RESEARCH CuSO4 Treatment of Nuisance Algal Blooms in Drinking Water Reservoirs

DIANE M. M c K N I G H T t SALLIE W. CHISHOLM DONALD R. F. HARLEMAN

Ralph M Parsons Laboratory Department of Civil Engineering Massachusetts Institute of Technology Cambridge, Massachusetts 02139

ABSTRACT / Since the turn of the century, copper sulfate has been used extensively as an algicide to control nuisance algal blooms in drinking water reservoirs Recent experimental results have shown

that the toxicity of copper to algae is determined by the activity of ionic copper, and not the total copper concentration The sensitivity of algae to ionic copper has been found to vary considerably for different algal species Chemical processes such as precipitation, complexation by dissolved organic substances, and adsorption by suspended material can be important in controlling the chemical speciation of copper added to drinking water reservoirs The copper sulfate dosage required for effective treatment of a reservoir is shown to depend on water chemistry and the copper sensitivity of nuisance algal species By evaluating copper chemistry and copper sensitivity of nuisance algal species it may be possible to obtain effective treatment with lower copper sulfate dosages in some reservoirs

Copper sulfate has been .used extensively to control algal blooms in lakes and reservoirs in the United States (Moore and Kellerman 1905) and is still considered to be the algicide of choice for control of most nuisance algal blooms. In this report, we outline the chemical and biological processes controlling the effect and fate of copper added to reservoirs. Further, it is suggested that in some cases improvements in algal control strategies can be made based on consideration of these chemical and biological processes. From the initial use of copper as an algicide, the free cupric ion has been thought of as the toxic copper species. Water supply operators have also been aware of differences in sensitivity to copper among algal species (Hale 1954). However, simple formulae that prescribe CuSO4 dosages based on pH, alkalinity, and nuisance algal species have not been very useful, probably because of their empirical nature. A more general formula for determining the effective CuSO4 dosage would calculate the cupric ion activity as a function of total copper added from the relevant chemical characteristics of the reservoir and compare that with the cupric ion activity that is toxic to the resident nuisance algal species. Research on the environmental chemistry of copper and the toxicity of copper to phytoplankton has progressed sufficiently that is now possible to obtain the data necessary for computing optimal CuSO4 dosages (for example, Morel and others 1979), but this is still not practical on a routine basis.
Toxicity of Copper to Freshwater Plankton

The ecology of freshwater algae in relation to the management of reservoirs was discussed by Lurid (1955), who pointed
KEY WORDS: Algalblooms;Coppertoxicity;Coppersulfatetreatment;Drinking water; Reservoirs

ITowhomcorrespondence shouldbe addressedat The U.S.GeologicalSurvey, Water ResourcesDivision,M.S. 407,5293Ward Road,Arvada,CO 80002.
EnvironmentalManagement,Vol 7, No 4, pp 311-320

out the importance of monitoring algal populations regularly to anticipate nuisance algal blooms, and the importance of organic substances in controlling copper toxicity by complexation. Since then, many researchers have shown that complexation by natural and synthetic organic compounds decreases the toxicity of copper (Fogg and Westlake 1955, Fitzgerald and Faust 1963, Manahan and Smith 1973). Gaehter and others (1973) showed that complexation by organic ligands in lakewater determined the toxicity of copper to Chlorella pyrenoidosa. They concluded that ionic copper is likely to be toxic at concentrations of 10-10 M, and that the high organic content of lakewater necessitates the use of much higher CuSO4 dosages. In a field study of a CuSO4 treatment, complexation by humie substances was shown to control the observed change s in the phytoplankton assemblage (McKnight 1981). Although it had been hypothesized that algae can release detoxifying metabolites that complex copper, recent laboratory data indicate that concentrations of algal-produced copper-eomplexing agent in freshwater ecosystems are probably too low to control the speciation of copper except during or after algal blooms (McKnight and Morel 1979, 1980). If a reservoir is treated with CuSO4 before a bloom has developed, complexation of the added copper by algal metabolites is unlikely to be significant. Although it had been believed that the toxic agent was the free cupric ion, it was not until recently that researchers quantified the toxicity of the free copper ion to phytoplankton. By using different concentrations of copper and chelating agents, the toxic response was shown to be a unique function of the cupric ion activity (pCu 2+ = -log [Cu 2+ ]) for four species of marine algae (Sunda and Guillard 1976; Anderson and Morel 1978, Sunda and Lewis 1978). Determination of toxicity in terms of the cupric ion activity allows for meaningful comparison of toxicity data from different researchers. In Figure 1 the growth rate (/~) expressed as a percent of the growth rate of control cultures (#max) is plotted as a function of cupric ion activity for seven algal species (McKnight 1979;
9 1983 Springer-Verlag New York Inc

312

o.M. McKnight

andothers

100 -

80-

= _o
_>
O" 0 r

60-

40--

20

06

9 P/Cu 2 " t

10

11

12

13

14

Figure 1. Inhibition of growth as a function of cupric ion activity for various species of freshwater (denoted by *), estuarine and marine algae.

