WATER POLLUTION CONTROL RESEARCH SERIES
13030 ELY 05/71-12
REC-R2-7I-I2
DWR NO. 17-4-1!
BIO-ENGINEERING ASPECTS OF AGRICULTURAL. DRAINAGE
SAN JOAQUIN VALLEY, CALIFORN IA
DESALINATION OF AGRICULTURAL TILE DRAINAGE
MAY 1ST I
ENVIRONMENTAL PROTECTION AGENCY»RESEARCH AND MONITORING
-------
BIO-ENGINEERING ASPECTS OF AGRICULTURAL DRAINAGE
SAN JOAQUIN VALLEY. CALIFORNIA
The Bio-Engineering Aspects of Agricultural Drainage
reports describe the results of a unique interagency study
of the occurrence of nitrogen and nitrogen removal treat-
ment of subsurface agricultural wastewaters of the San
Joaquin Valley, California.
The three principal agencies involved in the study are
the Water Quality Office of the Environmental Protection
Agency, the United States Bureau of Reclamation, and the
California Department of Water Resources.
A triplicate abstract card sheet is included in the report
to facilitate information retrieval. Space is provided
on the card for the user's accession number and for
additional uniterms.
Inquiries pertaining to the Bio-Engineering Aspects of
Agricultural Drainage reports should be directed to the
author agency, but may be directed to any one of the three
principal agencies.
THE REPORTS
It is planned that a series of twelve reports will be
issued describing the results of the interagency study.
There will be a summary report covering all phases of
the study.
A group of four reports will be prepared on the phase of
the study related to predictions of subsurface agricul-
tural wastewater quality — one report by each of the
three agencies, and a summary of the three reports.
Another group of four reports will be prepared on the
treatment methods studied on the biostimulatory
testing of the treatment plant effluent. There will be
three basic reports and a summary of the three reports.
The other three planned reports will cover (1) techniques
to reduce nitrogen during transport or storage, (2) possi-
bilities for reducing nitrogen on the farm, and (3) this
report, "DESALINATION OF AGRICULTURAL TILE DRAINAGE".
-------
BIO-ENGINEERING ASPECTS OF AGRICULTURAL DRAINAGE
SAN JOAQUIN VALLEY, CALIFORNIA
DESALINATION
OF
AGRICULTURAL TILE DRAINAGE
Study Conducted by
Robert S. Kerr Water Research Center
Treatment and Control Research Program
Ada, Oklahoma
The agricultural drainage study was coordinated by:
Robert J. Pafford, Jr., Regional Director, Region 2
UNITED STATES BUREAU OF RECLAMATION
2800 Cottage Way, Sacramento, California 95825
Paul DeFalco, Jr., Regional Director, Pacific Southwest Regior.
WATER QUALITY OFFICE, ENVIRONMENTAL PROTECTION AGENCY
760 Market Street, San Francisco, California 94102
John R. Teerink, Deputy Director
CALIFORNIA DEPARTMENT OF WATER RESOURCES
1416 Ninth Street, Sacramento, California 95814
PROGRAM #13030 ELY
May 1971
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $1.00
-------
REVIEW NOTICE
This report has been reviewed by the
Water Quality Office of the Environmental
Protection Agency, U. S. Bureau of
Reclamation, and the California Depart-
ment of Water Resources and has been
approved for publication. Approval does
not signify that the contents necessarily
reflect the views and policies of the
reviewing agencies nor does mention of
trade names or commercial products con-
stitute endorsement or recommendation for
use by either of the reviewing agencies.
ii
-------
ABSTRACT
Investigations were made to determine the technical feasibility of
desalination of tile drainage. The source of the tile drainage was a
400-acre field near Firebaugh, California. Reverse Osmosis (RO) and
Electrodialy.sis (ED) processes were studied. Two RO membrane stacks were
investigated. The first, a high salt rejection, low product yield, was
operated on variable quality (3000-7000 mg/1 TDS) irrigation return
water. In the 7-month investigation period TDS removal efficiencies
decreased from 93 percent to 80 percent salt rejection and the product
flux decreased from 12 gal/ft^/day to less than 9 gal/ft^/day. The
20 mg/1 of nitrate-nitrogen and 8 mg/1 of boron contained in the influ-
ent were not effectively rejected. The second RO stack and the ED unit
were operated on return waters that were controlled to have a 3000 mg/1
TDS. The second RO stack was designed for a high product rate and low
salt rejection. The TDS removal remained at 85 percent for a 3-month
run. Product flux decreased from over 19 gal/ft2/day to less than 12
gal/ft^/day. Nitrate and boron rejection was low. The ED data are
based on a single pass through the membrane stack. The TDS removal
varied from 35 percent to 15 percent. The nitrate removal rate was
greater than the TDS removal. Boron removal was negligible. It is
estimated that the costs for the two processes are approximately equal—
$320 per million gallons of product.
This report was submitted in partial fulfillment of Project No.
13030 ELY under sponsorship of the Environmental Protection Agency.
ill
-------
BACKGROUND
This report is one of a series which presents the findings of intensive
interagency investigations of practical means to control the nitrate
concentration in subsurface agricultural wastewater prior to its
discharge into other water. The primary participants in the program
are the Environmental Protection Agency, the United States Bureau of
Reclamation, and the California Department of Water Resources, but
several other agencies also are cooperating in the program. These three
agencies initiated the program because they are responsible for providing
a system for disposing of subsurface agricultural wastewater from the
San Joaquin Valley of California and protecting water quality in
California's water bodies. Other agencies cooperated in the program by
providing particular knowledge pertaining to specific parts of the overall
task.
The ultimate need to provide subsurface drainage for large areas of
agricultural land in the western and southern San Joaquin Valley has
been recognized for some time. In 1954, the Bureau of Reclamation
included a drain in its feasibility report of the San Luis Unit. In
1957, the California Department of Water Resources initiated an investi-
gation to assess the extent of salinity and high ground-water problems
and to develop plans for drainage and export facilities. The Burns-Porter
Act, in 1960, authorized San Joaquin Valley drainage facilities as a part
of the California Water Plan.
The authorizing legislation for the San Luis Unit of the Bureau of Recla-
mation's Central Valley Project, Public Law 86-488, passed in June 1960,
included drainage facilities to serve project lands. This Act required
that the Secretary of Interior either provide for constructing the San
Luis Drain to the Delta or receive satisfactory assurance that the State
of California would provide a master drain for the San Joaquin Valley
that would adequately serve the San Luis Unit.
