United States EPA-600/3-81-012
Environmental Protection February 1981
Agency
<&EPA Research and
Development
^Precipitation and
Inactivation of
Phosphorus as a Lake
Restoration Technique
Prepared for
Office of Water Regulations and
Standards
Criteria and Standards
Division
Prepared by
Environmental Research Laboratory
Office of Research and Development
Corvallis, OR 97330
-------
EPA-600/3-81-012
February 1981
PRECIPITATION AND INACTIVATION OF PHOSPHORUS
AS A LAKE RESTORATION TECHNIQUE
G. Dennis Cooke
Department of Biological Sciences
Kent State University
Kent, Ohio 44242
and
Robert H. Kennedy
U.S. Army Corps of Engineers Waterways Experiment Station
Environmental Laboratory
Vicksburg, Mississippi 39180
Project Officer
Spencer A. Peterson
Freshwater Division
Corvallis Environmental Research Laboratory
Corvallis, Oregon 97330
CORVALLIS ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
Protection
230 south Dearborn Street
Chicago, Illinois 60604
-------
DISCLAIMER
This report has been reviewed by the Corvallis Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or recom-
mendation for use.
n
-------
ABSTRACT
Many eutrophic lakes respond slowly following nutrient diversion because
of long water retention times, and the recycling of phosphorus from sediments
and other internal sources. Treatment of lakes with aluminum sulfate and/or
sodium aluminate is a successful method for removing phosphorus from the water
column and for controlling its release from sediments. Twenty-eight lake
projects treated with aluminum salts are reviewed and summarized. The tech-
nique is successful when sufficient doses of aluminum are applied, however a
few undesirable side-effects such as reduced planktonic microcrustacea species
diversity and increases in rooted plant biomass may occur. Two methods for
determining aluminum sulfate doses are compared. Both approaches are related
directly to treatment objectives and involve simple laboratory methods. One
approach emphasizes short-term reductions in water column phosphorus concen-
tration. The other emphasizes long-term control of sediment phosphorus
release while also assuring maximum removal of phosphorus from the water
column at the time of treatment. Maximum dose, dictated by the buffering
capacity of the particular lake, is defined as that dose which reduces pH to
6, a pH favorable for forming insoluble aluminum hydroxide and for assuring
that dissolved aluminum remains below potentially toxic concentrations. Lake
dosage can be easily determined prior to application by titrating several lake
water samples of varying alkalinity with aluminum sulfate, determining maximum
dose for each sample, and establishing a relationship between alkalinity and
maximum dose. This lake restoration technique is successful and long-lasting
when properly applied. It would become a standard means of treating small
lakes or ponds following diversion. Additional efforts should be directed
toward long term monitoring of selected projects, studies of application
methods and the potential toxicity of chemically treating lakes.
-------
CONTENTS
I. Introduction 1
II. The Chemistry of Phosphorus Inactivants/Precipitants 2
III. Criteria for Successful Lake Treatment 5
IV. Case Histories of Phosphorus Precipitation/Inactivation 23
V. Costs for Phosphorus Precipitation/Inactivation 33
VI. Summary 34
References 37
-------
ACKNOWLEDGEMENTS
The authors thank the Department of Biological Sciences and the Office of
Research and Sponsored Programs at Kent State University, the Corvallis Envi-
ronmental Research Laboratory of the USEPA (Grant R801936, K. W. Malueg,
Project Officer) and the Allied Chemical Company for their support of portions
of the research described herein. The manuscript was prepared during an
Intergovernmental Personnel Assignment between the Corvallis Laboratory and
Kent State University (S. A. Peterson, Project Officer). We thank three
anonymous reviewers for the helpful comments. We also acknowledge the
patience and skill of Marilyn Silvey and Carol Toncar, whose help throughout
was invaluable.
vi
-------
I. Introduction
Nutrient diversion alone does not always bring about prompt and suffic-
ient reduction in lake water concentration due to recycling from nutrient-rich
sediments (Larsen et al. , 1975; Cooke et al. , 1977). Certain lakes continue
to have nuisance algal blooms and require an additional restorative steps.
The phosphorus (P) precipitation/inactivation technique is a procedure to
remove P from the water column and to control its release from sediments in
order to achieve P-limiting conditions to algal growth following nutrient
diversion.
The salts of iron, aluminum, and other metals have long been used in
advanced wastewater treatment to remove P and this technology was extended to
lake rehabilitation. Iron (Fe III) was apparently first used at Dorrdrecht
Reservoir (Netherlands) in 1962 (Peelen, 1969) and the first use of aluminum
sulfate to precipitate P from the water column at Lake Langsjon, Sweden, in
1968 (Jernelov, 1970). Since those early lake treatments, considerable
advances in our knowledge of dose, effectiveness, costs, side-effects and
other factors have occurred. The purpose of this paper is to describe the
current state-of-the-art of this lake rehabilitation technique.
-------
II. The Chemistry of Phosphorus Inactivants/Precipitants
The chemical and physical bases for the use of phosphorus precipitants/
inactivants is their ability to form complexes, chelates, and insoluble
precipitates with phosphorus. Iron and calcium have distinct disadvantages in
this regard. During spring and fall circulation, or just after thermal strat-
ification, lake sediments are covered by an oxidized microzone which traps P
by precipitation with hydrated iron oxides. As the dissolved oxygen content
of the hypolimnion falls, the redox potential also decreases and P is released
during the reduction of ferric hydroxide and other iron complexes. Thus iron
is not a suitable inactivant for long-term control of P release. This rela-
tionship of iron, phosphorus, and redox is summarized by Morgan and Stumm
(1964) and Stumm and Morgan (1970). Calcium additions to lakes may lead to
the formation of apatite and hydroxylapatite, but effective P removal occurs
at pH values above those those found in natural waters (Stumm and Morgan,
1970; Wetzel, 1975), and thus addition of lime (CaO) for P removal may induce
damage to the biota.
Aluminum has been used most often in P inactivation/precipitation
projects because Al complexes and polymers are inert to redox changes, are
effective in entrapment and removal of inorganic and particulate P in the
water column when given sufficient contact time, and are apparently of low
toxicity at the pH and dose required to bring about lake improvement. The
most common forms of aluminum used are aluminum sulfate (alum) and sodium
aluminate, both of which produce aluminum hydroxide in aqueous solution. The
floe of aluminum hydroxide, once deposited, seems to "consolidate" with the
sediments within weeks (Cooke e_t al. , 1978). Phosphorus-rich water, including
interstitial water and groundwater, are presumably stripped of P as they move
through the floe.
Aluminum hydroxide, which is formed when aluminum salts are added to
water is unique among nontransition metal hydroxides in that it is amphoteric,
forms complex ions with other substances commonly found in natural waters, and
polymerizes (Burrows, 1977). It is these properties which have made aluminum
a valuable agent for the treatment of water and wastes.
The dissolution of aluminum salts in pure water brings about the coordin-
ation of six water molecules to form a hydrated trivalent aluminum ion (Fiat
and Connick, 1968). This ion then undergoes a series of pH dependent hydrol-
ysis reactions ultimately forming colloidal aluminum hydroxide:
Al+3 + 6H20 ? A1(H20)6+3 (1)
A1(H20)6+3 + H20 + A1(H20)50+2 + H+ (2)
etc.
A1(OH)+2 + H20 * A1(OH)2 + H+ (3)
A1(OH)2+ + H20 ? A1(OH)3 + H+ (4)
-------
The result is a decrease in solution pH, and in natural waters, a decrease in
total alkalinity.
Aluminum hydroxide, because of its amphoteric property, is converted to
the soluble aluminate ion in basic solutions:
A1(OH)3 + H20 * A1(OH)4 + H+ (5)
The distribution of hydrolyzed aluminum species is dependent on solution
pH (Figure 1), with settleable, polymerized A1(OH)3 predominating between pH+6
and 8. The formation of aluminate ion is favored above this range, while Al 3
predominates below pH 4. A number of other hydrolysis species have been
postulated and the reader should consult such reviews as that of Hayden and
Rubin (1974) for more complete discussions of the hydrolysis of aluminum.
Removal of P by aluminum can occur by precipitation of A1P04 (Recht and
Ghassemi, 1970), sorption of phosphates to the surface of aluminum hydroxide
polymers or floe (Eisenreich et al., 1977), and/or by entrapment/sedimentation
of P-containing particulates by aluminum hydroxide floe. The dominant mechan-
ism(s) for removal will depend on chemical and physical conditions under which
treatment occurs and the nature of the P species present.
The removal of inorganic P is dependent primarily on reaction pH and P
concentration. In general, high inorganic P concentrations (> 1 mg P/1), such
as those encountered during wastewater treatment, and low pH would favor the
formation of A1P04. Theoretically, under these conditions and in the presence
of excess P, the removal of one mole of P as A1P04 would require one mole of
aluminum. However, Al/P molar ratios for maximal removal during conventional
wastewater treatment are often greater than unity. For example, Linsted et
al. (1974) achieved maximal P removal from sewage at pH 5-6 using a treatment
Al/P molar ratio of 2.3.
At lower inorganic P concentrations (< 1 mg P/1) and higher pH, OH
competes with P04 3 for aluminum ions (Hsu, 1976), favoring the formation of
aluminum hydroxide-phosphates. Under these conditions, maximal P removal
efficiencies occur at even higher Al/P molar ratios. Maximum P removal from
Cline's Pond water (0.43 mg P/1 iniital soluble reactive P concentration),
which was found to be pH dependent, occurred at Al/P molar ratios ranging from
5.7 to 7.2 (S. A. Peterson et al., 1976). Al/P molar ratios in excess of 525
were required to achieve 90 percent P removal from ulfiltered Lake Mendota
epilimnetic water (0.01 mg P/1 initial soluble reactive P concentration) at pH
6.5 to 7.0 (Eisenreich et al_. , 1977).
Dissolved organic phosphates are removed considerably less effectively
due presumably to their complex molecular structure and chemical character-
istics (Browman et al. , 1973, 1977; Eisenreich et al., 1977). The effeciency
of removal appears to be related more to the types of organic phosphates
present than to treatment conditions. If the objective of an alum application
is to reduce P to limiting concentrations, as would be the case for lake
treatments, failure to remove dissolved organic P could be of significance
-------
^ CO
1C
o
I/)
s-
0)
-o
o
s-
o
Ol
+->
D.
O
T3
(0
to
CO
oil
jC
o
-------
since it has been shown that certain nuisance blue-green algae, under
P-limiting conditions, will produce an enzyme (phosphatase) which will remove
inorganic P from any organic phosphates at rates sufficient to support algal
blooms (Heath and Cooke, 1975).
Effective removal of particulate P depends on both the quantity and
quality of aluminum hydroxide floe produced during treatment. As indicated
above, aluminum solubility is pH dependent with maximum floe formation
occurring at approximately pH 6 to 8. Therefore, it would be expected that
the potential for entrapment of particulates would be greatest within this
range. Recht and Ghassemi (1970) observed maximal floe formation in the range
pH 5-7. Similar results have also been obtained for mineral suspensions with
the greatest reduction in turbidity occurring in the range pH 6.8 to 7.8
(Packham, 1962). Outside of the range of minimum solubility, dissolved alum-
inum concentrations increase, floe size and quantity decrease, and settle-
ability is reduced. Therefore, with pH controlled, particulate removal rates
are then dose dependent.
The amount of alum added to lakes and the manner in which it is applied
will have a marked effect on the chemical conditions at the time of treatment.
These conditions, particularly pH, will in turn influence the effectiveness of
P removal and dictate the mechanism by which it occurs. As will be discussed
in a subsequent section ("Dose Determination"), dosages sufficient to reduce
pH to approximately 6.0 will result in maximum P removal (primarily by sorp-
tion to particulate aluminum hydroxide floe) but more importantly, provide for
a maximum addition of aluminum to sediments.
Once deposited, aluminum hydroxide provides continued P control. Kennedy
(1978) monitored P concentrations in experimental enclosures above alum-
treated and untreated sediments and determined that treated sediments were
active in retaining P. Conditions at the sediment/water interface (i.e., low
pH and high P concentrations) will favor the formation of A1P04. Laboratory
elution experiments (Kennedy, 1978) using pre-formed floe provide data
suggesting that such formation occurs at molar ratios between 2 and 4. There-
fore, the P-trapping effectiveness of the floe layer will depend on the amount
of aluminum present, pH, the concentration of P and the rate at which P is
supplied to the floe surface.
