RESTORATION AND MANAGEMENT OF LAKES AND RESERVOIRS - CHAPTER 19 pot

Complete circulation mayalso reduce algal biomass by increasing the mixed depth, thereby reducing available light, and bysubjecting mixed algal cells to rapid changes in hydrostatic pres

19 Artificial Circulation

19.1 INTRODUCTION

Artificial circulation, also referred to as destratification, and hypolimnetic aeration/oxygenation(Chapter 18) are two general techniques for aerating lakes Circulation has been achieved by pumps,jets, and diffused air Complete lake circulation is usually the objective, and in the majority of casesexamined either stratification was prevented or destratification occurred Unlike hypolimnetic aer-ation/oxygenation, the temperature of the whole lake is raised with complete circulation; the greatestincrease in temperature occurs at depths that were previously part of the cooler hypolimnion.The principal improvements in water quality caused by complete circulation are oxygenationand chemical oxidation of substances in the entire water column (Pastorak et al., 1981, 1982).Similar to hypolimnetic aeration, its main benefit is enlarging the suitable habitat for aerobicanimals Complete circulation may reduce internal loading of P, if the principal P-release mechanismwas due to iron reduction in anoxic profundal sediments (Chapter 18) Complete circulation mayalso reduce algal biomass by increasing the mixed depth, thereby reducing available light, and bysubjecting mixed algal cells to rapid changes in hydrostatic pressure (Lorenzen and Mitchell, 1975;Fast, 1979; Forsberg and Shapiro, 1980) Although reduced internal P loading and decreasedphytoplankton biomass may be reasonable expectations, other factors such as nutrient availability

in the photic zone, may be more important to P availability, and actually be enhanced withcirculation In some instances, phytoplankton biomass and P content either did not change or wereincreased following circulation

Artificial circulation has been employed as a management technique since at least the early1950s (Hooper et al., 1953) Initially it was used to prevent winter fish kills in shallow, ice-coveredlakes (Halsey, 1968) Although not discussed here, refinements to winterkill prevention wereproposed recently (McCord et al., 2000; Miller et al., 2001; Miller and Mackey, 2003) Nearly all

of the reported applications of the technique to control eutrophication effects and to improve waterquality occurred later than the mid 1960s Complete circulation has been the most frequently usedtechnique to improve water quality (except for algicides and herbicides)

19.2 DEVICES AND AIR QUANTITIES

Introduction of compressed air through a diffuser or perforated pipe located at depth employs theair-lift method of circulating lakes and reservoirs, in which water is welled up by the rising plume

of air bubbles (Pastorak et al., 1981, 1982) Although techniques using pumps and water jets havebeen used successfully to circulate lakes, the air-lift method, through diffusion of compressed air,

is apparently the least expensive and is easiest to operate (Lorenzen and Fast, 1977) However,high efficiencies of oxygenation have been reported from pumped jets in some cases (Stefan and

Gu, 1991; Michele and Michele, 2002)

If the lake is already stratified, mixing is usually achieved only above the depth of air injection

If the lake is not stratified however, injection near the surface can prevent stratification (Pastorak

et al., 1981, 1982) The effect of an unconfined rising plume of air bubbles on water circulation

in an already stratified lake is illustrated in Figure 19.1 As the plume rises, the mixture becomesheavy, upward water flow ceases and the water plume spreads laterally or sinks to a neutral

(10 6 /m 3 )

Area (ha) Q Air/m 3 /min

Q Air/m 3

Max Mean Device

Boltz, KY 18.9 9.4 18.9 3.614 39.0 3.17 a 0.88 8.17 Symons et al., 1967, 1970; Robinson et al., 1969

Falmouth, KY 12.8 6.1 12.8 5.674 91.0 3.26 0.58 3.58 Symons et al., 1967, 1970; Robinson et al., 1969

29.9

55.0

Eufaula, OK 27.0 16.2 27.0 703.1 414.8 × 10 2 33.98 c 0.05 0.06 Leach et al., 1980

Pfaffikersee, Switzerland 35.0 18.0 28.8 56.5 325.0 6.0 b 0.11 1.85 Thomas, 1966; Ambuhl, 1967