Investigations by the Bureau of Reclamation and the Department of Water
Resources revealed that serious drainage problems already exist and that
areas requiring subsurface drainage would probably exceed 1,000,000 acres
by the year 2020. Disposal of the drainage into the Sacramento-San Joaquin
Delta near Antioch, California, was found to be the least costly alternative
plan.
Preliminary data indicated the drainage water would be relatively high
in nitrogen. The Environmental Protection Agency conducted a study to
determine the effect of discharging such drainage water on the quality of
water in the San Francisco Bay and Delta. Upon completion of this study
in 1967, the Agency's report concluded that the nitrogen content of
untreated drainage waters could have significant adverse effects upon the
fish and recreation values of the receiving waters. The report recommended
a three-year research program to establish the economic feasibility of
nitrate-nitrogen removal.
-------
As a consequence, the three agencies formed the Interagency Agricultural
Wastewater Study Group and developed a three-year cooperative research
program which assigned specific areas of responsibility to each of the
nitrogen conditions in the potential drainage areas, possible control of
nitrates at the source, prediction of drainage quality, changes in nitrogen
in transit, and methods of nitrogen removal from drain waters including
biological-chemical processes and desalination.
VI
-------
CONTENTS
SECTION PAGE
Abstract iii
Background v
Contents vii
Figures viii
Tables ix
I Summary and Conclusions 1
II Introduction 3
Theory of Process Operations 3
III Experimental Equipment and Procedures 7
Reverse Osmosis Unit 7
Electrodialysis Unit 10
Feed Water Blending System 10
Influent Water Quality 12
Sampling and Analytical Analyses 14
IV Results and Discussion 15
Reverse Osmosis 15
Electrodialysis 22
Cost Comparison for Reverse Osmosis 23
and Electrodialysis
V Acknowledgements 27
VI References 29
VII Publications 31
vii
-------
FIGURES
PAGE
1 The Principle of Reverse Osmosis 4
2 The Principle of Electrodialysis 6
3 Aerojet General Reverse Osmosis Water Purifier 8
Model 1-560B-1
4 Aerojet General Reverse Osmosis Unit Model 1-560-B-l 8
Flow Schematic
5 Water Flow Through Circular Desalination Plate 9
6 Ionics, Inc. Electrodialysis Unit Model 300-B-3
Membrane Stack 11
7 Ionics Electrodialysis Flow Schematic 12
8 Schematic of Blending System 13
9 Reverse Osmosis Data - Stack I 16
Total Dissolved Solids and Per Cent Removal
of TDS Versus Time
10 Reverse Osmosis Data Stack I Product Flux Versus Time 16
11 Reverse Osmosis Data Stack II Total Dissolved 19
Solids vs Time
12 Reverse Osmosis Data Stack II Product Flux 21
Versus Time
13 Electrodialysis Data Product Total Dissolved Solids 23
and Per Cent Removal of TDS Versus Time
viii
-------
TABLES
NO. PAGE
1 Characteristics of Tile Drainage Used at the IAWTC 13
and Average Mineral Concentrations of Irrigation
Waters
2 Reverse Osmosis Stack I Mineral Analysis 17
3 Reverse Osmosis Data - Stack I Product/Brine
Production 18
4 Reverse Osmosis Stack II - Mineral Analysis 20
5 Electrodialysis Data Mineral Analysis 24
6 Cost Analysis Summary for Reverse Osmosis and 25
Electrodialysis Units
ix
-------
SECTION I
SUMMARY AND CONCLUSIONS
Two reverse osmosis (RO) membrane stacks and an electrodialysis (ED)
unit were operated on agricultural tile drainage waters.
The initial RO stack received an influent with varying total dissolved
solids (TDS) content which ranged from approximately 3000 mg/1 to 7000
mg/1. When operated at peak performance, this set of membranes was
able to remove over 90 percent of the influent TDS; however, nitrate
and boron removal averaged less than 27 percent. The product recovery
for this stack averaged approximately 37 percent. The second RO stack
used was able to achieve 85 percent removal of a constant 3000 mg/1 TDS
influent, but nitrate and boron removal was negligible. Although low
nit:rate rejection was characteristic of the reverse osmosis membranes
used for these experiments, it is believed that reverse osmosis units
using current technology, especially new membranes, could be expected
to achieve a higher nitrate rejection than reported herein. The pro-
duct recovery for this stack averaged approximately 40 percent; however,
this lower-than-expected recovery was caused by internal damage to the
reverse osmosis stack.
The electrodialysis unit had an. average TDS removal of 23 percent, with
a maximum removal of 36 percent, when supplied with a 3000 mg/1 TDS
influent, based on a single pass through the membrane stack. Although
the ED unit did not remove boron at any time, nitrate was removed at
a rate averaging 1.98 times that calculated for TDS removal. Product
recovery based on a single pass remained constant at 75 percent.
Economic data taken from literature indicate that water produced by
both RO and ED costs approximately $0.32 per thousand gallons.
It was concluded from the experimental data that desalination of San
Joaquin Valley subsurface agricultural return flow is technically
feasible. However, it does not appear at the present time that direct
reuse of the water as an irrigation source is economically possible.
-1-
-------
SECTION II
INTRODUCTION
In October 1966, the Office of Saline Water and the Federal Water
Quality Administration (now the Environmental Protection Agency) began
cooperative research into the use of desalination processes to remove
salt from subsurface agricultural wastewaters (tile drainage). This
program was conducted from June 1967 to September 1968 in conjunction with
the cooperative nitrogen removal studies being performed at the Interagency
Agricultural Wastewater Treatment Center (IAWTC) near Firebaugh, California.
The primary objective was to determine the technical feasibility of
desalinating the drainage waters. Secondary objectives were to produce
desalting cost estimates and to provide performance data on the removal
of nitrate and boron from the wastewaters.
The information provided by these experiments can be used to consider
reuse of the water for agriculture and to compare reclamation costs and
costs of future imports of water into the San Joaquin Valley. Of par-
ticular interest when considering reuse is the removal of boron. Tile
drainage in the western portion of the San Joaquin Valley typically has
relatively high concentrations of this element. The removal of nitrogen
from the wastewater was also studied.
Theory of Process Operations
Reverse osmosis and electrodialysis were the two desalination processes
applied in this study. Both processes accomplished essentially the
same result; however, theories of their operation differ. A discussion
of their characteristics follows.
