United States
Environmental Protection
Agency
Robert S. Kerr Environmental
Research Laboratory
Ada, OK 74820
Research and Development
EPA/600/S2-85/016 Apr. 1985
Project Summary
Optimization of Nitrogen
Removal by Rapid Infiltration
E. R. Bennett, L. E. Leach, Carl G. Enfield, and David M. Walters
The objective of this research field
study was to evaluate the operational
methods for optimization of nitrogen
removal in the rapid-Infiltration proc-
ess. Previous studies (1,2) at the same
site had showed a high degree of bio-
logical nitrification but essentially no
nitrogen removal through denitrifica-
tion. A further objective of the study
was to evaluate the efficacy of
removal of other pollutlonal constitu-
ents under the operational conditions
necessary for enhanced nitrogen
removal.
The rapid infiltration field site con-
sisted of three basins with surface
areas of 0.36, 0.19 and 0.22 hectares
(0.87, 0.47 and 0.54 acres). The beds
had approximately 3 meters (10 feet)
of granular earth materials overlying
an impermeable shale strata. The
basins had underdrain pipes installed
just above the shale base. Primary ef-
fluent from the City of Boulder waste-
water treatment plant was applied to
the basins utilizing hydraulic loading
rates ranging from 4.4 to 42
meters/year (IB to 140 ft./yr.). Three
types of loading sequences were
used: flood loading every three and
one-half days, flood loading daily, and
sprinkler system loading based on soil
moisture sensors and computer
analysis and control. The influent and
effluent quality variations were
measured over a three-year period and
the performance of the system was
related to the operational parameters.
The quality parameters utilized were
nitrogen (total, organic, Kjeldahl, am-
monia and nitrite/nitrate),
biochemical oxygen demand, total
organic carbon, suspended solids,
total and fecal conforms, phosphorus.
and pH. Total nitrogen removals were
increased substantially over the
previous studies. Under optimum con-
ditions, sustained removals above
seventy-five percent were achieved
with values for individual weeks in
the mid eighty percent range. In-
creased denitrification resulted from
maintaining saturated soil conditions
for long periods, using low infiltration
rates and reduced hydraulic loadings.
The removal of phosphates was
shown to be directly related to a
critical phosphorus mass loading of
3.0 Kg/Ha-d (0.3 gm/m'-d) that repre-
sented the mineralization rate for the
soil chemistry of the field system.
Sustained removals greater than
ninety-five percent were achieved.
BOD and TOC removals were high
throughout the study with most BOD
values above ninety percent.
The project report is published in
two volumes based on each of the
two loading modes, the manual
loading and the automated loading
system with computer control.
This Project Summary was devel-
oped by EPA's Robert S. Kerr Environ-
mental Research Laboratory, Ada,
OK, to announce key findings of the
research project that is fully docu-
mented in a separate report of the
same title (see Project Report order-
ing information at back).
Introduction
The high-rate land treatment process
utilizing rapid-infiltration beds is an
economically attractive, low energy con-
suming process providing a high degree
of pollutant removal for municipal waste-
waters. When rapid-infiltration beds are
-------
designed with direct discharge to the
groundwater, nitrogen removal through
soil adsorption, biological nitrification and
denitrification becomes one of the key
parameters in the design and operation of
the process. In the earlier studies at the
same site, it was found that nearly com-
plete ammonia conversion to nitrate was
accomplished when applying either pri-
mary or secondary effluent but that the
removal of nitrogen from the wastewater
denitrification was relatively low, in the
range of fifteen percent. Evaluations have
been made on several rapid-infiltration
sites located in different regions of the
country and the results have been re-
ported in the literature. In general,
nitrogen removals in the range of thirty to
seventy percent have been attained.
This research was focused on evalua-
tion of the operational conditions that in-
fluence conversion and removal of the
nitrogen forms in the treatment of primary
wastewaters. In addition, the treatment
efficiencies for biochemical oxygen de-
mand (BOD), total organic carbon (TOO,
suspended solids, coliform bacteria and
phosphorus were evaluated under a range
of operating conditions.
Conclusions
Rapid-infiltration treatment utilizing
flood loading was shown to be capable of
removing seventy-five percent of the total
nitrogen from wastewater, continuously
producing an effluent with a concentra-
tion of less than ten mg/l of N. In addi-
tion, under reduced loading conditions,
effluent total nitrogen levels of 7 mg/l,
phosphorus of 0.5 mg/l and BOD of 5
mg/l were attained.
Nitrogen removal was enhanced by
creating reducing conditions in the soil
system. Maintaining a saturated soil
moisture condition with long periods of
continuous flooding of the basins pro-
duced the highest nitrogen reductions. In
addition, it was found that low infiltration
rates improved removals by increasing the
contact time of the water with the soil
and by extending the periods of saturation
of the upper portion of the soil. Lowering
the mass loading rate and the hydraulic
loading of the system reduced the am-
monia leakage from the system.
