Distance between sampling locations can be esti-
mated based on the relative change in toxicant
concentration desired between the stations. Table
2-11 summarizes travel times required for various
C/Co ratios (ratios of downstream to upstream
concentrations) and decay rates. Travel times greater
than approximately 2 days between locations are not
recommended.
The travel times shown in Table 2-11 can be found by
solving the following equation:
t= In (Co/C)
k
(2-16)
where t = travel time between locations where the
concentration changes from Co to C
k = first order decay rate, 1 /day
By selecting a C/Co ratio and an approximate decay
rate, the analyst can determine the travel time interval
between sampling locations. Given the travel time
the equivalent distance between two sampling sta-
tions (x) is approximately
x=Ut (2-17)
where U = stream velocity
t = travel time
For toxicants that are expected to act nearly con-
servatively, the distance increment is approximately
controlled by the longest travel time between samp-
ling points the water quality specialist is willing to
tolerate, but generally this should be less than two
days. For toxicants that decay, rapidly, travel time
between sampling points are on the order of 0.2 to 0.5
days for C/Co ratios of 0.5 to 0.7. This will generally
correspond to intervals of 5 to 9 km (3 to 5 miles).
2-30
-------
Figure 2-17
Typical concentration profiles of toxicants in
rivers
Figure 2-18 Sampling locations for toxicants during low
flow and high flow period
Point Source
Flow Direction
Point
Source
Point
Source
Tributary
c
V
u
c
o
CJ
Distance, km
la) Toxicant Profile That Reflects
Settling Or Decay
I
I
1
£ Recommended
Sampling Location
ft Recommended when
Solids Settle
• Flow
10
35 40 45
Concentration
Point Source
1
^^fc- Flow Direction
Distance km
(bl Toxicant Profile That Reflects
Scouring Of Contaminated Sediments
uistance. Km
la) Typical Low Flow Profile
c
c
o
Concentrat
Table 2-11. Travel TJm«» for Various C/Co Ratios
Pomt Point
ource Source Tributary
! 1 1
9 Recommended
Sampling Location
^ e *|
^^^ Flow
(UtM
Travel Time (Day«)
C/Co
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.05
2.1
-
-
-
-
-
-
-
-
0.1
v.o
2.2
-
-
-
-
-
-
-
0.2
0.5
1.1
1.8
2.6
-
-
-
-
-
0.3
0.4
0.7
1.2
1.7
2.3
_
-
-
-
k, (I/day)
0.4 0.5
0.3
0.6
0.9
1.3
1.7
2.0
-
-
-
0.2
0.4
0.7
1.0
1.4
1.6
-
-
-
0.6
0.2
0.4
0.6
0.9
1.2
1.5
2.0
-
-
0.8
0.1
0.3
0.4
0.6
0.9
1.1
1.5
2.0
-
1.0
0.1
0.2
0.4
0.5
0.7
0.9
1.2
1.6
2.3
1.5
0.07
0.1
0.2
0.3
0.6
0.6
0.8
1.1
1.5
2.0
0.05
0.1
0.2
0.3
0.3
0.5
0.6
0.8
1.1
10 15 20 25 30 35 40
Distance, km
(b) Typical High Flow Profile
45
Figure 2-18 shows profiles of toxicant concentrations
in a river during two seasons of the year (a summer
low flow period and a spring high flow period). During
high flow it is assumed that the toxicant is scoured
back into the water column, to produce increasing
concentrations with distance.
Also shown in Figure 2-18 are suggested sampling
locations for a minimal sampling program. Sampling
directly below point sources is not recommended.
Instead, samples should be collected upstream and in
the point source (the latter should include a meas-
urement of the discharge flow rate). At least one
sample per reach (between sources) is recommended;
two or more samples in one reach help add a degree
of certainty to the observed profile.
In short reaches on wide rivers, sampling should be
conducted at several points laterally across a river if
evidence (e.g., from a dye study) indicates that
complete lateral mixing has not been attained. See
Section 3.2 for general guidance on this subject. In
2-31
-------
one-dimensional models the predicted concentrations
are cross-sectional averages, which can be signifi-
cantly different from a single measured concentration
if large lateral concentration gradients exist across
the river.
When adsorption to suspended solids is important
(e.g., C/Cr < 0.9), suspended solids profiles versus
distance should be found to determine the signifi-
cance of settling or scouring of sediments and
adsorbed toxicants. Figure 2-19 shows example
profiles of suspended solids during a tow flow and
high flow period, and can be used to help make this
determination.
i r.c _ mling rate (w,) and resuspension rate {Wn) of
.;,. o lids are required for use in WLA modeling, and
can be generated from the profiles. For settling of
solids, w. can be computed as:
Figure 2-19.
w, =
-Hu
X
In
SS(x)
SS(0)
(2-18)
where w. = settling rate
h= stream depth
U= stream velocity
SS(0) = suspended solids concentrationat a ref-
erence location x = 0
SS(x) = suspended solids concentration at loca-
tion x
Based on Figure 2-19(a), the approximate settling
rate between km 50 and km 35 is:
w. =
0.5x0
30 x 10
.2 , /10\
— In —
33 \4/
x 86400 = 0.26 m/day
where the depth (0.5 m) and velocity (0.2 m/sec) are
taken from (5).
The resuspension velocity, w^, can be estimated as:
UH ASS
Ax-108
(2-19)
The term ASS/ Ax is the change in suspended solids
concentration over distance Ax and mi is the solids
density in the bed. From Figure 2-19(b), the resus-
pension velocity is computed as:
w,,=
1 -2-(25- 12)
0.2 -80 • 10s- 10
, 86400= 1.4-10"4 m/day
The data for velocity (1 m/sec), depth (2 m) and
bedded sediment concentration (0.2 kg/I) are also
from (5).
Typical suspended solids concentrations
during (•) low (low and (b) high (low periods
Key
Trend Line
Mean & Standard
Deviation of Data
£,
E
vT
•o
"o
in
1
c
£
i/>
3
16-
12-
8-
4-
n-
Flow Direction I
1
-* ! .
T 1 J.
_ I ___.,_. So
i ^ . . T j «P ^
'T' T---.V. ._ j i
•* T 'TR J !
1
05 15 25 35 45 55 65 75
River Kilometers
(a) Suspended Solids, mg I. During Low Flow Period
Key
Trend Line
I Mean & Standard
Deviation of Data
T 30
o
E24-|
tn
I18^
"g 12-
I 6-
Flow Direction
10
20
60
70 BO
30 40 50
River Kilometers
(b) Suspended Solids, mg. I During High Flow Period
Equations 2-18 and 2-19 should be used cautiously
since they give only the net settling rate and the net
resuspension rate.
Note that the values for w. and wra are specific to the
flow regime at the time of sampling. Thus, these
values can change between the calibration period,
the verification period, and the wasteload allocation
period.
Table 2-12 summarizes the topics presented in this
section. Sampling requirements can be intensive if
multiple fate processes, multiple point sources,
sediment interactions, and time variability need to be
evaluated. However, if steady-state conditions pre-
vail, one well quantifiable process controls the fate of
the toxicant being allocated, hydraulics are at steady
state, and there is a single waste discharge, then
stream survey requirements will be minimal.
232
-------
Table 2-12. Summary of Sampling Guidelines for Toxicant*
Seasonal Considerations
• Summer low flow condition* can result in high level* of toxicants in river*. Even if toxicant* tend to be highly adsorptive.
downstream transport in solution can still be significant due to low suspended solids concentration. This period is often used
for WLA analyses.
• High flow condition* can scour toxicants from streambeds and elevate total instream concentrations. This occurs only for
highly adsorptive toxicants, and elevated loadings can be offset by dilution from high flow rates during periods of scour.
Waste Loading Considerations
• Point Sources: Any diurnal variations in loadings should be established. If loading* vary significantly TOXIWASP can be used
for organic*. Otherwise multiple simulations with steady state models should be considered.
• Tributaries: Sampling within tributaries is required if sources that are subject to allocation are located there. Otherwise, sam-
pling at the mouth of the tributaries to establish boundary conditions is acceptable.
• UpstcMm Boundary and Or/Mr Background Locations: Background concentrations and flow rates are required at upstream
boundaries on rivers and tributaries. Concentrations of background organic toxicants are ofen negligible. However, background
metal concentrations can be significant.
Sampling Location Considerations
• Sample upstream of point sources.
• Typically sample at 1-2 locations per reach, with the stations located below the zone of complete lateral mixing.
• If samples are taken ai locations where complete mixing of upstream point sources has not occurred, several samples across
the width of the stream should be taken and averaged before comparing to one-dimensional model predictions.
• The maximum recommended travel time between sampling locations should be leas than two days for toxicants that appear to
be conservative.
• For toxicants with high decay rates, a spacing of 3-5 miles (5-8 km) between sampling stations is appropriate. This will vary
depending upon (ravel time.
• For pollutants that adsorb significantly, depth-averaged suspended solids profiles should be determined for each season ana-
lyzed for the waste load allocation. Fairly close spacing of sampling stations is recommended (5-8 km).
Temporal Considerations
• Sampling at 3-4 hour intervals over a day may be required when:
- source loadings are highly time variable.
