HAZARDOUS WASTE
MANAGEMENT SEMINAR
Prepared For
U.S. ENVIRONMENTAL PROTECTION AGENCY
ECONOMIC ANALYSIS DIVISION
Washington, DC
consultants in environmental manaoement
P.0.8ox40284,Nashville,TN 37204(615)794-0110
-------
HAZARDOUS WASTE MANAGEMENT
SEMINAR SCHEDULE
June 26, 1981
I. CHARACTERIZATION OF HAZARDOUS WASTES
A. Chemistry of Hazardous Wastes
B. Hazardous Waste Characterization Tests
II. ENVIRONMENTAL FATE AND EFFECTS
III. DESIGN AND UPGRADING OF HAZARDOUS WASTE MANAGEMENT
FACILITIES
A. General Site Selection Criteria
B. Liner and Cover Criteria for Landfills and
Impoundments
C. Gas Collection Systems
D. Leachate Collection
E. Monitoring Requirements
IV. OTHER DISPOSAL OPTIONS
A. Landfarming
B. Incineration
V. CASE HISTORY
-------
I. CHARACTERIZATION OF HAZARDOUS WASTES
(Oral Presentation)
-------
II. ENVIRONMENTAL FATE AND EFFECTS
-------
BIOASSAY CONCEPTS
Evaluation of bioassay data allows identification of the concentration
and/or loading of constituents which cause a measurable reaction in a
living organism. Bioassay methodology is easily influenced by factors
which are difficult to control. These factors include:
1. The characteristics of the dilution water.
2. The selection of suitable test organisms.
3. The condition of the test organisms.
4. The degree of acclimation of the test organisms to bioassay
testing conditions.
Under many situations the most important condition causing unacceptable
bioassay results is the quality of the dilution water. The dilution water
-------
130
should be selected on the basis of the following considerations:
1. The characteristics will not affect the response of the
selected organisms to the contaminants.
2. It is representative of the receiving water for the
particular discharge.
Static (batch) tests are commonly used to screen a wide variety of
testing alternatives providing information on the approximate dosages
for lethality. Thus, the batch test allows the range for continuous-
flow tests to be determined. Continuous-flow studies are used to
acclimate the test organism and to evaluate the effect of the contaminants.
Results of the bioassays with aquatic organisms are subject to a
great degree of variability depending on the quality of the dilution
water, the test species, the quality of the organisms and the testing
apparatus. These parameters must be considered before bioassay results
can be evaluated and interpreted with respect to the impact on the
aquatic biota.
Numerous techniques exist for establishing numerical values des-
cribing lethality. The terminology normally used is the lethal
concentration for 50 percent of the organisms tested, based on a
particular time interval. This parameter is designated as LC 50 and
must be accompanied by the time interval for which the lethal concentration
was established. The LC 50 value must also be accompanied by confidence
limits (normally based on a 95 percent confidence level). The techniques
for the determination of the confidence level will be described in more
detail in a subsequent section.
-------
131
Definition of Terms. In order to develop a more detailed
description of the procedures employed herein, a brief review of the
bioassay terminology is presented below:
Bioassay - A test in which the characteristics of a contaminant
are determined by the reaction of a test organism
to it.
Median Lethal Concentration - The usual method of reporting
results. The current trend is to use the symbol
LC 50 (lethal concentration for 50 percent
mortality). The symbol TL 50 (formerly TL ) meaning
the tolerance limit has also been used. The two
terms have the same numerical value.
Mean Lethal Dose - A measure of a toxicant which is sorbed by
the test organism (LD 50).
Acute - A stimulus, severe enough to bring about a rapid
response (usually within 4 days for fish).
Sub-Acute - Involving a stimulus which is less severe than the
acute stimulus which produces a response in a longer
time and may be chronic.
Chronic - Involving a stimulus which is lingering or continuous;
often signifies periods of about one-tenth of a life
span or more.
Mean Effective Time - Time of occurrence of LC 50 (ET5Q).
-------
132
LethaJ_ - Level of contaminants causing death or sufficient to
cause it by indirect action.
Sub-Lethal - Below the level which directly causes death.
Cummulative - An increase in effect on the organism brought about
by successive additions of the contaminants at different
times or in different ways.
Delayed - Symptoms which do not appear until an appreciable time
after exposure.
Long-Term - Chronic but more indefinite.
Short-Term - Acute but more indefinite.
GENERAL BIOASSAY PROCEDURES
The bioassay procedures employed during this study conformed to
established techniques and procedures outlined in Standard Methods for
the Examination of Hater and Hastewater, 13th Edition, and by the Environ-
mental Protection Agency3. After a field laboratory was established at
the Deer Park site, static screening tests were conducted to evaluate the
range of concentrations to be used during the continuous-flow bioassay
studies. These screening tests also provided qualitative information on
organism reaction which aided in the interpretation of the continuous-flow
a
"Methods for Acute Toxicity Tests with Fish and Macroinvertebrates,
Effluent Tests Using the Flow-Through Techniques", (Tentative Method)
September 11, 1973, EPA Committee on Methods for Toxicity Tests with
Aquatic Organisms.
-------
133
bioassay studies data. Brief descriptions of the testing procedures employed
in the bioassay are presented below.
Requirements for the Test Organisms. There is a tendency in bioassay
testing to use standard species in the evaluation of toxicological effects.
The disadvantage to this technique is that organisms which are not indigenous
to a particular environment are used as a basis for establishing toxicological
effects for the native population. The species selected to be representative
must be readily adaptable to laboratory conditions, respond naturally to
extended captivity, readily available in sufficient quantity, and obtained
from a common source with similar past history.
It is desirable to have approximately ten test organisms at each con-
centration of wastewater. Because of the size of the test organisms, it may
be necessary to divide them into two groups and test them in separate con-
tainers at the same concentration. The test organisms should be chosen
from the total population by random and also randomized among the available
test tanks.
Size of Test Organisms. Fish should be small, preferably less than
8 cm and 5 g. Size variation in the fish population should be maintained
so that the longest fish is no more than 1.5 times the length of the
shortest.
The containers used during the bioassay study should be inert to
prevent the solubilization of sorbable compounds which might interfere
with the bioassay evaluation.
Dilution Water. A dependable supply of representative dilution
water is necessary for conducting bioassay tests. Although test organisms
-------
134
may be native to the dilution water, the dilution water should be
characterized to determine the variation in pH, temperature, and
hardness.
Acclimation of Test Organisms. Acclimation of fish prior to
performing a bioassay should be continued for at least 2 weeks. This
period should be extended if cold weather conditions are to be studied.
Flow Rates for Continuous-Flow Studies. Flow rates should be
maintained so that at least 1.0 1/min will be provided for every 1000 g
of fish being held. This corresponds to a rate of 1.44 1/day-g of
fish. The tank volume should be at least 1.0 1/10 g of fish. The tank
volume should be equal to the total flow during 2 or 3 hr. Aeration
is not permitted within the test containers, thus the oxygen supplied
with the dilution water must be adequate to prevent a significant
lowering of the oxygen concentration, since the dissolved oxygen con-
centration should be maintained within 1.0 mg/1 of saturation.
Of the many bioassay dilution techniques, the "fail-safe" system,
which insures that wastewater flow-ceases if dilution water flow is
stopped, is typically used.
Materials of Construction for Testing Apparatus. The test tanks
and plumbing should be constructed of relatively inert materials .
such as glass, fiberglass, or unplasticized polyethylene or polypropylene.
Observation of Mortality. The method of determining the death of
a test organism must be defined. Normally, mortality is defined when
there is no respiration or other movement and no response to gentle
prodding.
-------
135
Mortality of Controls. For acceptance of bioassay on a particular
test organism, the control mortality must not be greater than 10 percent
of the control organisms.
-------
142
I III
cc
ID
LJ
o
cr
LU
Q.
O.I
95% Confidence:(cone.)
LC50=1.0(1.8-0.6)
5=1.93(3.93-0.95)
96 hr LC50
1.0%
u
I 111 J I I \ I
1.0 10.0
PERCENT WASTEWATER BY VOLUME
FIG 48 ESTIMATION OF ^yi£DIA^I LETHAL CON-
'GENERATION FOR RAW V\/ASTEVVATER
TO Fundulus heteroclitus
-------
143
10
.J
g 20
> 30
§ 40
en 50
h 60
70
80
LLJ
o
o:
LU
90
95
98
99
95% Confidence
LC50 = IOO (120-84)
S = l.42 (1.86-1.03)
96 hr LC50 = IOO%
I I L L
10 100
PERCENT WASTEWATER BY VOLUME
FIG.49.ESTIMATION OF MEDIAN LETHAL CON-
CENTRATION FOR BIOLOGICALLY TREATED
WASTEV.ATER TO Fundulus heferoclitus
-------
144
99.99
I 1 1iirn
0.01
10
100
TIME (hr)
1000
FIG.50.ESTIMATION OF MEDIAN EFFECTIVE
TIME FOR 1.0% AND 0.75% RAW
WASTEWATER TO Fundulus heteroclitus
-------
145
99.99
9S
98
95
580
g 70
§60
§50
40
,-
020
KlO
0.01
90%
ET50=260 hr
10
100
TIME (hr)
FIG. 51.ESTIMATION OF MEDIAN EFFECTIVE
TIME FOR 100% AND S0% ElOLOGiCALLT
TREATED VW\STEV\^TER TO Fun du I us
heferocljfus
-------
BASIC CONCEPTS OF
WATER QUALITY MODELING
Prepared by
F. G. Ziegler, Ph.D., P.E.
Director of Resources Management
AWARE, Inc.
Nashville, Tennessee
-------
BASIC CONCEPTS OF
WATER QUALITY MODELING
INTRODUCTION
Public Law 92-500 identifies the goal of achieving water quality
standards by July 1, 1977. To facilitate achievement of this goal, the
Act requires that each State identify those waters for which the minimum
effluent limitations required by section 301 are not stringent enough to
meet water quality standards. They must then establish the total maximum
daily load allowable to achieve the water quality standards for that
stream segment. This process, if properly implemented, should result in
the issuance of NPDES permits to municipal and industrial dischargers
consistent with the water quality goals of the Act. Most often, some
form of a mathematical model is used to determine the effects of a
discharge on water quality, and where needed, to determine the total
maximum daily load and wasteload allocations for that stream segment.
Since wasteload allocation, by definition requires waste treatment
levels more stringent than best practicable control technology (BPT)
for industries and secondary treatment for municipalities, signifi-
cant capital expenditures are necessary to implement these treatment
levels.
^..
It is only natural for those affected by the results of modeling
studies to question the predictive capabilities of these water quality
models. It is the modelers' responsibility to be sure that the
-------
model is properly developed and applied. Otherwise, wastewater
treatment requirements may be incorrectly predicted. It is the pur-
pose of this paper to briefly review the requirements for a properly
developed model and,.in doing so, to also discuss various problems
that the authors have encountered in the calibration and verifica-
tion of water quality models.
REASONS FOR USING MODELS
Before discussing the application of water quality models, it is
appropriate to consider the question of why one would want to use a model
in the first place. No doubt it is much easier to rely on intuition,
convenience, or some administrative decision rather than attempt the
difficult, if not sometimes impossible task of modeling certain situ-
ations. But without the cause-effect relationship which a model provides,
one is put in the position of making decisions on a purely arbitrary
rather than scientific basis. The very process of applying a model tends
to further one's understanding and knowledge of a particular river or
stream. The overall complex problem is broken down into many simpler,
discrete elements which are easier to understand and quantify. For
example, when modeling the dissolved oxygen concentration of a river,
one first deals with defining and quantifying such things as flow,
river velocity, water depth, temperature, waste load, rate of BOD
decay, etc. In considering each of these components in detail, one
often finds certain parameters to be more significant than originally
-------
anticipated. For example, the rate of decay of BOD may be consid-
erably less than what is normally found due to'inhibition by toxicity
or low pH. After each of the components of the model is defined...and
integrated into the overall model, it may become apparent that some
important process or effect has been overlooked. Using the dissolved
oxygen example, one might find that even after the discrete components
of the model are adequately defined, the predicted dissolved oxygen
levels do not match the observed values. Why? Possibly benthic
oxygen demand was not included as a component of the model and should
have been. In this way, the process of applying a model leads to a
better understanding of the actual, physical, chemical, and biological
processes that affect water quality.
In addition to providing a scientific basis for decision making
and developing an overall knowledge of the river or stream, modeling
permits the evaluation of many alternatives for a best solution. A
frequently overlooked application of models is their utility for
evaluating alternative sites for a proposed waste treatment plant or
industrial facility. Used in this manner, the best sites both economi-
cally and environmentally can be chosen. Potential problems can be
identified before the fact and then minimized or completely avoided.
BASIC FRAMEWORK FOR APPLYING WATER QUALITY MODELS FOR
WASTE LOAD ALLOCATION
Certain basic steps should be followed if water quality models
are to be successfully applied in the determination of waste load
allocations. These basic steps are:
-------
1. Define Water Quality Problem
2. Select Methods of-Analysis
3. Evaluate Existing Data
4. Preliminary Modeling
5. Data Development for Calibration and Verification
6. Calibration
7. Verification
8. Sensitivity Analysis
9. Model Application to Determine Allowable Waste Load
10. Monitor Water Quality and Refine Model (if needed)
Define Water Quality Problem
The logical starting point of the analysis is the definition of the
water quality problem, both present and future. All possible constituents
which relate to that problem should then be considered. For example, if
the water quality problem is low concentrations of dissolved oxygen, all
related constituents such as BOD, organic nitrogen, ammonia-nitrogen,
nitrite-nitrogen, nitrate-nitrogen, and temperature should be investigated.
Sources of these constituents, both natural and manmade must be identified.
It may turn out that some of the constituents initially considered
are not significant, and, therefore, need not be included in the
analysis. However, it is better to consider all possibilities ini-
tially arid then eliminate the insignificant constituents. Once the
significant water quality problems have been defined, one can proceed
to the selection of a method of analysis.
-------
Select Methods of Analysis
Having defined the water quality problem and related constitu-
ents, a method of analysis must be selected that is suited to the
specific problem. A great many techniques are currently available
which can be used in the determination of the assimilative capacity
of a river or stream. These techniques range in complexity from the
quick, simplified methods which consider only a few constituents, to
the very sophisticated multi-dimensional,'dynamic computer models
which attempt to consider complex processes such as eutrophication.
The degree of sophistication that one chooses should be consistent
with the complexity of the problem and the amount of risk one can
accept. The less complex the problem, the simpler the model can be.
Section 303 of PL 92-500 required the determination of the
"maximum daily load . . . necessary to implement the applicable water
quality standards" for all water quality limited streams within each
state. In response to this mandate, most states turned to the quick,
simplified modeling techniques in their basin planning and waste load
allocation efforts. Both time and resource limitations necessitated
this approach. The simplified techniques include:
1. EPA's Simplified Mathematical Modeling of Water Quality 1
2. Tennessee Valley Authority Assimilative Capacity Equations2
3. Streeter-Phelps equation using "textbook" values
-------
All three of these techniques require little or no field data
and are quick and easy to use. Unfortunately, the limitations and
applicability of these simplified techniques are quickly forgotten
once the allowable waste load is determined. It must be recognized
that the simplified modeling techniques were not meant for use in the
development of detailed waste load allocations including treatment
pjant design ef f icjencies. The use of simplified methods for these
purposes is a misapplication and subject to a great degree of
uncertainty. This very point is emphasized in the EPA publication
concerning its simplified technique. However, these techniques are
useful if applied in the proper context.
A simplified modeling technique can be an extremely valuable tool
when used as a screening device to evaluate and categorize a large
number of streams. For example, all the streams in a particular basin
could be easily evaluated using a simplified modeling technique. Based
on this "first cut" analysis, one would identify those stream segments
which require greater than the maximum treatment levels specified by
PL 92-500 to achieve water quality standards. Having identified these
"water quality limited" segments, more precise modeling analysis based
on actual field data could be used to determine the waste load alloca-
tions for that segment. Most states in EPA's Region III are now at
this point. Simplified methods have been used in identifying the "water
quality limited segments" and in determining "first cut" waste load
allocations. The next step should be to proceed to a more precise
-------
modeling analysis. There is the danger, however, that the "first cut"
nature of these allocations will be forgotten and no attempt made to
develop more precise models calibrated and verified with field data.
These more refined modeling efforts become economically justified
when relatively small differences in resultant waste treatment levels
involve large expenditures. This same reasoning should also be
applied in certain environmentally sensitive situations where a small
error in the allowable waste load could produce significant environ-
mental damage.
A good example of the problems that one might encounter in the
application of simplified modeling techniques involves the waste load
allocation for a municipal facility developed by one of the states in
EPA's Region III. The treatment plant was located on a small tributary
stream which had a 7-day, 10-yr low flow estimated to be 0.063 m /min
(0.037 cfs). The service area for this facility was located on the
drainage divide between two basins. As a result, an assessment of the
assimilative capacity of the stream was made in each of the two basin
plans prepared by different consultants for the State. One consul-
tant used the TVA "flatwater" equation to assess the assimilative
?
capacity of the stream.
The application of this equation yielded the following allowable
effluent limitations at the 7-day, 10-yr low flow:
treatment plant flow = 3,785 m /day (1.0 mgd)
effluent DO = 7.0 mg/1
BOD, =8.9 mg/1
-------
The basin plan for the adjacent river basin (prepared by a different
consultant) made use of the Streeter-Phelps equation to evaluate the
assimilative capacity of this same stream. "Textbook" values were assumed
for use in the equation. Using this technique, the following allowable
effluent limitations were determined at the 7-day, 10-yr low flow.
treatment plant flow = 3,785 m /day (1.0 mgd)
effluent DO = 7.0 mg/1
BOD5 ' = 15.0-17.0 mg/1
Another analysis of this same stream was then performed by State
personnel. The Streeter-Phelps equation was again applied using assumed
values for the variables in the equation. This analysis produced the
following waste load allocation.
treatment plant flow = 3,785 m /day (1.0 mgd)
effluent DO =7.0 mg/1
BOD, =20.0 mg/1
0
For comparison purposes, the "Simplified Mathematical Modeling
of Water Quality" was also applied. This technique indicated the
foTlowina waste load allocation:
treatment plant flow = 3,785 m3/day (1.0 mgd)
effluent DO =7.0 mg/1
approximate treatment level = advanced waste treatment
BOD5 =10.0 mg/1
TKN = 2.6 mg/1
-------
It should be pointed out that the users guide for this technique empha-
sizes that the answers obtained are approximations and by no means exact.
The results of these four simplified modeling techniques are compiled
in Table 1. All four methods require a treatment level more stringent
than secondary treatment. However, depending on the modeling analysis
chosen or the person performing the analysis, the allowable BOD5 effluent
concentration can range from 8.9 mg/1 to 20.0 mg/1 and the allowable TKN
effluent concentration can range from 2.6 mg/1 to no limitation at all.
It should be apparent that a treatment plant cannot confidently be
designed based on simplified modeling techniques such as the ones dis-
cussed.
TABLE 1
RESULTS OF VARIOUS SIMPLIFIED MODELING TECHNIQUES
Modeling Techniques
TVA Flatwater Equation
Streeter-Phelps Equation
(applied by Consultant)
Streeter-Phelps Equation
(applied by State)
EPA Simplified Modeling
Flow (m /day)
3,785
3,785
3,785
3,785
Allowable Waste
BOD5(mg/l)
8.9
15.0-17.0
20.0
10.0
Load
TKN (mg/1)
*
*
*
2.6
DO (mg/1)
7.0
7.0
7.0
7.0
*Not predictable from the model.
-------
Making the assumption that a simplified water quality modeling
analysis has been performed on all segments of the receiving streams
and that this analysis has resulted in the identification of those
segments which are water quality limited, the second phase of the
water quality and modeling analysis should be performed. During the
identification of the segments which are to be analyzed with a com-
plex water quality model, it is important to extend these segments so
that the characteristics of upstream or headwater quality can be estab-
lished. Model selection should be based on the preliminary analysis
of the water quality and experience. The model should consider the
most significant parameters affecting water quality.
The level of model sophistication should be commensurate with
the training of the water quality personnel. Models should be reviewed
to identify those which are susceptible to artificial variation of
r-
water quality as a result of modeler decisions. An example of such a
situation may occur in models whose grid size will create numerical
dispersion. Consideration should also be given to the complexity of
the phenomena influencing the water quality parameters of interest.
A greater chance for error in predicting waste load allocations may
result when models are selected which do not distinguish the intrinsic
water quality phenomena, thus preventing the modeler from dissecting
the significant influences. Nitrification is an example of such a
situation. The entire nitrogen series should be modeled. This
includes ammonia, organic, nitrite, and nitrate nitrogen concentra-
tions. Failure of nitrate nitrogen to increase downstream of a source
10
-------
does not necessarily prove that nitrification does not occur. However,
it does indicate to the modeler that further research to confirm nitri-
fication is necessary prior to an ultimate decision to limit nitrogen
discharges on the basis of reduction in dissolved oxygen.
Generally, it is appropriate to select a model which includes
more water quality parameters than can be calibrated. Those phenomena
which are either not significant or which cannot be calibrated because
of limited resources and data can be suppressed during modeling analysis,
yet will provide additional capability at a later date when data are
available. Such a model also provides additional opportunity to evalu-
ate the sensitivity of various parameters even though they cannot be
rigorously calibrated and verified. It is important for regulatory
agencies to develop a thorough understanding of specific models which
are adequate to describe the water quality phenomena occurring in their
region of the country. Most models require similar data, but in a
different format which increases the time for model development substan-
tially. By selecting a sophisticated model capable of expansion and
development as additional water quality data becomes available and
training water quality modeling personnel in the use of this specific
model, a more effective use of funds can be achieved. Obviously,
there are instances where a single model will not be adequate to
describe all the water quality and hydraulic conditions occurring with-
in a state or region.