Altoona, GA 1968–1969 46.0 9.4 42.7 453 4800 21.6 b –27.7 0.05–0.86 0.45–0.58 USAE, 1973; Raynes, 1975

Kremenchug, Poland 3.0 2.0 2.6 0.002 0.12 4.38 a 1750 3500 Ryabov et al., 1972; Sirenko et al., 1972

Tarago, Australia 23.0 10.5 14.0 27.6 360 3.0 c –9.0 0.08–0.24 0.83–2.50 Bowles et al., 1979

3.0 c –7.50 0.08–0.20 0.83–2.08

Arbuckle, OK 1975; 1977 24.7 9.5 6.0; 2.0 89.3 × 10 2 951.0 Axial-flow pump c Toetz, 1977a, b, 1979

a Flow rate produced destratification.

b Partly mixed.

c Flow rate inadequate to destratify.

d R.A Pastorak, personal communication.

Source: From Pastorak, R.A et al 1981; Pastorak, R.A et al 1982 Tech Rept No E-82-3 U.S Army Corps of Engineers; with additions.

buoyancy level However, the bubbles continue to rise with increased buoyancy having expandeddue to reduced hydrostatic pressure at shallow depth, repeating the water-entrainment process,until they reach the surface Assuming air flow is adequate, the process continues until the densitydifference above the diffuser is zero (Zic and Stefan, 1994; Sahoo and Luketina, 2002) Theoverall effect is that water is pulled from the hypolimnion into the epilimnion, breaking up thethermocline, producing generally homothermous, completely mixed conditions near the plume.

As mixing and entrainment continue, erosion of the thermocline proceeds away from the plume

so long as the energy applied through the airlift system exceeds the energy of resistance due tothermal (density) stability

Injection of compressed air at maximum depth usually affords the greatest rate of mixing,because flow of the entrained water is a function of depth of release and air-flow rate Lorenzenand Fast concluded that an air-flow rate per lake surface area of 9.2 m3/km2 permin(1.33 ft3/acreper min) should provide adequate surface reaeration and other benefits of circulation However,the areal air-flow rates approached or exceeded that critical value in only 42% of the cases cited

in Table 19.1 Effectiveness of that flow rate is substantiated by the cases in Table 19.1 wherebefore and after temperature data were provided (Pastorak et al., 1982) Figure 19.2 is a plot ofthe degree of destratification (percent reduction in Δt in the water column) related to air-flow rateper unit area Except for three observations, areal air-flow rates approaching or exceeding 9.2

m3/km2 perminproduced complete mixing, or 100% decrease in the surface to bottom Δt In two

of the three exception lakes to the right of the line in Figure 19.2, the final Δt was < 3°C, whichwas used as the criterion for satisfactory destratification (Pastorak et al., 1982) In 30 of the 45cases cited for the airlift technique, where temperature data were available, the presented air-flowrates were adequate to destratify or prevent stratification (Table 19.2)

The Lorenzen and Fast areal air-flow rate criterion has been more reliably followed in morerecent commercially installed systems The average areal air-flow rate for 21 systems installed byGeneral Environmental Systems in reservoirs and lakes > 23 ha during 1991–2002 was 7.8 m3/km2

permin (Geney, personal communication) Delivering the air to as much of the deep area of thewater body as possible is also important to attain and maintain destratification (Geney, 1994).The basis for the areal air-flow rate criterion of 9.2 m3/km2 perminis a relationship among air-flow rate, depth, and flow rate of up-welled water above an orifice (Lorenzen and Fast, 1977;Pastorak et al., 1982):

FIGURE 19.1 The process of destratification as a result of entrainment of water by a rising plume of air

bubbles Cooler, hypolimnetic waters from elsewhere replace the volume entrained near the plume, ultimately

eroding away the thermocline (From Davis, J.M 1980 Water Serv 84: 497–504 With permission.)

a range of lake reservoir areas, volumes, and depths, indicated an air-flow rate approaching

or greater than the 9.2 m3/km2 per min level (midpoint of a range 6.1–12.3 m3/Km2 per min)consistently achieved destratification (Table 19.1)

