6 Dilution and Flushing 6.1 INTRODUCTION Dilution and flushing can achieve improved quality in eutrophic lakes by reducing the concentration of limiting nutrient (dilution) and by increasing the water exchange (flushing) rate. Both processes can reduce the biomass of plankton algae, by reducing the inflow concentration of limiting nutrient, resulting in a decreased lake concentration, on which maximum biomass depends. By increasing the water input the flushing rate is increased, which in turn increases the loss rate of plankton algae from the lake. Dilution can be effective even when the increase in flushing rate is insufficient to cause a significant loss of algae. On the other hand, flushing rate increase can cause a significant loss without achieving a reduction in the limiting nutrient concentration. Other effects of dilution are also possible, such as increased vertical mixing and a decrease in the concentration of algal excretory products, which can influence the kinds and abundance of algae (Keating, 1977). Dilution is usually feasible only where large quantities of low-nutrient water are available. Treatment effectiveness is greatest when dilution water is low in limiting nutrient concentration relative to that in the lake and its natural inflow. Lake nutrient concentration can be more effectively lowered if dilution water is the dominant inflow. In some instances, improvements can be achieved by adding water with moderate to high nutrient content, but these results are less certain than with low-nutrient water, largely because lowering the lake concentration is more effective at reducing biomass than washout. Dilution and flushing have worked successfully in several lakes. Green Lake in Seattle, Wash- ington, was improved markedly by adding city water, beginning in the 1960s (Oglesby, 1969). Moses Lake in eastern Washington has received Columbia River dilution water on a regular basis since 1977, resulting in substantial improvement in lake quality (Welch and Patmont, 1980; Welch and Weir , 1987; Welch et al., 1989, 1992). Dilution was instituted in three other lakes in Washington State and was proposed for Clear Lake, California (Goldman, 1968). Lake Bled, Yugoslavia, was flushed intentionally with water from the River Radovna (Sketelj and Rejic, 1966). Lakes Veluwe and Donten in The Netherlands have been diluted with relatively low-P water during winter since 1979 (Hosper, 1985; Hosper and Meyer, 1986). Snake Lake, Wisconsin, was diluted by removing the equivalent of three lake volumes, which allowed low nutrient groundwater to refill the small seepage lake (5 ha, mean depth 2.3 m) (Born et al., 1973). In what must be one of the world’s first lake flushing experiments, water was diverted from Switzerland’s Ruess River to Rotsee in 1921 to 1922 to alleviate eutrophic conditions (Stadelman, 1980). The flushing rate of this 460-ha lake was increased from 0.33 to 2.5/yr (or about 0.1 to 0.7%/d) following construction of a canal between the Ruess River and the lake. The lake’s state did not improve because the increased flushing rate was insufficient to significantly washout biomass, and because of high concentrations of nutrients in the river water used for flushing. The nutrients originated in sewage effluent from the upstream city of Luzern. There was still no improvement following diversion of direct inputs of sewage effluent to the lake in 1933. There was considerable improvement, however, after nutrient removal from Luzern’s wastewater, in the 1970s, resulting in a tenfold P reduction in the Ruess River inflow water. Following is a discussion of the theoretical basis for the dilution and flushing, reviews of the Moses Lake, Green Lake, and Lake Veluwe cases and general guidelines for application of the Copyright © 2005 by Taylor & Francis technique. The latter includes quantity and quality of water, frequency of application, and project and operating costs. 6.2 THEORY AND PREDICTIONS Maintenance of low phytoplankton concentrations by high natural rates of dilution and flushing is a commonly observed phenomenon (Dickman, 1969; Dillon, 1975; Welch, 1969). The mechanisms involved in the control of nutrients and/or algal biomass in lakes are in many ways analogous to those in continuous culture systems. When low-nutrient dilution water is added to a laboratory continuous algal culture, the inflow concentration of limiting nutrient is reduced, the maximum biomass concentration possible in the reactor vessel is likewise reduced, and at the same time nutrients and algal biomass are more rapidly washed from the reactor vessel since the water exchange rate is increased. Concentration of limiting nutrient is the critical variable that determines algal biomass in many lakes, as well as in continuous culture systems. The in-lake nutrient concentration may sometimes be more, but is usually less than the inflow concentration, because sedimentation is greater than internal loading. However, increased rates of dilution/flushing will theoretically reduce loss through sedimentation, as