11
-------
Evaluate Existing Data
The development of a comprehensive water quality model requires
extensive water quality sampling and analysis. The cost of data develop-
ment may often be excessive. Therefore, the first stage in a comprehen-
sive model development should be to review all existing data. This
analysis should only be performed after the specific model has been
selected and modelers familiarized with its use. Employing this sequence
permits the modeler to evaluate the water quality data in light of the
capabilities of his model, thus, permitting him to extract from the
large quantities of water quality data available those data which are
appropriate for modeling. Additionally, by analyzing the water quality
data available for particular stream segments, the modeler will develop
insight into the significant phenomena occurring and altering appro-
priate water quality parameters. Attempting to model with existing water
quality data, even though these data do not meet all the requirements
for an ideal data base for the model, will also provide insight into
the data which should be obtained during calibration and verification
sampling surveys. Generally, existing water quality data surveys have
not been devised or obtained for water quality modeling and are of
little value other than general monitoring of water quality conditions
in the receiving stream.
Preliminary Modeling
Simultaneously with the assessment of existing water quality data,
the model should be set up with regard to segmentation and hydraulics.
12
-------
A thorough and careful development of the segmentation and hydraulics
of the model may be time-consuming, but once adequately developed, per-
mits this model to be used in the future without significant alteration,
However, if a different model is selected because the original one was
not adequate to describe water quality conditions, it may be necessary
to redevelop the hydraulic input data because the second model uses a
different format. Often river constrictions, such as impoundments,
substantially alter water quality relationships. These artificial
alterations of the hydraulic regime of a river system must be iden-
tified and considered during water quality sampling and analysis as
well as further model development. Once the hydraulic data for the
model, such as flow, velocity, depth, and cross-sectional area are
developed, the sampling program for calibration and verification of
the model can be devised. By operating the model with "best estimates"
on the various model constants required as input data, some insight
will be developed into the most appropriate locations for sampling
as well as segment selection. Recognizing that model segmentation is
normally based on hydraulics, biological alterations, and points of
discharge, it is most appropriate to sample and segment boundaries.
Data Development For Calibration And Verification
Past experience has indicated that comprehensive surveys prior to
implementing the steps described above often result in relatively large
expenditures without the development of the most appropriate data for
13
-------
the model. A preliminary survey should be conducted to familiarize
the samplers with the river system as well as to practice sampling
techniques. A list of parameters is provided in Table 2 which can
be used as a guide for selection of those parameters which are most
significant for the specific water quality modeling application. By
reviewing this list of parameters and assessing the impact by sensi-
tivity analysis of the model prior to calibration, some insight into
the significance of each of them may be obtained and, thus, identify
those that should be sampled.
The costs of a field survey vary substantially depending on those
parameters to be sampled, the extent of the river system, the number
of segments and samples, complexity of sample acquisition, and the
time of travel of the river.
Typically, the first comprehensive survey will include parameters
which will prove to be insignificant during calibration and may be
ignored in subsequent surveys. Calibration is generally defined as
that phase of model development in which the numerous constants of a
model are adjusted to reproduce the water quality data obtained during
the survey. Carbonaceous and nitrogenous oxygen demands are generally
considered independent variables and are fitted first. If dissolved
levels drop below 1 or 2 mg/1, carbonaceous and nitrogenous BOD's are
no longer independent of dissolved oxygen concentration. Algal concen-
trations may also significantly alter the independent relationships of
14
-------
TABLE Z
LIST OF PARAMETERS SAMPLED IN WATER QUALITY
ANALYSIS STUDIES
These parameters should be analyzed for headwaters, tributaries, dis-
charges, and at all river segments.
Time of Travel
Flow
Temperature
Dissolved Oxygen
Short- and Long-Term BOD's
Suppressed and Non-Suppressed
BOD's
Total Nitrogen Series (TKN, ammonia
nitrogen, nitrite nitrogen, nitrate
nitrogen)
Nitrifier Concentrations
Temperature Effect on Decay Rates
Benthic Demand
Temperature Effect on Benthic Demand
Chlorophyll or Algae Concen-
trations
Diurnal Variations in Dis-
solved Oxygen
Phosphorus
Chlorides
Other Conservative Elements
and Toxic Substances
Tidal and Sieche Cycles
Solar Insulation
Temperature
Turbidity
Suspended Solids
Operation of Dams and Power
Stations
15
-------
the carbonaceous and nitrogenous materials. Some reassessment of
"independent variables" after establishment of a dependent (dis-
solved oxygen.) may be required for model refinement.
Water quality data developed for both calibration and verifica-
tion of the dissolved oxygen portion of a model should be obtained
during warm weather, low flow conditions so that the data are similar
to conditions used for waste load allocation. Model calibration may
also identify additional parameters which either need to be sampled
or need further amplification during the second survey. As an example,
it may be easily demonstrated that nitrification is not occurring in
the surface waters, but a substantial benthic oxygen demand occurs
partially as the result of nitrification. Therefore, a second survey
may concentrate on nitrifier concentrations in the benthos as well as
the impact of the benthos on the dissolved oxygen concentration in the
river system. If the results of the calibration survey indicate that
a substantial alteration of the survey program is required prior to
verification of the model, it will be necessary to conduct a third
survey to confirm the model's predictive abilities.
Calibration and Verification
Fitting a natural phenomenon to .a. mathematical relationship is
difficult. Model calibration requires some license on the part of the
modeler to make assumptions with regard to the adequacy of the water
quality data and the adequacy of the analytical techniques. Often this
16
-------
license is exceeded as the result of inadequate data evaluation.
As an example, BOD data are often pronounced as inadequate because
of extensive scatter. The assumption is made that the scatter is
the result of the analytical accuracy of a testing program. Often
this is only partially true. The scatter is often the result of
failure to maintain a steady-state sampling program or errors in
sampling compositing. Alteration of time of travel as a result of
regulation of downstream reservoirs frequently produces misinter-
pretations of data.
All sampling techniques and assumptions should be tested prior
to the start of a comprehensive survey. It is often very enlight-
ening to have each sampling crew make a traverse of a river to deter-
mine the average dissolved oxygen concentration. Frequently, this
evaluation will result in dissolved oxygen concentrations varying as
much as 1 mg/1. Sampling techniques can then be modified to obtain
dissolved oxygen values most representative of the river.
The second comprehensive survey often used for verification of
models should be performed under a different set of low flow, warm
temperature conditions assuming these are the conditions which will be
identified as critical for waste load allocations. It should also be
emphasized that sampling techniques are also dependent upon the use
of either steady-state or dynamic models. Actually, a single sampling
survey conducted over a time period which is longer than the time of
travel will permit adequate data for dynamic analysis and result in
only a single mobilization cost.
17
-------
Verification is defined as that period in which a second set of
discharge and boundary data are applied to a model and water quality
predicted which can then be compared with actual measured data. If
this can be accomplished without modification of the constants deter-
mined during the calibration phase of the model, it is reasonable to
believe that the model is capable of reproducing water quality condi-
tions.
During the development .of the water quality data for calibration
and verification, it is important to obtain comparable data for tribu-
taries, discharges, and water quality. The assumption that accurate
water quality data for the river and its tributaries can be combined
with permit discharge numbers to calibrate and verify a model is incor-
rect. Correct assessment of discharge data is mandatory in any accurate
calibration and verification of a model. Certainly, this requires
extensive cooperation between industrial and municipal dischargers and
water quality modelers and sampling teams in order to obtain a compre-
hensive and accurate data base. The assumption of discharge loads which
are lower than those actually discharged to the river system can work
against the discharger if this information is applied to the calibra-
tion portion of the model. Conversely, the assumption by regulatory
agencies of discharges greater than those actually being discharged may
result in prediction of waste assimilative capacity greater than actually
occurring in the river. Additionally, both upsets in treatment facilities
and discharges of toxics or inhibitory substances may cause substantial
18
-------
alteration in water quality conditions. These occurrences must be
accounted for if a reasonable assessment of the water quality data
and, thus, an accurate calibration of the model is to occur.
CASE STUDIES
Inappropriate Interpretation of Time-of-Travel Data. In 1975 an
extensive study was performed on the river system in which the chloride
concentrations served as a useful tracer. Chloride measurements were
made at numerous locations within the river system, its tributaries,
and all point source discharges. A dynamic version of the QUAL-II
Model was used to predict the chloride concentrations within the river.
Results of the prediction on the second and fifth days of the simula-
tion are presented in Figures 1 and 2. The results demonstrated
excellent agreement with measured data. Therefore, it was concluded
that the model accurately predicted the hydraulics and dispersion with-
in the river system. Based on this conclusion, the model was then
calibrated for carbonaceous and nitrogenous BOD and dissolved oxygen
concentrations. Once the model had been calibrated, it was compared to
the accumulation of all the water quality data which were obtained during
the period of the investigation. The results of this comparison are
presented in Figure 3.
The measured data in Figure 3 are presented in a form which would
have been used if these data were processed for calibration of a
model. The dynamically calibrated model does not adequately predict
19
-------
1.00
'.90
.80
.70
o. .60
UJ ,-*
Q .50
en
3
O
40
.30
.20
.10
.00
73.70 66.34 58.98 51.62 44.26 36.90 29.54 22.18 14.32 740 .10
RIVER MILE TO HEAD OF REACH
FIG. I. DYNAMIC MODEL SIMULATION OF CHLORIDE
8-14-75
-------
1.00
.90
.80
.70
^ .60
S
uj -50-
§ 40
_i
IE
.30
.20
.10
.00
T
A Measured Data
Calculated Data
70
60
10
0
50 40 30 20
RIVER MILE TO HEAD OF REACH
FIG. 2. DYNAMIC MODEL SIMULATION OF CHLORIDE
8-17-75
-------
ro
ro
10.00
9.00
8.00
7.00
6.00
5.00
4.00
3.00
2.00
1.00
.00
A. Measured Data
Calculated Data
70
FIG. 3.
60 50 40 30 20
RIVER MILE TO HEAD OF REACH
10
0
COMPARISON OF STEADY STATE SIMULATION
OF EPA MODEL TO MEASURED DISSOLVED
OXYGEN DATA-8-13-17-75
-------
the measured data. Obviously, this sampling period cannot be used
as an average to simulate steady-state conditions.
Diurnal Variation. An important aspect in the calibration and verifi-
cation of dissolved oxygen models is the effect of photosynthesis and
respiration. The importance of the photosynthetic activity of chlorophyll -
containing plants on the oxygen balance of streams has long been recognized.
A brief historical review of this subject has been presented by O'Connor
and DiToro . Nonetheless, this process has frequently been ignored in
the analysis of small, free-flowing streams. There has been the tendency
to associate photosynthetic effects primarily with phytoplankton which
generally are not found in sufficient numbers to create a noticeable impact
on small, free-flowing streams. Numerous studies4'5'6'7, however, have
pointed to the fact that there are large diurnal dissolved oxygen fluctu-
ations occurring in small, free-flowing streams due to periphyton and/or
macrophytes. The relatively high surface area to water volume ratio in
these small streams acts to magnify the effects of photosynthetic activity
on the dissolved oxygen balance.
The South River, near Waynesboro, Virginia, provides a good example
of the diurnal dissolved oxygen variations that one might find on a
relatively small stream. The South River has a drainage area of
23
-------
2 2
approximately 549 km (212 mi ). During lower flow periods of the summer
and fall, the average river depth varies from 0.06 m (0.2 ft) to 1.47 m
(4.5 ft), with a mean of about 0.61 m (2.0 ft). There are significant
contributions of ammonia-nitrogen, nitrate-nitrogen, and phosphorus
from point sources located at Waynesboro, Virginia, between river mile
24.2 and river mile 22.5. Attached algae and macrophytes are the dominant
aquatic plants with little or no phytoplankton being observed, as indi-
cated by chlorophyll 'a' measurements of less than 10.0 ug/1. As a
result of these conditions, large diurnal fluctuations in dissolved oxygen
can be observed. The magnitude of the diurnal dissolved oxygen fluctua-
tion that was observed by EPA in a survey of the South River is pre-
sented in Figure 4. With a dissolved oxygen fluctuation of approxi-
mately 5 mg/1, it is clear that photosynthesis and respiration of attached
algae and macrophytes play a significant role in the dissolved oxygen
balance of the river. Failure to consider these effects will result in
an incorrect analysis. Consider, for instance, what would happen if data
for calibration of a dissolved oxygen model were collected with no consid-
eration of the diurnal effects. A sampling run made in the morning would
yield a substantially different dissolved oxygen profile than one made
later in the day.
The results of two EPA dissolved oxygen sampling runs at differ-
ent times of the day in relation to the maximum, minimum, and daily
averages observed are presented in Figure 5. As expected, the
24
-------
ro
LU
X
o
Ct
LJ
O
CO
b
1.0
10.0
9.0
8.0
7.O
6.0
i i
i r
SOUTH RIVER, VA
JULY 7-8, 1976
RIVER MILE 21.3
J 1 1 1 I ' ' ' ' I I I I
0400 OGOO O&OO 1000 1200 WOO 1600 1800 2000 2200 2400 0200 0400 0600
FIG.4. DIURNAL VARIATION OF DISSOLVED OXYGEN
-------
U
o
>-
X
o
Q
UJ
O
CO
CO
Q
12.0
11.0
10.0
9.0
8.0
7.0
6.0
5.0
4.0
i
-
X
,' (
/
T
' I
-k
X
X
X
X
X
^
1 1
SOUTH RIVER, VA
JULY 7. 1976
v,
"X
) ^-,
(
-
X _
X
) vx
«^
C
>*
^fc
-p ~
_ _ _ - 1030-1240
- ~~ _ _
6
>
_JL -0710- 0835
. ""^
-
1 1 1
2.5.0
20.0
15.0
10.0
RIVER MILE
FIG.5. DISSOLVED OXY3EN PROFILES OF THE SOUTH RIVER
-------
dissolved oxygen profile observed during the 0710-0835 sampling run was
much lower than the 1030-1240 sampling run. Neither of these sampling
runs represent average daily conditions. If either of these sampling
runs were used to calibrate a steady-state, dissolved oxygen model,
the resulting model would not predict average daily conditions as it
should. In order to avoid this problem, one should first define
whether photosynthesis and respiration is significant. If it is
indicated that the impact is significant, then intensive diurnal dis-
solved oxygen sampling should be included in all future water quality
surveys to be used for model calibration and verification.
The results obtained from a modeling analysis that includes a
component for the photosynthesis-respiration effect will differ consid-
erably from an analysis that does not explicitly include this effect.
For example, a detailed, intensive sampling program may adequately
define the average daily dissolved oxygen concentration as well as the
diurnal variations. If the dissolved oxygen model that is applied to
the stream does not include a component that specifically accounts for
net additions of dissolved oxygen due to photosynthesis, then this
effect will be implicitly included by another model calibration param-
eter, the reaeration coefficient. In this fashion, it is possible to
calibrate a model to produce results which match the observed average
dissolved oxygen stream concentrations. It may also be possible to
obtain a reasonable verification with a different set of data. However,
this should be difficult to do because the reaeration process is
27
-------
distinctly different from the photosynthetic processes and each will
respond differently to the changed conditions (stream flow, temperature,
DO deficit, etc.)
A steady-state dissolved oxygen model which did not specifically
include algal effects was calibrated and verified using seven indepen-
n
dent sets of data for the Jackson River, Virginia. Another steady-
state model which did incliude a component for algal effects was also
calibrated using this same data. Although both models appeared to
simulate the calibration conditions rather well, there was quite a
difference in the answers obtained when the models were applied to
determine an allowable waste load. Preliminary results showed that
without the specific inclusion of algal effects the allowable waste
load was 3,423 kg/day (7,540 Ib/day) BODg. With the inclusion of
algal effects the allowable waste load was 4,168 kg/day (9,180 Ib/day)
BODg. The theoretically more correct model specifically includes algal
effects and proved the more conservative in waste load allocation deter-
minations.
MODEL APPLICATION TO DETERMINE ALLOWABLE WASTE LOAD
Perhaps the most controversial issue confronting water quality
modeling is the establishment of the water quality conditions which
will occur for waste load allocations. It is beyond the scope of this
paper to discuss in detail the possible critical water quality condi-
tions. However, it is appropriate to identify those decisions which
23
-------
are frequently the most significant and often incorrectly developed as
a result of water quality assessments. Once the modeler has calibrated
and verified the model, he must modify it to describe those future condi-
tions under which the waste load allocations will occur. Many issues
arise when one projects future conditions. These are summarized in
Table 3.
A slight error in the establishment of low flow conditions may
substantially alter waste load allocations and, thus, create an
unreasonable hardship on the discharger. Thorough review of the means
by which the minimum flow conditions were determined should be per-
formed for each waste load allocation study.
A rigorous analysis of temperature should also be performed to
determine the river temperature occurring during this low flow, warm
weather condition. Often, the low flow and warm temperatures which
will be used for waste load allocation are independently determined
and then merely combined in order to project waste load allocations.
This can result in an unreasonable hardship on dischargers. A more
reasonable analysis would be to statistically determine an acceptable
waste load allocation based on the analysis of an extended period of
associated flows and temperatures. In any event, care is necessary in
selecting a temperature and flow condition to establish waste load allo-
cations. Often, within the United States, the low flow conditions occur
in September and October which are not the warmest months.
29
-------
TABLE 3
FUTURE CONDITIONS FOR WASTE LOAD ALLOCATIONS
River Low Flow Conditions
River Temperature
Alteration in Decay Rates
Alteration in Benthic Demand
Influence of Eutrophication
Alteration in Relationship
Between Carbonaceous and
Nitrogenous Oxygen Demands
30
-------
Carbonaceous and nitrogenous decay rates are often misinterpreted.
Conceptually, an alteration in decay rate will occur as treatment
levels improve and, therefore, biologically less reactive compounds
are discharged to river systems. As these decay rates are altered, the
waste load allocations will also change. It is difficult or impossible
to predict the change in the decay rate assuming the calibration and
verification of the river are done under conditions substantially differ-
ent than those which will occur when water quality is achieved. As an
example, in the river system in which raw wastewaters are being dis-
charged, decay rates established by calibration of the model may be
substantially higher than those predicted if secondary treatment has
been applied to these wastewater discharges. Therefore, projection of
waste load allocations based on a raw wastewater discharge model cali-
bration may predict lower treated discharge loads than actually possible.
On the other hand, calibration of a model which is influenced by the
discharge of inhibitory substances may overpredict waste load alloca-
tions recognizing that these inhibitory substances will most likely be
removed during treatment processes. Improvement in water quality condi-
tions may also stimulate nitrification in benthic deposits which had
been anaerobic prior to the improvement in wastewater quality.
Eutrophication of sluggish systems may increase the diurnal cycle
of dissolved oxygen and perhaps alter the relationship between minimum
and average dissolved oxygen concentrations, thus altering the waste
load allocations for the river system. This phenomenon is depicted in
Figure 6.
31
-------
00
ro
^aiFuture Cycle
X
o
o
LJ
O
to
CO
Q
Average Dissolved Oxygen Standard
at Future Time
\ \ Average Dissolved Oxygen
Standard at Time Zero
/<
-------
From this discussion, it is obvious that in the river systems
in which a substantial improvement in water quality is anticipated,
a step approach to the problem is necessary in. order to adequately
manage the environment and yet not require treatment levels which
are either substantially less than or greater than those actually
needed to maintain water quality standards.
Benthic demands are frequently altered as a result of the improve-
ment in water quality conditions in the river system. The reduction in
benthic demand as a result of this improvment which should occur in the
step between calibration and waste load allocations in modeling may
result in a substantial additional oxygen resource available for assimi-
lation of wastewaters once volatile solids have been removed from the
discharges.
MONITOR WATER QUALITY AND REFINE MODEL
Once the waste load allocations have been determined and achieved
through the installation of the appropriate treatment facilities, it
is always desirable to conduct a water quality sampling survey to update
and refine the model. Various changes to the river system can and fre-
quently do occur as previously mentioned. The water quality sampling
survey will indicate the various changes that have occurred and the
model can be updated accordingly. An updated, scientific tool based
on the best information available will be the result of this approach.
This fits nicely with the concept that planning is not a static, but
rather a dynamic process being continuously updated and refined.
33
-------
REFERENCES
1. Environmental Protection Agency, "Simplified Modeling of Water
Quality," March, 1971.
2. Ruane, R. J., "Statistical Equation for Estimating the Assimilative
Capacity of a Stream for BOD," Tennessee Valley Authority,
Water Quality Branch (Unpublished Report).
3. O'Connor, D. J. and 0. M. DiToro, "Photosynthesis and Oxygen Balance
in Streams," Journal of the Sanitary Engineering Division,
ASCE, April, 1970.
4. Hydroscience, Inc., "Water Quality Analysis of the Jackson River,"
June, 1976.
5. Streeter, H. W. and E. B. Phelps, "A Study of the Pollution and Natural
Purification of the Ohio River," U.S. Public Health Service,
Public Health Bulletin No. 146, 1925.
6. Cahill, T. H., et. al., "A Math Model of Dissolved Oxygen for the
Brandywine Basin," Tri-County Conservancy Technical Publication,
No. 4, 1975.
7. Cairns, J. and K. L. Dickson, "An Ecosystematic Study of the South
River, Virginia," Bulletin 54, Water Resources Research Center,
Virginia Polytechnic Institute and State University, July, 1972.