- the fates of the toxicants of concern are controlled by temperature dependent processes such as volatilization and water
temperature fluctuate considerbly (e.g., 5-10*0 or more over a 24 hour period).
- photolysis (which depends on solar radiation) is an important process for the toxicant.
Rate Constant and Equilibrium Process Considerations
• Generally rate constants for toxicants are determined based on theoretical considerations. When multiple processes ere opera-
tive, the processes may be difficult to segregate based on instream measurements alone. However, some supporting instream
data are required (see Tables 2-2 and 2-3) nevertheless.
• The total rate constant for toxicant* can be determined by plotting concentration v*. distance. Settling and scour are included
in this approach.
• The total instream rate constant can be compared with the theoretical rate constants for validation.
• For metals, if speciation is to be predicted, major cations and anions, plus pH, ic required. Ref (55) provides appropriate data
for'major rivers through the United Stales. MEXAMS is a computer model which will make these predictions for arsenic, cad-
mium, copper, lead, nickel, silver and zinc. On a screening level, the methods of (27) can be used for the same seven metals
plus mercury and chromium.
2-33
-------
Coefficients
where 02 = dissolved oxygen concentration
0Ui - dissolved oxygen saturation concen-
tration
L - carbonaceous BOD concentration
NH3 = ammonia concentration
NOj = nitrite concentration
NOj = nitrate concentration
A = algal concentration
K. = reaeration rate
Ka = carbonaceous BOD decay rate
K, = carbonaceous BOD settling rate
K, - total removal rate for carbonaceous
BOD
KSOD - sediment oxygen derruiiu
h = water depth
KNi = ammonia oxidation rate
KNJ = nitrite oxidation rate
benthic source rate for ammonia
(j - algal growth rate
r - algal respiration rate
V, - algal settling velocity
a\ - oxygen consumed per unit of ammonia
oxidized
ai = oxygen consumed per unit of nitrite
oxidized
03 = oxygen produced per unit of photo-
synthesis
at - oxygen consumed per unit of respi-
ration
a-, - ammonia produced per unit of algal
respiration
ae = ammonia uptake per unit of algal
growth
a? = nitrate uptake per unit of algal growth
2 34
-------
Chapter 3
Whole Effluent Approach
3.1 Overview
Guidelines have been established for the use of the
whole effluent approach to waste load allocation (6).
These tjuiaennes are bas>?d on information concern-
ing an c"!ijc.v'» ci.-imit^i constituents and whether
single or multiple effluents discharge into the stream.
the whole effluent approach should be used if:
• effluent constituents are not well characterized
• known effluent constituents have not been evalu-
ated as to potential effects on stream biota
• the mixture of effluent constituents may produce a
complex (additive, antagonistic or synergistic)
instream effect
• multiple dischargers may create complicated ef-
fluent mixtures instream
In the case of multiple dischargers, or nonpoint source
pollution, it may not be possible to characterize the
chemical constituents of each individual effluent.
From the above considerations, it might be concluded
that only whole effluent testing should be conducted
in this situation. However, if any of the effluents has
been characterized as containing chemicals with bio-
accumulative, carcinogenic, teratogenic, or muta-
genic potential, the USEPA suggests the use of an
integrated approach where both chemical-specific
analyses and whole effluent analyses are conducted.
The chemical-specific approach is discussed in
Section 2. In this section, the data requirements for
whole effluent toxicity testing are addressed.
The whole effluent approach to waste load allocation
involves testing of effluent toxicity as well as ambient
toxicity testing in streams. Two tiers of effluent
toxicity testing are defined. Tier 1 consists of screen-
ing methods and may be used to identify potential
water quality impact situations. Where this potential
impact is minimal, further evaluation is not required
and the process can proceed directly to wasteload
allocations. Tier 2 effluent testing is used to develop
the data necessary to quantify potential effluent
impacts. In some cases, effluent toxicity assessment
may bypass the screening level and proceed directly
to the Tier 2 analysis. This determination is made as
the first step of the screening process. The overall
process of effluent toxicity testing is presented in
Figure 3-1.
Ambient toxicity testing is conducted to identify areas
in the receiving waters where ambient toxicity exists.
These procedures consist of exposing test organisms
to receiving water samples and can be used to
determine whether or not the effluent has a meas-
urable toxicity after mixing and undergoing instream
alteration/decay processes. An overview of ambient
toxicity testing procedures is presented in Figure 3-2.
3.2 Tier 1 Effluent Testing—Screening
The first step of the screening process consists of
determining the amount by which an individual
effluent may be diluted by stream flow and mixing.
Effluent dilution ratios (R) are calculated on the basis
of average effluent flow and the critical low-flow
conditions in the following manner:
13-1]
where Q« =Stream critical low-flow defined by the
state
QE = Average effluent flow.
The determination of effluent dilution under condi-
tions of complete mixing requires information on the
average effluent flow and the critical low-flow of the
stream. Effluent flows can be obtained from plant
operating records or NPOES permits. Stream critical
low-flow is state-specified and may be based on a
variety of water quality parameters. The critical low-
flow typically selected is the rQio Stream flow data
vailable from the USGS's WATSTOR data base.
The instream distance required for complete mixing
of the effluent to be achieved must also be deter-
mined This information is used to determine both the
need for toxicity testing and the type of testing to be
conducted as part of the Tier 1 screening procedure.
Waste water discharged into rivers does not instan-
taneously mix across the entire channel. Although
vertical mixing occurs quickly, considerable distance
is sometimes necessary for complete transverse
mixing to occur (Figure 2-2). The following methods
3-1
-------
Figure 3-1 Overview o< effluent toncity testing procedures
|
Effluent Toxicily Screening I
Effluent Dilution &
Mixing Characteristics
earn &
Effluent Flows
Yes ^Dilution • 10.000 1
Dilution -.100:1
No ^^ Yes
: Dilution >1000 1 >
Chronic Testing
; Mortality 2 50% >
^
Yes
Effluent Toxicitv • Definitive Data Generation
Chronic Testing
LC50 Level of
IWC Uncertainty
I No 'LC50 > Level of
^~*^>v. IWC Uncertainty
Waste Load Allocation
Exposure Assessment
3-2
-------
Figure 3-2.
Overview of ambient toxicitv testing proce-
dure*.
Dilution
Characteristics
CM Each Effluent
(Dye Studies)
Short-term
Chronic Toxicity
Testing
Waste Load
Allocation Process
I Permit Process |
I I
fW"J
can be used to estimate mixing distances. The
estimated mixing distances obtained using these
methods can be used to determine the appropriate
toxicity tests.
Distances below an outfall where complete vertical
and lateral mixing are achieved can be estimated by
the following expressions (6,56):
x, =
and
Xi =
[ 0.4 h2/c,, for a bottom discharge
; 0.1 hVc,, for a mid-depth discharge
10.4 wa/ct, for a side discharge
[0.1 w2/c, for a centerline discharge
(3-2)
(3-3)
where x, = distance required for vertical mixing to
be approximately 95 percent complete.
feet
Xi = distance required for transverse mixing
to be approximately 95 percent com-
plete, feet
h = average river depth, feet
W = river width, feet
f,,fy = mixing coefficients in vertical and trans-
verse directions, respectively, ftVsec
U = stream velocity
The mixing coefficients t, and e, are typically
expressed as:
and
f, = ohU.
(3-4)
(3-5)
where U. = (ghS)12 = friction velocity, ft/sec
S = slope, ft/ft
ft = 0.05-0.07, typically
o= 0.3-1.00, typically.
To help estimate x, and xt. Equations (3-2) and (3-3)
are plotted in Figure 3-3. The distance to vertical
mixing (Figure 3-3a) is plotted as a function of river
depth for two different riverbed slopes. For river
depths of 3 to 10 feet, x, is typically between 10 to 300
feet. Even for very deep rivers, vertical mixing is
typically accomplished within 500 feet. For most
rivers, then, vertical mixing is completed quite rapidly.
For transverse mixing (Figure 3-3b), x, is plotted
against W2/h18, which indicates that river width and
depth are both required to estimate x,. Typical W'/h1 *
combinations are between 500 to 5000, so that Xi can
vary from several hundred feet to many miles. For a
river 100 feet wide and 3 feet deep, for example, the
mixing distance is likely to be about 2 miles for a side
discharge.
For particularly wide rivers. Equation 3-3 is likely to
overestimate XL because other factors which en-
hance mixing are not considered in the equation. For
wide rivers then, dye studies (tee Section 3.4) are
more likely needed to characterize the multi-dimen-
•ional aspects of mixing.
For the purposes of determining the appropriate
screening test methods, four categories are defined
based on the effluent dilution ratios (6):
Category
1
2
3
4
Dilution Ratio
R>10000
1000
-------
Each of these categories is evaluated differently
within the whole effluent toxicity procedure (see
Figure 3-1) If the effluent is diluted by a factor of
greater than 10,000 to 1 and there is a reasonably
rapid mix of the effluent outside of the zone of initial
dilution in the receiving water, then no further
evaluation is necessary. The effluent is assigned a
low priority in the assessment of any potential toxic
impacts on stream biota. If an effluent dilution is less
than 100 to 1, then further screening is not con-
ducted; definitive data generation, Tier 2 testing, is
begun instead. Thus, the toxicity tests are conducted
as part of the screening process only for dilution
categories 2 and 3
Figure 3-3 Z>* . ...... „».-.,.. r—nt source discharges re-
quired for 96 percent verticil and transverse
mixing.
