WATER POLLUTION CONTROL RESEARCH SERIES • I6060DRV 03/72
RENOVATING SECONDARY SEWAGE
BY GROUND WATER RECHARGE
WITH INFILTRATION BASINS
U.S. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes
the results and progress in the control and abatement
of pollution in our Nation's waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Environmental
Protection Agency through inhouse research and grants
and contracts with Federal, State, and local agencies,
research institutions, and industrial organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications
Branch (Water), Research Information Division^ R & M,
Environmental Protection Agency, Washington, D. C.
20460.
-------
RENOVATING SECONDARY SEWAGE BY GROUND
WATER RECHARGE WITH INFILTRATION BASINS
by
Herman Bouwer, R. C. Rice, and E. D. Escarcega
U. S. Water Conservation Laboratory, Agricultural Research
Service, U. S. Department of Agriculture, Phoenix, Arizona
and
M. S. Riggs, Salt River Project, Phoenix, Arizona
for the
OFFICE OF RESEARCH AND MONITORING
ENVIRONMENTAL PROTECTION AGENCY
Project No. 16060DRV
(formerly WPD-140-03)
March 1972
For sale by the Superintendent ot Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $1.00
-------
EPA Review Notice
This report has been reviewed by the Environmental Pro-
tection Agency and approved for publication. Approval
does not signify that the contents necessarily reflect
the views and policies of the Environmental Protection
Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for
use.
11
-------
ABSTRACT
The feasibility of renovating secondary sewage effluent by ground
water recharge was studied with six infiltration basins in the loamy
sand of the Salt River bed. The ground water was at a depth of 10 ft
and observation wells for sampling renovated water were installed
inside and outside the basin area.
Infiltration rates generally ranged between 2 and 3 ft/day at 1-ft
water depth. They were highest in grass-covered basins and lowest in
a gravel-covered basin. Flooding periods of 2 to 3 weeks alternated
with dry ups of 10 days in the summer and 20 days in the winter yielded
maximum long-term infiltration, i. e., about 400 ft/year. Directional
hydraulic conductivities of the aquifer were evaluated by resistance
network analog and field tests. The resulting values were used to
predict water table profiles and underground detention times in an
operational system which would produce renovated water at about
$5/acre-foot.
Suspended solids, BOD, and fecal coliform were essentially completely
removed as the water seeped through the soil. With proper inundation
scheduling, significant removal of nitrogen could also be obtained,
especially below vegetated basins. Most of the phosphates and fluorides
were also removed from the water. Boron was not removed.
111
-------
CONTENTS
Section Page
I CONCLUSIONS 1
II RECOMMENDATIONS 3
III INTRODUCTION 5
IV DESCRIPTION OF PROJECT 1
V INFILTRATION STUDIES 13
1. Basin management and infiltration rates 13
2. Pressure-head and water-content measurements 25
3. Basin management for maximum hydraulic loading 29
VI AQUIFER STUDIES 37
1. Hydraulic conductivity of aquifer and flow system 37
2. Underground detention times 4-2
3. Water table response 4-5
4. Effect of recharge on hydraulic conductivity 4-7
of aquifer
VII WATER QUALITY STUDIES ij-9
1. Biochemical and chemical oxygen demand 4-9
2. Nitrogen 52
3. Phosphate 63
4. Boron 65
5. Fluoride 65
6. Total dissolved salts 68
7. pH TO
8. Coliform and soil bacteria 70
9. Effect of grass filtration on effluent quality 75
10. Water temperatures 78
v
-------
Section Fag£
VIII DESIGN AND OPERATION OF LARGE-SCALE SYSTEM 79
1. General design aspects 79
2. Underground flow system 8l
3. Economic aspects 9-2
4. Future projects 93
IX ACKNOWLEDGMENTS 97
X REFERENCES 99
XI PUBLICATIONS 101
VI
-------
FIGURES
PAGE
1 PLAN OF FLUSHING MEADOWS PROJECT 9
2 SCHEMATIC OF INFLOW AND OUTFLOW STRUCTURES FOR 10
INFILTRATION BASINS
3 INFILTRATION RATES FOR RECHARGE BASINS IN 1967 lk
4 INFILTRATION RATES FOR RECHARGE BASINS IN 1968 l6
5 DIAGRAM OF SERIES FLOW WITH MEASURING AND SAMPLING POINTS 19
6 INFILTRATION RATES FOR RECHARGE BASINS IN 1969 21
7 ACCUMULATED INFILTRATION FOR RECHARGE BASINS IN 1969 2U
8 PRESSURE HEADS OF WATER IN SOIL BENEATH BASIN 5 DURING 2.6
FLOODING AND DRY UP
9 PRESSURE HEADS OF WATER IN SOIL BENEATH BASIN 6 DURING 2?
FLOODING AND DRY UP
10 INFILTRATION RATE AND K AT DIFFERENT DEPTHS BELOW 28
u
BASIN 2
11 K AS A FUNCTION OF SOIL-WATER PRESSURE-HEAD FOR SOIL 30
u
IN BASIN 2
12 RELATION BETWEEN WATER CONTENT AND PRESSURE HEAD FOR 31
SOIL IN BASIN 5
13 SCHEMATIC PRESENTATION OF INFILTRATION DECREASE DURING 32
INUNDATION AND RECOVERY DURING DRY UP
14 LONG-TERM INFILTRATION RATE IN RELATION TO LENGTH OF 33
INUNDATION PERIODS
15 RESPONSE OF WATER LEVEL IN ECW TO INFILTRATION 38
16 GROUND WATER FLOW SYSTEM AT STEADY-STATE DURING RECHARGE kl
17 WATER-LEVEL PROFILES IN OBSERVATION WELLS DURING k6
INUNDATION AND DRY UP
18 COD OF SECONDARY EFFLUENT AND OF RENOVATED WATER FROM 50
EAST CENTER WELL IN 1968
VII
-------
PAGE
19 COD OF SECONDARY EFFLUENT AND OF RENOVATED WATER FROM 51
ECW, WELL 1-2, AND WELL 5-6 IN 1969
20 TOTAL NITROGEN OF SECONDARY EFFLUENT, AND NITRATE N AND 5^
AMMONIUM N IN RENOVATED WATER FROM ECW IN 1968
21 TOTAL NITROGEN OF, SECONDARY EFFLUENT, AND NITRATE N AND 55
AMMONIUM N IN RENOVATED WATER FROM ECW IN 1969
22 NITRATE N IN RENOVATED WATER FROM WELLS 1-2, ECW, AND 58
5-6 IN 1968
23 NITRATE N IN RENOVATED WATER FROM WELLS 1-2, ECW, AND 59
5-6 IN 1969
24 COLIFORM BACTERIA IN RENOVATED WATER FROM ECW IN 1968 73
25 COLIFORMS AND FECAL STREPTOCOCCI IN RENOVATED WATER jk
FROM ECW IN 1969
26 PLAN AND CROSS-SECTION OF TWO PARALLEL STRIPS WITH 80
RECHARGE BASINS WITH WELLS IN CENTER FOR COLLECTING
RENOVATED WATER
27 GROUND WATER MOUND OF FLOW SYSTEM IN FIGURE 16 WITH 82
VERTICAL SCALE EXAGGERATED
28 LINEAR FLOW BENEATH RECHARGE STRIP (ABCD) AND NODE 85
ARRANGEMENT IN PORTION OF FLOW SYSTEM (ABEF) ANALYZED
BY RESISTANCE NETWORK ANALOG
29 DIMENSIONLESS GRAPH OF 2 WSI/T AH 8?
30 DIMENSIONLESS GRAPH OF AH ,,/AH. „ VERSUS L/S 88
D ~~ F A~ r
31 STREAMLINES AND EQUIPOTENTIALS (IN FEET ABOVE WATER 89
TABLE ADJACENT TO WELL) FOR HYPOTHETICAL SYSTEM IN
SALT RIVER BED
32 EQUIPOTENTIALS (AB = 100, F = 0) IN REGION ABEF FOR 90
DIFFERENT VALUES OF L/S
33 MODEL OF RIO SALADO PROJECT, SHOWING SEWAGE WATER 95
RENOVATION SYSTEM (FRONT) AND RECREATIONAL LAKES (BACK).
vxii
-------
TABLES
No. Page
1 Driller's Log for East Well and West Center Well 8
2 Effect of Surface Condition on Infiltration Rate (1968) 18
3 Results of Hydraulic Conductivity Measurements IfO
4 Occurrence of Distinct Nitrate Peaks in ECW in Relation l|i|.
to Start of Inundation Period and Average Infiltration
Rate
5 COD in Mg/Liter for Various Wells 53
6 Nitrate Nitrogen Concentrations in Mg N per Liter for 6l
Various Wells
7 Ammonium Nitrogen Concentrations in Mg N per Liter for 62.
Various Wells
8 Phosphate-Phosphorus Concentrations in Mg P per Liter 64
of Effluent, Renovated Water, and Native Ground Water
9 Boron Concentration in Mg B per Liter for Effluent, 66
Renovated Water, and Native Ground Water
10 Fluoride Concentration in Mg F per Liter for Effluent, 6j
Renovated Water, and Native Ground Water
11 Total Salt Concentration in Effluent and in Water from 69
Wells, Mg/Liter
12 pH of Effluent and Water from Wells (1969) 71
13 Aerobic Bacteria Counts in Recharge Basin Soil 72
14 Effect of Flow Through Basins on COD, NH, , and NO., 76
Content of Effluent (1968)
15 Effect of Flow Through Grass Basins and Grass Mat on 77
COD, NH. and N0_ of Effluent
4 3
IX
-------
SECTION I
CONCLUSIONS
1. The pilot project has shown that secondary sewage effluent can be
effectively renovated for unrestricted irrigation and primary-contact
recreation by ground water recharge with infiltration basins in the Salt
River bed. In an operational system, the basins would be located on
both sides of the river bed and the renovated water would be pumped
from wells in the center of the river bed. Total cost of renovating
sewage effluent in this manner is estimated at $5/acre-foot at the well.
2. An accumulated infiltration of 400 ft/yr can be obtained with
flooding periods of 2 to 3 weeks alternated with dry-up periods of
10 days in the summer and 20 days in the winter. Thus, 1 acre of
recharge basin can handle about 0.35 mgd. Grass covered basins had the
highest infiltration rates. However, water depths are restricted in
grass covered basins and short, frequent floodings must be employed in
the spring and early summer to get a tall, dense stand of the grass.
3. The response of the ground water table beneath the infiltration
basins enabled the evaluation of the horizontal and vertical hydraulic
conductivity of the aquifer. These hydraulic conductivities were used
in the design of an operational system to predict underground detention
times and water table profiles for various geometries of recharge
basins and wells for pumping renovated water.
4. Short, frequent inundations of the basins (2-4 days wet, and 3-5
days dry, for example) yielded almost complete conversion of the
nitrogen in the effluent water to nitrate in the renovated water. With
2-3 week floodings alternated by 10-20 day dry ups, however, 50 to 80%
of the nitrogen was removed following the passage of nitrate peaks.
These peaks occurred shortly after the start of a new inundation period
due to leaching of nitrified effluent from the soil held as capillary
water during the preceding dry up. Nitrogen removal was greater below
vegetated basins than below nonvegetated basins.
5. Prolonged operation of the infiltration basins for nitrogen removal
eventually caused an increase in the ammonium level of the renovated
water. Thus, short, frequent inundations, preferably in combination
with growing a crop, should be used every other year or so to restore
the nitrogen-removing capability of the recharge system.
6. The soil filtration process yielded essentially complete removal
of suspended solids, BOD, and fecal coliform. Significant removal of
phosphates and fluorides also took place. Boron, however, was not
removed.
7. There was no indication of gradual clogging of the soil profile or
the aquifer, or of a decrease in the renovation efficiency of the
system. Thus, a ground water recharge system for renovating secondary
sewage effluent should have a long, useful life.
-------
SECTION II.
RECOMMENDATIONS
Because of the favorable results of the experimental project, larger
projects of an operational nature can be started with confidence. The
renovated water could, for example, be used for unrestricted irrigation
and recreational lakes. The design procedures presented in this report
could be used to determine the most favorable layout of recharge basins
and wells, and to predict the water table positions, pumping lifts, and
underground detention times for the system.
The infiltration basins for an operational system are likely to be from
several acres to 10 acres or more, which is much larger than the long,
narrow basins of the experimental project. Field investigations should
therefore be carried out to determine if the large basins have
restricted escape of air from the soil when a new inundation is started.
The resulting buildup of air pressure beneath the advancing wet front
in the soil can then reduce infiltration rates. Slow filling of the
basin from one side to obtain a low rate of advance of the water over
the bottom of the basin, and installation of air relief pipes in the
soil may be effective in minimizing the buildup of air pressure in the
soil when the soil is wetted.
To avoid spread of renovated sewage water into the ground water basin
of the Salt River Valley, all the effluent infiltrated in the basins
should be pumped out of the aquifer as renovated water. This makes it
possible to determine how many pounds of nitrogen, oxygen demand,
organic carbon, phosphorus, etc., are removed by the system. When a
larger system is installed, the important quality parameters should be
monitored, as well as the ground water levels along the periphery of
the system to make sure that no renovated water spreads into the rest
of the ground water basin. Underground detention times and water table
profiles should be measured and compared with the predicted values to
check the validity of the design procedures.
Operation of the pilot project should be continued to permit more
detailed studies of the nitrogen and coliform behavior. Also, the
effectiveness of various types of vegetation on nitrogen removal should
be studied. Soil clogging, particularly of the surface layer, should
be examined in more detail, as well as the effect of rainfall during
dry up on the recovery of the infiltration. The suitability of various
crops for stimulating denitrification, and the possibility of the
addition of artificial carbon sources to obtain more denitrification
should be investigated. Some of these studies can best be carried out
in the laboratory on soil columns. Because of the interest in using
renovated sewage water for recreation, the biostimulation of such water
in impoundments should be studied so that guides for the optimum manage-
ment of recreational lakes can be developed. Eventually, the existing
experimental project should be operated on a maintenance basis after
most research information has been obtained. Periodically, intensive
measurements should then be carried out to determine the long-term
behavior of the system.
-------
SECTION III
INTRODUCTION
The Salt River Valley with Phoenix, Tempe, Mesa, Scottsdale, and
Glendale as major cities and a total population of close to one million,
is a water-deficient region. In the area served by the Salt River
Project alone, about one-third of the roughly one million acre-feet
of water annually used by agriculture and municipalities is pumped from
the ground water. This is essentially a non-renewable water resource
and, consequently, ground water levels in some parts of the valley are
declining at about 10 ft/yr.
With the rapid urbanization of the valley, sewage is becoming increas-
ingly significant as a pot^ ial water resource. Most of the sewage is
treated at the 23rd and 91t> L. Avenue Treatment Plants in Phoenix, which
presently handle about 20 and 60 mgd, respectively, or a total of
almost 100,000 acre-feet/yr. This volume is expected to increase to
about 300,000 acre-feet by the year 2000. Both plants use the activated-
sludge process for secondary treatment.
Potential uses of the sewage effluent would be for irrigated agriculture,
recreational lakes, and certain industries. Since the average water use
by crops is about 4.5 ft/yr, the 300,000 acre-feet of effluent expected
by the year 2000 could irrigate about 70,000 acres. Because of the
varied agriculture and dense population in the valley, and because canal
water is commonly used for irrigation of parks, playgrounds, and private
yards, large-scale return of the effluent to the canal system requires
tertiary treatment so that it will be aesthetically acceptable and
suitable for unrestricted irrigation and primary-contact recreation.
The Salt River bed, which is normally dry and traverses the entire
valley from east to west at widths of about one-fourth to one-half
mile, offers an excellent opportunity to achieve this tertiary treatment
by ground water recharge with infiltration basins.
The technical and economical feasibility of tertiary treatment by soil
filtration in the Salt River bed, as well as the design and management
criteria for an operational system, could only be evaluated with a
pilot project. Plans for such a project were formulated and refined
in the period 1964-1966. After receipt of a demonstration grant in
December 1966 from the then Federal Water Pollution Control Administra-
tion, construction of the pilot project, known as the "Flushing Meadows
Project," began in February 1967 and was completed in August 1967-
-------
AERIAL VIEW OF FLUSHING MEADOWS PROJECT
-------
SECTION IV
DESCRIPTION OF PROJECT
The Flushing Meadows Project is located in the Salt River bed about
1 1/2 miles west of the 91st Avenue sewage treatment plant. This is
an activated sludge plant which discharges unchlorinated secondary
effluent into a channel on the north side of the (dry) river bed. The
effluent flows westward to Buckeye where some of it is used by the
Buckeye Irrigation District.