The extent of phosphorus removal was
found to be correlated to a distinct max-
imum mass loading of 3.0 Kg/Ha-d (0.3
gm/m2-d). It was concluded that this
value represented the mineralization rate
for the soils employed in the study. When
loadings exceeded this level, the minerali-
zation reaction was not capable of regen-
erating the sites for the phosphorus ad-
sorption at a fast enough rate to maintain
a balance between the two reactions and
the treatment efficiency declined rapidly.
BOD reductions were high for all load-
ing conditions studied. A large portion of
the BOD was removed in the solids mat
produced at the surface of the beds by
the accumulation of the solids present in
the primary effluent. BOD removals were
improved at lower hydraulic loading rates.
The rapid-infiltration basins provided
major reductions in coliform bacterial den-
sities. However, the removals were not
sufficient to meet some environmental
water quality standards and disinfection
processes may be a required part of the
process for discharging of the effluent
into a stream.
Weather-dependent operational prob-
lems with the flood loaded beds were
minimal throughout the year. The use of
primary effluent did not create odor prob-
lems. Operation of the sprinkler loading
system was not possible when the am-
bient air temperature dropped below
freezing.
Several different loading rates and pat-
terns were evaluated. The results for the
best conditions are based on the four-
weeks' result in the maximum removals.
Optimum four-week nitrogen removals
were found to be seventy-seven percent
for the flood loaded beds utilizing the
conditions of daily application of 4.67
cm/day. Nearly the same optimum remov-
al, seventy-six percent, was achieved with
loading twice per week at a rate of 10.3
cm/day. In both cases, reducing condi-
tions were maintained by keeping the
beds flooded over for several weeks dur-
ing the loading sequence. Nitrogen
removals were somewhat less for the
sprinkler loaded system with a maximum
four-week value of sixty-five percent at a
loading rate of 3.84 cm/day. Phosphorus,
BOD, and suspended solids removals
were all greater when the sprinkler loading
system was utilized.
Several conditions can be defined for
the rapid-infiltration system studied that
resulted in an optimum operational range.
The optimization can be defined in terms
of an operating range that will produce
the best combination of removals for four
constituents: BOD, phosphorus, ammonia
and nitrate. The effluent requirements for
a rapid-infiltration system may require
removal of some or all of these constitu-
ents.
Enhanced BOD removals were found to
result primarily from lowering the
hydraulic loading rate of the system.
Phosphorus removals were found to be
high when the mass loading rates on the
long and short term were less than the
mineralization rate of the soil material. In
this study the value was found to be 3.0
Kg/Ha-d (0.3 gm/m2-d). With an influent
wastewater phosphorus concentration of
approximately 7.5 mg/l, the hydraulic
loading rate should not exceed about 4
cm/day (50 ft/yr, 1 gpd/ft2). Ammonia
leakage was also found to be reduced at
low mass application rates. A value of 4
cm/day for the hydraulic loading rate was
also found to be advantageous for im-
proved nitrogen removal operating condi-
tions.
While lowering the loading rates
enhanced the removals of these three
parameters, there are two factors that
provide the lower limit of hydraulic
loading that can be used for an optimum
system. One of these is system cost, but
more importantly from the standpoint of
performance is that nitrate removals re-
quire a high enough loading rate to pro-
duce saturated soil conditions. This is
governed by the grain size distribution of
the soil and the infiltration rate through
the solids mat that forms at the surface of
the beds. In this research, it was found
that it required approximately 1000 Kg/Ha
(10 mg/cm2. 0.022 Ib/ft2) of wastewater
suspended solids captured at the surface
to produce an infiltration rate low enough
for the continuous flooding of the bed
surfaces to create reducing conditions.
This would require that the beds be
loaded for more than seven weeks at a
rate of 4 cm/day after each drying period
before reducing conditions could be fully
reestablished. It was difficult to attain
these conditions during the low loading
rate phase of the flood loading studies on
bed 1 and the sprinkler loading phase for
beds 2 and 3. Two approaches could be
used to improve the system operation.
Resting periods, without scarification,
should be short to prevent drying out of
the solids mat on the surface of the beds.
New designs should use tighter, more
organic soils with high cation exchange
capacities in order to reduce infiltration
rates and enhance the adsorption capacity
for ammonia and allow somewhat higher
hydraulic loading rates while maintaining
saturated conditions. Shallow scarification
should be used only when the percolation
rate of the beds cannot be restored by
resting. Compaction of the soil to a
prescribed Proctor density could be used
after the scarification operation.