10- -
• s = oooot
, Bottom Discharges 5 = 0001
„.—-^trrrr —s = 00001
\ Mid-depth Discharges
/ S =0001
^J— •»»—'~ "'
S - Bed Slope
20 30 40
River Depth Feel
la) Vertical Mixing Distance
10' -3
10
100 1000
(Width)V(Deplh)'s, (Feet)05
(b) Transverse Mixing Distance
10.000
The decision concerning the type of toxicity testing to
conduct as part of the screening procedure is based
on the level of dilution achieved(refer to(6), page 18).
If dilution is between 1,000 to 1 and 10.000 to 1
(dilution category 2) or a poorly mixed effluent plume
is of concern, then acute toxicity tests should be
conducted. If dilution is between 1,000 to 1 and 100
to 1 (dilution category 3), chronic toxicity tests are
appropriate.
When -either chronic or acute toxicity testing is
performed, effluent samples must be collected. The
selection of sample type (grab or composite) and time
of collection should be based on information concern-
ing variability of effluent characteristics. Guidance
for effluent sampling as well as toxicity testing
methods are provided elsewhere (6,57,58).
The evaluation criterion for the results derived by
screening toxicity tests is based on the level of
observed mortality. If mortality exceeds 50 percent in
any sample, the potential for toxicity is assumed and
Tier 2 toxicity testing is required. If less than 50
percent mortality is observed for all samples, the
discharge should be given a low priority for further
analysis.
3.3 Tier 2 Tasting—Definitive Data
Generation
Once screening has indicated the potential for toxic
impact, further testing is conducted to determine
whether or not the discharge causes unacceptable
impact. Initially "baseline" acute toxicity testing is
conducted using whole effluent and two species of
test organisms. Then a simple relationship can be
applied to determine whether to require more data
(i.e., additional chronic and/or acute testing (see
Figure 3-1) or whether to stop testing and begin the
process of establishing permit conditions. The evalu-
ation criterion for the results of the Tier 2 toxicity tests
is given by the following formula:
LC50
IWC
> Level of Uncertainty
where LC50 = concentration of effluent producing 50
percent mortality in toxicity tests
IWC = Instream Waste Concentration
The level of uncertainty is determined by a number of
factors, e.g., effluent variability, species sensitivity
variability and the type of toxicity test conducted. All
of these factors are defined in (6). If it is determined
that additional data are required, further testing,
including acute or short-term chronic testing, may be
required to reduce the level of uncertainty by elimi-
nating the identified sources of variability in a
stepwise manner.
3-4
-------
Where more than one effluent is contributing to toxic
impact, additional toxicity testing may be required.
Additional testing is only required if the regulatory
agency decides not to treat each effluent separately. If
effluents are considered as portions of an interactive
system, testing must be conducted to ascertain the
potential for additive, antagonistic or persistent
toxicity. Either chronic or acute toxicity testing may be
necessary depending on whether the receiving water
body is considered to be:
• "effluent-dominated" or
• "stream-dominated"
Guidelines for this determination have been devel-
oped and are shown in (6)
When multiple effluents affect a receiving-water
system it is also necessary to determine the "relative"
and "absolute" effects of each effluent. The appro-
priate procedure for conducting these effluent toxicity
tests is described in the technical support document
(6).
3.4 Ambient Toxicity Testing and Dye
Studies
Ambient toxicity testing can be used to determine
instream toxicity levels resulting from individual
discharges. The same test organisms used in the
tiered-testing procedures are exposed to receiving
water samples collected from selected sampling
stations above, at, and below the discharge points).
Chronic toxicity tests are generally conducted since
the primary concern following dilution of the effluent
is the effect of chronic, lowdose exposures on the
aquatic community.
The number and location of sampling stations should
be based on a knowledge of the mixing characteristics
of the effluent including the influence of other point
and nonpoint sources. The best way to characterize
the mixing and dispersion of the effluent is to conduct
dye studies. The information developed in these
studies can be used to determine which instream
concentration isopleths correspond to concentrations
used in the dilution series in effluent tests (see
Sections 3.2 and 3.3), assuming the toxicants behave
conservatively.
A widely used dye for this type of study is rhodamine
WT in 20 percent solution The dye can be purchased
as a liquid so that mixing of powder and water is not
required. Often the dye is not locally available, but can
be purchased from its manufacturer, Crompton and
Knowles Corporation, in Skokie, Illinois.
Rhodamine WT 20 percent solution has a specific
weight of about 1.19. However, because of its high
solubility it mixes rapidly with the river water and
soon becomes neutrally buoyant Consequently,
mixing of the dye with solvents less dense than water
(e.g., methanol) is not required.
The dye should be injected continuously into the
effluent of the discharger so that it is completely
mixed with the waste stream when it is discharged
into the receiving wat«r. The injection rate of the dye
should remain constant over the duration of the
study. Based on the expected study duration, the
quantity of dye required can be estimated, and
prepared beforehand. When possible, the effluent
discharge rate should also be kept constant.
Some time is required before the dye isopleths attain
their steady-state concentrations in the river. Figure
? -4 r,-ovi.-!0s ciiideJinesfqr distances of up to 10 miles
Lilowa di«rh*rg« SecBuse of stream dispersion, the
time exceeds the travel time to the location in
question. For example, for a stream velocity of 0.3 fps
and a distance of 3 miles, approximately 21 hours of
continuous release is required to establish steady-
state dye isopleths. The travel time is 15 hours and by
then the dye has attained about 60 percent of its
steady-state concentration, based on predictions
from the advection-dispersion equation using a
dispersion coefficient of 500 ftVsec.
Figure 3-4. Time required for a continuous release of dye to
reach steady-state concentration* »t selected
location* below the point of discharge. Not*:
the curves are based on a solution to the
advection-dtspersion equation which is used to
predict when dye concentration*are 95 percent
of steady-state levels.
48
40-
s 32"
I 24 ^
t>
j '
8-
V = Stream Velocity
01234567B910
Distance Below Point of Dye Injection. Miles
Instream sampling should begin upstream of the
outfall and progress downstream. Thus, sampling
near the outfall can commence before downstream
dye levels have attained steady state.
Typical background fluoresence in rivers is equivalent
to about 0.1 ^g/l as rhodamine WT, so dye concen-
trations should be above background levels, but also
within levels calibrated for fluorometers (typically
less than 200-300 i/g/l). Consequently a dilution of
3-5
-------
2000 or 3000 to 1 can normally be measured. If
dilutions greater than this are required for the study.
two separate continuous releases may be required.
one using a higher effluent concentration so that dye
concentration isopleths can be measured further
downstream
The dye injection rate should be selected so that the
dye is not visible after it has begun to mix with the
river water The USGS plans dye sHidies so that
concentrations do not exceed 10^/g/l at water treat-
ment plant intakes and other diversions. If the dye is
visible, concentrations will be high enough so that
instrument readings will be inaccurate and adverse
public reaction may be generated as well.
Figure 3-5 shows example dye isopleths that might
result from injection into wide and narrow rivers. For
the wide river the two-dimensional profile can be
maintained for large distances.
Figure 3-5 Dye uoplethi in wide and narrow river*.
Flow Direction
1000
Key 500 Dilution Isopleth
Flow Direction
I
50)
2000
3000
3bOO;
3600
3620
I Potentially
Long
Distance
ia| Wide River
1010
1010
1010
1010
1010
1011:
101 li
1011!
(b) Narrow River
Dye isopleths should be generated from the point of
discharge to below the no observable effects level
(NOEL) as determined from toxicity tests Figure 3-6
shows example limits of observable toxicity For the
narrow river, the NOEL extends to a dilution of
approximately 1010 to 1. Once complete mixing is
attained the concentration isopleths change very
slowly with distance. However the NOEL may have a
distinct downstream location that indicates toxicity is
decreasing for reasons other than dilution, as sug-
gested in Figure 3-6b
Figure 3-6. Region* of observable toxicity in wide and
narrow river*.
Key 500 - Dilution Isopleth
:•:•:•:;:; Region of Observable
Toxicity Effect
Flow Direction
95
•y*j
1010
1010
1010
1010
lOlOi
1011!
• f\ • •
1 O1 1 :
10111
I Potentially Long
Distance
•««
*•
-"
**•
—
IslWtde River
(b) Narrow River
Figure 3-7 illustrates a typical sampling network for
narrow and wide rivers. Sampling of dye concentra-
tions at a number of transects is required. For wide
rivers, samples should be taken from 4 to 5 points on
the transect. By putting the fluorometer in a boat and
moving it across the river and starting on the side that
the outfall is located, the lateral extent of the plume
can be readily determined. In shallow rivers, a flat
bottomed canoe can be used to move the fluorometer.
On a wide river, sampling may be required only 100
feet or so below the outfall, even though the river may
be 500 feet wide.