Sunda and Guillard 1976, Sunda and Lewis 1978). Toxicity data for four additional algal species where parameters other than growth rate were used are included in Figure 1 by expressing the data as percent of algal cells that were motile for the dinoflagellates Gonyaulax tamarensis (Anderson and Morel 1978) and Ceratium hirundinella (a), as percent 14C uptake for the marine blue-green alga Trichodesmium theibautic (Rueter and others 1979) (b),' and as the lag phase duration of control cultures divided by the lag phase duration of copper-stressed cultures for the coastal diatom Thalassiosira pseudonana (Reuter 1979) (c). The toxic cupric ion activities range from greater than 10 -6 to 10 -11 M for species of diatoms, dinoflagellates, green algae, and blue-green algae. The conclusion that there are order of magnitude differences in copper sensitivity among algal species is also supported by earlier experimental results (Mandelli 1969, SteemannNeilsen and others 1969, Steemann-Neilsen and WiumAnderson 1970, Steemann-Neilsen and Brunn-Larsen 1976) and by observations of changes in the phytoplankton following CuSO4 treatment episodes. In the context of control of nuisance algal blooms in drinking water reservoirs, the copper sensitivity of blue-green algae is especially important. The ecology of blue-green algae is complicated by their unique physiology, which includes formation of gas vacuoles for buoyancy, fixation of atmospheric nitrogen, and excretion of siderophores that can complex copper. Blue-green algae appear to be more sensitive to copper

than some eucaryotic algae (Whitton 1973), but this generalization needs further experimental support. Suppression of nitrogen fixation by Anabaena and Aphanizomenon has been observed after copper additions of 5 to 10 #g/L, which was much less than the copper-complexing capacity of the lakewater (Horne and Goldman 1974, Elder and Horne 1978). In a study of low level CuSO4 treatment (10 /zg/L Cu) of Casenovia Lake, New York, Anabaena flos-aquae was observed to be more affected than species of coccoid colonial blue-green algae (Effler and others 1980). Maintenance of sufficient copper concentrations to prevent fixation of atmospheric nitrogen has been suggested by Horne and Goldman (1974) as an alternative to a single higher dosage CuSO4 treatment for prevention of blue-green algal blooms. Addition of copper to a reservoir will also affect the resident zooplankton and fish populations, either by direct toxicity or by depletion or change in the algal biomass available for grazing. A desirable feature of an algicide is that it be nontoxic to zooplankton, fish, and other non-target organisms. Changes in zooplankton populations are also important because they may affect the longer term success of CuSO4 treatment; for example, removal of grazing pressure may allow a rapid increase in surviving phytoplankton populations. Zooplankton populations were not affected by a low level CuSO4 treatment in the Casenovia Lake study (Effler and others 1980). However, significant decreases of populations in Bosmina, Tetramastix, and Keratella occurred following treatment of Mill Pond,

CuSO4Water Treatment

313

Massachusetts, with a higher level CuSO4 dosage (60/~g/L Cu) (McKnight 1981). There are little field data on the survival of fish following a typical CuSO4 treatment, although instances have been reported where treatment of fully developed algal blooms resulted in anoxic conditions and fishkills (Muchmore 1976). Laboratory studies show that the cupric ion activity is again the critical parameter determining toxicity to zooplankton and fish (Pagenkopf and others 1974, Andrew and others 1977, Waiwood and Beamish 1978) (Figure 2). Clearly more data, preferably for co-occurring species, are necessary before any generalizations on the relative copper sensitivity of phytoplankton, herbivores and planktivores can be made.

~, ec

, a

~x40
9

2O

Chemistry of Copper in Freshwater

In the previous discussion, the response of aquatic organisms to copper was shown to be a function of the cupric ion activity. There are several chemical processes that can regulate the cupric ion activity resulting from addition of copper to a lake or reservoir: (1) inorganic complexation (for example CuCO3 (aq), Cu(CO3)22 - , Cu(OH)22- , CuOH + ); (2) precipitation (for example, amorphous Cu(OH)2, malachite (Cu(OH)2CO3), tenorite (CuO), and covellite (CuS); (3) complexation by dissolved organic compounds (for example, amino acids, polypeptides, humic and fulvic acids); (4) adsorption by days, hydrous metal oxides, or other particulates; and (5) biological uptake. All of the above processes could be important at some time after CuSO4 treatment. The key parameters are pH, alkalinity, the relative concentrations of copper, dissolved organic ligands, surface binding sites on particulates, and the potential for copper assimilation by the biota.
O, 7 8 8 P~Cu 24"} 10 11 12

F i g u r e 2. Inhibition of growth as a function of cupric ion activity for several freshwater algae (Sunda and Lewis 1978; McKnight 1979), a freshwater cladoceran (Andrew and others 1977), and rainbow trout (Waiwood and Beamish 1978).

InorganicComplexationand Precipitation
If only inorganic complexation and precipitation are considered, the major copper species and the cupric ion activity can be computed from chemical equilibrium models. Figure 3 shows the major copper species as a function of pH and total copper for two different concentrations of total carbonate, CT = 10-3"4 M as in Mill Pond reservoir (McKnight 1981) and CT = 10 -2"ts M as in lake 885 (Whitaker and others 1978). The pH and total copper concentrations were chosen to be representative of the CuSO4 treatment of drinking water reservoirs. Below neutral pH the free copper ion is the major species and above neutral pH copper carbonate complexes, malachite, and tenorite are predicted to be the major copper species. These diagrams provide a partial explanation for the need for higher CuSO4 dosages in reservoirs with high pH and alkalinity. At the higher carbonate concentration, CuCO3(aq) is the major copper species for a greater range of total copper and pH. The