III. Criteria for Successful Lake Treatment
No common set of procedures, dose determinations, and matters related to
application has emerged from the several field tests of the P precipitation/
inactivation technique. The following factors must be considered; 1) dose;
2) choice of dry of liquid chemical; 3) depth of application; 4) application
procedures; 5) season; 6) side-effects; and 7) lake types best suited for the
technique. Definite statements about these factors cannot always be given,
due in part to an absence of information, but also because the user will have
to evaluate the individual situation, the objective of the treatment, and
varying social-economic factors of importance. The following is a summary of
current knowledge and experience of those who have worked with the technique.
-------
1. Dose Determination
An objective of all alum treatments is the control of P release from
bottom sedmients. However, in early treatments this was often considered to
be secondary to P removal from the water column (e.g., J. 0. Peterson et aj. ,
1973; S. A. Peterson et alI., 1974). More recently, lake treatments have been
conducted with control of P release from sediments as the stated primary
objective (e.g., Cooke et aj., 1978; Kennedy, 1978). Two different approaches
to dose determination, both related directly to treatment objective, have been
followed. In the first, dosage is optimized for P removal from the water
column with little attention given to the quantity of floe ultimately
deposited on bottom sediments. Dose is determined by jar tests in which
aluminum salts are added until a desired P removal is achieved. This labora-
tory determined dose is then used directly to calculate dose on a lake volume
basis (e.g., J. 0. Peterson et a]_. , 1973; Ellis, 1975). Alternatively, dose
is expressed as an Al/P molar ratio by dividing moles of aluminum added by
moles of P removed (S. A. Peterson et aT_. , 1974) and dosage to the lake is
calculated, based on a knowledge of the P content of the lake volume to be
treated. Usually the dose of aluminum chosen is small enough that drastic
shifts in pH and residual dissolved aluminum (RDA) do not occur.
The second approach to dose determination allows maximum application of
aluminum to bottom sediments and thus emphasizes long-term control of P
recycling. Again, laboratory jar tests are employed but dose is determined by
changes in pH and RDA, with P removal as a secondary consideration (Kennedy,
1978; Cooke et al., 1978). Initially, dissolved aluminum concentrations are
high (Figure 2). As aluminum dose is increased, pH and alkalinity decrease.
In the range pH 7 to 5.5, hydroxide floe is formed and dissolved aluminum
concentrations are minimal. As pH and alkalinity continue to decrease with
increased dose, dissolved aluminum concentrations increases exponentially and
then linearly with dose. A dose producing an acceptable pH and RDA concentra-
tion is then chosen. Using this method, Kennedy (1978) and later Cooke et al.
(1978) defined a "maximum" dose as that dose above which dissolved aluminum
concentration exceeds 50 ug Al/1, a concentration Everhart and Freeman (1973)
had shown to be safe for rainbow trout. Titration of lake water samples from
several depths and thus of varying alkalinity allowed the establishment of a
linear relationship between RDA, alkalinity, and dose, which was then employed
for lake-scale appliations of alum to Dollar and West Twin Lakes, Ohio.
Although a secondary consideration, P removal from the treated water column
exceeded 90%.
It should be noted that this approach is not suitable for lakes of low
alkalinity because the dose would be too low to exert control of P release.
An alternative approach, perhaps first suggested by A. R. Gahler and C. F.
Powers of the Pacific Northwest Laboratory of the USEPA and used by Dominie
(1978), among others, is to added both alum and sodium aluminate at the same
time, calculating how much of each is needed to maintain a pH at which RDA
will not increase. Theoretically as much aluminum as desired or as affordable
could be added under this procedure.
-------
Total Alkalinity (mg CaCOj/L)
s-
0)
-p
m
O)
^.
a
-a
c
ro
3
C
rO
-a
ai
o
in
(0
3
-a
r
10
O)
i-
OJ
to
o
E
o S 5 2 S!
-. o. Q Q q
00
10
c +->i
O O)|
p ai
fD -V
i O
a> o
ce: o
c\j
ai
paA|ossiQ
-------
When aluminate is used alone, pH will rise. This can be avoided by
neutralizing it with HC1 before addition to soft water (Sanville et al. ,
1976), as was done at Cline's Pond, Oregon.
The dose requirements to achieve a desired period of control of P release
are not now known, and this must be a topic for further research. The main
source of P to the water column of many eutrophic lakes during the summer is
internal P release (e.g., Larsen et al. , 1975; Cooke et aj. , 1977; Dominie,
1978). While there are biotic sources of internal P (fish, macrophytes), much
of it probably comes from anaerobic sediments and P must be controlled there
to achieve P limitation of algal populations. Therefore, the Kennedy and
Cooke method or the alum-aluminate balance method should be used, and aluminum
should be applied well in excess of that needed to remove P from the water
column. P removal as the principal objective of the treatment should be
reserved for special cases such as the interception of P from decaying macro-
phytes (Funk et aK , 1977).
Suggested Dose Determination Method
Aluminum solubility is minimal between pH 6 to 8, a range also favorable
for removal of inorganic and particulate P. Therefore, a dose of aluminum
sulfate sufficient to reduce pH to 6.0 is considered an "optimal" dose. RDA,
which is independent of dose at this pH, will remain below toxic levels (i.e.,
< 50 pg Al/1) and the amount of aluminum hydroxide applied to the sediments
will be maximized. Outlined below is a simplified method for determining such
a dose. Non-metric units are used for the alum since it is supplied in pounds
or gallons. Similarly, construction materials are supplied in feet or inches.
This method requires minimal laboratory facilities and can be used to deter-
mine alum doses for lake treatments in which sediment "sealing" for P control
is the primary objective. Prior to initiation of such a lake treatment,
decisions concerning the area(s) and/or depth(s) of the lake to be treated
must be made. Since this dose drastically reduces pH, potential short-term
toxic effects in treated areas should be considered. Many of these can be
avoided if treatment is confined to the hypolimnion.
The P precipitation/inactivation technique should be used in lakes with
moderate to high retention times (several months or longer). Applications
without sufficient diversion of nutrients, such as occurred at Bluff Lake,
Illinois (Kothandaraman e_t aJL , 1978), will be futile. Lakes with low alka-
linity will exhibit excessive pH shifts unless the lake is buffered or a
mixture of alum and sodium aluminate are used.
Procedure:
(1) Obtain water samples from several areas and depths. The number of
samples needed will vary from lake to lake. For lakes exhibiting
wide variation in alkalinity, the number of samples should be
sufficient to span the entire range of alkalinities.
(2) Determine the total alkalinity of replicate subsamples by titration
with 0.02 N H2S04 to a pH 4.5 endpoint.
-------
(3) Determine the dose of aluminum sulfate required to reduce the pH of
replicate samples to pH 6.0. Since granular aluminum sulfate
dissolved slowly, stock solutions of liquid or pre-dissolved alum
should be used. Additions can be made using a burette or graduated
pipette. The concentration of stock solutions should be such that
the maximum dose to each one liter lake water sample can be reached
by an addition of between 5 to 10 ml. Reaction vessels should be
stirred using an overhead stirring motor until pH at final dose
(i.e., pH 6.0), determined by continuously monitoring with a pH
meter, has stabilized (approximately 2 to 5 minutes). Convert the
volume of alum stock solution used to a mass per unit volume dose
(i.e, mg Al/1) for each sample.
(4) Determine the linear relationship between dose and alkalinity using
the information obtained from the above treatments. This can be
accomplished by simple regression analysis or by carefully plotting
dose vs. alkalinity. This relationship can then be used to deter-
mine the dose at any alkalinity within the range tested. The rela-
tionship obtained for a particular lake should not be applied to
other lakes.
(5) As described in subsequent paragraphs (see "Application Proce-
dures"), the lake is divided into convenient treatment areas for
ease and accuracy of dose. The total amount of alum to be added to
each area is a function of the area's volume and the alkalinities of
each stratum (usually each meter). Alkalinity of each one meter
stratum is measured. Based upon the relation between dose and
alkalinity, the maximum dose for each depth interval is calculated
from the maximum dose in mg Al/1 to alum/m3, using a formula weight
of 594.19 (A12(S04)3 14 H20) and a conversion factor of 0.02428 to
change mg Al/1 to pounds (dry) alum/m3.
(6) If liquid alum is to be used, further calculations are necessary to
express the dose in gallons of alum/m3. Details are given in Cooke
et al. (1978). Briefly, commercial alum ranges from 8.0 to 8.5%
A1203, equivalent to 5.16 to 5.57 pounds dry alum per gallon at
60°F. Alum is shipped at temperatures near 100°F and will thus have
lower density. The percent A1203 (at 60°F) is supplied by the
shipper and this is converted to density, expressed as degree Baume'
using Figure 3. A temperature correction is applied against this
Baume1 number, using Figure 4, to account for the decrease in
density. The supplier can estimate the temperature, or it can be
checked at the delivery site. Adjusted Baume' is obtained by
subtracting the correction factor from the 60°F Baume1. Pounds of
alum at shipping temperature are obtained using the adjusted Baume1
and Figure 5.
(7) Maximum dose at each depth interval, calculated earlier as pounds of
alum (dry)/m3, is converted to gallons/m3 by dividing by the value,
in pounds per gallon, obtained from Figure 5. Total dose is finally
-------
o
in
co
o
co
U)
CM
q
O)
q
co
CO
CM
Q.
n3
T3
00
r*.
CT)
ml
-M
OJ
m -i-
Ol i
o:
-------
TEMPERATURE CORRECTIONS
(for 32-36° Be7 Liquors)
2.8 r
2.4
O
i-
o
2.0
1.2
o
LU
Q£
Od
8
0.8
0.4
(Allied Chem. Corp.)
60
80
100 n 120
°F
140
160
Figure 4.
Temperature correction factors for 32-36° Be' liquors (from Cooke
et aj. , 1978 as adapted from Allied Chemical Corp. Technical
Bulletin).
11
-------
obtained for a treatment area by knowing the volume of each stratum,
multiplying the volume by the maximum dose in gallons/m3, and summing.
The sum is applied to the area (e.g., top of hypolimnion to be treated.
6.0
5.0
3,4.0
3.0
2.0
(Allied Chem. Corp.)
20
25
30
35
40
ADJUSTED °Be
Figure 5. Curve to determine pounds of alum/gallon based on adjusted Baume'
(from Cooke et a\_. , 1978 as adapted from Allied Chemical Corp.
Technical Bulletin).
2. Chemical Form
Nearly all treatments since Lake Langsjon (Jernelov, 1970) in 1968 have
been with liquid alum or sodium aluminate, applied with extensive mixing to
surface or hypolimnetic waters (Table 1). Dry alum does not form a floe as
12
-------
p
cz
cu
E
TO
CU
}
C.
o
TO
-P
Q-
U
CU
s_
c
o
+J
TO
U
TO
C
S-
0
jr
Q.
O
_C
Q_
4-
o
S-
TO
3
to
<
-p .p
TO U
S- CU
3 4-
Q 4-
LLJ
C
0
T3 -P
C 03
TO U
CU r^
> O.
r- Q.
-P 0
o cu
to
cu
in
0
TO
TO
u
cu
(_)
cu
-p
TO
TO
-P
1 TO
TO
(J r
r 03
E U
-C I/)
jr:
0)
TO
TO
-P
TO
Q
0
z.
TO
TO
Q
O
z.
TO
-P
TO
O
O
z.
CU
u
TO
to
r
TO
O
E
CU
S-
Q_
E
CU
LL.
O)
CSJ
CSJ
10
cn
X
TO
-P
TO
Q
O
z.
i-
0
S-
01
cu ^
Q£ cn
P -CJ CT>
.C C
tj nj \^
CU r
S- t- C
-o o> cu
S- .C r
S- -P CU
O CD CU
Q z a.
,
-a
cu
o
o.
cu
c
o
z
TO
-p
TO
Q
o
z.
c
1
c
c
13
cu
u
TO
4-
r
cu
S-
a_
E
*£ 'TO
r S-
O) 3
c
r**- TO
^" C3
CO
to
"E
(0 X
O TO
TO
S=. X E
LT) LO O
ro . .
CO CO
II
II II
o
< > N
c.
cu
CU O
cu 5 r^
-^ co cn
TO r
J r.~*^S
E
CZ i >
^0 O O
cn u c
C 0 L-
TO -P ai
I to -3
CSJ
-a
cu
£-
o
a.
cu
cu
c
o
z.