At the Flushing Meadows Project, effluent is pumped from the channel
into a constant-head pipe, from where it flows through an underground
concrete pipe to concrete boxes located at the head of each experimental
basin (Figure 1). These head boxes are placed above the concrete pipe-
line and connected thereto with a concrete riser pipe (Figure 2). The
flow from the riser into the box is controlled with a valve commonly
used in irrigation, a so-called "alfalfa" valve. The water leaves
the head box through a triangular, critical-depth flume (10) for
measuring and recording the inflow into each basin (Figure 2).
The basins are 20 x 700 ft each with 1:1 side slopes and horizontal
bottoms. They are approximately 3 ft deep and spaced 20 ft apart
(Figure 1). In February 1968, gravel dams 2 ft high were placed across
each basin about 50 ft from the inflow end to provide a presedimentation
reservoir at the head of each basin (Figure 2). This was necessary to
protect the infiltration basins against high suspended solids loads due
to poor effluent quality.
An overflow structure was installed at the downstream end of each basin
to control the water depth in the basins. Boards could be inserted or
removed to increase or decrease the water depth, respectively (Figure 2).
After spilling over the boards, the water passed through a critical
depth flume for measuring and recording the outflow. These flumes were
of the same type as the flumes at the inflow end of the basins. Thus,
constant water depth could be maintained in each basin and the infil-
tration rate for each basin was calculated as the difference between
the inflow and the outflow. Inflow rates were commonly set at 0.4 cfs,
the outflow was usually in the 0 to'0.2 cfs range, and the water depths
in the basins were maintained at 0.5 or 1 ft. The flow from the outflow
end was collected by a concrete pipeline and returned to the effluent
channel.
The soil below the infiltration basins consisted of a layer of fine,
loamy sand with an average thickness of 3 ft (range 1.3 to 5 ft),
underlain by coarse sand and gravel layers to a depth of about 247 ft,
where a clay deposit began. This clay deposit is essentially imperme-
able and it forms the lower boundary of the system.
Mechanical analysis of the loamy-sand top layer indicated that 10% of
the soil particles are less than 0.06 mm in diameter and 50% of the
-------
Table 1. Driller's log for East Well and West Center Well.
East Well
ft
0-3 fine loamy sand
3-27 sand, gravel and boulders
27-30 clean sand, gravel, and boulders
30-49 clean, fine sand with occasional
cobbles
49-81 clean, fine sand with occasional
thin gravel strata
81-123 clean, fine sand
123-126 fine sand with trace of clay
126-136 clean, fine sand
136-146 clean sand and gravel
146-197 clean, fine sand
197-200 fine sand and gravel
200-247 fine sand
247 start of clay layer
West Center Well
ft
0-3 fine loamy sand
3-33 sand and gravel
33-44 boulders and gravel
44-50 sand and gravel
50-57 sand and traces of clay
57-63 coarse, clean gravel
63-72 sand, gravel, traces of clay
72-86 coarse gravel and boulders
86-98 sand, gravel, and traces
of clay
98-100 fine sand
-------
constant-head
structure
supply line
flume
gravel dam
drainage line
• EW
Figure 1. Plan of Flushing Meadows Project.
-------
o
trapezoidal
flume
INFLOW
alfalfa
valve
supply line
boards
sedimentation
basin
infiltration basin
OUTFLOW
infiltration basin
return of outflow
to channel
Figure 2. Schematic of inflow and outflow structures for infiltration bas
ins.
-------
particles are less than 0.18 mm in size. The clay content is small and
about 2 or 3%. The hydraulic conductivity, K, of the loamy sand at or
near saturation is 3 to 5 ft/day and the air-entry value is about -30
cm water, as measured with the double-tube and air-entry permeameter
techniques (1, 2, 4, 5).
The sand and gravel layers beneath the loamy-sand top layer are
described in Table 1, which is taken from the driller's log for the
100-ft-deep West Center Well and the 247-ft-deep East Well (Figure 1).
The static ground water table was at a depth of about 14 ft below the
bottom of the basin at the start of the project in 1967. In the first
year of operation, the static water table rose about 4 ft and remained
at about 10-ft depth thereafter. The rise of the water table may be
attributed to the ground water recharge from the Flushing Meadows Project
and to seepage from the effluent channel, which was newly constructed in
1966. Also, the main channel of the Salt River, which is a few hundred
yards south of the Flushing Meadows Project, carried water several times
in the spring of 1968, which could have recharged the ground water.
Observation wells consisting of 6-inch-diameter steel pipe in wells
drilled with the cable-tool technique, were installed in a line normal
to the basins in the center of the project (Figure 1). Another well,
the East Well, is located on the east side of the basins. The East
Center Well (ECW), West Center Well (WCW), and East Well (EW), were
installed in 1967, the other wells in 1968. The depth of the wells is
as follows:
Well Depth in ft
1 20
1-2 20
ECW 30
WCW 100
5-6 20
7 20
8 20
EW 247
All wells have solid casings open at the bottom with the exception of
EW, which was plugged at the bottom and perforated from 10 to 30 ft to
yield renovated sewage water for some experimental fish ponds that were
installed in 1970.
11
-------
SECTION V
INFILTRATION STUDIES
1. Basin Management and Infiltration Rates
The effect of basin management on infiltration rate was studied by
using different inundation and dry periods, different surface conditions
of the soil (bare, gravel-covered, and different types of vegetation),
and different water depths. Harrowing was sometimes included as a
variable, but usually all basins were cleaned and harrowed at the same
time. This was normally done in the spring.
An important factor in the infiltration behavior was the suspended
solids content of the secondary effluent, which was low in the summer
(5-20 ppm) but high in the winter period November-May (50-100 ppm) .
These solids were not completely removed by the presedimentation
reservoir and a layer of sludge often accumulated on the basin bottoms
during the winter period. Upon drying, the sludge layer shrank and broke
up into curled-up flakes of about 1 to 5 inches in diameter. Usually,
these flakes were removed in the spring, ("shaving" the bottom with a
front-end loader was the most effective way), after which the basins
were harrowed.
The infiltration rates for the six basins in 1967 are shown in Figure 3,
all for a water depth of about 7 inches. The basins were first flooded
on 30 August, but equipment failure forced a halt to the pumping. Thus,
22-27 September was the first "real" inundation period. The secondary
effluent during this period contained very few suspended solids, and
the infiltration rates remained essentially constant. These infiltration
rates could be taken as the potential or bench-mark infiltration rates,
affected only by the permeability of the soil and not yet by clogging
and microbiological processes . The bench-mark infiltration rates were
Basin Infiltration rate in ft/day
1 3.2
2 4.0
3 3.3
4 3.6
5 3.9
6 3.1
These rates agreed with the range of 3.6 to 6 ft/day calculated by
applying Darcy's equation to the 3-ft thick top layer of the loamy
sand using the hydraulic conductivity values determined with the
double-tube and air-entry permeameter techniques.
The infiltration rates for the next inundation period, 3-6 October
were lower, probably because of a 0.5-inch rain which fell just before
the basins were flooded. On 12 October, all basins were harrowed,
13
-------
RAIN ,
IN./DAY .5
0
3
2
0
3
2
< '
Q 0
i ' i
"I M ' I ' ["' I ' I i | I I I
I ' I ' ] i I ' I ' [' » I ' I ' | ' I ' I ' 'I
"'I'"
TT '"I
'I*
1967
BASIN
BASIN 2
|
UJ °
CC 2
O 0
BASIN 3
\ \
WATER DEPTH 7 INCHES
BASIN 5
2
— 3
2
1
BASIN 6
4-
JAN. FEB. MAR. APRIL MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC.
Figure 3. Infiltration rates for recharge basins in 1967.
-------
except basin 5 to see how effective harrowing was in restoring infil-
tration rates. Expressing the initial infiltration rates for the next
flooding period, 12-24 October, as a percentage of the bench-mark
rates for the first period 22-27 September, shows that the lowest
percentage is obtained for basin 5, which was not harrowed prior to
flooding (77% compared to 80-94% for the other basins). Thus, the
low infiltration rates for 3-6 October were likely due to surface
sealing and crusting caused by the rain, and harrowing was effective
in breaking up this crust. The infiltration rates for the period
12-24 October decreased almost linearly with time. This decrease
was probably mainly due to the solids content in the effluent, which
had started to increase to its winter level.
On 3 November, all basins were harrowed again and different inundation
schedules were used for the rest of the year (Figure 3). The suspended
solids content of the effluent was high and had a significant effect on
the infiltration rates. The infiltration rates again decreased almost
linearly with time during inundation and dry-up periods of at least
9 days were required to obtain reasonable infiltration recovery
(Figure 3) .
Infiltration rates for 1968 are shown in Figure 4. On 6 and 7 February,
the basins were raked to remove the sludge flakes. On the same day,
gravel dams were placed about 50 ft from the inlet end to create
presediinentation basins (Figure 2). These basins were effective in
reducing the solids load on the infiltration basins. The sludge removal
and drying gave excellent infiltration recovery, as shown by the high
infiltration rate for the period 7-10 February. On 20 February, all
basins were harrowed with a tooth harrow. The low infiltration rate
for basin 1 for the next infiltration period cannot be explained and
may be erroneous.
On 16 April, the water depth in the basins was increased from the
7 inches used so far to 13 inches by adding a board in the overflow
structure (Figure 2). This almost doubled the infiltration rate
(Figure 4), indicating that the infiltration reduction in the basins
was mainly caused by clogging of the surface layer of the soil.
The basins were swept on 7 May with a power lawn sweeper to remove the
sludge flakes. Next, all basins were harrowed with a tooth harrow.
Basins 3, 4, 5 and 6 were seeded with a mixture of giant and common
bermudagrass and then irrigated with about 6 inches of effluent every
2 or 3 days. These irrigations, for which the infiltration rates were
estimated at 2 ft/day, are shown as dots in Figure 4. Test strips of
other vegetation, including sudangrass, tifway, fescue, and blue
panicum, were also planted.
Basin 2 was covered with a 2-inch layer of coarse sand ("concrete" sand)
and topped with a 4-inch layer of 3/8-inch gravel on 23 May. Basin 1
was left in bare soil condition. All basins received 1 ft of water on
-------
RAIN 10
IN./DAY .'5
0
3
2
I
0
3
2
j
L BASIN I
BASIN 2
-
\
|968
%
- — ^ Nl
h-
UJ
UJ
1 BASIN 3
\
\
coo ooao © - - -
'_ BASIN 4
/
\
A\
ooooooo ©- • - N v \ ~. _
2 3
*™ •>
< 2
'_ BASIN
: \
\\*\
ooo oo oo
1 BASIN 6
X
\
WATER DEPTH 7 INCHES
WATER DEPTH 13 INCHES
© 6-INCH IRRIGATION
® S2-INCH IRRIGATION
\
OOO OOOO©---
1 1 1
1 1 1 ( 1 1 1 1 1 1 1
1 , 1 1
I I I 1 t 1 t I 1
I I I I I t I 1 i I 1.1,1
JAN. FEB.
MAR. APRIL MAY JUNE JULY AUG. SEPT.
Figure 4. Infiltration rates for recharge basins in 1968.
OCT. NOV. DEC.
-------
7 June and a sequence of short inundation periods was started for all
basins on 12 June, using a water depth of 7 inches.
The grasses in basins 3, 4, 5, and 6 reached a mature stand in August
with the giant bermudagrass emerging as the dominant species. To
evaluate the effect of the grass cover on the quality of the effluent
as it flowed through the grass, "series" flow was started to increase
the distance of overland flow for the effluent. For this purpose, the
basins were split into two groups of three with the flow going in
serpentine fashion through the three basins in each group, as shown in
Figure 5. Flumes were installed where the effluent flowed from one
basin to the next, so that inflow and outflow were measured for each
basin. The water depths in the basins during this period of series flow
were as follows:
Basin Water Depth in Ft
1 0.7
2 0.7
3 0.85
4 0.85
5 0.85
6 0.95
The water depth in the gravel basin (basin 2) refers to the original
soil surface. The grass in the vegetated basins was sufficiently tall
to be well above the water level.
Rather long inundation periods were selected for the series-flow
sequence because the effect of flow through the grass on the quality
of the effluent could be expected to increase with time, as biologically
active films developed on the grass stems. For the period 15 August to
17 December, the secondary effluent was of good quality and had a low
suspended solids content. Nevertheless, the grass basins 3 and 6,
which received effluent directly from the channel, had fairly low
infiltration rates (Figure 4).
The effect of surface condition on infiltration rate was evaluated for
basins 1, 2, 4, and 5, which received effluent that had traveled
through at least one basin of grass and therefore had lost most of its
suspended solids. Table 2 shows the accumulated infiltrations for the
period 20 August-17 September. In the second column, these values are
adjusted to a common water depth of 0.7 ft in the basins (assuming a
linear relation between infiltration rate and water depth). In the
third column the bench-mark infiltration rates for the period 22-27
September 1967 are listed. The fourth column shows the ratio between
adjusted and bench-mark infiltration rates and the fifth column shows
these ratios with the highest ratio arbitrarily set at 100.
The highest relative infiltration index is obtained for the grass
basins with values of 100 and 94 (Table 2). Next comes the bare soil
IT
-------
Table 2. Effect of surface condition on infiltration rate (1968).
Basin
1 (bare soil)
2 (gravel cover)
4 (grass)
00 5 (grass)
Accumulated
infiltration for
20 August- 17 December
feet
72
45
111
127
Adjusted
accumulated
infiltration
feet
72
45
91
105
Bench-mark
infiltration
rates
feet/day
3.2
4.0
3.6
3.9
Ratio of
adjusted to
bench-mark
infiltration
22.5
11.2
25.3
26.9
Index of relative
infiltration
84
42
94
100
-------
BASIN
A
4
A
__. JL_
700 FEET
= MEASURING FLUME
A SAMPLING POINT
Figure 5. Diagram of series flow with measuring and sampling points.
19
-------
basins with 84. The gravel basin is lowest with 42. The higher
infiltration rates in the grass basins are probably due to reduced
clogging of the soil surface by suspended solids and algae. The nega-
tive effect of the gravel layer on infiltration is probably due to a
"mulching" action of the gravel layer and resulting poor drying of the
underlying soil. Also, dust and other solids may have entered the
gravel layer and settled on the gravel-soil interface, creating a
clogged layer which could not be dried or broken up mechanically.
The giant bermudagrass appeared to be the most suitable grass. It
dominated the common bermuda and developed a vigorous stand that was
sufficiently dense for filtration and shading, yet sufficiently open
for the effluent to flow through, rather than over, the grass. This
made it possible to use inundation periods as long as a month without
adverse effects on the condition of the grass. Excellent growth was
also exhibited by tifway. However, the sod was too dense to permit
much flow of the effluent through the grass. The resulting increase
in water depth and complete inundation then caused the tifway to die,
particularly at the inflow end where solids settled on the grass. The
sudangrass also grew well, but the stand was not sufficiently dense
to provide good filtration and shading. Also, it died after reaching
maturity. The other grasses tested (blue panicum and fescue) failed
to survive. None of the grasses were mowed in 1968.
At the beginning of 1969- basin 1 was still in bare soil condition,
basin 2 had a gravel layer, and basins 3, 4, 5, and 6 had the dead
bermudagrass left from the 1968 growing season. The bermuda straw lay
flat and became covered with sludge due to the poor effluent quality
in the first half of 1969. Because of this, the infiltration rates
were relatively low, even when the water depth was 13 inches (Figure 6)
Just before the flooding period starting 3 March, basins 3, 4, 5, and 6
were burned. Only the top of the bermuda straw, which formed a mat
about 4 inches thick, was dry enough to burn effectively.
In the first week of April, basin 1 was swept with a power lawn sweeper
to remove most of the dry sludge flakes, and then harrowed with a spike
tooth harrow. Basin 2 was left unchanged and basins 3, 4, 5, and 6
were burned several times and harrowed between burnings to speed up
drying of bermuda straw. This procedure was fairly effective and when
the next flooding period started on 7 April, the previously vegetated
basins were essentially in bare soil condition.
On 14 April, basin 6 was harrowed and seeded with rice (variety
Caloro) . All basins were then inundated for 2 days. On 21 April,
basins 1, 3, 4, and 5 were harrowed again and basin 5 was seeded
with rice. For the next few months, short, frequent inundation periods
were used for basins 5 and 6 to promote germination and early growth
of the rice. In July, the rice was sufficiently tall to permit longer
inundation periods for the rest of the growing period.
20
-------
RAIN
IN/DAY1
0
3
2
I
0
3
2
5 '
Q 0
: BASIN i
® © a ®
1969
\\
BASIN 2
\ " - N
ro
I BASIN 3
LJ
LJ
i 0
I
z °
O 3
H 2
CC I
!j o
iZ 3
E 2
I
0
1BASIN 4
1 BASIN 5
®—o
CM!