-------
Recommendations
The nitrogen and phosphorus removal
mechanisms in land treatment systems are
dependent on the adsorption capacity of
the soil for these constituents and by the
regeneration rate of the sorption sites by
biological reactions for nitrogen and
chemical precipitation reactions for
phosphorus. Studies relating the proper-
ties of different soils with removal effec-
tiveness over a range of loading condi-
tions would provide a greater understand-
ing of the mechanisms of the treatment
process and provide direction for the
selection of the land treatment location
for the design engineer.
Nitrogen removal effectiveness has
been shown to be related to maintaining
saturated soil moisture conditions that
produced the reducing environment. Fur-
ther development of the operational
parameters for creating optimal soil
moisture conditions, particularly with the
sprinkler loading method, would be
useful.
Most rapid infiltration systems have
been designed with the concept that the
effluent percolates to a ground water
aquifer. The alternate system utilizing
beds designed with underdrain pipes and
discharge to a surface stream, such as the
system used in this research project, can
be designed to be a highly effective
means of providing advanced wastewater
treatment with a high degree of reliability
while requiring low energy consumption
and ease of operation. The underdrained
system concept should be developed fur-
ther as a contained system for small com-
munities or cluster home developments.
Facilities
The rapid-infiltration system utilized in
this field investigation was located on the
site of the 75th Street wastewater treat-
ment plant in Boulder, Colorado. The 75th
Street plant is a trickling filter plant
situated on the south bank of Boulder
Creek. Construction of the rapid-
infiltration site was accomplished in the
spring of 1976. The site consisted of three
basins, designated one, two, and three,
moving from south to north. The con-
figuration of the beds is shown in Figure
1. The earth materials in the beds were
approximately three-meters (10-ft) deep
and the base material was an imperme-
able shale strata. Earth berms, approx-
imately 0.8 meters (2.5 ft) high, separated
the basins, while a clay dike surrounding
Earth Berm with Clay Dike ,,
the entire basin area isolated the system
from groundwater interference.
The rapid-infiltration site was located in
the Piedmont section of Boulder County.
These deposits were composed of loose
sand, loose gravel, and clayey sand
underlain by bedrock of oceanic origin.
Each of the beds originally had approx-
imately 75 cm (2.5 feet) of finer textured
material at the surface overlying coarse
sand and gravel in the lower portion of
the system. Prior to this study, 60 cm (2
ft) of the finer material was removed from
beds two and three so that 15 cm (0.5 ft)
of finer textured materials remained. All of
the beds had twelve or more percent clay
and two and one-half percent or more
organic matter in the surface layer.
Primary effluent wastewater was
pumped through a 0.35 meter (14 in.)
delivery pipe to the rapid-infiltration site.
Two methods of bed loading were used.
From December, 1980 to January, 1983,
the beds were flood loaded. During the
summer of 1982, a fixed sprinkler distribu-
tion system was installed on beds two
and three and this loading method was in-
itiated in January 1983. The system con-
sisted of a network of 10 cm (4 in.)
diameter aluminum irrigation pipes with
Primary
Clarifier
0.22 Ha. 0.54 Ac
Splitter Box
Valve
Surface Sprinkler System
Effluent
Sampling
Manhole
0.13 Ha, 0.47 Ac
Flood Loading Pipes
Bedl
0.35 Ha, 0.87 Ac
Trailer with
Sprinkler Pump
18 cm dia. Underdrains
O.O04 Slope
and Flow Controls
N
Figure 1. The rapid infiltration system layout.
-------
risers installed at 6.2 meter (20 ft) inter-
vals. Rotating sprinkler nozzles were in-
stalled at the top of the risers with each
nozzle having a 4.4 mm (11/64 in.) orifice.
There were 13 rotating nozzles per bed.
The spray loading system was designed to
operate at a gage pressure of 2585 mm of
Hg (50 psi). The pipeline from the primary
clarifier to the beds was modified so that
it passed through a trailer in an inverted
"U" shaped loop. A tee was installed in
the pipe so that the flow could be di-
rected to the flood loading of the beds or
diverted to a separate pump and valving
system that would allow the sprinkler
systems to be operated.
The primary purpose of the sprinkler
system was to use a computer-operated
loading pattern based on sensing in-
struments at the site to control the
moisture content of the soil in order to
optimize the conditions for the nitrifying
and denitrifying bacteria in the soil. Each
bed had three sets of electrodes at 5 cm
and 30 cm depths that measured soil
temperature. Three moisture-content
probes were also installed in each bed at
the 5 cm depth. The temperature and
moisture content data were measured on
computer command and sent by
telephone to the control computer located
in the Robert S. Kerr Environmental
Research Laboratory in Ada, Oklahoma.