The diffuser should be a pipe with multiple orifices, usually located at the deepest point in thelake, but suspended sufficiently well off the bottom (1 to 2 m) to minimize sediment entrainment.Orifice spacing should be about 0.1 times the depth of air release, because the rising water plumewill spread horizontally at 0.05 m/m of rise (Lorenzen and Fast, 1977)

Another approach to designing an air-lift system to destratify lakes and reservoirs was described

in detail by Davis (1980) This approach requires the following information/steps:

1 Obtain surface area and volume as function of depth

2 Determine or assume temperature or density profile

3 Existing stability and added heat input and theoretical energy required to overcome itare calculated

4 Calculate free air-flow rate at the compressor

5 Calculate perforated (diffuser) pipe length (50 m suggested as minimum)

FIGURE 19.2 Percent destratification, based on surface to bottom temperature differences (Δt) before and

after circulation, related to free air flow (Data from Pastorak, R.A et al 1982 Environmental Aspects ofArtificial Aeration and Oxygenation of Reservoirs: A Review of Theory, Techniques, and Experiences Tech.RepT No E-82-3, U.S Army Corps of Engineers, Vicksburg, MS; from Cooke et al 1993 With permission.)

Free air flow, m3 min −1 km−2

62 114 200

6 Select diffuser pipe and hole diameters (0.8 mm suggested) and hole spacing (0.3 msuggested).

7 Determine internal pipe diameter and air pressure at compressor considering losses due

to hydrostatic pressure, excess pressure at pipe end, friction in the pipe, the pipe bends,valves, etc

8 Recheck diffuser length, considering pressure losses and free air flow through a singlehole

9 Calculate anchor weight

Stability is calculated first, as the difference between the unmixed, existing density gradient,and the mixed condition:

(19.2)

where S = stability, joules (kg m2/s2), g = acceleration due to gravity, m/s2, ρi = density of layer i,

kg/m3, V i = volume of layer i, m3, h i = height of centroid of layer i, m, m = mixed, and s = stratified.

The energy required for destratification is calculated by

The volume of water entrained by the air bubbles from a perforated pipe is recommended to

be 2.5 times the volume of the lake or reservoir to be destratified and can be calculated according to

10 4

.ln

Pastorak et al (1982) compared the calculated flow rates required by the two procedures,using an example from Davis (1980) for a body of water with a volume of 20 × 106 m3, a maximumdepth of 20 m, and an area of 1.2 × 106 m2 The flow rate recommended by the Davis procedurewould be 70 L/s (3.5 m3/km2 permin) By the Lorenzen and Fast (1977) procedure the rate would

be 6 m3/km2 permin, or 120 L/s, nearly twice the Davis rate The rate used here is the the lowerend of the range (6.1 to 12.3 m3/km2 permin) because deeper lakes generally require less air tomix than do shallow lakes (Pastorak, personal communication)

According to Equation 19.6, the diffuser pipe length needed to destratify is inversely related

to air-flow rate Thus, pipe length would be 216 m, based on a 70 L/s air-flow rate and 182 mbased on 120 L/s for destratification to occur in 5 days For the example lake, Davis (1980) selected

a high-density polyethylene pipe of diameter 50.8 mm, perforated with 1-mm diameter holes spaced

at 0.3 m An air pressure of 5.3 bar (5.5 kg/cm2) at the compressor was calculated by summingthe hydrostatic pressure represented by the water depth over the pipe, mean excess pressure abovethe hydrostatic pressure at the end of the pipe (related to pipe length), friction loss in the pipe(related to pipe diameter) and pressure drop from bends in the pipe An air-flow rate of 108 L/swas recalculated for pipe length and pore size and number of holes with that compressor pressure(5.3 bar) That exceeded the calculated 70 L/s so the nominal pipe length of 250 m was consideredadequate A longer pipe length than the minimum calculated facilitates destratification with greaterair distribution These estimates can be obtained from nomographs in Davis (1980)