^ BASIN 6
-®- WATER DEPTH 13 INCHES
WATER DEPTH 7 INCHES
I I I I I I I I I I I I . I I I I I I i I i I I I I I . 1 I I I I
JAN. FEB. MAR. APRIL MAY JUNE JULY AUG. SEPT.
Figure 6. Infiltration rates for recharge basins in 1969.
OCT
NOV.
DEC.
-------
Because of the continuous high suspended solids content of the
secondary effluent, short inundation periods were also maintained in
basins 1, 2, 3, and 4 for the April-July period. At the beginning of
a flooding period, the infiltration rates were relatively high, but
they declined rapidly because of the buildup of a sludge layer on the
bottom of the basins. Drying was effective in restoring infiltration
rates.
The giant bermudagrass in basins 3 and 4 made a slow comeback. The
tifway had completely died and there were only sporadic tufts of giant
bermudagrass. During the frequent short inundations in May and June,
however, the bermudagrass spread rapidly, especially from the banks.
Also, native grasses, mainly Mexican sprangletop (Leptochloa uninervia)
and barnyard grass (Echenochloa crusgalli) came in and exhibited a
luxurious growth, particularly in basin 1, which previously had been
kept in bare soil condition. A third grass, probably blue panicum,
also volunteered in basins 1, 3, and 4. The condition of the basins
in summer and fall of 1969 was as follows:
Basin 1, dense stand of sprangletop, barnyard grass, and
blue panicum
Basin 2, gravel layer
Basins 3 and 4, giant bermudagrass with sprangletop and
barnyard grass and some blue panicum
Basins 5 and 6, rice with giant bermuda and some sprangletop,
barnyard grass, and blue panicum.
A 2-week dry-up period was held in the second half of June for basins
1, 2, 3, and 4. This was effective in restoring the infiltration rates
which had become low because of the sludge accumulation in the basins.
The effluent quality improved at this time, so that normal infiltration
rates were again maintained.
On 9 August, a pump failure occurred so that the rice was dry for
almost 6 days. This happened during a very hot period and it caused
visible damage to the rice. Although the rice seemed to recover fully,
not much grain was formed and also because of lodging later on, no
attempt was made to harvest the grain.
A sequence of long inundation and dry-up periods was started in August,
with the rice basins receiving shorter dry ups than the other basins
until 15 October. On 20 November, the vegetation in basins 1, 3, and
6 was mowed, baled, and removed. Basins 4 and 5 were not mowed. Thus,
the condition of the basins at the start of the winter was as follows:
Basin 1, stubble of sprangletop and barnyard grass
Basin 2, gravel layer
22
-------
Basin 3, stubble of bermudagrass
Basin 4, bermudagrass straw (about a 4-inch thick layer)
Basin 5, rice straw (about a 2- to 3-inch layer)
Basin 6, rice stubble.
For the next flooding period, starting 21 November, a water depth of
13 inches was employed. On 4 December, the water depth was reduced
to 7 inches. After the decrease in the water depth, the infiltration
rates were on the average 57% of the values prior to reducing the
water depth. Theoretically, a reduction of close to 50% can be expected
if surface clogging is the restricting factor. If the fine-sandy-loam
top layer had become less permeable over its entire thickness, Darcy's
equation shows that the infiltration rates at a water depth of 7 inches
would be 80% of the rates at a water depth of 13 inches. Since the
actual reduction is closer to 0.5, most of the infiltration decrease
during flooding is thus due to clogging of the surface layer of the
soil. Applying a similar analysis to the basin covered with a gravel
layer shows that the reduction of the infiltration in the gravel-covered
basin is due to clogging of the surface of the original soil, just
below the soil-gravel interface.
Rainfall, which adversely affected the infiltration rate during flooding
when it occurred during the preceding dry-up period and the basins were
in bare soil condition, did not affect the infiltration rate when the
basins were vegetated. For example, on 15 September, a 0.6-inch rain
fell just before the start of a new inundation period. As shown in
Figure 6, this rain had little or no effect on the infiltration rate.
Thus, the reduction in infiltration rate due to rain must be attributed
to the mechanical impact of the raindrops on the soil surface and the
resulting translocation of fine particles and sealing. In areas with
high rainfall or where the sewage effluent or other wastewater is
applied with sprinklers through the land, a vegetated surface may be
required to avoid direct impact of the drops on the soil surface.
The accumulated infiltration for 1969 is shown in Figure 7. Dividing
the annual infiltration by the bench-mark infiltration for each basin
and setting the ratio for the basin with the highest relative infiltra-
tion rate at 100, according to the procedure used in Table 2, yields
the following indexes of infiltration for 1969:
Basin Index of Relative Infiltration
1 95
2 65
3 92
4 100
5 92
6 85
23
-------
250
JAN. FEB. MAR. APRIL MAY JUNE JULY AUG. SEPT.
Figure 7. Accumulated infiltration for recharge basins in 1969
OCT.
NOV. DEC.
-------
The vegetated basins all appear to have relative infiltration indexes
between 85 and 100. The rice basins are in the lower part of the range,
probably because of the short dry-up periods used in the summer and
fall. The gravel basin again had the lowest relative infiltration
with an index of 65. The index for basin 1 was 95, which is higher
than the index of 84 obtained in 1968 when no vegetation was allowed
to develop in this basin.
2. Pressure-Head and Water-Content Measurements
To evaluate flow regimes in the loamy-sand top layer, pressure heads
were measured at different depths with tensiometers during inundation
and dry-up periods. Water contents of the loamy sand were measured
with the neutron method. The loamy-sand layer is the top layer of the
soil profile beneath the basins and its average thickness is 3 ft.
Measurements taken in November and December 1967 indicate that the
pressure heads in the loamy-sand top layer were always negative,
except for the first few inches or so shortly after the start of a
new inundation period. The pressure heads decreased (became more
negative) as inundation progressed. Sometimes the pressure heads
decreased rapidly during flooding (Figure 8) indicating severe clogging
of the soil surface above the tensiometer profile. Sometimes the
decrease was more gradual (Figure 9). For both cases, the pressure
heads also decreased from one inundation to the next, indicating
increased clogging at soil surface.
The pressure-head measurements at different depths enabled the calcu-
lation of the unsaturated hydraulic conductivity, K , during inundation
for each depth increment between tensiometers. This was done for
basin 2 in the fall of 1967 when the basin was not yet covered by a
gravel layer. The infiltration rates in these calculations were taken
as the average infiltration rates for the entire basin as calculated
from the inflow-outflow records. The results (Figure 10) show that
K of the 8- to 40-inch layer remained essentially equal to the infil-
tration rate. Thus, the hydraulic gradient in this layer was essentially
one and the increasing desaturation of this layer was due to a decrease
in infiltration rate caused by clogging of the soil above the 8-inch-
depth level. The value of K of the 4- to 8-inch layer was less than
the infiltration rate, and K of the 0- to 4-inch layer was even
lower. Thus, the top layer had the lowest hydraulic conductivity and
most of the K-reduction was concentrated in the surface layer of the
soil. This is also indicated by the essentially linear relationship
between infiltration rate and water depth in the basins previously
discussed.
When a new flooding period is started, "final" infiltration rate and
essentially unit gradient in the loamy-sand top layer should be reached
in less than a day. If soil clogging has not yet begun, the "final"
infiltration rate can thus be used as an estimate of K of the loamy
sand at saturated or near-saturated conditions. The bench-mark infil-
tration rates can be used for such a purpose, which would yield a
-------
4-20
r- r—i
BASIN 5
-cm DEPTH
O 20-cm DEPTH
INUNDATION PERIOD
I RAINFALL
0
JO 14 18 22 26 30 4
NOVEMBER
Figure 8. Pressure heads of water in soil beneath basin 5 during flooding and dry up.
8 12 16
DECEMBER (1967)
20
-------
I ' I
BASIN 6
• 40-cm DEPTH
O 60-cm DEPTH
INUNDATION PERIOD
I RAINFALL
14 18
NOVEMBER
8 12 16
DECEMBER (1967)
Figure 9. Pressure heads of water in soil beneath basin 6 during flooding and dry up.
-------
ro
co
I —
0
BASIN 2
A Ku OF 0-4 INCH LAYER (0-10 cm)
n Ku OF 4-8 INCH LAYER (I0-20cm)
• Ku OF 8-40 INCH LAYER (20-100 cm)
O INFILTRATION RATE
12 17 22 27 2
NOVEMBER
Figure 10. Infiltration rate and K^ at different depths below basin 2.
7 12 17
DECEMBER (1967)
-------
K-value of 4 ft/day for the loamy sand in basin 2. The unsaturated
hydraulic conductivity, K , of the loamy sand can be taken from
Figure 10 and related, for example, to the tensiometer readings used
to calculate K . The resulting relation between K and (negative)
pressure head is shown in Figure 11. u
Combining the tensiometer readings with the water-content measurements
yields the relation between water content and pressure head of the loamy
sand. This relationship is shown in Figure 12 for three depths in
basin 5.
The curves in Figures 11 and 12 describe the unsaturated hydraulic
conductivity and water content characteristics of the loamy-sand top
layer. This information would enable the theoretical analysis of the
flow system in the top layer under intermittent inundation. Such
analyses may yield predictions of water and air movement into soil
and they may be used, for example, to develop guides for inundation
and dry-up schedules whereby oxygen demand of the system is matched
to oxygen supply under a variety of conditions.
3. Basin Management for Maximum Hydraulic Loading
The general pattern of the infiltration behavior of the recharge
basins was an essentially linear reduction in infiltration rate from
2-3 ft/day at the beginning of an inundation period to 1-2 ft/day
after approximately 3 weeks inundation, using a water depth of about
1 ft. Dry-up periods of less than a week were generally not effective
in restoring infiltration rates. Most of the infiltration recovery
seemed to take place in the second week of the dry up. Thus, the
infiltration recovery curve is S-shaped, which was confirmed with
infiltrometer measurements. The linear reduction in infiltration rate
during flooding and the S-shaped recovery during dry up are
schematically shown in Figure 13, where the infiltration rate at the
start of an inundation period was arbitrarily set at an index of 100.
Assuming full infiltration recovery after 12 days dry up in the summer
and 20 days in the winter, long-term infiltration rates (which include
the time that the basins are not inundated) were calculated for
different lengths of the inundation period. The resulting curves,
shown in Figure 14 with the infiltration rate at the start of the
inundation period again taken as 100, have flat peaks with a maximum
long-term infiltration rate, or hydraulic loading, at inundation
periods of about 24 days in the summer and 30 days in the winter.
Because the infiltration recovery may be slower after a long inundation
than a short inundation, it would probably be best to use inundation
periods that are on the left side of the maximum. Thus, maximum
long-term infiltrations will probably be obtained with inundation
periods of 16 to 24 days, alternated with dry-up periods of 12 days in
the summer and 20 days in the winter.
29
-------
O
UJ
Ijj
u_
I
I
1
BASIN 2
-10 -20 -30 -40 -50
PRESSURE HEAD AT 100-cm DEPTH IN cm WATER
Figure 11. K as a function of soil-water pressure-head for soil in basin 2.
-------
0.4
0.3 -
0.2 -
o.
H
O 0.4
o:
0.3
O
> 0.2
LU
H O.I
O
O
o: 0.4
LJ
0.3
0.2
O.I
»•—I—••
60-cm DEPTH
a
•H
CO
cfl
,0
•H
O
O
M-l
•O
tti
0)
D
CO
CO
OJ
M
a,
c
CL)
J-l
c
o
o
C
cu
ai
C
o
•H
0)
Pi
oo
•T-l
-20 -40 -60 -80 -100
PRESSURE HEAD.cm
31
-------
soo
tr
H>
LsJ
>
h-
UJ
CC
^/ /*•"
o / /^
«/^
II
//
.^i/
10
20
30
40
DAYS
Figure 13. Schematic presentation of infiltration decrease during inundation and
recovery during dry up.
-------
100
CO
LO
cc
UJLJ
HH
I <
occ
50 -
CC.—
I—I—I—I—I—\—I—
0
10 20 30
INUNDATION PERIOD , DAYS
40
Figure 14. Long-term infiltration rate in relation to length of inundation periods.
-------
The average height of the maxima of the summer and winter curves in
Figure 14 is at an infiltration index of about 45. The Infiltration
rate at the start of a new flooding period is of the order of 2.5
ft/day, using a water depth of 13 inches in the basins. Thus, the
average long-term infiltration rate for the inundation schedules
mentioned in the previous paragraph will be about 1.12 ft/day. This
yields an annual loading rate of 410 ft (410 acre-feet per acre per
year). This loading was essentially achieved in 1970, when the average
accumulated infiltration for the six basins was 400 ft.
Since complete infiltration recovery to the original values may not
always be obtained, it may be necessary to periodically include an
extra long dry-up period. The best time for such a dry up would be
when drying conditions are favorable, such as in the summer or during
periods of low rainfall.
The gravel layer clearly yielded lower infiltration rates in the basins,
contrary to what has been observed for other recharge facilities such
as the Peoria recharge pit in Illinois and the Whittler Narrows recharge
basin near Los Angeles. Covering the basins with a gravel layer is,
therefore, not recommended for systems similar to the Flushing Meadows
Project.
Vegetation in the basins yielded higher infiltration rates in the
summer and fall, when the vegetation had reached a mature stand.
However, the water depth in vegetated basins is restricted to avoid
complete submergence of the vegetation. Thus, the higher infiltration
rate of the vegetated basins could also have been obtained with bare
soil basins, simply by increasing the water depth. Also, vegetated
basins require a sequence of short, frequent, and shallow inundations
in the spring to get the vegetation established, whether from seed or
from winter dormancy. The total infiltration during such a sequence
will tend to be less than that for a sequence of longer inundation and
dry-up periods, as could be used with a bare soil basin. Thus, bare-
soil basins may yield equal or even higher annual infiltration amounts
than the vegetated basins, because of the greater water depths and
longer spring inundation periods that can be employed.
Vegetation may be desirable in areas with high rainfall or where the
waste water is supplied with sprinklers, to protect the surface of
the soil against the impact of the water drops. Also, vegetation may
be preferred because of considerations other than infiltration rates,
for example, nitrogen removal from the soil or utilization of the
infiltration basins for agricultural production. Thus, the choice
between vegetated or nonvegetated basins is governed by a number of
factors, which should be considered for each individual system.
Scraping, sweeping, harrowing, or disking the basin bottoms is necessary
when sludge accumulations restrict the infiltration rates, and particu-
larly the infiltration recovery during dry ups. "Shaving" the bottom
-------
with a front-end loader or sweeping with a power sweeper effectively
removes dried sluage flakes. If the sewage effluent has a low suspended
solids content (for example, less than 10 mg/1), sludge removal and
harrowing may only be necessary once a year or once every few years.
Suspended solids concentrations of 50-100 mg/1 in the effluent may
require sludge removal every few months.
-------
SECTION VI
AQUIFER STUDIES
1- Hydraulic Conductivity of Aquifer and Flow System
Analog Studies. Hydraulic properties of the aquifer beneath the
infiltration basins were evaluated so that the flow system below the
water table could be determined and the underground detention times of
the renovated water from the various observation wells could be
estimated. Also, knowledge of the hydraulic properties of the aquifer
was necessary for the design of the large-scale effluent recharge and
reclamation system in the Salt River bed envisaged for the future.
Because of the stratified nature of the alluvial deposits (Table 1),
the aquifer could be expected to behave as an anisotropic medium with
the hydraulic conductivity in horizontal direction, K , greater than
the hydraulic conductivity in the vertical direction, K . Values of
IL and K were obtained from the response of the water levels in ECW
and WCW to infiltration of effluent in the basins.
The water levels in the observation wells rose with the start of a new
inundation period, reached a pseudo-equilibrium level after a few days,
and declined slowly as the infiltration rates in the basins decreased
due to soil clogging. When the inundation of the basins was stopped,
the water levels in the observation wells dropped in typical "decay"
fashion to about the same levels as before the inundation period. This
behavior is illustrated in Figure 15 for ECW. The drop of the water
level in ECW between 3 and 5 December was caused by reduction in
infiltration rate due to changing the water depth in the basins from
13 to 7 inches on 4 December. The elevations in Figure 15 are with
respect to a local bench-mark. The elevation of the bottom of the
recharge basins with respect to the bench-mark is about + 8 ft.
For the inundation period 22-27 September 1967, the infiltration rates
were high and remained essentially constant at about 3.5 ft/day (average
for the 6 basins). Thus, conditions during this period were favorable
for evaluating 1C and K from the water level rises at pseudo-
equilibrium. These rises were 2.7 ft for ECW, and 0.7 ft for WCW.