Reverse Osmosis
When two solutions of differing concentrations (but with a common
solvent) are separated by a semipermeable membrane.!/, solvent will
flow from the weaker solution to the more concentrated solution.
The process is known as osmosis and is illustrated in Figure 1.
Osmotic pressure (p) is a measure of the difference between the
diffusion pressure of solvent (water) molecules in the two solu-
tions. Solvent tends to flow from an area of high diffusion
pressure to an area of low diffusion pressure until equilibrium
is established. If a pressure (P) greater than osmotic pressure (p)
is applied to the more concentrated solution side of the membrane,
solvent (water) is forced through the membrane in a direction
opposite to normal osmotic flow. This is the reverse osmosis
I/ A membrane more permeable to solvent than to solute molecules
is said to be differentially permeable or semipermeable.
-3-
-------
SEMI PERMEABLE
MEMBRANE
•:;X;X;X*X;X;X;XvX'X
NORMAL OSMOSIS
1
P
i
OSMOTIC
PRESSURE
=5L£
IE
OSMOTIC EQUILIBRIUM
^APPLIED PRESSURE (P)
REVERSE OSMOSIS
FIGURE I -THE PRINCIPLE OF REVERSE OSMOSIS
-4-
-------
phenomenon used in desalination. In theory, the flow of water through the
membrane is approximately proportional to the net pressure (applied
pressure minus osmotic pressure), and the amount of salt passing through
a less than ideal semipermeable membrane depends primarily on the gradient
in salt concentration between the two solutions. In practice, however,
flow limitations have been shown to result from osmotic pressure increases
due to concentration build-up in the liquid boundary layer on the brine
side of the membrane, thereby decreasing the net pressure, and from salt
precipitation or other deposits onto the membrane surface (1). The salt
precipitation is an operational problem. Depending on the composition
of the brine water, precautions such as pH control and chemical additions,
may be necessary to prevent the accumulation of 'precipitate onto the
membranes.
Electrodialysis
The removal of ions from a saline solution by electrodialysis depends
on the basic principle that positively and negatively charged poles attract.
Therefore, if a direct current potential is applied across a solution of
salt in water by means of two electrodes inserted in the solution, the
cations will be attracted toward the cathode and the anions will be
attracted toward the anode. This movement of ions shown in Figure 2 can
be used to advantage if ion selective membranes were so placed that they
isolate a purified zone from which the ions had been removed. For this
purpose, cation and anion permselective membranes were developed, each
membrane allowing only cations or anions to pass through respectively.
Use of these membranes to form watertight compartments in a salt solution
and the electrical potential will result in a demineralized central com-
partment. As in reverse osmosis, an over-concentration of salts in the
compartments receiving the ions will lead to precipitation of salts onto
and possibly in the membrane. Precautions were taken during operation to
minimize such occurrences.
-5-
-------
A-ANION PERMEABLE MEMBRANE
C-CATION PERMEABLE MEMBRANE
Cl"
©
Na
cf
1234
NO ELECTRIC FLOW
-o—o-
c Ci
cr
cf
i \
C
1234
ELECTRIC FLOW
FIGURE 2- THE PRINCIPLE OF ELECTRODIALYSIS
-6-
-------
SECTION III
EXPERIMENTAL EQUIPMENT AND PROCEDURES
The two units used to demonstrate these processes were an Aerojet-General
Reverse Osmosis Water Purifier (Model 1-560B-1) and an Ionics Electrodialysis
Demineralizer (Model 300-B-3). The features of these units, the quality of
the irrigation return waters, and the methods of evaluating the processes
are described in this section.
Reverse Osmosis Unit
The main components of the unit were a pressure vessel, a control console,
a high-pressure positive displacement pump, and two banks of nine 5-micron
cartridge filters (Figures 3 and 4). Within the pressure vessel were
600 square feet of semipermeable cellulose acetate membranes. The membrane
and support plate combinations were divided into five modules, which
composed the reverse osmosis membrane stack. The membrane and support
plate were fabricated to separate the feed water into a product stream
and a brine or waste stream (Figure 5). The control console consisted of
an instrumentation panel, a pump control panel, and an externally mounted
pressure switch. The instrumentation panel included the following: pump
suction line pressure indicator, pump discharge line pressure indicator,
feed water temperature indicator, product water conductivity indicator, and
a differential pressure indicator to monitor the difference between the
pressure vessel inlet and outlet pressure. Flows were monitored by flow
meters installed on the effluent lines of the product and brine streams.
The maximum feed flow possible by means of the positive displacement pump
was 13 gpm. However, this flow was varied, depending on the recovery
desired. Percent recovery was increased by reducing the quantity of
influent and maintaining a constant quantity of product effluent.
Because of the ionic makeup of the tile drainage, there existed a
distinct possibility that either calcium carbonate or calcium sulfate,
would precipitate onto the membranes. To prevent this, the tile drainage
was pretreated by controlling the pH to prevent calcium carbonate scaling
and by adding a chemical to inhibit precipitation of calcium sulfate. The
pH of the influent irrigation return waters was lowered from approximately
7.4 to a range of 5.5 to 5.8 by the injection of concentrated sulfuric
acid. The chemical inhibitor was Cynamer P-35 —' , which was injected at
an average rate of 3 mg/1. Both chemicals were added by positive dis-
placement pumps.
Two membrane stacks were tested at the Center. The first was manufactured
to have a "tight" membrane. This stack (Stack I) was operated at 750
pounds per square inch gage pressure (psig). Its main characteristics
j./ Manufactured by the American Cynamid Company
-7-
-------
FIGURE 3-AEROJET GENERAL REVERSE OSMOSIS WATER PURIFIER
MODEL I-560B-I
INFLUENT
FEED PUMP-
ACCUMULATOR
POSITIVE
DISPLACEMENT
PUMP
F
CARTRIDGE
FILTER
UNITS
BACK PRESSURE
CONTROL VALVE
FIGURE 4 - AEROJET GENERAL REVERSE OSMOSIS UNIT - MODEL I-560-B-I
FLOW SCHEMATIC
-8-
-------
PRODUCT COLLECTION GROOVES
PRODUCT SHAFT
CONCENTRATED FEED OUTLET
SPIRAL BAFFLE
PRODUCT FLOW CHANNEL
SUPPORT PLATE
SUBSTRATE & MEMBRANE
FEED INLET
FIGURE 5 -WATER FLOW THROUGH CIRCULAR DESALINATION PLATE
-------
were a low flux and high salt rejection capacity. The second (Stack II)
was designed to have a "loose" membrane. It was operated at an applied
pressure of 350 psig. The membrane was characterized by a relatively high
flux and a low salt rejection ability.