TO
.p
TO
Q
o
z
<-
-p
c
0
E
1
cu
0
TO
4-
3
CO
r_
TO
0
E
a.
CO
\ E
TO
cn
S-
r- TO
r-- 3
1 C
CSJ TO
. i-
ro O
00 cn O ** LO
cn cn cn cn cn
E
in X
CD TO
*"" CJ
X E E <-
TO O -P
.C CSJ O O U
. . . 00 -i-
co i ro csj E
II >>
II II II II i
O X O
< > Nl N Q. Q.
C
cu *~*
T3 CT1
cu r^-
J OT
CU tO r-
TO -
-1 E E
c ro i-
:o .c -P
yi U i
-P O i
:o -P a>
CO
o
cu
S-
cu
o
cu
c
0
z.
\ ro
I- c: co TO
O TO -C
4- E i \
S- in
TO CSJ O >>
-C i 4- U TO
\ -r- -a
o 3 ^"a. c
03 a. TO
«- TO -O TO E
in
s-
TO
01
ro
i
CSJ
cu
u
03
4-
t-
co
,
TO
O
E
CU
s-
°-
Tsl
i TO
-o
kO O"
CSJ _J
o
r-
CO \
E g)
in X
CD TO OO
TO CSJ .
-C 0 E 1 CO
r- E CO I
cn r-» i oo u
to o csi . -r-
ro r .3- u u
II II -r-
II II II -^ E
0 r- X T-
< > NI x < a. -a
TO
cu
cu - c
-^ . o
I O t-
OJ
eu u c -P
O TO -r- CU
.c j i/i a.
in o c '-^
in -r- u o h*
S- C V) CT>
*
T3
0>
.fj
0
0.
Ol
a:
01
c
o
ro
.£=
o
CM
«
(0
01
"~
oi
O)
OJ
o
c a.
^ o
ro
> i Cu
0 0 U
E i- ro
CU -4-> 4-
i- C !-
0 3
Q- u on
_
E T3
3 C
r- ro
rt "D
e o oi
r- ro
a: -a c
Ol 3 g
cr 3
O «- i
r _l ro
,_
\
;
x i1
-c e in P-. u
E i «-
^ o o r-* -P
o cn «tf- ro i u
o CD . . o -r-
CT> ^ CSJ || . E
II r*-- o
II II II -^ C
0 i ZC 0
Nl N >
i- i- C
i O TO
O CJ CO
in
T3
a>
> T3 C
t- C 0
0) ro +J
JD -C C
o ul ro
01 t- Q.
c o
0 O O
2 +J N
ro
4^
rO
O
O
z.
i
c
c
^
cu
u
TO
4-
S-
3
to
,_
TO
O
S-
a.
'E^^E 3
r r r TO
I
i csj ro
r*. r*. r^-
E
m X \
O TO O)
TO i E E
-C
X E E 0
*t in o
. CD i Lfl i
O 00
0 CSJ i ||
II II
II II II -^
O i ZC
«X > Nl N < Q.
C
cz
C -f*^
o cu cn
Q. r- r
CU I- O
-C CU T3
(/I 4~> C
'<- TO TO
uu 3: co
CO
cu
-p
S-
o
CL
cu
cu
c
o
z.
TO
-p
TO
CD.
0
z.
in
s-
ro
cu
CSi
cu
u
TO
4-
S-
3
to
r
TO
o
E
O)
$-
Q.
E
'TO
in t-
c: 03
o
P 3
c
O TO
to i.
C3
(_
P-.
cn
1
X
TO
TO E
-C
LT)
r CSJ
CSJ
II
II
o
r cu cu o
O O l/> U .Q
E i- TO TO
ai -P cu t- ~o
S- C i i- CZ
O CU 3 O
ex u i- in a.
,__
« « TO
cu s-
i- 4-
r-> 3 cn
* CT >,
O '^ r- S-
r 1 CO Q
r-
r-.
r**"
X
TO
E
E
oo
II
II
N O.
c 3:
o .
O CO
ai .
TO -Z.
"«^>
o o i-^
o o cn
£.3****'
!~ t- TO
CD CQ Z
00
13
-------
-a
11
3
C
-P
C
o
u
CU
-Q
TO
1
in
-P
0)
4-
4-
LU
CU
-a
CO
o
CJ
in
4- U)
O QJ
c.
C CU
0 >
-p -p
TO U
S- CU
a 4-
LU
c
0
"D -P
c: TO
TO U
CU i
> Q.
r- Q-
P 0
O CU
-p
CO
a>
U)
o
Q
-o
C
ro
ro
u
E
CU
.C
o
a>
TO
CD
rO
P
r CD
TO
U i
-r- TO
E U
CU -r-
-C in
CL.
QJ
TO
"Cf in
11 CU
> -P
CU 1 S-
m o .a
_a s- cu
o u P
TO S-
o o c:
Z P -r-
TO
4-)
TO
CD
0
Z
V
c
D
11
S-
a.
0
I
cu
u
TO
4-
-a s-
C O- 3
TO CO
4-
i O
TO
> i CU
§o in
s- TO
ai P cu
i- c r
o cu
0- U S-
~ ru
cu c
-o -P o
C TO «-
TO C P
« E E r-
E 3 3 T-
\ . i -D
r TO TO
~ *r-
a: N P
S- TO
CU 1 S-
l/> O .Q
JD t- CU
O U P
TO L.
CU E CU
C >
0 O C
Z -P -
TO
TO
CD
0
Z
c
o
c
c
13
p >>
O -P
(1) .-
4- C
4- -r-
11 i
TO TO
TO C
> TO
0
E :n
CU O.
j_
CL. a
11
TD P
TO C
E E
3 3 0»
co < f (J
\ 4-
-O E S_
TO - 3 3
cn 3 -i- in
cr-a
«d- -r- O P
( _i in TO
CM
LO
TO cn
-C E
co x r--
CSI TO
II
II -^
O r
< M <
4-
O S-
O TO
Z
cu
"D
TO D_
cn
TO
1 -Q
c
TO
D
TO P
-J S- JD
. -r- 3
cn u 3: O.
cm c
O -r- .3
_j 3: ) s^
o
c
o
c
c
ID
TO
O
CSJ
C
o
c
c
13
CU
u
TO
4-
S-
3
CO
TO
0
E
CU
£-
o_
CO
1 (J
O 'I-
u_ cu
c
co E
CO O C
+ a.---
cu >> c
u_ j= o
CTT3 CX
C <-
UD TO t/>
CM
LO
X
TO
TO E
E
UO
CM
II °°
II
o
^ N
o cn «^^s
S- r- CSJ
11 K-'
D TO|
co TO cn
i- pi cn
i- r CU| CU
D TO C
-C U P > TO 3
s: i o
ro
"0 1
cu cu
> -P
s- u s-
CU -r- D
in .c >
.a -P c
o c: -i- in
cu o cu
U JD E- P
C O TO
O C TO S-
z o S -a
TO
p
TO
a
o
z
c
0
c
c
13
cu
u
TO
4-
s_
3
CO
TO
0
E
D
s-
°-
E 3
a
3
ro cr
r-* i
CO
r--
*
X \
TO cn
E E
TO U
-C E "- 0
P CM r
O ID O U r
CM . . *r- CO
*3- ro E 11
II >> II
U II -^
O O X i
iTOTOTO
-C CT.C .C O TO E -C
\ CU (J ^v -r- TJ . \
^- i E I E LO in
Or -DID D c nro>>
CM r C i -C TO -C -TO
tfl-TO TO«y*U E U--O
ai
c
TJ -i-
C L-
cu o
'-' -p
S- C
TO 0
CU E
4-
< o
TO
o
E
D
s-
°-
co S
E 3
i TO
< CU
a u
cn-^ TO
3 4-
LO cr s-
. -r- 3
CM i in
ro
LO
co i
ex "-S.
TO cn
(D E E
TO O
-c f E E cn
0 U
sf CM o o - -
r- . . . -P
ID cn cn u u
H u H ^ e
O r <-
Nl IM
s- .
r- TO ^1
TO -P TO |
C C
TO O -PI
o cut
X) -a c
ceo
TO TO C
i r C
r- i TO
CU 11 -C
3: rs co
LO
-Q
11
s~
D
in
o
cu
c
o
z
TO
TO
0
O
z.
c
o
c
c:
ID
CU
u
TO
4-
S_
3
CO
TO
O
E
CU
s-
O-
TD
C
TO
CO r
E CJ
\, S-
!- M
cno
TO
LO Z
^.
ro
X
TO
TO E
-c e
E
- o cn
CD ID
cn *3-
II
II II
o
fr^o
E 1 -r-
O O 4J
^i- ro . u
r^- «-
CM II E
II 0
II -* c
i I O
Ml IM O)
i- S- 3s
r O O
c_> c_> cu
J£
14
-------
-a
Ol
3
C.
-P
C
0
(J
1 -
O)
XI
re
t
in
-P
u
01
4-
4-
LLJ
0)
T3
to
P
in
O
in
4- in
O o>
c
c 01
0 >
-p -p
re u
S- 01
Q 4-
LU
C
0
TJ -P
c re
re u
01 ^
> Q.
r- Q.
-p <:
u
^01 4-
XI
O Ol
-p
l/>
CL-
IO
0
O
T3
C
re
re
u
E
Ol
CJ
OJ
p
re
TO
P
i re
( Q
re
U r
,- TO
E U
x: in
o >,
x:
Q_
0)
re
i
c 01
O i O
r- XI -P
-P *r- re
re in >> oi
3 o - i re
U Q. (J 0 -P
U !- S- in
O -O X U 3
r C 0 -<- S-
u_ re -P E u
TO
.p
re
a
0
X
S-.
0
p
in
c
re
i
0)
u
re
4
3
re
s.
a.
c
re
E
- E
CQ \ re
-< S-
m re
0) O)r-
w D
O CsJ C
a re
o s_
CSJ CSI C3
^
E
-CO E
cn o
cn x 10
in . ai
II "~" """ II
II II
o z:
i N a.
co
cn
c
-O CD
to V)
.C
£££
O)"O C
C Ol C
>> :* Xt
1 tO LO
("V.
1
-a
o>
-P
t-
0
o.
O)
0)
c
o
z
re
.p
re
Q
o
z.
c
o
c
c
^
a>
u
o re
i. 4-
C 3
o to
U
0
P in
o>
r- -P
re >,
> x:
E o
re
a- E
E 3
r- ro
13
3
CM CT
r- _J
^
i
Ol
X
(0 (0 CO CM
r cn
CO E I 1
CO VO O
* II
II II
II -*
o or
CM -
'""a S
t/l -r-
a; s- -a
^ 0 C
(0 i rtj
_ J U.
c
10 - C
III
o £ n:
CO
(D
o
Q.
Ol
01
c
o
z.
re
-P
re
0
c
o
c
c
ID
o>
o
ro
4-
1/1
TO
°-
in
Q.
O
co E C
E 3 re
r ro c
E £
^j- r- . . re
r ID CSJ S-
II 4->
II II 11 m
0 C
< > Nl N 3
s-**
in
E
- XI
10 3
in Q_
D ro c
c s: 3
O v '
X c
in i o
s in w
01 a> a.
(/> E
s- o
o oi x:
cn
i
a>
S-
0»
in
o
01
c:
o
ro
s- ^j- ro
o x:
4 in \
TO ro >,
x: u TO
SE i
01 C
to x: ro
t/J- U E
+j
c
a>
I/)
0)
O-
o
H-
. - .
O) Ol
in in
ro o
01 -o
O) 4- C
S- 0 0
a *r-
c d. ^e c
ro o E
r O *» * ^~"
fO 0
> ( o) a,
0 0 U >>
E s- ro x:
O) -P 4-
S- C S- TD
O 3 C
CO
E E
NIN 0.
S-
01
in
XJ
o
01
c
o
z.
ro
-P
o
in
S-
TO
QJ
CSJ
1 QJ 1
S- in o
O) re s-
-P 0) U
c i ro o
r- OJ £ U
s- ro
O O)4-
c a_ c s-
re -^3
4- >,CO
o re
re
> c ai
o o "O in
£ -r- OJ
0) -P E -P
S- Q. 0 >>
en s- x:
Q_ U 4- Q.