While calculation of required free air-flow rate at the diffuser end and the initial estimate ofminimum diffuser length to accommodate that rate are relatively straightforward, determining therequired pressure at the compressor, and a more precise estimate of diffuser length incorporatingall the pressure losses, is not straightforward and involves an iterative process (Meyer, 1991).Consistent with the above procedure, first obtain an initial estimate of diffuser length (Equation19.6) Then, determine hydrostatic and internal pipe pressures to obtain a new estimate of free airflow from a single diffuser hole From that air flow and knowing the diffuser hole-spacing andtotal air flow required, a new pipe length can be determined With that pipe length, pressures can

be recalculated and the process repeated until the optimum diffuser length is obtained To simplifythe process, Meyer (1991) incorporated the equations and charts from Davis (1980) into a spread-sheet, which allowed an iterative process of changing variables and formulas to arrive at an optimumdiffuser length Results using the spreadsheet procedure for a hypothetical reservoir are summarized

as follows:

• Surface area: 1,011,750 m2

• Diffuser depth: 10 m

• Volume above diffuser: 10,117,500 m3

• Time to destratify 5 days: 432,000 s

• Temp range from 30°C @ surface to 21.8°C @ 25 m

• Theoretical energy required (E) = stability (S) + solar input (R) (Equation 19.3):

1.9 × 108 J + 0.25 × 108 J = 2.15 × 108 JAir flow required (from Equation 19.4):

1 2 /

Diffuser length, initial calculation (from Equation 19.6):

= 89 m

Selected:

• Supply line: 500 m

• Internal diameter supply line: 45 mm

• Internal diameter diffuser: 35 mm

Through iteration, an optimum diffuser length of 339 m and compressor pressure of 9.7 kg/cm2

(135 psi) were determined

The iterative approach was used to estimate air-flow pressure and diffuser length for East SidneyLake, New York, 85 ha, 15.7 m maximum depth and 4.9 m mean depth (Meyer et al., 1992) Therespective values by using the Davis nomographs were 1.53 m3/min, 3.4 kg/cm2, and 107 m Thoseusing the iterative process were 2.19 m3/min, 3.9 kg/cm2, and 135 m A destratifying time of 5days was used with both procedures

To gain flexibility and control over the long and narrow reservoir, 244 m of total diffuser lengthwas installed, with 8 separate 30-m lines spread through the reservoir A 15 hp compressor wasused to deliver 1.8 m3/min air flow at 3.6 kg/cm2 pressure

The system operated satisfactorily during 1989–1990 to maintain destratified conditions (< 2°Cdifference surface to bottom) in the near field, but temperature difference was greater in the farfield or whole lake, despite the extended lines Also, bottom DO levels dropped below 3 mg/L.Use of a diffuser longer than calculated, i.e., “underloading” the diffuser, may have accounted forand restricted destratification capacity However, the total air delivery per area to the reservoir,which was 2.1 m3/km2 permin, relative to the Lorenzen and Fast criterion, was not discussed Thatrate for East Sydney Lake was well below their median criterion and probably accounted for some

of the less-than-expected water quality response, discussed later in this chapter While a successfuloutcome for complete circulation depends on the size and length of diffuser pipes, results indicatethat for best results in improvement of water quality, as well as achieving destratification, adherence

to the Lorenzen and Fast criterion is also advisable

Mechanical mixing devices have been used less frequently than compressed air (Table 19.1).Two types of pumps have been developed for destratifying reservoirs: (1) axial-flow pumpswith a large propeller (6 to 15 ft diameter) that generates a low velocity jet (Punnet, 1991),and (2) direct drive mixer with a small propeller (1 to 2-ft diameter) that generates a highvelocity jet (Stefan and Gu, 1991; Price, 1988, 1989) Design of a pumping system to destratify

a lake or reservoir depends on the desired time to destratify (or rates of circulation) and depth

of hydraulic jet penetration Time to destratify in turn depends on the degree of stratification

s ln 1+10 m10.4

or resistance to mixing The number of pumps needed to achieve a given depth of penetrationand time to mixing can be calculated (Holland, 1984; Gu and Stephan, 1988; Stefan and Gu,1991) Destratification was complete (Δt < 3°C) for 4 of the 10 cases for pumps and jets cited