Electrodialysis Unit
The main feature of the Ionics* Electrodialysis (ED) unit was the
electrical membrane stack (Figure 6) which consisted of 150 membrane-
spacer combinations. Each combination consisted of an anion and cation
membrane separated by water flow guides. The water flow guides separate
the water flow through the stack into two main streams, the dilute
(product) and brine streams. A pressure differential of approximately
1 psig was maintained between the two streams. The brine stream was
confined under the lower pressure. Thus, if a membrane became damaged
water could only flow from the dilute to the brine, thereby removing
the possibility that the dilute stream could become contaminated.
Other features of the unit were a control panel to monitor pump
pressures, stack pressure differential, and product conductivity, and a
rectifier that included a voltmeter and an ammeter to monitor voltage
and amperage across the membrane stack. A constant electromotive force
of 275 volts was maintained across the stack for all experiments; the
amperage varied, depending on feed salinity, water temperature, and
membrane conditions.
Basically, the unit operates by separating the water flow through the
unit into the two streams previously mentioned. A flow schematic
showing the routing of these streams is presented in Figure 7. The dilute
stream is formed by water from which the salt ions are removed as it passes
through the membrane stack at the rate of 25 gpm. Ions removed from the
dilute stream were collected by a recirculating brine stream. Feed water
was constantly added to the brine stream so that the ion concentrations
remained below their saturation limit. This addition was referred to as the
brine stream blow down.
Precautions were necessary to avoid calcium carbonate and/or calcium
sulfate scaling on the brine side of the membranes. Calcium carbonate
scaling was prevented by injecting a 20 percent sulfuric acid solution to
keep the brine stream pH in a range of 5.5 to 5.8. Calcium sulfate
scaling was prevented by using a brine stream blow down rate of 6 to 8
gpm which was sufficient to prevent saturation of calcium and sulfate ions.
However, had it been necessary, a chemical inhibitor similar to that used
in the RO could have been injected.
Feed Water Blending System
The tile drainage that came directly from the field was used as feed water
for evaluation of Stack I of the RO and preliminary testing of the ED unit.
.10-
-------
FIGURE 6-IONICS, INC. ELECTRODIALYSIS UNIT
MODEL 300-B-3 MEMBRANE STACK
REPRODUCED WITH THE
PERMISSION OF IONICS, INC.
-11-
-------
The salinity of these waters varied seasonally from 2500 to 7600 mg/1
total dissolved solids (TDS). The variations in influent TDS made
evaluations of the units difficult; therefore, a constant salinity feed
was provided by installing a blending system which mixed water from a
source having low TDS with the tile drainage to provide a constant salinity
of 3000 mg/1 TDS. A schematic of this system is shown in Tigure 8.
Suspended solids were removed from the low salinity water before it was
mixed with the tile drainage by the use of a pressure sand filter and a
diatomaceous earth filter. In actual practice, the sand filter was usually
bypassed because it caused pressure fluctuations that affected the per-
formance of the diatomaceous earth filter. Varying flow rates of blended
water were used, depending on the TDS concentration of the tile drainage.
Influent Water Quality
The water used for the experimental work was taken from a tile drainage
system that serviced a 400-acre field. The land was used primarily to
cultivate rice during the summer and to grow barley during the winter.
The amount of irrigation water needed by these two crops differs greatly.
This difference causes a seasonal change in tile flow and in concentration
of the dissolved minerals. The lower concentrations occur in the summer.
The mineral concentration ranges of the tile drainage used at the Inter-
agency Agricultural Wastewater Treatment Center are listed in Table 1.
FEED
STORAGE
TANK
t
fjst _ PRODUCT
JL, TRANSFER
FILTER hY PUMP
ELECTRC
ISO CEL
PAIR
nr
LJ-
_| * P
-------
DIATOMACEOUS
EARTH
FILTER
DIATOMACEOUS
EARTH SLURRY
TANK
'ASTE
TO DESALINATION UNITS
FIGURE 8- SCHEMATIC OF BLENDING SYSTEM
TABLE 1
CHARACTERISTICS OF TILE DRAINAGE USED AT THE IAWTC
AND AVERAGE MINERAL CONCENTRATIONS OF IRRIGATION WATERS
CONSTITUENT
RANGE OF MINERAL
CONCENTRATIONS IN
TILE DRAINAGE
mg/1
AVERAGE MINERAL
CONCENTRATION OF
IRRIGATION WATER
me/1
Bicarbonate
Boron
Calcium
Chloride
Magnesium
Nitrogen
Phosphate
Potassium
Sodium
Sulfate
Pesticides (CHC)
Total Dissolved Solids
5 Day BOD
COD
Dissolved Oxygen
280-330
4-15
160-390
310-640
70-230
5-25
0.13-0.33
4-11
620-2050
1500-3900
0.001
2500-7600
1-3
10-20
7-9
90
0.3
20
60
10
1
0.5
3
50
65
300
-13-
-------
The nitrogen in the tile drainage was essentially in the form of nitrate.
Nitrite-nitrogen was not detected in concentrations greater than 0.1 mg/1;
organic nitrogen was generally constant at a concentration of approximately
0.4 mg/1; and ammonia nitrogen concentrations were usually not detectable.
A characteristic of this water is the high concentration of sulfate,
which is the dominant anion. A high concentration of sulfate in conjunc-
tion with calcium and carbonate make the pretreatment previously described
a necessity. Another characteristic of the San Joaquin Valley tile
drainage is the high concentration of boron, an element which is toxic to
many crops in concentrations over 1 mg/1.
Sampling and Analytical Analyses
All operational parameters (pump pressures, electrical conductivity, pH,
etc.) for both units were monitored at four-hour intervals. Electrical
conductivity was monitored by a Wheatstone Bridge and conductivity cell; pH
was monitored with a glass electrode. A correlation between electrical
conductivity and total dissolved solids was determined on a weekly basis.
The following ions were determined periodically; calcium, magnesium,
sodium, potassium, boron, sulfate, carbonate, chloride, nitrate, total
iron, silica, and total alkalinity. All analyses were determined by the
procedures described in "Standard Methods for the Examination of Water and
Wastewaters" (2).