importance of copper carbonate complexes in freshwater was shown by Stiff (1970) and used by Andrew and others (1977) to explain the observation that copper is less toxic in hard waters. At both values of total carbonate, malachite is predicted to precipitate in neutral or alkaline waters at total copper concentrations that are commonly used in CuSO4 treatment. Although there is some uncertainty in the solubility constant for malachite, in terms of the efficacy of CuSO4 treatments, the uncertainty associated with the time course of malachite precipitation, which may take several days or longer, (Stiff 1971) is the most significant. Because of the uncertainty in the rate of copper precipitation, it is difficult to compare field data from CuSO4 treatment episodes with chemical equilibrium calculations. Button and others (1976) studied two CuSO4 treatments of the Hoover reservoir in Ohio (pH 7.8 and alkalinity 96 rag/L). The added copper concentrations (10 -5 and 5 x 10 -5 M) exceeded the solubility of malachite, and the particulate copper was observed to increase and settle to deeper waters over 24 hours. Although these data are consistent with the removal of copper from the surface by malachite precipitation, without direct evidence of malachite formation, other explanations, such as uptake of copper by dying diatom populations, must still be considered. Whitaker and others (1978) studied the CuSO4 treatment of an alkaline farm pond in Manitoba and reported measured cupric ion activities several orders of magnitude higher than those predicted by malachite precipitation or formation of

314

o.M.McKnightand others

'[
u I r

~,

Malachite

Cut+

~i

=(OH)=cO3(s)
Tenorite

CuO(s)

~h,

CuCOs(ocl.) ?
pH

Molochlte

Cu~(OH )zCO~ ($]

Tenorite

CuO (s) g 7 ~ Cu= CuCO=[oq,)

I'

pH

Figure 3. -Log (total copper) versus pH diagram showing predicted precipitation of malachite and tenorite. A. Total carbonate (CT) = 10-3# M. B. CT = 10-215 M. CuCO3(aq). Interference in the measurements with the cupric ion selective electrode may explain the discrepancy. Although precipitation of covellite (CuS) or chalcocite (Cu2S) is unlikely to be important in surface waters, precipitation of copper sulfides may control the transport of copper in reservoirs with anoxic bottom waters (McKnight 1979). Precipitation of copper sulfides in the metalimnion may accelerate the transport of the added copper to the sediments and decrease the time during which toxic concentrations of copper are maintained in the epilimnion.
Organic Complexation

The eomplexation of copper by dissolved organic compounds is difficult to model in chemical equilibrium calculations because the nature of the organic compounds in lakewater and their copper-complexing properties are poorly understood.

There are three categories of organic substances that can act as complexing agents: 1) low molecular weight compounds such as amino acids; 2) polymeric compounds such as polypeptides, polysaccharides, and humic and fulvic adds; and 3) colloidal organic compounds and organic compounds bound to inorganic colloids (Stiff 1971, Stumm and Baccini 1978). Since it is not feasible to quantify the copper-complexing properties of the individual organic compounds present in lakewater, the approach has been to determine the total complexing capacity experimentally either by bioassay experiments (Gillespie and Vaccaro 1978, Gachter and others 1973), by voltametric methods (Chau and others 1974, Smith 1976) or by potentiometric methods (Sunda and Hansen 1978, McKnight 1981). With potentiometric methods it is possible to determine the complexing capacity and the cupric ion activity as a function of total added copper, which are the critical data needed for the optimization of CuSO4 treatment strategies. For organic complexation to be the dominant chemical process controlling the cupric ion activity during a CuSO4 treatment, the added copper concentration much be less than or equal to the equivalent ligand concentration. Copper-complexing capacities for freshwaters have been reported from 10 -8 to 10 -5 M, and when these values are compared with the usual dosages of CuSO4 (from 5 x 10 -v to 10-s M), organic complexation would appear to be important only in lakewaters high in organic substances. The maintenance of elevated soluble copper concentrations following CuSO4 treatment reported by Elder and Home (1978) and Whitaker and others (1978) can be explained by formation of copper-organic complexes or CuCO3(aq). Complexation of copper by humic substances was shown to control copper speciation following CuSO4 treatment of Mill Pont reservoir (McKnight 1981). However, two other reservoirs commonly treated with CuSO4 were found to have complexing capacities less than 10 -v M (McKnight 1979). Even in reservoirs enriched in dissolved organic substances, the upper limits of equivalent ligand concentrations are in the range of common CuSO4 dosages. Therefore, unlike carbonate and hydroxide complexation, complexation by organic compounds can be saturated by addition of reasonable amounts of CuSO4 and this is probably the optimal control strategy. Titrating the organic ligands will give the maximum increase in the cupric ion activity for the minimum copper addition. Mill Pond reservoir is an example of this strategy; as a result of experience and coincidence, the added copper concentration was equal to the equivalent ligand concentration of the humic substances in the water. The presence of organic ligands may increase the effectiveness of treatment by stabilizing soluble copper and retarding precipitation and adsorption, which has been the major advan-