in Table 19.1

Mixing devices powered by solar and wind energy are available commercially, but published results

of effectiveness were unavailable for inclusion here

19.3 THEORETICAL EFFECTS OF CIRCULATION

19.3.1 DISSOLVED OXYGEN (DO)

The principal, and probably the most reliable, effect of circulation is to raise the dissolved oxygen(DO) content throughout the lake over time If the lake is destratified, the DO content in what wasthe hypolimnion will increase, and that in the epilimnion will decrease, at least at first This canoccur from simple dilution Additional reasons why the surface water DO may decrease are thetransfer of oxygen-demanding substances toward the surface and a decrease in photosynthesis inthe photic zone due to increased mixing depths (Haynes, 1973; Ridley et al., 1966; Thomas, 1966)

DO will continue to increase as circulation is maintained, largely because water undersaturatedwith oxygen is brought into contact with the air While the vertical transport of water is achieved

by entraining water through releasing compressed air at some depth, little oxygen increase isachieved through direct diffusion from bubbles (King, 1970; Smith et al 1975)

19.3.2 NUTRIENTS

Internal loading of P theoretically can be decreased through increased circulation This would occur

in situations where the dominant mechanism of P release was from iron-bound P in anoxichypolimnetic sediments By aerating the sediment-water interface of lakes where iron is controlling

P solubility, P should be adsorbed from solution by ferric-hydroxy complexes (Mortimer, 1941,1971; Stumm and Leckie, 1971; Chapters 8, 18, 20) Thus P would be prevented from migratingfrom high concentrations in sediment interstitial water to the overlying water Calcium may control

P solubility in hardwater lakes, rather than iron, or the iron/phosphorus ratio may be too low tocontrol P release (Jensen et al., 1992), in which case the release rate could be due largely to afunction of aerobic decomposition of organic matter (Kamp-Nielsen, 1975) In that event, internal

P loading may actually increase as temperature at the sediment-water interface is raised in thecirculation process Also, some sediments with a low Fe:P ratio have a high organic and watercontent and are very flocculent, and may have a high loosely bound P fraction (Boström, 1984)

In that latter situation as well, internal loading could actually increase from such sedimentsfollowing circulation P exchange rates are dependent upon circulation at the sediment-waterinterface and that process could be enhanced by mixing (Lee, 1970) Degree of wind mixing had

a dominant effect on summer internal loading of P in shallow Moses Lake, Washington (Jones andWelch, 1990)

Internal loading of P may be high in unstratified, shallow, eutrophic lakes in which the water interface is usually oxic (Jacoby et al., 1982; Kamp-Nielsen, 1975; Søndergaard et al., 1999).Therefore, reduced internal P loading probably cannot be expected to result from artificial circu-lation Internal loading and whole-lake TP may decrease in shallow stratified lakes followingcirculation (Ashby et al., 1991), but the concentration available for growth in the photic zone mayincrease, as has been observed (Brosnan and Cooke, 1987; Osgood and Stiegler 1990) Thus, depth

sediment-is an important criterion in determining the candidacy of shallow lakes for complete circulationfrom not only phytoplankton production related to available light, but also internal P loading Unlessoxic conditions will substantially reduce P internal loading, maintaining stratified conditions may

be preferable for limiting P availability in the photic zone

Other potential changes in chemical content resulting from complete circulation are the version of ammonium to nitrate and the complexation and sedimentation of trace metals such asmanganese and iron Ammonium decrease can largely be attributed to increased nitrification, whichrequires aerobic conditions (Brezonik et al., 1969; Toetz, 1979) This effect will be greater thelonger that duration and completeness of hypolimnetic deoxygenation proceeded prior to circula-tion The decrease in trace metals like manganese and iron should also be greater in lakes withlarger oxygen deficits prior to aeration increases Because these metals diffuse from the sediment

con-in their reduced, soluble forms, aeration will promote their oxidation and subsequent complexationand precipitation This can be an important benefit in lakes used for drinking water supplies