To evaluate K and K of the aquifer from the pseudo-equilibrium water
level rises in ECW and WCW, a vertical cross section of the aquifer
was simulated on a resistance network analog, assuming a horizontal
water table and taking the impermeable boundary at 247-ft depth. The
cross section was taken normal to the basins, in the line of observation
wells 1 through 8 (Figure 1). The portion of the horizontal water table
beneath the recharge basins was treated as a source, and the rest of the
water table as a sink. The aquifer was assumed to be of infinite
lateral extent, which was represented by a termination zone according
to the procedure described in a previous paper (6). Assuming that the
aquifer was uniformly anisotropic, different ratios of K,/K were
37
-------
U)
Co
CO
o
o
_J
Id
LJ
U.
O
r -2
AVERAGE INFILTRATION
RATE
WATER LEVEL
ECW
WATER DEPTH
CHANGE FROM
FEET,
20 25
NOVEMBER
30
10
DECEMBER
1969
DAY
EL EVA
1
INUNDATION
STARTED
i
13 TO 7
i
INCHES
r ^
1 . ,
INUNDATION
STOPPED
• 1 . , , , I ,
15
Figure 15. Response of water level in ECW to infiltration.
-------
simulated until the voltage rises at the points representing the 30-ft
and 100-ft depths in the center of the system (corresponding to the
bottom of ECW and WCW, respectively) agreed with the water level rises
of 2.7 ft and 0.7 ft, respectively, measured in the field. This
agreement was achieved when K /K was taken as 16.
With the proper value of the K /K -ratio thus established, the pseudo-
equilibrium shape of the ground water mound for the average infiltration
rate in the basins was determined according to the analog procedure
presented in an earlier paper (3). In this analysis, the flux at the
water table mound beneath the basins was considered uniformly distri-
buted and taken as the average infiltration rate for the gross area of
the recharge basins, which includes the dry strips between the basins.
Since the x\ridth of each basin was 20 ft, the average rate for the six
basins was calculated as 120 x 3.5/220 =1.91 ft/day, where the number
3.5 indicates the average infiltration rate for the six recharge basins.
The electrical current in the analog model at pseudo-equilibrium corres-
ponded to the average recharge rate of 1.91 ft/day. Thus, the K-
components could be calculated from the horizontal and vertical resis-
tance values in the analog model in accordance with the procedure in
(6). The resulting values were IL = 282 ft/day and K =17.6 ft/day.
Equipotentials and streamlines for the flow system at pseudo-equilibrium
with an average infiltration rate of 1.91 ft/day are shown in Figure 16.
Points A and B in this system refer to the bottoms of the 30-ft ECW
and the 100-ft WCW, respectively. The potential at these points in the
analog model indicated water level rises of 2.62 ft in ECW and 0.78 ft
in WCW. These values closely agree with the actual rises of 2.7 ft
and 0.7 ft observed in the field. Thus, the 16-fold ratio between
horizontal and vertical conductivity evaluated for the horizontal water
table assumed in the first analysis, and the resulting values of TL and
K , are valid.
v
Field Studies. The values of 1C and K were also evaluated from field
measurements of K at the observation wells. The hydraulic conductivity
K of the material at the bottom of each observation well was determined
with the tube-method developed by Frevert and Kirkham (9). With this
method, the water level in the well is lowered a few feet below its
equilibrium position and the subsequent rate of rise is measured. The
value of K is then calculated from the rate of rise and a factor
expressing the geometry of the flow system. The results (Table 3) show
that a wide range of values was obtained.
Previous measurements with the double-tube method in the Salt River bed
indicated that K of the sandy material usually ranges from 1.7 ft/day
to 30 ft/day, depending on the texture of the sand (5 and references
therein). Thus, the values of K in Table 3 that are 34 ft/day or less
apparently belong to sandy materials, whereas those of 173 ft/day and
more would pertain to the more gravelly materials. The average of the
values of 34 ft/day or less in Table 3 is 12.2 ft/day, which can be
considered as the average K of the sandy materials. Similarly, the
average K of the gravelly materials can be calculated as 458 ft/day.
39
-------
Table 3. Results of hydraulic conductivity measurements.
Well Depth, in feet K, in feet per day
1 20 2.5
1-2 20 173
ECW 30 34
WCW 100 270
5-6 20 0.1 (estimated)
7 20 775
8 20 620
-------
Figure 16. Ground water flow system at steady-state during recharge.
-------
Assuming that the geologic profile consists of a regular succession of
equally thick sandy and gravelly layers, the hydraulic conductivity in
horizontal direction can be calculated as the arithmetic mean of the
K-values of the two materials, and the hydraulic conductivity in vertical
direction as the harmonic mean, of these two values (7). Since the
average value of K for the sandy materials is 12.2 ft/day, and that
for the gravelly material 458 ft/day, this procedure yields K = 236
ft/day and K =23.7 ft/day. This is in good agreement with the values
of 282 ft/day and 17.6 ft/day, respectively, obtained with the analog
procedure, despite the fact that replacement of the actual profile by
a profile of alternating sand and gravel layers of equal thickness is
a drastic simplification.
The ratio K /K calculated from the field measurements is lower than the
ratio obtainedVwith the analog (10 versus 16). This may be due to the
fact that in the calculation of K, and K from the hydraulic conductivity
measurements at the observation wells the individual layers were con-
sidered isotropic. In actuality, however, each layer may be anisotropic
in itself due to particle orientation and microstratification. This
anisotropy was demonstrated by double-tube measurements on sandy deposits
in the Salt River bed, which showed that although the sand was seemingly
uniform, the hydraulic conductivity in horizontal direction was 7 times
that.in vertical direction (5). Nevertheless, the good agreement
between the values of K. and K evaluated with the analog and with the
n v
field measurements, which are two completely independent and widely
different techniques, lends validity to the results of both.
The flow system below the water table shown in Figure 16 was used to
predict underground detention times of the renovated water pumped from
ECW and WCW (see next section). The flow system of Figure 16 was also
used to evaluate the effective transmissibility of the aquifer for
ground water recharge. This effective transmissibility was then used
in an analog model of an operational system to predict water table
positions and underground detention times for various designs (See VIII.
DESIGN AND OPERATION OF LARGE-SCALE SYSTEM) .
2. Underground Detention Times
The hydraulic gradients along the streamlines in Figure 16 can be used
to predict underground velocities, and hence travel times, of the
renovated water. This procedure was applied to the vertically downward
streamline in the center of the system, which is the symmetry line of
Figure 16. The hydraulic conductivity along this line is 17.6 ft/day.
The underground detention time is inversely proportional to the
infiltration rate and directly proportional to the porosity of the
soil material. Assuming an infiltration rate of 1 ft/day, and a poro-
sity range of 20% to 30%, the calculations yielded travel times below
the water table of 9.6 to 14.4 days for the water pumped from ECW
(point A in Figure 16), and 189 to 284 days for that pumped from WCW
(point B in Figure 16). The time for the water to travel from the
-------
bottom of the basins to the water table is about 2 days per 1 ft/day
infiltration rate. This time must be added to the travel times below
the water table to get the total underground detention time.
In calculating the underground detention times from the gradients and
hydraulic conductivity in the flow system of Figure 16, the assumptions
were made that (1) the aquifer is uniformly anisotropic (thus ignoring
the effect of any distinct layers of low permeability that may have an
overriding effect on the travel time), (2) the difference in density
between the reclaimed water and the more saline native ground water has
no effect on the flow system, and (3) dispersion can be ignored and the
native ground water is displaced in piston-like manner.
The calculated underground detention times can be checked by observing
the arrival of nitrate peaks in the renovated water from the observation
wells. The nitrogen in the sewage water held as capillary water in the
upper portion of the profile during the dry-up period is completely
nitrified. When a new inundation period is started, this nitrified
capillary water is pushed ahead of the newly infiltrated water, and
when it arrives at the intake of an observation well it causes
temporarily a high nitrate concentration, or nitrate peak, in the
renovated water.
The underground detention times can also be checked by monitoring the
salinity level of the observation wells. Since the native ground water
has a salinity that is 2 to 4 times as high as that of the sewage
water, a reduction in the salinity of the well water from a few thousand
parts per million to about 1000 parts per million indicates that native
ground water has been displaced by renovated sewage water.
The nitrate-peak technique was used for determining underground
detention times for the water from ECW. Table 4 shows how many days
after the start of a new inundation period nitrate peaks were observed
in ECW in 1969. The average infiltration rates during those days are
shown in the third column. The last column of Table 4 shows the under-
ground detention time as if the infiltration rate had been 1 ft/day.
The average of the underground detention times per unit infiltration
rate, i. e., 11.5, is close to the predicted value of 11.6 days for
20% porosity.
Since the calculated underground travel time for the water from WCW
is 191-286 days (adding 2 days for travel from the basin bottom to the
water table) for an infiltration rate of 1 ft/day, displacement of
the native ground water by renovated sewage water at the bottom of the
well should take place after 191 to 286 ft of water have infiltrated
in the basins. At the end of July 1969, the salt content of WCW-water
started a gradual drop from its 3000-4000 ppm range, indicating that
renovated water had arrived at the bottom of the well. The average
accumulated infiltration for the six basins since the start of the
project was 315 ft on 19 July 1969. This is higher than the predicted
range of 191-286 ft, but considering the simplifications and assump-
tions made in the calculations, the agreement is still reasonable.
-------
Table 4. Occurrence of distinct nitrate peaks in ECW in relation to start
of inundation period and average infiltration rate.
Day
N0»- peak
was observed
17 Jan 1969
12 Feb 1969
13 Mar 1969
25 Sep 1969
21 Oct 1969
26 Nov 1969
30 Dec 1969
Number of days
after start of
inundation
11
9
10
8
6
5
7
Average
infiltration
rate, ft/day
0.83
1.12
1.00
1.65
1.98
2.50
1.94
Number of days
times
infiltration rate
9.1
10.1
10.0
13.2
11.9
12.5
13.6
Av.
11.5
-------
The long underground detention time of almost 2 years for the renovated
water from WCW is due to the anisotropic nature of the aquifer, which
offers much more resistance to water movement in vertical direction
than in horizontal direction.
The observation wells 1, 1-2, 5-6, 7, and 8 were installed in April
1968. At that time, these wells were already yielding renovated sewage
water, with the exception of well 8. The underground detention time
for the water from wells 1-2 and 5-6 is about half that for ECW. The
underground detention time for the water from wells 1 and 7 is 1 to 2
months, as deduced from the nitrate behavior in these wells. Well 8
continued to yield native ground water until January 1970, when the
salt content started to drop from about 2000 ppm to about 1500 ppm.
Thus, the underground detention time for renovated sewage water from
well 8 is almost 2 1/2 years.
3. Water Table Response
Profiles of water levels in the observation wells are shown in Figure
17, which applies to the inundation period 21 November-8 December 1969.
Since x^ells 1, 1-2, 5-6, 7, and 8 are all 20 ft deep, the water level
readings from the 30-ft-deep ECW were corrected so that they would
also apply to the 20-ft depth. This was done using the vertical gradient
in the center of the flow system as evaluated by analog (Figure 16).
Thus, the water levels as shown in Figure 17 all apply to the piezo-
metric head at 20-ft depth, which is about 10 ft below the water table.
The water level profile on 21 November, prior to the inundation period,
is essentially horizontal, indicating static ground water conditions
in the north-south direction. The horizontal profile also indicated
negligible effect of seepage from the effluent stream north of the
project area (Figure 1) on the ground water flow system. The water
level in the stream is about 2 1/2 ft above the static ground water
level. The profile on 1 December shows the water levels at their
highest positions, which is also the pseudo-equilibrium level for the
first part of the inundation period. Because basins 3 and 4 exhibited
the highest infiltration rates of this period, the water level in ECW
is probably higher than it would have been in case the infiltration
rates had been the same for all basins. The profile on 8 December
shows the water levels in the wells at the end of the inundation period,
after the water depth in the recharge basins was changed from 13 to
7 inches on 4 December. On 11 December, 3 days after the flow into
the basins was stopped, the water level profile in the observation
wells was essentially horizontal, and on 15 December it had returned
to about the same static position as before the start of the inundation
period.
Comparing the profiles on 1 and 8 December shows that the water level
in well 5-6 is slower to rise and slower to fall than, for example,
the water level in well 1-2. This slower response of well 5-6 is due
to the materials of relatively low hydraulic conductivity around this
well (Table 3) .
-------
8
CQ
o
o
-J
LU
UJ
Lu
WELL NO. !
1-2 ECW 5-6
0
I- -I
UJ
_J
UJ
-2 -
STREAM LEVEL
21 NOV.
1969
8
I
I
200 80 0 80 200
DISTANCE IN FEET FROM CENTER
400
Figure 17. Water-level profiles in observation wells during inundation and dry up.
-------
The rise of the water levels to semi-equilibrium positions and the
relatively rapid decline to the original static levels after inundation
is stopped are typical for the Flushing Meadows conditions where the
aquifer is of considerable lateral extent and of high hydraulic
conductivity in horizontal direction.
4- Effect of Recharge on Hydraulic Conductivity of Aquifer
Since the water table rises during infiltration were small compared to
the total height of the flow system, the water level rises in the
observation wells at pseudo-equilibrium conditions should be in direct
proportion to the infiltration rate in the basins. Thus, the water
level rise per unit infiltration rate should remain constant if the
hydraulic conductivity of the aquifer does not change. An increase in
the water level rise per unit infiltration rate would indicate a
decrease in the hydraulic conductivity of the aquifer, and vice versa.
To study the possibility of a change in the hydraulic conductivity of
the aquifer, the water level rise at pseudo-equilibrium in ECW was
measured each fall and divided by the average infiltration rate in the
basins for the flooding period in question. The following results
were obtained
Static water level in ECW Equilibrium rise per
prior to recharge unit infiltration rate
£t local BM £t/ft/day
September 1967 - 6.0 0.77
October 1968 - 2.0 0.71
November 1969 - 1.5 0.72
These data indicate that from 1967 to 1968, the hydraulic conductivity
of the aquifer actually increased. This increase may, at least in part,
be due to the 4-ft rise in the static ground water level from 1967 to
1968, which could have included some permeable strata into the aquifer.
From 1968 to 1969, the hydraulic conductivity of the aquifer remained
essentially constant.
-------
SECTION VII
WATER QUALITY STUDIES
Composite samples of the secondary sewage effluent, which was the
influent for the Flushing Meadows Project, were obtained almost daily
with a continuous sampler placed on the head box at the upper end of
one of the basins (Figure 1) . The sampler was of the dipstick type,
taking approximately 0.5 ml effluent every minute and yielding about
0.7 liter (about 1/5 gallon) per 24 hrs . The bottle collecting the
sample was in a thermally insulated box cooled with icepacks.
Grab samples were obtained from the observation wells. The wells
were bailed or pumped several times before collecting the sample to
make sure that a fresh sample was obtained. The wells within the
basin area (1-2, ECW , and 5-6, Figure 1) were usually sampled daily.
The other wells were sampled weekly, monthly, or bimonthly. Well 1-2
yielded water that had mainly infiltrated in basins 1 and 2. The
water from ECW came essentially from basins 3 and 4, and water from
well 5 from basins 5 and 6, provided that all basins were inundated
at the same time.
All samples were transported to the U. S. Water Conservation Laboratory,
a distance of about 25 miles, and kept under refrigeration until
analysis . The chemical and bacteriological analyses were performed
using the procedures of Standard Methods (11) as a guide. Although
analyses were started in 1967, the analytical laboratory did not
function smoothly until 1968. Therefore, results obtained after
1 January 1968 will be reported only.
1 . Biochemical and Chemical Oxygen Demand
The BODr of the secondary effluent was usually in the 10-20 ppm range.
After traveling through 30 ft of sand and gravel (10 ft above and 20 ft
below the water table), the BOD_ of the renovated water, as sampled
from ECW, ranged from 0 to 1.2 ppm in 1968 with an average of about
0.3 ppm. In 1969, the BOD of the samples from ECW did not exceed
2 ppm and was less than 0.5 ppm for the summer months (Figure 19) .
The BOD-values of the renovated sewage water were determined by the
laboratory of the 91st Avenue Phoenix Sewage Treatment Plant. Although
the BOD of the renovated water was essentially zero, the organic
carbon content of the renovated water from ECW was in the 2-10 ppm
•range, as was determined later with the total carbon analyzer (Beckman
model 915) .
The chemical oxygen demand was determined with the dichromate technique.
The COD of the effluent was determined on the supernatant liquid in the
sample bottle so that it would be characteristic of the effluent as it
would move into the soil, relatively free of suspended material. The
COD of the effluent usually was in the 30-60 ppm range with the values
in the first half of the year being somewhat higher than in the second
-------
VJ1
o
70
60
50
40
Q 30
O
O
20
10 -
INUNDATION PERIODS
„-*-., li
.-""" ^ . j!