demonstrated by predictions using the Vollen- weider equation (Figure 3.8), where at short detention times, the lake P concentration equals the inflow concentration. That situation is most typical of reservoirs, which tend to have shorter detention times than lakes and would, therefore, behave more like a continuous culture system with algae (McBride and Pridmore, 1988). To predict the response of a lake to the addition of low-nutrient water on a day-to-day or year- to-year basis, without consideration of sedimentation or internal loading, the following “dilution only” first order, integrated equation is useful: (6.1) where C 1 is the concentration at time t; C i is the concentration in the inflow water; C o is the initial lake concentration; and ρ is the water exchange or flushing rate. This equation assumes the lake is well mixed, that no other nutrient sources exist, and that the limiting nutrient or “percent lake water” can be treated as conservative. Because this equation does not include a sedimentation term, it is normally useful only in the short term as a tracer for nutrient behavior and under conditions of rather large water exchange rates (several percent per day or more). However, in some instances (e.g., Lake Norrviken, Figure 4.6) the response of lake nutrient concentration could follow a simple dilution model if there were no retention of nutrient in sediments, i.e., sediment release equals sediment uptake. In most instances, however, it could be used to estimate the potential for reducing average lake concentrations with a given source of water, and the time necessary for that reduction. Such predictions can be compared with the observed distribution of dilution water — as percent lake water — indicated by a conservative variable, such as sodium or specific conductance. For more realistic predictions, sediment–water interchange of nutrients must be considered. Equations given in Chapter 3 are pertinent in that regard where flushing rate and external loading terms are modified for the added dilution water. Increased flushing has an indirect effect on lake P concentration, as shown with a Vollenweider steady state, mass balance P model. Adding more water with lower nutrient content also increases nutrient loading, while the resulting increased flushing rate decreases nutrient loss through sedi- mentation (Uttormark and Hutchins, 1980). These processes could be counteracting the dilution effect in some instances, because, as the authors stated, “a reduction in the influent concentration tends to reduce in-lake concentration, but a reduction in phosphorus retention tends to increase in- lake concentration.” They showed that a large increase in the combined flushing rate obtained by adding low-nutrient water (40% of the normal inflow nutrient content) could theoretically increase CC CCe ioi t 1 =+ − − () ρ Copyright © 2005 by Taylor & Francis the lake nutrient concentration if the original flushing rate ρ is low enough, e.g., 0.1/yr. If the flushing rate is relatively large (≥ 1.0/yr) initially, the effect of reduced sedimentation rate is minimized and a reduction in lake concentration will result, but large quantities of water are necessary. Of course, the amount of water needed to achieve a given reduction in inflow concen- tration is a function of the concentration difference between the normal inflow and dilution water source. In an actual case, the effect of diminished sedimentation was not significant, but the effect of dilution was reduced because of enhanced sediment release of P (Jones and Welch, 1990). Additions of dilution water (plus normal inflow) to Parker Horn of Moses Lake, Washington resulted in average flushing rates of 8%/d from April to September during 12 years of treatment. Based on a calibrated and verified steady state, mass balance P model, where net internal loading of P was indicated by a negative sedimentation rate coefficient (Equation 3.10), increased dilution water input predicted a much more gradual reduction in lake P concentration than was expected from the diluted inflow concentration alone (Figure 6.1). The minimizing effect of dilution water could not have been caused by reduced sedimentation, because the lower curve of inflow concentration is the highest possible lake concentration with no sedimentation. 6.3 CASE STUDIES Two lakes where the dilution and flushing technique was implemented can be used as guides for application elsewhere. Moses Lake lies in eastern Washington, and has a surface area of 2,753 ha and mean depth of 5.6 m. Dilution water has been added to one arm of Moses Lake during spring and early summer since 1977. Transport to previously undiluted portions of the lake, by pumping, was begun in 1982. Green Lake, in Seattle, has an area of 104 ha and a mean depth of 3.8 m. The lake received dilution water from the city domestic supply at relatively high rates from 1962 through the mid 1970s, but inputs were subsequently variable resulting in worsening lake quality. The effect of low-nutrient water was to dilute internal loading, the primary cause for summer algal blooms. The cost and effectiveness of other, more reliable dilution-water sources and other controls on internal loading were evaluated and alum (Chapter 8) was implemented in 1991 (URS, 1983, 1987; Jacoby et al., 1994). The suitability of dilution water sources for treating these lakes is apparent FIGURE 6.1 Predicted and observed (solid circles) column-weighted mean TP as a function of dilution water input in Lower Parker Horn (7) during May–September for the 9-year average TP loading, flushing rate, and range of RTR values observed. (From Jones, C.A. and E.B. Welch. 