19.3.3 PHYSICAL CONTROL OF PHYTOPLANKTON BIOMASS

Circulation can reduce phytoplankton biomass through light limitation, brought about by providing

a greater depth of mixing of plankton cells in the water column so that the total light received duringtheir brief period in the photic zone is insufficient for net photosynthesis (photosynthesis in excess

of respiration) and thus any growth or increase in cell mass This is known as the “critical depth”concept, first formulated to predict the timing of the spring diatom bloom in the ocean (Sverdrup,1953) By knowing light at the surface, compensation depth, and the extinction coefficient, the criticaldepth can be calculated as the point above which net production is possible; when that calculateddepth exceeds the mixed-layer depth, a bloom can occur This model is dependent upon somerelationship between light intensity and gross photosynthesis, assuming a constant rate of respiration.The same concept applies in lakes (Talling, 1971) The combination of low surface light intensityand deep mixing prevented net photosynthesis during winter in relatively deeper lakes (> 30 m) ofthe English Lake District, but not in the shallower lakes (10 m) Growth rate during the springphytoplankton maximum was directly related to light intensity in a long-term data series (Neale etal., 1991) Normally, lakes are shallow enough to allow some net photosynthesis even in winter,but decreasing mixing depth, as stratification develops and surface light intensity increases in thespring, usually accounts for the large increase in net photosynthesis and the spring diatom bloom

in deeper lakes

Light can limit maximum phytoplankton biomass even in shallow eutrophic lakes (Sheffer,1998) A 35-year data base from Lake Võrtsjär (270 km2, mean depth 2.8 m), Estonia, showed thatthe water level change produced a 2.5 times difference in mean depth resulting in biomass levelssignificantly lower in high water level years (Nõges and Nõges, 1999; Nõges et al., 2003) Thus,artificial circulation may produce light-limiting benefits in shallow, eutrophic lakes with normallyhigh particulate matter concentrations and light extinction

The concept of physical control of phytoplankton growth was extended to the effects of artificialcirculation in eutrophic lakes (Lorenzen and Mitchell, 1975; Murphy, 1962; Oskam, 1978) Forsbergand Shapiro (1980) and Shapiro et al (1982) integrated the effects of nutrients with those of physicalfactors By increasing the depth of mixing, a lake potentially can be returned to a winter conditionwhere light is limiting, assuming maximum depth and light attenuation are sufficient Increasingmixing depth would not be great enough in most cases to prevent net biomass production completely,which is not expected This effect of mixing depth is clearly shown in results from Kezar Lake(Figure 19.3; Lorenzen and Mitchell, 1975) Increased mixing depth though complete circulation

is expected to substantially reduce algal biomass due to light limitation alone However, nutrientsmay initially be limiting in the epilimnion, so that a slight increase in mixed depth may entrainwater with higher nutrient content from below and biomass may increase (point A to point B inFigure 19.3) At some point light will limit and productivity and biomass will decrease (point C

to point D) Note that biomass is plotted as mass per area (g/m2), which was expected to decrease

by only 38% for a mixing-depth increase of 2 to 6 m Biomass concentration (g/m3), however, wasexpected to decrease by 80%, which would also include the effect of water column dilution Thismodel predicted only the potential productivity without nutrient limitation and included no losses

from sinking, grazing, parasitism, or washout Actual values may therefore fall below the line inFigure 19.3, as was the case for Kezar Lake.

Little change in biomass may occur in oligotrophic lakes following circulation, because theslope of the ascending line in Figure 19.3 (nutrient limitation) would be less for such lakes (Pastorak

et al., 1981, 1982) Because that line represents the maximum nutrient-limited biomass, anydisplacement of the biomass vertically by circulation would bring about a smaller change in biomassconcentration in oligotrophic than in eutrophic lakes That was not the case in experiments in deepplastic bags in an oligotrophic lake (15 μg/L TP) in which biomass increased with mixing depth

up to 15 m so long as background turbidity was low (Diehl et al., 2002) Results verified thehypothesis that increased mixing depth reduces growth rate, but at the same time reduces cell andnutrient loss However, light attenuation should be greater and nutrient conservation less importantunder eutrophic conditions, as was demonstrated with increased background turbidity; i.e., biomassdecreased beyond a mixing depth of 6 m