The fluorometer can be used to assist in selection of
downstream transects. Generally, change in dye
3-6
-------
concentration (based on measurements taken on the
same side of the river as the discharge) should not
exceed a factor of three to four between adjacent
transects so that detailed concentration isopleths can
be generated
The following formula can be used to estimate the
number of required transects:
N =
log
logRf
(3-6)
where N= number of transects
QR = river flow rate
Qw = point source flow rate
Rr = ratio of fluorometer readings between
two adjacent transects, measured on
the same side of the river as the
discharge.
For example if QR = 500, Qw = 0.3, and RF is specified
to be 2, then
log
N =
500
0.3
log 2
= 10.7 = 11 transects
Figure 3-7 Example sampling locations in wide and narrow
rivers.
Kev 500 - Dilution Isopleth
• Sampling Station
Flow Direction
Flow Direction
50
10OO
10103
ioio
"
(a) Wide River
loior'
1010
1011
1011|—
1011 *
(b) Narrow River
. Potentially
I Lone
* Distance
Once the dye readings along a transect are uniform
(say less than 5 percent difference between readings)
then complete transverse mixing has almost been
attained, so one reading per transect is sufficient
further downstream
Sampling at multiple depths may be necessary just
below the outfall Since vertical mixing is rapid (see
Figure 3-3), vertical profiles probably are not required
at a large number of locations. The fluorometer itself
is the best method of determining if sampling at
multiple depths is required To simplify this aspect of
sampling, a preselected standard can be used, where,
for.example, samples 1 foot off the bottom are
uniformly taken
For multiple discharges, ne aye studies and pro-
cedures outlined above are r^Hc_.uo Scpai-tely for
each discharge. The dye is injected in the downstream
discharge first, and then at the next upstream
discharge, and so on This will prevent upstream dye
from contaminating earlier surveys
The delineation of effluent plume configurations
using the results of the dye studies provides a basis
for comparing instream effluent concentrations with
the toxicity concentrations determined in Tier 1 and
Tier 2 toxicity tests. Where dye study results indicate
that effect-level concentrations are exceeded in-
stream, ambient toxicity tests should be conducted
Receiving-water samples should be obtained from
sampling locations within the potential impact zones
to conduct static-renewal exposure tests Sampling
stations should be placed at instream locations which
correspond to concentrations measured in the dilution
series in the effluent tests. For example, where
effluent testing shows the effluent NOEL is 10
percent, an instream station should be placed where
dilution is estimated to create a 10 percent instream
waste concentration. The results of the ambient
toxicity testing can be used to evaluate the persis-
tence of effluent toxicity and the decay rate of toxicity
This supplementary information is of value in setting
waste load allocations
3-7
-------
Chapter 4
Example Application
4.1 Dissolved Oxygen
Figure 4-1 shows an 80 mile (1 30 km) stretch of the
Eel River below the City of Dublin. Also shown on the
figure are Cache Creek, the Dublin wastewater
treatment plant, and historical water quality data
cotlected during the summer of 1 981 . The data show
that dissolved oxygen levels in the river have been as
low as 3.5 mg/l. The dissolved oxygen standard is 6.0
mg/l, expressed as a daily average. The state has
mandated that the municipality reduce their waste
loadings to be in compliance with the water quality
standard for dissolved oxygen. Consultants for the
municipality have been retained to design a summer
low flow survey so that data can be gathered for a
dissolved oxygen model of the river.
Figure 4-1.
Eel River end environs ihowing summer of
1981 water quality results.
•"•v
6 ) Locauor\ 6
17 Julv 1981
DO -81 mg I
OiOl 5 (10l 12 [20'
Location C
20 July 1981
DO = 35 mg/l
Location A
10 Augusi 1981
* --' 22° C
DO ^ 7 9 mg/l
Miles (km)
Location D
13 August 1981
_ = 21C C
DO - 5 1 mg I
Before deciding on their modeling approach, the
consultants first review the historical data. Based on
the data, they conclude that high loadings of CBOD
and NBOD from trig treatment plant are primarily the
causes of the depressed dissolved oxygen levels. Th»
data show that algal activity has been minimal ant'
the river is large enough so that diurnal temperature
changes are no more than 2 to 3°C Based on their
assessment of the problem, the consultants intend to
use a steady-state approach to dissolved oxygen
prediction, where the processes of CBOD, NBOD,
sediment oxygen demand (SOD), and reaeration are
simulated.
A rQio summer low flow is selected for the wasteload
allocation period. A stream survey will be conducted
during a summer low flow period to provide the
necessary data to calibrate the model The model will
then be applied to simulate the rQio conditions. The
sampling locations selected are shown in Figure 4-2.
They include locations to characterize:
• background levels in the river above the treatment
plant
• the treatment plant effluent and tributary
• the river just prior to mixing with the tributary
• intermediate locations in the river necessary to
locate the dissolved oxygen sag and to determine
the CBOD and NBOD profile
• water quality at the end of the reach
Based on historical data, and a preliminary model
application, the minimum dissolved oxygen level is
expected to occur near location 4. Locations 3,4, and
5 will help to accurately establish the shape of the
dissolved oxygen sag curve. Location 3 is far enough
below the treatment plant that the effluent is
expected to be well mixed before that location;
consequently multiple samples across a section are
not needed.
Table 4-1 summarizes the data that are to be
collected. Diurnal variations of efflueht loading
(station 2) and of instream quality at stations 3 and 4
will be quantified. Diurnal variations are needed to
predict daily average dissolved oxygen levels to
compare with the state standard. Instream diurnal
variations are expected to be due to wasteload
variation, and not to temperature and algal effects.
4-1
-------
Figure 4-2 Location of sampling itations on Eel River
STP
Denotes Sampling
Location 1
Additionally, a plug flow sampling event will be
conducted between stations 1 and 5 to help better
estimate NBOD and CBOD decay rates. Diurnal
loading variations are expected to make the range of
CBOD and NBOD concentrations at specific locations
quite large that accurate decay rates will be difficult to
estimate otherwise.
In-situ sediment oxygen demand rates will be deter-
mined at stations 1, 3, and 7. Station 1 represents
background conditions, station 3 is expected to show
the influence of the treatment plant discharge, and
station 7 is located in a recovery zone.
Because the river is fairly deep (4 ft or greater even
during low flow), the consultants intend to use an
historical reaeration rate.expression characterized by
a depth-velocity relationship. Specific tracer studies
are not planned. The water temperature is expected
to remain fairly constant over time, so that water
temperature simulation techniques are not needed.
Rather, water temperature effects will be considered
indirectly in terms of temperature effects on rate
constants and temperature effects on dissolved
oxygen saturation. Consequently meteorological data
are not needed.
The judgement and experience of the consultants and
water quality specialists employed by the munic-
ipality have been combined to design this particular
sampling program. Review of historical data, pre-
liminary model applications to the river, and under-
standing the behavior of rate coefficients such as the
reaeration rate constant, were all used to design the
survey.
Table 4-1. Summary of Beta to be Collected During Stream Survey for Otoeotved Oxygen Waate Load A»ocatk)o
Sampling Station Parameters Frequency Comments
1.
2.
3
Background station,
Eel River above
Dublin STP
Effluent of Dublin
STP
In Eel River 8 miles
below Dublin STP
• CBOD, NBOD, DO.
Temperature
• Row
• CBOO, NBOD, DO.
Temperature
• Flow
• CBOD, NBOD, DO.
Temperature
• 1 per day for
7 days
. USGS gage
• Every 3 hour* for
7 days
• Continuously
• Every 3 hours for
24 hours, plus
sample for plug
flow analysis
Station 1 is used to establish background level.
The diurnal variability is used to establish daily average
loads, and to help explain time variability in BOO and
DO at specified location downstream.
Stations 3. 4, and 5 show the diurnal response to waste
load variations. The plug flow sample is taken to corre-
spond with the passage of the centroid of dye released
at the treatment plant.
4. In Eel River 16 miles
below Dublin STP
5. In Eel River just
above Cache Creek
6. Mouth of Cache
Creek
same as station 3
same as station 3
same as station 1
7. In Eel River 4 miles same as station 3
below Cache Creek
same as station 3
same as station 3
same as station 1
same as station 1
Single flow rate estimates at the beginning and end of
survey will be sufficient if a continuous gage is not
available.
Flow rates are not needed.
4-2
-------
4.2 Organic Toxicant
Figure 4-3 shows two wastewater treatment plants
that discharge to the El Cahon River, which flows into
Lake Chabot. A limnological investigation has shown
that surficial sediments of Lake Chabot are contam-
inated with the polycyclic aromatic hydrocarbon
pyrene. Subsequent investigations in the river also
revealed high concentrations of pyrene in the bed and
occasional high pyrene concentrations in the water
column as far upstream as the Bently sewage
treatment plant. Sampling of the effluent from the
Bently and Vallejo plants has confirmed that these
two plants are sources of pyrene. To meet water
quality standards, the state has decided that the
loading of pyrene to the river is to be reduced and
allocated between the two sources.
Figure 4-3. El Cahon River. Lake Chabot, and environ*.
6(10)
Prior to collecting supplemental stream data to use in
the WLA analysis, the state first selects a modeling
approach and a sampling period. Since historical data
have indicated that pyrene levels have been highest
during the low flow period, the state has selected both
a ?Qio period (for chronic criteria) and a iQto period
(for maximum criteria) to perform the WLA. This
example problem, therefore, deals with sampling
during a low flow period.