CuSO4Water Treatment

315

tage put forth by proponents of synthetic copper chelates as algicides (Anon. 1976, Fair and others 1971). The optimal strategy with synthetic ligands is still to add enough copper to titrate the ligand. Since synthetic chelators such as EDTA, NTA, and triethanolamine will bind copper more strongly than natural ligands such as humic and fulvic acids, the cupric ion activity will be lower in the case of the synthetic ligands for similar concentrations of ligand and copper. In fact, if the added copper is less than equimolar with the added chelator, it is possible that the cupric ion activity will be lower than before additon of the algicide. Use of synthetic chelating agents has improved the control of algal blooms primarily in hard-water alkaline reservoirs, and their advantage probably results from decreasing the supersaturation of malachite and tenorite and thereby the rate at which equilibrium is approached. Slower equilibrium would result in maintenance ot toxic cupric ion activities and inhibition of algal growth for longer periods after addition of the copper. Because of the increased cost and possible public health hazards, the actual changes in the cupric ion activity resulting from use of synthetic organic complexing agents should be thoroughly assessed.

of such nuisance species as Synedra and Asterionella, which may become silicon limited. The Hoover reservoir study (Button and others 1977) is an example in which the formation of particulate copper can be explained by precipitation of malachite, biological uptake by dying diatom populations, adsorption on dead diatom frustules, or adsorption on clays, hydrous metal oxides, or other organic particulates such as bacteria, fecal matter, and zooplankton exoskeletons. None of these processes can be ruled out by theoretical arguments or by simple analysis of particulate material after a CuSO4 treatment episode, and any one of these processes may be dominant in a given reservoir.

Accumulation of Copperin Sediments

Many of the chemical and biological processes important in CuSO4 treatments will eventually transport copper to the sediments. There has been concern that copper accumulated in the sediments will be toxic to benthic organisms and act as a source of copper after treatment has been discontinued. Paleolimnological technqiues were used by Brugam (1978) to study the sediments of Linsley Pond in Connecticut which had been treated once with CuS04 in 1938. He found no evidence of increased copper deposition in 1938, and in fact found highest deposition of copper from 1900 to 1920, when Bordeaux mixture (lime and CuSO4) was used as a pesticide on apples in the surrounding countryside. In a study of the chemistry of copper added to Lake Monona in Wisconsin, 1.5 x 106 lb over a 30 yr period, Sanchez and Lee (1978) found that copperenriched sediments (600 mg/kg dry mauer at 60 cm depth) were not interacting with the more recent sediments deposited, since CuSO4 treatment was discontinued or with the overlying waters. In both Lake Monona and Linsley Pond, copper concentrations in recent sediments (250 and 180 mg/kg dry matter) were found to be higher than in sediments from the turn of the century (50 mg/kg); even though CuSO4 is not currently being used in either lake.

Adsorption and Biological Uptake

From the perspective of controlling algal blooms, adsorption by clays and hydrous metal oxides detoxifies copper, whereas uptake of copper by algae and adsorption on dead algal eeUs are part of a successful CuSO4 treatment. Unfortunately, it is difficult to separate analytically, in a field sample, copper that has been precipitated, adsorbed by nonliving particulates or assimilated by living plankton. Adsorption and biological uptake are both dependent on the pH and the cupric ion activity of the surrounding medium, and may not be as significant in acidic waters (Schindler and others 1979i or in waters where the copper is strongly complexed (Vueeta 1976). The dependence of algal copper assimilation on the cupric ion activity is responsible for the observed dependence of copper toxicity on cupric ion activity. Anabaenaflos-aquae was shown to take up more copper per dry weight at a given copper concentration than Scenedesmus quadricuada, but toxocity occurred at the level of 1% dry weight for both species (Gibson 1972). S. quadricuada has been observed to replace A. flosaquae as the dominant algae after CuSO4 treatment. Foster (1977) also found that toxicity occurred at the same concentrations of cellular copper in copper-tolerant and copper-sensitive strains of Chlorella vulgaris. In experiments with a marine diatom, Rueter and others (1981) have found that copper uptake is also dependent on silicon nutrition, with more copper being assimilated by silicon limited cells. This result has implications for the CuSO4 treatment of spring diatom blooms

Examples of CuS04 Dosage Control

How the biological and chemical processes discussed in the previous sections determine effective CuSO4 dosages can be summarized by considering different combinations of water chemistries (Figure 4) and sensitivities of algae to copper (Figure 5). As a first approximation, the effective concentration of copper added to the reservoir can be computed initially by considering the epilimnion to be well mixed and isolated from the hypolimnion. With the typical treatment practice this appears to be a good representation, as demonstrated by a field study in Mill Pond (McKnight 1981). In that study, the copper added to the pond was evenly distributed in the epilimnion