Oskam (1973, 1978) developed a model to express the effect of mixing-depth change on

productivity and maximum biomass Because net productivity (Pnet, mg C/m2 per day) is thedifference between gross productivity and respiration in the mixed layer, the following equationshould hold:

(19.7)

where C = chlorophyll (chl) a in mg/m3, Pmax = maximum photosynthetic rate in mgC/mg chl per

hour, F(i) = dimensionless function of light intensity (expands Pmax to total areal rate), λ = daylighthours, εw = extinction for water in 1/m, εc = specific extinction coefficient per unit algae in m2/mg

chl, Z m = depth of mixing in m, 24 = 24 h/d, r = respiration/Pmax

According to this equation, as the depth of mixing increases, assuming uniform distribution ofalgae, net productivity decreases The mixing depth can be increased by artificial circulation Critical

depth can be calculated without knowing Pmax by setting Pnet = 0 and solving for Z m:

FIGURE 19.3 Theoretical and observed peak biomass of algae in Kezar Lake (see text for explanation of points

A–D) Solid circles: theoretical values; solid square: 1968, stratified; triangle: 1969, destratified; open square:

1970, destratified (From Lorenzen, M.W and R Mitchell 1975 J Am Water Works Assoc 67: 373–376 With permission; Pastorak, R.A et al 1981 Evaluation of Aeration/Circulation as a Lake Restoration

Technique 600/3-81-014 USEPA; Pastorak, R.A et al 1982 Environmental Aspects of Artificial Aeration

and Oxygenation of Reservoirs: A Review of Theory, Techniques, and Experiences Tech Rept No E-82-3.U.S Army Corps of Engineers, Vicksburg, MS.)

2 60 50

ε ε 24

Maximum biomass (mg chl/m3) can also be estimated from Equation 19.7) as a function of

mixing depth by setting Pnet = 0, and solving for Cmax:

maximum would be correspondingly less These relationships show the sensitivity of potentialmaximum biomass to mixing depth in shallow lakes and may offer a first approximation of thefeasibility for circulation to reduce algae in a particular, non-nutrient-limited lake

Forsberg and Shapiro (1980) and Shapiro et al (1982) developed an expanded model to includenutrient limitation and losses Their equation for maximum biomass in the mixed layer is

The basis for the nutrient effect in Equation 19.10 is an expression of cell nutrient quota, which

is approximated by the ratio of TP to chl a:

FIGURE 19.4 Relation of maximum chlorophyll concentration to mixing depth for different levels of nonalgal

attenuation of light (From Oskam, G 1978 Verh Int Verein Limnol 20: 1612–1618 With permission.)

ε θ ))Pmaxsat Kq] /TP

Pmaxsat

where Pmax = maximum specific daily rate of photosynthesis at saturating nutrient level and Kq′ =

minimum ratio of TP/chl a required for photosynthesis to occur (1.8 in Forsberg and Shapiro, 1980)

The relationships between maximum biomass (chl) per unit volume and per unit area in themixed layer, and the depth of mixing, based on this model, are shown in Figure 19.5 Clearly, theconcentration of limiting nutrient determined the maximum biomass at any depth of mixing This

is an important point, and should be considered with predictions of improvements followingcirculation because of the great potential for increasing the nutrient available to algae followingdestratification Of course, if nutrient content is relatively high already and increases do not occurwith mixing, then biomass concentration should decrease, with the greatest decrease occurring atmixing depths less than 10 m (Figure 19.5a)

As Pastorak et al (1982) indicate, there are several problems with application of this model.The most serious would appear to be difficulties in estimating loss rates, which would decreasewith mixing depth increase (Diehl et al., 2002), as well as the effects of shifts in species composition

FIGURE 19.5 The effect of changes in the mixed depth and TP on: (a) the maximum concentration of chl a

and (b) the maximum aerial standing crop of chl a in the mixed layer of Twin Lake, Minnesota, as predicted

by the model (closed circles indicate observations and connecting lines indicate the deviation between predicted

and observed results) (From Shapiro, J et al 1982 Experiments and Experiences in Biomanipulation —

Studies of Biological Ways to Reduce Algal Abundance and Eliminate Blue Greens USEPA-600/3-82–096.)