The computer was programmed to
analyze the data and relay a signal back
to the control trailer at the Boulder site if
the spray pumps were to be turned on
based on the sensing of soil moisture
contents. The system was programmed to
be inoperable when the temperature was
below -5°C or when the wind velocity
exceeded 56 Km/hr (35 m.p.h.). It was
also possible to operate the sprinkler
system manually at the site.
After the wastewater passed through
the soil materials in the beds, it was col-
lected in an underdrain system consisting
of two perforated 18 cm (7 in) diameter
pipes running the length of each bed at a
soil depth of 2.5 to 3.2 meters (8 to 10 ft),
just above the shale layer at the bottom
of the beds. The line from each of the
beds had V-notch weirs installed in the ef-
fluent sampling manhole for flow rate
measurements.
For the first two years of the study, the
influent pipe led directly to an in-line,
totalizing, propeller flow meter and then
to a distribution spitter box. Lift plate
gates were used to divert the flow to bed
1, bed 2, or bed 3 during flood loading. In
the third year of the study, a section of
the influent pipe was removed, reducers
were attached to each end of the remain-
ing pipe and a pipe was installed in a 'U'
shape, through the control trailer and
back to the original pipe. The pipe had a
tee connection to a 7 cm (3 in) diameter
pipe which led to a shutoff valve and the
sprinkler pump. The pump discharge
branched into two pipes each having a
pressure actuated, rate of flow controller.
Each of the two pipes exiting the control
trailer was connected to a 10 cm (4 in)
diameter aluminum irrigation pipe that led
to the west edge of beds 2 and 3 and
continued down the center of the beds to
the east end. The ends of the irrigation
pipes were capped and 2 cm (0.75 in)
diameter risers were installed at 6.2 meter
(20 ft) intervals. Rotating sprinkler heads
at the top of the risers had an application
radius of approximately 12.3 meters (40
ft).
Operation
Two effluent samples, acidified and
unacidified, were taken at the outlet
streams of the individual beds in the ef-
fluent manhole on intervals of 3, 7, 10,
24, 34, 48 and 72 hours after loading was
initiated and effluent flows were measured
at each time. The acidified aliquots were
used for the analysis of ammonia, total
Kjeldahl nitrogen, nitrate, phosphorus,
and TOC. The unacidified volumes, taken
at effluent peak flow, were for pH, BOD
and suspended solids analyses. Special
grab samples of influent and effluent
flows were obtained using sterilized bot-
tles for the coliform tests and using nitric
acid acidified bottles for the calcium,
magnesium, and sodium series. All chem-
ical analyses were accomplished using
methods from Standard Methods for the
Examination of Water and Wastewater
(3), or Methods for Chemical Analysis of
Water and Wastes (4), using blanks and
spikes for quality control.
Nitrogen removal involved biological
nitrification and denitrification processes
and for this reason, wastewater treatment
plant primary effluent was used as the in-
fluent to the process to provide an ade-
quate carbon source for the biological
growth. Another reason for the use of
primary effluent was to demonstrate that
rapid infiltration was an effective and
reliable means of achieving high removal
efficiencies for BOD, suspended solids,
phosphorus and other pollutional param-
eters while operating in a mode intended
for optimum nitrogen removals.
The research was conducted as a field
study using several different loading rates,
loading-resting patterns and two applica-
tion techniques: flood loading and
sprinkler loading. At times, the combina-
tion of these parameters did not result in
optimum conditions. For this reason, the
time-concentration curves are presented
to illustrate the effect of the variables.
The average removal percentages are
used to summarize the treatment achieved
by the rapid-infiltration process and values
for the best four-week period during the
study are shown and give an indication of
the performance that could be expected
for a system operating continuously at
near optimum conditions. Individual week-
ly removal percentages displayed some
degree of variability, due to the changing
nature of the pollutional strength of the
influent stream, which is typical of a field
study. For this reason, longer term four-
week trends in removal percentages are
more significant.
During the study, the Boulder waste-
water treatment plant was involved in a
major construction project. Relocation of
the rapid-infiltration influent pipe was
necessary and this caused some down-
time for the beds which was not a
planned part of the research. The inactive
period for beds 2 and 3 during weeks
118-135 was caused by the construction
project.
Three different modes of wastewater
application were used during the study.
Initially, the beds had a target value for
flood loaded of approximately 40 cm (16
in.) of primary effluent on three and one-
half day intervals. This pattern was
followed for the first 78 weeks of the
study, encompassing the period from
December of 1980 through March of
1982. The second loading sequence,
which involved only bed 1, utilized a
lower loading rate of approximately 5 cm
(2 in) of flooding application each day.