\ « i
M\ r
^i
• EFFLUENT
- EAST CENTER WELL
70
60
\ I
\ I
\ J
\ i H
\ I
50
40
30
20
10
Figure 18. COD of secondary effluent and of renovated water from East Center Well in 1968.
-------
__.__ COD OF EFFLUENT
COD OF ECW
0
NOV.
DEC.
Figure 19. COD of secondary effluent and of renovated water from ECW, Well 1-2, and Well 5-6 in 1969.
-------
half of the year (Figures 18 and 19). This is probably due to the
poorer quality of the effluent in winter and spring.
The COD of the renovated water from ECW was generally in the 10-20
ppm range (Figures 18 and 19). The length of the inundation period and
the surface condition of the basins (grass, gravel and bare soil)
apparently had no significant effect on the COD of the renovated water.
Figure 19 shows that the COD of the renovated water from well 1-2 was
slightly higher than that of ECW. This may be due to the fact that
well 1-2 is only 20 ft deep, compared to 30 ft for ECW. The COD of
the water from well 5-6 was slightly lower than that of ECW, which
could be due to the finer-textured soil around the bottom of this well.
The COD of the renovated water from wells 1, 7, 8, and EW, was about
the same as that for wells 1-2, ECW, and 5-6, i. e., an average of
14 ppm (Table 5). Thus, the additional underground travel of the
renovated water to the "outlying" wells had little or no effect on
COD.
The COD of the native ground water was slightly higher tiian that of the
renovated sewage water, i „ e., an average of about 20 ppm in 1969 for
the three wells yielding native ground water (Table 5). The 91st
Avenue well mentioned in this table is an irrigation well located
about 1 1/2 miles east of the Flushing Meadows Project. The well
pumps from a depth of about 100 to 200 ft.
2. Nitrogen
Nitrogen content of the water samples was determined with the brucine
method for nitrate, the diazotization method for nitrite, the Nessler
technique for ammonium (later supplanted by the distillation technique),
and the Kjeldahl method for organic nitrogen.
The total nitrogen content of the effluent generally ranged from 20 to
35 ppm with the concentrations in the summer being somewhat lower than
in the winter (Figures 20 and 21). Almost all the nitrogen in the
effluent was in the ammonium form. The organic nitrogen content of
the effluent was about 1 ppm (about 3 ppm when the suspended solids
content was high), and the nitrate-nitrogen concentration was about
0.1 ppm.
If short, frequent inundation periods were used (for example, 2 or 3
days wet, 2 to 5 days dry), almost all the nitrogen of the effluent
was converted to the nitrate form in the renovated water (Figure 20).
If longer flooding periods were used (for example, 2 to 4 weeks wet
and 1 to 2 weeks dry), the nitrate-nitrogen concentration in the
renovated water was considerably lower (April-May and September-December,
Figure 20), except for distinct "peaks." These nitrate peaks occurred
5 to 10 days after the start of a new inundation period and they are
apparently caused by the arrival of nitrified effluent water that was
held as capillary water in the soil during the preceding dry-up period.
-------
Table 5. COD in mg/liter for various wells
Date
1968
23 Jan
20 Feb
12 Mar
26 Mar
2 May
6 Jun
16 Jul
15 Aug
10 Sep
10 Oct
14 Nov
18 Dec
1969
1969
1969
1969
1969
1969
1969
1969
1969
1969
1969
1969
Av. 1969
la
10-12
12
12
12
6
12
17
16
25
23
16
8
14
7a EWa
12-16
12
10
16
6
7
12
16
12
25
18
16 15
13 13
14 14
wcwb
11-23
31
22
19
24
20
35
37
29
29
21
13
25
8b
10-12
11
12
16
8
7
8
27
22
30
26
20
17
17
91st Avenue
10-17
24
14
14
11
29
21
19
renovated sewage water
native ground water
53
-------
40
30
VJl
f-
CL
Q_
S
bJ
CD
O 20
CC
10
INUNDATION PERIODS
•!
o o TOTAL EFFLUENT NITROGEN
• • NITRATE N OF ECW
» « * •- AMMONIUM N OF ECW
40
30
20
10
Figure 20. Total nitrogen of secondary effluent, and nitrate N and ammonium N in renovated water
from ECW in 1968.
-------
45
40
35
30
LU 20
O
O
2f 15
10
INUNDATION PERIODS
7
BASINS 1,2,3,84
BASINS 586
—•— TOTAL N OF EFFLUENT
—•-- NO - N OF EC W
O
NH4-N OF ECW
JAN.
NOV.
DEC.
Figure 21. Total nitrogen of secondary effluent, and nitrate N and ammonium N in renovated water
fVnm F.CW in 1969.
-------
Because of the aerobic conditions in the upper soil layers during dry
up, the ammonium nitrogen in the effluent and the ammonium adsorbed to
the soil and organic matter could be converted to nitrate. When a new
flooding period was started, the nitrified capillary water was then
pushed downward by the newly infiltrating water. Arrival of this
nitrified water at the intake of an observation well caused a nitrate
peak in the water sampled from that well.
The nitrate behavior in the renovated water from ECW in 1969 was
similar to that in 1968 (Figure 21). During the short, frequent inun-
dation periods from April to about July, most of the nitrogen was
converted to the nitrate form. With the long inundation periods in
the first and last parts of the year, nitrate-nitrogen contents were
close to zero, except for the peaks which again occurred 5 to 10 days
after the start of the inundation periods.
Nitrogen can be removed from waste water moving through soil by various
processes, including adsorption of ammonium to the clay and organic
fraction of the soil, fixation of ammonium by the organic fraction,
fixation of nitrogen in tissue of soil microorganisms, nitrogen uptake
by vegetation, volatilization of ammonia, and denitrification. Of
these processes, only uptake of nitrogen by crops (if the crop is
harvested), volatilization of ammonia, and denitrification cause a net
removal of nitrogen. The other processes merely store nitrogen in
the soil.
Quantitative analysis of the nitrogen-removing capability of each
of these processes at the Flushing Meadows Project indicated that the
sustained removal of nitrogen observed at the project must mainly be
attributed to denitrification. This process requires the simultaneous
occurrence of nitrates and organic carbon under anaerobic conditions.
The main end product of denitrification is free nitrogen gas, which
escapes to the atmosphere or dissolves in the downward moving water.
That storage of nitrogen in the soil was insignificant became evident
at the end of 1969, when the soil of the recharge basins was analyzed
for total nitrogen. The analyses showed that the loamy sand contained
only about 0.1 mg nitrogen per gram of dry soil more than the virgin
soil outside the recharge basin. This nitrogen could account for only
a small fraction of the nitrogen removed from the 450 ft of sewage
effluent which the recharge basins had received by the end of 1969.
As regards the mechanism of nitrogen removal by denitrification, when
a new flooding period is started, there is entrapped air in the soil
for nitrification of ammonium during the initial part of the flooding
period. As the oxygen will be consumed, however, an aerobic pocket
will begin to develop where nitrate and organic carbon can both be
present, creating conditions favorable for denitrification. With
continued flooding, all oxygen will eventually be used up and the
nitrogen of the effluent will stay in the ammonium form, which can be
adsorbed by the clay and organic fraction of the soil.' If the inunda-
tion is not stopped before the cation exchange complex in the soil is
-------
saturated with ammonium, increased ammonium levels in the renovated
water can be expected. When the inundation is stopped, air will enter
the soil and the resulting aerobic conditions in the upper soil layers
will enable nitrification of the adsorbed ammonium. Some of the
nitrates formed in this process may diffuse to anaerobic micro-
environments in the same soil and denitrification can occur if organic
carbon is also present. Some of these nitrate ions may also mix later
with the newly infiltrating water when a new flooding period is
started and move down to anaerobic environments where denitrification
may occur.
Conditions for denitrification tend to be more favorable in root zones
than in soils without roots (13). This is because roots contribute
organic carbon to the soil by direct exudation from living roots
and decomposition of dead roots. Also, the oxygen uptake by the
roots will help to create anaerobic environments necessary for
denitrification.
Evidence that denitrification is more complete below the vegetated
basins than below the nonvegetated basins of the Flushing Meadows
Project is presented in Figure 20. This figure shows that the nitrate-
nitrogen levels in the renovated water from ECW during long flooding
periods were lower in the fall when the bermudagrass was fully
developed, than in April and May when basins 3 and 4 from which ECW
mainly receives its water were still bare. Also, the nitrate-nitrogen
levels in the water from well 1-2, which receives water from the
nonvegetated basins 1 and 2, were higher in the fall of 1968 than the
nitrogen levels in the water from ECW and well 5-6 (Figure 22). Well
5-6 receives its water mainly from basins 5 and 6, which were also
seeded to bermudagrass in the spring of 1968. Thus, nitrogen removal
from waste water seems to be greater below vegetated basins than below
nonvegetated basins.
In 1969, the nitrate peaks and nitrate levels between peaks of the
renovated water from well 1-2 were higher than those from ECW and
well 5-6 in the beginning of the year (Figure 23). However, later
in the year, when basin 1 was covered with a full stand of native
grasses, the nitrate peaks and the nitrate concentrations between the
peaks for well 1-2 were more in line with those for ECW and well 5-6.
This is another indication that the development of vegetation may
have contributed to increased denitrification in the soil.
Since direct uptake of nitrogen by the plant roots is insignificant
when compared to the total nitrogen loading of the recharge basins,
which is about 100 Ibs of nitrogen per acre per day during flooding,
the lower nitrate-nitrogen levels in the renovated water below
vegetated basins must be attributed to more complete denitrification
in the root zone.
The ammonium nitrogen concentrations in the renovated water were
initially low but gradually increased until about July 1969
-------
50
40
vn
OD
CL
Q_
30
LU
O
O
K
t 20
UJ
h-
<
a: 10
INUNDATION PERIODS"^
. . EAST CENTER WELL
o -o WELL N° i -2
« » • WELL Ne 5-6
968
Figure 22. Nitrate N in renovated water from wells 1-2, ECW, and 5-6 in 1968.
-------
INUNDATION
PERIODS
JAN.
SEPT.
OCT.
NOV.
DEC.
Figure 23. Nitrate N in renovated water from wells 1-2, ECW; and 5-6 in 1969.
-------
(Figures 20 and 21). This increase was apparently due to saturation
of the cation exchange complex in the soil with ammonium, particularly
during sequences of long flooding periods. A sequence of short
flooding periods should restore the capacity of the soil to adsorb
ammonium, because the predominantly aerobic conditions in the upper
layers of the soil during short flooding sequences would favor nitri-
fication of'the adsorbed ammonium. That this may indeed occur is
evidenced in Figure 21, which shows that when the sequence of long
flooding periods used from September 1968 to April 1969 was changed
to a sequence of short flooding periods, the ammonium nitrogen concen-
trations in the renovated water from ECW began to decrease in June
1969 (the delay is probably due to the underground detention time and
the time necessary for a population of nitrifying bacteria to develop).
Another factor that may have contributed to the 1969 summer reversal
of the upward trend in the ammonium nitrogen concentration of the ECW
water is that in the summer, drying conditions are better and the upper
soil layers have a lower water content than in the winter. This lower
water content causes an increase in the oxygen diffusion rate and
hence in the depth to which the soil profile becomes aerobic during
dry-up periods. Thus, adsorbed ammonium can be nitrified to greater
depths and the adsorptive capacity for ammonium can be better restored
in the summer.
Nitrate and ammonium nitrogen concentrations for the wells outside
the basin area are shown in Tables 6 and 7, respectively. The nitrate-
nitrogen levels in the renovated sewage water from the outlying wells
also show the effect of the length of the inundation periods. The
change to longer inundation periods in July 1969, for example, must
have been responsible for the drop in nitrate levels in the renovated
water from wells 1 and 7 in August and September 1969, respectively.
Ammonium levels in the renovated water from wells 1 and 7 were lower
than those for the wells inside the basin area, indicating additional
removal of ammonium as the renovated water travelled laterally beneath
the water table.
Organic nitrogen in the renovated water was about 0.9 ppm or less for
the wells inside the basin area, and about 0.5 ppm for wells 1 and 7.
Nitrite-nitrogen concentrations in the renovated sewage water were
generally in the 0-1 ppm range, with the values being close to zero
most of the time.
The results of the nitrogen studies show that the form and concentration
of the nitrogen in the renovated water can be controlled by the length
of the inundation period and, to a lesser extent, by the use of vegeta-
tion in the basins. If short inundation periods of a few days each
are used, almost all the nitrogen in the sewage effluent is converted
to the nitrate form in the renovated water. With very long inundation
periods, for example, several months, the nitrogen will mostly stay in
the ammonium form. With inundation periods of a few weeks, almost zero
nitrate-nitrogen concentrations can be expected in the renovated water,
after distinct nitrate peaks which occur when nitrified effluent held
60
-------
Table 6. Nitrate nitrogen concentrations in mg N per liter for various wells.
Date
1968
23 Jan
20 Feb
12 Mar
26 Mar
2 May
6 Jun
16 Jul
15 Aug
10 Sep
10 Oct
14 Nov
18 Dec
28 Nov
3 Dec
9 Dec
17 Dec
22 Dec
29 Dec
la 7a
0.2-16 1.3-3.2
1969 14.2 7.6
1969 16.0 15.6
1969 14.0 18.4
1969 11.5 11.8
1969 23.2 4.2
1969 21.6 10.8
1969 22.4 17.6
1969 2.0 14.8
1969 1.3 5.0
1969 0.3 3.4
1969 0.2 0.0
1969 0.2 0.5
1969
1969
1969
1969
1969
1969
8 EWS WCW 91st Avenue
0.2-4.2 0-7.6 6.4-6.7
2.0 3.5 6.0
2.1 4.1
0.3 4.7 5.8
0.3 1.8 5.9
0.2 2.2 4.3
0.2 2.8
0.1 3.5 5.4
0.5 3.5
0.3 3.3
0.1 2.5 6.4
0.3 3.7
0.5 4.4 5.5
8.5
6.4
3.3
0.2
4.8
8.3
renovated sewage water
native ground water
61
-------
Table 7. Ammonium nitrogen concentrations in mg N per liter for various wells.
Date
EW
WCW
91st Avenue
1968
0.8-2.9 0.2-0.5
0.1-0.5
0-0.6
0-0.6
23 Jan 1969 3.3
20 Feb 1969 2.8
12 Mar 1969 2.9
26 Mar 1969 2.8
2 May 1969 3.4
6 Jun 1969 2.8
16 Jul 1969 2.9
15 Aug 1969 2.0
10 Sep 1969 3.8
10 Oct 1969 4.1
14 Nov 1969 3.8
18 Dec 1969 2.8
28 Nov 1969
3 Dec 1969
9 Dec 1969
17 Dec 1969
21 Dec 1969
29 Dec 1969
1.1
1.2
0.9
0.7
0.4
1.3
0.2
1.2
1.5
1.5
1.1
0.2
0.2
0.2
0.1
0.2
0.0
0.0
0.4
0.1
0.1
0.1
0.1
2.2
2.8
3.6
4.9
3.5
5.6
0,1
0.4
0.0
0.0
0.0
0.0
0.5
0.0
0.1
0.1
0.1
0.0
0.1
0.0
0.0
0.0
renovated sewage water
native ground water
62
-------
as capillary water in the soil during dry up arrives at the intake of
the well. Also, nitrogen removal is greater below vegetated basins
than below nonvegetated basins, which is most likely due to increased
denitrification in the root zone.
Nitrogen removals of 90% after the passage of the nitrate peak have
been achieved. If the nitrate peaks are included, the total nitrogen
removal is probably of the order of 30%. The remaining 70% of the
nitrogen, however, is concentrated in the nitrate peaks which constitute
only a small volume of the renovated water. This small volume could be
recycled through the basins, it could be used where high nitrate levels
in the water are not undesirable (for example, irrigation of parks or
forage crops), or it could be pumped to a plant for denitrification
with methanol or other carbon source.
Continued use of long inundation periods for nitrogen removal apparently
causes the cation exchange capacity of the soil to become saturated with
ammonium, after which the ammonium levels in the renovated water begin
to increase. Additional lateral movement of the renovated water below
the water table may be effective in removing some of this ammonium.
However, it is probably more desirable to restore the nitrogen-removing
capacity of the recharge system by changing to a sequence of short,
frequent inundation periods. The resulting nitrification of the
adsorbed ammonium then restores the ability of the cation exchange
complex in the soil to adsorb ammonium. The same may be achieved by
using some extra long dry-up periods in the summer so as to allow the soil
to dry to lower water contents, which increases the oxygen diffusion
rate and hence the depth of the aerobic zone. Growing a crop may also
be effective in restoring the nitrogen-removing capability of the soil
system, not only because the crop will remove nitrogen from the soil
but the moisture uptake by the roots will increase the oxygen diffusion
rate and hence the depth of the aerobic zone. The root zone could also
contribute to denitrification.