CO
E E
\ 3
c
oi-a o
f- C
csj 3 a>
. -f- O
0 i (-
]*
O
CO
E -
CD Ol
C3 E
i ro
re ID
x: x cn IJD
E 1 I
OO CSJ l£> CM (J
00 .0 r .-
CSJ O IX» -P
csj r- M u
II II --
II II -^ E
0 ^ . T-
l N Q.> Ol C
P c ro
s- ro
o> -^ -^
XI O C
.,- Q. 3
i I/) LJ_
,
CSJ
C
re >
Ol Ol -P
in u (
TO i TO in
a) o -P £-
s- s_ in a)
U U 3 >
01 -r- S- -
o E u -a
o
c
s- TO «t re
o cn x:
4- in . \
- o in
TO re >»
x: u -TO
^^ -i- Q.T3
ID E *r- I
CM o> 3 c
«j- x: o- TO
*A U Ol E
-P
O)
m
Ol
i-
0
t
O)
in
re
0)
O)
cu
c
4- O
O >-
C
E
0-r-
S- ,
-P 0
C. Q.
O >^
o £
E
CO 3
E -
^ TO
< -o
O) 3
CSJ _l
ID
r*-
-
J i
o TO cn -a
r E «± Ol
i +j
re x E i re
-C E CSJ O)
CSJ iD O (J S-
OO *J- i . +J
^ M u ro
II "~ x:
H II It -^ E
Nl N
ro
o "re
r- x: -P
3 o oi
I
- Ol
-p -P -*
in c o
O) Ol O
OJ
CM
Ol
-P
s_
o
Q.
0)
Ol
£1
0
z.
re
x:
CSJ
c
3
O
c:
Z3
O)
(J
re
4-
S-
3
IO
ro
0
E
01
CL.
E
3
CO i
E re
3
Ol O"
ID I]
ID
r-
>-
CO 01
E E
m X CO
ro CD TO i (£>
x: r- E -CM
cn i
Csj X E E I O
CO ro ro n i
^- CM CM ||
II M
II II II -^
O Z r
^ > Nl N O- -i- in
-^ > c
re s- o
1 O C_3
ro
CM
-o
0)
ZJ
c
c
o
u
15
-------
o
0)
c:
P
c
o
u
OJ
£
l
01
CJ
Ol
4-
UJ
1
01
o
IS>
01
o
o
01
*4 tn
O O>
C
o >
re u
S- 01
Q 4-
LiJ
C
o
o +->
c re
re u
01 F^
> Q-
r- Q.
P >
.c
Q_
_2
re
O)
S-
0
o.
o>
c
0
re
re
Q
o
Z.
Ol
re
0
CD
01
re
CD
CJ
s-
c ca-
re
F 0
re
o o u
E s- fO
o> -P *+
S- C S-
0 3
CL o to
3
E 're
< '5
en cr
00 -J
r-
r-.
00
CO
<0 X \
re o re CT
-C r E E
CM X E E r^ u
r*. CNJ csi CM i . -F-
oo ... oo 4->
F 00 CO || U
II II "-
II II II -* E
o x -F-
< > Ml N <. Q.-D
w r
«r- re
o
C -P
r- 0)
i C
>~4 re
E
OJ - HJ
^ c. s-
re -i~ m
I fO -O
J= C
4- O U)
3 X -P
r- O O
CO U- ^
CSJ
Ol
i.
o
Q.
O)
(U
C
o
z.
re
o
0
z.
c
O)
o>
S-
ci.
o
F
QJ C
01 O
re T-
« nj
O) U
j- 'F- -a
-o i c
c a. a. re
re CL c
*i- re 01 o
r O U !-
re 01 re c
> F F 4- £
i£.?3^
g^*;«.o
0 3 0 >,
Q- u z j-> x:
CO
^ 3
« 1o
CM 3
. CT
CM -r-
i _J
f;
CT1
00
n
(O "^
O X DlLfi
F re E -
re E en
-C X 01
E E un un
to * oo o oo
II II
II II II -*
Q , -f-
Nl N
0 0 QJ 00
^ ^ oi O^
01 01 re F
£ S CU s- "
CSJ
O)
+J
0
CL
01
Ol
C
0
Z£
re
\
re
a
c
re
E
^
C
0) E
O) Q- O
-a
c
nj
E 3
F 'm c
< 0
^ re
^ CT S-
. <- Nl N < O- £
JD
C Q.
F- C
O> 3
O
U ^
01 00
i- re c 01
z S *: E
esj
o
o>
S-
o
a.
0)
O)
o
z.
i
c
E -c
i Ol
*£> >.
re
0 TJ
c
01
in
Ol
CL.
o
H-
O)
Ol
re
o>
o>
i.
a
c o_
re c
4- O
1 O !-
ro c
> F e
00-.-
E t. i
CJ 4^ 0
S- C Q.
O >>
Q- U X
E 3
r re
n3 *xx
re o E 01
x e oo
^f- 00 XI N
O)
c:
o
re
re
o
"Z.
4->
C
a)
Ul
a>
S-
Q_
0
1
OJ
Ol
re
0)
Ol
s_
cx.
c:
4- O
O *r-
c
r- E
O 'i-
p 'o
c a.
o >,
0 X
S-
3
O
CO C
"x O
< E
O)O
un i
csj r--
00
00
E -a
0)
CD +->
o re
^ X "^ 0>
re re O) S-
^ X E _ E 4->
mi E ~or--ure
P-* 1|- CSJ . >r- ^
LO F CSJ - UD 4->
CO r- UT || (J ^~
II II "- .
|| || || ^ E
O f X -i- Cst
Nl M < Q.T3 F
r>
0) f»
Js^ C
re 3
_J O) ^ '
c
^ -r- 00
o re r^-
o s en
O F
re -"- '
01 CL
01 O Ol
J!£'E~
CO +J -r- .
C C E Ul
C 'f- O 10
<. 3 Q E=
oo
CM
16
-------
well and is less effective. In instances where dry alum is the only form
available, it can be mixed in tanks to form a slurry before application (e.g.,
J. 0. Peterson et al.. , 1973).
3. Depth of Application
Depth of application is directly related to the treatment objective. A
surface treatment is required for P removal, whereas control of P release from
sediments usually involves a hypolimnetic application. Most of the pre-1974
projects were surface treatments, but since then hypolimnetic application has
been favored (Table 1) and is recommended since control of P release is the
primary objective in most situations and because application to this zone will
avoid an inadvertent increase in aluminum or pH in productive areas of the
lake. It should be noted however that surface applications are far less
costly than hypolimnetic, and could be as effective in controlling P release
if sufficient chemical is added.
4. Application Procedures
The methods for application of aluminum salts to lake waters have been
remarkably similar over the twelve years in which this technique has been
employed in the United States. The basic application system was reported by
J. 0. Peterson et a]. (1973) for Horseshoe Lake, Wisconsin, the site of the
first treatment in the United States. Dry alum was mixed on board the vessels
by pumping lake water into a slurry tank. A second pump moved the slurry to a
manifold pipe (3 m long, performed with 8 mm holes at 30 cm intervals), which
was suspended behind the craft just below the surface and perpendicular to the
path of travel. Nearly all subsequent applications have used this basic
design, although liquid alum has been used in most cases thus avoiding the
need for the first pump and mixing tank, but requiring an on board holding
tank.
Figure 6 is an illustration of the basic equipment design for the appli-
cation of liquid alum, modeled after the system used at Dollar and West Twin
Lakes, Ohio in 1974 and 1975 (Kennedy, 1978; Cooke et aj. , 1978). The system
consisted of three components: a storage facility, delivery to the barges,
and the barges.
Alum was delivered by tank truck to an onshore above-ground swimming pool
with a capacity of 28.7 m3 (7,600 gallons), or about two truck loads. The
pool was 10-15 m above lake level, giving gravity assisted pumping out to the
loading platform. At poolside, a 12 hp, 2^ inch pump delivered alum through a
hose from the pool to a floating platform in the lake. The piping system was
a series of 20 foot 2 3/8 inch I.D. PVC rigid wall pipes, floated under
anchored 55 gallon drums spaced 10 m apart. The pipe terminated at a
butterfly valve. A 6.1 m section of flexible hose, connected to the outflow
side of the valve, was used to fill barge tank. Communication from this
loading platform to the pump operator at the pool was 2-way radio.
The application barges were designed by Robert H. Kennedy, and consisted
of a series of five 55 gallon drums welded end to end. The buoyancy of one
drum, subtracting its own weight, is about 400 Ibs. These barges had a
17
-------
LU
J
^*
DZ
-l<
< h-
LJJ LU
QO
v>%
LU DC
^ O
< H
_J CO
I
m
03
M-
O
cn
O)
a
03
DO
ID
ai
S-
O)
18
-------
buoyancy of 10,000 or 12,000 Ibs (4,500 to 5,400 kg). The rows of barrels
were lashed with 3/8 inch cable to a steel frame made from 4 inch channel
iron. A seven meter long application manifold made of 2 3/8 inch I.D. PVC
pipe, was suspended below the barge at the depth of application by attaching
it to two steel A-frames, which were bolted fore and aft on either side of the
barge. (The barges were powered by a 35 hp outboard motor with a work prop.)
For travel in shallow water, the A-frames and manifold could be raised to the
surface behind the barge. The manifold pipe was capped at the ends and
drilled with 1/4 inch holes, oriented bottom and back.
Alum was pumped with a m hp pump to the manifold from two 250 gallon
tanks, coupled by their bottom drains with 2h inch galvanized pipe. The pump
was located on the barge deck and fitted with a tee coupling at the intake,
allowing alum from the tanks and lake water to be drawn simultaneously.
Damage to pump fitting was slight when they are periodically flushed with
fresh water. Flow from each source was maintained by valves at 50/50 to
provide initial mixing. Mixing was provided by turbulence from the manifold.
The quantity of alum in the tanks was monitored by a graduated plastic stand-
pipe.
In most surface alum applications, the manifold has trailed behind the
barge and the only mixing has been that provided by the manifold itself. Funk
et al. (1977) used a frontal distribution system to take advantage of the
mixing and spreading action of the barge's pontoons and the fast mixing action
of the propellers of the motor. This modification may provide far more
effective sorptive action of the aluminum hydroxide.
Special application designs have been employed for the simultaneous
addition of sodium aluminate and aluminum sulfate to soft water lakes, and for
the addition of ferric alum and dry aluminum sulfate to small ponds. Dominie
(1978) describes the use of a 6,000 gallon (22.7 m3) three-compartmented tank
truck, mounted on a 40 x 25 ft (1,000 ft2; 9,219 m2) barge. Each compartment
delivered material via pumps to a completely segregated dual diffuser or
manifold, made of 2 inch blank iron pipe which hung below the barge at the 8 m
level (hypolimnetic treatment of Annabessacook Lake, Maine). Sodium aluminate
and aluminum, sulfate had to be pumped separately since contact prior to
release to the water would clog equipment with the precipitate formed. May
(1974) added aluminum sulfate to ponds by mixing dry alum in drums and then
returning the mixture to the pond. Ferric alum was then added by placing
blocks of it in the water, suspended by floats, and allowing it to dissolve.
Blocks were replaced over the year as they dissolved.
Most of the equipment described above was designed to give even dose and
surface coverage, or to add alum to the hypolimnion. These application
methods can be tedious and expensive, particularly when the cost of building
equipment is considered (see later section on cost estimates). An alternative
procedure would be to spray the alum on the lake surface with large diameter
hoses and a large velocity pump. This would result in considerable savings in
manpower and equipment costs. Several serious problems could arise, including
localized lethality due to high aluminum or hydrogen ion concentrations from
uneven distribution of chemical, and poor mixing also leading to locally very
19
-------
heavy amounts of floe. On small lakes or ponds this method of application
could be effective, particularly if boats could be used to create turbulence
and mixing.
An even coverage of lake surface or hypolimnetic area has usually been
accomplished with a grid design. This procedure is well illustrated by the
work of Funk et al. (1977), who divided Liberty Lake, Washington in four
areas, which were then divided into subsections and marked by buoys (Figure
7). Each section could be identified by the barge operator from the color of
the buoys, and from an accompanying enlarged sectional map carried on board.
The map was also marked with the number of bags of alum required for each
section. This application plan has been the basic type since it was first
used at Horseshoe Lake (J. 0. Peterson et aT_. , 1973). For hypolimnetic treat-
ments, the area to be treated is marked at the lake surface, but each
subsection within that area receives a different amount of liquid alum
(assuming a maximum dose as earlier defined is to be applied), depending upon
the subsection's volume and total alkalinity. These data are marked on the
onboard map so that barge operators can apply a full dose (Cooke et al. ,
1978).