316

o.M. McKnight and others

o.ot

o.lo

~.o

ppm

Cu

0.8

o
p

Pmax

o.s

0,4
0.2

olou,iI
z+) 7 5 6 7 8 9 I0 II

if
/

l~ r +
p Cu T
Figure 5. Copper spedation in three hypothetical freshwaters: A

.I cu + I
Figure 4. Copper sensitivity of three hypothetical nuisance algal species: I copper tolerant, 2 intermediate copper tolerance, and 3 copper sensitive.
within 24 hours of the treatment with no losses to the hypolimnion or littoral zone on this time scale. In Figure 4, three hypothetical copper sensitivities are presented which are representative of a copper tolerant alga (line 1), an alga with intermediate copper sensitivity (line 2), and a copper-sensitive alga (line 3). In Figure 5, line A represents an acidic water (pH ~ 6) where almost all added copper is present as free copper ions, line B represents a water of neutral pH with sufficient dissolved humic substances (~-4 mg C / l ) to complex copper added at concentrations less than 0.10 ppm, and line C represents an alkaline water (pH -< 8.5) where precipitation of copper carbonates or hydroxides is predicted for copper additions in the range used in CuSO4 treatment. In Figure 6, the effective CuSO4 dosages for these hypothetical water chemistries and algal sensitivities are shown. For the copper-sensitive alga, low dosages of CuSO4 (0.01 ppm) should be effective in controlling algal growth regardless of whether the reservoir is aeidic, alkaline, or has significant organic eomplexing agents. These low dosages may not be practical, however, because of difficulty in uniformly distributing copper in the epilimnion of a reservoir. In the case of a copper-tolerant alga, in all but the alkaline water, control should be possible with dosages of about 1 ppm (at the high end of dosages currently used). In the alkaline water, precipitation of copper carbonates and hydroxides prevents achievement of cupric ion activities sufficiently high to limit the growth of a copper-tolerant nuisance alga, and other techniques such as destratification may be necessary.

acidic, B containing organic complexing agents, and C alkaline; reference line for p[Cu2+ ] = p C u T is included. The effective dosage for an alga with intermediate copper sensitivity shows the most dependence on water chemistry. In acidic waters, low dosages (_<0.10 ppm) should be effective. In a water with significant organic content enough copper must be added to titrate the humic complexing agents (0.10 ppm) before a toxic effect will be obtained. Similarly, dosages grater than 0.10 ppm are required in alkaline waters because of copper precipitation.

Potential for Improvements in CuS04 Treatment Practices

Although there as been some action in several states (for example, Maine and Illinois) to discontinue the use of CuSO4 as an algicide, CuSO4 is nonetheless expected to remain the algicide of choice for the next fifty years (McKnight 1979). As the need for potable water increases, the use of more marginal water supplies and an increased need for algal control measures can be anticipated. The recent development of experimental procedures to study the chemistry and toxicity of copper in natural ecosystems is only one of the many advantages of CuSO4 in comparison with other potential algieides such as potasium permanganate, silver quaternary amonium, and organic mercury compounds. The main advantages are low cost, proven effectiveness, and lack of public health hazard. The fact that detailed field studies of CuSO4 treatment episodes are feasible does not imply that they are worthwhile.

CuSO, Water Treatment

3 17

o.o= o.lo
5

I.o ppm Cu

,/

:l'
B
9 I0

o.oe o.lo t.o ppm Cu

5 C

o.ol o.ao i.o

/
6

ppm Cu

,
9
I0 3 2 1

///
/ I

'
9 IO 2

p CUT Figure6.

p Cul.

p CUT

Relationshipbetween copper sensitivityof nuisance algal species and copper sulfate dosagefor three representativefreshwaters. A, B, and C and 1, 2, and 3 are describedin Figures 4 and 5. Application of low CuSO4 dosages (less than 100 ppb) when algal counts and weather conditions indicate an incipient nuisance algal bloom is currently the treatment strategy followed by water supply managers in the Boston metropolitan area and in many other localities (Faucher 1978). From the previous discussion it can be seen that for a typical reservoir with nuisance algal species of intermediate copper tolerance this strategy provides algal control with minimal impact on other aquatic organisms and with low risk of inadequate water supply to the public. The success of this strategy depends on the ability of water supply managers to anticipate nuisance algal blooms, and in this they rely on past experience and detailed algal counts. At reservoirs where this strategy is effective, it is unlikely that more complete knowledge of the chemistry of copper would lead to major changes in how or when CuSO4 is applied. There is a possibility that in acidic reservoirs or in reservoirs with copper-sensitive nuisance algal species, very low CuSO4 dosages (10 ppb) may be effective. However, there is a risk; treatment with copper insufficient to inhibit rapidly growing algae will subsequently necessitate use of much higher dosages to break up a fully developed algal bloom. Detailed study of the chemistry and biology of copper may help improve treatment practices in reservoirs where the single shot strategy is not practiced or does not provide adequate prevention of algal blooms. Unsuccessful treatments may result from failure to predict algal blooms from patterns of algal growth. The ecology of blue-green algae, the most common nuisance algae, is very complex (Whitton 1973), and although no one parameter can reliably predict a bloom, monitoring of the concentrations of major nutrients (nitrate and phosphate) may be helpful. Unsuccessful control of algal blooms by addition of moderate CuSO4 dosages (about 100 ppb or 10 -6 M) may also result from detoxifying chemical processes such as precipitation, adsorption, and complexation that maintain low cupric ion activities (about 10 -1~ M) or from the ability of the particular nuisance alga species to tolerate high cupric ion activities (about 10 -7 M). Experimental methods based on potentiometric measurement of cupric ion activity can be used to determine the chemical processes controlling copper speciafion and the copper tolerance of the nuisance algal species (see Appendix I). In large reservoirs with serious algal problems, these methods and detailed field studies may be worthwhile. For example, if experiments show that the added copper precipitates rapidly, then the use of synthetic chelating agents may be justified; or if the nuisance algal species is found to tolerate high cupric ion activities, it may be necessary to use other treatment strategies such as destratification or reduction of nutrient inputs to the reservoir (Shapiro and others 1975).