The state selects a dilution or mass balance approach
to allocate pyrene from the Bently treatment plant.
Above the Vallejo treatment plant, however, the state
believes that pyrene concentration is not predictable
by pure dilution alone, based on the presence of
pyrene in the stream sediments.
The state decides to perform a preliminary analysis of
the fate of pyrene in the river, and to use the computer
model MICHRIV for the WLA (if needed) to simulate
the transport and transformation of pyrene in the
river between the two treatment plants. Table 4-2
summarizes data the state has collected on the fate of
pyrene. The data show that hydrolysis is probably
negligible, but that the biodegradation rate, while
unknown, is likely to be significant. The volatilization
rate is not shown in the table, but its importance can
be determined from Henry's Constant:
KH =
/VMW
760 -Sw
(4-1)
where P« - saturation vapor pressure, torr
MW = molecular weight
Sw = solubility in water, mg/l
For pyrene.
KH =
(6.9-IP'7)(202)
(760) (.140)
• = 1.3 • 10"* atm • mVmole
Table 4-2. Ptopecttea and Fata Pro cm M for Pyrane
_ are from (T7) untaaa otherwise noted)
• Molecular weight - 202
• Octanol-water partition coefficient. K^. -2-10*
• Saturation vapor prenura (lorr at 20*C), P, - 6 9 1CT7
• Solubility (mo/1 at 25*C), S» - 140 m»/l
• Biodegradation rate (1/dav): unknown but probably signifi-
cant from (49)
• Hydrolysis rate (1/day): unknown, but probably negligible
• Near surface direct photolysis rate (I/day at a light intensity
to • 2100 langleys/day), k^ - 24
• Wavelength of maximum light absorption (nm), X* - 330
This a very small KH, and indicates that volatilization
is negligible (probably between 0.01 /day a nd 0.001 /
day based on the two-f i Im theory of volatilization (27).
The high octanol-water partition coefficient (Km, = 2 x
10*) indicates that pyrene adsorbs to suspended and
bedded sediments, and will settle out in the stream-
4-3
-------
bed along with solids that are deposited there,
consistent with historical observations.
The near-surface direct photolysis rate is 24/day. The
expected photolysis rate in the stream can be
approximated by (6).
01
Dot.
(4-2)
where k«o= near surface rate. 1 /day
Mo = intensity of radiation from sunlight and
from laboratory source, respectively
D.Do = distribution coefficients in river and in
clear water, respectively
Z= water depth, m
MA*) = light attenuation in water at wavelength
A", 1/meter
The light attenuation term in Equation 4-2 can be
estimated from Table 4-3, excerpted from Mills (26).
For the El Cahon River during low flow conditions, the
attenuation factor is on the order of 0.1 for water type
C with depth of 1 m. SinceI0 = 2100 langleys/day and I
= 540 langleys/day.
540
2100
= 0.6/day
Depth of Water (m)
•(nm)
300
340
•Water
A
B
C
D
Water Type* 1 2
A 0.9 0.8
B O.S 0.4
C 0.1 0.06
0 0.03 0.01
A 0.9 0.9
B 0.7 0.5
C 0.2 0.08
D 0.04 0.02
Type Chi a (mg/l)
0.0
0.001 loligotrophic.
e.g.. Lake Tahoe)
0.01 (eutrophtc)
0.1 (highly «utrophic)
3 5
0.8 0.6
0.2 0.14
0.04 0.03
0.009 0.005
0.9 0.8
0.4 0.2
0.06 0.03
0.01 0.007
DOC
Miles (km)
Denotes Sampling
Location 1
The stations between the two point sources are
selected based on an assumed travel time of about
0.7 days between stations (it is assumed that the
state had previously determined travel times), and
considering that pyrene may photolyze and biode-
4-4
-------
Table 4-4. Summary of Data to be Collected During Streem Survay
Sampling Station Parameters Frequency
1. Background station in
El Canon River above
BentrySTP
2. Effluent of Bently STP
3. In El Canon river 6 mi
(10 km) below Bently
STP
In El Canon River
12 mi (20 km) below
BentlySTP
6. In El Canon River jusl
below Vallejo STP
6. Effluent of Vallejo
STP
Other:
Plug flow sampling:
between stations 3
and 4 (approxi-
mately), depending
on the travel time
corresponding to
sunrise
• Suspended solids
• Pyrene, total
• Flow rate
• Suspended solid*
• Pyrene, total
• Flow rate
Suspended solids
Pyrene, total
Pyrene, dissolved
Water temperature
Cross-sectional area
Water depth
• Same as at Station 3,
plus:
• Chlorophyll a
• Dissloved organic
carbon
• Same as Station 3
Same as Station 2
Dye
Total pyrene
Dissolved pyrene
Suspended solids
• Three times during 7-day survey
• Three times during 7-day survey
• Continuously (USGS gage)
• Every 3 hours for 7 days
• Every 3 hours for 7 days
• Continuously
Twice during 24-hour period
Every 3 hours for 24 hours
Twice during 24-hour period
Every 3 hours
Once
Once
Twice during 24-hour period
Twice during 24-hour period
• Same as Station 3
• Same as Station 2
Every two hours .from sunset to
sunrise beginning near Station 3
Uaad to establish background levels
used to confirm that background
pyrene concentrations are negligi-
ble
The frequency for suspended solids
sampling can be relaxed if time
variability of suspended solids Is
small, or if the suspended solids
concentrations In the river are in-
sensitive to affluent suspended
solids.
Samples for suspended solids and
dissolved pyrene should be taken at
the same time, and along with total
pyrene, used to find the partition
coefficient (Sea Table 2-1).
Suspended solids versus distance
profile* should be used to analyze
the importance of solids settling on
total pyrene in the water column.
If the state has the resources ch a
and DOC can a found at station 3
as well.
Same as Station 3
Same as Station 2
The dye is injected into the El
Canon River near the Berrtty STP so
than it is well mixed by the time It
; Station 3.
grade fairly rapidly. Based on the mixing character-
istics of the river, it has been established that
complete mixing of effluent and stream water is
achieved upstream of station 3. Thus, the state does
not need to take multiple samples laterally across a
transect.
The state chooses to sample at three hour intervals,
over a 24-hour period at stations 3, 4 and 5. Due to
manpower limitations, the stations are sampled
sequentially, beginning with station 3. The state is
aware that this is not as desirable as sampling
simultaneously at the three stations because of the
time variability of the waste loadings. The effluent
quantity and quality of the treatment plants are
monitored for a period of one week, beginning the day
before the instream sampling begins at station 3.
At the upstream boundary station, a nearby USGS
gauge continuously records the flow. Because pyrene
contamination has never been found in the river
above the Bently treatment plant outfall, only three
background grab samples are taken during the
sampling period (one every two days).
Once sampling is completed at stations 3,4 and 5, the
plug-flow sampling event is begun. Dye is injected
into the stream at the Bently treatment plan} so that it
is well-mixed at station 3 and arrives near sunset. The
effluent loading of pyrene at the time of dye injection
is recorded. Sampling for pyrene then begins at
station 3 as the peak dye concentration passes.
Samples are collected approximately every two hours
based on passage of peak dye concentrations, and,
continues through the night. Suspended solids
4-5
-------
concentrations are also taken to see if settling of
solids is significant. The state realizes tha1 the plug-
flow sampling event has to be carried out accurately
in order to determine the decay rate, because of the
time limitation (approx. 12 hours) before photolysis is
again active.
Once the state has completed the seven day sampling
program, enough information has been collected to
analyze the fate of pyrene in the river, and to calibrate
MICHRIV. The agency intends to run MICHRIV a
number of times, with different loading rates to see
how well the predictions match the envelope of
instream concentrations observed at locations 3, 4
and 5.
Following model calibration, it is expected that the
state will conduct a second survey for model verifica-
tion. Different conditions will intentionally be chosen
between the calibration and verification periods. For
example, if the calibration survey were conducted
under cloudy or rainy conditions when the solar
radiation is suppressed by as much as 50 to 70
percent, the verification survey would be conducted
under clear sky conditions.
4-6
-------
Chapter 5
References
1. Roesner, LA., P.R. Giguere, and D.E. Evenson.
1981 a. Computer Program Documentation for
trie Stream Quality Model QUAL-II. U.S. Envi-
ronmental Protection Agency, Athens, GA.
EPA-600/9-81-014.
2. Roesner, L.A., P.A. Giguere, and D.E. Evenson.
1981 b. User's Manual for Stream Quality Model
(QUAL-II). U.S. Environmental Protection Agency
Environmental Research Laboratory, Athens,
GA. EPA-600/9-81-015.
3 Brown, L.C., and T.O. Barnwell, Jr. 1985.
Computer Program Documentation for the En-
hanced Stream Water Quality Model QUAL-2E,
U.S. Environmental Protection Agency Envi-
ronmental Research Laboratory, Athens, GA.
EPA-600/3-85/ 065
4. Ambrose, R.B.. S.I. Hill, and L.A. Mulkey. 1983.
User's Manual for the Chemical Transport and
Fate Model TOXIWASP. Version 1. EPA-600/3-
83-005. USEPA Athens, GA 30613.