5. Optimum Time for Application
The optimum time for application has been debated. If P removal is the
objective then early spring is ideal since, as pointed out by Browman et al.
(1977) and Eisenreich et aj. (1977), most of the P in the water column at thTs
time is inorganic P, a form almost completely removed by the floe. In the
summer months a large fraction of total P is in the particulate and dissolved
organic fractions, and dissolved organic P is efficiently removed with
aluminum. If control of P release is the objective of the treatment then time
of application appears not to be as critical since it is the sediments which
are the target and not P in the water column. Since barges are slow and
unmaneuverable, and may sail on windy days, application should be on calm
days.
6. Toxicity of Aluminum
The toxicity of aluminum to aquatic biota has been reviewed by Burrows
(1977). The relatively sparse data on freshwater organisms allow no generali-
zations about toxicity to taxonomic or habitat-related groups. Freeman and
Everhart (1971) and Everhart and Freeman (1973) carried out one of the most
thorough studies of aluminum toxicity and pointed out that few investigators
have accounted for the complex chemistry of aluminum in water. This was noted
as well by Borruws (1977). The amount of residual dissolved aluminum (RDA) is
pH dependent and some test waters could receive large amounts of aluminum
before RDA became sufficiently high to be toxic. Thus reports of safe dose
limits, unless actual RDA is measured in the test and alkalinity-pH reported,
are of limited value since they report aluminum added, not dissolved aluminum
in the water. There is very little direct laboratory or field evidence about
the short or long-term effects of aluminum on aquatic biota or aquatic commun-
ities.
20
-------
Bl
Al
FLAG COLOR
O GREEN
A RED
0 YELLOW
n WHITE
AIO
Figure 7. Division and delineation of treatment sections for aluminum sulfate
treatment of Liberty Lake, Washington (from Funk and Gibbons,
1978).
21
-------
There are at least three laboratory studies in which the actual chemistry
of this element in water was considered. Biesinger and Christensen (1972)
reported that Daphm'a magna had a 16% reproductive impairment at 320 |jg Al/1.
Freeman and Everhart (1971), Everhart and Freeman (1973), and Freeman (1973)
used a constant flow bioassay to test toxicity to rainbow trout. A concentra-
tion of 5,200 ug Al/1, whether at pH 9.0 where it is totally soluble or at pH
7.0 where it is nearly insoluble, seriously disturbed trout if present longer
than 6 weeks. At 520 ug Al/1, symptoms appeared after a few weeks exposure,
suggesting that the usual short-term bioassay might have missed this response
entirely. At 52 ug Al/1, there was no obvious effect on growth or behavior,
leading Kennedy (1978) and Cooke et aj. (1978) to adopt this value as the
upper RDA limit for lake treatment. S. A. Peterson et al_. (1974, 1976), using
static bioassays, reported that Chinook salmon survived an RDA of about 20 ug
Al/1. Higher concentrations were not tested. Daphm'a magna did not reach a
96 hr TL with concentrations up to 80 ug Al/1 RDA, but percent survival was
reduced fo about 60%.
Several investigators have reported an apparent absence of negative
effects on fish (Kennedy and Cooke, 1974; Bandow, 1974; Sanville et a]_. , 1976)
or benthic invertebrates (Narf, 1978) after full-scale lake treatments.
Narf's report is of particular importance since it represents monitoring of
Horseshoe, Long, Pickerel, and Snake lakes, Wisconsin, the earliest full-scale
aluminum treatments in the United States (Table 1). Cooke and Myers (unpub.
mss.), and Moffett (1979) found a significant decline in the Shannon H1
diversity (Shannon and Weaver, 1959) of planktonic microcrustacea in West Twin
Lake, Ohio, after a hypolimnetic aluminum sulfate treatment, when post-
treatment samples (1976 and 1978) were compared to pre-treatment (1969)
samples or to the untreated downstream lake. The diversity decline was
apparently not due to aluminum toxicity in the water column (see also West
Twin case history) but to changes in type of algae cells or to low pH and/or
high RDA in interstitial waters where ephippia, or other resting stages might
be found. Their results could not have been predicted from the standard
laboratory bioassay, pointing out that future toxicity studies should be
directed toward the actual level of biological organization to which the
aluminum treatment is directed.
Aluminum toxicity does not appear to be a significant problem, as long as
pH is controlled and/or RDA is not allowed to reach levels in the area of 50
ug Al/1. The long-term effects seem to be small, at least to most benthic
invertebrates which live directly in the aluminum-enriched sediments. Effects
upon the community level organization are essentially unknown but the report
of Cooke and Myers, and Moffett (1979) suggest that the response of this level
must be further investigated. In areas where lakes have low alkalinity and
acid rainfall is significant, lowering of lake pH could occur years after an
aluminum treatment with a sudden increase in RDA and possible toxic effects to
lake biota.
22
-------
IV. Case Histories of Phosphorus Precipitatiorrlnactivation
A. Introduction
There have been at least 28 reported uses of the phosphorus inactivation/
precipitation technique, all of them since 1962, and nearly all with aluminum.
Many of these projects have little or no documentation of long-term effects.
All are listed in Table 1, along with available data on lake characteristics,
dose, method of application, cost and side-effects. The reader is urged to
consult reports or one or more of the persons listed for detailed guidance for
a particular lake type or problem.
Seven different treatments have been selected for more detailed review
because they are representative of a particular approach, or have a long
period of monitoring of the effects, or because they seem to best illustrate
the strengths and shortcomings of the technique. In general there have been
few treatments for which published documentation is available and even fewer
for which sufficient post-treatment monitoring has occurred.
B. Case Histories
1. Horseshoe Lake, Wisconsin
The first reported phosphorus precipitation in the United States was at
Horseshoe Lake, Wisconsin (J. 0. Peterson et al. , 1973). Their results are
important because they describe the mechanics of a surface application, have
been monitored for effects longer than any other study, and because this
project illustrates the need for adequate diversion of nutrient income and
application of sufficient aluminum to control P release from the sediment.
Pertinent limnological data are listed in Table 2. The lake had experi-
enced algal blooms, dissolved oxygen depletions and fish kills prior to treat-
ment. The authors attribute the high nutrient levels to agricultural and
natural drainage, and to a cheese-butter factory which discharged wastes to
the lake before closing in 1965.
Table 2. Physical-Chemical Data for Horseshoe Lake (from J. 0. Peterson et
an. , 1973)
Location: Manitowac County, Wisconsin
Watershed Area: 700 ha
Lake Area: 8.9 ha
Maximum Depth: 16.7 m
Mean Depth: 4.0 m
Volume: 3.6 x 105 m3
Water Residence Line: 0.7 yrs
Thermal History: dimictic
pH Range: 6.8-8.9
Total Alkalinity Range: 218-278 mg/1 as CaC03
23
-------
The goals of the Horseshoe Lake project were to demonstrate that alum
could be applied safety and effectively to a lake, to document the effects of
the alum on water chemistry and algae, and to determine the costs. It was
expected, based upon the Swedish experience (Jernelov, 1970), that there would
be at least a short-term decreased in total P and improved oxygen conditions.
The object of the treatment was to remove P from the water column. The
dose was determined by jar tests and it was found that 71% removal of total P
could be achieved with a dose of 225 g A12(S04)3 14 H:20/m3 (16.7 g Al/m3).
While a dose of 18 g Al/m3 was proposed, 10.2 metric: tons were actually
applied to the lake surface (2.1 g Al/m3), as liquid alum..
Three applicator vessels of different sizes were used to distribute the
alum, and these are described in detail in their report.
The lake was divided into 9 plots, each 1 hectare, and a predetermined
amount of alum was added to each, just below water surface, on 20 May 1970.
The P content in 1971 and 1972 was much lower than 1966, and significant
decreases were observed in both the epilimnion and hypolimnion. While the
normal summer increase in P did not occur in 1970, there was no decrease below
pre-treatment levels until after fall circulation, at which time there was a
substantial decrease. A similar observation has been made since then in the
cases of the hypolimnetic treatments at Medical Lake and Annabessacook Lake
(Gasperino, 1978; Dominie, 1978, respectively).
Hypolimnetic P in Horseshoe Lake has increased slightly every year since
treatment (up to 1978) although it has never reached the; levels found before
the applications (Born, 1979). Their treatment gave about 8 years of control
of hypolimnetic P concentration.
Secchi disc transparency increased, few algal blooms occurred in 1970,
and there were no fish kills through 1972. Lake shore residents were reported
to be pleased. No data have been reported about algae since 1970. Nitrogen
fractions were higher than anticipated after treatment, and dissolved oxygen
conditions were greatly improved. Surveys through 1978 have indicated no
detrimental side effects to benthic insect larvae.
2. Dollar Lake-West Twin Lake, Ohio
The alum treatment of Dollar Lake in July 1974 and West Twin in July,
1975 differed from earlier applications in these respects: 1) application was
hypolimnetic; 2) the objective was control of P release from anaerobic sedi-
ments; 3) dose was not based on P removal in jar tests but upon the maximum
amount the hypolimnion could receive before a pre-defined level of RDA was
approached; and 4) the West Twin treatment was compared to changes in a
similar, adjacent and untreated lake.
The Twin Lakes and Dollar Lake are located in a small residential water-
shed in northeastern Ohio. Pertinent features of the watershed and lakes are
listed in Table 3 and thorough descriptions are given in Kennedy (1978) and
Cooke et al. (1978).
24
-------
Table 3. Physical-Chemical Data for Dollar-Twin Lakes
Location:
Watershed Area:
Watershed Area:
Lake Area:
Maximum Depth (m):
Mean Depth (m):
Volume (m3):
Water Residence Time:
Thermal History:
pH Range:
Alkalinity Range (mg/1 as as CaC03):
West Twin
34.02
11.50
4.34
14.99 10s
1.28
6.9-8.5
East Twin
rortage Lo, unio
.__ OCR U-,
OCJD na
26.88
12.00
5.03
13.50 105
0.58
all are dimictic
6.9-8.3
Dollar
2.22
7.50
3.89
1.86 105
6.7-8.6
Data From: Cooke e_t aJL (1978), Kennedy (1978)
Application Area (ha): 16 -- 1.39
The lakes became very eutrophic in the late 1960s due to septic tank
drainage and urban runoff, and had a Carlson (1977) Tropic State Index about
65 (hyper-eutrophic). In 1971-72, septic effluent was diverted but recovery
was slow, due in part to internal release of P (Cooke e_t a^L , 1977), and the
lakes continued to exhibit blooms of blue-green algae and extensive macrophyte
areas.
Kennedy and Cooke (1974) suggested that alum dose could be based upon
alkalinity and the criterion that alum could be added to lake waters until
there is sufficient pH change to bring about a residual dissolved aluminum
concentration of 50 ug Al/1, a level shown to be safe for trout (Everhart and
Freeman, 1971). A step-by-step evaluation of this concept followed, including
a test of long term effectiveness in enclosures in the lake (Kennedy, 1978),
and toxicity tests to the Northern Fathead Minnow (Wilbur, 1974). A pilot
application to Dollar Lake followed this work.
Dollar Lake received 10.2 tons of alum (2,896 gallons or 11 m3), 10% of
which was added to the surface, on 19 July 1974. Details of dose calculation
and application are given in Kennedy (1978), and described in Section III of
this report. Application of aluminum sulfate to West Twin was made to the 5
meter contour, an area of 16.04 ha. A total of 36,919 gallons (140 m3) were
applied, based upon hypolimnion alkalinity and volume, in 3 days (29-31 August
1975). Details are given in Cooke et a_L (1978).
The alum applications had an immediate and dramatic effect on the total
phosphorus content of the lakes, illustrated in Figure 8 for West and East
Twin (the untreated reference lake). West Twin has continued to have low P
content and improved water transparency through summer, 1980. Dollar Lake
responded similarly (Figure 9), and hypolimnetic concentration in it has
remained low 4 years after treatment, although not as low in 1978 as in 1976
(Cooke, 1979). The effectiveness of the floe may be beginning to decline in
Dollar Lake.
25
-------
KILOGRAMS- P
LU
<
J
Z
. ?.. 1
.. s
* Uu
* -
Q
. * . Z
* C
V.
* . . w <
: . r- _;
2 -
*
:' *'* ^
'' LL.