34
-------
SPILL RESPONSE*
Although it is not the purpose of this manual to provide a
treatise on spill management or spill control, it is important to
relate those aspects of spill management that influence directly
modeling as discussed in this handbook. Additionally, much of
the information available through modeling is directly related to
spill management, and therefore, considered herein. Spill manage-
ment as considered in this manual includes:
a. The effect on calibration and verification data of
inadvertent spills occurring during the time of the
field surveys.
b. The use of data obtained for calibration and verifi-
cation of dissolved oxygen deficit models to assist
in the management of spills.
When water quality surveys are designed for use in calibra-
tion and verification models, consideration should be given to po-
tential spills which could invalidate the water quality survey. It
is not always possible to identify spills through monitorinc of
outfalls during the water quality surveys because of the time
cycles between samples. Additionally, because of longitudinal dis-
persion, toxic or inhibitory spills may alter the decay rates even
%
Excerpted from a Modelling Manual prepared for the State of Virginia
by F. 6. Ziegler, et. al.
III-l
-------
though they occurred prior to or after the field survey commenced.
The potential of such an occurrence must be judged relative to the
complexity of the specific water system. If potential for a spill is
high within a particular water system, excessive monitoring of the
discharge should occur in order to establish the variability of the
wastes during a period of time in which field surveys are being con-
ducted. As previously indicated, spill prevention is not a subject
of this manual. However, spill management once a contaminant has
entered a receiving stream is directly related to this manual, and
therefore, considered herein. Much of the data obtained during acqui-
sition of data for verification of dissolved oxygen deficit models is
directly apolicable to assist the modeler in managing spills.
Until recently the primary emphasis on spill prevention has been
the discharge of oils into receiving waters. The objectionable aesthe-
tics of such a discharge resulted in severe public criticism and,
therefore, generated regulations by various authorities to control anc
prevent further abuses. Within the last few years, however, the smcna-
sis on toxics and the need for determining the magnitude of the problem
of toxics in the environment has generated extensive monitoring. The
increased emphasis on drinking water quality has resulted in exten-
sive monitoring of drinking water intakes and thus the realization of
the possible contamination of water supplies by discharges. Recognizing
III-2
-------
the possibility of inadvertent spills to receiving watars, which must
then be managed until they have been adequately diluted or degraded
to the point of being no longer significant from the standpoint of
human health or biological survival, has become a critical concern
of most dischargers and officials responsible for providing potable
water.
Although the responsible industrial community recognizes the need
for prevention of any discharge to receiving watars, the potential for
the incident must be recognized. Therefore, it behooves those respon-
sible for discharges or for the maintenance of water supplies to have
access to readily available information which will provide them with
the knowledge to determine the impacts and influences of an accidental
spill.
Spills may occur from many sources: (1) through commerce (barges),
(2) from roadways (accidents), (2) point source discharges, and (4)
rainfall runoff events. This section will present techniques which can
be usad to enable a discharger or a potable watar supplier to evaluate
potential spill impact, using available data. Additionally, these same
techniques permit the environmental agency to identify those segments
of the receiving watars which may receive damage to the biological com-
munity as a result of a spill. The step-by-step development of these
data and the preparation of the methodology to predict the impact in-
cludes but is not limited to:-
1. evaluation of the magnitude and characteristics of a spill ;
III-3
-------
2. determination of time cf travel through steady-state and
modified techniques considering the hydraulic variability
or the river systems-
3. evaluation of possible chemical reactions which may occur
within the receiving streams.;
4. identification of the impact on the biological corn?,unity;
5. analysis of the influence and characteristics of watar
treatment facilities which may alter tine contanrinantV
5. identification of the characteristics of the water cr==^-
ment_facilities which must be known in order to determine
the impact in the community if inadvertently csntamina^^
water is drawn into the system.
Such investigations and acquisitions of the appropriate informa-
tion to provide spill management have been prepared for industries
and municipal water users.
Sefore discussing the location of spills, it is appropriate to
consider the various types of spills which may occur.
1. Accidental Spill - Such an occurrence is normalIv consi-
dered the most frequent class of ssill because it*recr*-
sents an occurrence completely beyond the control of' ^he
individuals associated with the operations contributing
the waste to the receiving waters. A spill cf this natur-
T^S frequently the result of ships! collisions, ruptured
.uei :anks, or on a more limited scale, sloppy operations
of the maintenance around loading off-lcadinc of'ships.
It may also occur by discharges to storm sewers, which'are
mere dirricult to trace. As will be discussed la^=" th»
ability to trace a spill, and thereby identify its~quali-
ty and quantity, is most critical to any accurate sredic-
tions of a significant on the receiving" waters and the
aquatic environment. If the source of'accidental spills
can be determined, it is frequently possible to detarmine
time, quantity, and characteristics of the material discharged.
2. Process Start Up - An operation .which often creates a spin
is the start-up of certain industrial processes which may
generate large quantities of unusable product or by-products.
III-4
-------
Process equipment which has not been tastsd under opera-
ting conditions may fail creating inadvertant discharges.
There-ore, it is most appropriate for regulatory person-
nel to identify start-up periods for industries*who have
potential toxic or hazardous wastes which may inadvertently
be^discharged to receiving streams. Obviously, all aporo-
pnate safeguards should be taken but it must be recognized
thai the potential for inadvertant discharge is higher for"
start-up than during normal operating conditions.
Process Shut-Down - Although it is often possible to have
better control of chemical and physical processes during
shut down of an operating facility, the cleaning and main-
tenance associated with such an operation may causa upsets
in^treat^ent facilities as well as genera-ion of high'Quan-
tities or waste cleaning fluids which require disposal Ac-
cidental discharge or wastewater treatment upsets'nav result
in spilis from these operations.
Sustained Discharge - It is csrmonly accepted that a sus-
tained discharge from a process does not constitute a
spill because the Quantities of discharge remain constant.
However, under exzreme variations in river flow, an acc*p-
table concentration of waste products may become unaccep-
rably high as a result of reduction in dilution. Such
phenomena are common in air pollution considerations but
also occur routinely in wastewater and in water quality
problems. At least three situations might contribute to
such a phenomenon. These are:
a. regulated streams in which extremely low flows
may occur as a result cf dam regulation;
b. warm weather low flow conditions. These conditions
are not conducive to measurement of concentrations
in receiving streams which appear as though SBIMS
occur. Nevertheless, extremely lew flow conditions
may result in concentrations of pollutants being
higher than what might normally be anticipated;"
c. winter low flow conditions. Often winter low
flow conditions occur as a result of severe icino
which substantially reduces river flow. Rapid
thaws may also cause very substantial variations
in flow thus causing waste discharges to bs diluted;
III-5
-------
d. hydraulic variations. Resuspension of sattlable
solids as a result of increased hydraulic load
and thus axcsedence of scour velocity may cause
increased concentrations of certain pollutants.
Additionally, significant turbulence within
estuaries_may also result in suspension of anae-
robic sediments causina suppression of dissolved
oxygen;
s. rainfaU runoff. The discharge of pollutants carried
by rainfall runoff from surface contamination to
receiving streams may often appear as a spill or
under certain circumstances may actually cause a
spill through excessive rainfall causing seepage
or overtopping of dams containing wasta'products.
Additionally, rainfall seepage may increase leachate
contamination from landfills"or slude dr/ beds,
ultimately resulting in discharges to receivina
streams.
Obviously, it should be the objective of all regulatory personnel
as well as dischargers to receiving streams to prevent any spills
from any of the above or other mechanisms. However, often it is impos-
sible to prevent spills from occurring. Once a spill occurs, it is
necessary to control the use of waters and to protect the public and
wherever possible, aquatic and terrestrial environment as much as pos-
sible until the spill has been diluted to safe levels. To orotaci all
users of the water resource, it is essential that the water quality
modeler be able to predict the location and approximate concentration of
tne waste so that effective recommendations may be made to minimize
damage. To have any success with predicting the location of a spill,
extensive information must be available and processed prior to the inci-
dent. Developing information on the day of the spill is, in most cases,
useless and can result in costly errors due to misinterpretation of the
III-6
-------
data;, nevertheless, ft should be recognized that, for many streams,
Information presently exists which can be processed into a form useful
in the event of a spill. In the areas where discharge records are
not available, time of travel measurements can be made during water
quality studies which will provide useful and accurate information
in predicting time of travel transport of spills if they should occur.
Therefore, on streams which have a high potential for spills or which
are characterized by extremely sensitive environmental conditions or
heavy water supply usage, it is important to have adequate infornaticn
with regard to the hydraulics and in particular time and travel within
the system. It should be recognized that knowing the location of a
spill is often much more valuable than being able to predict specific
concentrations. Often measurement of high concentrations is possible
once the location of the spill is known to tie field survey crews. When
spilled materials have either a characteristic color or are immiscible,
idenfincaticn of the location of the spill is relatively easy. However,
often spills are diluted to a point that they are not recognizable
or are colorless and soluble, preventing easy recognition. Under such
conditions, a careful calculation of the location based en time-of- tra-
vel analysis is necessary in order to make appropriate measurement in
regard to the concentration and thus the significance of the spill.
Additionally, it is often impossible to predict the concentration
of the spill because no information is available with regard to the
actual amount of material discharged.
III-7
-------
INFORMATION REQUIREMENTS
Waste characteristics - To completely describe a spill, exten-
sive information must be available with regard to the characteristics,
quantities, and quality of the wastes inadvertantly discharged. Such
parameters as toxicity (acute and chronic) density, solubility,
potential synergisms, time of the occurrence, biodegradability,
adsorption, and volatility must be established for the wastes if one
is to adequately predict its location and its effects on aquatic and
terrestrial environments.
Iu is often impossible to have previously established characteris-
tics of each potential pollutant which may be inadvertantly discharged
in the receiving system. However, recent data retrival systems available
through the Environmental Protection Agency and other service orcanira-
tions can provide a relatively rapid identification of the potential
hazard of particular pollutants.
Specific details associated with a spill are also necessary in
order to accurately predict quantities of waste which may have entered
the receiving waters. As an example, spills of a highly volatile sol-
vent may not actually reach the receiving stream as a result of evaco-
ration to the atmosphere before there are contacts to the water system
occurs. Therefore, a relatively large spill on a surface may result
in only limited quantities discharged through stem sewers to receiving
streams. To adequately predict the location of the pollutant using
time of travel analyses, it is absolutely necessary to have a good
idea of the time of occurrence of the spill. Additionally, duration
III-8
-------
of the spill is critical in order to establish the dosage or
exposure of specified concentrations to aquatic organisms.
Receptive characteristics - In order to effectively manage a
drainage basin and have accurate information available if a spill occurs,
it is appropriate to have established all points within the river system
which are sensitive with regard to spills. These key locations include
water treatment plant intakes and breeding grounds for aquatic organisms.
Additional information with regard to water treatment systems include
modes of operation, the detention time of water treatment systems and
water distribution systems, the ease or difficulty in purging such
water systems. General information in regard to additional techniques
which may be employed on an emergency basis to treat water systems to
prevent contamination of drinking water supply or tc remove contaminates
during water treatment are most important. Such basic knowledge as
boiling to remove volatile compounds may be extremely useful.
Transport characteristics - It is impossible to have too much
information with regard to transport characteristics within - specific
stream. Sefore beginning the discussion of transport characteristics,
it should be emphasized that a little informa-ion in such instances is
often much more dangerous than no information at all. Recorrmendina
corrective measures or cessation of drafting of water from a water supply
because of errors in calculation of the time and/or magnitude associated
with potentially hazardous spill events, often causes more damage than
.taking no. action at all. Information necessary to make accurate predic-
tions falls in the following categories.
III-9
-------
River flows before, during, and after soills. Immediate
availability of up-to-da.ta flow data is most important in the ultimata
prediction of time of travel. In addition to knowledge of the flows
occurring in the immediata vicinity of a spill, it is most important
to have the knowledge of the type of low measurement being used to
predict these numbers. As an example, slope gauges, commonly used on
impounded river systanis ars often of limitad value or useless at low
flow conditions. Therefore, if data is obtained from such gauges,
time of travel based on their measurement may be inaccurate. Other
types of gauges such as acoustical gauges can provide more accurate
information if they are adequately calibrated.
Time of Travel - Time of travel means the time for a particular
particle of water to move from one location to another within a river
system. Detailed descriptions of the precadurss used in calculation
of time of travel will be discussed below.
Tributary flow - To adequately predict concentration of waste
within a receiving stream, it is important to have adequate information
with regard to tributary flow. Often tributaries are not adequately
gauged as are main river systams, therefore, calculation of the area
of the tributary basin and the potential quantity of flow based on a
unit area is often necessary in order to predict tributary contribution.
Additionally, estimates of tributary now based on a proportion of
che main stam flow may also be used to oreduct quantities. If
tributary flew is relatively small, compared to main stam flow,
111-10
-------
than 10 percent) then errors associated with such flows are
often not significant.
Longitudinal dispersion - If it is anticipated that a spill will
affect a river for a great distance then longitudinal dispersion or the
mixing ahead and behind the centroid of the spill may become significant.
Dispersion forward of the initial spill is often limited and the dis-
persion after the centroid is often extended as a result of dead zones
and eddies which cause portions of the spills to delay in their trans-
port downstream. The knowledge of dispersion is most significant
when dosage or exposure of aquatic organisms to the contaminant must be
determined. Additionally, dispersion is most significant when cessation
of drafting of drinking water from river systems is necessary in order
tc prevent contamination of water supplies.
Lateral/Vertical dispersion - If the primary area of concern cf a
spill is only for a short distance downstream, then lateral/vertical
dispersion may become significant. Normally discharges occur at one
bank or the other. In such situations a delay occurs before the inad-
vertant discharge is completely mixed within the system thus contami-
nating the area from one bank to the other. This concept also applies
to the vertical dispersion and mixing of a pollutant. Therefore, if
water intakes exist on the opposite bank at approximately the same
river mile as the spill, they may not be contaminated.
Time of travel prediction - Certain aspects of time of travel pre-
diction and other hydraulics are presented in other phases of this
III-ll
-------
document. However, for the purposes of writing accurate measurement
of spill management the information is reproduced for the convenience
of the reader. In order to develop hydraulic data useful in the pre-
diction of spills through time of travel calculations it is most
efficient to acquire all existing information available from the U. S.
Geological Survey, Corps of Engineers, state agencies, etc. mis
inronnation is normally characterized as stage discharge records which
are obtained by measuring velocities, depths, and widths of river sys-
tems at a specific flow condition. From these data, the stage and
corresponding discharge can be plotted on a semi-log paper. This plot
is frequently a straight line, which will permit the calculation of
flows within a river system based on the measurement of the depth of
water at a particular monitoring station. Additionally, the same
stage discharge records can be used to predict the semi-log relation-
ship between velocity and flow. From these data it is possible to
predict time of travel within a river system. Relationships betv/een
flow, depth and velocity are as follows:
v - aQb
D = a'Qb'
There are certain limitations inherent in using these data. The
limitations are primarily asscicated with the recognition that stage
discharge records are ootained at a specific point within a river system
and might not be representative of the rsach of river over which the
spill occurs. Therefore, stage discharge records at several locations
111-12
-------
within a rivsr system are often necessary in order to draw an accurate
prediction of the relationship between velocity and flow. A more
accurate prediction of this relationship can be obtained by performing
time of travel studies employing a discharge of tracer to the receiving
stream which can be monitored on its way downstream. If these tests
are performed under numerous flow conditions, a more accurate predic-
tion of time of travel or average velocity within a river segment for a
specific flow can be obtained. Once that relationship is obtained, based
upon time of travel analysis, figures can be developed which permit the
integration of the overall time of travel within a river system as
shewn in in Figure 2.F-S. 3y use of the Information presented in
Figure 2.F-5 or other relationships of time of travel, it is possible
to predict accurately the location of a wasta spill. Assuming this
information has been developed into easily interpreted figures or
nomographs prior to the incident. Other examples of the prediction
of time of travel based on dye study are presented in Figure 2.F-5. This
figure describes the time of travel in days from a specific Ice-ion
(river miles) to any particular river mile downstream for any identi-
fied flew. The alteration in the slope of the curve at specific loca-
tions is a result of man-made obstructions within the river system.
It can be seen that information such as this readily available for
a specific industrial complex within a river system can be most useful
and readily available in the time of emergency.
111-13
-------
-o
c
o
o
Ol
LO
u
O)
Q,
O)
4-
l/) I '
It- .1
u
o
o
.(III
.01
0.0000 U <>
1
NT
I I
n
I
"\ "\
o
1 <^ >4_
I I I I 1 I I I I III |
MVI'CAStuJy Itculic
*p >» o\ «o v
f «A u,
-------
7(1
1.0
50
141
tn
30
'JO
10
0
10
70
HO
'M 30 -Kl 50 f.O
KIVI.K Mil I
FIQ. 2.F 0. EFFECT OF FLOW ON TIME OF THAVEL IN AN IMPOUNDED RIVER
-------
Significant tributaries can substantially altar the time of travel
and thus create significant errors with regard to prediction to time of
travel unless they are accurately included in such predictions. Figure
2.F-7 is a presentation of a method of calculation of time of travel
when the ratio of the tributary flow to that of the main stem is as
identified. The ratio is based upon the flows within the two systems.
The prediction is of time of travel to two specific locations down-
stream of an industrial complex. In addition to the information with
regard to tirne of travel, it is also important to have basic calcu-
lations of "concentrations within the receiving stream for various flows
and various potential discharges. Such information is obtained
through bisic calculations and mixing theory and is presented in Figure
' C_0
-.i-G.
Using stage discharge records instead of dye studies, reasonably
accurate prediction of time of travel can be made. These predictions
permit one to calculate the centroid of the spill but not estimate ac-
curately dispersion either laterally or longitudinally. 3ecause dis-
persion is not calculated accurately it is not passible to predict
actual concentrations associated with such a spill occurrence. There-
fore, such procedures do net accurately predict concentrations and
may result in predictions of concentrations of factor of two or rare
different than the actual measured values.
111-16
-------
5.COO
Ui
U.
O
2,000
1,COO
500
200
100
50
20
10
CITY 3
CITY A
5 10 20
RIVER 2 FLOW = Q (en x iO'3)
FIG. 2.F-7. TIME OF TRAVEL
50
100
111-17
-------
i.non
I
00
ino -
O
i
t 20 -
U-l
o
z
o
CJ
10 -
1
\
10 20 50 |()() 200 500 1,000 2.000 5.0110 10.000 20,0(Ml Stt.UOO 100.000
WASTE LOADING. Ib/day
FIG. 2.FO. RELATIONSHIP 8ETWEF DISCHARGE AND RIVER
C( .EN ITU .. FOf. .. .DIC...JDI ^
-------
SUMMARY
Hydraulics are the basis for all water quality and spill manage-
ment. Without an accurate and carsful analysis of hydraulics, no
model can be considered valid with regard to projection of waste load
allocations or identification and location of spills. As turbulance is
directly related to reaeration as well as stimulation of organic demands
through biochemical decomposition, all modeling efforts should be able
to demcnstrata accurate development of hydraulics. It should be recog-
nized that once the hydraulic relationships ara developed, they are not
altered by varying composition of waste discharges or other chemical
and biological parameters. It should also be recognized by the modeler
that many receiving bodies of water have been investigated by the U. S.
Geological Survey or Coast and Geodetic Service with regard to measure-
ment of cross-sections and t1me-of-travel and/or stage discharge records.
Therefore, before extensive hydraulic analyses are performed, all
existing data should be exhausted.
RECOMMENDATIONS
All water quality raveling investigations must demonstrate an
accurate treatment of hydraulics. All models must be able to demon-
strata that they effectively predict tirne-of-travel considerations
and, if more sophisticated, demonstrate accurate predictions of dis-
persive mixing through the use of dispersion equations. Whenever
field surveys are performed, t1me-of-travel and other hydraulic
measurements must be included for demonstration of accurate survey
111-19
-------
technique as well as for calibration of the hydraulic portions of
water quality models. Additionally, performing accurate time-of-
travel analysis will provide valuable information for spill
management. -
It is inappropriate to recommend a specific sampling technology
associated with hydraulic and time-of-travel measurement. However,
the level of effort for hydraulic measurement should be equivalent to
the overall level of effort for the water quality investigation. It
should be remembered that accurate hydraulic analysis is seldom invali-
dated by time, while other chemical and biological assessments may be
altered. Therefore, an accurate measurement of time-of-travel and
hydraulic cross-sections should occur. Data for at least three flow
conditions are necessary to characterize the hydraulic parameters.
When initiating sampling for calibration and verification of water
quality models, flow and water temperature should be similar to those
which will exist during the predictions of waste load allocations.
Normally, high temperature and low flow conditions are necessary unless
ice cover creates a more critical situation. Water temperature is
seldom difficult to achieve but low flow may be. As a general rule
flows should be less than twice the waste load allocation conditions
when sampling occurs.
This recommendation must be evaluated for the specific system
considering chemical, biological and physical factors and Employing
good professional judgment. Time constraints may prevent wating for
111-20
-------
a low flow event of such a magnitude. The modeler must assess the
potential alteration of the system at higher flow conditions.
When the tnodeler is confronted with authorization for a field
survey to commence, he should make sure that the flew conditions under
which the survey will occur are approaching those of the 10-yr 7-day
low flow conditions. As a general rule of thumb, the flow conditions
should not be more than twice that lew flow condition. Such a recom-
mendation may be impractical considering the number of surveys which must
be performed and the meteorological conditions which may prevent such
a condition from recurring frequently.
If it appears to the modeler that it is impractical to wait until
a low flow condition lass than twice the 10-yr 7-day low flow exists
two actions will be necessary: (1) the conduction of more than two field
surveys under varying flow conditions to calibrate and verify (with per-
haps two sur/eys) the water quality model. Additionally, in order tc
ensure the accurate prediction of the hydraulics of the water system, it
may be necessary to perform additional hydraulic (time-of-travel) studies.