3. Phosphate
Inorganic phosphorus in the secondary effluent and in the renovated
water from the observation wells was mainly in the ortho-phosphate
form and was analyzed with the Murphy-Riley technique. The results
(Table 8) indicate phosphate-phosphorus concentrations of about 13 ppm
in the effluent and essentially zero in the native ground water. When
renovated sewage water arrived at the observation wells, the phosphate-
phosphorus concentration increased to about 5-7 ppm, but remained
relatively constant thereafter. Preliminary indications are that
additional underground travel further reduces the phosphate-phosphorus
concentrations, but time will tell how long the low phosphate levels
in WCW and well 8, which began to yield renovated water after 1969,
will continue.
The phosphate removal at the Flushing Meadows Project seems to occur
gradually as the effluent water moves through the sands and gravels.
-------
Table 8. Phosphate-phosphorus concentrations in mg P per liter of effluent, renovated water, and native
ground water.
Effluent
1968
Feb 1969
May 1969
Aug 1969
Nov 1969
8-18
14
15
12
15
Renovated water
1-2 5-6
5-7 0.1-0.
8 0.1
6.5 0.5
3.5 1.2
6.5 1.5
ECW EW 1 7
5 2.4-5 0.05-0.15 0.5-0.8
7 12
5 1.7 1
4 24
6 1.7 1 3
Native ground water
8 WCW 91st Ave.
0-0.1 0-0.2
00 0
0.1 0 0.1
0.05 0.1 0
0.2 0.1 0.2
-------
This is probably because of the low clay content and absence of iron
and aluminum oxides in these materials. The removal of phosphate is
probably due to precipitation of calcium-phosphate complexes, which
are formed in the slightly alkaline and calcium-rich environment of
the effluent water as it moves through the sands and gravels.
Phosphates are the only constituents of the effluent that precipitate
in quantity in the soil. Thus, it would be of interest to calculate
how this precipitate may affect the porosity, and hence the hydraulic
conductivity, of the sands and gravels in the future. The phosphate
will be assumed to precipitate as oxy-apatite, formula 3Ca (PO ) .CAO,
of which the density is 3.2. Assuming that the amount of P precipitated
in the soil is 10 rag/liter of effluent, that this precipitation takes
place in a soil body 33 ft deep and 131 ft wide (considering two-
dimensional flow only), and that the infiltration rate is 330 ft/year,
the volume of apatite after 120 years of recharge would occupy about
0.5% of the total soil volume. At a pore space of 20%, the apatite
would then occupy about 2.5% of the pore space. This is not likely
to have a significant effect on the hydraulic conductivity. The above
calculations are based on a very simplified process, but they show
that it will probably take a very long time before phosphate accumula-
tion in the sands and gravels will have an effect on the hydraulic
performance of the system.
4. Boron
Boron concentrations, determined with the curcumin technique, were
generally in the 0.4-0.5 ppm range for the secondary effluent as well
as for the renovated water (Table 9). The boron concentration of the
native ground water was also in about the same range.
Boron is not removed as the sewage water moves through the sands and
gravels, indicating that such boron-fixing materials as iron and
aluminum oxides are essentially absent. Boron concentrations of more
than 0.5 ppm are undesirable for irrigation of boron-sensitive crops,
such as citrus. With the increased use of low- or no-phosphate
detergents, some of which contain borax or perborate, the boron concen-
tration of the sewage effluent, and hence that of the renovated water,
can be expected to increase in the future.
5. Fluoride
Fluoride concentrations, determined with the SPADN method, were about
4.5 ppm for the effluent (Table 10). This high concentration is
probably due to certain industrial wastes. More than 50% of the
fluoride ions are removed by the time the water has reached ECW, and
additional removal takes place as the water travels laterally to the
outlying well (Table 10) . The fluoride removal parallels the phos-
phate removal, indicating that the fluorides may be precipitated with
the calcium-phosphate complexes as fluorapatites.
-------
Table 9. Boron concentration in mg B per liter for effluent, renovated water,
and native ground water.
Date Effl.
1968 0.43-
0.50
23 Jan 1969 0.40
26 Mar 1969
25 Apr 1969 0.50
16 Jul 1969
15 Aug 1969 0.50
14 Nov 1969
18 Dec 1969 0.36
Renovated water
1 1-2 ECU 5-6 7 EW
0.43-
0.56
0.38 0.39 0.39 0.38
0.45 0.41
0.45 0.40 0.39
0.50
0.50 0.50 0.50 0.50
0.46 0.45 0.40 0.44
0.44 0.46 0.45 0.44 0.47
Native ground water
8 WCW 91st Ave
0.63 0.64
0.38 0.50
0.35
0.50
0.58
66
-------
Table 10. Fluoride concentration in mg F per liter for effluent, renovated water,
and native ground water.
Date
1968
23 Jan
20 Feb
26 Mar
6 Jun
10 Sep
10 Oct
14 Nov
18 Dec
Av.
1969
1969
1969
1969
1969
1969
1969
1969
1969
Effl.
5
5
2
5
4
5
4
4
4
.0-
.5
.7
.6
.3
.3
.3
.7
.5
Renovated
1
1.2-
1.7
1.3
1.3
1.5
1.5
2.1
2.0
1.8
2.2
1.7
1-2
2.1-
2.8
3.5
3.1
2.9
2.6
2.4
2.3
2.5
2.8
2.8
ECW
2.1-
2.5
1.9
2.1
2.6
2.3
2.6
2.6
2.7
3.1
2.5
water
5-6
1.8-
2.3
1.4
1.6
1.8
1.8
2.6
2.6
2.8
3.0
2.2
7
1.
1.
2.
1.
1.
1.
2.
2.
2.
2.
2.
EW
8-
9
0
9
5
2
7
9
3
7 2.6
1 2.6
Native ground water
8 WCW 91st Ave.
0
0
0
0
0
0
0
0
0
0
.8 0.8
.8 0.8
.8 0.8 0.6
.8 0.9 0.6
.6 0.8
.8 0.8
.7 0.6 0.6
.8 0.6
.8 0.7 0.6
.76 0.75 0.6
67
-------
6. Total Dissolved Salts
Total dissolved salts were determined by measuring the electrical
conductivities of the water samples, and multiplying millimhos per
centimeter by 640 to obtain ppm.
The salt concentration of the secondary effluent was about 1020 ppm,
that of the renovated water about 1060 ppm, and that of the native
ground water in the 2000-4000 ppm range (Table 11). The fact that
the salt content of the renovated water was about 4% higher than that
of the effluent may be largely due to evaporation from the water sur-
face in the basins and from the soil during dry up. The evaporation
from a free water surface in Central Arizona is about 6 ft/yr. Since
the average infiltration for 1969 was about 220 ft, evaporation
accounts for about a 3% salt increase. The remaining 1% may be
attributed to solution of salt in the soil caused by a lowering of
the pH due to CO -production by bacteria in the soil.
The salt content of WCW water began to drop in August 1969, indicating
the start of the arrival of renovated water. The gradual nature of
the reduction (Table 11) indicates that the renovated-water "front"
has become diffuse after the long time and distance of underground
travel.
The concentration of the main ions in the secondary effluent was as
follows:
ion
381
j
Cl~ 213
SO,™ 107
co3 o
Na+ 200
Ca 82
Mg 36
K+
Total 1,027
The ionic composition of the renovated water can be expected to be
about the same. The sodium adsorption ratio (SAR), defined as
_l_ / j I j i
Na / V(Ca + Mg )/2 with the ionic concentrations in miIll-
equivalents per liter, was about 4.6. This is well below the range
-------
Table 11. Total salt concentration in effluent and in water from wells, mg/liter.
Date
1968
13-14 Jan 69
23 Jan 69
10-11 Feb 69
14-17 Feb 69
10-11 Mar 69
26 Mar 69
14-15 Apr 69
28-29 Apr 69
6 May 69
26-27 May 69
6 Jun 69
30 Jun 69
8-9 Jul 69
28-29 Jul 69
5-6 Aug 69
19-20 Aug 69
9-10 Sep 69
25-26 Sep 69
8-10 Oct 69
25-26 Oct 69
14 Nov 69
24 Nov 69
5-8 Dec 69
29-30 Dec 69
Av. 1969
Effl.
800-
992
1056
922
1088
1024
979
1024
1088
1024
966
966
960
973
960
1152
1152
1216
1056
864
960
928
1018
Renovated
1
1024-
1182
1152
1152
1152
1056
1203
1216
1152
992
1024
1088
992
960
1-2
909-
1216
1216
1088
1120
1088
1152
960
1011
1216
1152
1216
1280
1152
1184
1056
1030
1043
992
1088
1056
1056
845
960
1024
992
1082
ECW
832-
1158
1088
1024
1088
1024
1120
992
1024
1088
1216
1184
1248
1210
1184
1024
1024
1018
896
1152
1088
992
928
960
896
1043
1063
water
5-6
1024-
1088
1011
960
1024
1056
992
1037
1024
960
1120
1152
1248
1184
1152
1088
1030
1082
960
1088
1056
1024
896
960
992
947
1043
7 EW
1024-
1152
1056
1120
1152
1216
1792
1696
1613
1408
1120
1050
992
1024
1024
1152
960 1024
960
Native ground water
WCW 8 91st Ave.
2291- 2048-
3712 2176
3800
1984
2048
3200 1856 2560
1664 2432
3600
1677 2624
4000 1798
1920
3850 2560
3500 2048
2048
3150
1948 2720
2400
1888
2100
2528
1950 1856
-------
of 8 to 18 whereby damage to the structure of the soil can occur if
the water is used for irrigation of soil containing clay. High
values of SAR will cause the clay to deflocculate, which reduces the
permeability of the soil and adversely affects the structure of the
soil.
7- £H
The pH of the effluent was usually between 7.7 and 8.1, and of the
renovated water between 7.0 and 7.6 (Table 12). The lower pH of the
renovated water was probably due to bacterial activity in the soil and
resulting CO -production. The pH of the native ground water ranged
from 7 to 8 with an average of around 7.5.
8. Coliform and Soil Bacteria
Presumptive, confirmed, and fecal coliform densities were determined
with the multiple-tube fermentation technique. Fecal coliform densi-
ties in the secondary effluent, which is not chlorinated, were usually
of the order of 100,000 to 1,000,000 per 100 ml.
The fecal coliform density of the renovated water from ECW was generally
less than 10 per_100 ml (Figures 24 and 25). If sequences of long
inundation periods were held, the fecal coliform density in the reno-
vated water tended to rise when newly infiltrated water arrived at
the well, and then to decrease to essentially zero as inundation
continued. The coliform removal was probably due to "filtering" at the
surface and mortality in the hostile and competitive soil environment.
Filtering and competition, and hence the fecal coliform removal, can
be expected to increase with continued inundation, because of increased
surface clogging and bacterial population in the soil. Later studies
showed that essentially all the removal of fecal coliforms took place
in the first 3 ft of the soil.
Fecal coliform densities in the renovated water from well 7 were lower
than those from ECW, whereas fecal coliforms could not be detected in
well 8 when it started to yield renovated sewage water. Thus, while
a few feet of underground travel is sufficient 'to remove almost all
fecal coliform bacteria, a distance of 100 to 200 ft is necessary to
obtain renovated water with consistent absence of fecal coliforms.
Presumptive coliform densities were higher than the fecal coliform
densities (Figures 24 and 25), indicating the presence of soil coliform
bacteria such as Aerobacter aerogenes. The median density of fecal
streptococci in the renovated water from ECW was 10/100 ml (Figure 25).
Plate counts of aerobic bacteria in theRsoil below the recharge basins
indicated total populations of about 10 per gram of dry soil near
the surface, and 10 to 10 per gram of dry soil at depths of about
3 ft (Table 13).
70
-------
Table 12. PH of effluent and water from wells (1969).
Date
1968
13-14 Jan 69
27 Jan 69
20 Feb 69
10-11 Mar 59
20 Apr 69
6 May 69
9-10 Jun 69
26-27 Jun 69
8-9 Jul 69
22-23 Jul 69
4-5 Aug 69
19-20 Aug 69
9-10 Sep 69
25-26 Sep 69
8-10 Oct 69
23-24 Oct 69
14 Nov 69
24 Nov 69
5-8 Dec 69
29-30 Dec 69
Av. 1969
Effl.
7.9-
8.1
7.24
8.1
7.8
7.7
7.7
8.1
8.0
7.6
8.0
8.1
8.1
8.1
8.2
7.8
8.1
8.0
7.8
7.9
1 1-2
7.9 7.1
7.9 7.6
7.8 7.2
7.4
7.6 7.1
7.3
7.1
7.4 7.0
7.3
7.6 7.1
7.7 7.2
7.3
7.9 7.6
7.1
7.8 7.2
7.2
7.7 7.4
7.6
7.3
Renovated
ECW
7.3-
8.4
8.1
7.2
7.3
7.1
6.9
7.0
7.2
7.1
6.9
7.1
7.2
7.4
7.0
7.3
7.3
7.2
7.2
7.2
water
5-6 7 EW
7.7 7.6
8.0 7.8
7.7 7.4
7.8
7.8 7.5
7.9
7.6
7.7 7.0
7.7 7.4
7.6 7.2
7.4
7.4 7.2
7.5
7.7 7.4
7.3
7.5 7.5
7.4 7.4
7.4 7.5 7.5
8.0 7.2
7.6
Native ground water
WCW 8 91st Ave.
7.4- 7.9
8.3
7.6 8.0
7.9 7.7
7.4 7.8
7.4
7.4 7.7
7.8
7.3
7.4 7.1
7.2 7.5
7.1
7.3 7.5
7.3 7.5
7.6 7.4
7.3
7.4 7.4
7.4
7.5 7.3
7.5 7.6
71
-------
Table 13. Aerobic bacteria counts in recharge basin soil.
Date
2 July 1969
21 July 1969
21 July 1969
4
5
6
6
6
1
1
Basin
(grass)
(rice)
(rice)
(rice)
(rice)
(bare area)
(bare area)
Depth
inches
0-1
0-1
0-1
12
36
0-1
36
Population
per gram
of dry soil
9.2 x 107
3.7 x 108
4.2 x 108
4.3 x 108
4.0 x 106
1.9 x 108
2.2 x 107
72
-------
150
s
o
a:
LJ
QL
50
0
I ' I ' I ' I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
INUNDATION PERIODS
EAST CENTER WELL
-*— PRESUMPTIVE
© CONFIRMED
-•-- FECAL
l < i ' i ' | ' i ' i ' i ' i > i ' i
JAN. FEB. MAR. APRIL MAY JUNE JULY AUG. SEPT. OCT.
Figure ,'.'*. Co Li form bacteria in renovated water from ECW in 1968.
MOV.
DEC.
-------
1000
700
1611
NUNDATION PERIODS
BASINS 5 a 6-—/
BASINS 1,2,3 8 4
PRESUMPTIVE
O CONFIRMED
--o-- FECAL COLIFORMS
FECAL
STREPTOCOCC
n-n-'l
JAN.
FEB.
Figure 25. Coliforms and fecal streptococci in renovated water from ECW in 1969.
DEC.
-------
9 • Mfect of Grass Filtration on Effluent Quality
Passing waste water through grass or other vegetation as overland
flow is sometimes successfully used as a treatment process. To inves-
tigate the effect of the flow through the grass basins on the quality
of the effluent at the Flushing Meadows Project, the basins were split
in two groups of three with the effluent flowing serpentine fashion
through the three basins of each group (Figure 5). This "series"
flow resulted in an overland flow distance of about 2100 ft through
the grass in basins 4, 5, and 6.
Continuous samples of the effluent in the basins were obtained at flow
distances of 0, 700, 1400, and 2100 ft, as indicated in Figure 5.
Basin 1 had no vegetation and basin 2 was covered with a gravel layer.
Thus, the flow for basins 1, 2, and 3 had passed through 700 ft of
grass before entering the gravel basin No. 2, after which it entered
the bare soil basin No. 1. The sampling devices consisted of a syphon
tube which conducted water from the basin to a constant-head (overflow)
cylinder. A long, small diameter tube, mounted on the cylinder wall
near the top of the cylinder, conveyed water from the constant head
device to a bottle at a slow enough rate to collect a sample of about
2 gallons in 24 hrs. The sample obtained of the effluent as it left
basin 6 and entered basin 5 is referred to as sample "6-5." A similar
notation was used for the other basins. The samples obtained at the
outflow from basins 4 and 1 are referred to as "4-out" and "1-out."
The grass filtration studies were carried out in the last half of 1968.