1990. J. Water Pollut. Cont. Fed. 62: 847–852. With permission.) 130 120 110 100 90 80 70 60 50 20 40 30 TP (μg/l) 2 6 0 2 4 0 2 2 0 2 0 0 1 8 0 1 6 0 1 4 0 Dilution inputs (m ^ 3 * 10 ^ 6) 1 2 0 1 0 0 8 0 6 0 4 0 2 0 0 0 Upper limit Upper limit Inflow Mean Copyright © 2005 by Taylor & Francis from the large ratios of nutrient content in the lakes relative to the dilution inflows, ranging from 5:1 to 10:1. A third case to be discussed is LakeVeluwe in The Netherlands (Hosper, 1985; Hosper and Meyer, 1986), where dilution was used together with wastewater P removal. 6.3.1 MOSES LAKE Dilution water from the Columbia River has been added to Moses Lake’s Parker Horn via the U.S. Bureau of Reclamation’s East Low Canal and Rocky Coulee Wasteway (Figure 6.2). The pumping of dilution water from Parker Horn to the previously undiluted Pelican Horn began in 1982, and sewage effluent was diverted from Pelican Horn in 1984. The effects of dilution water only on lake quality can be evaluated for Parker Horn for all years. Because South Lake was affected to some extent by sewage effluent, the added improvement for that area after 1984 was partly due to diversion. The principal effects in Pelican Horn, however, were from sewage diversion. The ashfall from the eruption of Mount St. Helens reduced the lake’s internal P loading during 1980 to 1981, so effects of dilution are obscured during those years when dilution water additions were also low. Therefore, periods of evaluation are for 1977 to 1979 and 1986 to 1988 in Parker Horn (dilution only) and 1977 to 1979 (dilution only) and 1986 to 1988 (dilution plus sewage diversion) in South Lake. The patterns of dilution water addition have been varied but not in a systematic way conducive to determining optimum quantity and seasonal distribution. The average amount of dilution water added from 1977 through 1988 was 169.4 × 10 6 m 3 /yr, which represented a flushing rate in Parker Horn of 17%/d for the 971 d of actual inflow. The average input for April to September (includes days with and without input) was 130 × 10 6 m 3 /yr or 5.8%/d. With dilution water plus the normal input, the flushing rate averaged 7.8%/d for Parker Horn. For the whole lake, these inputs repre- sented a flushing rate of less than 1%/d. Thus, dilution water input created flushing rates in Parker Horn that could have caused some washout of algal cells, but such an effect would not have been significant in the remainder of the lake. Dilution water addition continued at a slightly higher rate during the 1990s, with input averaging 221 × 10 6 m 3 /yr over the subsequent 13-year period through 2001 — a 30% increase over the previous 12 years. Inputs during the wet years of 1996 and 1997 were only 75 and 32 × 10 6 m 3 /yr. So while dilution has not been constant from year-to-year or evenly distributed in time, water was delivered every year except one during the past 25 years. Nevertheless, low dilution water years (e.g., 1997) still produced large algal blooms and poor lake quality. Columbia River water was nearly ideal for dilution (Table 6.1). Because the P and N concen- trations in Crab Creek (Parker Horn’s natural inflow) were so high, apparently due to irrigation and fertilization practices in the watershed, relatively large quantities of Columbia River water were needed to significantly lower the composite inflow concentration, and thus to lower the in-lake concentration. This resulted in larger exchange rates than would otherwise have been necessary without the Crab Creek inflow. Unfortunately, diverting Crab Creek was not economically feasible in this case, but such a manipulation could be considered for other lakes to obtain more efficiency from dilution water quantities. The addition of dilution water to Moses Lake predictably and rapidly replaced lake water, as judged by tracing a conservative parameter, specific conductance. Values for percent lake water were calculated, assuming that 100% was represented by the conductance of Crab Creek and 0% by the conductance of Columbia River water. For example, using a typical value of 460 μmhos/cm for Crab Creek (CCW), 250 μmhos/cm for lake water (LW), and 120 μmhos/cm for East Low Canal dilution water (ECDW), the percent lake water would be (6.2) 100 100 250 120 46 ()/( ) ()/( LW ECDW CCW ECDW LW−−= − % 00 120 38−=) Copyright © 2005 by Taylor & Francis Percent lake water, which decreases with dilution, is used instead