(2) Inform headquarters that it is impossible to obtain conditions
similar to the 7-day low flow and that surveys must be performed and
request guidance. It is most important, however, that temperatures
of the receiving streams be selected and surveys perferred at conditions
similar to those which will be selected for inclusion in waste load
allocation studies. This is often more easily accomplished than
the low flow condition.
111-21
-------
III. DESIGN AND UPGRADING OF HAZARDOUS
WASTE MANAGEMENT FACILITIES
-------
OUTLINE
HAZARDOUS WASTE MANAGEMENT
INTRODUCTION
A. TECHNICAL
B. LEGAL
C. PUBLIC RELATIONS
D. PLANT MANAGEMENT
E. FINANCIAL
F. INTERRELATIONSHIPS (REGULATORY, POLITICAL)
G. PROJECT CONTROL
H. SCHEDULE
-------
FIGURE 1
COMPANY TASK GROUP
SENIOR CORPORATE MANAGEMENT
TECHNICAL
LEGAL
PRIMARY CONTACT
& COORDINATOR
PUBLIC RELATIONS
POLITICAL '
REGULATORY
PLANT
MANAGEMENT
FINANC
IAL
-------
A. TECHNICAL
1. IN-HOUSE
2. SINGLE CONSULTANT
3. SINGLE "LEAD" CONSULTANT
4. CONSORTIUM OF CONSULTANTS
5. TECHNICAL
A. WHAT TO DO
1. PROCESS
2. DEEP WELL
3. LANDFILL
4. INCINERATION/EMISSION
5. TREATMENT (WWT) DISCHARGE
B. PROFESSIONS
1. BIOLOGISTS
2. CHEMISTS
3. CIVIL ENGINEERS
4. CHEMICAL ENGINEERS
5. MECHANICAL ENGINEERS
6. GEOHYDROLOGISTS
7. TOXICOLOGISTS
8. PATHOLOGISTS
9. LABORATORY
10.
6. NORMALLY USED EVERY JOB
-------
B. LEGAL
1. IN-HOUSE
2. OUTSIDE STAFF
3. AT START DETERMINE IF TO BE USED; IF YES, INCLUDi
FROM START
-------
C. PUBLIC RELATIONS
1 . IN-HOUSE NORMALLY B.ETTER
2. PROFESSIONAL
3. INCLUDE FROM BEGINNING
OFTEN APPROACH FROM BOTH
CORPORATE
PLANT
4. DEVELOP OVER LONG TERM
-------
D. PLANT MANAGEMENT
1. CONSIDER PREVIOUS RELATIONS
REGULATORY
PUBLIC
TECHNICAL
2. RELATIONSHIP IN PROJECT
E. FINANCIAL
1. INVOLVED LATER
2. NECESSARY FOR CORPORATE DECISIONS
-------
F. INTERRELATIONSHIPS WITH OUTSIDE AGENCIES
1. PMS ANALYSIS
REASON FOR USE IS MANAGEMENT
2. CPM ANALYSIS
3. DISCUSSION
G. PROJECT CONTROL
REASONS TO HAVE IT
REASONS NOT TO HAVE IT
HOW TO GET IT
HOW TO LOSE IT
WHO ARE THE COMPETITORS
1. INDUSTRY CORPORATE
2. STATE
3. EPA
4. CONSULTANT
H. SCHEDULE
WEATHER
CORPORATE
REGULATORY AGENCY
AGENCY TURNOVER
-------
WATER TREATMENT TECHNOLOGY
by
W. Wesley Eckenfelder, Jr.
Distinguished Professor
Vanderbilt University
GENERAL
Treatment technology can be applied to the control of hazardous
waste spills either in situ, transfer of the spill material to a
portable treatment system, or controlled discharge to a wastewater
treatment plant. Selection of disposal technology will depend on the
nature of the spill material, available technology and the geographical
location of the spill.
ALTERNATIVE TECHNOLOGIES
There are a number of technologies which can be applied to the
neutralization and detoxification of a hazardous waste spill. Some of
these may be employed for in situ treatment, some for external treatment
and some for either alternative. The primary treatment alternatives are
shown in Figure 1 and are listed in Table 1.
The selection of treatment technology as shown in Figure 1 would
depend on the characteristics of the spill material. Depending on the
material, it could be discharged either to a water course or to a waste-
water treatment plant at any stage in the treatment sequence.
-------
Coagulation
Precipitation
To receiving water
toSTP
FIG. 1. ALTERNATIVE TREATMENT TECHNOLOGIES
-------
TABLE 1
TREATMENT ALTERNATIVES FOR HAZARDOUS WASTE SPILLS
Material
Acids/Alkal ies
Ammonia
Suspensions
Heavy Metals
Colloidal Dispersions
Organics
Organics
Technology
Neutral ization
Neutral ization
Sedimentation
Precipitation
Ion Exchange
Coagulation/Filtration
Adsorption
Chemical Oxidation
In Situ
Yes
Yes
Yes
Yes
No
No
No
No
External
Yes '
No
Yes
Yes
Yes
Yes
Yes
Yes
-------
NEUTRALIZATION
Acid or alkaline spills can usually be neutralized in situ. Acid
spills can be neutralized using lime, limestone or soda ash. Weak bases
have the advantage that an overdose will not result in an excessively
high pH. Alkaline spills can be neutralized with HC1 (or in some cases
H2S04). The quantity of alkali required for an acid spill can be estimated
/
from stochiometry. For example, the quantity of 90 percent lime required
to neutralize 95 percent H^SO, can be calculated:
Ca(OH)2 - » CaS04
96 74
9 T" = °'81 lbs lime/lb H2S04
If the acidity of the material is not known, a sample of the spill material
should be titrated with a standard lime solution to pH 7.0 (or other alkali
if it is to be used). The quantity of lime/gallon of spill can then be
directly calculated.
Ammonia requires special consideration since un-ionized ammonia (NhL)
is extremely toxic to aquatic life. Since the percentage of ammonia as NH^
increases with increasing pH, it is important to reduce the pH to below pH
7.0 before discharge. Caution should be exercised in neutralization since
the reaction generates considerable heat.
SEDIMENTATION
Sedimentation, i.e., the. removal of suspended particles by gravity
separation, can be accomplished in situ or by external treatment in a gravity
-------
separation basin or tank. The time required for separation in situ to
occur can be estimated by observing the subsidence in a beaker. External
treatment in a continuous-flow basin is related to the overflow rate in
the basin expressed as gal/sq ft/day. This can be roughly estimated by
observing the time required for the particles to settle 4 ft in a cylinder.
The settling rate in ft/hr can be computed:
4 ft ft
hrs to settle hr
The overflow rate in gal/sq ft/day is:
w x 18°
To compensate for turbulence and short circuiting, the overflow rate
estimated above should be divided by two.
COAGULATION
Spills containing inorganic or organic colloidal suspensions can be
treated by chemical coagulation. Coagulation can be defined as the addition
of a chemical to a colloidal dispersion which results in particle destabilize-
tion and the formation of complex hydrous oxides which form flocculent
suspensions. The flocculent suspensions are subsequently removed from the
liquid by sedimentation.
The most common coagulants in use today are alum, iron salts, and lime.
In some cases organic polyelectrolytes (cationic, anionic, or nonionic) can
be effectively used as a primary coagulant or in conjunction with alum, iron
or lime.
-------
In-the case of iron or alum, the charge on the colloidal particle is
neutralized by the A1'T or Fe ion and by positively charged microflocs
which are rapidly produced when the coagulant is added to water. Floccula-
tion of the mixture for 20 - 30 minutes will result in the production of
large floes which can subsequently be removed by sedimentation, flotation,
or filtration.
Effective coagulation is a function of dosage of coagulant and pH.
Sufficient alkalinity must be present to react with the added coagulant. '
The coagulation sequence is shown in Figure 2.
Coagulation is functionally an art and a series of jar tests should be
run to determine the optimum pH and coagulant dosage. The jar test procedure
can be readily done in the field. The test procedure involves varying the
pH (usually over a range of pH 4 to pH 10) with a constant coagulant dosage
which will produce a floe. Having established the optimum pH, the coagulant
dosage is varied to define that which yields the optimal removal. Details
of the test procedure can be found in reference (1).
PRECIPITATION
Most heavy metals can be precipitated as the hydroxide (Me(OH) ) by the
1 ^
addition of caustic soda (NaOH) or lime (Ca(OH)2). The reaction which occurs
is:
Me+X + Ca(OH)? > Me(OH) + Ca++
£ A
Most metals can also be precipitated as the sulfide. Precipitation of the
metal to an insoluble form eliminates the problem of seepage into the soil
or the surrounding area. In order to define the alkali requirements, the
following procedure is suggested:
-------
RAIMD MIX
FLOCCULATION
SEDIMENTATION
COLLOIDS
m O
r- >
8 32
33 d
O O
51
ANIONIC On NONIONIC
POLYELECTROLYTE
FIG. 2. MECHANISM OF COAGULATION PROCESS
-------
1. Titrate a sample with a standard lime solution to the terminal
pH of minimum solubility (see Table 2). Depending on the
volume of spill material, the quantity of lime required can
be calculated.
2. Spread the lime over the spill area, insuring contact between
the spill material and the lime for precipitation.
3. If external treatment is employed, lime in slurry form should
be fed and mixed with the waste at a rate determined from (1)
above.
4. It should be noted that for concentrated metal solutions, the
quantity of sludge produced may be equal to the quantity of
spill material. This will require on-site dewatering or disposal
as a wet slurry.
ION EXCHANGE
Heavy metals and other salts can also be removed by ion exchange in
which the metal is exchanged for sodium ion:
Me++ + Na2Z > MeZ + 2 Na
Ion exchange is a process in which ions held by electrostatic forces to
functional groups on the surface of a solid are exchanged for ions of a
different species in solution. This exchange takes place on a synthetic
resin. Various kinds of resins are available including weakly and strongly
acidic cationic exchangers and weakly and strongly basic anion exchangers.
The ions are exchanged until the resin is exhausted at which time the resin
is regenerated. The capacity of resins vary so that the necessary data on
-------
TABLE 2
REMOVAL OF HEAVY METALS BY PRECIPITATION
Metal
Arsenic
Barium
Cadmium
Copper
Lead
Mercury
Nickel
Selenium
Zinc
Process
Precipitation with S~
Carbon Adsorption
(low levels)
Fe(OII)3 coprecipitation
BaSO^,
Cd (OH)2
Fe(OH3) coprecipitation
ILO., oxidation
Cu (Oll)2
Fe(OII)^ coprecipitation
Pb(OH)2
Pb(OH)3
Pb S
Fe(OH)3; A1(OII)3 coprecipitation
N1(OH)2
Se S
Zn(OH)2
Effluent Level
0.05 mg/1
0.06 mg/1
0.05 mg/1
0.5 mg/1
0.1 mg/1
none
-
0.2 mg/1
0.3 mg/1
0.5 mg/1
0.001 mg/1
-
0.1 mg/1
0.15 mg/1
0.05 mg/1
-
Constraint
pll 6-7
pH 10.0
complexing ions e.g. CN~ require
pretreatment; pH 8.5
oxidizes CN~ and Cd to oxide
pll 9.0 - 10.3
pll 8.5
pll 10.0
pll 8.0 - 9.0
pH 7.5 - 8.5
Na?S added
pll 10.0
pll 6.5
ptl 8.5
-------
the resin must be obtained by the manufacturer. Since the resin usually
will not be regenerated on site, the necessary quantity of resin should be
established by chemical tests.
This is an external treatment in which the spill material is pumped
through an ion exchange column. The system can appropriately be considered
as a detoxification process since the resulting concentrated salt solution
will require a controlled discharge to a receiving water or a municipal
sewer or removal to a suitable disposal site.
CARBON ADSORPTION
Many organics can be removed by carbon adsorption. In the adsorption
process, molecules attach themselves to the solid surface through attractive
forces between the adsorbent and the molecules in solution. Adsorption
continues until equilibrium is established with the concentration in solution
The ability for organics to be adsorbed on carbon depends upon such factors
as molecular structure, solubility and the substitute groups in the molecule
(A general guideline is shown in Table 3.). Extensive adsorption studies
have recently been conducted by Dobbs et al (2) on a wide variety of toxic
organics and priority pollutants. Table 3 shows organics susceptible to
adsorption on carbon. The capacity mg/g at 1.0 mg/1 influent concentration
shows the relative capacity for adsorption of the organic on carbon.
Organic removal on carbon will usually employ external treatment through
granular carbon columns brought to the site. The spill material is pumped
through multiple columns in series as shown in Figure 3. A breakthrough
curve of the type shov/n in Figure 4 will result. When breakthrough occurs,
-------
TABLE 3
COMPOUNDS NOT ADSORBED BY ACTIVATED CARBON
1. Acetone cyanohydrin
2. Butyl amine
3. Choline chloride
4. Cyclohexylamine
5. Diethyleneglycol
6. Ethylenediamine
7. Hexamethylenediamine
8. Morpholine
9. Triethanolamine
-------
TABLE 3 (cont'd)
ORGANICS REMOVED ON ACTIVATED CARBON
Compound mg/gm
Hexachlorobutadiene 360
Anethole 300
Phenyl mercuric acetate 270
p-Nonylphenol 250
Acridins yellow 230
Benzidine dihydrochloride 220
n-Butylphthalate 220b
N-Nitrosodiphenylarnine 220
Dimethylphenylcarbinol 210
Bromoform 200
S-Naphthol 100
Acridine orange 180
ct-Naphthol 180
a-Naphthylamine 160
Pentachlorophenol 150
p-Nitroaniline 140
l-Chloro-2-nitrobenzene 130
Benzothiazole 120
Diphenylamine 120
Guanine 120
Styrene 120
Dimethyl phthalate 97
Chlorobenzene 93
Hydroquinone 90
p-Xylene 85
Acetophenone 74
a
-------
TABLE 3 (cont'd)
ORGANICS REMOVED ON ACTIVATED CARBON
a
Compound mg/gm
1,2,3,4-Tetrahydronaphthalene 74
Adenine 71
Nitrobenzene 68
Dibromochloromethane 63
Ethyl benzene 53
o-Anisidine 50
5-Bromouracil 44
Carbon tetrachloride 40
Ethylene Chloride 36
2,4-Dinitrophenol 33
Thymine 27
5-Chlorouracil 25
Phenol 21
Trichloroethylene 21
Adipic Acid 20b
Bromodichloromethane " 19
bis-2-Chloroethylether 11
Chloroform 11
Uracil 11
Cyclohexanone 6.2
5-Fluorouracil 5.5
Cytosine 1 .1
EDTA 0.86
Benzoic Acid . 0.80
Benzene 0.70
^capacity at C« = 1 mg/1
adsorption capacities at pH 3
-------
FILTER
I ,__J
]
FIG. 3. CARBON COLUMN CONFIGURATION
-------
1.0
o
o
CO
O
LU
CC
a
o
o
C/l
LU
_l
co
<
00
tr
O
CO
a
u.
O
2
g
H
u
<
DC
0.8
0.6
0.4
0.2
JBL
CHLORINATED HYDROCARUONS
(DICIILOnOETHAfJE)
10
15
20
TIME, dayi
25
30
35
FIG. 4. GRANULAR CARBON COLUMN BREAKTHROUGH CURVE
-------
carbon is replaced in Column 1 and it becomes Column 4, as shown in
Figure 3. This procedure is continued until all of the spill material
has been treated.
If there is a question as to the applicability of carbon to the
organic in question, a laboratory batch study can rapidly be made to
define carbon effectiveness:
In this test, various quantities of powdered carbon are mixed with
the spill material in a shaker assembly and mixed for one hour. The
mixture is then filtered to remove the carbon and the concentration of
organic remaining measured. The results are plotted as shown in Figure 5.
Since the exhausted carbon removed will be in equilibrium with the influent,
the quantity of carbon required for the treatment can be estimated from
Figure 5 by extrapolation to the influent concentration of organic and
selecting the Ibs organic removed/lb carbon. The total carbon requirement
can then be computed from a knowledge of the volume of spill chemical to be
treated.
In one case, powdered activated carbon has been applied in a slurry in
a mixing chamber for the removal of PCSs. By adjusting the carbon dosage,
the PCB concentration in the discharged water was maintained less than 1 ppb.
OXIDATION-REDUCTION
Chemical oxidation or reduction can be applied to a variety of spill
materials. Hexavalent chromium, Cr , can be reduced to the trivalent state
Cr , by the addition of a reducing agent such as SO- or sodium metab.isul f i te
-------
CAPACITY AT INFLUENT
CONCENTRATION
Z
o
33
O.
<
u
o
X
Q
UJ
c
U
2
<
O
O
o
u
<
<
O
^
O
a
c:
<
CJ
a
o
Co
LOG EQUILIBRIUM CONCENTRATION REMAINING C, mgfl
FIG. 5. CARBON CAPACITY ESTIMATION FROM LABORATORY DATA
-------
(NaS.,05). The reaction requires a pH of 2.0, so in some cases acid must
be added. The trivalent chromium can then be precipitated by the addition
of lime at pH 8.3 as the insoluble Cr(OH).,. The reactions are
Cr+6 +
Cr+3 +
so2
Na2S2°5
Ca(OH)2
PH=2.0 . +3 .
* LI T
> (vfni-n
' LI \ un i T
so4
+
Ca++
Sulfur dioxide (S02) is fed as a gas from cylinders while metabisulfite is
a dry powder. S02 has the advantage that it hydrolyses in water to the
acid H2$0., so that additional acid for pH adjustment is rarely necessary.
Removal of the chrome hydroxide sludge is necessary.
Spills of chlorine can be reduced with sulfite to the chloride ion.
Cyanide can be oxidized to harmless end products (N- and C-OJ by oxidation
with chlorine under alkaline conditions (pH 8.5). The reactions are:
CN + 20H" T C12 » CNO" + 2C1" + H20
CNO" + 40H" + 3C12 f 2C02 + N2 + 6C1" + 2H20
It is important to maintain the pH in the alkaline range to avoid the
production of noxious byproducts.
Other oxidants such as hydrogen peroxide (K?0?) and ozone (0.,) offer
£ C, O
some future promise but available data on their application is insufficient
at this time.
-------
BIOOEGRADATION
It is estimated that one-half of the 650 designated hazardous
materials will biodegrade. When considering hazardous waste spills, however,
there are a number of factors which must be considered. A majority of the
hazardous chemicals are complex organics and require long periods of acclima-
tion before effective biodegradation will occur. In many cases, concentration
limits exist to avoid inhibition and large dilution would be required.
Excessively long periods of aeration will usually be required to reduce the
contaminant to a level suitable for discharge to a water course or a sewer.
All of these factors would usually mitigate against biological treatment at
the spill site.
The possibility exists, however, of discharging the spilled chemical to
a municipal sewer at a controlled rate which will avoid shock loading of the
wastewater treatment plant and insure degradation of the chemical in the
biological treatment plant. Table 4 lists the biodegradability of various
organic compounds. Information on specific compounds is available in various
published sources. If the chemical is deemed biodegradable, the suggested
procedure is to control the discharge rate such that the organic loading rate
(F/M) as Ibs BOO/day/lb MLVSS does not exceed 0.2 (based on the chemical
discharged). This will therefore be in addition to the organic loading
normally received by the plant. (It should be checked that the total loading
to the biological plant does not exceed a loading of 0.5.) The following
example will illustrate:
-------
TABLE 4
RELATIVE BIODEGRADABILITY OF CERTAIN ORGANIC COMPOUNDS
Biodegradable Organic Compounds
Acrylic Acid
Aliphatic Acids
Aliphatic Alcohols
(normal, iso, secondary)
Aliphatic Aldehydes
Aliphatic Esters
Alkyl Benzene Sulfonates
w/exception of propylene-
based Benzaldehyde
Aromatic Amines
Dichlorophenols
Ethanolamines
Glycols
Ke tones
Methacrylic Acid
Methyl Methacrylate
Monochlorophenols
Nitriles
Phenols
Primary Aliphatic Amines
Styrene
Vinyl Acetate
Compounds Generally
Resistant to Biological
Degradation
Ethers
Ethylene Chlorohydrin
Isoprene
Methyl Vinyl Ketone
Morpholine
Oil
Polymeric Compounds
Polypropylene Benzene
Sulfonates
Selected Hydrocarbons
Aliphatics
Aromatics
Alkyl-Aryl Groups
Tertiary Aliphatic Alcohols
Tertiary Benzene Sulfonates
Trichlorophenols
Some compounds can be draded biologically only after extended
periods of seed acclimation.
-------
Spilled Chemical - Phenol
Biological Treatment Plant:
Aeration Volume - 2 million gallons
Aeration VSS - 3,000 mg/1
P/M - n ? - IDS BOD Applied/day
' ~ U^ " Ibs MLVSS
IDS BOD Applied/day
3,000 2 8.34
Ibs BOD5 Applied/day = 10,000
Ibs phenol applied/day (based on 1.87 mg BOD/mg phenol)
= 10,000/1.87
= 5,345
or discharged at a rate of 222 Ibs/hr of phenol
The possibility exists that pre-acclimated cultures in dry form could
be employed in some cases in an external treatment plant, or that biological
sludge from an industrial wastewater treatment plant treating the same or
similar chemicals could be employed. As a rule of thumb, the organic loading
of the spill material fed to the biological process should be adjusted to 0.1
Ibs BOO/day/lb MLVSS to insure a low effluent concentration of the spill
pollutant.
-------
REFERENCES
1. Adams, C., Ford, D., and Eckenfelder. Development of Process Design
Criteria for Wastewater Treatment Processes, Enviropress, Inc. 1979.
2. Dobbs, R.A., Middendorf, R.J. and Cohen, J.M. Carbon Adsorption
Isotherms for Toxic Organics, MERL EPA, Cincinnati, OH, May 1978.