This was used from April, 1982 through
February of 1983. The third mode, the
sprinkler distribution system was ready for
use on beds 2 and 3 after January, 1983,
and it was necessary to discontinue the
research operation of bed 1 because the
overflow of the sprinkler pumping system
was directed to bed 1 and made it im-
possible to maintain a consistent and
measured loading amount. The sprinkler
system operated from January through
December of 1983, although it was inter-
rupted from March through June by the
construction in the treatment plant. The
average loading rates and loading patterns
are summarized in Table 1. The loading
rate for the 3.5-day pattern averaged 123
ft/yr based on the weeks that loading ac-
tually occurred. Resting of the beds was
-------
necessary and occupied 28 percent of the
weeks of this portion of the study. The
overall average loading rate over time in-
cluding resting periods is given in Table 1 .
No intentional resting periods were util-
ized with the 1-day or sprinkler-system
patterns. Downtime periods caused by
construction in the wastewater treatment
plant were not included in the calculation
of loading rates.
Results
The average concentrations for the
rapid-infiltration bed influents and ef-
fluents, as well as removal percentages
for the total period of each loading mode,
are given in Table 2. Some of the meas-
ured chemical constituents of the waste-
water including pH, sodium, and mag-
nesium were essentially unchanged during
the rapid-infiltration process. The slight
increase in calcium probably resulted from
dissolving of calcium carbonate deposits
in the bed material.
Coliform removals were significant, but
relatively high concentrations remained in
the process effluents. The results of this
study show that the long-term use of
underdrained rapid-infiltration beds may
require disinfection of the effluent prior to
discharge into a surface receiving water.
The values shown in Table 2 include
periods of study when the variables tested
were not at optimum. In order to judge a
more optimized system performance,
Table 3 has been constructed with a sum-
mary of the removal percentages for the
important parameters shown and with the
efficiencies obtained under the best condi-
tion studied based on four continuous
weeks of operation.
The first ten weeks of the study were
not included in the selection of the best
four weeks of operation because the beds
had been rested for more than a year and
loading during this period did not repre-
sent normal operation.
The use of rapid-infiltration beds to
remove nitrogen from wastewater involves
me creation OT aerouic conoiuons ior
nitrification, and reducing conditions for
denitrification simultaneously or sequen-
tially in such a manner as to prevent
pulses of ammonia or nitrate from being
released in the effluent during any portion
of the loading and infiltration cycle.
The beds, which had been standing idle
for a year, were scarified prior to any
wastewater application. For the first ten
weeks of loading, the influent percolated
through the beds very rapidly. Infiltration
rates were high (>50 cm/d) and the bed
Table 1. Loading Depths
Operating Mode 1 23
flood (3'A day) flood (1 day) sprinkler
Loading rate, cm/d
including resting ft/yr
gal/d-ft1
Table 2. Average Constituent Concentration
Constituent Bed influent
Mode 1 (3% day)
Nitrogen
Ammonia-N 16.5 mg/l
Nitrate-N 0.27 mg/l
Organic-N 7.05 mg/l
Total-N 23.82 mg/l
T. Phosphorus-P 7.65 mg/l
BOD-5 101 mg/l
TOC 67.3 mg/l
Suspended so/ids 51.2 mg/l
Coliforms/ 100 ml
Total 58.5 x 10'
Fecal 13.4 x 10'
pH 6.85
Calcium 41.6 mg/l
Magnesium 21.5 mg/l
Sodium 56.6 mg/l
Mode 2 (1 day)
Nitrogen
Ammonia-N 13. 75 mg/l
Nitrate-N 0.36 mg/l
Organic-N 8.40 mg/l
Total-N 22.51 mg/l
T. Phosphorus-P 8,94 mg/l
BOD-5 54.5 mg/l
TOC 53.1 mg/l
Suspended solids 53.9 mg/l
Coliforms/ 100 ml
Total -
Fecal 14.3 x 10s
Mode 3 (sprinkler)
Nitrogen
Ammonia-N 17.85 mg/l
Nitrate-N 0. 13 mg/l
Organic-N 6. 76 mg/l
Total-N 24.74 mg/l
T. Phosphorus-P 7. 13 mg/l
BOD-5 80.4 mg/l
TOC 54.8 mg/l
Suspended solids 53.5 mg/l
Coliforms/ 100 ml
TV**a/
Total —
Fecal —
7.42
88.6
1.8
and Removals
Bed effluent
6.5 mg/l
1.85 mg/l
0.85 mg/l
9.20 mg/l
1.33 mg/l
8.8 mg/l
10.2 mg/l
14.7 mg/l
1.6 x 10"
0.65 x 10'
6.85
58.6 mg/l
21.5 mg/l
53.0 mg/l
3.23 mg/l
4.58 mg/l
1.06 mg/l
8.87 mg/l
0.53 mg/l
3.8 mg/l
6.7 mg/l
10.2 mg/l
—
0.93 x 10'
1.28 mg/l
10.50 mg/l
0.64 mg/l
12.42 mg/l
0.55 mg/l
2.2 mg/l
5.0 mg/l
6.0 mg/l
4.67
56
1.15
% removal
60
—
88
61
83
91
85
71
97
95
—
-41
0
6
77
—
87
61
94
93
87
81
—
94
93
91
50
92
97
91
89
3.84
46
0.95
Table 3. Comparison of Three Loading Modes Average and Best 4-Week Results
Operating Mode 1
flood (3'A d)
% removal
avg. 4-week
Total Nitrogen 61 76
Total Phosphorus 83 87
BOD-5 91 91
TOC 85 85
Suspended solids 71 73
2
flood (1 d)
% removal
avg. 4-week
61 77
94 97
93 96
87 91
81 89
3
sprinkler
% removal
avg. 4-week
50
92
97
91
89
65
98
99.5
96
98
-------
effluents were highly nitrified but with
less than forty percent nitrogen removal.