Since the grass was not mowed, a dense mat of about 6 inches thick and
consisting of flat grass stems several feet in length had formed in
basins 3, 4, 5, and 6 during the summer. Most of the overland flow
of the effluent took place above this mat, through the green growth
of the grass. However, the effluent that infiltrated into the soil
had to move through the mat. To evaluate the quality of the effluent
after it had moved through the grass mat, a perforated copper tube of
about 3 ft length was placed in horizontal position on the soil surface
below the mat. The location of this tube was near the flume connecting
basin 6 to basin 5 and the sample is referred to as "6-5 tube."
Continuous water samples from this tube were obtained using a device
similar to that used in the basins.
The results of the analyses (Tables 14 and 15) show that, contrary to
what might be expected, there was little effect of the overland flow
through the grass basins on the quality of the effluent, even when the
flow distance was about 2000 ft and the corresponding travel time 16
hrs. There seemed to be a slight tendency for the COD to increase
as the effluent flowed through the grass, and for the ammonium and
nitrate concentrations to decrease. The flow through the nonvegetated
basins 1 and 2 also seemed to have little or no effect on the COD and
ammonium content (Table 14).
75
-------
Table 14. Effect of flow through basins on COD,, NH , and NO content
of effluent (1968).
Sample
Effluent (pump)
6-5
5-4
4-out
3-2
2-1
1-out
COD, ppm 0
M
3
CM
1
CO
CM
50
46
50
57
46
46
?
ft
o>
CO
LO
48
49
53
58
49
56
55
ft
0)
CO
LO
CM
1
CM
28
47
40
39
?
42
40
ft
QJ
CO
CM
vD
CM
47
56
43
43
?
43
42
™,
ft
0)
CO
LO
1
-d-
26.5
24.9
26.1
26.0
26.8
25.5
24.9
ft
-------
Table 15. Effect of flow through grass basins and grass mat on COD,, NH
4
and NCL of effluent.
COD,, ppm 0
Effluent (pump)
6-5 tube
5-4
4 -out
NH, , ppm N
Effluent (pump)
6-5 tube
5-4
4-out
Kjeldahl, pprn N
Effluent (pump)
6-5 tube
5-4
4-out
NO ppm N
Effluent (pump)
6-5 tube
5-4
4-out
4-1
u
o
u~l
r— 1
1
O
oj
r—l
1
r—l
, [
43
7
40
52
24.9
7
35.0
29.5
0.00
7
0.02
0.05
o
o
P
ro
r— I
1
C-l
i— 1
7
50
56
51
31.7
26.4
30.5
31.6
0.02
1.9
0.02
0.02
o
0)
O
o
i— i
i
ro
r—l
61
7
7
54
31.2
7
7
34.3
0.05
7
7
0.00
o
QJ
P
r^
r—l
1
•^o
,— 1
49
48
?
53
27.7
27.9
7
28.5
0.02
0.00
7
0.00
—]
-------
There was a tendency for the COD to increase as the effluent moved
through the grass mat before it infiltrated into the soil (Table 15).
This COD increase is probably due to release of organic matter from
decaying grass stems and leaves. Also, there seemed to be a reduction
in the ammonium content, probably because of nitrogen fixation by
bacteria in the decaying mat. The nitrate content was not appreciably
affected. The increase in COD as the effluent moved through the grass
mat may have contributed to the increased denitrification below the
vegetated basins, as discussed under the section "Nitrogen."
10. Water Temperatures
Temperatures of the secondary effluent and renovated water from ECW
and WCW were measured in 1969. The temperature of the effluent at the
inflow end of the basins ranged from 20 C in the winter to 32 C in the
summer. The effluent temperature at the outflow end ranged between
13 and 30 C. The renovated water from ECW had temperatures of about
18 C in winter and 30 C in summer. For WCW-water, the temperature
range was 21 to 28 C.
-------
SECTION VIII
DESIGN AND OPERATION OF LARGE-SCALE SYSTEM
1- General Design Aspects
The infiltration studies of the Flushing Meadows Project have indicated
that 1 acre of recharge basin can infiltrate at least 300 acre-feet of
secondary effluent per year. Thus, reclaiming the present effluent
flow of about 90,000 acre-feet per year would require about 300 acres
of recharge basins. Renovating the sewage flow of 300,000 acre-feet
per year projected for the year 2000 would require about 1000 acres of
recharge basins. The recharge basins would be fairly large (about
300 x 500 ft each) . They would probably best be located on both sides
of the Salt River bed, one chain of recharge basins on the north side
of the river bed, and another chain on the south side (Figure 26).
The effluent would be distributed into the basins by a channel which
would run along one side of the recharge strip.
After infiltration and lateral movement below the water table, the
sewage water would be pumped as renovated water from a series of wells,
which would be located in the center of the river bed (Figure 26).
Since the river bed is about 2000 ft wide, the recharge strip could
be about 500 ft wide, leaving a distance of about 500 ft between the
recharge strips and the wells to insure adequate time and distance of
underground travel.
Since a low suspended solids content in the effluent is required to
obtain high infiltration rates with minimum maintenance of the basins,
a sedimentation reservoir should be placed at the head of the channel
system carrying the effluent to the recharge basins. Such a reservoir
could also serve as a storage facility to even out diurnal fluctuations
in the effluent discharge.
The design of a system for renovating sewage effluent by ground water
recharge with infiltration basins should be based on the following
criteria:
a. The water table below the recharge basins should be at a depth of
at least 5 to 10 ft during infiltration. This is necessary to maintain
an unsaturated zone of sufficient thickness for aerobic decomposition
of organic material, and to permit rapid drainage, and hence aeration,
of the soil profile between inundations.
b. The renovated water should have a certain minimum time and distance
of underground travel before it is pumped from the well. This is
necessary for "polishing" treatment, taste and odor removal, and for
precaution against breakthrough of pathogenic organisms. While reliable
data regarding desired time and distances of underground travel are
lacking, an underground detention time of a few weeks and travel dis-
tances of a few hundred feet may be sufficient, depending on the aquifer
79
-------
RECHARGE STRIPS
PLAN
YTWTTTT ww
////////////////,
///.
Figure 26.
CROSS SECTION G~G
Plan and cross-section of two parallel strips with recharge basins with
wells in center for collecting renovated water.
-------
material, the quality of the effluent, and the desired quality of the
renovated water.
c. The spread of the renovated water into the aquifer outside the
system of recharge basins and wells should be avoided because the
chemical composition of the effluent and the various processes in the
soil are not completely known. Some compounds (hormones, enzymes,
biocides, etc.) may resist biodegradation in the soil and if future
research should show that some of these refractory compounds are toxic
or carcinogenic, a difficult situation would exist if the renovated
water had been allowed to spread uncontrolled into the aquifer. Thus,
the system of Figure 26 should be operated so that all the effluent
water that has infiltrated into the soil will be pumped from the, wells
in the center of the river bed. Also, the water table below the edges
of the river bed (such as at points C and D in Figure 26) should,not
rise above the ground water level adjacent to the river bed. As a
matter of fact, the ground water levels at points C and D should remain
slightly lower than the water table outside the recharge system, so
that there is a slight gradient from the aquifer to the recharge system
to make sure that renovated sewage water cannot spread into the ground
water outside the system.
2. Underground Flow System
In order to design a system in accordance with the above criteria, it
must be possible to calculate water table positions and underground
detention times that can be expected in the recharge system. For this
purpose, analyses of the underground flow system for various system
geometries and aquifer conditions were carried out with an electrical
resistance network analog. Horizontal models of the flow system were
simulated on the analog. This required knowledge of the effective
transmissibility of the aquifer for ground water recharge.
The effective transmissibility, T , for recharge is less than the
transmissibility of the entire aquifer because the recharge flow system
is concentrated in the upper portion of the aquifer, often called the
"active" region. The deeper portion of the aquifer, which does not
contribute much to the recharge flow system, is called the passive
region.
The effective transmissibility of the aquifer beneath the Salt River
bed was evaluated by applying the Dupuit-Forchheimer assumption of
horizontal flow to the recharge flow system of Figure 16. The shape
of the ground water mound in this system is shown in Figure 27 with the
vertical scale exaggerated.
According to the Dupuit-Forchheimer assumption, the horizontal flow
beneath the ground water mound at a certain distance from the center
can be described as
ix = - T £L
e dx
81
-------
CO
ro
RECHARGE
BASIN
PERCOLATION
ZONE
ORIGINAL
WATERTABLE
100
200
FEET
Figure 27. Ground water mound of flow system in Figure 16 with vertical scale exaggerated,
-------
where I - infiltration rate for recharge basin area (ft/day)
Te = effective transmissibility of aquifer (ft2/day)
h = height of ground water mound above static water table
x = horizontal distance from symmetry line (Figure 27) .
The factor I represents the average infiltration rate for the entire
basin area, including the dry portions, such as dikes, roads, and
basins undergoing a dry-up period. Of course, the inundated basins
should be sufficiently frequent to permit treatment of the basin area
as one infiltration unit with an average infiltration rate, I.
Integrating equation (1) for the region beneath the recharge area
yields
where h h at center of mound
c
h = h at edge of mound (x = W/2)
W = width of infiltration area (Figure 27) .
For the ground water mound in Figure 27, h = 3.80 ft, h = 2.35 ft,
I = 1.91 ft/day, and W = 220 ft. Substituting these values in equation
(2) yields a value of 7,970 ft /day, rounded to 8000 ft /day, for the
effective transmissibility of the aquifer. Since the hydraulic conduc-
tivity in horizontal direction is 282 ft/day, the effective height of
the flow system in Figui •- 16 is about 28 ft. According to Figure 16,
this depth corresponds approximately to the 60% streamline. The total
height of the unconfined aquifer was much greater, i. e., about 240 ft
(Table 1) .
The effective transmissibility for ground water recharge may be a
constant, or it may depend on the width of the recharge basin area.
If the transmissibility of the aquifer is mainly due to a single layer
of very high hydraulic conductivity, the effective transmissibility
will be essentially constant. If, however, the aquifer is uniform,
or uniformly anisotropic, to appreciable depth, the effective trans-
missibility will vary in direct proportion to the width of the recharge
basin area, until the effective height of the flow system is about
equal to the height of the aquifer (vertical distance between water
table and impermeable boundary) .
For the unconfined aquifer beneath the Salt River bed, with its sand
and gravel layers at great depth, the effective transmissibility should
be taken in direct proportion to the width of the recharge basin area.
83
-------
If the hydraulic properties of the aquifer are similar to those at
the Flushing Meadows site, the effective transmissibility for systems
of different widths can be calculated as
Te = io 800° ft
where W is the width of the basin area or recharge "strip" for which
the effective transmissibility is to be calculated. The factors 8000
and 220 in this eq'uation refer to T in ft /day and W in ft, respec-
tively, of the Flushing Meadows Project.
Analyses of the flow system for the large-scale system (Figure 26) were
performed for different geometries of recharge strips and wells and
different effective transmissibilities . The analyses were based on
steady-state conditions. Thus, all the water that infiltrated in the
basins was assumed to be pumped from the wells and the well discharge
could be calculated as
Q = 2 SWI (4)
in which 0 = discharge per well
S = distance between wells
W = width of recharge: strip
I = average infiltration rate for recharge strip (Figure 26)
Because of the symmetry of the system in Figure 26, only one quadrant
of the flow system around each well, such as section CDEF, needed to
be analyzed. The analysis was based on the assumption that the water
table below the outside edges of the recharge strip (CD in Figure 26)
would be maintained at the same elevation as the water table in the
aquifer outside the recharge system, so that there would be no flow
from the recharge system to the outside aquifer, or vice versa.
Assuming that the ground water flow beneath the recharge strip is
rectilinear and parallel to CB (Figure 28) , the elevation difference,
AH , between the water table at C and B could be calculated with
L-B
equation (2) as
AHC-B • yf-
e
Because there is no symmetry , line in the center of the recharge strip
for the system of Figure 26, the factor 8 in the numerator of equation
(2) was replaced by a factor 2 in equation (5). The water table
difference, AH between D and A is equal to AH , because of the
jU~"/V U — Jj
assumption of parallel flow in section ABCD .
84 ,
-------
B
CD
VJ1
*S (
\ 1
**sk i
' w w w w * w w '
•
-------
To determine the elevation difference between the water table adjacent
to the well (point F in Figure 26) and the water tables at A and B,
the section ABEF was simulated on a resistance network analog. The
line AB was taken as a line source with uniform flux and the well as a
point sink. The analog analyses were performed for different ratios of
L/S, in which L = distance between recharge area and well, and S =
distance between wells (Figure 26). The node arrangement for the case
where L/S = 0.5 is indicated in Figure 28. The results of the analog
analyses are shown in dimensionless form in Figure 29, where the water
table elevation drop, AH , from A to F is plotted versus L/S in terms
J\~~ J.
of the dimensionless parameter 2 WSI/T AH . Another dimensionless
graph (Figure 30) shows the ratio AH /AH versus L/S. As long as
L/S > 0.5, this ratio is essentially equal to 1, so that the water
table elevations at A and B are nearly the same. Thus, the assumption
of one-directional flow in section ABCD is valid if L/S > 0.5.
Equation (5) and Figure 29 can be used to calculate the elevation
difference between the water table at the wells (point F) and the water
table beneath the outer edges of the recharge strips (points C and D).
This elevation difference can be used to determine the dynamic lift
for pumping the renovated water from the wells, adapting the water
table below the outer edges of the infiltration strips (C and D,
Figure 26) to the original water table.
Since the well was simulated as a point sink in the analog analyses,
the actual diameter of the well was not taken into account and the
results are approximate only. To adapt the results to< a well of given
diameter, the theory of steady, radial flow to a well (see, fo>r example,
Todd (12) , Chapter 4) can be applied to the region around the we'll
where the equipotentials are essentially circles (Figure 32).
The underground detention time consists of time for downward travel
to the ground water and time for lateral travel to the well. The
downward travel time can be calculated from the infiltration rate, the
soil porosity and water content in the percolation zone, and the
distance of vertical travel. The horizontal travel time depends on
the distance from the point of infiltration to the well. The water
that has infiltrated at A will be of most interest, because it will
have the minimum horizontal travel time occurring in the system
(Figure 26).
To evaluate the minimum underground travel time, equipotentials and
streamlines in the region ABEF must be known. An example of the equi-
potentials and streamlines in this region obtained by electrical
resistance network analog is presented in Figure 31. To predict the
travel time from A to F in this system, the flow rate in the stream
tube between AF and the next streamline is calculated by multiplying
the width of the tube along AB by IW. This flow rate is then divided
by the cross-sectional area of the tube (width of tube x height of
86
-------
2WSI
TeAHA-P
i i i
I 2
L/S
Figure 29. Dimensionless graph of 2 WSI/T
-------
H.
1.6
\
B-F
,-F
1.5
1.4
1.3
1.2
0.5
I I I
L/S
1.5
Figure 30. Dimensionless graph of AH,, «/AH. versus L/S.
B-F A-F
-------
,<
Q
O
O
O
o"
R E
-t H
A R
G E
-26-
S T
O
O
O
o"
10
200
Figure 31. Streamlines and equipotentials (in feet above
water table adjacent to well) for hypothetical
system in Salt River bed.
-------
B A
B A B
— 100 -—
'00.
&A
B A
•100-
100-
H,
B A
80-
100
B
80'
60
V°\
(•—i—i-
E F
f • °-625
-- = 0.833
80
•60
•40-
Figure 32. Equipotentials (AB = 100, F = 0) in region ABEF for different values of L/S.
-------
aquifer) and by the porosity of the aquifer material to yield the
macroscopic velocity. This velocity can be computed for each section
of the stream tube between successive equipotentials. Dividing the
distance between these equipotentials by the macroscopic velocity for
the section in question then yields the travel time increment for that
section. The total travel time from A to F is obtained by summing the
time increments for each section in the stream tube.
To enable calculation of travel times for different system geometries,
equipotentials for the the region ABEF were determined for different
ratios of L/S (Figure 32). These equipotentials are expressed in
percent head loss between A and F and they were evaluated by resistance
network analog using node arrangements similar to the one shown in
Figure 28. Streamlines can be sketched as orthogonals to the
equipotentials, as was done in Figure 31.
The effective transmissibility- T , for the flow region near the well
may be greater than T for the flow region below the recharge area.
This is true if the aquifer is unconfined and the wells penetrate the
aquifer to a greater depth than the height of the active region of
the aquifer for the recharge flow system. In that case, it will be
desirable to treat the flow in the region ABDF (well-flow region) with
a higher T -value than the one for the recharge flow in region ABCD.
This is permissible, since most of the head loss between the recharge
area and the well occurs in the vicinity of the well (Figures 31 and
32) . The value of T for the well-flow system would then be evaluated
on the basis of the well depth or of pumping tests, whereas the value
of T for the recharge-flow system would be based on equation (2),
corrected for the width of the recharge area by equation (3) if
necessary.