5. Delos, C.G., W.L. Richardson, J.V. DePinto, R.B.
Ambrose, P.W. Rodgers, K. Rygwelski, J.P. St.
John, W.J. Shaughnessy, T.A. Faha, W.N.
Christie. 1984. Technical Guidance Manual for
Performing Waste Load Allocations. Book II
Streams and Rivers. Chapter 3 Toxic Sub-
stances. EPA-440/4-84-022. USEPA Office of
Water, Washington, DC 20460.
6. U.S. Environmental Protection Agency. 1985.
Technical Support Document for Water Quality-
Based Toxics Control. USEPA Office of Water,
Washington, DC 20460.
7 Mount, D.I., N.A Thomas, T.J. Norberg, M.T.
Barbour, T.H. Roush, and W.F. Brendes. 1984.
Effluent and Ambient Toxicity Testing and
Instream Community Response on the Ottawa
River, Lima, OH. EPA-600/3-84-080.
8 Grenney, W.J. and A.K. Kraszewski. 1981.
Description and Application of the Stream
Simulation and Assessment Model: Version IV
(SSAM IV). Instream Flow Information Paper.
U.S. Fish and Wildlife Service, Fort Collins. CO,
Cooperative Instream Flow Service Group.
9. Raytheon Company. Oceanographic & Envi-
ronmental Services 1974 New England River
r*--;-.'. •"--'•Ving Project, Vol. Ill—Documenta-
tion Rpnort r>ar, i—RECEIV-II Water Quantity
and Quality Model. For Office of Water Pro-
grams, U.S. Environmental Protection Agency,
Washington, DC.
10. Smith, D.I. 1978. Water Quality for River-
Reservoir Systems. Resource Management
Associates, Inc., Lafayette, CA. For U.S. Army
Corps of Engineers, Hydrologic Engineering
Center (HEC), Davis, CA.
11. Driscoll, E.D, J.L. Mancini, and P.A. Mangarella
1983. Technical Guidance Manual for Perform-
ing Waste Load Allocations, Book II Streams and
Rivers, Prepared for Office of Water Regulations
and Standards, Monitoring and Data Support
Division, Monitoring Branch, USEPA, Washing-
ton, DC.
12. Theurer, F.D. and K.A. Voos. 1982. Instream
Water Temperature Model Instream Flow In-
formation Paper No. 16. Instream Flow and
Aquatic Systems Group, U.S. Fish and Wildlife
Service, Fort Collins, CO.
13. National Oceanic and Atmospheric Administra-
tion (NOAA). 1974. Climatic Atlas of the United
States.
14. Duke, J.H., Jr. and F.D. Masch. 1973. Computer
Program Documentation for the Stream Quality
Model DOSAG3, Vol. I. Water Resources
Engineers, Inc., Austin, Texas. For U.S. Envi-
ronmental Protection Agency, Systems Devel-
opment Branch, Washington, DC 20460.
15. Johanson, R.C., J.C. Imhoff, and H.H. Davis,
1980. User's Manual for Hydrological Simula-
tion Program—Fortran (HSPF). Hydrocomp, Inc.,
Mountain View, CA. For U.S. Environmental
Protection Agency, Athens. GA. EPA-600/9-
80-015.
16. O'Connor, D.J. and D.M. Di Toro. 1970. Photo-
synthesis and Oxygen Balance in Streams.
ASCE, Journal of Sanitary Engineering Division,
ASCE, Vol. 96, No. SA2, pp. 547-571.
5-;
-------
17 Deb. A.K and D Bowers 1983. Diurnal Water
Quality Modeling—A Case Study Journal Water
Pollution Control Federation, Vol 55, No. 1 2. pp.
1476-1488
18 Bowie, G.L., W.B Mills, D.B Porcella, C.L.
Campbell, J.R Pagenkopf, G.L. Rupp, C. Cham-
berlin.K M.Johnson, S A .Gherini. 1985. Rates.
Constants, and Kinetics Formulations in Surface
Water Quality Modeling Edition 2 US Envi-
ronmental Protection Agency, Environmental
Research Laboratory, Athens, GA.
19. Zison, S.W., W.B Mills, D. Deimer. and C.W.
Chen 1978 Rates, Constants and Kinetics
Formulations in Surface Water Quality Model-
ing Prepared by Tetra Tech, Inc., Lafayette, CA.
For Environmental Research Laboratory, U.S.
Environmental Protection Agency, Athens, GA.
EPA-600/3-78-105 335 p.
20. Jorgensen, S.E. (ed.). 1979 Handbook of Envi-
ronmental Data and Ecological Parameters.
International Society for Ecological Modeling.
21 Steele, J.H 1965. Notes on Some Theoretical
Problems in Production Ecology. In: Primary
Production in Aquatic Environments. C.R.
Goldman (ed.). University of California Press,
Berkeley, CA. pp. 393-398
22 Texas Water Development Board. 1970
DOSAG-I Simulation of Water Quality in
Streams & Canals Program Documentation &
Users Manual. EPA OWP TEX-DOSAG-1.
23. Braster, R.E.. S C Chaptra, and G.A. Nossa
1975. SNSIM-1 /2 A Computer Program for the
Steady State Water Quality Simulation of a
Stream Network. 4th Edition
24. Streeter, H.W. andE.B.Phelps. 1925. A Study of
the Pollution and Natural Purification of the
Ohio River. U.S. Public Health Service, Washin-
gton, DC, Bulletin 146.
25. Willinoham, W.T. 1976. Ammonia Toxicity.
EPA-908/3-76-001 USEPA Washington, DC
20460.
26. National Council of the Paper Industry for Air
and Stream Improvement, Inc. (NCASI). 1982.
The Mathematical Water Quality Model QUAL-II
and Guidance for its Use Revised Version,
Technical Bulletin No. 391.
27. Mills, W.B., Porcella, D.B., Gherini, S.A., Ungs.
M.J., Summers, K.V., and Haith, D.A. (1985).
Water Quality Assessment: A Screening Pro-
cedure for Toxic and Conventional Pollutants in
Surface Waters and Ground Waters. Prepared
for U.S. Environmental Protection Agency,
Athens, GA EPA/600/6-85/002ja,b.
28. DiToro. D.M..J.J Fitzpatrick. and RV Thomann
1981. Water Quality Analysis Simulation Pro-
gram (WASP) and Model Verification Program
(MVP)—Documentation Hydroscience, Inc ,
Westwood, NJ. For U.S Environmental Pro-
tection Agency, Duluth. MN
29. DiToro, D.M., D.J. O'Connor, R V Thomann, and
J.L. Mancini. 1975 Phytoplankton-Zooplankton
Nutrient Interaction Model for Western Lake
Erie. In: Systems Analysis and Simulation in
Ecology, Vol. III. B.C. Pattpn (ed.) Academic
Press, Inc., New York, NY, 423 pp
30. Chapra, S.^and G.A. Nossa. 1974 Documenta-
tion for HAR03. 2nd Ed -II"' "-„.-'•" New
York, NY.
31. Nossa, G.A. Nov, 1978 FEDBAK03—program
Documentation and User's Guide. USEPA
Region II, New York, NY.
32. Baca. R.G., W.W. Waddel, C.R. Cole, A Brand-
stetter, and D.B. Clearlock. 1973. EXPLORE-1: A
River Basin Water Quality Model Battelle, Inc.,
Pacific Northwest Laboratories, Richland, WA
33, Rich, L.G. 1973. Environmental Systems Engi-
neering, McGraw-Hill, Inc. New York, NY
34. Tsivoglou, E.C., R.L. O'Connell, C.M. Walter, P.J.
Godsil, and G.S. Logsdon. 1965. Tracer Meas-
urements of Atmospheric Reaeration-l. Labora-
tory Studies, Journal Water Pollution Control
Federation, Vol. 37. No. 10. pp. 1343-1362
35. Tsivoglou, E.C., J.B. Cohen, S.D Shearer, and
P.J. Godsil. 1968. Tracer Measurement of
Stream Reaeration. II. Field Studies. Journal
Water Pollution Control Federation, Vol. 40, No
2. Parti, pp. 285-305.
36. Tsivoglou, E.G., and J.R Wallace 1972. Char-
acterization of Stream Reaeration Capacity,
EPA-R3-72-012, Prepared for Office of Research
and Monitoring, USEPA, Washington, DC
37. Tsivoglou, E.C., and LA. Neal. 1976. Tracer
Measurement of Reaeration: 3. Predicting the
Reaeration Capacity of Inland Streams, Journal
Water Pollution Control Federation, Vol. 48. No.
12. pp. 2669-2689.
38. Rathbun, R.E., D.W. Stephens, D.J. Shultz, and
D.Y. Tai. 1978. Laboratory Studies of Gas
Tracers for Reaeration, ASCE, J. Environmental
Engineering Division, Vol. 104, No. EE1. pp.
215-229.
39. O'Connor, D.J., and W.E. Dobbins. 1958.
Mechanism of Reaeration in Natural Streams,
ASCE Transactions, Paper No. 2934. pp. 641-
684.