NOTE: Space does not permit a detailed presentation of the theory and
application of the alternative control technologies. The reader
is referred to the following references:
1. Eckenfelder, W.W., Water Quality Management - Principles and Practices,
CBI Publishing Co., Boston, Mass 1979.
2. Metcalf and Eddy, Wastewater Engineering, Treatment, Disposal & Reuse,
McGraw Hill, New York 1979.
-------
GROUNDWATER CONSIDERATIONS
Prepared by
F. G. Ziegler, Ph.D., P.E.
Director of Resources Management
Associated Water and Air Resources Engineers, Inc
Prepared for
National Hazardous Materials Training Course
Vanderbilt University
Nashville, Tennessee
September 25-29, 1978
-------
GROUND WATER
1. Introduction
2. Legal Requirements
3. Scope/Magnitude of Problem
4. Information Sources
5. Basic Definitions and Concepts
6. Chemical/Physical Considerations
7. Monitoring As A Tool
8. Model Basics
9. Available Models
10. Prevention-Containment-Correction
-------
1. INTRODUCTION
Major Thrust Modeling
Review Outline
-------
2. LEGAL REQUIREMENTS
Acts Influencing:
PL 93-523 Safe Drinking Water Act
PL 94-469 Toxic Substances Control Act
PL 94-580 Resource Conservation and Recovery Act
-------
Existing Federal and state programs address many of the sources of
potential contamination, but they do not provide comprehensive pro-
tection of ground water.
- Existing Federal programs administered by EPA which address
ground water are (1) the Federal Water Pollution Control Act
Amendments of 1972; (2) the Safe Drinking Water Act of 1974;
and to a lesser degree (3) the Solid Waste Disposal Act of 1965;
and (4) the National Environmental Policy Act of 1969.
- The FWPCAA provide for a statewide and areawide waste treatment
management planning function which may include identifying and
controlling pollution from mine runoff, the disposal of re-
sidual waste, and the disposal of pollutants on land or in sub-
surface excavations.
- FWPCAA also include (1) a program to issue permits for point
sources of water pollution, including some wells; (2) best
practicable treatment standards for municipal sewage effluent
disposal which must address ground-water protection; (3) guide-
lines for land spreading of municipal sludges; and (4) munic-
ipal waste treatment facilities planning for areas where
septic systems pose potential adverse ground-water impacts.
- FWPCAA do not address the discharge of contaminants to ground
water from surface impoundments, land disposal of solid wastes,
septic systems, or most wells.
The basic and primary and essentially sole interest of the Water
Pollution Control Act Amendments of 1972 is directed to control the
pollution of surface waters. In fact, sections of the Act that require
-------
consideration of new and alternative waste treatment and disposal
methods are used to support and promote the use of land applications
of sewage and other liquid wastes and their residuals as a method of
achieveing zero discharge. When a land application methodology is
used, no discharge permit is required unless runoff is collected for
discharge from a point source to a surface stream. In the absence of
a permit, there is no legal responsibility or authority in the Federal
program to require discontinuance of the practice, with one exception-
the Administrator may act under his emergency powers to protect the
public health.
- The SDWA provides for a Federal/state cooperative effort to
prevent endangerment of underground drinking water sources
from industrial and municipal waste disposal wells, oil-field
brine disposal wells and secondary recovery wells, and eng-
ineering wells. At present, surface impoundments are not included
in this program, but some types of impoundments may be included
at a later time.
- SDWA also provides that EPA may review any commitment of Federal
financial assistance in an area designated as having a sole
source aquifer.
- SDWA cannot be used to regulate land disposal of solid wastes,
land application of sludges and effluents, or septic systems
except under the emergency powers provisions of the Act.
- The Solid Waste Disposal Act contains no specific reference
to ground water, however, guidelines developed under the Act
provide for ground-water protection from pollution activities
-------
and surface drainage. There are also site development guide-
lines which consider the impact on ground water. These guide-
lines are only mandatory for Federal agencies.
The NEPA requires Federal agencies to prepare environmental
impact statements on major actions. Ground-water protection
is a significant need for writing an EIS.
While site selection is an important parameter in preventing
ground-water contamination, there are no direct Federal controls
in this area. States are encouraged to develop site selection
programs within the context of their land-use planning and
control authorities.
Most state laws give broad authority to protect all waters of
the state, including ground water. Such language, plus de-
ficiencies in budget and staffing, force state and local agencies
to act on cases of contamination only after the fact.
States are beginning to develop programs which encourage pre-
vention of contamination from some waste disposal sources.
Because clean-up of contaminated ground water is rarely eco-
nomically or technically feasible, action by the states has been
directed toward condemning the affected water supply.
Legal action is seldom taken against a specific source of
contamination because individuals, private organizations, and
public agencies seldom have the resources required to prove a
specific source as the source of contamination.
-------
3. SCOPE/MAGNITUDE OF PROBLEM
Discussion and Transparencies
-------
Ground water is a high quality, low cost, readily available source of
drinking water.
- Half of the population of the United States is served by
ground water. (29% Delivered by Communities, 19% Private Well)
- In many areas, ground water is the only high quality, economic
source available.
- The use of ground water is increasing at a rate of 25 percent
per decade.
Waste disposal practices have affected the safety and availability
of ground water, but the overall usefulness has not been diminished
on a national basis.
- Current data indicate that there are at least 17 million waste
disposal facilities emplacing over 1,700 billion gal. (6.5
billion cu m) of contaminated liquid into the ground each year.
Of these, 16.6 million are domestic septic tanks emplacing
about 800 billion gal. (3 billion cu m) of effluent.
- Ground water has been contaminated on a local basis in all
parts of the nation and on a regional basis in some heavily
populated and industrialized areas, precluding the development
of water wells. Serious local economic problems have occurred
because of the loss of ground-water supplies.
- Degree of contamination ranges from a slight degradation of
natural quality to the presence of toxic concentrations of such
substances as heavy metals, organic compounds, and radioactive
materials.
- More waste, some of which may be hazardous to health, will be
-------
oo
too -400 eoo aoo
KILOMETERS
LtOEND
STATES WITH MAJOR RELIANCE
OM OHOUHO WATER
Figure Dependence of United States population on ground water as a source of drinking water.
-------
IIOUH
smic
UHK
OK
CtS'.KJQL
GROUND
WATER
INTENTIONAL HOUT£
UNKJTtN I I(J(IAI. HUUIE
WAS It UlSHOSAL fltACHCt'S COVttlEC IN THE Mt'POHT
Figure Wasle diipcmil pracMces and tlie routes of contaminants from solid and liquid wastes.
-------
PUUPIMG WUL X I ANOFIll. DUMP
OK KifUSl Pllf
LAND SPREAOIHC
ON IM Hit* 11(111
SEPTIC 1ANK
ON CESSPOOL slwl(,
e
DISCIIANCC LEAKAGE
-^S.~.
(VAPOINANSPIRAIIOM
> /
i^»c==X
/ A\\
' sf'Sft I x N
Vfty I ' A(,OON. PI I
LE A**ai
AQUIFER (FRESH)
CONFINING 20NE
ARTCSIAM AQUIflu (FHESM)
CONFINING ZONE
ARTESIAN AQUIFER (SALINE)
PUMPING Wll I
=0
DISLIIANUt
OH
IN Jt CIlOM
t(fiEND
<^ I INTENIIONAI IKPOI
<^* ' UHlNIEIinONAL INPUT
PINECriON Of cftoUHh - w* IIH
MOVE ME Ml
IDNAWIHG NOT TO SCALD
Figure - How waste disposal practices contaminate the ground-water system.
-------
ARKANSAS- WHITE and KED ,,
ILLIONS QAL. MILLIONS CU
a *oa
KM.OHCICH
< 10 < 37.8
rrrr^ 10-30 37.8-189.2
51-100 193 0-378.S
> 100 > 378.3
Figure - Tofal industrial waste wafer rreated in ponds and lagoom, 1968.
-------
-ESTIMATED NUMBER OF FACILITIES, VOLUMES OF WASTE, AND
LEAKAGE TO GROUND WATER.
Estimated
total
Waste disposal sractics number
Industrial impoundments
Treatment lagoons NA
All impoundments 50,000
Land disposal of solid wastes
Municipal 18,500
Industrial NA
Septic tanks and cesspools
Domestic 16,600,000
Industrial 25,000
Municipal waste water
Sewer sysrems 12,000
Treatment lagoons 10,000
Lend spreading of sludge
Municipal NA
Ma nufcc Turing NA
Petroleum exploration end
development
Wells 60,000
Pits NA
Mine waste
Coal
Waste water 277
Solid waste NA
Other 691
Disposal end injection wells
Agricultural, urban run-
off, cooling wafer and
s«wag« disposal wells 40,000
Industrial and municipal
injection wells < -00
Animal feeding operations
Cattle ' 140,000
Other NA
Estimated Estimated
total amount
sizs of wcsre
NA 1,700
NA NA
SCO, COO acre. 135
NA . NA
300
NA
470, COO mi 5, COO
20,COOacrej 3CO
NA NA
NA NA
260
NA 43
77
173, OOOac.-es 100
360
NA
NA
50,OCOocr-s 33
NA 7
bgy
mty
tw
bgy
bgy
'bgy
bsy
bgy
mry
bgy
mry
mry
Estimated
leakage
to ground
100 bgy
NA
90 bgy
NA
300 bgy
NA
250 bey
19 bgy
NA
NA
260 bgy*
43 bgy
8 bgy
600 m Ibs/y =C'd
1 00 bgy
NA
NA
NA
NA
bgy - billion gallons per yrar
mry - million tons per yecr
m Ibs/y - million pounds per year
- not opplics'ole
* - almost all returned to salt-warer aquifers
NA - insufficient data for esfirrcrs
12
-------
Table - SUMMARY OF DATA ON 42 MUNICIPAL AND 13 INDUSTRIAL LANDFILL
CONTAMINATION CASES.
Findings
Type of Landfill
Munieipol Indusfrial
of principal dcmcge
Canfamincfion of cqulfer only
Wafer lupply well(j) cfr'ec.'ed
Conrcminafion of surface wcrer
Principal aquifer affec?ed
Unconsolic'cfed deposit
Sedimenfcry racks
Crystalline rockj
*yp« of pallufcnr observed
General ccnrcminarion
Toxic jubsrancss
Observed distance Traveled by polluronr
Lsa fhan ICO feef
100 to 1,000 feer
.More fhcn I,OCO fesr
Unknown cr 'jnreporred
.vtaximum oajerved c'eprh ,o«nerrafed =y :ollufanf
Lsu then 20 fe=r
30 to 100 fear
More fhan 100 fesf
Unknown or unreoorred
Acfion raken regarding source of canraminarion
Landfill abandoned
Landfill removed
Confainmenr or ,'rearmenr of leac.-.cre
No lenown ccrion «
Acrion faken -eaarding groond-wcrer resource
Wafer jupply well(5) abandoned
Ground-warer moniroring progrcm eircbKshed
No known acrt'cn
Lifigcn'on
Lirigafion involved
No
-------
going to the land because of increased regulation against, and
the rising costs of, disposal of potential contaminants to the
air, ocean, rivers, and lakes.
- Removing the source of contamination does not clean up the
aquifer once contaminated. The contamination of an aquifer can
rule out its usefulness as a drinking water source for decades
and possibly centuries.
Almost every known instance of ground-water contamination has been
discovered only after a drinking-water source has been affected.
- Few state or local agencies systematically collect data on
contamination incidents, water supply wells affected, and
drinking-water supplies condemned as unsafe.
- Effective monitoring of potential sources of ground-water
contamination is almost non-existent.
- Typical water-well monitoring programs traditionally have not
been directed toward protecting public health because water
analyses normally do not include complete coverage of such
significant parameters as heavy metals, organic chemicals,
and viruses.
- There are potentially millions of sources of contamination
and isolated bodies of ground-water contamination nationwide.
- While detailed national inventories of all potential sources
of ground-water contamination have not been carried out, EPA
and some states have begun some inventories and assessments
of some waste disposal sources.
14
-------
Waste disposal practices of principal concern are those related to
industrial and urban activities.
- For every waste-disposal facility documented as a source of
contamination, there may be thousands more sited, designed,
and operated in a similar manner.
- The opportunity for severe contamination of ground water is
greatest from industrial waste-water impoundments and sites
for land disposal of solid wastes.
- Septic tanks and cesspools discharge large volumes of effluent
directly to the subsurface. In many cases, the degree of
treatment is not adequate to protect ground-water supplies.
- Contamination resulting from the collection, treatment, and
disposal of municipal waste-water exists but the magnitude is
unknown.
- Because there is a known potential for contamination from the
land spreading of industrial and municipal sludges, there is
concern about the expected increase in sludge generation over
the next decade.
- There have been far fewer reports of contamination of potable
ground-water supplies by the several hundred industrial and
municipal wells injecting into saline aquifers than from thous-
ands of shallow wells used to dispose of sewage, runoff, and
irrigation return flow to aquifers containing potable water.
Other waste-disposal practices, whose distribution is dependent
upon geology, climate, and topography, have also contaminated ground
water.
15
-------
- Contamination from oil and gas field activities is caused
primarily by improperly plugged and abandoned wells and, to a
lesser degree, poorly designed and constructed operating pro-
duction and disposal wells.
- Although specific case histories of ground-water contamination
related to the disposal of mine wastes do exist, adequate
documentation of the problem is unavailable.
- Ground-water contamination from the disposal of animal feedlot
wastes is a relatively new environmental problem, and few
cases of ground-water contamination have been reported.
Existing technology cannot guarantee that soil attenuation alone will
be sufficient to prevent ground-water contamination from a waste
disposal source.
- Proper site selection as well as proper operation and main-
tenance of facilities, is the principal technique available
for minimizing ground-water contamination problems.
- Such technologies as advanced treatment and physical containment
play a major preventive role where economics dictate that sites
be located in areas of critical ground-water use.
- Land disposal of wastes is not environmentally feasible in
many areas and such alternatives as waste transport, resource
recovery, ocean disposal, and surface-water or air discharge
should be investigated and may be more environmentally acceptable.
- Federal demonstration grants and technical assistance are pro-
vided to assist the development of new technology and facilitate
the application of existing technology.
16
-------
4. INFORMATION SOURCES
Sources presented here pertain primarily to Geology, Soils
and Groundwater quality and quantity.
1. EPA - State and Federal
2. USGS - State and Federal
3. Soil Conservation Service - State and Federal
4. COE
5. Well Drilling Records
6. Well Drillers
7. State Geologists
8. Department of Agriculture
9. Universities
Geology
Civil Engineering
Agriculture
10. Contractors
11. Construction Evaluations
12. Planning Agencies
Comments:
1. Old versus New Drawings
2. Field Surveys
3. Additional Testing
-------
5. BASIC DEFINITIONS AND CONCEPTS
-------
An aquifer performs two important functions'- a storage function
and a conduit function. It stores water, serving as a reservoir, and
transmits water like a pipeline. The openings or pores in a water-
bearing formation serve both as storage spaces and as a network of
conduits. The ground water is constantly moving over extensive dis-
tances from areas of recharge to areas of discharge. Movement is
very slow, with velocities measured in feet per day or even feet
per year. As a consequence of this and of the great volume represented
by its porosity, an aquifer detains enormous quantities of water in
transient storage.
Our prior discussion has indicated that openings in subsurface
geologic formations are of three general classes:
1. Openings between individual particles, as in formations
of sand and gravel.
2. Crevices, joints, or fractures in hard rock which have
developed from breaking of the rock.
3. Solution channels and caverns in limestone, and openings
resulting from shrinkage and from the evolution of gas
in lava.
Two properties of an aquifer related to its storage function
are its porosity and its specific yield.
Porosity
The porosity of a water-bearing formation is that part of its
volume which consists of openings or pores - the proportion of its
19
-------
volume not occupied by solid material. Porosity is an index of how
much ground water can be stored in the saturated material. Porosity
is usually expressed as a percentage of the bulk volume of the material.
For example, if one cubix foot of sand contains 0.30 cu ft of open spaces
or pores, we say that its porosity is 30 percent.
While porosity represents the amount of water an aquifer will hold,
it does not indicate how much water the porous material will yield.
Permeability
The property of a water-bearing formation which is related to
its pipeline or conduit function is called its permeability. Perme-
ability is defined as the capacity of a porous medium for transmitting
water. Movement of water from one point to another in the material
takes place whenever a difference in pressure or head occurs between
two points. Permeability may be measured in the laboratory by noting
the amount of water that will flow through a sample of sand in a
certain time and under a given difference in head.
20
-------
Land surfacs
o
"o
Caatllajy
* J
>" -\
Soil water
ad
-§
^J
rt3
1
No fres water
r!
FIGURE Divisions of subsurface watar
21
-------
j^Ptt/oma/ric fur fact
Wo|ec
liv.l
* t«»l 1»l
fit dillnlllcni.
CM
C\J
Figure - Ground-water Relationships
-------
ro
GJ
FIGUnE " perched Watej^Talbj^ Commjon, Leachate percolates «o the perched water
water lahle and flows dowrigrad ient lo the end of the confining layer where
it moves downward to the actual water t.ible.
-------
ro
POT£NTIOMELRI£-;
FIGURE - Two-Aquifer System With Opposite Flow Directions. Leachate first moves
into and flows with the ground water in the upper aquifer. Some of the
leachate eventually moves through the confining bed into the lower
aquifer where it flows back beneath the landfill and away in the other
d i rect ion.
-------
6. PHYSICAL/CHEMICAL CONSIDERATIONS
-------
TAHI.E '
CONTHIHUTION OF I.ANDF 11.1, I.I-ACIIATK INOrCATOKS TO
CKOIINH WATF.Il UY OTIITJl S(
Ind lea Lor
Phosphate
Cii'l c I.uin
M.'ignes I.uin
Sod I.uin
Pot iifi.'i.iuni
AMIIMOII 1 inn
Chloride
Stil f an;
Nitrate
11 Irarlionarc
1 ron
M-ini'anosc
Boron
SC 1 (Ml Illlll
'/. i.iir.
Lead
Or her li. in.
MI1AS
Pliunols
PCll
Orf, N
I'AII-IIC
|t)C
IVOI)
Ct> 1 J f orm
Virus
Highway I.caky
deJ.cJii(> aowt>rs
M
M
II 1.
r-t
u i.
M
II
M
II
1.
M
P
M
II
II
P
P
Land dj.s-
Septic Trrl- posa.l.
Lanka M In in p. gat ion sludge
M P >'
M L
M 1.
1. M L
M
M
I. 11 I-
M U M
II 1. "
M
MM II
II U
1.
M M
M I-
M 1-
P P
I.
P
'
M M
II '
M
P ''
P P
Petro-
leum ex pi Feed-
it dev. lota
P
M
II
II I-
M
M H
I.
P
1.
M
M
l>
f
M-Moiliiral.e
l.= l.nw
P=Potei»tlal
-------
TABLE
PROBABILITIES OF LANDFILL LEACHATZ INDICATORS FROM
GIVEN SOURCES CONTAMINATING GROUND WATER
Indicator
Phosphate
Calcium
Magnesium
Sodium
Potassium
Ammonium
Chloride
Sulfate
Nitrate
Bicarbonate
Iron
Manganese
Boron
Selenium
Zinc
Copper
Lead
Other a.m.
MBAS
Phenols
?C3
Org N
PAH- KG
TOC
BOD
Colif orm
Virus
Waste lagoons
and ponds
II
III
III
T
x
III
II
x
I
T
III
T
±.
I
II
II
II
II
II
II
III
I
II
II
III
II
II
III
III
Buried pipelines
and tanks
III
III
III
I i
III
III
II
II
III
III
III
III
II
II
I
III
III
I
I
I
III
III
I=Highly probable
II=Probable
m=Unlikelv
27
-------
TAHLK
LEACIIA'IT. INDICATORS
Physical
Chemical
Iliologl cal
IX)
CO
Appearance
pll
Ox Jda tJon-RcdiicL Ion
Potent: la 1
Conduct l.vl Ly
Color
Tnrbildity
Tempo ralu re
(Kl or
ORGANIC
Phenols
CheinJcal. Oxygen
He ma ml (COP)
Total Or gnu Jr.
Carbon (TOC)
Volatile Adds
Tann Ins, l.j(>nlns
Or j>nn l.c-N
Kdier Soluble
(oJ 1. & grease)
MIIAS
Orgnnlc Pit net: l.onn.l
Hroups as Kcqii I. rod
(!hl or InaLcil
llyil rocarbons
IMORCANT.C
Total Ul.carbonate
Solids (TSS, TDS)
Volatile Sol Ltls
Chloride
Su l.fnte
Phosphate
AlkallnJty and
Acidity
Nitratti-N
N(.trlte-N
Ammon ia-M
Sod linn
Pot ass linn
Ca 1 cliiui
M;i|;nes linn
Ha I'dnoss
Heavy Metals (I'M), Cn,
NJ. Cr, /.n, Cd, Fe,
Mn. SI, ll(.. As, Se,
»n, AR)
CyanIdc
FinorIdo
Uloclicinlcal
Oxygen Demand
(HOD)
Collform
Ha cLerln
(Total, fecal;
fecal
streptococcus)
Standard Plate
Count
-------
r>o
UD
TAULli
SUSCE1TIE11LITY OF l.liACHATLi CONST.1TUENTS TO UTFFEUKNTIAL
ATfliNUATl ON
At teniuitetl Cong t J CULM if.