of percent dilution water, to represent the behavior of nutrients in the lake. Remaining lake water in Parker Horn, where water enters (Figure 6.2), was reduced to values around 20%, much less than in other parts of the lake (Figure 6.3). This was expected (dashed line, for Equation 6.1), because the average dilution rate during the period from April to June, described here, was 15%/d for Parker Horn, which is a small (8%) portion of the lake volume. The dilution rate decreased, as the water moved through other parts of the lake. As the dilution water input declined in June, the fraction of lake water in Parker FIGURE 6.2 Moses Lake, Washington, Note source of dilution water from East Low Canal via Rocky Coulee Wasteway and point of pumped transfer of dilution water from Parker Horn to Pelican Horn. (From Cooke, et al. 1993. With permission.) TABLE 6.1 Nutrient Concentration (μg/L) in Inflow Water to Parker Horn from May to September, 1977 and 1978 Total P Total N SRP NO 3 –N Crab Creek inflow without dilution 148 1331 90 1096 East Low Canal dilution water 25 305 8 19 Source: From Cooke et al., 1993. With permission. Present dilution input from East Low Canal Rocky Coulee Wasteway Scale 1 km Scale 4 km Main Arm Lewis Horn Moses Lake State Park Parker Horn City of Moses Lake Pelican Horn Lower Lake Lake outlets I-90 Dilution water transfer Copyright © 2005 by Taylor & Francis Horn quickly rose to between 50% and 60% (Figure 6.3). Part of that increase was probably caused by wind pushing water from the Main Arm and South Lake into Parker Horn. Moses Lake is dendritic in shape and most of the lake’s volume (63%) is in the Main Arm, out of a direct path from the dilution water inflow. Thus, dilution water was expected to have little effect in the main arm compared to Parker Horn and South Lake, which together represent 29% of the lake volume. However, the lake water residual decreased similarly to that for the lower lake, even if the whole lake volume was used to calculate residual lake water and water exchange rate (solid line and Equation 6.1). Lake water residuals in the whole lake and lower lake reached levels between 50% and 60% in late May and early June and then began a more gradual return to normal as dilution input declined (Figure 6.3). Improvement in lake quality during the first three years (1977 to 1979), compared to predilution years 1969 to 1970, was near or in excess of 50% for TP, SRP, and chlorophyll (chl) a. Secchi transparency increased markedly, not only for Parker Horn but for South Lake as well (Table 6.2). Total N decreased by about the same magnitude, although predilution data for N are incomplete. By 1986 through 1988, lake quality improved even more. The further improvement in Parker Horn was due to reduced inflow P (50%) in Crab Creek during the 12-year study period. The further improvement in most of the lake volume was due partly to sewage diversion, which primarily affected South Lake (Table 6.2). Pelican Horn, which was influenced entirely by groundwater and sewage effluent, showed little effect of dilution water in 1977 to 1979. The extensive improvement there in 1986 to 1988 (Table 6.2) was due primarily to sewage diversion. While dilution water was distributed throughout the lake, improvement in quality was greatest in Parker Horn, where the fraction of dilution water was highest (Figure 6.3). Wind was probably the main force causing the transport of dilution water from Parker Horn into the Main Arm, with the fraction of dilution water, which reached half the distance through the main arm, being dependent largely upon the fraction existing in Parker Horn (Welch et al., 1982). About half the natural surface inflow and P load entered the Main Arm via Rocky Ford Creek. Phosphorus concentration in that source did not decline during the 12-year study, as it had in Crab Creek. Thus, the continued trend in quality improvement was not as apparent in the Main Arm. Means for the May to September periods obscure the extreme conditions, such as a maximum Secchi transparency of 3 m in June (4 m was reached in 1982) throughout most of the lake. Chl a reached peaks near 50 μg/L in late July and August after dilution water input was curtailed for two to four weeks. Unless dilution water was added continually, blooms returned as the fraction of dilution water in the lake declined. This “boom and bust” phenomenon, promoted by large inputs followed by no input at all, did not produce the optimum effects that would have occurred with a continual input at low rates throughout the summer, while employing similar total amounts of water. The large quantities added over a short period of time — exchanging water in Parker Horn at the rate of about 20%/d and in most of the lake at 2–3%/d — are probably unnecessary, considering that induced flushing rates throughout most of the lake were insufficient to cause significant washout of algal cells and most improvement was due to dilution. Determining the optimum quantity of dilution water and its distribution over time requires