(see below) and the nutrient history on the growth-rate response of algae Nevertheless, rathergood agreement between the model predictions and experimental results were observed (Figure19.5; Forsberg and Shapiro, 1980) The low level of complexity in this model makes it appealing

as a tool to guide the application of the circulation technique However, a separate prediction for

TP is necessary

19.3.4 EFFECTS ON PHYTOPLANKTON COMPOSITION

There are several hypotheses to explain the dominance of blue green algae (cyanobacteria) ineutrophic lakes (Welch and Jacoby, 2004) There are three that may explain a shift from dominance

by bloom-forming blue-greens to dominance by more desirable diatoms or green algae as a result

of complete circulation These involve changes in (1) CO2 and pH, (2) distribution of buoyant cells,and (3) grazing by zooplankton, all of which could be results from increased circulation

Blue-green algae-dominated cultures shifted to dominance by green algae in response todecreased pH and associated increases in free CO2 concentration (King, 1970, 1972; Shapiro, 1973,

1984, 1990; Shapiro and Pfannkuch, 1973; Shapiro et al., 1975) Blue-greens apparently absorb

CO2 at lower concentrations, compared to green algae, giving them an advantage at higher pH.Green algae may have a competitive advantage over blue-greens with respect to nutrients at lower

pH The observed rapid die-off of blue-greens following pH decrease, however, may have beencaused by lysing of the blue-greens by viruses that were favored by low pH (Shapiro et al., 1982).King introduced the CO2 hypothesis based on comparisons of algal populations and chemicalconditions existing in sewage lagoons, and suggested that the potential for lakes to promote blue-green dominance increases as alkalinity (buffering capacity) decreases at any given P loading.Shapiro (1984, 1990) was able to shift dominance from blue-greens to greens in bag experiments

in situ with either HCl or CO2 addition, but the shift was more complete if nutrient additions werealso included

Increased circulation can cause CO2 to increase and pH to decrease in the euphotic zone byvertical transport of bottom water, in which CO2 content is high due to respiration in the absence

of photosynthesis, as well as by increased contact with the atmosphere For circulation to promotethe shift from blue-greens, the surface waters should not be nutrient-limited, because high content

of N and P also exists in bottom water, that, with vertical entrainment, could increase blue-greenbiomass already present

A large-scale experiment in Squaw Lake, Wisconsin during summer 1993 produced some doubtabout the role of the CO2/pH hypothesis (Shapiro, 1997) The lake is naturally divided into twobasins; south (9.1 ha, 2.55 m mean depth) and north (16.8 ha, 2.92 m mean depth) The south basinwas artificially circulated and enriched with CO2 The pH exceeded 10 in the north basin, butremained steady at around 7 in the enriched south basin during circulation Despite the contrastingpH/CO2 condition, populations of Aphanizomenon and Anabaena reached levels exceeding 300 μg/L chl a in both basins.

Because the blue-green algae have gas vacuoles that permit buoyancy, the increased stabilitybrought about by thermal stratification and calm weather will allow blue-greens to produce surface

“scums” and a decreased light environment for non-buoyant algae Therefore, increased circulationcan favor non-buoyant algae, which otherwise tend to sink rapidly (especially diatoms) under stable

conditions Anabaena succeeded Tabellaria as thermal stratification developed in summer, although

specific growth rate of the blue-green was not different from that of a spring dominating diatom(Knoechel and Kalff, 1975) This change was explained as a physical effect based on sinking-ratedifference, rather than one based on growth-rate differences related to nutrient changes Hadstratification been prevented by artificial circulation the diatom may have persisted Artificialdestratification in a Thames River reservoir in late July promoted a second bloom of the diatom

Asterionella, which had previously bloomed in the spring and had subsequently declined (Taylor, 1966) Such a decline in Asterionella has been attributed to the combination of nutrient limitation