^
o o o a
in o u>
O O
12 e § o
....,., c
z
Q
^ c/i
'.* -^
'2^
.' 5
^
^
LL
>
VIVO ON-ZZ61
Q
z
: P
CO
* ~?
z /" ~>
2 ..* 5
H" ^ ^
< S -
O 5
IJ * . u.
0- I -i
0. ^
< -. c
5 > . ^
1 .' P
< w
^ . ."*">'*
"" . S "
* 2
. 5
* <
5
LL
Q
z
* ^N
c/
. " « S <-
. o> ->
'." 1
.» 1
. ' 5
> LL
LU -5
: ^ z
.
z c^
^ « <
* . ( *
* S" *" *" 5
%. . '* 2 «i
*. " LL"
..,*, , , , . ~>
O O O O
in O 10
at
-^
o
o
o
o
en
LO
a>
Q.
E
ro
LO
01
re
i
c
5
I/)
0)
-a
c
(0
*->
l/l
(0
Q-
CO
rO
S-
D5
O
C
QJ
O
U
S-
o
to CT1
o r^-
00
OJ
en
26
-------
.
*
.
VIVO ON - LLGl
. '
' .
\
%
*
I
«.
z
2
^x ^_ **
L ^~ *
* y
~
K Q.
< Q.
d *'
0 5 "|
Q D V
-i /
«r /
I1'
.
.
.
*
.
«
";.:'
. .' .
.
.
Q
\
o
CO
<
->%.
-,«
<
-)
Q
O
CO
<
->£
-,0)
>
<
2
U-
5
Q
Z
0
CO
<
"O 1^
-, 0)
1
2
u.
-1
Q
Z
0
CO
^
->*
-J °>
5^
<
^
u.
->
Q
Z
O
<»
>
>
2
o « o "> o «
«««»-'-
/ N
CD
CD
i
01
o
o
o
0
J-
^~s
CD
c
r
in
a>
'a.
ro
in
o
.
ai
ro
i
S-
m
o
a
<4_
O
Q_
in
E
s_
D5
0
|
^
^^
i ^
C
OJ
4-3
O
0
in
S-
o
a.
in
O
^.
Q_
CD
ai
s_
3
r
1 i
27
-------
Internal P release was not completely controlled in West Twin (Table 4).
Cooke et aj. (1978), Cook and Kennedy (1977, 1978) and Cooke (1979) believe
that unknown but significant internal P sources were in the littoral zone,
since they had earlier (Kennedy and Cooke, 1974; Kennedy, 1978) demonstrated
that the aluminum hydroxide floe was effective in controlling most of the P
release from the treated sediments. This suggests that alum treatment will be
more effective in lakes without a large littoral area or that the littoral
area will have to be treated in conjunction with the hypolimnion. Unfortun-
ately, a floe over littoral sediments could be unstable in wave-swept areas.
Table 4. Net External, Internal and Total
(mg P/m2/day) During Summer
Phosphorus Income to the Twin Lakes
West Twin
Year
1972
1973
1974
1975
1976
1978
External
0.200
-0.127
0.435
0.637
0.334
Internal
2.007
2.668
0.831
Calculation -- Al
0.689
0.163
Total
2.207
2.541
1 . 226
urn Application
1 . 326
0.497
Days
126
126
116
77
101
East Twin
Year
1972
1973
1974
1975
1976
1978
External
-0.054
0.207
0.336
__ M-
NO
0.663
0.161
Internal
2.917
2.804
0.750
Calculation -- Al
1.022
1.351
Total
2.863
3.011
1 . 086
urn Appl ication
1 . 685
1.512
Days
98
112
125
72
101
Note: 1.
2.
Negative external income due to negative groundwater income.
Days of summer defined as date of spring low phosphorus content to
summer high phosphorus content. This interval corresponds to date
of thermal stratification in spring and to beginning of destratifi-
cation in fall.
(from Cooke, 1978).
Blue-green dominance was reduced in West Twin to about 80%, but East
Twin, the downstream untreated lake, exhibited a much greater reduction (to
30% by 1978) in blue-green dominance. In both lakes total cell volume
declined dramatically after the alum treatment, and the importance of greens,
diatoms, and dinoflagellates increased. Very similar changes were noted in
Dollar Lake (Kennedy, 1978; Cooke, 1979).
28
-------
Changes in the Carlson Trophic State Index (Carlson, 1977), an index
based on algal biomass, are listed in Table 5. The index is scaled so that
each change of 10 units (e.g., 40 to 50) represents a doubling (or halving) of
algal biomass. Total phosphorus changes are the best illustrators of change
in trophic state due to the heavy use of herbicides and algicides in the Twin
Lakes in the early 1970s, which temporarily increased transparency and
decreased chlorophyll. Thus, Dollar Lake has had a 3-fold decrease in algal
biomass and West Twin a 2.4-fold decrease 4 and 3 years after the aluminum
sulfate treatments, respectively, and the lakes are now mesotrophic (a TSI
value between 41 and 51 is usually found in mesotrophic lakes). The down-
stream untreated lake, East Twin, changed little from 1975 to 1976 but is now
also mesotrophic due to the income of nutrient-poor water from West Twin.
An important side-effect was noted in 1979. After three years of
increased water clarity the biomass and outward distibution of macrophytes
from the shore has increased.
A significant increase in the number of planktonic microcrustacea species
and Shannon H1 (Shannon and Weaver, 1949) diversity occurred in West Twin, and
these measures of community structure have remained significantly lower than
East Twin through 1978 (Moffett, 1979). The dominant species also shifted
from Cladocera to Copepoda. Since no residual dissolved aluminum was observed
in the water column after treatment, it is speculated that these changes in
community structure may be due to low pH or high aluminum in sediments where
resting stages of these invertebrates may be found, or to the shift in the
phytoplankton from blue-green dominance to diatoms and green algae, a change
which may favor herbivorous copepods according to McNaught (1975). A higher
copepod-cladoceran ratio is found in less eutrophic water (Gannon and
Stemberger, 1978). The long term implications of the shift are not known. To
date the treatment and these changes in pelagic zooplankton appear not to have
been detrimental to sport fishing.
3. Medical Lake, Washington
Medical Lake, near Spokane, Washington (see Table 6 for features), main-
tained eutrophic conditions (algal mats, dissolved oxygen depletion) through
internal recycling of nutrients from decomposing algae and from anaerobic
sediments. It receives no surface water but presumably became enriched from
groundwater flows which were contaminated by septic fields. This lake
appeared to be ideal for an aluminum sulfate treatment since it is a seepage
lake with a long water residence time and small littoral area. However, the
alkalinity of the water is very high (about 750 mg/1 as CaC03) and jar tests
revealed that large doses would be required for P removal from the water
column. It was determined from laboratory tests that multiple treatments and
a combination of surface and hypolimnetic applications could be more effective
than a single surface application for removal of phosphorus. During the 41
day period from 3 August to 13 Sej cember 1977, 936 metric tons of liquid alum
(12.2 g Al/m3) were added to Medical Lake for the purpose of P removal. There
were 7 entire treatments of the hypolimnion and 4 at the surface. The appli-
cation equipment was essentially the same as described for other lake treat-
ments (Gasperino and Solter, 1978).
29
-------
/"N
-P
C
cu
E
cu
3
I/)
rO
CU
e
cu
u
rO
(4
S_
i/)
c
0
s-
n-
T3
CU
4_>
rO
f^
3
U
'rO
O
v^x
>^
CU
TJ
C
1 (
Ol
-p
rO
-P
OO
U
r-
-C
O.
0
S-
1
p
f~
o
(/)
1
ro
CJ
/,
S-
cu
E
CU
-p
D.
CU
i,,
s:
s»^>
c
rO
CU
2:
.
in
cu
r
O
ro
t
0)
3
i
ro
>
i i
oo
1
S-
rO
^~
r~-
O
a
i/>
cu
3
^
ro
^5»
1 I
(/)
j__
C
r
3f
r""
-P
U)
rO
1 1 1
1/5
CU
3
'itj
^>
f 1
00
l_
^^
c
r
3T
1
p
cu
->
-^
^
H-
^
r~
O
t/5
C
rO
S-
1
^
1
r
jC
O
I/)
C
fo
S-
*~
rt
^~
f~
o
c
rO
l-
1
K
i c\j in en i LD m
fo CM ID r-~- cr> m en
csj csj oo en r-~ r*^
1 C^ Q) ^^* fit*) f in f^) C\J
o o o * o f~ Ln LO un ^ ^ Ln
z z z c
cu
SX
K
CU * *
r^ooiDcnr^-r^r^- o
oo o r~~ r~ ro 00 ro
ID in ^- ^" U3 rf
oo ^"~ oo 0*1 ^^ ^sj r^ o^ LO r"
(^O O") ^J" OO f^ ^-.P '^J O^ O*J f~*>
(Y) QQ l^Q QQ f^. t^p f^^ tp p ^
LO 1^) LO LO LO LO ^J" LO LO
03
4_) ^^- p**. I^Q ^- CI5 CD CD LO LO 00
03 ooc\JO^csjoo<3-,>,o ro
t^*i C^D tf^ oo ^^ ^^J cr^ r^"» f*" ^~
O ^^ ijO '^^ LO ^£3 LO ^^ LO LO
z
r~~ ^^ 00 ^O ^^ ^O ^^ ^~" ^O 00 00
If) ^J* f*^, Q^ ^^ Q^ ^J* LO 00 ^^" ^^"
r- ocsjooor-i inooo o
ininLn^ininLnsj-ir)^-
mi oocooooocsjooro «i-
r^csjrcnincsj«i-ooo CM
LO I*O tO LO LO LO ^" tD ^"
ro
.fj C3 ^J" r**- tO 00 00 O^ ^^ ^i" C3
ro ' cnr^LOLor^cooooo o
oOP^csjcDOOcncncncn csj
O tD LO LO tD LO ^t" ^J1 LO ^t1
"Z.
CD CD oo r«~ oo ^* r*** co ^f to r^
g^^^^^^^g^ °
00
r***
en
to ^~
to
C ro .
i- E t/>
^00 O >
r**1* p"** QJ *^~
cr> en to QQ ^~
r r~- O3 r*"
1 1 CU i 1
OOCni Csjro^t-LOlIDOOi IX) I- rOr
to tjD r^* r^- f^^ r*^ r*^ r^ r^^ r^ r^> o 01 r^«
en en en en en en en en en en en cu *~~ en
>>rd
c: ro 00
o -a r^
en
ro 00 i
-p f^*>
ro en "
'Or- CU
i -^
co in o
r--- r^ o
O1 O
r r
£
II II 0
i.
*c * *-
* s^
30
-------
Table 6. Physical-Chemical Data for Medical Lake"
Location: 25 km Southwest of Spokane, Washington
Watershed Area: Closed Basin
Lake Area: 64 ha
Maximum Depth: 18 m
Mean Depth: 10 m
Volume: 6.07 x 106 m3
Water Residence Line: Unknown
Thermal History: dimictic
pH Range: 8.5-9.5
Alkalinity (mg/1 as CaC03): 750
* From Gasperino and Soltero (1978, 1979).
Total and ortho-P in the water column were lowered after fall turnover in
1977, a response similar to Horseshoe Lake (J. 0. Peterson et al_. , 1973) and
Annabessacook Lake (Dominie, 1978). The cause of this delayed response is not
known. Concentrations of P fractions have remained low to date.
Before treatment, dissolved oxygen conditions in the lake were extremely
poor and the lake did not support a fishery. After treatment, conditions
improved greatly with no anoxic periods during ice cover in 1978 and 1979, and
summer 1978 oxygen depletion occurred only a depths below 10 m. The lake now
supports a rainbow trout fishery. The growth rate and condition factor of
fish in Medical Lake exceeds that of most surrounding lakes (Gasperino and
Soltero, pers. comm.).
Associated with improved oxygen conditions was an increase in trans-
parency, decrease in chlorophyll, and a shift in the phytoplankton community
from an assemblage of species dominated by blue-greens to a more diverse one
with greens and flagellates. The phytoplankton became P-limited, as indicated
by an algal bioassay. No negative impacts on the biota were observed despite
an increase in dissolved aluminum to 700 pg Al/1 during treatment. Current
aluminum levels are 30-50 pg Al/1 and are equal to the pretreatment concentra-
tions (Gasperino and Soltero, pers. comm.).
The Medical Lake experience indicates that applicators may extend the
period of treatment over many days and this could constitute a significant
savings in labor costs.