-14-
-------
SECTION IV
RESULTS AND DISCUSSION
The overall experiments with the two desalination units were limited
to unit performance, with little emphasis given to process cost or
brine disposal methods.
Reverse Osmosis
The first part of this section deals with the performance of the initial
RO stack and the second part describes the performance of the second RO
stack that received the blended 3000 mg/1 TDS water. Cost evaluation of
operating the reverse osmosis unit was performed only on the second stack.
Reverse Osmosis - Stack I
This reverse osmosis stack was operated from early June 1967 to mid-
January 1968. The first two months of operation were mainly devoted
to familiarization and correction of operational problems. These problems,
which were mostly mechanical, did not appear to greatly affect the initial
performance of the unit.
TDS Removal. The unit received an influent TDS concentration that varied
from less than 3000 mg/1 to approximately 7000 mg/1. For the first three
months the membranes produced a product containing less than 500 mg/1 of
TDS with an average salt rejection of more than 90 percent. The changes
in influent and product TDS and also the percent of TDS removed are
illustrated in Figure 9. Typical analyses of the minerals in the unit's
influent, product, and brine stream are presented in Table 2.
During the unit's operation, the product salinity fluctuated with the
influent salinity. Such fluctuation is normal because of increased
passage of dissolved minerals through the membrane by pore transport.
However, the continued decrease in product quality after October 1967
shown in Figure 9 was attributed to biological fouling and subsequent
membrane deterioration. Biological degradation of the membranes may have
begun when operations were suspended for extended periods of maintenance.
Variations in Membrane Flux. The flux for this stack reached its highest
point, 13 gallons per square foot of membrane per day (Figure 10), on the
first day of operation and then generally declined throughout the experi-
mental period. Only in October 1967, when a higher-than-normal pressure
was applied, did a significant rise in flux again occur. The flux
decreased partly because a rise in influent TDS caused the osmotic
pressure to increase. However, the degree to which flux dropped is
probably due in part to another factor. It is known that under the
-15-
-------
800O]
—
E
, 6000-
>
o
o
UJ 4000-
O
M
W
5
£ 2000^
p
-
^^g- PER CENT TDS REMOVAL
/ ~~~~— -**£?*
' / ~~^*V*rtX» /_j-INFLUENT
^\^ /
"*•*•>. i
~~^~^-^^ /
^^
-^"^
^•PRODUCT . ^^
JUNE ' JULY ' AU6 ' SEPT ' OCT ' NOV ' DEC " JAN
1967 1968
-100
-
H
- 80 2
u
UJ
0.
i
M
-60-
u.
O
^
-4o|
UI
K
-
FIGURE 9-REVERSE OSMOSIS DATA - STACK I
TOTAL DISSOLVED SOLIDS AND PER CENT REMOVAL OF TDS VERSUS TIME
1
.60
UJ
ac.
m
I4O
i
K
UJ
HIO.O
3
O
0.
u.
o
w 8.0
o
6.0
JUNE
JULY
AUG
OCT
SEPT
1967
FIGURE 10 - REVERSE OSMOSIS DATA
STACK I
PRODUCT FLUX VERSUS TIME
NOV
DEC
JAN
1968
-16-
-------
TABLE 2
REVERSE OSMOSIS
STACK I
MINERAL ANALYSIS
Operating Conditions
Date of Sample - - -
Operating Pressure, psig
pH of feed water - - - -
Flow Rates
GPM
Influent
Product
Brine
July 14, 1967
750
5.5
12.5
4.2
8.3
CONCENTRATION - mg/1
CONSTITUENT
Calcium
Magnesium
Sodium
Potassium
Boron
Sulfate
Bicarbonate
Chloride
Nitrate
Iron (total)
Silicon
Total Alkalinity
TDS
pH*
INFLUENT
230
111
862
3.4
8.2
2360
98
314
34
0.02
42
80
4230
6.4
PRODUCT
3.3
1.7
82
0.0
6.1
28
16
94
29
0.02
13
13
304
5.4
BRINE
347
215
1350
5.1
8.4
3720
140
440
37
0.02
58
136
6400
6.5
PERCENT
REJECTION
99
98
90
100
25
99
84
70
15
0
69
84
93
—
* pH of samples at time of analysis.
normally high operating pressures, reverse osmosis membranes do show
a decreasing flux, due to what is usually called "compaction". This
is most likely the main reason for the observed drop in flux.
Product Recovery. Of particular importance in any desalination experiment
is the amount of brine that must ultimately be disposed of. Depending on
geographical location, the cost of such disposal may equal the cost of
producing the desired product water. Table 3 summarizes the product
recovery achieved with this stack. The percent product recovery was
controlled by varying the quantity of influent of the unit. Therefore,
-17-
-------
TABLE 3
REVERSE OSMOSIS DATA - STACK I
PRODUCT/BRINE PRODUCTION
MONTH
PRODUCT FLOW
GPM
PRODUCT RECOVERY
PERCENT
BRINE/PRODUCT
RATIO
June
July
August
September
October
November
December
January
Average
4.1
3.9
3.8
3.4
3.8
3.4
3.3
3.2
3.6
37
37
37
37
33
36
38
40
37
1.7
1.7
1.7
1.7
2.0
1.8
1.6
1.5
although the amount of product generally decreased, the percent recovery
was maintained by decreasing the total flow through the unit. This
particular stack produced from 1.5 to 2 times as much brine as product.
Nitrate and Boron Removal. The nitrate removal of this stack varied from
0 to 49 percent, with an average of 27 percent. The boron removal ranged
from 8 to 35 percent, with an average of 21 percent. Both ions were
removed at substantially lower percentage than the general TDS removal.
In a separate experiment conducted by Aerojet-General with the same type
of membranes, the rejection of nitrate was shown to be highly dependent
on the presence or absence of the sulfate and/or chloride ion (3). In a
solution containing 2000 mg/1 calcium sulfate, 2000 mg/1 sodium chloride,
and 50 mg/1 sodium nitrate, approximately 82 percent of the nitrate and
approximately 98 percent of the calcium sulfate and sodium chloride were
rejected. In a second solution containing 2000 mg/1 calciam sulfate and
50 mg/1 sodium nitrate less than 20 percent of the nitrate was rejected.