defining the cause(s) for the quality improvement, in this case algal reduction. The reduction of N concentration was the most probable cause for the dilution effect on algal biomass following the initiation of dilution and prior to the Mount St. Helens ashfall. Nitrate, rather than SRP, was the nutrient that most frequently limited algal growth rate (Welch et al., 1972). Nitrate per se was not appreciably reduced by dilution, because it was limiting and remained rather low in the lake water during summer both before and after dilution. Although control of biomass was discernable at total N concentrations in lake water below about 600 μg/L (Welch and Tomasek, 1981), the best relationship was found between flow weighted NO 3 concentration in the inflow (Crab Creek) and the average chl a concentration in Parker Horn and South Lake (r = 0.97; Welch et al., 1984). Controlling algal biomass by reducing the inflow concentration of limiting nutrient, which in its Copyright © 2005 by Taylor & Francis soluble, available form continues to remain at low concentration in the lake (seemingly unrelated to biomass), is analogous to the functioning of a continuous culture system (see Welch, 1992). Prior to implementation of this project, it was thought that TP would be the most important nutrient to control in order to reduce algal biomass. Although NO 3 limited growth rate during summer, fixation of atmospheric N by blue-green algae should have supplied enough N to com- plement the available P supply. N fixation is not a rapid process; maximum rates of cell N replacement and growth of about 5% and 10%/d have been reported (Horne and Goldman, 1972; Horne and Viner, 1971). A growth rate by N fixation of only 2.4 ± 1.8%/d was determined for Moses Lake algae (Brenner, 1983). With such slow growth, increasing the flushing rate by about tenfold may well have prevented N uptake by fixation from fully utilizing the available P. In any event, there was a close relationship between inflow NO 3 and chl a. Using data before and three years after dilution started (1977 to 1979), that relationship allowed estimates of optimum dilution water input. The Mount St. Helens ashfall produced a seal over Moses Lake sediments and stopped internal P loading for two years (Welch et al. 1985; Jones and Welch, 1990). That event, coupled with the continued reduction in inflow P concentration to Parker Horn, resulted in a trend toward P limitation (Welch et al., 1989, 1992). Sewage diversion from Pelican Horn also contributed to that trend in the lower lake. For 1986–1988 when P was limiting, the Jones and Bachmann (1976) relationship provided an adequate fit of the data for predicting chl a from TP. A steady state model for TP was developed in which internal loading was predicted as a function of flushing rate and relative thermal resistance to mixing (RTRM) (Welch et al., 1989; Jones and Welch, 1990). The resulting relationship between dilution water input and lake TP in Parker Horn shows the moderating effect of internal P loading on the effectiveness of dilution water input (Figure 6.1). TABLE 6.2 Average April–September Dilution Rates through Discrete Sections of Moses Lake and Resulting Average May–September Values for TP, SRP, Chl a and Secchi Transparency for before (1969–1970) and after (1977–1979) Dilution (except for Pelican Horn) and after Dilution and Sewage Diversion (1986–1988) Years/Lake Area Dilution Rate (%/d) TP (μg/L) SRP (μg/L) Chl a (μg/L) Secchi (m) Parker Horn 1969–1970 1.6 152 28 71 0.6 1977–1979 7.8 68 15 26 1.3 1986–1988 8.0 47 6 21 1.5 South Lake 1969–1970 1.1 156 48 42 1.0 1977–1979 3.5 86 35 21 1.7 1986–1988 3.6 41 7 12 1.7 Pelican Horn 1969–1970 0.0 920 634 48 a 0.40 1977–1979 0.0 624 441 39 a 0.45 1986–1988 7.7 77 6 12 0.65 a Chl a: biolvolume ratios one half rest of lake. Note: Samples from 0.5 m depth transects. Source: From Cooke et al., 1993. With permission. Copyright © 2005 by Taylor & Francis Goldman (1968) suggested that reducing the N content in Clear Lake, CA (a hypereutrophic, shallow lake in an arid region, like Moses Lake), by adding Eel River water, should reduce algal content (as it did initially in Moses). There was uncertainty about the buffering effect of increased release of N from sediments once the lake water N content decreased and a larger gradient between sediment interstitial N and the overlying water occurred. That effect did occur in Moses Lake, as indicated earlier, although the effect of increasing RTRM was 2.8 times more important, based on the steady state P model, than increasing the flushing rate of dilution water (Jones and Welch, 1990). Reducing the inflow concentration of P, either by adding dilution water or by removing P from the external input, results in an increased gradient in P concentration between sediment and water and, thus, a greater diffusive flux from sediments (Poon, 1977; Sas et al., 1989). A decrease in internal loading as sediments become depleted of P is possible with continued dilution as with wastewater P removal or diversion. The