4. Annabessacook Lake, Maine
The aluminum application to Annabessacook Lake illustrates the use of
phosphorus inactivation/precipitation to control algal blooms in soft water.
The nutrient budget methods of Cooke et aj. (1977) was used by Dominie (1978)
to demonstrate that 85% of summer P increase in this soft water (alkalinity =
20 mg/1 as CaC03) lake was due to internal P release, presumably from sedi-
ments. The objective of the treatment was to control this P source. In order
to maintain pH at near normal levels and thereby also prevent the appearance
31
-------
of dissolved aluminum when aluminum salts were added, a mixture of aluminum
sulfate and sodium aluminate, in a ratio of 1:1.6 was added to the hypolimnion
of the lake in August, 1978. This ratio was determined empirically from jar
tests to give good removal and little pH shift. This technique for soft
waters was apparently first utilized at Snake and Long Lakes, Wisconsin in
1972 (R. Narf and T. Wirth, pers. comm.).
Annabessacook Lake has a surface area of 575 hectare and a hypolimnion
area of 130 ha. The entire hypolimnetic area was treated in August, 1978,
with the 8-10 m contour receiving a dose of 25 g Al/m3, and the 10 m contour
and below a dose of 35 g Al/m3, using application procedures similar to
earlier hypolimnetic treatments (Dominie, 1978; Dominie, pers. comm.). Had
aluminum sulfate been used alone, the total dosage possible before dissolved
aluminum approached the 50 ug Al/1 level set by Cooke et a!. (1978) would have
been only about 4 g Al/m3 (using the empirical dose relationships for West
Twin and assuming that Annabessacook waters had similar properties to it), a
dose which would not exert long-term control of P release.
There was little immediate P reduction at Annabessacook after the appli-
cations, as was found at Horseshoe and Medical Lakes, and it was not until
September that a large decline in the lake's P content was observed. The
treatment appeared to be successful in that pH shifts were minimal, no
dissolved aluminum could be detected, and no adverse side effects were
observed (Dominie, pers. comm.).
5. Braidwood Lagoons, New South Wales
All previous descriptions of the P inactivation/precipitation technique
have been about lakes. May (1974) has reported the successful use of liquid
alum and blocks of ferric alum (Fe2(S)4)3 24 H20) in a shallow (max. depth 1
m) pond which had experienced severe blooms of toxic blue-green algae. The
objective was to stop P release from anaerobic sediments by applying the
chemicals to the pond surface, first in mid-winter (July 1971) and by adding
the ferric alum at quarterly intervals as groups of suspended blocks (508 kg)
which were replaced as they dissolved. A total of 1,067 kg of alum (10.60 g
Al/m3) and 2,540 kg ferric alum (31 g Fe/m3) were added.
The pond's pH ranged from 5.0-5.5, phosphorus concentration remained low,
and blooms of Anacystis cyanea and Anabaena circinalis did not occur during
the subsequent summer. Monitoring ended in 1972.
6. Cline's Pond, Oregon
Lanthanum rare earth chlorides and zirconium were investigated by S. A.
Peterson et aJL (1974, 1976) for their effectiveness in P removal and for
their toxicity. While lanthanum proved to be unacceptably toxic, zirconium,
in doses ranging from 0.5 to 10 mg Zr/1 in a static bioassay, was non-toxic to
Coho salmon for up to 10 days of exposure but did exert small chronic
mortality and impairment of reproduction in Daphnia magna. This lead the
authors to conclude that while zirconium was as much as 3.5 times more effic-
ient than aluminum in removing P from pond water, further experiments were
needed before zironcium could be used by the public.
32
-------
Powers et al. (1975) continued the work of S. A. Peterson et aj. by a
pilot scale test of the effectiveness of zirconium chloride addition in
removing P. On 26-27 March 1974, a dose of 5 g Zr/m3, buffered with NaOH to
counteract the pronounced pH drop produced by hydrolysis of ZrCl4, was added
to half of Cline's Pond (0.5 ha), after dividing the pond with a polypropylene
curtain.
A bloom of Anabaena circinalis developed on the control side in July, but
the treated side was clear, with 5 times less chlorophyll a and 2.5 times less
P. Their data suggest that Zr inhibited recycling of P from the sediments and
that the treatment was not toxic to algae but made them P-limited. Kumar and
Rai (1978), in laboratory studies with Chlorella, also noted that inhibition
of growth was due to P-limitation and not toxicity of ZrOCl4. Macrophytes
spread over one-third of the experimental area of Cline's Pond. This was
attributed to the greatly increased water clarity.
The Cline's Pond and the Braidwood Lagoon experiments illustrate the
efficacy of P inactivation/precipitation in smaller systems, and the use of
inactivants other than aluminum. At present, the cost of zirconium is high
and it is not generally available.
7. Wahnbach Reservoir, Germany
Bernhardt et a!. (1971) and Bernhardt (1978, personal communication)
describe a special use of phosphorus precipitation to improve the quality of
water entering the drinking water treatment plant at Wahnbach Reservoir.
After impoundment in 1957, the reservoir was very eutrophic, with blooms of
Oscillatoria rubescens, and treatment for drinking purposes became costly and
increasingly difficult. Since the primary sources of nutrients were non-
point, it was decided to allow water to flow into a pre-reservoir, precipitate
P, flocculate and filter this water, and then allow it to flow into the main
reservoir. The treatment plant was completed in 1977 and treats income to the
pre-reservoir up to 8 mVsec without flooding. Pumps, operating at+18,000
m3/hr, lift water out of the reservoir and P is precipitated with Fe 3, and
the iron phosphate is then removed by flocculation with alum. Negatively
charged particles and colloids are destabilized, the floe agglomerated,
treated with a polyelectrolyte, and then filtered. Total P concentrations in
the emerging water averages 4 pg P/l, a 95% reduction. Plankton development
in the reseroir has become much less.
V. Costs for Phosphorus Inactivation/Precipitation
Funk and Gibbons (1979) have summarized the costs for materials,
supplies, transportation, construction, and application for four lakes. Their
summary indicates high cost variability among treatments, in part due to
changing prices of labor and chemicals.
The most expensive portion of phosphorus precipitation/inactivation may
be labor. Labor costs (expressed as man-days or one person working 8 hours)
for several aluminum treatments are summarized in Table 7 along with dose and
area tested. Few relationships appear from these data, due primarily to the
low doses and surface treatments of Horseshoe and Liberty Lakes and the
33
-------
Welland Canal. If these three are omitted, there appears to be a curvilinear
relationship between area treated and man-days. There was a difference of
only 6 man-days labor for application between the surface-hypo!imnetic treat-
ment of Medical. Lake (64 ha treated) and the hypolimnetic treatment of
Annabessacook (121 ha treated), but a difference of 56 man-days between the 16
ha hypolimnetic treatment of West Twin and Medical Lake.
More data are needed before a quantitative relationship between treatment
area and man-days can .be stated. For planning purposes, persons considering a
hypolimnetic treatment designed to control P release will find that labor
costs for application will be higher for small treatment areas. The cost of
construction could be offset because application vehicles can be used for
subsequent treatments of other lakes or modified for other purposes.
Table 7. Dose, Area Treated, and Man-Days for Application of Aluminum Salts
Lake
1.
2.
3.
4.
5.
6.
1.
Horseshoe
Well and Canal
Dollar
West Twin
Liberty
Medical
Annabessacook
Dose
g Al/m3
2.
2.
20.
26.
0.
12.
1
5
9
1
5
2
-30**
Area Treated
(ha)
8.
74.
1.
16.
277.
64.
121.
9
0
39
0
0
0
4
Man-Days for* Kilograms
Application Al
11.
100.
6.
73.
36.
130.
136.
8
0
0
8
0
0
0
15
1
8
95
75
11
946
,490
,797
,649
,300
,853
,654
Man-Days/
ha
1.
1.
4.
4.
0.
2.
1.
33
35
3
61
13
03
12
* 1 person working 8 hours (note: most treatments were done with 12-14 hours
working days, increasing labor costs but also increasing efficiency).
** Different doses were given to shallow and deep hypolirnnetic waters.
VI. Summary
1. The purpose of the phosphorus inactivation/precipitation technique
for lake rehabilitation is to lower the phosphorus concentration in the water
column. The expected result is an increase in transparency, fewer algae, and
increased recreational potential. Aluminum has been the inactivant/precipi-
tant of choice to date.
2. The technique is effective in removing inorganic (precipitation) and
particulate (entrapment) P from the water column, thereby bringing about an
immediate decrease in algal blooms and an increase in transparency. If
sufficient aluminum hydroxide floe is formed over the sediments, P release
during anaerobic conditions is retarded and long-term control of P concentra-
tion in the water columns may be achieved (at least 5 years) if there has been
nutrient diversion. Treated lakes have promptly changed from a "eutrophic" to
34
-------
a "mesotrophic" classification and have remained in that state for at least 5
years after a sediment treatment to control P release. Ponds can be effect-
ively treated also.
3. Some lake treatments with aluminum have been ineffective or have had
short-term effectiveness only. Failures can usually be traced to incomplete
nutrient diversion and to insufficient dose.
Lakes with large littoral areas may continue to have higher P concentra-
tion than expected after an open water application due to P generation from
littoral sediments by macrophytes, animals, and groundwater. This apparently
cannot be halted by an aluminum application to littoral sediments, except
perhaps for internal P "loading" brought about by flow of groundwater through
P-rich littoral sediments.
Lake treatment will not be effective if nutrient diversion is insuffic-
ient, or if there is rapid flow-through of water. In both of these instances,
as well as with "internal P loading," a water-P budget is needed to assess
their significance before application.
In some cases the dose of aluminum has been adequate to remove P from the
water column but too low to affect lasting control of P release from sediment.
The amount of aluminum hydroxide needed to achieve such control is unknown but
it will usually be in excess of the amount needed for P removal.
4. Persons contemplating an aluminum application should address these
pertinent problems:
a. How much aluminum should be added? An answer to this is not
now clear but it appears that a dose in excess of an A1:P molar ratio (but not
to exceed toxic levels of 50 ug Al/1), as determined from jar tests is
required for control of P release. Monitoring data are scarce, and nearly all
high-dose lakes have been treated in 1977-78, so that the relation between
dose and duration of effect is unknown. In the West Twin-Dollar Lakes experi-
ences, the only lakes for which there is any published long-term monitoring,
effective P control was maintained for at least 5 years with doses over 20 g
Al/m3.
b. What area of the lake should be treated? At a minimum, the
area of anaerobic sediments should be covered with floe.
c. Will there be unacceptable side effects? Experience to date
indicates that as long as pH remains in an acceptable range and residual
dissolved aluminum does not exceed 50 ug Al/1, there will be no toxic effects
to fish. There is evidence of significant reduction in the species diversity
of planktonic microcrustacea. Increased water clarity may stimulate macro-
phyte growth, particularly in ponds and shallow lakes.
35
-------
5. Unit cost for treatment has not been established, but is related to
area and to amount to be added. For those treatments for which detailed costs
have been published, the mean (N = 5) man-days/metric ton applied is (± s.e.)
10.8 ± 4.83, for actual application. Construction, monitoring, and cleanup
will be additional.
There is apparently no significant social impact with phosphorus precipi-
tation/inactivation since disruption of lake use is minimal. There are no
disposal problems and no restriction on subsequent water use.
6. There are a considerable number of gaps in our knowledge about this
restoration technique. The relationship between dose and longevity of control
is poorly known. Only a few treatments have had sustained monitoring to
assess effectiveness and side-effects. A number of questions remain. How
often is reapplication necessary? Is a surface application as effective as a
hypolimnetic one in controlling P release? More toxicity studies, directed
primarily towards the actual level of organization to which the aluminum is
applied rather than to species, are also needed. What effect would an early
spring littoral application have on subsequent summer growth of macrophytes
and internal P release from the littoral?
36
-------
REFERENCES
1. Anderson, G. and E. Z. Arledge. 1962. The adsorption of inositol phos-
phates and glycerophosphates by soil clays, clay minerals and hydrated
sesquioxides in acid media. J. Soil Sci. 13:216-224.
2. Bandow, F. 1974. Algae control in fish ponds through chemical control
of available nutrients. Minn. Dept. Nat. Res., Div. Fish Wildlife.
Invest. Rept. 326. 22 pp.