Example. To illustrate the use of equation (5) and Figures 29, 30,
and 32, the height of the water table at A, B, C, and D (Figure 26)
above that of the water table adjacent to the well, and the travel time
from A to F, will be calculated for a hypothetical system in the Salt
River bed.
Taking W = 600 ft, T for the flow system beneath the recharge basins
(ABCD, Figure 26) can be calculated with equation (3) as about 21,850
ft2/day, which will be rounded to 20,000 ft2/day. Because the wells
will probably penetrate the aquifer more than the height of the active
zone for the recharge flow system, T& for region ABEF will be taken as
30,000 ft2/day.
The annual recharge rate will be taken as 300 ft, giving an I-value
of 0.82 ft/day which will be rounded to 0.8 ft/day. The pumping rate
will be taken as 768,000 ft3/day for each well, which corresponds to
a well discharge of about 4000 gpm.
Substituting the values for Qw, W, and I in equation (4) yields
S = 800 ft. This gives L/S = 0.625, so that according to Figure 29,
91
-------
2 WSI/T AH = 1-38. Substituting the values for W, S, I, and T£
for the well-flow system into this parameter yields AH = 18.6 ft.
Figure 30 shows that for L/S = 0.625, AH /AH = 1.02, so that
r>—r A—r
AH = 18.9 ft. The next step is to calculate AH with equation (5),
^Q — _p U .D
which yields AH = 7.2 ft. Thus, AHD_F = 18.6 + 7.2 = 25.8 ft, and
AH = 18.9 + 7.2 = 26.1 ft. The water table adjacent to the well is
thus about 26 ft lower than the water table beneath the outside edges
of the recharge strips.
Equipotentials and streamlines for this example are shown in Figure 31.
The equipotentials for the recharge flow system in region ABCD were
calculated with equation (5), and those for the well-flow system in
section ABEF were obtained from Figure 32. The streamlines were
sketched as orthogonals to the equipotentials. To calculate the
minimum travel time, Kh will be taken as 282 ft per day (Figure 16)
so that the effective height of the aquifer for the well-flow system
is 30,000/282 = 106 ft. Assuming a porosity of 15%, the macroscopic
velocity between the 18-ft and 16-ft equipotentials in the stream tube
between AF and the next streamline can be estimated as (0.8 x 600 x
100)/(123 x 100 x 0.15) = 26 ft per day. In this calculation, the
numerator is the product of recharge rate and recharge area feeding
the stream tube, the factor 123 is the average height of the aquifer
between the 18-ft and 16-ft equipotentials (106 + (16 + 18)/2) , the
factor 100 in the denominator is the average width of the stream tube
between the 18-ft and 16-ft equipotentials, and the factor 0.15
represents the porosity. Dividing the average distance between the two
equipotentials, which is 115 ft, by the macroscopic velocity, yields a
travel time of 4.4 days. Applying this procedure to the rest of the
potential intervals of the stream tube and summing the time increments
yields a total travel time of 2 weeks from A to the well. If a longer
minimum travel time is desired, S can be decreased, L increased,
I decreased, and W decreased.
Prediction of systems of underground water movement is fraught with
uncertainty where nonuniform hydrogeologic conditions exist. Thus,
design calculations may have to be followed up by field measurements to
determine how the actual performance of the system compares with the
predictions, and what modifications in the operation of the system may
be desirable. Such field studies should include measuring water levels
in observation wells and determining the quality of the water as it
moves to and below the water table. Travel times can be evaluated by
following the movement of a characteristic ion or compound present in
the waste water, or by adding a tracer.
3. Economic Aspects
The total cost of renovating sewage water with the operational system
of Figure 26 is estimated at about $4.50 per acre-foot. This figure
is calculated as follows:
92
-------
Cost of land * $ 2000 per acre
Construction (sedimentation reservoir, channels,
distribution works, dikes, etc.) ' 2000 per acre
Wells ($30,000 per well, one well per 15 acres) 2000 per acre
Total capital cost $ 6000 per acre
* Some land in the Salt River bed is in private ownership. The rest
is state or federal land.
Annual fixed cost (10% of capital costs) $ 600 per acre
Annual maintenance 150 per acre
Total annual cost $ 750 per acre
With an annual infiltration rate of 300 ft, the above cost amounts to
$2.50 per acre-foot. Assuming a pumping lift of 100 ft and pumping
cost of 2 cents per acre-foot per foot of lift, the pumping costs would
be $2.00 per acre-foot. Thus, the total cost of infiltrating the
effluent and pumping the renovated water would be $4.50 per acre-foot.
A more detailed cost analysis of renovating sewage water with the
system of Figure 26 was made by Buxton (8), who arrived at a total-cost
range of $4.83 - $5.87 per acre-foot.
Because the hydrogeologic conditions of the Salt River bed and the
climate of the lower Sonoran Desert are favorable for renovating sewage
water by soil filtration and ground water recharge, renovated sewage
water appears to be the cheapest and most readily available "new" water
resource for Central Arizona. The cost of renovating the sewage water
by in-plant tertiary treatment (phosphate precipitation, ammonium
stripping or nitrification-denitrification, activated carbon adsorption,
and disinfection) would be at least $50 per acre-foot, or about 10 times
the cost of renovating the sewage water by ground water recharge.
4. Future Projects
Because of the favorable results from the Flushing Meadows Project,
larger projects can be embarked upon with confidence. One such project
would be at the Phoenix 23rd Avenue Sewage Treatment Plant, where a
40-acre oxidation pond could be split into four 10-acre infiltration
basins. These basins would be intermittently flooded to infiltrate
about 15 mgd of the secondary effluent (activated sludge). The renovated
water could be pumped from wells in the center of the basin area, and
delivered to an irrigation-district canal for unrestricted irrigation.
Another project involving use of renovated sewage water is the Rio Salado
Project, proposed in 1966 by Arizona State University's Architecture
Department. This project would consist of converting the Salt River
93
-------
bed into aquifer parks, housing and industrial developments, etc.
Part of the river bed would be used for infiltration basins. The
renovated water would be pumped upstream to be cycled through chains
of recreational lakes. The remainder of the renovated water could be
pumped into the canal system for unrestricted irrigation. Some of
this water could also be used for ground water recharge where the
ground water table is low because of too much pumping. A model of
the proposed system, showing the waste water renovation system and
the recreational lakes, is shown in Figure 33.
-------
Figure 33. Model of Rio Salado Project, showing sewage water
renovation system (front) and recreational lakes (back)
95
-------
SECTION IX
ACKNOWLEDGMENTS
Meetings between various agencies in the Salt River Valley to discuss
means for upgrading and reusing the valley's municipal waste water were
held in the early 1960's. Participating in these meetings were, among
others, H. Shipley (Salt River Project), D. Travaini (City of Phoenix),
and L. E. Myers (U. S. Water Conservation Laboratory). Plans for a
pilot project to study renovation of the effluent by ground water
recharge were developed in 1964 by Herman Bouwer (U. S. Water Conserva-
tion Laboratory) , and an Effluent Recharge Committee was formed to
carry the project to reality. The committee consisted of:
Herman Bouwer, U. S. Water Conservation Laboratory (Chairman)
C. A. Pugh, Bureau of Reclamation
H. Shipley, Salt River Project
D. Travaini, City of Phoenix
J. J. Weinstein, Maricopa County Health Department
L. G. Wilson, University of Arizona
Numerous meetings were held and an application for a demonstration grant
from the Environmental Protection Agency (then Federal Water Pollution
Control Administration) was forwarded by the Salt River Project in
March 1966. The grant was approved in December 1966 and construction
began in early 1967 by the Salt River Project. H. Shipley, Associate
General Manager of the Salt River Project, was the project officer for
the grant. Tne research program, which was carried out by the U. S.
Water Conservation Laboratory, started in September 1967.
Many persons have contributed to the project. The "effluent team"
of the U. S. Water Conservation Laboratory consisted of Herman Bouwer,
R. C. Rice, and E. D. Escarcega. It was later joined by J. C. Lance
and F. D. Whisler. The water analyses were done, in succession, by
John Krebs, Paul Kuechelmann, and M. S. Riggs of the Salt River Project,
stationed at the U. S. Water Conservation Laboratory. F. S. Nakayama
of the U. S. Water Conservation Laboratory developed the initial
procedures for the water analyses. J. A. Replogle of the U. S. Water
Conservation Laboratory designed the flumes for measuring flow into
and out of the infiltration basins. Of the Salt River Project,
H. Shipley, R. W. Teeples, G. Garrison, 0. Hatcher, R. L. Juetten,
D. E.'Womack, R. Pristo, and W. L. Simser rendered valuable services
to the development and construction of the project. Of the City of
Phoenix, the cooperation of Dario Travaini, E. H. Braatelien and
A. E. Watson is acknowledged.
97
-------
X. REFERENCES
1. Bouwer, Herman, "A Double-Tube Method for Measuring Hydraulic
Conductivity of Soil In Situ Above a Water Table," Soil Sci.
Soc. Amer. Proc., 25, pp 334-339 (1961).
2. Bouwer, Herman, "Field Determination of Hydraulic Conductivity
Above a Water Table with the Double-Tube Method," Soil Sci.
Soc. Amer. Proc.. 26, pp 330-335 (1962).
3. Bouwer, Herman, "Analyzing Ground-Water Mounds by Resistance
Network,"" Journal Irrigation and Drainage Division, ASCE, 88,
No. IR3, Proc. Paper 3276, pp 15-36 (1962).
4. Bouwer, Herman, "Rapid Field Measurement of Air Entry Value
and Hydraulic Conductivity of Soil as Significant Parameters
in Flow System Analysis," Water Resources Res. _2_,_ No. 4,
pp 729-738 (1966).
5. Bouwer, Herman, and Rice, R. C., "Modified Tube Diameters for
the Double-Tube Apparatus." Soil Sci. Soc. Amer. Proc., 31,
pp 437-439 (1967).
6. Bouwer, Herman, "Analyzing Subsurface Flow Systems with Electric
Analogs," Water Resources Res. 3, pp 897-907 (1967).
7- Bouwer, Herman, "Planning and Interpreting Permeability Measure-
ments," Journal Irrigation and Drainage Division, ASCE, 95,
No. IR3, Proc. Paper 6775, pp 391-402 (1969).
8. Buxton, John Lacy, "Determination of a Cost for Reclaiming
Sewage Effluent by Ground Water Recharge in Phoenix, Arizona,"
MS Thesis, Arizona State University, College of Engineering
Science (1969) .
9. Frevert, R. K., and Kirkham, Don. "A Field Method for Measuring
the Permeability of Soil Below a Water Table," Proc., Highway
Research Board, 28, pp 433-442 (1948).
10. Replogle, J. A., "Critical-Depth Flumes for Determining Flow
in Canals and Natural Channels," Trans. Am. Soc. Agr. Eng., 14,
pp 428-433, 436 (1971) .
11. Standard Methods for the Examination of Water and Waste Water,
13th Ed^, Am. Public Health Assoc., New York (1971).
12. Todd, D. K., "Ground Water Hydrology," John Wiley & Sons, Inc.,
New York and London (1960).
-------
13. Woldendorp, J. W., Dilz, K. , and Kolenbrander, G. J., "The Fate
of Fertilizer Nitrogen on Permanent Grassland Soils," Proc.
First General Meeting European Grassland Federation,'Wageningen,
The Netherlands, pp 53-76 (1966).
100
-------
XI. PUBLICATIONS
1. Bouwer, Herman, "Water Quality Improvement by Ground-water Recharge,"
Second Seepage Symposium Proc.. Phoenix, Arizona, 25-27 March 1968,
USDA-ARS 41-147, pp 23-27 (1969).
2. Bouwer, Herman, "Returning Waste to the Land: A New Role for
Agriculture?," Jour. Soil and Water Conserv., 23, No. 5, pp 164-168
(1968). — ~
3. Bouwer, Herman, "Putting Waste Water to Beneficial Use — The
Flushing Meadows Project," Arizona Watershed Symposium Proc.,
Phoenix, Arizona, 17 Sept. 1968, pp 25-30 (1968).
4. Bouwer, Herman, "Ground Water Recharge Design for Renovating Waste
Water," Jour. Sanitary Engin. Div., Amer. Soc. Civil Engin, Proc.,
96, SA 1, pp 59-74 (1970).
5. Bouwer, Herman, "Water Quality Aspects of Intermittent Infiltration
Systems Using Secondary Sewage Effluent," Artificial Groundwater
Recharge Conf. Proc., Water Research Assn., Reading Univ., Berkshire,
England, Paper 8, pp 199-217 (1970).
6. Bouwer, Herman, "Waste Water Purification," IN McGraw-Hill Yearbook
of Science and Technology, pp 434-436 (1971).
7. Bouwer, Herman, and Lance, J. C., "Reclaiming Municipal Wastewater
by Groundwater Recharge," Amer. Assoc. Adv. Sci. Symposium on
Urbanization of Arid Lands Proc. (Dec. 1970), In press.
8. Bouwer, Herman, and Mann, Russell F., "Agricultural and Urban Waste
Water Reuse," presented at Amer. Soc. Civil Engin. Water Resources
Engin. Conf., Phoenix, Arizona, January 1971. (Multilithed for
handout at conference, only.)
9. Bouwer, Herman, Lance, J. C., and Rice, R. C., "Renovating Sewage
Effluent by Ground-Water Recharge," Proc. 16th New Mexico Water
Conference, Las Cruces, pp 32-47 (1971).
10. Bouwer, Herman, Lance, J. C., and Rice, R. C. , "Land Disposal of
Sewage Effluents," Proc. Symposium on Nitrogen in Soil and Water,
Univ. of Guelph, Canada, pp 110-136 (1971).
11. Bouwer, Herman, Lance, J. C. , and Rice, R. C. , "Renovating Sewage
Effluent by Ground-water Recharge," Proc. Annual Meeting Arizona
Section Am. Water Res. Assoc. and Hydrology Section, Arizona
Acad. of Science, Tempe, Arizona, (April 1971). Hydrology and
Water Resources in Arizona and the Southwest, I, pp 225-245.
101
-------
12, Bouwer, Herman, Lance, J. C., and Rice, R. C., "Renovating Sewage
Effluent by Ground-water Recharge," Proc. Land Spreading Conf.,
Orlando, Fla., East Central Florida Regional Planning Council,
pp 6.1-6.20 (July 1971).
102 -'U.S. GOVERNMENT PRINTING OFFICE: 1972 UBIi-l(fl7/31fi 1-3
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
', Rep ft No.
w
4 Tit!e 5, K OrtD-'.'?
Renovating Secondary Sewage by Ground Water Recharge with 6
Infiltration Basins
S. P'--;'ortnir- f
AUi^r< Herman Bouwer, R. C. Rice, E. D. Escarcega (U. S.
Water Conservation Laboratory; M. S. Riggs (Salt River Project)
Salt River Project, P. 0. Box 1980, Phoenix, Arizona 85001,
in cooperation with U. S. Water Conservation Laboratory,
4331 East Broadway, Phoenix, Arizona 85040
16060DRV
J2.
13. Type r.f lepor and
Period Covered
rganization
A field project demonstrated the feasibility of renovating secondary sewage effluent
by ground water recharge with infiltration basins. Maximum loading rates were
obtained with cycles of 20 days flooding rotated with dry periods of 10 days in the
summer and 20 days in winter. With these schedules the system could infiltrate
300-400 ft/year using a water depth of 1 ft. Grassed basins had higher infiltration
rates, and a gravel covered basin had a lower infiltration rate than a bare soil
basin. Essentially complete removal of BOD and fecal coliform, and significant
removal of phosphorus, nitrogen and fluoride were obtained. Hydraulic properties
of the aquifer were evaluated by analog from the response of piezometric heads in
the ground-^water system to infiltration. These properties were then used in the
design of a prototype system, which would yield renovated water at an estimated
total cost of about $5 per acre-foot at the pump.
17a. Descriptors
*sewage disposal, *tertiary treatment, *water reuse, *ground water recharge,
*water spreading, aquifer characteristics, water table, design, infiltration,
oxygen demand, nitrogen, denitrification, phosphates, fluorides, boron, coliforms
17b. Identifiers
Flushing Meadows Project, Salt River bed, Phoenix, Arizona
17c. COWRRField &
18, Availability
Group
19. Mmn^ty C!.siSS.
"20. Security Class.
21,
21.
No. of
Pages
Price -4
*
Send To :
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
US DEPARTMENT OF THE INTERIOR
WASHINGTON, D. C. 2024O
a Bouwer instrution U. S. Water Conservation Laboratory
,/RSI C i ''- ,d (REV . J UNE 1971)
------- |