In a third solution containing approximately 500 mg/1 sodium chloride and
50 mg/1 sodium nitrate, both nitrate and chloride were rejected at
approximately a 97 percent level. These experiments demonstrated that it
is possible to observe a range of nitrate permeations, depending upon the
-18-
-------
presence of sulfate or other highly rejected anions. The results from
the second solution agreed well with the results obtained from the stack
operated on strictly tile drainage.
Reverse Osmosis - Stack II
The second stack of membranes used in the experiments was operated from
July to September 1968 for a total of approximately 10 weeks. The
operation was terminated because an error in factory assembly increased
the differential pressure through the membrane stack. Inspection of the
stack showed that an aluminum washer had disintegrated and blocked the
brine stream flow paths. This shortened the expected life of the membranes
considerably.
TDS Removal. This stack was operated with the TDS blending system
previously described. It received an influent having a more constant
level of TDS concentration at approximately 3000 mg/1 (Figure 11). The
o»
6
o
V)
o
UJ
o
(O
<
o
TVW
Tlf\f\f\_
3UOU-
2000-
1000-
O-
v^ _
5- IN FLUENT
^^^.
^ ^^
' • ^
<-PRODUCT
f
JULY AUG SEPT
1968
FIGURE II-REVERSE OSMOSIS DATA-STACK TJ
TOTAL DISSOLVED SOLIDS vs TIME
-19-
-------
unit's salt rejection remained constant at 85 percent which produced a
product containing a IDS of between 400 and 600 mg/1. The only variation
in product T-DS occurred as the influent IDS concentration varied; thus,
the minor increase in product TDS could be increased salt transport through
the membrane. A typical influent product and brine stream analysis for
this stack is presented in Table 4. Due to the lower salt rejection
capability of the membranes in this stack, its nitrate and boron removals
were essentially zero.
TABLE 4
REVERSE OSMOSIS
STACK II
MINERAL ANALYSIS
Operating Conditions
Operating Pressure
pH of feed water -
Flow Rates
GPM
CONSTITUENT
Calcium
Magnesium
Sodium
Potassium
Boron
Sulfate
Bicarbonate
Chloride
Nitrate
Iron (total)
Silicon
Total Alkalinity
TDS
pH*
, psig -
Influent
JTITOQUCL
Brine
FEED
157
85
665
4.3
7.0
1650
55
320
22
0.05
30
46
2930
6.8
CONCENTRATION - me/1
PRODUCT
3.3
1.2
140
1.0
7.0
6.9
21
196
24
0.02
17
17
381
6.8
• - AUgUS L
O7C
5O
1 ^
C Q
— 7 9
BRINE
331
161
1270
7.0
7.3
3480
79
400
20
0.04
48
65
5900
7.0
y, JL^OO
PERCENT
REJECTION
98
98
82
77
0
99
62
39
0
60
77
63
87
^"
*pH of samples at time of analysis
-20-
-------
Variations in Product Flux. Product flux varied widely (Figure 12).
The variations occurred under constant operating conditions, apparently
independently of any exterior operational changes. As mentioned
previously, disassembly of this reverse osmosis stack disclosed that the
brine flow paths on the desalination plates were blocked, causing the
differential pressure to increase from 35 psig to more than 95 psig
through the stack. This blockage reduced the effective pressure by 20
percent which directly influenced the flux.
UJ
a:
m
u.
o
I
<
§
w£j
0:0.
K!
o
s
a.
u.
o
3
10
JULY
SEPT
FIGURE 12-REVERSE OSMOSIS DATA
STACK IE
PRODUCT FLUX VERSUS TIME
-21-
-------
It was also postulated that the closed brine stream reduced the
effective membrane area of the unit by bridging the spiral flow path
with precipitated salt. This reduced exposure of the water to the
membranes and again caused a decrease in product flux.
Product Recovery. This reverse osmosis stack was designed to have a
high product recovery ratio, and initially it did. On the first day of
operation it recovered approximately 50 percent of the influent;
however, within 14 days, this percentage decreased to approximately 36
percent. By the end of the experiment, product recovery was less than
31 percent. Such drastic changes in such a short period were assumed
to be the result of the internal damage and not typical of expected
results. Nevertheless, even at the highest recoveries, the amount of
brine produced would equal the product and would involve brine disposal
problems.
Electrodialysis
Although the electrodialysis unit (ED) was operated in the summers of
1967 and 1968, a more consistent operation was possible in 1968; this
report is concerned primarily with that period. During 1968, the unit
was operated on essentially the same blended water as was the reverse
osmosis unit (Figure 11).
TDS Removal
The percent TDS removal and effluent TDS concentrations for the
electrodialysis unit are shown in Figure 13. The variations in TDS
removal from 36 percent to less than 20 percent were highly dependent
on the physical condition of the membranes. In general, any sudden
increase in TDS removal was due to a cleansing of the membranes. The
general decline in efficiency from early August through September was
attributed partially to a 10°C decline in influent water temperature and
partially to a general chemical and/or biological fouling of the membranes,
Substances that frequently accumulated within the stack were precipitated
salts, biological slimes, and suspended solids. Any fouling of the
membranes increased the electrical resistance in the stack, which lowered
the TDS removal capacity of the unit. A typical mineral analysis of the
three flow streams in the unit is shown in Table 5.
Nitrate was removed at an average rate 1.98 times that calculated for
total dissolved solids. This factor compares favorably with the removal
range of values reported by Ionics, Incorporated, which was 1.47 to 2.47
times the TDS removal (4). No significant boron removal was observed
at any time.
-22-
-------
o»
E
i
W
0
-I
o
(/>
o
UJ
o
to
g
UJ
o
UJ
Q.
UJ
oc.
v>
o
1.5
JUNE
JULY
AUG
SEPT
10
1968
FIGURE 13 - ELECTRODIALYSIS DATA
PRODUCT TOTAL DISSOLVED SOLIDS
AND PER CENT REMOVAL OF TDS VERSUS TIME
Product Recovery Rates
The design of this unit permitted product recovery to vary only through
increases or decreases in the amount of water needed to prevent satu-
ration of salts in the recirculating brine stream. For the experiment,
this requirement was kept constant at 8 gpm; thus, with a constant
product flow of 25 gpm, the product recovery was approximately 75 percent.
On the surface, this appears to be a better ratio than that achieved by
the reverse osmosis unit; however, it should be remembered that to achieve
a product of similar quality, the product stream must be passed through
a series of stacks, thus ultimately yielding a larger quantity of brine.