Conclusions
Copper sulfate has been the algicide of choice since 1905. Consideration of the chemical and biological processes acting in a CuSO4 treatment episode shows how CuSO4 dosages can be optimized, but does not point to major changes in the engineering practice of CuSO4 treatment of drinking water reservoirs. Experiments characterizing the copper tolerance of nuisance algae and the copper chemistry of reservoirs may result in reduction of copper sulfate dosages from 10 -5 to 10 -6 or 10 -7

318

o.M. McKnight andothers

M in some reservoirs. Further improvement could come from better prediction of algal blooms by water supply managers, based on more frequent algal counts or monitoring the concentrations of major nutrients.

Acknowledgments
This research was supported by International Copper Research Association, Inc., Project No. 252. The authors wish to thank Professor Francois M. M. Morel for his varied and essential contributions to this work.

Appendix I Experimental Approaches to More Effective Treatment

Here we outline an experimental approach that could be used to identify chemical and biological processes important in the CuSO4 treatment of a particular reservoir. In the future, greater demands may be placed on marginal water supplies where copper sulfate treatment is currently ineffective or requires high dosages. In these cases, the data from such experiments may be necessary to upgrade algal control practices. Although these experiments will not provide quick, precise information on how much CuSO4 to use, if carried out during one spring/summer season the data could be used to improve treatment practices the following year. The experiments presented here are scaled down from the ones presented in McKnight (1979) and can be carried out in a laboratory with standard equipment and access to an atomic absorption spectrophotometer for copper analyses.

filtered nutrient-enriched reservoir-water. This determination requires a cupric ion selective electrode, a reference electrode, a pH electrode, and a pH meter. All are standard equipment except for the cupric ion selective electrode, which can be purchased for less than five hundred dollars. The detection limits of the cupric ion selective electrode (10 -7 M or 10 ppb total copper) are equal to the lowest practical CuSO4 dosages. The electrode is standardized in copper solutions at pH 4 (McKnight and Morel 1979). A suggested series of total copper additions to the reservoir samples is 10 -7, 10 -6, 5 10 -6, 10 -5, and 2 x 10 -s M. The pH is maintained at the pH of the reservoir for all copper additions by addition of acid or base. The signal from the cupric ion electrode is converted to cupric ion activities using the Nernst equation: Eh = Eo - A log (Cu 2+) where A = R T / z F = 29.5 mv for copper, Eh is measured potential, and Eo is the standard potential determined from measurements in the standard copper solutions. Variation in A between 27 and 32 my/decade is acceptable. It is important to wait until the signal has become stable (less than 4 mv/hr) before recording the measurements. Failure of the signal to stabilize may indicate slow precipitation of copper and measurements of particulate copper should be made. A log-log plot is made of cupric ion activity (Cu 2+) versus total copper (CUT) for both unfiltered and filtered nutrient-enriched water with the line (Cu 2+) = CUT included for reference. Significant differences between filtered and unfiltered samples indicate that either adsorption or biological uptake are important.

Sensitivity of Nuisance Algal Species to Copper

The next step is to measure the growth of the nuisance algal species in filtered medium with four or five evenly spaced cupric ion activities ranging from no copper addition to 1 x 10 -s M (0.63 ppm) added copper. Clean glassware is recommended, but there is no need for exacting trace metal clean techniques at these elevated copper concentrations. Replicate cultures at each copper addition are inoculated with the nuisance algal species and growth is monitored daily for 5-10 days by either cell counts, optical density, chlorophyll a, dry weight, or another convenient measure of biomass. A final cupric ion activities measurement is made at the end of the experiment. Measurement of assimilated copper may also be informative. The growth rate is computed from the growth curves (Guillard 1973) and plotted as a function of the negative logarithm of the cupric ion activity, p(Cu2+). From the results of these experiments, sound decisions can be made about the changes in CuSO4 treatment practices at the reservoir. The cupric ion activity which will prevent a bloom of the nuisance algal species with minimum impact on non-target organisms will be about twice the cupric ion activity of the

Nuisance Algal Species

The first step is to obtain cultures of the important nuisance algal species in the particular reservoir. Techniques for isolating algae are described by Guillard (1973), Hoshaw and Rosowski (1973), and Allen (1973). If culture of the nuisance algal species is not feasible, experiments can be designed to use concentrated phytoplankton samples from the reservoir. Although sterilization of growth medium is advisable, axenic cultures are not required. The best medium for algal culture is filtered (0.4 #m) reservoir-water enriched with nitrate, phosphate, silicate (for diatoms), and trace metals with no synthetic chelating agents--see Guillard (1975) for suggested concentrations of nutrients. Cultures can be grown on a sunny laboratory bench. Temperature regulation is only necessary for species that bloom at low temperatures.

Determination of Cupric Ion Activity

The second step is to determine the cupric ion activity as a function of the total copper added to reservoir-water and

CuSO4Water Treatment

319

breakpoint of the growth rate (~) versus (Cu 2+) plot. The corresponding copper addition can be taken from the plot of (Cu 2+) versus added copper in unfiltered reservoir water, and that value can be used to compute how much copper to add to the reservoir for effective control of nuisance algal growth. The relationship between growth rate and (Cu 2+) is not expected to change significantly from year to year for the same algal species. However, the chemistry of the water in the reservoir may change during a spring/summer season and from one year to the next. The determination of (Cu 2+) for different copper additions in unfiltered samples is not a lengthy process and could be repeated prior to treatment if time perimts.