These results are shown as the beginning
of the time-history curves of Figure 2. The
results shown in the figure are for bed 1
for the first 120 weeks of flood loading
and for bed 2 with manually controlled
sprinkler loading during weeks 109-117,
followed by an inactive period caused by
inplant construction during weeks 118-135
and then for bed 2 with computer-
controlled loading for the remaining
sprinkler loading period. The results of the
other beds were similar. The top curve on
each graph shows the loading rate as a
function of the week of loading. The next
curve shows the infiltration rate, and the
third curve relates the total nitrogen con-
centration of the influent and effluent
(solid dark lines) and the ammonia nitro-
gen concentration of the influent and ef-
fluent (lighter solid lines) and the nitrate
nitrogen concentration of the effluent
(dashed lines). Nitrites were included in
the nitrate measurements. Organic nitro-
gen and influent nitrate nitrogen are not
shown on the curves except that they are
included in the total nitrogen values. The
bottom curves show the percent removal
of total nitrogen.
> 20
•^ 15
-------
total nitrogen loading of 24.5 Kg/Ha-d
(21.5 Ib/ac-d), caused an overloading of
the system resulting in continued increase
in ammonia leakage into the effluent.
When the loading was reduced to 17.0
meters/year (55.8 ft./yr.) with a total
nitrogen loading of 11.1 Kg/Ha-d (9.9
Ib/ac-d) beginning with week 71, the am-
monia leakage showed a steady decline.
The removal of nitrate was accom-
plished by maintaining flooded-over condi-
tions on the basins. This was done for
periods of more than a month without
adversely affecting the total nitrogen con-
centration in the effluent. The low infiltra-
tion rates that resulted from the formation
of solids mats on the bed surfaces were
necessary for the denitrification reaction
to occur. The only times when poor nitro-
gen removals were observed were im-
mediately after resting periods with
scarification for the first three loading se-
quences when nitrates were released from
the system due to the fact that it took
several loadings to reduce the infiltration
rate and establish reducing conditions.
The operation with continuously saturated
bed surface conditions was possible due
to the field layout and the soil profile. It
was possible for air to enter the lower,
coarse strata of the beds from the sides
and from the underdrain system.
The major factor affecting the effluent
nitrogen concentration appears to be the
nitrogen mass-loading rate with respect to
the soil cation exchange capacity. The
cation exchange capacity is usually
greater for soils of higher organic content.
The cation exchange capacity for the up-
per layer of soil in the beds of this study
ranged from 15 to 24. The coarse ma-
terials in the lower portions of the beds
had measured values in the range of 2 to
3. Soil organic matter, including that
disked in during scarification provides the
energy source for the denitrifying bacteria
and this may be an important source
along with the BOD of the wastewater
being treated.
Phosphorus removal from wastewater
was evaluated in conjunction with the
operating conditions needed for nitrogen
removal. Land treatment of phosphorus
involves the two sequential reactions of
adsorption of the phosphate ion and
precipitation of a solid that is retained in
the soil matrix. The precipitation usually
appears to be the rate-limiting step. When
a system is overloaded, the rate of pre-
cipitation becomes inadequate to con-
tinually renew the sites for the adsorption
reaction, and phosphorus leakage occurs
in the process effluent.
Phosphorus removals in this study are
summarized in Figure 3. The lower curves
in each part of the figure show that the
treatment system was severely overloaded
during the initial 32 weeks of the
research. Long resting periods in the
following weeks caused the system to
completely recover by week 50. Lower
loading rates and more frequent resting
periods maintained efficient phosphorus
removals for the remainder of the study.