Cost Comparison for
Reverse Osmosis and Electrodialysis
The costs of power, chemicals, and supplies for both the RO and ED units
for operation on the 3000 mg/1 TDS blended feed water are summarized in
Table 6. The costs presented for the ED are based on a single pass
through demineralizing stack and, therefore, can not be readily compared
to the costs of the RO unit.
-23-
-------
TABLE 5
ELECTRODIALYSIS DATA
MINERAL ANALYSIS
OPERATING CONDITIONS
isauc uj. i-> diup j.e
pH of brine str
Temperature .of
Applied E. M. F
Stack current -
Flow Ra
GPM
VJJ. .LA
Influent -
S Pi-nHiirt-
CONCENTRATION -
CONSTITUENT
Calcium
Magnesium
Sodium
Potassium
Boron
Sulfate
Bicarbonate
Chloride
Nitrate
Iron (Total)
Silicon
TDS
pH*
FEED
167
68
625
4.0
6.7
1430
336
167
26
0.02
29
3010
8.1
PRODUCT
105
51
515
2.5
6.5
1140
230
157
14
0.02
30
2240
7.9
- - — jua.y ±3,
5Q
9A
- — 97A
- — IT
- — ^T
oc
ma/1
BRINE
321
138
1060
7.6
6.7
2580
147
593
54
0.02
29
5240
7.6
±yoo
PERCENT
REJECTION
37
10
18
38
3
20
32
6
46
0
0
26
"~
pH* of samples at time of analysis,
-24-
-------
TABLE 6
COST ANALYSIS SUMMARY FOR
REVERSE OSMOSIS AND ELECTRODIALYSIS UNITS
ITEM
Sulfuric
Acid
QUANTITY USED PER MILLION
COST GALLONS OF PRODUCT
PER UNIT
$31.60*/Ton
RO
2.74 Tons
ED
1.04 Tons
COST PER MILLION GALLONS
OF PRODUCT
RO
$86.50
ED
$32.86
Filter
Cartridges $0.75/Each
Electrical
Power $0.01/KWH
Cyanamer $1.00/Ib
P-35
2.50
100
6700 KWH 4250 KWH
59 Ibs
$197.50
$67.00
$59.00
$79.00
$42.50
TOTALS
$410.00
$154.36
*Based on tankcar lots.
The amount of materials used in these calculations was based on actual
quantities used to achieve the product water produced at the Interagency
Agricultural Wastewater Treatment Center. No consideration was given
to capital cost, cost of operation and maintenance, or cost of brine
disposal. Because of the small scale of the units and because the primary
objective was determination of technical feasibility of the units, the
costs for supplies and chemicals alone are high. Other published estimates
have shown the cost of large-scale reverse osmosis plants to be considerably
less than the prorated costs found in these experiments.
An economic evaluation study performed by Kaiser Engineers (5) has published
costs for a 50 mgd plate and frame module type of reverse osmosis plant based
on the following factors:
-25-
-------
Plant 30 years
Membrane Life 1 year
Stream Factor 90 percent
Fixed Charge Rate
@ 3 1/4% Interest 5.27 percent
Insurance 0.25 percent
Taxes 1.50 percent
Brine Disposal $0.02/KGAL
Product Recovery 50 percent
Power Cost $0.010/KWH
Influent TDS 3000 mg/1
The product water cost for such a plant was approximately 32 cents per
1000 gallons. These figures were based on present membrane technology.
Costs may be reduced by 30 to 40 percent with projected improvements in
membrane technology.
Cost data have been published for a 50 mgd electrodialysis plant with a
demineralizing stack flow path similar to that of the smaller scale plant
at the Interagency Agricultural Wastewater Treatment Center and designed
under the following criteria: (6)
Plant Life 30 years
Membrane Life 3-5 years
Stream Factor 90 percent
Fixed Charge Rate
@ 3 1/4% 5.27 percent
Insurance 0.25 percent
Brine Disposal $0.020/KGAL
Power Cost $0;007 - $0.01/KWH
Influent TDS 3638 mg/1
The product water cost for the above plant was also approximately 32
cents per 1000 gallons. Technical improvements in critical components,
materials, and process design could result in a 33 percent cost reduction
as compared to the present state-of-the-art.
-26-
-------
SECTION V
ACKNOWLEDGMENTS
The desalination investigations were performed under the direction of
Percy P. St. Amant, Sanitary Engineer, Environmental Protection Agency.
The field work was the responsibility of Bryan R. Sword, Sanitary Engineer,
Environmental Protection Agency.
The Office of Saline Water provided the desalination equipment and
technical advice for the Federal Water Quality Administration's use.
The assistance given by Mr. Warren H. Bossert, formerly of the
Aerojet-General Corporation, proved invaluable in maintaining and
operating the reverse osmosis unit.
The assistance given by the Field Operations Office of Ionics, Inc. in
providing guidelines for operation of the electrodialysis unit is
gratefully acknowledged.
A major contribution to the studies was given by the efforts of the
United States Bureau of Reclamation technicians, Messrs. Norman W.
Cederquist, Gary E. Keller, Gary L. Rogers, and Mathew C. Rumboltz.
This report was prepared by Bryan R. Sword, Sanitary Engineer, Environmental
Protection Agency.
-27-
-------
SECTION 71
REFERENCES
Merten, U., Lonsdale, H. K., Riley, R. L., Vos, K. D.,
Reverse Osmosis for Water Desalination. Office of Saline Water,
Research and Development, Progress Report No. 208, September
1966.
Standard Methods for the Examination of Water and Wastewater.
American Public Health Association, Inc., New York, 12th
Edition (1965).
Correspondence with Aerojet-General Corporation, February
5, 1968.
Katz, William E., Nitrate Removal by Electrodialysis - A Brief
Review, Ionics, Incorporated, October 25, 1966.
Harris, F. L., Engineering and Economic Evaluation Study of
Reverse Osmosis, Office of Saline Water, Second Symposium on
Reverse Osmosis, April 1969.
Christodoulou, G. R., Olsson, G. R., Monnik, H. J., Parametric
Economic and Engineering Evaluation Study of the Electrodialysis
Process for Water Desalination, Office of Saline Water, Research
and Development Progress Report No. 488.
-29-
-------
SECTION VII
PUBLICATIONS
SAN JOAQUIN PROJECT. FIREBAUGH CALIFORNIA
1968
"Is Treatment of Agricultural Waste Water Possible?"