year-to-year variability in internal load was too great to detect any trend in Moses Lake with the 12 years of data through 1988. However, a 2001 P budget shows a negative P internal loading for the April–September period (Carroll, personal communi- cation), compared with internal loading being 53% of the total for that period in 1988. The physical loss of algal cells by washout has probably contributed to reduced biomass in Parker Horn, where high rates of exchange (20–25%/d) existed for short periods. For example, biomass in upper Pelican Horn decreased from 80 to 10 mm 3 /L in about 1 month following increased water exchange by pumping Parker Horn water through Pelican Horn (Carlson and Welch, 1983; Welch et al., 1984). The greatest decrease occurred with an exchange rate of 9%/d, although FIGURE 6.3 Residual lake water, in percent, remaining in Parker Horn (open circles), South Lake (closed circles), and the whole lake (triangles) compared with that predicted (based on average inflow from mid-April to mid-June) for the whole lake (solid line) and Parker Horn (dashed line) in response to dilution water input in 1978. Parker Horn, South Lake, and the whole lake represent, respectively, 8%, 21%, and 100% of the lake volume. (Reprinted from Welch, E.B. and C.R. Patmont. 1980. Water Res. 14: 1317–1325. With permission from Pergamon Press Ltd., Oxford.) 100 80 60 40 20 0 April May June Percent lake water m 3 sec −1 Dilution water Copyright © 2005 by Taylor & Francis subsequent exchange rates were higher (19%/d). Persson (1981) observed that Oscillatoria biomass was greatly influenced by flushing rate during the period (about a month) of maximum growth in a hypereutrophic brackish-water bay. At a flushing rate of 8.1% and 9.4%/d the average biomass was about one half the level at 4.7%/d. At 20.7%/d biomass was only about one third the level at 4.7%/day. For Parker Horn, therefore, where the mean flushing rate was about 8%/d from April to September (Table 6.2), washout of cells probably represented part of biomass control. In the remainder of the lake, where flushing averaged less than 1%/d, washout was of less significance when compared with a 50%/d maximum growth rate, which was observed for Aphanizomenon in the lake. If the N source is fixation, however, even relatively low flushing rates may effectively reduce the biomass of that alga through washout. Instability of the water column, as indicated by a decreased vertical density gradient (low relative thermal resistance to mixing, RTRM), has contributed to the crash or prevention of blue- green algal blooms in Moses Lake (Welch and Tomasek, 1981). Because the buoyancy capability of blue-greens provides advantages over greens and diatoms when mixing is poor, decreased stability hinders dominance by blue-greens (Knoechel and Kalff, 1975; Paerl and Ustach, 1982; Chapter 19). Daily monitoring showed that blue-green biomass increases and surface accumulations become more pronounced under quiescent conditions, but biomass disperses with increased mixing from wind > 4.9–7.6 m/h (Bouchard, 1989; Welch et al., 1992). Wind has more effect on water column stability in Moses Lake than dilution water input. While biomass was substantially reduced by dilution, algal composition did not change during the 12 years dilution. Blue-greens dominated the phytoplankton throughout the summer (Welch et al., 1992). The blue-green fraction decreased initially (Welch and Patmont, 1980), but that did not persist. This was unexpected since decreased blue-green dominance has accompanied decreased TP content in most situations (Sas et al., 1989). However, the first intensive monitoring since 1988, conducted during 2001, the fifth highest dilution water input year, showed low mean surface TP (20 μg/L), low chl a and a near absence of blue-green algae in Parker Horn and South Lake (Carroll, personal communication). The optimum use of dilution water for Moses Lake is a moderate, but continuous input from May through August. Water added too early (February to March) would be largely replaced by high-nutrient Crab Creek water by June, when algal blooms begin. Replacement with Crab Creek water creates a problem if dilution water is stopped in June or early July. Unfortunately, the lack of irrigation demand during wet years reduces space available in the downstream impoundment for dilution water routed through Moses Lake. Thus, dilution water transport through the lake is higher in dry years. Nevertheless, based on the earlier relationship between May to August average inflow NO 3 concentration and June to August average chl a, a dilution water volume of about 100 × 10 6 m 3 from May through August should control chl a to an average of about 20 μg/L, whether that quantity comes as 10 m 3 /s for the whole period or is divided into 25 m 3 /s for May and 5 m 3 /s for June through August. If water input is not maintained through August, blooms with very high chl a levels will result. Also as indicated in Figure 6.1, that volume should result in a TP concen- tration of