Three periods were selected from the
data array (weeks 15-19, 74-78 and 92-96)
as steady-state condition for the points
(squares) in Figure 4. A line was con-
structed through the points to illustrate
the removal capacity of the soil in this
research. The reported results of other
researchers have also been shown. The
curve illustrates that there was a very
discernible maximum loading rate that
gave high removals and there was very lit-
tle tolerance for overloading. Analysis of
the phosphorus removal curves for each
of the beds during the period of the study
showed that when the long-term
(months) or short-term (days) phosphorus
loading was greater than 0.3 g/m2-d, the
removal efficiency declined quite rapidly,
and when the phosphorus loading was
reduced below that level, the phosphorus
removal improved. It was concluded that
the value of 0.3 g/m2-d represented the
mineralization rate (phosphorus precipita-
tion rate) for the soil of the beds used in
this study. The other researchers points
on Figure 4 show that the mineralization
rate may have been quite different for the
various soils encountered in the different
research projects and that the value may
become less with many years of operation
of a system.
BOD and TOC gave similar patterns of
high removal under all operating condi-
tions for the rapid-infiltration process. The
most significant parameter affecting BOD
removal was the hydraulic loading rate.
The effluent suspended solids concen-
trations were higher than expected for an
earth filtration system. Suspended solids
concentrations followed the same removal
pattern as that of BOD, except that the
^20
£ 10
^ °0
-J 0
3.5 Day Cycle
1
Loading Rate: Bed 1
1 Day Cycle
Sprinkler
Manual , Computer
Construction
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150160
Week Number
Phosphorus: Bed 1
0 10 20 30 40 50
60 70 SO _90 100 110 120 130 140 150 160
Week Number
Phosphorus Removal: Bed 1
(Bed 2)
2/00
Q
E 80
«
* 60
0 40
£ 20
8 0
W \ A.
\ 1
xA/^A 1
\ i /^
M
f__S^T«
\
I
I 1 1 1 1 1 1 1 1 1
\
\
\
,
, ,
— ~*\ f ^
y
i i i
fc 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Q.
Week Number
Figure 3. Phosphorus analysis.
-------
influent values were lower and the ef-
fluent values were two to three mg/l
greater. Inspection of the underdrain
system revealed extensive biological slime
growth, which was thought to be the
source of the paniculate matter in the ef-
fluent. For systems discharging directly to
the ground water, this would not occur.
One of the major considerations in the
use of rapid-infiltration beds with primary
effluent is the potential of odors near the
site. Odors were not a problem during the
period of this research. Most of the time
during the study, standing water, primary
effluent, was present on the beds. Natural
surface aeration was sufficient to prevent
septic conditions and the ponding did not
cause offensive odors.
The highly nutriated soils in the beds
produced voluminous weed growth on
the surface of the beds during the warmer
months. Weed cutting was a necessary
part of the system maintenance program
at this site for aesthetic reasons. Cutting
was done about every six weeks in the
summer, during the period when the beds
were rested. Excessive algae growth oc-
curred when the ambient air temperature
was above 33°C (90°F) for extended
periods of time. This condition caused
clogging of the solids mat on the beds ac-
companied by greatly reduced infiltration
rates. When this condition became ex-
cessive, it was necessary to rest the beds,
which tended to necessitate shorter, more
frequent intervals between resting periods
of the beds in the summer than in the
winter.
Cold weather conditions had very little
effect on the operation of the flood
loaded beds. Temperatures as low as
-23°C (-27°F) were encountered for
short periods of time. Under the worst
conditions, an ice layer several inches
thick formed on the beds but seemed to
have little effect on the overall functioning
of the system. Problems were encoun-
tered with the sprinkler system during
severe winter conditions. The system ex-
perienced operating problems when the
Middleville. Mi. (7) Helen. Ga. (9)
\WardCo.. (6)/ ^Tallahassee. Ft. fW)
CCa
100
I
90
80
70
60
50
40
30
20
10
.Caddilac. Mi. (8)
Brook ings. SD.fl 1)
FortDevens. Ma. (12)
This Study
Phoenix, Ar. (13)
Holister. Ca.<5)
Figure 4.
.1 .2 ,3 A .5 .6 .7
Long Term Phosphorus Removed(g/m2.d)
Phosphorus removal comparison with other studies.
8
.8
.9
1.0
ambient air temperature went below freez-
ing. Although the wastewater was warm
enough to flow through the sprinkler
nozzles in the normal fashion, some of
the water would wet the outside of the
sprinkler head housing. When frozen, this
water would prevent the sprinkler head
from turning, causing it to spray in the
same position throughout the loading
cycle.
The thin-walled aluminum irrigation pipe
was joined with connections having neo-
preme flap gaskets to assure that the pipe
was water tight. When pressurized the
pipe would drain at the joints when the
pump was stopped between loadings.