Chloride-
Snlf.-ite
Snl fide
Plios|>lia«:e
HI. l. rat «i
Aimunii i vim
Soil Linn
Potass! urn
('.» 1 Cl Mill
Miiftiies'liini
Heavy Me un 1 An Ions
(Cr, V. Se. l», As)
Heavy Metal Cations
(I'l) Cn, HI, 7.n, Ctl. I'e, Mn. H|»)
Or (5. -in i c Hit: rogen
con
lion
Volac.l le Ac \i\s
I'licnoJs
MIlAfi
Solid Waste Zone
O
(5-11
C
0
0-1)
O-ll
O-ll
0
0
0
0-11
U-A-C
0
0
0
0
0
0
UiiunCtirntcit Zone
O
0-1)
C-li
A-C
O
A-n
0
A
A
A
O-ll
A-C
it
n
it
11
A-!t
A-ll
Aquifer
0
0
C
A-t:
0
A
0
A
A
A
O
A-C
0
0
O
0
0-A
0-A
0- Nfi Ar.tr. n tin I Ion A= Absorption
(.'= Cliomlcii I I'rer Ipl tatlori
lt= 11:1 oclioinl.c;i I. Uagrnda tlou On Convention
-------
7. MONITORING AS A TOOL
-------
L EGEND
D
CASING
A.D.C -
SCREEN .
M ON (TORINO
WELLS
:S A
^ft:xfi:v^fo^f/^ S ^|^^^|^^^ft:-^fe^-'?^i^^
^NlV/^Vs^s^sV/xC-VV^^
N p'£ji:;!-:-'.':';;
(£^^
^BB^
FIGURE , MONITORING NETWORK FOR AQUIFERS WITH
INTERGRANULAR POROSITY - VERTICAL FLOW PATTERNS
-------
CA?
LAND SURFACE
BOREHOLE
SCHEDULE 40 ?VC
CASING
SLOTTED SCHEDULE
40 ?VC SCREEN
FIGURE
LOW PEHMEA3ILITY
3ACXFILL
GRAVEL ?4CX
WATER TA3L£
TYPICAL MONITORING WELL SCREENED
OVER A SINGLE VERTICAL INTERVAL
32
-------
CEMENT OR
SENTONITE GROUT
SLOTTED SCHEDULE
40 PVC PIPE
REMOVABLE
PVC CAP
CONCRETE PLUG
SCHEDULE 40
PVC PIPE
SAND OR
GRAVEL PACK
FIGURE
PIEZOMETER WELL INSTALLATION FOR
SHALLOW GROUND-WATER MONITORING
(After Clark, 1S75)15
33
-------
PRESSURE-VACUUM
LINE
DISCHARGE LINE
LAND SURFACE
LOW PEP
MATERIAL
90REHCL3E
POROUS GR SLOTTED
PVC PIPE
CHECX VALVE
SANO 3ACXF1LL
POLYETHYLENE T'JSING
"T"ANOEL30W FITTINGS
SAMPLE COLLECTION
CHA.MSES
ENOCA?
FIGURE
DETAILS OF A LOW-COST PIEZOMETER MODIFIED
FOR COLLECTION OF WATER SAMPLES
34
-------
to
en
200-
GAMMA
LOO
ELECTRIC
LOG
( RESISTIVITY)
DRILLERS
LOG
::: '.'' S A M n ';;.: Vy.
-CLAY-
.{.v
FIGURE
"gamma log is Included because it indicates
that the leacliate plume is not actually a clay
layer.
DETECTION OF A LEACHATE PLUME USING AN ELECTRIC WELL LOG
-------
8. AVAILABLE MODELS
-------
Figure Coefficients of Payability and Trsnsmissii
(Frorn Theory of Aquifer Tcs*s b
U. S. " ' -
«
s: 5i- "62,
37
-------
THEORETICAL ASSUMPTIONS
The basic equations that are solved numerically by the finite
element method are derived on the basis of a number of assumptions as
to the physics of motion and achievement of continuity of mass. These
assumptions must at least be approximately fulfilled in order to apply
the method, the following brief discussion outlines the basis equations
and associated assumptions.
Darcy's Law
Oarcy's law is the dynamic equation expressing the conditions which
control the flow rate under low and medium velocity, laminar flow conditions
in a porous medium where inertial forces (i.e., forces associated with the
mass of the fluid times its acceleration) can be neglected. It can be
conveniently stated in vector component form for two dimensions as:
v - -K
x xx 5x
where:
v = the component of the bulk velocity or discharge per unit
A
area in the x direction,
v = the component of the bulk velocity in the y direction,
K = a principal component of the conductivity ellipse which
A A
corresponds in direction to the x direction,
a principal component of the conductivity el"
corresponds in direction to the y direction,
38
-------
f\ l>*
TT = the component of hydraulic gradient in the x direction,
^h
y- = the component of hydraulic gradient in the y direction,
h = hydraulic head (a potential function) = + z,
p = fluid pressure,
Y = unit weight of the fluid, and
z = elevation above a datum (z = y if a vertical flow cross
section is being used).
The form of Darcy's lav; given by equation 1 assumed (a) that the porous
medium is fully saturated, (b) that it is anisotropic with two principal
values of hydraulic conductivity K and K , oriented in the x and y
xx yy
directions (or isotropic if K = K ), and (c) that these and inter-
xx yy
mediate values of hydraulic conductivity are the same for flow in
opposite directions.
The Continuity Equation
This equation expresses the idea that fluid may be neither created
nor destroyed. Flow is assumed to be steady state (i.e., the components
of velocity do not vary with time). One form of the continuity equation
for steady state flow in two dimensions is:
3(v m) 3(v m)
- ^- + -3y- =W(x'y)
where:
m = the local thickness of the flow region, and
W(*,y) = a source or sink term in units of discharge per unit area;
it is positive for a sink and negative for a source.
39
-------
In addition to steady state conditions, equation 2 assumed (a) that
variations in fluid density throughout the porous medium are negligible,
(b) that variations in thickness of the flow region (measured normal to
the x,y plane) are small enough that they produce negligible flow components
in the direction normal to the x,y plant, and (c) that the flow system
is in dynamic equilibrium or is only slowly unsteady. An example of a
system in dynamic equilibrium would be a ground-water basin in which
fluctuations in the water table position are cyclic with seasons of
the year, are small enough that they constitute a small percentage of
the total flow region, and do not change the relative configuration of
the water table. A mean potential distribution could then be estimated
using the program.
Basic Flow Equations
The equations to be solved result from combining equations 1 and 2:
Ix (Kxxm £> * fy ' "<»»> (3)
If a cross section of a three dimensional flow system is being studied
then m = 1, and the equation is:
where:
W'(x,y) = source or sink in units of discharge per unit area per
unit thickness.
For constant, isotropic conductivity or for flow of an ideal, incompressible
fluid.
40
-------
32h , 32h W'(x.y)
"~~
where K is the constant conductivity (K = 1 for flow of an ideal,
incompressible fluid).
The program will also solve the equation describing flow which is
symmetric around a central axis. For this case the Cartesian coordinates
(x,y) are transformed to axisymmetric cylindrical coordinates (r,z), and
the appropriate flow equation is:
F IF * f? ° (6)
It should be noted that the directions of the x and y axes do not
have to be constant throughout the flow region. As will be shown further
on, the directions need be constant only within each element (although
they must always be normal to one another). This allows for analysis of
porous media with varying directions of anisotropy. In applying this
idea to the (r,z) coordinate system, it must be remembered that the
properties of the porous medium must be axisymmetric. For all practical
purposes this limits analysis of axisymmetric ground-water flow problems
to vertically stratified porous media so that the symmetry axis is
vertical.
41
-------
Techniques of Wafer-Resources Investigations
of the United States Geological Survey
FINITE-DIFFERENCE MODEL FOR
AQUIFER SIMULATION IN
TWO DIMENSIONS
WITH RESULTS OF
NUMERICAL EXPERIMENTS
BOOK 7
CHAPTE3 Cl
-------
FINITE-DIFFERENCE MODEL FOR AQUIFER SIMULATION IN
TWO DIMENSIONS WITH RESULTS OF
NUMERICAL EXPERIMENTS
By P. C. Trescatt G. F. Finder, and S. P. Larson
Absiraci
Tha model STL! jimuiata groand-°ratar Sow in an
artesian aquifer, a. Tracer-able aquifer, or a com-
bined artesian and wuter-cable aquifer. Tie aquifer
may be heterogeneous and aniaoeropic and hare ir-
regular boundaries. The jonixa tern in the 3crw equa-
tion 337 include w^il diaciarjs. constant rechar?s.
from connnin? beds ia -sTiica the effects of
s are considered, ind erapotranapiracoa as
a Linear function of depch to Tatar.
The theoretical development ir.clndea presentation
of the appropriate daw equacocs and derivation of
tha inite-ditfersncs approximations (^r-tien for a
variable grid). The docussentacion emphasizes the
numerical tachniques that <»" be used for solving the
simultaneous ^quarior.s and describes the results of
numerical experiments csing: thesa tachnicraea. Of
the three numerical leciniqnea available in the aodei,
the stron^iy implicit procsdura, ia jsssral, requires
leas compuiar drae and fca? 'eTfer numerical diS-
cnltiea than do ie iterative aitematiny dir^cdon im-
piict procsdurs and line jucsassiva overreiaxaiion
(which inciudea a ^o-diraensional corr««rdoa pn>
csdnre to accelerata conver^enca).
The documentadon includes a low char^ jrograra
listing, an example simulation, and secrons on de-
jigniEg an aquifer model and rsqtiirem«nta :'or data
boat. It illustratea how modal rsaults can be prs-
aented on the line printer and pen plotters -witli a
program iai atilires ie graphical display soft. *are
available from the GrMiogtcaJ Sorve7 Computer
Center Division. In addition the ncdei includes op-
tions for rsadia? input data from a disk and -arrltinj
intermediate results on a disk.
Introduction
The finite-difference aquifer model docu-
mented in this report is designed to simulata
in two dimensions the response of an aquifer
to an imposed stress. The aquifer may be
artesian, water table, or 2. combination of
artesian and water table ; it ma/ be hetero-
geneous and anisotropic and have irrsgtdar
boundaries. The model permits leakage from
confining beds in which the sifsca of storage
are considered, constant recharge, evapo-
transpiration as a linear junction of depth
to water, and well discharge. Although it was
not designed for cross-sectional problems, the
model has been used with some success for
this type of simulation.
The aquifsr simulator has evolved froni
Finder's (1970) original model and modifica-
tions by Finder (1969) and Trescott (1973).
The model documented by Treacoct (1973)
incorporates several features described by
Prickett and Lonnquist (1971) and has been
applied to a variety of aquifer simulation
problems by various users. The model de-
scribed in chis report is basically the same as
the 1973 version but includes minor modifi-
cations to the logic and data inpu;. In addi-
tion, the user may choose an equation solving
scheme from among the altarrating direction
implicit procedure, line successive overreiax-
ation, and the strongly implicit procedure.
The program is arranged 20 that other tech-
niques for solving simultaneous equations can
be coded and substituted for the iterative
techniques included with the model.
The documentation is intended to be r
sonably self contained, but it assumes that
the user has an elementary knowledge of the
physics of ground-water uow, finite-differ-
ence methods of solving parriai differencial
43
-------
TECHNIQUES OF WATES-SESOURCES INVESTIGATIONS
equations, matrix algebra, and the FOR-
TRAN IV language.
Theoretical Development
Ground-water flow equation
Tie partial differential equation of ground-
water Sow in a confined aquifer in rwo di-
mensions may be written as
m wruch
r.-, T-j, T,,, T.n are the components of
the transmissivity tensor
'" is hydraulic head (L) ;
5 is the storage coefficient
(dimensionless) ;
W(~,y, 0 is the volumetric flux of re-
charge or withdrawal per
unit surface area of the
aquifer (Li~l).
The reader is referred to Finder and Brede-
hoeft (1963) for development and discussion
of equation I. In the simulation model, equa-
tion 1 is simplified by assuming that the
Cartesian coordinate axes - and y are alined
with the principal components of the trans-
missivity tensor, T.x and Ty
3 ,
3=
(2)
In water-table aquifers, transmissivity is a
function of head. A. -tuning that the coordi-
nate axes are co-linear with the principal
components of the hydraulic conductivity
tensor, the flow equation may be expressed as
(Brecehoeft and Finder, 1970)
3 _. ,3/t
3A
(3)
in which
^u» K;V are the principal components of
the hydraulic conductivity
tensor (Lt~l) ;
S7 is the specific yield of the aqui-
fer (dimensionless) ;
b is the saturated thickness of the
aquifer (L).
44
-------
9. PREVENTION-CONTAINMENT-CORRECTION
-------
LINERS FOR LAND DISPOSAL SITES
An Assessment*
One potential environmental impact of landfills is contamination
of ground and surface waters which can occur from improperly located,
designed, or operated land disposal sites. The potential for con-
tamination occurs because within a land disposal site various physical,
chemical, and biological processes take place which produce compounds
that can be dissolved or suspended in water percolating through the
solid waste. Waters contaminated in this manner are called leachate.
The occurrence of leachate does not mean that ground and surface
water will be polluted. Methods to control leachate are available.
One of these methods is to collect the material and treat it to
remove the harmful constituents. Collection of leachate requires
that a barrier exists between the solid waste which produces leachate
and the water that would become polluted. The barrier can be made
from existing impervious soil or by importing other construction
materials. The most common barrier is made by building the land
disposal site so that a "bathtub" is formed. The sides and bottom of
this type of site must be impermeable in order to contain the leachate.
Also, provisions must be made so that the leachate collected can be
removed for treatment. Sloping the bottom to a sump where a pump
is installed is the most common way of providing for removal to a
treatment facility.
*
Mr. Geswein is a Sanitary Engineer in the Systems Management
Division of EPA's Office of Solid Waste Management Program.
46
-------
The installation of impermeable liners in a solid waste land
disposal site is a recent development, so very little is known
about long term effects. The base of a landfill can be a hostile
environment for these materials. Anaerobic, reducing conditions
are encountered so the durability and integrity of the barrier can be
questioned, particularly long term integrity. Even materials, such
as layers of clay and polymeric membranes, which are usually considered
inert may react with the leachate, resulting in liner failure.
The construction of a large impermeable barrier can be a diffi-
cult task. The special techniques that are required for each diffi-
cult material type are presented later in this report.
Cost is a major consideration in any construction project.
Because none of the proposed materials have been judged superior to
another, cost will very likely be one of the considerations that can
be examined closely during the design process. Cost estimates are
presented later in this paper.
ASPHALT
Several types of asphalt liners have been used at various land-
fill sites. One of the first installations was constructed in 1971
at Montgomery County, Pennsylvania. The liner was a three inch
thick tar cement pavement. The aggregate for this liner was the
same as is commonly used for street paving except tar was used as the
binder rather than asphalt. A one-eighth of an inch thick coating
of hot tar was then sprayed over the pavement as a sealer. The
pavement was then protected by a two to three inch cover of crushed
47
-------
Table
TYPICAL SANITARY LANDFILL LEACHATE COMPOSITION'
Analysis
pH
Hardness (carbonate)
Alkalinity (carbor.-ts)
Calciura
Magnesium
Sodium
Porassiun
Iron (tocai)
Chloride
Sul rate
Phosphate
Organic nitrogen
Anunorua nitrogen
Cond'-icr ivity
SOD
COD
Suspended solids
Ranae
Low
5. 7
55
510
240
64
S5
23
6
96
40
1.5
2.4
0.2
100
7,050
800
15
of Values"
High
8.5
5,120
9,500
2,570
410
5,300
1,360
1,640
2,550
1,220
130
550
845
1,200
32,400
50,700'
:&, son
* Source: Leonard S. Wegman Co., Inc. lycical specifications
of an impermeable membrane. Lycoming County Board of Commissioners,
Pennsylvania. Unpublished data, 1974.
T Values are given in milligrams per liter except pH (pH units}
and conductivity (micromhcs per centimeter).
48
-------
rock (maximum size three-eighths of an inch) and an additional 12 to
18 inches of incinerator residue was placed over the pavement. The
base for the pavement consisted of four feet of broken stone, four
feet of backfill, and a six inch layer of crushed stone (maximum size
one inch). The base was built very thick because the site was an
abandoned quarry and the contractor was building up to escape ground
water. Tar was used as the binder in the bituminous paving because
it, unlike asphalt, is heavier than water.
POLYMERIC MEMBRANES
Six polymeric liner materials have been proposed as sanitary
landfill liners. They are PE, PVC, butyl rubber, Hypalon, EPDM, and
CPE. PVC is the most popular of these materials. It has been used
at Romeo, Michigan, North Hemstead, and Brookhaven, New York, and has
been selected for use in Lycoming County, Pennsylvania. (Further
information on the Lycoming County project is given in Appendices B
and C). Harrisburg, Pennsylvania, has installed a butyl rubber liner
at a disposal site used for incinerator residue. The SHWRL has
installed both Hypalon and CPE liners at the Boone County field site
(Walton, Kentucky). There are no known full scale liner applications
using either PE or EPDM.
TREATED SOILS
One commercial firm offers refined montmori1lonite, a naturally
occurring clay mineral, as an admixture to be used with native soils
to provide a liner. The material is sold under the commerical names
49
-------
Bentonite and Volclay. Crystal Lake, Illionis, has an operating site
with this type of liner, and Toronto, Canada, is currently building one,
CONSTRUCTION METHODS
The construction of a sanitary landfill liner requires close
attention by the field engineer. Three distinct phases of con-
struction have been identified. These are: subgrade preparation,
liner installation, and liner protection.
SUBGRADE PREPARATION
Any landfill liner must be built on a firm base in order to
prevent significant differential settlement of the subgrade and
subsequent loss of liner integrity. The specifications for the sub-
grade preparation should include the appropriate soil tests to
insure that optimum compaction is achieved.
Wet and/or cold weather make the construction of the subgrade
and liner more difficult and should be avoided when possible. When
liners are built during adverse weather conditions, more efficient
monitoring and control procedures should be used by the field
engineer to insure the installation of a quality product.
LINER PROTECTION
None of the proposed liner materials should be used as a pavement.
While some of these materials can easily support rubber-tired con-
struction equipment, no manufacturer recomnends allowing collection
vehicles to use the liner as a pavement because of the high wheel
loadings. Equipment with crawler treads should not be allowed to
50
-------
operate directly on the liner. Manufacturers recommend protecting
the liner with an earth cover one to two feet thick. This material
should not contain jagged rocks or other sharp objects that could
damage the liner. Similarly, the first lift of solid waste placed
in the fill site should not contain items such as bulky wastes,
pipe or white goods that could puncture the liner during the filling
operation. Such quality control is difficult to achieve, considering
the heterogeneous nature of solid waste delivered in compactor trucks.
COSTS
The cost of liner materials is difficult to establish. Many of
the proposed materials are petroleum products which are increasing
in cost. The relative costs given in Tables 2 and 3 are as meaningful
as the absolute values.
51
-------
Table
COST FOR VARIOUS SANITARY
LANDFILL LINER MATERIALS*
Material
Installed cost"
(S/sq yd)
Polyethylene (10 - 20i"mils^)
Polyvinyl chloride (10 - 30~ mils)
Sutyl rubber (31^3 - 52. 5? mils)
Hypalon (20 - 45^ mils)
Ethyl ene proovlane diene monomer
(31 .3 - 52*5? mils) t
Chlorinatsd polyethylene (20 - 3Q~mils)
Paving asphalt with sealer coat (2 inches)
Paving asphalt with sealer coat C4 inches)
Hot Sprayed asphalt (1 gallon/yd )
Asohalt Sorayad on Polyoroovlene fabric
' (100 mils)
Soil -bentonite (9.1 T:s/yd 1
Soil-bantonita (13.1 lbs/yd2)
Soil-cement with sealer coat (5 inches;
0.90 -
1.17 -
3.25 -
2.38 -
2.43 -
2.43 -
1.20 -
2.35 -
1.50 -
1.25 -
0.
1.
1.44
2.15
4.00
3.06
3.42
3.24
1.70
3.25
2.00
1.37
72
17
25
( includes
earth cover)
* Source: Haxo, ri. £. Jr. Evaluation of liner materials.
U.S. E?A Research Contract 53-03-0230. October 1973.
* Cost does not include construction of subcrade ncr the
cost of earth cover. These can range from 30.10 to SO.50/yd^/ft
of degth.
= Material costs are the same for this range of thickness.
§ One iiii 1 = 0.001 inch.
52
-------
lable
COST OF TAILINGS POND LINERS*
Liner material Installed cost'
(S/sq yd)
Bentonite
18 Ib/sq yd 1.25
Asphalt
Asphalt membrane 1.26
Asphalt concrete 1.80
Rubber
Butyl
1/15" 3.78
3/64" 3.24
1/32" 2.70
Ethylane propylane diena monomer
1/15" ' ' 3.69
3/54" 3.15
1/32" 2.61
Synthetic membrane
Polyvinyl chloride
10 mils 1.17 (includes
20 mils 1.62 earth
30 mils 1.98 cover)
Chlorinated polyethylene
20 mils 2.34
30 mils 3.06
Hypalon
* 20 mils 2.34
30 mils 3.06
* Source: Clark, 0. A., and J. E. Moyer. An evaluation
of tailings ponds sealants. Environmental Protection Technology
series EPA-560/2-74-065. Washington, U.S. Government Printing
Office, June 1974. p. 22-23.
* Includes material and labor. Cost of subgrade preparation
and, except where noted, earth cover is not included.
53
-------
4. HAZARD AREA ESTIMATES
4.1 DEFINITION OF HAZARD AREA
The hazard area that results from a release of contaminant into the
atmosphere is an area downwind of the source where ambient concentrations
endanger the health and safety of persons exposed to the material for a
relatively short period of time, on the order of an hour or less. It is
also called an "exclusion area" since all persons within its delineated
boundaries should be evacuated as quickly as possible.