about 55 μg/L and a chl a content of about 24 μg/L if P is limiting. Also as indicated by Figure 6.1, there is a diminishing value to more dilution water and any further improvement in the quality of Moses Lake must involve control of internal P loading. However, if the lack of internal P loading observed in 2001 is representative of conditions in recent years, effectiveness of long-term dilution is greater than expected from Figure 6.1. Although the input of dilution water has not been distributed over the summer as desired, the water has cost the Moses Lake Irrigation and Rehabilitation District nothing, because water diverted through Moses Lake is used for downstream irrigation. The primary project cost has been the pumping facility for Pelican Horn, at $577,000 (2002 U.S. dollars) plus planning, administrative, and monitoring and research cost. To guarantee a flow of 5 m 3 /s during July and August, however, it would be necessary for the District to buy water. Although this is not planned and costs and Copyright © 2005 by Taylor & Francis liabilities have not been considered, it is instructive to assume a cost (2002 dollars) similar to that encountered for diluting Green Lake ($0.13/m 3 ; see Green Lake section). For two months the volume would be about 26 × 10 6 m 3 , with a cost of nearly $3.5 × 10 6 . The cost for dilution water at Moses Lake would probably not be as high as that for the domestic water used in Green Lake. However, it becomes clear that to buy dilution water may be a practical restoration alternative only for relatively small lakes, such as Green Lake, that require much lower rates of dilution water input. Green Lake is not fed naturally by a large input of high-nutrient water that must be diluted in order to dilute the lake nutrient concentration. In that respect, Moses Lake may be a rather unique case of a large lake existing at a location where large amounts of low-nutrient water are available at no cost. 6.3.2 GREEN LAKE This is another example of the benefits of dilution. The setting is the Seattle metropolitan area; 47,000 people live within 1.6 km of the lake, and as many as 1,162 people per hour use the 4.5- km path around the lake. Perhaps the smaller Green Lake (100 ha) represents a more practical example of dilution than does Moses Lake (2,700 ha). Dilution was proposed as the primary treatment in 1960 (Sylvester and Anderson, 1964) and was instituted in 1962. In contrast to the high rate of water input to Moses Lake’s Parker Horn, the dilution of Green Lake represented a much lower rate of flushing, even less than the whole of Moses Lake (2–3%/day). The average combined flushing rate was increased nearly threefold by adding low-nutrient water from the Seattle domestic supply, which comes from diversions near the source of two Cascade mountain streams. The addition of dilution water to Green Lake from 1965 to 1978 produced a flushing rate, based on dilution water only, ranging from 0.88 to 2.4/yr (0.24–0.65%/d). A marked improvement in chl a, TP, and Secchi transparency (SD) was noted during the first few years of dilution. Only one year’s predilution data existed for comparison with the three years of postdilution monitoring. Water transparency during the summer increased nearly fourfold, to an average of 4 m (because the mean depth is 3.8 m, most of the lake bottom was visible), and chl a decreased more than 90% from 45 to 3 μg/L. The summer mean TP decreased from 65 to 20 μg/L. A substantial decrease in the blue-green algal fraction was observed, particularly during spring and early summer. Regular monitoring was terminated in 1968, but the lake was again studied intensively for the purpose of proposing a new restoration plan (Perkins, 1983; URS, 1983). Mean chl a and TP had increased to 38 and 55 μg/L, respectively, during the summer of 1981; the lake quality had degraded markedly during the late 1970s, primarily due to declining dilution water inputs. No water was added in 1982, resulting in massive blue-green algal blooms. Dilution water was subsequently added in modest amounts on a regular basis to avoid deterioration in lake quality. Future limitation in the availability of Seattle domestic water necessitated developing a long-term solution. The percent decrease in TP concentration following initiation of dilution was about what would be expected from Equation 3.10. The expected TP concentration in Green Lake prior to dilution, calculated from estimated external loading, should have been about 80 μg/L, but was only 65 μg/L. Following dilution, the steady state concentration should have been about 35 μg/L; however, it actually declined to 20 μg/L by 1967 (Welch, 1979). The pre- to post-dilution decrease was the same (45 μg/L) for the expected and the actual TP values. The discrepancy is most likely due to overestimating external loading. Mass balance analysis in the 1980s showed that much of the lake’s problem was due to internal P loading, which was unrecognized earlier. Internal loading was high despite nearly the whole lake being unstratified and oxic during summer (Perkins, 1983; URS, 1983). Internal loading accounted for 21% of the total annual P loading during 1981 (determined by difference in the annual mass balance of P). During the three summer months, however, when chl a averaged 38 μg/L, with a Copyright © 2005 by Taylor & Francis [...]