Literature Cited
Allen, M. M. 1973. Methods for Cyanophyceae. Pages 127-139 in J. R. Stein (ed.) Handbook for phycological methods. Cambridge University Press. New York, NY. Anderson, D. M., and F. M. M. Morel. 1978. Copper sensitivity of Gonyaulax tamerensis. Limnol. Oceanogr. 23:283-295. Andrew, R. W., K. E. Biesinger, and G. E. Glass. 1977. Effects of inorganic complexation on the toxicity of copper to Daphnia magna. Water Resour. 11:309-315. Anonymous. 1976. How to identify and control water needs and algae. Applied Biochemists, Inc. Mequon, WI. Brugam, R. B. 1978. Human disturbance and historical development of Linsley Long. Ecology 59:19-36. Button, K. S., H. P. Hostetter, and D. M. Mair. 1977. Copper dispersal in a water-supply reservoir. Water Resour., 11:539-544. Chau, Y. K., R. Gachter, and K. Lum-Shuc-Chan. 1974. Determination of the apparent complexing capacity of lakewaters. J. Fish. Res. Bd. Can. 31:1515-1519. Davey, E. W., M. J. Morgan, and J. j . Erikson. 1973. A biological measurement of copper complexion capacity of seawater. Limnol. Oceanogr. 18:993-1977. Effler, S. W., S. Litten, S. D. Field, T. Tong-Ngork, F. Hale, M. Meyer, and M. Quirk. 1980. Whole lake responses to low level copper sulfate treatment. Water Resour. 14:1489-1499. Elder, J. F., and A. J. Home. 1978. Copper cycles and CuSO4 algicidal capacity in two California lakes. Environ. Manage. 2:17-30. Erickson, S. J., N. Lackie, and T. E. Malone},. 1970. A screening technique for estimating copper toxicity to estuarine phytoplankton. J. W.P.C.F. 42:R271-278. Fair, M. G., C. J. Geyer, and D. A. Okun. 1971. Elements of water supply and wastcwater disposal. John Wiley and Sons, Inc. New York, NY. Faucher, F. 1978. Personal communication. Fitzgerald, G. P., and S. L. Faust. 1963. Factors effecting the algicidal and algistatic properties of copper. Appl. Microbiol. 11:345-351. Fogg, G. E., and D. F. Westlake. 1955. The importance of extracellular products of algae in freshwater. Int. Assoc. Theoret. Appl. Limnol. 12:219-232.

Foster, P. L. 1977. Copper exclusion as a mechanism of heavy metal tolerance in a green alga. Nature 269:332-333. Gachter, R., K. Lum-Shuc-Chan, and Y. K. Chau. 1973. Complexing capacity of the nutrient medium and its relation to inhibition of algal: photosynthesis by copper. Schweizer. Aeits. fur Hydrol. 35:253-260. Gibson~, C. E. 1972. The algicidal effect of copper on a green and blue'green alga and some ecological implications. J. Appl. Ecol. 9:512-518. Gillespie, P. A., and R. F. Vaccaro. t978. A bacterial hioassay for measuring the copper-chelation capacity of seawater. Limnol. Oceanogr. 23:543-548. Guillard, R. R. L. 1973. Methods for microflagellates and nannoplanktons and, Division rates. Pages 87-104, 289-312 in J.R. Stein (ed.) Handbook of phycological methods. Cambridge University Press. New York, NY. Guillard, R. R. L. 1975. Culture of phytoplankton for feeding marine invertebrates, pages 2-60 in W.L. Smith and M.H. Chanley (eds) Culture of marine invertebrate animals. Plenum Pub. Co. New York, NY. Hale, F. E. 1954. The use of copper sulfate in control of microscopic organisms. Phelps Dodge Refining Corp. New York, NY. Home, A. J., and C. R. Goldman. 1974. Suppression of nitrogen fixation by blue-green algae in a eutrophic lake with trace additions of copper. Science 183:409-411. Hoshaw, R. W. and J. R. Rosowski. 1973. Methods for microscopic algae. In J. R. Stein (ed), Handbook of phycological methods. Cambridge University Press. New York, NY. Lund, J. W. G. 1955. The ecology of algae and waterworks practice. Proc. Soc. Wat. Treat Exam. 4:83-109. Manahan, S. E., and M. J, Smith. 1973. Copper micronutrient requirements for algae. Environ. Sci. Technol. 7:829-833. Mandelli, E. F. 1969. The inhibitory effects of copper on marine phytoplankton. Contrib. Mar. Sci. 14:47-57. Martin, D. F., and W. K. Olander. 1971. Effects of copper, titanium and zirconium on the growth rates of the red tide organisms, Gymnodinium breve. Environ. Lett. 2(3):135-142. McKnight, D. M. 1979. Interactions between freshwater plankton and copper speciation. Ph.D. thesis. Mass. Inst. Technol., Cambridge, MA. 284 pp. McKnight, D. M. 1981. Chemical and biological processes controlling the response of a freshwater ecosystem to copper stress: A field study of the CuSO4 treatment of Mill Pond Reservoir, Burlington, Massachusetts. Limnol. and Oceanogr. 26(3):518-531. MeKnight, D. M., and F. M. M. Morel. 1979. Release of weak and strong copper-complexing agents by algae. Limnol. Oceanogr. 24:823-837. McKnight, D. M., and F. M. M. Morel. 1980. Complexation of copper by siderophores from filamentous blue-green algae. Limnol. Oceanogr. 25:62-71. Moore, G. T., and K. F. Kellerman. 1905. Copper as an algieide and disinfectant in water supplies. Bull. Bureau Plant Indus. U.S.D.A. 76:19-55. Morel, F. M. M., N. M. L. Morel, D. M. Anderson, D. M. MeKnight, and J. G. Rueter, Jr. 1979. Trace metal speeiation and toxidty in phytoplankton cultures. In F. S. Jacoff (ed), Advances in