3. Bernhardt, H. 1978. The oligotrophication of the Wahnbach Reservoir by
elimination of the phosphorus from the main tributary. Unpub. mss. 8
pp.
4. Bernhardt, H. , J. Clasen and H. Schell. 1971. Phosphate and turbidity
control by flocculation and filtration. Jour. Amer. Wet. Wks. Assoc.
63:355-368.
5. Biesinger, K. E. and G. M. Christensen. 1972. Effects of various metals
on survival, growth, reproduction and metabolism of Daphm'a magna. J.
Fish. Res. Bd. Can. 29:1691-1700.
6. Born, S. M. 1979. Lake rehabilitation: A status report. Environ.
Manag. 3:145-153.
7. Browman, M. G. , R. F. Harris and D. E. Armstrong. 1973. Lake renewal by
treatment with aluminum hydroxide. Draft report to Wisconsin Department
of Natural Resources. Madison, Wise.
8. Browman, M. G. , R. F. Harris and D. E. Armstrong. 1977. Interaction of
soluble phosphate with aluminum hydroxide in lakes. Tech. Rept. No.
77-05, Water Resources Center, Univ. of Wisconsin, Madison, Wise.
9. Burrows, W. D. 1977. Aquatic aluminum: chemistry, toxicology, and
environmental prevalence. CRR Critical Reviews in Environmental Control
7:167-216.
10. Carlson, R. E. 1977. A trophic state index for lakes. Limnol.
Oceanogr. 22:361-369.
11. Cooke, G. D. 1979. Evaluation of aluminum sulfate for phosphorus
control in eutrophic lakes. OWRT Proj. No. A 053-OHIO. Final Report.
Ohio Water Resources Center, Columbus, Ohio.
12. Cooke, G. D. , R. T. Heath, R. H. Kennedy and M. R. McComas. 1978.
Effects of diversion and alum application on two eutrophic lakes.
EPA-600/3-78-033.
37
-------
13. Cooke, G. D. and R. H. Kennedy. 1977. The short-term effectiveness of a
hypolimnetic aluminum sulfate application. Conference on Mechanics of
Lake Restoration. Madison, Wise.
14. Cooke, G. D. and R. H. Kennedy. 1978. Effects of a hypolimnetic appli-
cation of aluminum sulfate to a eutrophic lake. Verh. Int. Ver. Limnol.
20:486-489.
15. Cooke, G. D., M. R. McComas, D. W. Waller and R. H. Kennedy. 1977. The
occurrence of internal phosphorus loading in two small, eutrophic,
glacial lakes in Northeastern Ohio. Hydrobiol. 56:129-135.
16. Cooke, G. D. and D. W. Myers. Effects of a hypolimnetic alum treatment
on the planktonic microcrustacea of a eutrophic lake. Unpub. mss.
Department of Biological Sciences, Kent State University.
17. Dominie, D. 1978. Cobbossee watershed district lakes restoration
project. Progress report #4. Cobbossee watershed district. Winthrop,
Maine.
18. Dominie, D. Personal Comm. Letter dated March, 1979.
19. Dunst, R. C. et aJL 1974. Survey of lake rehabilitation techniques and
experiences. Tech. Bull. No. 75. Department of Natural Resources,
Madison, Wisconsin.
20. Eisenrich, S. J. , D. E. Atmstrong and R. F. Harris. 1977. A chemical
investigation of phosphorus removal in lakes by aluminum hydroxide.
Tech. Rept. Wise. Water Resources Center, No. 77-02. Univ. of Wisconsin,
Madison, Wise.
21. Ellis, J. D. The application of aluminum sulphate for the improvement of
water quality in lakes. Ontario Ministry of the Environment. 66 pp.
22. Everhart, W. H. and R. A. Freeman. 1973. Effects of chemical variations
in aquatic environments. Vol. II. Toxic effects of aqueous aluminum to
rainbow trout. EPA-R3-73-011b.
23. Fiat, D. and R. E. Connick. 1968. Oxygen-17 magnetic resonance studies
of ion solution. The hydration of aluminum (III) and gallium (III) ions.
J. Amer. Chem. Soc. 90:608-615.
24. Findenegg, I. 1972. Dos Phytoplankton des Reither See (Tirol,
Osterreich) im Jahre 1961. Ber. nat.-med. Ver. Innsbruck 59:15-24. Only
abstract seen.
25. Freeman, R. A. 1973. Recovery of rainbow trout from aluminum poisoning.
Trans. Amer. Fish. Soc. 102:152-154.
26. Freeman, R. A. and W. H. Everhart. 1971. Toxicity of aluminum hydroxide
complexes in neutral and basic media to rainbow trout. Trans. Amer.
Fish. Soc. 100:644-658.
38
-------
27. Funk, W. H. and H. L. Gibbons. 1979. Lake restoration by nutrient
inactivation. In: Lake Restoration, Proc. of Nat!. Conf., Minneapolis,
Minn. EPA-440/5-79-001. 141-151.
28. Funk, W. H. , H. R. Gibbons and S. K. Bhagat. 1977. Nutrient inactiva-
tion by large scale aluminum sulfate treatment. Conf. on Mechanics of
Lake Restoration, Madison, Wise., April, 1977.
29. Gahler, A. R. and C. F. Powers. Program proposal and cost evaluation for
lake restoration by nutrient inactivation. 36 p. U.S. Environmental
Protection Agency, Corvallis, Oregon.
30. Gannon, J. E. and R. S. Stemberger. 1978. Zooplankton (especially
crustacean and rotifers) as indicators of water quality. Trans. Amer.
Micros. Sco. 97:16-35.
31. Gasperino, A. F. and R. A. Soltero. 1978. Restoration of Medical Lake:
engineering design and preliminary findings. BN-SA-807. Battelle North-
west, Richland, Washington.
32. Gasperino, A. F. and R. A. Soltero. 1979. Personal Comm. Letter dated
January, 1979.
33. Haumann, D. and T. D. Waite. 1978. The kinetics of phosphate removal in
small alkaline lakes by natural and artificial processes. Water, Air,
and Soil Poll. 10:291-313.
34. Hayden, P. L. and A. J. Rubin. 1974. Systematic investigation of the
hydrolysis and precipitation of aluminum (III). Pages 317-381 In: A.
Rubin, ed., Aqueous-envi ronmental Chemistry of Metals. Ann Arbor
Science, Ann Arbor, Michigan. 390 p.
35. Heath, R. T. and G. D. Cooke. 1975. The significance of alkaline phos-
phatase in a eutrophic lake. Verh. Internat. Ver. Limnol. 19:959-965.
36. Hellstrom, B. G. 1979. Personal Communication. Letter dated January,
1979.
37. Hsu, P. H. 1965. Fixation of phosphate by aluminum and iron in acidic
soils. Soil Sci. 99:398-402.
38. Hsu, P. H. 1976. Comparison of iron (III) and aluminum in precipitation
of phosphate from solution. Water Research 10:903-907.
39. Jernelov, A. 1970. Aquatic ecosystems for the laboratory. Vatten
26:262-272.
40. Kennedy, R. H. 1978. Nutrient inactivation with aluminum sulfate as a
lake restoration technique. Ph.D. Dissertation, Kent State University.
292 pp.
39
-------
41. Kennedy, R. H. and G. D. Cooke. 1974. Phosphorus inactivation in a
eutrophic lake by aluminum sulfate application: a preliminary report of
laboratory and field experiments. Conference on Lake Protection and
Management, Madison, Wise.
42. Knauer, D. Personal Communication. Department of Natural Resources,
Madison, Wisconsin.
43. Kothandaraman, V., D. Roseboom and R. L. Evans. 1978. Pilot lake res-
toration investigations in the Fox Chain of Lakes. 111. State Water
Survey, Urbana. 44 pp.
44. Kumar, H. D. and L. C. Rai. 1978. Zirconium-induced precipitation of
phosphate as a means of controlling eutrophication. Aquat. Bot. 4:357-
366.
45. Larsen, D. P., K. W. Malueg, D. W. Schults and R. M. Brice. 1975.
Response of Shagawa Lake, Minnesota, USA to point-source phosphorus
reduction. Verh. Internat. Ver. Limnol. 19:884-892.
46. Linstedt, K. D. , E. R. Bennett, R. L. Fox, Jr. and R. D. Heaton. 1974.
Alum clarification for improving wastewater effluent quality. Water Res.
8:753-760.
47. May, V. 1974. Suppression of blue-green algal blooms in Braidwood
Lagoons with Alum. J. Aust. Inst. Agric. Sci. 40:54-57.
48. McNaught, D. C. 1975. A hypothesis to explain the success from calan-
oids to cladocerans during eutrophication. Verh. Internat. Ver. Limnol.
19:724-731.
49. Moffett, M. 1979. Changes in the microcrustaceari communities of East
and West Twin Lakes, Ohio, following lake restoration. M.S. Thesis.
Kent State University.
50. Morgan, J. J. and W. Stumm. 1964. The roll of multivalent metal oxides
in limnological transformations, as exemplified by iron and manganese.
In: 0. Jagg (Ed.), Advances in Water Pollution Research. Pergamon
Press, N.Y. pp. 103-118.
51. Narf, R. P. 1978. An evaluation of past 'aluminum sulfate lake treat-
ments: present sediment aluminum concentrations and benthic insect
renewal. Wise. Dept. Nat. Res., Madison, Wise.
52. Narf, R. and T. Wirth. Personal communication. Department of Natural
Resources, Madison, Wisconsin.
53. Packham, R. F. 1962. The coagulation process. I. Effect of pH and the
nature of the turbidity. Jour. Appl. Chem. 12:556-564.
54. Peelen, R. 1969. Possibilities to prevent blue-green algal growth in
the delta region of The Netherlands. Verh. Internat. Ver. Limnol. 17:
763-766.
40
-------
55. Peterson, S. A., W. D. Sanville, F. S. Stay and C. F. Powers. 1974.
Nutrient inactivation as a lake restoration procedurelaboratory invest-
igations. EPA-660/3-74-032.
56. Peterson, S. A., W. D. Sanville, F. S. Stay and C. F. Powers. 1976.
Laboratory evaluation of nutrient inactivation compounds for lake restor-
ation. J. Wat. Poll. Cont. Fed. 48:817-831.
57. Peterson, J. 0., J. J. Wall, T. I. Wirth and S. M. Born. 1973. Eutro-
phication control: nutrient inactivation by chemical precipitation at
Horseshoe Lake, Wisconsin. Tech. Bull. No. 62, Wise. Dept. of Nat. Res.,
Madison, Wise.
58. Powers, C. F. , F. S. Stay, W. D. Sanville, W. L. Lauer and G. S.
Schuytema. 1975. Lake restoration: zirconium inactivation of phos-
phorus in a eutrophic pond. USEPA. Corvallis Environmental Research
Laboratory Report 033.
59. Recht, H. L. and M. Ghassemi. 1970. Kinetics and mechanism of precipi-
tation and nature of the precipitate obtained in phosphate removal from
wastewater using aluminum (III) and iron (III) salts. Wat. Poll. Cont.
Res. Ser. 17010 EKI. 77 pp.
60. Sanville, W. D. , A. R. Gahler, J. A. Searcy and C. F. Powers. 1976.
Studies on lake restoration by phosphorus inactivation. EPA-600/
3-76-041.
61. Shannon, C. G. and W. Weaver. 1949. The mathematical theory of communi-
cation. Univ. of Illinois Press, Urbana. 117 pp.
62. Shannon, C. G. and W. Weaver. 1949. The mathematical theory of communi-
cation. Univ. of Illinois Press, Urbana. 117 pp.
63. Sdnnichsen, T. 1978. Toxicity of a phosphate-reducing agent (aluminum
sulphate) on the zooplankton in the lake Lyngby Stf. Verh. Int. Ver.
Limnol. 20:709-713.
64. Stumm, W. and J. J. Morgan. 1970. Aquatic Chemistry. An Introduction
Emphasizing Chemical Equilibria in Natural Waters. Wiley-Interscience.
New York. XV + 583 pp.
65. Wetzel, R. G. 1975. Limnology. W. B. Saunders Co., Philadelphia. XII
+ 743 pp.
66. Wilbur, D. L. 1974. The effect of aluminum sulfate application for
eutrophic lake restoration on benthic macroinvertebrates and the Northern
Fathead Minnow (Pimephales promelas Ref.). M.S. Thesis. Kent State
University.
41
i US GOVERNMENT PRINTING OFFICE 1981-757-064/0293
------- |