5-2
-------
40 Owens. M., R.W Edwards, and J.W. Gibbs
1964. Some Reaeration Studies in Streams, Int.
J. Air Wat. Poll., Vol. 8, pp. 469-486.
41 Churchill, M.A., H.L. Elmore, and R A. Bucking-
ham. 1962 The Prediction of Stream Reaeration
Rates, ASCE, Journal Sanitary Engineering
Division, Vol. No. 88, SA4, pp. 1-46.
42. Whittemore, R.C. Implementation of In-Situ and
Laboratory SOD Measurements in Water Quality
Modeling, in Sediment Oxygen Demand: Pro-
cesses, Modeling, and Measurement. K. Hatcher
(ED), Institute of National Resources, U Georgia,
(In Press,.1986)
43. Whittemore, R.C. Recent Studies on the Com-
parison of In-Situ and Laboratory SOD Meas-
urement Techniques. Proceedings, Stormwater
and Water Quality Management Modeling
(SWMM) Conference, Dec 5-6, 1985. Toronto,
Canada. William James, McMaster Univ, (Ed).
44. Whittemore, R.C. A Review of In-Situ and
Laboratory SOD Measurement Comparisons,
NCASI Technical Bulletin No. 386, New York, NY
(Nov. 1982).
45 Whittemore, R.C. A Review of Uncertainty in the
In-Situ Measurement of Sediment Water
Column Interactions. NCASI Technical Bulletin
No. 467, New York, NY. (Aug, 1985).
46 Hatcher, K.J., and D. Hicks (Eds). 1986. Sedi-
ment Oxygen Demand; Processes, Modeling
and Measurement. Proceedings of the WPCF
Symposium, Published by Institute of National
Resources, Univ of Georgia, Athens, GA.
47. Velz, C.J 1984. Applied Stream Sanitation.
John Wiley and Sons. New York, NY 10016.
48 Hern, S.C., G.T Flatman, W.L. Kinney, F.P. Beck,
J.E. Pollard, A.B. Crockett. 1983. Guidelines for
Field Testing, Aquatic Fate and Transport
Models. U.S. Environmental Protection Agency.
EPA-600/4-85-030.
49 Lyman, WJ , W.F Reehl, and D.H. Rosenblatt.
1982. Handbook of Chemical Property Estima-
tion Methods: Environmental Behavior of
Organic Compounds. McGraw-Hill Book Com-
pany
50. Leo. A., C. Hansch, and D. Elkins. 1971. Partition
Coefficients and Their Uses. Chemical Reviews,
Volume 71, No. 6.
51. Donigian, A.S., T.Y.R. Lo, and E.W. Shanahan.
1983 Rapid Assessment of Potential Ground-
Water Contamination Under Emergency Re-
sponse Conditions. EPA Report USEPA Athens,
G A 30613
52 Thibodeaux. 1979. Chemodynamics Environ-
mental Movement of Chemicals in Air, Water.
and Soil. John Wiley & Sons, New York p. 501
53. Mabey, W., and T. Mill. 1978. Critical Review of
Hydrolysis of Organic Compounds in Water
Under Environmental Conditions J. Phys. Chem.
Ref. Data 7(2):383-415
54. Hubbard, E.F., F.A. Kilpatrick. L.A. Martens, and
J.F. Wilson, Jr. 1982. Measurement of Time of
Travel and Dispersion in Stream by Dye Tracing,
USGS
55. Briggs, J.C. and J.F. Ficke. 1977. Quality of
Rivers of the United States, 1975 Water Year-
Based on the Nationjl Stream Quality Account-
ing Network(NASQAN) LfSGS Open-File Report
78-200.
56. Fisher, H.B., E.J. List, R.C.Y. Koh, J. Imberger,
and W.H Brooks. 1979. Mixing in Inland and
Coastal Waters. Academic Press. 483 pp
57. Peltier, W., and C.I. Weber. 1985 Methods for
Measuring the Acute Toxicity of Effluents to
Aquatic Organisms. 3rd Ed. Office of Research
and Development, Cincinnati, OH. EPA-600/4-
85-013. April, 1985.
58. Horning, W., and C.I. Weber. 1985. Methods for
Measuring the Chronic Toxicity of Effluents to
Aquatic Organisms. Office of Research and
Development, Cincinnati, OH. EPA-600/4-85-
014.
5-3
-------
Chapter 6
Appendix
References for Instream Data Collection and
Laboratory Techniques for Analysis of Water and Waste Water
Techniques of Water Resources investigations
til.S. Geological Survey)
Barnett, P.P. and E.G. Matlory Jr. 1971. Determina-
tion of Minor Elements in Water by Emission
Spectroscopy 31 p. Bk 5, A2
Benson, M.A. and Tate Dalrymple. 1967 General
Field and Office Procedures for Indirect Discharge
Measurements. 30 p. Bk 3, A1
Bodhaine, G.L. 1966. Measurement of Peak Dis-
charge at Culverts by Indirect Methods. 60 p. Bk 3,
A3
Buchanan, T.J. and W.P. Somers. 1968. Stage
Measurements at Gaging Stations. 28 p. Bk 3, A7
Buchanan, T.J. and W.P. Somers. 1969. Discharge
Measurements at Gaging Stations 65 p. Bk 3, A8
Carter, R.W. and Jacob Davidian. 1968. General
Procedure for Gaging Streams. 13 p Bk 3, A6
Craig, J.D. 1983. Installation and Service Manual for
U.S Geological Survey Manometers. 57 p. BkS. A2
Dalrymple, Tate and M.A. Benson. 1967. Measure-
ment of Peak Discharge by the Slope-Area Method.
12 p. Bk3, A2
Davidian, Jacob 1984. Computation of Water-Surface
Profiles in Open Channels. 48 p. Bk 3, A15
Friedman, L.C and D.E Erdmann. 1982. Quality
Assurance Practices for the Chemical and Bio-
logical Analyses of Water and Fluvial Sediments.
181 p. Bk5, A6
Goerhtz, D.F. and Brown. 1972. Methods of Analysis
of Organic Substances in Water. BkS. A3
Greeson, P.E., T.A. Ehlke, G.A Irwin, B.W. Lium, and
K.V. Slack (editors). 1977. Methods for Collection
and Analysis of Aquatic Biological and Microbio-
logical Samples. 332 p Bk 5, A4
Guy, H.P. 1969. Laboratory Theory and Methods for
Sediment Analysis 58 p. Bk 5, C1
Guy, H.P. 1 970. Fluvial Sediment Concepts. 55 p. B\
3. C1
Guy, H.P. and V.W. Norman. 1970. Field Methods for
Measurement of Fluvial Sediment. 59 p. Bk 3, C2
Hubbard, E.F., Kilpatrick, F.A., Martens, L.A., and
Wilson, J.F., Jr., 1982. Measurement of Time of
Travel and Dispersion in Stream by Dye Tracing. 44
p. Bk 3, A9
Hulsing, Harry. 1967. Measurement of Peak Dis-
charge at Dams by Indirect Methods. 29 p.
Jenkins, C.T. 1970. Computation of Rate and Volume
of Stream Depletion by Wells. 17 p. Bk 4, D1
Kennedy, E.J. 1983. Computation of Continuous
Records of Streamflow. 53 p. Bk 3. A13
Kennedy, E.J. 19B4. Discharge Ratings at Gaging
Stations. 59 p. Bk 3. A10
Kilpatrick, F.A. and V.R. Schneider. 1983. Use of
Flumes in Measuring Discharge. 46 p. Bk 3, A14
Kilpatrick, F.A., and Cobb, E.D. 1985. Tracer Dis-
charge Measurement. Bk 3, A16.
Laenen, Antonius. 1985. Acoustic Velocity Meter
Systems, TWRI. 38 p., Bk 3, A16
Matthai, H.E. 1967. Measurement of Peak Discharge
at Width Contractions by Indirect Methods 44 p Bk
3, A4
Porterfield, George 1972. Computation of Fluvial-
Sediment Discharge 66 p Bk 3, C3
Riggs, H.C. 1968. Some Statistical Tools in Hydrology.
39 p. Bk4, A1
Riggs, H.C. 1968 Frequency Curves. 15 p Bk 4, A2.
Riggs. H.C 1972. Low-Flow Investigations. 18 p Bk
4, B1
Riggs, H.C and C.H. Hardison 1973 Storage Anal-
yses for Water Supply. 20 p. Bk 4, B2
6-1
-------
Riggs, H.C. 1973. Regional Analyses of Streamflow
Characteristics. 15 p. Bk4, B3
Skougstad. M.W. and others (editors). 1979. Methods
for Determination of Inorganic Substances in Water
and Fluvial Sediments. 626 p. Bk 5, A1
Smoot, G.F. and C.E. Novak. 1968. Calibration and
Maintenance of Vertical-Axis Type Current Meters.
15 p.
Smoot, G.F. and C.E. Novak. 1969. Measurement of
Discbarge by Moving-Boat Method. 22 p. Bk 3, A11
Stevens Jr., H.H., J.F. Ficke, and G.F. Smoot. 1975.
Water Temperature Influential Factors, Field
Measurement and Data Presentation. 65 p. Bk 1,
D1
Thatcher, L.L., V.J. Janzer, and K.W. Edwards. 1977.