320

D.M. McKnight and others

marine research. US Environmental Protection Agency Environ. Res. Lab., Narraganset, RI. Muchmore, C. B. 1976. Algae control in water supply reservoirs. PB 22-275. Pagenkopf, G. K., R. C. Russo, and R. V. Thurston. 1974. Effect of complexation on toxicity of copper to fishes. J. Fish Res. Bd. Can. 31:462-465. Rueter, J. G., Jr., 1979. The effects of copper and zinc on growth rate and nutrient uptake in the marine diatom Thalassiosira pseudonana. Ph.D. thesis. Mass. Inst. Technol., Cambridge, MA 155 pp. Rueter, J. G., Jr., J. J. MaCarthy, and E. J. Carpenter. 1979. The toxic effect of copper on OsciUatoria (Trichodesmium) theibautii. Limnol. Oceanogr. 24:558-561. Rueter, J. G., Jr., S. W. Chisholm, and F. M. M. Morel. 1981. The effect of copper toxicity on silicic acid uptake and growth in Thalassiosira pseudonana (BaciUariophyceae). J. Phycol. 17:270278. Sanchez, I., and G. F. Lee. 1978. Environmental chemistry of copper in Lake Monona, Wisconsin. War. Res. 12:889-903. Schindler, P. W., B. Furst, R. Dick, and P. U. Wolf. 1979. Ligand properties of surface silanol groups. I. Surface complex formation with Fe3+, Cu 2+, Cd2+, and Pbz+. J. Colloid. Inter. Sci. 55:469475. Shapiro, J., V. Lamarra, and M. Lynch. 1975. Biomanipulation: An ecosystem approach to lake restoratioin. Pages 85-96 in P. L. Brezonik and J. L. Fox (eds) Proceedings of a symposium on water quality management through biological control. University of Florida and US Environmental Protection Agency. Smith, R. G., Jr. 1976. Evaluation of combined applications of ultrafiltration and complexation capacity techniques to natural waters. Anal. Chem. 48:74-76. Steemann-Nielsen, E. and J. Bruun-Larsen. 1976. Effect of CuSO4 on the photosynthetic rate of phytoplankton in four Danish Lakes. Oikos 27:239-242. Steemann-Nielsen, E., L. Kamp Nielsen, and S. Wium Anderson. 1969. Influence of deleterious concentrations of copper on the photosynthesis of Chlorella pyrenoidosa. Physio. Plant. 22:11211133. Steeman Nielsen, E. and S. Wium Anderson. 1970. Copper ions as poison in the sea and freshwater. Mar. Biol. 6:93-97. Stiff, M. J. 1970. Copper/bicarbonate equilibria in solutions of bicarbonate ion at concentrations similar to those found in natural waters. Wat. Res. 5:171-176. Stiff, M. J. 1971. The chemical states of copper in polluted freshwater and a scheme of analysis to differentiate them. Wat. Res. 5:585599. Stokes, P. M., T. C. Hutchinson, and K. Krauter. 1973. Heavy-metal tolerance in algae isolated from contaminated lakes near Sudbury, Ontario. Can. ]. Bot. 51:2155-2168. Stumm, W., and P. Baccini. 1978. Man-made chemical perturbauon of lakes. Pages 91-123 in A. Lerman (ed), Lakes--Chemistry, gcologj, physics. Springer-Verlag New York, NY. Sunda, W. G., and R. R. L. Guillard. 1976. Relationship between cupric ion activity and the toxicity of copper to phytoplankton. J. Mar. Res. 34:511-529.

Sunda, W. G., and P. J. Hanson. 1979. Chemical speciation of copper in river water: effect of total copper, pH, carbonate and dissolved organic matter. In E. A. Jenne ed, Chemical modeling--spedation, sorption, solubility, and kinetics in aqueous systems. Am. Chem. Soc. Envir. Geochem. Health. Sunda, W. G., and J. A. M. Lewis. 1978. Effect of complexation by natural organic ligand on the toxicity of copper to a unicellular alga, Monochrysis lutheri. Limnol. Oceanogr. 23:870-876. Vuceta, J. 1976. Adsorption of Pb(II) and Cu(II) on a-quartz from aqueous solutions: Influence of pH, ionic strength and complexing ligands. Ph.D. thesis. Calif. Inst. Technol., Pasadena, CA. Waiwood, K. G., and F. W. H. Beamish. i978. Effects of copper, pH and hardness on the critical swimming performance of rainbow trout (Salmo gairdneri Richardson). Wat. Res. 12:611-619. Whitaker, J., J. Barica, H. Kling, and M. Buckley. 1978. Efficiency of copper sulfate in the suppression of Apahizomenon flos-aquae blooms in prairie lakes. Environ. Pollut. 15:185-194. ~'Vhitton, B. A. 1973. Freshwater plankton. Pages 353-367 in N. G. Carr and B. A. Whitton (eds) The biology of blue-green algae. Univ. Calif. Bot. Monogr. 9.