... Estuary U.S Geol Surv Water Suppl Paper 1873-A Seattle, WA p 62 Welch, E.B 1979 Lake restoration by dilution In: Lake Restoration USEPA-400/ 5-7 9-0 01 pp 133–139 Welch, E.B and C.R Patmont 1979 Dilution effects in Moses Lake In: Limnological and Socioeconomic Evaluation of Lake Restoration Projects USEPA -6 0 0/ 3-7 9-0 05 pp 187–212 Welch, E.B and C.R Patmont 1980 Lake restoration by dilution: Moses Lake, Washington... Water Res 14: 1317–1325 Welch, E.B and M.D Tomasek 1981 The continuing dilution of Moses Lake, Washington In: Restoration of Lakes and Inland Waters USEPA-440/ 5-8 1-0 10 pp 238–244 Welch, E.B and E.R Weiher 1987 Improvement in Moses Lake quality by dilution and diversion Lake and Reservoir Manage 3: 58 65 Copyright © 2005 by Taylor & Francis Welch, E.B., J.A Buckley and R.M Bush 1972 Dilution as an algal... Carlson, K.L and E.B Welch 1983 Evaluation of Moses Lake Dilution: Phase II Water Res Tech Rept 80 Dept Civil Eng., Washington Dept of Ecology, Olympia Personal communication Carroll, J 2004 Moses Lake Total Maximum Daily Load Phosphorus Study Pub No 0 4-0 3-0 , WA Dept of Ecology, Olympia, WA Cooke, G.D., E.B Welch, S.A Peterson and P.R Newroth 1993 Restoration and Management of Lakes and Reservoirs, ... 232: 417–418 Hosper, S.H 1985 Restoration of Lake Veluwe, The Netherlands, by reduction of phosporus loading and flushing Water Sci Technol 17: 757–7 86 Hosper, H and M.L Meyer 19 86 Control of phosphorus loading and flushing as restoration methods for Lake Veluwe, The Netherlands Hydrobiol Bull 20: 183–194 Jacoby, J.M., H.L Gibbons, K.B, Stoops and D.D Bouchard 1994 Response of a shallow, polymictic lake... 1 969 Some effects of lake renewal on phytoplankton productivity and species composition Limnol Oceanogr 14: 66 0 66 6 Dillon, P.J 1975 The phosphorus budget of Cameron Lake, Ontario: the importance of flushing rate relative to the degree of eutrophy of a lake Limnol Oceanogr 29: 28–39 Entranco Engineers, Inc., Bellevue, WA personal communication Copyright © 2005 by Taylor & Francis Goldman, C.R 1 968 ... Water Pollut Cont Fed 62 : 847–852 Jones, J.R and R.W Bachmann 19 76 Prediction of phosphorus and chlorophyll levels in lakes J Water Pollut Cont Fed 48: 21 76 2182 McBride, G.B and R.D Pridmore 1988 Prediction of [chorophyll a] in impoundments of short hydraulic retention time: Mixing effects Verh Int Verein Limnol 23: 832–8 36 Mesner, N 1985 Use of a Seasonal Phosphorus Model to Compare Restoration Strategies... 44: 2245–2 265 Welch, E.B., K.L Carlson, R.E Nece and M.V Brenner 1982 Evaluation of Moses Lake Dilution Water Res Tech Rept 77 Dept Civil Eng., University of Washington, Seattle Welch, E.B., M.V Brenner, and K.L Carlson 1984 Control of algal biomass by inflow nitrogen In: Lake and Reservoir Management USEPA-440/ 5-8 4-0 01 pp 493–497 Welch, E.B., M.D Tomasek and D.E Spyridakis 1985 Instability of Mount... Expectations, Experiences and Extrapolations Academia-Verlag, Richarz, St Augustin, Germany Sketelj, J and M Rejic 1 966 Pollutional phases of Lake Bled In: Advances in Water Pollution Research Proc 2nd Int Conf Water Pollut Res Pergamon, London, pp 345– 362 Stadelman, P 1980 Der zustand des Rotsees bei Luzern Kantonales amt fur Gewasserschutz, Luzern Sylvester, R.O and G.C Anderson 1 964 A lake’s response... even if only moderate- to high-nutrient water is available, through physical limitations to large algal concentrations However, the principal limitation for use of this technique is the availability of low-nutrient dilution water REFERENCES Barbiero, R.P and E.B Welch 1992 Contribution of benthic blue-green algal recruitment to lake populations and phosphorus Freshwater Biol 27: 249– 260 Born, S.M., T.L... Aspects of Clear Lake, California with Special Reference to the Proposed Diversion of Eel River Water through the Lake Rept Fed Water Pollut Control Admin Horne, A.J and C.R Goldman 1972 Nitrogen fixation in Clear Lake, California I Seasonal variation and the role of heterocysts Limnol Oceanogr 17: 67 8 69 2 Horne, A.J and A.D Viner 1971 Nitrogen fixation and its significance in tropical Lake George, Uganda . Horn 1 969 –1970 1 .6 152 28 71 0 .6 1977–1979 7.8 68 15 26 1.3 19 86 1988 8.0 47 6 21 1.5 South Lake 1 969 –1970 1.1 1 56 48 42 1.0 1977–1979 3.5 86 35 21 1.7 19 86 1988 3 .6 41 7 12 1.7 Pelican Horn 1 969 –1970. Phosphorus Study. Pub. No. 0 4-0 3-0 , WA Dept. of Ecology, Olympia, WA. Cooke, G.D., E.B. Welch, S.A. Peterson and P.R. Newroth. 1993. Restoration and Management of Lakes and Reservoirs, 2nd. ed. CRC. 14: 1317–1325. Welch, E.B. and M.D. Tomasek. 1981. The continuing dilution of Moses Lake, Washington. In: Restoration of Lakes and Inland Waters. USEPA-440/ 5-8 1-0 10. pp. 238–244. Welch, E.B. and E.R. Weiher.