Methods for Determination of Radioactive Sub-
stances in Water and Fluvial Sediments. 95 p. Bk 5,
A5
Wershaw, R.L., M.J. Fishman, R.R. Grabbe, and I.E.
Lowe. 1984. Methods for the Determination of
Organic Substances in Water and Fluvial Sedi-
ments.
Wilson Jr., J.F., Ernest D. Cobb, and Frederick A.
Kilpatrick. Fluorometric Procedures for Dye Trac-
ings. 1984. Bk3, A12
Instream Flow Information Publications
(U.S. Fish and Wildlifa)
Bayha, K.D. 1978. Instream Flow Methodologies for
Regional and National Assessment. Instream Flow
Information Paper No. 7. U.S.D.I. Fish and Wildlife
Service. FWS/OBS-78/61. 98 p. Available from
NTIS(PB80181100).
Bovee. K.D. and T. Cochnauer. 1977. Development
and Evaluation of Weighted Criteria. Probabilrty-of-
Use Curves for Instream Flow Assessment: Fish-
eries. Instream Flow Information Paper No. 3.
U.S.D.I. Fish and Wildlife
Service. FWS/OBS-77/63.39 p. Available from NTIS
(PB 286 848).
Bovee, K.D. 1978. Probability of Use Criteria for the
Family Salmonidae. Instream Flow Information
Paper No. 4 U.S.D.I. Fish and Wildlife Service.
FWS/OBS-78/07 53 p. Available from NTIS (PB
286 849)
Bovee, K.D. and R.T. Milhous. 1978. Hydraulic
Simulation in Instream Flow Studies: Theory and
Techniques. Instream Flow Information Paper No.
5. U.S.D.I. Fish and Wildlife Service. FWS/OBS-
78/33 143 p. Available from WELUT and NTIS (PB
287015)
Bovee, K.D. 1982. A Guide to Stream Habitat Analysis
Using the Instream Flow Incremental Methodology
Instream Flow Information Paper No. 12. U.S.D.I
Fish and Wildlife Service. FWS/OBS-82/26 248
p. Available from NTIS (PB 83 131 052).
Grenney, WJ.andA.K. Kraszewkt. 1981. Description
and Application of the Stream Simulation and
Assessment Model Version IV(SSAM IV). Instream
Flow Information Paper No. 17. U.S.D.I. Fish and
Wildlife Service. FWS/OBS-81/46 199 p. Out of
print. Available from NTIS (PB 82 241 712).
Hyra, R. 1978. Methods of Assessing Instream Flows
for Recreation. Instream Flow Information Paper
Nr * U P P i P'«h and Wildlife Service. FWS/
0 tS-78/34. 52 p. wailable from NTIS (PB 285
9o /1 or C»K> (024-010-00469-0).
Lamb. B.L. and D.A. Sweetman. 1979. Guidelines for
Preparing Expert Testimony in Water Management
Decisions Related to Instream Flow Issues. In-
stream Flow Information Paper No. 1. Revised.
U.S.D.I. Fish and Wildlife Service. FWS/OBS-
79/37. 33 p. Available from NTIS (PB 80-162761).
Lamb, B.L. (editor). 1977. Protecting Instream Flows
Under Western Water Laws: Selected Papers.
Instream Flow Information Paper No. 2. U.S.D.I.
Fish and Wildlife Service. FWS/OBS-77/47. 65 p.
Available from NTIS (PB 272 691).
Milhous, R.T., D.L. Wegner, and T. Waddle. 1981.
Users Guide to the Physical Habitat Simulation
System (PHABSIM). Instream Flow Information
Paper No. 11. U.S.D.I. Fish and Wildlife Service.
FWS/OBS-81 /43 (revised). 475 p. Available from
NTIS (PB 84 199 736).
•
Olive, S.W. 1981 a. Protecting Instream Flows in
California: An Administrative Case Study. Instream
Flow Information Paper No. 14. U.S.D.I. Fish and
Wildlife Service. FWS/OBS-82/34.32 p. Available
from WELUT and NTIS (PB 83 169 482).
Olive, S.W. 1981b. Protecting Instream Flows in
Idaho: An Administrative Case Study. Instream
Flow Information Paper No. 15. U.S.D.I. Fish and
Wildlife Service. FWS/OBS-82/35 Available from
WELUT and NTIS (PB 83 169 490).
Olive, S.W. 1983. Protecting Instream Flows in Iowa:
An Administrative Case Study. Instream Flow
Information Paper No. 20. U.S.D.I. Fish and Wildlife
Service. FWS/OBS-83/18 35 p. Available from
WELUT.
Pruitt, T.A. and R.L. Nadeau. 1978. Recommended
Stream Resource Maintenance Flows on Seven
Southern Idaho Streams. Instream Flow Informa-
tion Paper No. 8. U.S.D.I. Fish and Wildlife Service.
FWS/OBS-78/68. 67 p. Available from NTIS (PB
287 849), or GPO (024-010-00496-7).
6-2
-------
Sweetman, D.A. 1980 Protecting Instream Flows in
Montana Yellowstone River Reservation Case
Study. Instream Flow Information Paper No. 10.
U.S.D.I. Fish and Wildlife Service. FWS/OBS-
79/36 75 p. Available from NTIS (PB 81 236 069)
Theurer, F.T., K.A. Voos, and W.J. Miller. Instream
Water Temperature Model. Instream Flow Informa-
tion Paper No. 16. U.S.D.I. Fish and Wildlife Service.
FWS/OBS-84/15. 372 p. Available from WELUT
and NTIS.
Wassenberg. P.S., S. Olive, J.L. Demon, and C.B.
Stalnaker. 1979. Elements in Negotiating Stream
Flows Associated with Federal Projects. Instream
Flow iiiiuu.-iotiun Paper No. 9. U.S.D.I. Fish and
Wild) ««Se-»»-c rws/OBS-79/03 41 p. Available
from NTIS (PB 80 146 202}.
U. S. Environmental Protection Agency
Plumb, R.H 1984. Characterization of Hazardous
Waste Sites, a Methods Manual Volume III.
Available Laboratory Analytical Methods. EPA-
600/4-84-038. USEPA, Environmental Monitoring
Systems Laboratory, Las Vegas, NV.
U.S. Environmental Protection Agency. 1977. Samp-
ling and Analysis Procedures for Screening of
Industrial Effluents for Priority Pollutants. USEPA
Washington, DC 20460
U.S. Environmental Protection Agency. 1978a. Data
Collection Quality Assurance for the Nationwide
Urban Runoff Program Water Planning Division,
USEPA Washington, DC 20460 41 p.
U.S. Environmental Protection Agency. 1979. Hand-
book for Analytical Quality Control in Water and
Wastewater Laboratories EPA-600/4-79-019
Environmental Monitoring and Support Laboratory,
Cincinnati, OH
U.S. Environmental Protection Agency. 1979b {re-
vised March, 1983). Methods for Chemical Analysis
of Water and Wastes. EPA-600/4-79-020. Envi-
ronmental Monitoring and Support Laboratory,
Cincinnati, OH.
U.S. Environmental Protection Agency. 19826
Handbook for Sampling and Sample Preservation of
Water and Wastewater EPA-600/4-82-029. Envi-
ronmental Monitoring and Support Laboratory,
Cincinnati, OH.
Miscellaneous
American Public Health Association, American Water
Works Association, Water Pollution Control Feder-
ation. 1981. Standard Methods for the Examination
of Water and Wastewater—1 5th Edition of Ameri-
can Public Health Association, Washington, DC
1139 p.
American Society for Testing and Materials 1982
Annual Book of ASTM Standards. Part 31. Water
ASTM, Philadelphia, PA. 1544. p.
American Society for Testing and Materials 1983.
Annual Book of ASTM Standards. Water and
Environmental Technology Volume 11.01. ASTM,
Philadelphia, PA. 752 p.
USCOE, USEPA. 1981 Procedures for Handling and
Chemical Analysis of Sediment and Water Sam-
ples. Waterways Experiment Station, P.O. Box631
Vicksburg, MS 39180
Ingram, W.M., Mackenthun, K.M., and Bartsch, A.F.
1967. Biological field Investigative Data for Water
Pollution Surveys FWPCA Superintendent of
Documents, U.S. Government Printing Office.
USGS. Development and Testing of Highway Storm-
Sewer Flow Measurement and Recording System
Water-Resources Investigations Report 85-4111.
Hem, J.D. Study and Interpretation of the Chemical
Characteristics of Natural Water. Third Edition.
USGS Water-Supply Paper 2254
Gordon, A.B., and Katzenbach, M 1983. Guidelines
for the Use of Water Quality Monitors USGS Open-
File Report 83-681.
Rantz, S.E., and others. 1982. Measurement and
Computation of Streamflow; Vol 1, Measurement
of Stage and Discharge; Vol 2. Computation of
Discharge USGS Water-Supply Paper2175 631 p.
U.S. Environmental Protection Agency 1979. Moni-
toring Requirements. Methods and Costs for the
Nationwide Urban Runoff Program Water Planning
Division, USEPA, Washington, DC 20460
63
QOVCftNMCNT
OWCt 198b-6'.6-116/ "» 0 6 I.
------- |