This prevented the freezing of the water
within the pipe between loadings. This ar-
rangement worked satisfactorily except
during large snowstorms. The snow and
ice surrounding the pipes would seal the
joints, causing the pipes to remain full of
water, in some instances causing irriga-
tion pipe rupture. Spray loading systems
must be used with care in winter weather
and provisions made for cold periods
when the system cannot be operated.
References
1. Smith, D.G., K.D. Linstedt and E.R.
Bennett. Treatment of Secondary Ef-
fluent by Infiltration-Percolation,
EPA-600/2-79-174, U.S. Environmen-
tal Protection Agency, Ada, Okla-
homa, 1979. 104pp.
2. Hartman, R.B., K.D. Linstedt, E.R.
Bennett and R.R. Carlson. Treatment
of Primary Effluent by Rapid Infiltra-
tion. EPA-60072-80-207, U.S. En-
vironmental Protection Agency, Ada,
Oklahoma, 1980. 104pp.
3. Standard Methods for the Examina-
tion of Water and Wastewater. Four-
teenth and Fifteenth Editions,
Academic Press, New York. APHA,
AWWA.WPCF, 1976-80. 1193 pp.
4. U.S. Environmental Protection Agen-
cy. Methods for Chemical Analysis for
Water and Wastes. 1979.
5. Pound, C.E., R.W. Crites and S.C.
Reed. Land Treatment: Present
Status, Future Prospects. American
Society of Civil Engineers, Civil
Engineering, 48, (61:98-102, 1978.
6. Sturdevant, C. Evaluation of Forest
Treatment of Wastewater in an Alpine
Environment. M.S. Thesis, Univ. of
Co., Boulder, Colorado, 1984. 152pp.
7. Sutherland, J.C., J.H. Cooley, D.G.
Neary and D.H. Urie. Irrigation of
Trees and Crops with Sewage Stabili-
zation Pond Effluent in Southern
-------
Michigan. Proceedings of Wastewater
Use in the Production of Food and
Fiber. EPA-660/2-74-041. Washing-
ton, D.C. p295-313, 1974.
8. Urie, D.H. Phosphorus and Nitrate
Levels in Groundwater as Related to
Irrigation of Jack Pine with Sewage
Effluent. Recycling Treated Municipal
Wastewater and Sludge Through
Forest and Crop Land, Penn State
University Press, University Park,
Pennsylvania, p176-183, 1973.
9. Nutter, W.L., R.C., Shultz and G.H.
Brister. Land Treatment of Municipal
Wastewater on Steep Forest Slopes
in the Humid Southeastern United
States. Proceedings of Symposium
on Land Treatment of Wastewater.
Hanover, New Hampshire. 1978.
10. Overman, A.R. Wastewater Irrigation
at Tallahassee, Florida. U.S. En-
vironmental Protection Agency,
EPA-600/2-79-151. 1979.
11. Dornbush, J.N. Infiltration Land
Treatment of Stabilization Pond Ef-
fluent. Technical Progress Report 3.
South Dakota State University,
Brookings, South Dakota. 1978.
12. Satterwhite, M.B., B.J. Condike and
G.L. Stewart. Treatment of Primary
Sewage Effluent by Rapid Infiltration.
U.S. Army Corps of Engineers, Cold
Regions Research and Engineering
Laboratory. 1976.
13. Bouwer, H., W.J. Bauer and R.D.
Dryden. Land Treatment of Waste-
water in Todays Society. American
Society of Civil Engineers,
48(11:78-81, 1978.
E. R. Bennett is with the University of Colorado. Boulder, CO 80309; the EPA
authors L. £. Leach (also the EPA Project Officer, seebelowl. CarlG. Enfield,
and David M. Walters, are with the Robert S. Kerr Environmental Research
Laboratory, Ada, OK 74820.
The complete report consists of two volumes, entitled "Optimization of Nitrogen
Removal by Rapid Infiltration"
"Volume I. System Description and Evaluation, "(Order No. PB 85-173 938/AS;
Cost: $13.00)
"Volume II. Remote Computer Operating System,"* (Order No. PB 85-173
946/AS; Cost: $26.50)
The above reports will be available only from: (cost subject to change)
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Robert S. Kerr Environmental Research Laboratory
U.S. Environmental Protection Agency
P.O.Box 1198
Ada, OK 74820
"Most readers would not find it necessary to purchase Volume II of this report since the environmental
evaluation and supporting data of the computer-operated system are included in Volume I. Volume II
contains the descriptions of the various hardware components, how they were interfaced, and computer
software for operating the system.
.Government Printing Office: 1985 — 559-111/10824
9
-------
75
;:• to
Is
a>
to
xuim
MO
o. m
O Z
*» o
o
(?
o
0) «
- • o
0D
I
A
U1
to
O)
oo
i)
'
c
<»
»
O
! r
------- |