4.2 THE PREDICTION MODEL
The basic expression for maximum (center line) concentration downwind
from a source of airborne contaminant is:
Q
*CL = .ayo2u
where Xn = the concentration, usually expressed in micrograms of contam-
3
inant per cubic meter of air (ug/m )
Q = source strength or rate of contaminant generation, expressed
as grams of airborne contaminant (gas or vapor) per second
(g/sec)
o = standard deviation of concentration crossv/ind throuah the
y
toxic cloud, given in meters (.71)
c = standard deviation of concentration vertically through the
toxic cloud, given in meters (~,)
u = wind speed averaged over a discrete period of time, say
30 seconds, expressed in meters per second (1 m/sec = 2.3 mph)
- = 3.1416
The two standard deviations increase as the cloud spreads during
downwind movement. Hence, the computed concentration, ^CL refers to a
given point downwind at which the cloud measurements are considered.
4-1
-------
Downwind cloud dimensions also depend upon the intensity of turbu-
lence in the atmospheric layers in which the cloud is moving. Turbulence
results when the air is healed by the sun (thermal turbulence) or when it
blows over rough ground or around obstacles (mechanical turbulence). Gn
a typical sunny afternoon with gentle to moderate winds, the lower layer
of the atmosphere exhibits pronounced turbulence, and is said to be
unstable. Under these conditions a cloud of contaminants is buffeted
laterally and vertically by turbulent gusts. It becomes diluted by the
ambient air so that concentrations diminish rapidly downwind. On a
typically clear, quiet night, the atmosphere is said to be stable, marked
by the absence of turbulence. Under these conditions, a cloud of contami-
nants is carried by a light wind as a narrow plume in which concentrations
at considerable distances from the source are still comparable with those
near the source.
4.3 TGXICITY FACTOR
For hazard area estimates, the additional factor of toxicicy must be
considered. Given two gasecus contaminants, for example, acrylonitrile
and hycrocen cyanide, escaping into the atmosphere at the same rate and
traveling downwind over a given terrain under identical = "ospner:c
conditions, the pattern and values of ambient concentration will be tr,e
same but the more tcxic substance, hydrogen cyanide, will oe a threat to
nealth anc safety farther downwind and ever a larger area t.v.n will icrylc-
- nitrile. For the fifteen chemicals that rated highest in potential hazard,
the ccncent-ticns listed in Tabie 4-1 were considered to define tne
r.azarc area boundaries.
-------
Threshold Limit Hazard Ar^a
Value (1970) Valu*
Chemical (rcg/m3)* . (|ng/m3)
Formaldehyde 3c 3
Chlorine 3 5
Hydrogen cyanide 11 (skin) 17
Hydrogen sulfide 12 18
Anhydrous ammonia 18 27
Acrylonitrile 45 (skin) 68
Benzene 80c (skin) 80
Ethylene oxide 90 135
Methanol 260 315
Acetaldehyde 360 450
Vinyl chloride 770c 770
Propane (L.P.G.) 1800 2250
Butanes (estimated) 2000 2500
Acetone 2400 3000
Ethyl chloride 2600 3250
Table 4-1. Hazard Area Limits for the Fifteen Highest Rated Chemicals,
*American Conference of Government Industrial Hygienists; includes
intended changes for 1970.
c = ceil ing value
4-3
-------
4.4 HAZARD AREA TEMPLATES
For a first approximation of potential hazard area downwind frcm
the site of an accidental episode, the use of templates is recommended.
Three templates labeled Figures 4-1, 4-2, and 4-3 will be found inside
the back cover of this manual. They are designed for Unstable, Neutral,
and Stable atmospheric conditions, respectively, and for use as overlays
on 1:24,000 U. S. Geological Survey (naps. They assume a source strength,
q, of 1 kg/sac and, downwind speed u" of 3 m/sec (7 mph). The templates
will be easier to handle and maintain if they are mounted onto plexiglass
or similar hard, transparent surface.
Downwind concentrations were computed frcm Turner's Workbook* for
Pasquill stability categories 3, 0, and F, respectively. For each case,
wind direction variability was considered in accordance with the roll owing
total angular range:
e For Category B, 80 degrees
9 For Category D, 30 degrees
For Category F, 15 degrees
These are approximately four standard deviations of the lateral wir.c
fluctuations that are observed for each typical case?*
To use the template, simply place the zero distance cc'nt over tha
accidental-episode location and rotate the tsmolata so that its car/:ar lir.a
is oriented in the downwind direction. Trace onto the basa nsc t."= rczar-
area identified by the name of the chemical involved, inclucing ali or tne
delineated area between zero distance and the labeled isopleth.
* See Aopendix 3 , Reference 16
** See SI ad*, 0. H. (ed.), "Meteorology and Atomic Energy." 'J. S. Atonic
Enerav'coranission, TIO 24190, Ju1y"l963; p. 130. Table 4-5 refers to
near-ecuivalent stability categories of the Brookhaven Trace ;yce System.
4-4
-------
4.5 SOURCE STRENGTH CONSIDERATIONS
4.5.1 Source Strengths Other Than 1 kg/sec.
The templates have been constructed on the basis of an assumed value
for Q, source strength, of 1 kg/sec, (about 4 tons per day). This emission
rate is typical of what has been observed or calculated for rocket fuel
spills. It is used here because no other estimates are available, and
because exclusion areas based on a more accurately known value of source
strength can readily be determined by means of the templates. If source
strength in a given accidental episode is known or estimated to be other
than 1 kg/sec., the shape of the appropriate template would not be altered
but the downwind distance would have to be adjusted accordingly. The
method is as follows:
1. From Table 4-1, read the Hazard Area Value for the chemical
of interest.
2. Divide this value by the actual emission rate 1n kg/sec.
Call this quotient the equivalent concentration.
3. Select the template for the actual stability condition.
4. Find the isopleth position for the equivalent concentration.
5. Trace the hazard area out to the isopleth of equivalent
concentration.
EXAMPLE:
Hydrogen sulfide is released at the rate of 3 kg/sec, uricer
Neutral Conditions. To what distance from the source does
the actual hazard area extend?
PROCEDURE:
1. From Table 4-1, the Hazard Area Value for Hydrogen Sulfide
1s 18 mg/m .
2. Divide that value by 3 (kg/sec), giving an equivalent concentra-
tion of 6 mg/m .
4-5
-------
3. On Figure 4-2, for Neutral Conditions, the 6 mg/m3 isopleth
is found by interpolation to lie directly downwind at about
the 4.0 km mark.
4. The hazard area is all that is enclosed by the template out
to 4.0 km.
NOTE: It is incorrect to estimate the actual area by applying the
factor of 3 to the downwind distance. !n the example, the
downwind distance for 1 ka/sac. is about 2.2 km, for 3 kg/sec.
about 4^0 km.
The relation between concentration and downwind distance is not
linear. Concentration is linearly related to source strength, Q, and is
inversely proportional to wind speed, u. If the values of these parameters
are different from those used in constructing the templates, the effect
on concentration must first be determined and the downwind distance then
located on the basis of where the limiting iscpleth would be drawn for
conditions as given. See other examples tnat follow.
4.5.2 Release Duration
From an assumed value of source strength it is possible to estir.ace
how long it will take for all the toxic material to flow out of its container.
For example, chlorine is shipped as a liquid under pressure in stsel con-
tainers. Cylinders normally hold 100 to 150 pounds, larger containers .";o'.d
2000 pounds, and single-unit tank cars hold 15, 30, and 55 tons. .Newer
tank cars that hold 35 tons and 90 tons are now in use as well. Chlorine
is also shiooed in -ultitank barges, uo to 1100 tons, anc in tank tr_c:
-------
4.5.3 Spillage of Volatile Liquids
Hazardous substances that are transported as liquids may have been
either liquid or gaseous at normal pressure and temperature. In the latter
case, the gases are liquefied and pressurized to several atmospheres within
their container. When vented to the atmosphere, pressurized liquid rapidly
returns to the gaseous state, carrying along aerosolized liquid particles
that vaporize in a few seconds. In such cases, the source strength may
be assumed to be 100* of the material emitted.
However, other hazardous substances that are liquid though volatile
at normal pressure and temperature may spill from their container to the
ground, and enter the atmosphere only through vaporization. In such cases
the source strength for the same quantity of spill as a liquefied gas will
be somewhat less than that of the gas by a factor that depends on its rate
of evaporation. The rate of evaporation is not constant; it depends upon
the temperature, pressure, and wind speed of interfacing ambient air, and
in the case of hygroscopic material like anhydrous ammonia, on the relative
humidity as well. It also depends upon its vapor pressure curve. Figure
4-4, adapted from Siewert*, shows the relation between source strength and
vapor pressure as determined for a variety of rocket fuels, some of which are
transported under cryogenic conditions. His original chart gives source
strength in pounds per second for a liquid spill that covers 500 square feet,
assuming a wind of 10 mph (4.3 m/sec.), air temperature of SC'F, and no
absorption or heating by the ground.
In the event of a volatile liquid spillage, vapor source strength,
Q, may be determined by the following procedure:
1. Estimate the rate of liquid spillage, expressed in kg/sec.
2. From Appendix A , Table A-5, read the vapor pressure at 20°C
for the chemical in question (last column).
*R. D. Siewert, personal communication, 1971
4-7
-------
VAPOR SOURCE STRENGTH @ 20 °C
% of Liquid Spillage Rate (g/sec)
?n 30 an in ?n inn.
VAPOR
T^LLljgn 50 PRESSURE
40
of
30 760 mm Hg.
Zff"-
10
Figure i-4. Vapor Source Strength, expressed as ;'a Percentaae
of Liquid Spillage Rate, shown as a function of
Vapor Pressure at 20 C (adapted from sleverc)
4-8
-------
3. Express the vapor pressure as a percentage of 760 mm. of Hg.
4. From Figure 4-4, find this -percentage along the ordinate,
move left to where this value intersects the curve, and read
the abscissa.
5. Determine vapor source strength from the percentage of liquid
spillage rate.
EXAMPLE:
A tank truck filled with acetone at ambient pressure is ruptured,
and the chemical spills to the ground at an estimated rate of
1 gal/sec. Specific gravity of acetone is about 0.8. What is
the vapor source strength, Q?
PROCEDURE:
1. Spillage rate of 1 gal/sec. = 3.8 liters/sec. = 2.8 x 0.8,
or about 3 kg/sec.
2. Table A-5 gives vapor pressure of acetone at 20°C as 175
mm Hg.
3. Acetone vapor pressure is therefore 175/750 = 22;J, for
ordinate value.
4. From Figure A-4, abscissa is 56%.
5. Vapor source strength, Q, is 56" of 3 kg/sec, or 1.7 kg/sec.
4.6 WIND SPEED CONSIDERATIONS
As previously noted, the templates are constructed on the basis of
a mean wind speed, u, of 3 m/sec. (7 mph). Should the reported wind be
other than 3 m/sec., the hazard area estimate is adjusted as follows: _
1. From Table 4-1, read the Hazard Area Value for the chemical of
interest.
2. Divide this value by 3.
4-9
-------
3. Multiply the quotient by the reported wind speed. Call this
product the equivalent concentration.
4. Select the template for the actual stability condition.
5. Find the isopleth position for the equivalent concentration.
6. Trace the hazard area out to the isopleth of equivalent
concentration.
EXAMPLE:
Anhydrous ammonia is released under Stable Conditions at the rate
of 1 kg/sec. The wind speed is measured as 2 rn/sec. To what
distance from the source does the actual hazard area extend?
PROCEDURE:
1. From Table 4-1, the Hazard Area Value for Anhydrous Amironia is
27 m
2. Dividing 27 by 3 gives 9.
3. Multiplying 9 by 2 gives 18 mg/m3, which is the equivalent
concentration.
4. The appropriate template is Figure 4-3.
5. The isopleth position for 13 mg/m3 is approximately 5.5 k,-
downwind.
5. .The hazard area is the total area inside the template betv/eer.
zero distance and 5.5 km.
NOTE: At 2 m/sec., the vaporous cloud will reach the 5.5 !
-------
1. Winds under 5 m/sec (11 to 12 mph) tend to follow paths
of least resistance: river channels, valleys, city avenues,
etc.
2. Winds are diverted by obstacles such as prominent hills, and
in cities by large buildings. With stronger winds that are
forced upward, around, and over these obstacles, reverse
eddies tend to appear an the lee side in which there are
zones of increased and decreased concentrations of airborne
contaminants. Hence, the lee side of a hill or building
may not necessarily provide temporary safety from a toxic
cloud.
3. If the wind travels upslope, gases that are significantly
heavier than air may not be transported as far as the template
shows. The pattern of concentration would be foreshortened,
and contamination in the valley would be heavier than indicated
by the template.
This correction may be required for all of the 15 gases on the
priority list with the exception of ammonia and hydrogen cyanide
4. If the wind travels downslope, cases that are significantly
lighter than air, chiefly armonia and to a lesser extent hydro-
gen cyanide, would appear in lower concentrations downwind
than shown by the template.
4.8 ADJUSTMENTS FOR FIRE
The fire hazard has not been considered in construction of the templates.
Many of the hazardous chemicals are highly flammable. In some cases ignition
caused by a spark or open flame some distance downwind may flash back to
the source, resulting in an initial explosion or fire of wide dimensions.
Atmospheric dispersion models that include this phenomenon would have to
consider the "stack effect" of the fire and the local atmospheric instability
condition that rapidly develops near the source. Once in progress the fire
tends to be concentrated at the source only, and the emanating cloud consists
4-11
-------
of combustion products that may or may not be toxic but are usually
suffocating or highly irritating. Evacuation of people from an area
farther than about 1 km downwind of the fire is seldom required.
4-12
-------
The disposal site is located in the southern section of the plant property.
The area in the vicinity of the plant property is heavily developed and
urbanized; however, the lined landfill is more than 1,000 ft from any plant
property boundary. Several settling basins and old disposal sites are
located northwest and north of the lined landfill with an additional small
sludge disposal area being located immediately east of Cell I.
The geologic materials underlying the area are primarily sands con-
taining varying amounts of silt and gravel. Discontinuous pockets and
lenses of clay are typically encountered especially south and southeast of
the disposal facility. Localized deposits of bog iron are also found.
Depth to ground-water generally ranges from 25 to 45 ft on the site with
perched water zones being locally encountered over discontinuous clay
lenses. Perched water conditions are reported south and southeast of the
lined landfill site with monitoring wells 0123 and 0107 located immedi-
ately south and east of Cell I reflecting such conditions. The direction
of ground-water flow is essentially to the east
The velocity of ground-water flow has been calculated to range from 1.0 to
1.35 ft per day with the permeablity of the underlying deposits ranging
from 1,000 to 1,400 gal/day/sq ft. Domestic wells utilized for drinking
water purposes are not known to be present in the developed areas to the
south-east of the plant property.
-------
DESIGN AND PERFORMANCE EVALUATION
In order to properly evaluate the performance of Cell II at the
it is necessary to make the evaluation in the
perspective of state-of-the-art technology in landfill design as well as
the actual field performance of similarly designed facilities. Therefore,
an analysis and evaluation of was made of the current state-of-the-art in
landfill technology as reflected in proposed state and federal hazardous
waste regulations pertaining to hazardous waste disposal. In addition, an
analysis and evaluation was made of the current performance of Cell I at
the disposal facility, since it represents the best
available example of true field performance for a design identical to that
of Cell II. The evaluation of Cell I would be expected to be much more
representative of the actual performance of Cell II rather than the extreme
test conditions imposed upon Cell II to date.
Performance of Cell I
The available data indicate that only very minimal seepage is cur-
rently occurring through the upper of the two liners of Cell I. Initial
problems were encounterd at Cell I with intrusion of rain water around the
manhole in Cell I; however, this problem has been resolved through repairs
to the manhole. Subsequent data show that considerable volumes of rain
water and leachate are being handled by the primary collection system with
only a very small quantity of leakage through the upper liner into the
secondary collection system. Flow measurements of approximately 3 gallons
per day have been documented. Problems were
evidently encountered initially in Cell I as a result of excessive seepage
around a leaking manhole. However, this problem was resolved in May of
1979 with the manhole being properly sealed.
-------
In order to assess the significance or insignificance of the current
rate of seepage into the secondary collection system, it becomes necessary
to evaluate the actual performance of the primary liner in terms of its
effective permeability. Although the leakage currently occurring is the
result of point sources in the liner material, the only practical approach
to evaluating effectiveness of the liner is to consider a uniform distribu-
tion in the seepage rate or permeability across the entire liner. Making
this assumption permits the usage of the following modified form of the
Darcy equation to evaluate the performance of the primary liner.
where:
QL = leakage through the liner in gpd
K1 = coefficient of vertical permeability of the liner, in
gpd/sq ft
Ah = difference between the head in the overlying collection bed
and the base of the liner (assuming the head in the lower
collection bed is below the base of the primary liner) in
feet
A = area of the liner through which leakage occurs, in square
feet
m = thickness of the liner through which leakage occurs, in
feet
The above equation will allow determination of the effective permeability,
K, of the primary liner based upon the known values of Q, A, and m.
However, the head, Ah, acting upon the liner is not known. In order to
derive an approximation of the potential head above the primary liner, we
-------
must first evaluate the quantity of discharge through the primary collec-
tion system as a result of seepage from the overlying sludge cake.
Data compiled during April and May, 1979, pertaining to the
quantity of discharge from the leachate pumping sump of Cell I provides
insight into the quantity of seepage of rain water from the sludge cake.
Although the quantities pumped from the leachate sump represent direct
runoff of rain water as well as seepage of rain water from the sludge cake,
the minimum quantity pumped during this time period can be assumed to
represent a "base flow" quantity indicative of seepage of the rain water
through the sludge. This quantity, which is approximately 1,400 gpd,
translates into an effective permeability of approximately 8 x 10"7 cm/sec
for the sludge cake. As a conservative measure, a seepage quantity of
3,000 gpd can be used to estimate the head in the primary collection system
by utilizing the following form of the Darcy equation:
Q = KIA (2)
where:
Q = quantity of seepage, in gpd
K = coefficient of permeability, in gpd per sq ft
I = hydraulic gradient, in ft/ft
A = area normal to the direction of flow, in square feet
The primary collection system was constructed with a 0.5 percent slope
which would represent the hydraulic gradient of the system. The permeabil-
ity of the sand comprising the primary collection system is assumed to have
a permeability of approximately 2,000 gpd/ft based upon the range of
permeability values reported for sands in this area*
This permeability value would also be expected for
-------
sands utilized in similar applications. The area normal to the direction
of flow would be the width of the primary collection system in Cell I,
approximately 400 ft, multiplied by the thickness, t, required to transmit
the observed quantity of rainwater seepage through the sludge cake. By
substituting the above values into Equation 2, the required thickness of
saturated sand for a seepage rate of 3,000 gpd is calculated to be approxi-
mately 0.75 ft. This thickness would be the expected head on the upper
liner once equilibrium has been established (Figure I).
Having determined the approximate head acting on the primary liner,
it is now possible to calculate the approximate effective permeability of
the primary liner utilizing Equation 1 and the measured flow rate of 3
gallons per day through the upper liner:
Q, = i^M
L m
Q,m
V I _ <-
K " AhA
= (0.4 ft3/day)(0.0025 ft)
(0.75 ft)(87,120 ft2)
1.53 x 10"8 ft/day
K1 = 5.39 x 10"12 cm/sec
Therefore, the permeability of the upper liner is determined to be
extremely low based upon the current rate of flow into the leak detection
manhole. In addition, once the cell has been properly capped and closed,
the rate of seepage from the sludge cake will decrease with time resulting
in a continued decline in head and subsequent seepage through the upper
liner.
A similar evaluation can be conducted for the lower collection and
liner systems to assess the potential for leakage through the lower
-------
en
Cb
i.i uui, Q . A\njvru ruoi . - ()-{)(
PVC LINERS
FIG. I . CELL I SECTION LOOKING SOUTH. PERFORMANCE OF CELL I UNDER FIELD CONDITIONS
-------
liner. Utilizing a coefficient of permeability equivalent to that of the
sand in the primary collection system and a seepage quantity of 3 gal/day
it can be determined that a potential head of only 0.00075 foot would be
exerted on the secondary liner. Such a small head on the secondary liner
would result in essentially no loss to the environment. The calculated
quantity of leakage from the entire cell based upon this small head value
~12
and a permeability of 5.39 x 10 cm/sec would be approximately 0.003! gpd
(Figure L).
An additional evaluation of liner performance can be made to obtain
some idea of the size of opening that might be required to transmit the
current leakage rate of 3 gal/day through the upper liner. If a singular
circular opening is considered as an orifice for flow, the following equa-
tion can be utilized to evaluate the approximate size of the opening or
orifice:
Q = 448.83 (CdA0/2g)(/R) (3)
where:
Q = discharge through the orifice, in gal/min
Cd = a discharge coefficient (dimensionless)
A = area of the orifice, in square feet
g = acceleration due to gravity, 32.2 ft/sec/sec
H = head above the orifice, in feet
The discharge coefficient typically ranges from 0.61 to 0.65 and for this
particular application a value of 0.61 will be used. If the discharge
through this opening is assumed to equal the total leakage of 3 gpd and the
head acting on the liner is the previously calculated value of 0.75 ft,
then the area of the opening or orifice would approximately 0.0000011 sq
-------
ft. This area would result in an orifice diameter of approximately
0.001183 ft or 0.0142 in. Such an opening with this total area would be
essentially impossible to detect and repair. In addition, this total area
would most likely be represented as several much smaller openings distri-
buted across the liner.
Analysis of Cell II
Since completion of Cell II, observations have been made and limited
tests conducted in order to evaluate the expected performance of this cell.