Louis A. Beck and Percy P. St. Amant, Jr. Presented at Fourth
International Water Quality Symposium, San Francisco, California,
August'14, 1968; published in the proceedings of the meeting.
1969
"Biological Denitrification of Wastewaters by Addition of Organic
Materials"
Perry L. McCarty, Louis A. Beck, and Percy P. St. Amant, Jr.
Presented at the 24th Annual Purdue Industrial Waste Conference,
Purdue University, Lafayette, Indiana. May 6, 1969.
"Comparison of Nitrate Removal Methods"
Louis A. Beck, Percy P. St. Amant, Jr., and Thomas A. Tamblyn.
Presented at Water Pollution Control Federation Meeting, Dallas,
Texas. October 9, 1969..
"Effect of Surface/Volume Relationship, C02 Addition, Aeration, and
Mixing on Nitrate Utilization by Scenedesmus Cultures in Subsurface
Agricultural Waste Waters".
Randall L. Brown and James F. Arthur. Proceedings of the
Eutrophication-Biostimulation Assessment Workshop, Berkeley,
California. June 19-21, 1969.
"Nitrate Removal Studies at the Interagency Agricultural Waste Water
Treatment Center, Firebaugh, California"
Percy P. St. Amant, Jr., and Louis A. Beck. Presented at 1969
Conference, California Water Pollution Control Association,
Anaheim, California, and published in the proceedings of the
meeting. May 9, 1969.
"Research on Methods of Removing Excess Plant Nutrients from Water"
Percy P. St. Amant, Jr., and Louis A. Beck. Presented at 158th
National Meeting and Chemical Exposition, American Chemical
Society, New York, New York. September 8, 1969.
"The Anaerobic Filter for the Denitrification of Agricultural
Subsurface Drainage"
T. A. Tamblyn and B. R. Sword. Presented at the 24th Purdue
Industrial Waste Conference, Lafayette, Indiana. May 5-8, 1969.
-31-
-------
SAN JOAQUIN PROJECT, FIREBAUGH. CALIFORNIA (Continued)
1969
"Nutrients in Agricultural Tile Drainage"
W. H. Pierce, L. A. Beck and L. R. Glandon. Presented at the
1969 Winter Meeting of the American Society of Agricultural
Engineers, Chicago, Illinois. December 9-12, 1969.
"Treatment of High Nitrate Waters"
Percy P. St. Amant, Jr., and Perry L. McCarty. Presented at
Annual Conference, American Water Works Association, San Diego
California. May 21, 1969. American Water Works Association
Journal. Vol. 61. No. 12. December 1969. pp. 659-662.
The following papers were presented at the national Tall Meeting
of the American Geophysical Union, Hydrology Section, San Francisco,
California. December 15-18, 1969. They are published in Collected
Papers Regarding Nitrates in Agricultural Waste Water. USDI, FWQA,
#13030 ELY December 1969.
"The Effects of Nitrogen Removal on the Algal Growth Potential of
San Joaquin Valley Agricultural Tile Drainage Effluents"
Randall L. Brown, Richard C. Bain, Jr. and Milton G. Tunzi.
"Harvesting of Algae Grown in Agricultural Wastewaters"
Bruce A. Butterfield and James R. Jones.
"Monitoring Nutrients and Pesticides in Subsurface Agricultural
Drainage"
Lawrence R. Glandon, Jr., and Louis A. Beck
"Combined Nutrient Removal and Transport System for Tile Drainage
from the San Joaquin Valley"
Joel Goldman, James F. Arthur, William J. Oswald, and Louis
A. Beck.
"Desalination of Irrigation Return Waters"
Bryan R. Sword.
"Bacterial Denitrification of Agricultural Tile Drainage"
Thomas A. Tamblyn, Perry L. McCarty and Percy P. St. Amant.
"Algal Nutrient Responses in Agricultural Wastewater"
James F. Arthur, Randall L. Brown, Bruce A. Butterfield, Joel
C. Goldman
-32-
-------
1
Access/on Number
w
5
2
Subject Field & Group
Q5D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization , . _,-,..
Water Quality Office
Environmental Protection Agency
Washington, D.C.
Title
DESALINATION OF AGRICULTURAL TILE DRAINAGE
•J Q Authors)
Sword, Bryan R.
16
21
Project Designation
Project #13030
ELY
Note
22
Cita tion
Agricultural Wastewater Studies
Report 13030 ELY 05/71-12
Pages 32 Figures 13 Tables 6 References 6
23
Descriptors (Starred First)
*Desalination, *Irrigation Waters, *Return Flow
Reverse Osmosis, Electrodialysis, Salinity, Nitrate, Boron, Tile Drains
25
Identifiers (Starred First)
*San Joaquin Valley, California
27
Abstract investigations were made to determine the technical feasibility of desalination
of tile drainage. The source of the tile drainage was a 400-acre field near Firebaugh,
California. Reverse Osmosis (RO) and Electrodialysis (ED) processes were studied. Two RO
membrane stacks were investigated. The first, a high salt rejection, low product yield,
was operated on variable quality (3000-7000 mg/1 TDS) irrigation return water. In the
7-month investigation period TDS removal efficiencies decreased from 93 percent to 80
percent salt rejection and the product flux decreased from 12 gal/ft^/day to less than
9 gal/ft^/day. The 20 mg/1 of nitrate-nitrogen and 8 mg/1 of boron contained in the
influent were not effectively rejected. The second RO stack and also the ED unit were
operated on return waters that were controlled to have a 3000 mg/1 TDS. The second RO
stack was designed for a high product rate and low salt rejection. The TDS removal
remained at 85 percent for a 3-month run. Product flux decreased from over 19 gal/ft /day
to less than 12 gal/ft2/day. Nitrate and boron rejection was low. The ED data are based
on a single pass through the membrane stack. The TDS removal varied from 35 percent to
15 percent. The nitrate removal rate was greater than the TDS removal. Boron removal
was negligible. It is estimated that the costs for the two processes are approximately
equal—$320 per million gallons of product.
This report was submitted in partial fulfillment of Project No. 13030 ELY under
sponsorship of the Environmental Protection Agency.
Abstractor
Sword
| Institution
Environmental Protection Agency
WR:I02 (REV. JULY 1S8B)
WRSIC
SEND. WITH COPY OF DOCUMENT. TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. 20240
* SPO: 1970-388-930
* V. S. GOVERNMENT PRINTING OFFICE : 1972-514-148/67
------- |