A tracer test utilizing MgSO^ was conducted in July, 1979 to verify leakage
through the upper liner. An additional leak test was performed in
December, 1979, and January, 1980 to compare changes in leakage rates
through the upper liner with water levels in the primary collection system.
Repairs have been made to several suspected problem areas in the upper
liner; however, the flow rate into the leakage detection manhole remains in
the range of 50 to 60 gpd.
The leak test performed on Cell II during December, 1979 and January,
1980 does show an expected increase in leakage through the upper liner in
response to increased head on the liner. However, the response is not that
which would be expected of a material exhibiting uniform permeability and
leakage rates. The fairly quick response to the increased head indicates
leakage occurring in close proximity to the sump; however, subsequent
fluctuations of water levels in the primary sump resulted in only more
subdued responses in the leakage rate. Even though water was periodically
removed from the primary sump, the rate of flow into the leak detection man-
hole did not drop below approximately 48 gpd. It should be noted that these
test conditions are considerably more severe than those to be expected during
-------
and following filling operations at Cell II since excessive heads would not
be permitted to build up. However, it would be worthwhile to evaluate the
performance of Cell II based upon this test data.
An approach similar to that taken for Cell I can be used to evaluate
the potential performance of Cell II based upon current test data. Using
the most conservative conditions consisting of the highest leakage rate of
110 gpd under a hydraulic head of 3 ft, Equation 1 can be utilized to
determine the effective permeability of the upper liner:
0 = K'AhA
wi m
K' = V
= (14.7 ft3/day)(0.0025 ft)
(3 ft)(87,120 ft2)
1.40 x 10"7 ft/day
K1 = 4.96 x 10"11 cm/sec
If a more average leakage rate of 60 gpd were used, the effective permea-
bility of the upper liner would be approximately 2.70 x 10"11 cm/sec.
Utilization of even the 60 gpd leakage rate for determination of liner
permeability is still considered to be excessive and not representative of
conditions to be encountered during operation of the cell.
Carrying the evaluation one step further to consider potential leak-
age through the lower liner requires determination of the expected head
occurring in the secondary collection system. This is accomplished by
using Equation 2 and values of 2,000 gpd/ft for permeability of the sand,
0.5 percent for the gradient of the system and 400 ft for the width of the
system. The head of water above the lower liner required for a seepage
rate of 110 gpd would be:
-------
Q = KIA
110 gpd = (2,000 gpd/ft2)(0.005)(400 x t)
t = 0.0275 ft
Therefore, an approximate head of 0.0275 ft would be exerted on the top of
the secondary liner (Figure II). Assuming an average leakage rate of
60 gpd would result in a head of 0.015 ft on the secondary liner. Using
Equation 1, leakage rates of 110 gpd and 60 gpd through the upper liner
could potentially result in leakage through the lower liner at rates of
approximately 1.0 gpd and 0.3 gpd, respectively (Figure II).
Assuming that under normal operating conditions the performance of
Cell II approaches that of Cell I and the head in the primary collection
system of Cell II declines to' a level of 0.75 ft, then the rate of leakage
through the upper liner would decline significantly. Using the previously
calculated permeability for the primary liner in Cell II, the quantity of
leakage would be:
n = K1 Ah A
L m
_ (2.70 x I0ncm/sec) (0.75) (87,120 ft2)
0.0025 ft
= 2.0 ft3/day
QL = 14.96 gpd
This constitutes a significant decrease from the average seepage quantity
of 60 gpd obtained during the liner tests. If this quantity of leakage is
used to determine the quantity of potential leakage through the secondary
liner, a potential head of approximately 0.0037 ft is calculated with a
resulting leakage rate of about 0.074 gpd.
Determination of the approximate size of a circular opening necessary
to transmit the maximum detected leakage rate through the primary liner of
10
-------
«=&
w
V .-
f
» * .
1 Foot
.'.'".. ' '.' ..'..-..'... ./
.-..'-': ''" '. ''''.'.;...- ''.'.'''.- -y?\
1;Foot ;./r;o276f^t- . Jpi-lW "..''.' '.'-'. .K:2obo gwl/tt2. -/ ;._^/^
jinn Fa**
/?
PVC LINERS
FIG. II . CELL II SECTION LOOKING SOUTH. PERFORMANCE OF CELL II UNDER EXTREME TEST CONDITIONS
-------
Cell II can be made by utilizing Equation 3 in a similar fashion as for
Cell I. Considering the maximum leakage rate of 110 gal/day and a hydrau-
lic head of 3 ft acting on the liner, the opening or orifice area is
calculated to be approximately 0.00002 sq ft. The diameter of a circular
opening of such an area would be approximately 0.005 ft or 0.060 in.
Again, such a very small opening would be essentially impossible to detect
and repair, especially since this total area of leakage would be comprised
of several smaller openings distributed across the liner.
Although the above data indicate that the performance of Cell II is
not currently equivalent to that of Cell I, it should be kept in mind that
leak tests were not performed on Cell I prior to initiation of disposal
operations, and that the test performance of Cell II would not be expected
to be a representative performance once the cell is placed into operation.
It is reasonable to assume that at ihe same stage of completion, Cell II
might well show an equivalent performance to that of Cell I. As the
filling operation progresses across the cell, the increased load on the
upper liner could result in continually decreased leakage through the
seams. In addition, much lower heads will be exerted on the upper liner
during and following filling of the cell which would also tend to decrease
seepage rates.
STATE-OF-THE-ART
The state-of-the-art in industrial waste landfills or lagoons appears
to hinge on the utilization of "impermeable" synthetic membranes or
natural soil mixtures having permeabilities of not greater than 10
cm/sec to prevent leachate migration from disposal facilities. Since the
disposal facility utilizes synthetic membranes for
11
-------
containment, an attempt was made to determine the performance record of
this accepted state-of-the-art technique. However, this was not an easy
task, since there are evidently very few facilities in the country that
have such a comprehensive containment and collection design, -
In a 1975 EPA publication assessing the use of liners for land
disposal sites, the following statement was made:
"Because the liners have been used in other applications to form
an impermeable structure, the landfill designers have assumed
that the materials can be used to construct impermeable land
disposal sites."
This assumption seems to have prevailed for the most part to the present
time. To date very little effort has been put forth to verify the effec-
tiveness of synthetic membranes under actual field conditions. Although
synthetic materials used for liners do have such low permeabilities as to
be considered essentially impermeable, the major problem in field applica-
tions is not the permeability of the material itself but the secondary
permeability resulting from inadequate seams, tears, and punctures occur-
ring during installation and operation. It is unreasonable to assume that
the structural integrity of synthetic liners can be absolutely maintained
or guaranteed during construction and operation of a waste disposal site
when so many variables can influence its ultimate performance.
A search conducted by AWARE, Inc. beyond its own experience in this
area revealed essentially no documentation on the field performance of
synthetic liners. Contact was also made with representatives of synthetic
liner manufacturers, liner installers, state solid waste offices, U.S. EPA
Atlanta and Washington solid waste offices, and the EPA Municipal Environ-
mental Research Laboratory. No existing or anticipated data were availa-
ble from any of these contacts concerning liner effectiveness. A fairly
12
-------
recent EPA publication entitled, "State-of-the Art Stqdy of Land Impound-
ment Techniques" (December, 1978) did involve an extensive literature
search on impoundment techniques as well as a survey of existing facilities
utilizing various types of lining materials. This report concluded that:
"The literature contains few meaningful engineering and perfor-
mance data on which to base an engineering analysis of lined land
impoundment sites that contain the industrial wastes of
concern."
No examples or references pertaining to actual field performance of lined
facilities were contained in the report. The only studies known to have
been conducted or on-going by the U.S. EPA involve the laboratory evalua-
tion of the performance of various lining materials when exposed to various
waste materials.
Actual documentation of leakage through a synthetic membrane liner in
an industrial waste disposal site has been reported,.
"Doe £y"-Prt*Ji has been provided with infor-
mation indicating a leakage rate of approximately 100 gpd through a
Hypalon liner at an industrial landfill site of similar design.
This particular disposal site is approximately five
acres in size. This leakage rate closely approximates the combined leakage
rates of Cells I and II, The
only other report of leakage in a double liner and collection system was
presented at a recent conference on hazardous waste disposal in Atlanta,
Georgia sponsored by the Chemical Manufacturers Association. At this
conference a representative of the Monsanto Company reported leakage
occurring through the primary liner at the Monsanto facility near Bridge-
port, New Jersey. However, the rate of leakage was not reported. It
should be noted that clay liners were used at this facility rather than
synthetic membranes.
13
-------
Actual state-of-the-art performance and design standards for indus-
trial waste disposal sites can be expected to be equivalent to existing and
proposed federal and state regulations governing the disposal of hazardous
wastes. In the December 18, 1978, Federal Register, the U.S. EPA proposed
regulations to control the storage, transportation, and disposal of
hazardous wastes. These proposed regulations contain design options for
waste disposal sites that are very similar to that of the
double liner and collection system. Based upon the design and
performance to date, it would appear that this facility
would meet the intent of these proposed regulations. However, since these
regulations are currently undergoing extensive modifications, it would be
of little use to make a detailed evaluation of t^-s facility based upon
the federal proposed regulations.
The proposed rules governing special waste facilities, including
chemical and hazardous wastes, for the State of New Jersey require similar
performance and design standards as those originally proposed by the U.S.
EPA. These proposes regulations are expected to be finalized and adoped
within the next few months. Proposed requirements for secure landfills
include a minimum of two bottom liners with a leachate collection system
above the primary liner and a leakage detection system between the primary
and secondary liners. The permeability of each in-place liners is also
required not to exceed 1 x 10" cm/sec at the maximum anticipated hydrosta-
tic head. Based upon the maximum head of 3 ft placed upon Cell II during
14
-------
its testing program, the leakage rate for a liner having a permeability of
1 x 10" cm/sec would be:
n K1 Ah A
gL m~~
(10"7 cm/sec) (3 ft) (87,120 ft2)
.0025 ft
= 29,634 ft3/day
QL = 221,666 gpd
This considerable quantity of leakage is a result of excessive hydraulic
gradients that can build up across the thin synthetic membranes. However,
even if it is assumed that a 5 ft clay liner of 1 x 10" cm/sec permeabil-
ity is subjected to the same 3 ft of head, the quantity of leakage would
be:
10"7 cm/sec) (8 ft) (87,120 ft2)
gl 5~7t
=39.5 ft3/day
QL = 295.5 gpd
Both Cells I and II . comply with the
proposed design standards and greatly exceed the required permeability for
the in-place liners. Permeability values determined for the in-place
liners in Cells I and II were less than 5 x 10 cm/sec. A review of all
requirements for secure landfills contained in Section 8.4 of the proposed
regulations indicates that the disposal facility, Cells I and II, would
meet or exceed all design and performance standards.
Potential Impact of Current Design
Based upon the current performance evaluations of Cells I and II, the
potential for adverse environmental impact as a result of leakage from the
15
-------
disposal facility would be minimal. The maximum discharge rate of
1 gal/day from the lower liner of Cell II based upon its performance under
extreme test conditions would be insignificant when considered in view of
the total quantity of groundwater moving beneath the cell. If any leakage
in the previously calculated quantities were to occur from the disposal
facility, it would be very difficult to detect a quantifiable impact on the
ground-water system. Even if larger rates of discharge were to occur at
some future time, the ground-water flow conditions in the vicinity of the site
are reported to be such as to minimize potential impact to the environment
and the public. The absence of ground-water supplies in the direction of
defined ground-water movement, and the flow of ground-water to the nearby
River are positive aspects of the disposal facility location.
CONCLUSIONS AND RECOMMENDATIONS
Based upon the available data and the preceeding evaluation of the
performance of Cell II at the disposal
facility, it is our opinion that the current design and operation does
represent state-of-the-art performance for waste disposal sites. The
double liner and collection system design employed at the facility are
matched by apparently very few other landfill sites in the country. In
fact, this limited utilization of such an extensive design has resulted in
the absence of comparative performance data for such facilities. The past
assumption that disposal facilities lined with "impermeable" synthetic
membranes are inherently impermeable or that they must be totally imper-
meable needs to be closely scrutinized by state and federal regulatory
agencies as well as private industry.
The calculations made for the liner evaluation indicate that
16
-------
extremely low effective permeabilities are being attained. These perme-
ability values are orders of magnitude lower than those required in the
proposed New Jersey Rules for Special Waste Facilities. It is our opinion
that the design and performance of Cell II of the
disposal facility would meet the intent of proposed state and federal
regulations pertaining to hazardous waste disposal facilities.
It is expected that the performance of Cell II would improve as
filling progresses in the cell as a result of increased loading and lower
hydraulic heads acting on the primary liner. This is of particular signi-
ficance in view of the fact that tests conducted in Cell II to date are
determined not to be representative of expected operating conditions for
the cell. The actual field performance of Cell II would be expected to
more closely approximate the performance currently observed for Cell I
rather than performance under test conditions.
17
-------
IV. OTHER DISPOSAL OPTIONS
-------
NOTES
DISPOSAL OPTIONS, INCINERATION
-------
THEORY AND DEFINITION OF TERMS
The following discussion presents a description of the principles
and concepts involved in incineracion of liquid waste. Because of the
preliminary nature of this evaluation, consideration of principles
of incineration is provided to aid in management decisions.
Combustion
Combustion is normally regarded as the oxidation of hydrocarbons
with release of heat. However, when considering liquid waste numerous
other compounds may be present which may be oxidized. Cvanide aas is
an example of an inorganic compound which is oxi.dized to water, carbon
dioxide and nitrogen.
Table I describes the tynes of data necessary for nroner design cf an
incinerator. If halogen, or sulohur comoounds and inorganic salts are
not contained in the liquid, the determination of combustion time,
temoerature and turbulence for comoletG oxidation of the organic com-
oounds is simolified. Obviously, once the incinerator has been con-
structed only tenioerature may be readily controlled while time and
turbulence can only be controlled with in a limited range based on the
design.
Control of Temperature
Incineration temperatures are normally controlled within a ranos
of 1800 to 20GO°F. Higher temperatures in the range of 2400°F require
special refractories.
Temocrature ,T,?.y be controlled by one o~~ four techniques which are
a. Excess ai r control
b. Radiant heat transfer
C. TwD-^t.nP:- rr.nhii-ti rn
-------
TABLE I
BASIC DATA CONSIDERATIONS
Type(s) of Waste
Ultimate Analysis
Ash Characteristics
Metals
Halogens
Heating Value
Solids
Liquids
Gases
Special Characteristics
Disposal Rates
Supply Conditions
Transportation
Liquid, solid, gas, or mixtures.
Carbon, hydrogen, oxygen and nitrogen,
water, sulfur and ash on an
"as-received" basis.
If appreciable and significant.
Calcium, sodium, copper, vanadium, etc.
Bromides, chlorides, fluorides.
BTU/lb on an "as-received" basis.
Size, form and quantity to be received.
Viscosity versus temperature, specific
gravity and impurities.
Density and impurities.
Toxicity end corrosiveness, other
unusual features.
Peak, average, minimum (present & future)
Temperature and pressure available.
Handling Requirements
-------
TABLE I
BASIC DATA CONSIDERATIONS
Typs(s) of Waste
Ultimate Analysis
Ash Characteristics
Metals
Halogens
Heating Value
Solids
Liquids
Gases
Special Characteristics
Disposal Rates
Supply Conditions
Liquid, solid, gas, or mixtures.
Carbon, hydrogen, oxygen and nitrogen,
water, sulfur and ash on an
"as-received" basis.
If appreciable and significant.
Calcium, sodium, copper, vanadium, etc.
Bromides, chlorides, fluorides.
BTU/lb on an "as-received" basis.
Size, form and quantity to be received.
Viscosity versus temperature, specific
gravity and impurities.
Density and impurities.
Toxicity and corrosiveness, other
unusual features.
Peak, average, minimum (present & future)
Temperature and pressure available.
-------
d. Direct heat transfer.
Control of Turbulence
Turbulence is normally controlled and established during the design
ohase of the incineration development. Generally, mechanical or
aero-dynamic means are employed in the mixing of air and fuel.' Measure-
ment of turbulence is difficult and may vary depending on the designer.
Thus, a detailed definition of the terms describing turbulence should
be provided in order to thoroughly identify the significance of the
design parameter.
Mechanical means for producing turbulence apply primarily to
incinerators used primarily for disposal of solid wastes and include
rotary kilns and moving grates.
Aero-dynamic turbulence is more often emoloyed in the destruction
of liquid waste utilizing high velocity jets converoent nozzles or*
air registers. Air registers are vein arranaments , usually surrounc'inp
a fuel injection nozzle. The veins r,ay be flexible or fixed and may
be adjusted to create an actual rotation of the combustion gases.
Time
Sufficient time must be provided within a combustion process to
allow slow burning particles or drcolets to completely burn before dis-
charge. Combustion chambers with heat releases between 20.000 and
60,000 BTU/cu ft-hr are common although design should be based on the
specific waste to be incinerated. If slow burning wastes are present
secondary combustion chambers may be required to assure complete combustion
-------
Liquid waste must be changed to a mist nrior to incineration-this
requires heat and may be accomolished by emoloying wicks or flame holders
to move the flame closer to the pool of liquid or by cascading the
liauid over an extended surface and increasing the valorization rate.
The most commonly used techninue is by atomizing the liruid in crooiets
Smaller than 40 microns. Atomization may be accomplished by emoloying
a rotary cup or pressure atomizer. The rotary cuo atomizer consists
of an open cup mounted en a hollow shaft. The ranidly soinninc CUD
and liquid film which is transferred through the hollow shaft creates
a thin film of liquid which is atomized centrifugall.y at the lio of the
CUD. High velocity air jets may be directed "axially around the cup
to create cone shaoed flames and increasing turbulence.
Pressure atomization may be achieved at moderate oressures of
between 100 and 150 lb/sq in. The disadvantages of this technique
include limited flow variation at low pressures and a tendency to olug
with solid materials in the waste. However, technology nas advanced
to the ooint v/here solids particles uo to a Quarter of an inch in diameter
may be oassed through a snecially design nozzles to satisfy licuid waste
incineration requirements.
Nozzles may imoince a combustion gas or air stream on the liquid
waste by either sonic or kinetic principles.
Corrosion
Numerous compounds react with refractory materials within incin-
erators resultinc in accelerated corrosion and thus increase maintenance
-------
costs. Substances which are noted to cause corrosion or other main-
tenance problems are listed below:
1. sodium
2. DOtassium
3. vanadium
4. calcium
5. zinc
6. phosohorus
7. i ron
8. cobalt
Of these compounds, sodium, ootassium, and vanadium are noted as the
most destructive. They attack refractory materials by reacting with
the alumina-silica refractories creating comoouncs whtch melt at low
temperatures or compounds which spall easily. When these corrosive
compounds are found, detailed evaluation of the reactions which
they may have with refractories must be made to prevent high main-
tenance ccs-s.
Emission Control
Perhaos the greatest limitation of incineration with resoect to
either solid or liouid waste disoosal is the nollutant emissions which
may result from combustion of otrrer than ornanic materials. Acids,
fluorides, caustic, sulphur oxides and metal oxides must be controlled
prior to emission of combustion cases to the atmosnhere. In certain
limited instances stacks mav be used to disnerse the waste products
-------
assuming that regulations, tonograohy, stack height, discharge velocity,
discharge temperature, and contaminant nature will not violate ambient
air quality standards. More often, solid oarticulates must be removed
from stack cases by various techniaues which include mechanical, scrubbing,
riltering and precipitation. For gaseous pollutants scrubbing
with water or other solutions may be required. As an examole removal
of hydrochloric acid (HC1) may be accomplished by water scrubbing in
packed rowers or submerged exhausts. Obviously, the reduction in the
temperature of the gases during scrubbing operations precludes the use
of a stack to disnerse other contaminants within the atmosphere.
Satisfactory removal of contaminants is achieved by creating additional
wastewaters which must then be treated by neutralization and oercipitation .
Because of the possible carryover of HC1 from the scrubbers high per-
formance rnist eliminators may be required to reduce discharge of con-
taminants .
ComDlete oxidation of organics including halogenated hydrocarbons
will result in the production of chlorine and fluorine as oroducts of
combustion. Chlorine which is relatively insoluble in water, .must com-
bine with hydrogen before they can be readily removed through scrubcina
processes. In many instances, insufficient hydrogen orecludes the
comolete conversion to a halogen acid and thus considerable Quantities
of the chlorine may be released to the atmosphere. The utilization of
auxiliary fuels may provide adequate hydrogen to completely convert the
halooens to the acid form.
-------
Metal oxides which are formed during combustion of many waste-
waters are usually of a submicron size and are difficult to control with
normal emission control equi orient. Only filters are canable of efficient
removal of submicron size oarticles. As a result of the difficulty of
removing the submicron size particles visible emissions are often observed
in the incineration of waste liauids containing inorganic salts. Sub-
merged exhaust and/or high energy scrubbers may also be emnloyed in the
removal of small sized oarticulates. However, the efficiency of these
techniques drops off raoidly with reduction in narticulate size in
the submicron range.
A mist eliminator may also be reauired to prevent the emission of
steam plumes created during combustion, evancraied cooling or stack gas
scrubbing.
As described above performance of emission control devices varies
deoenc'inc on the size and density of oarticulates which are to be
removed. Basically, cyclones which are normally emnloyed in the removal
of large particulates are effective with particulates greater than 10 microns
in size. Scrubbers are effective in removing particulates greater than
approximately 1 to 2 microns in size. Filter systems are effective
for submicron size particles. These limitations are dependent not
only on particulate size but also density and chemical reactions which
might occur during the combustion processes. Thus, pilot testing and
particulate size distribution analysis must be performed prior to
completion of design of emission control equipment.
-------
V. CASE HISTORY
------- |