Ground Water Recharge Using Waters of Impaired Quality (1994)

Although much can be learned from discussing the general characteristics of ground water recharge technologies, the knowledge becomes most useful when seen in light of actual examples. Examples provide an opportunity to see how theory translates into on-the-ground activity. This chapter provides brief descriptions of existing ground water recharge projects. The examples were selected to illustrate the common techniques used, show a variety of the purposes for which recharge is planned, and give concrete examples of the problems such projects sometimes face. The seven sites discussed are

• Water Factory 21, Orange County, California

• Montebello Forebay, California

• Phoenix, Arizona

• El Paso, Texas

• Long Island, New York

• Orlando, Florida

• The Dan Region, Israel

These examples are illustrative and brief, and the committee did not attempt to make recommendations from these site-specific cases. Instead, the committee hopes that these descriptions will show that artificial ground water recharge is not a "technology of the future" but rather something in use today in relatively diverse settings. These examples show that properly planned and operated artificial recharge projects can increase our water management options and flexibility.

The Water Factory 21 project in Orange County, California, is the first injection project involving highly treated municipal wastewater. The injected water provides a barrier to seawater intrusion, but also enters potable water supply aquifers. The project has been in operation since 1976 and has provided significant data on the capability and reliability of advanced wastewater treatment processes to remove microbiological and chemical constituents, ground water quality, and monitoring techniques.

Another California example is the Montebello Forebay project in south-central Los Angeles County. This project demonstrates indirect potable reuse via surface spreading of reclaimed water. The Montebello Forebay project has been in operation since 1962 and has been the subject of extensive research to investigate health-related issues.

The Phoenix, Arizona, example illustrates the extensive research undertaken to demonstrate the capability of soil-aquifer treatment (SAT) to treat relatively low quality treated municipal wastewater to levels acceptable for many nonpotable applications upon extraction.

The El Paso, Texas, project is the first injection project in the United States where the sole intent of the project is to augment the potable water supply aquifer using reclaimed municipal wastewater. It is a relatively new project and will provide important data as it builds an operational history.

The Long Island, New York, example demonstrates the effectiveness of artificial recharge in a more urbanized, eastern setting, where climate and water availability are significantly different than in the West. Stormwater runoff is recharged into infiltration basins to replenish the ground water withdrawn for use by Long Island residents, thereby also helping to retard seawater intrusion into the aquifers that provide the primary source of drinking water for the area.

Another eastern project, the stormwater drainage wells in Orlando, Florida, is included to illustrate another approach to using excess stormwater runoff for artificial recharge, thus helping to solve a wastewater disposal problem as well as a water supply problem.

Finally, one international example is provided. The Dan Region project in Israel provides information on a large-scale recharge operation that incorporates SAT of treated municipal wastewater and subsequent extraction of the water for extensive agricultural irrigation. The project is well documented and has been in operation for almost 20 years.

The Orange County Water District (OCWD) was formed by a special act of the California legislature in 1933 for the purpose of protecting the Orange County ground water basin. In 1955, OCWD was given the added responsibility of water management. Early in its history, OCWD secured the right to all water in

FIGURE 6.1 Orange County ground water recharge facilities and saltwater intrusion barriers.

the Santa Ana River. Over 3 million acre-feet of the river's flow has been captured to recharge the Orange County ground water basin. In addition, more that 2.5 million acre-feet of water imported from northern California and the Colorado River has been recharged (Orange County Water District, 1991). The location of OCWD and its recharge facilities is shown in Figure 6.1.

The Orange County ground water basin is the depositional plain of the Santa Aria River. The principal features of the region are surrounding hills and a broad, poorly drained alluvial plain with alternating gaps and minor hill systems along the coast. A major fault system parallels the coastline, which apparently

seals the basin from the sea at deeper levels. However, in several gaps along the ocean front there is hydraulic continuity between seawater and ground water in the upper 45 to 60 m (150 to 200 ft) of recent alluvial fill (Argo and Cline, 1985).

The aquifers in the area are composed of fine-to coarse-grained sand, separated by silt and clay layers (aquicludes and aquitards). The Talbert aquifer, the principal zone of production in the area, is of recent age and overlies Pleistocene deposits within the gap created by the Santa Aria River. The Talbert aquifer is the only aquifer in direct contact with the Pacific Ocean. The three lower zones of local production are subject to intrusion by virtue of their contact with the Talbert aquifer.

The base of the freshwater-bearing sediments is more than 1,200 m (4,000 ft) deep in some inland locations but rises to a depth of 60 m (200 ft) along the coast, where seawater intrusion has occurred. Seawater intrusion was first observed in municipal wells during the 1930s as a consequence of basin overdraft. Overdrafting of the ground water continued into the 1950s. Overpumping of the ground water resulted in seawater intrusion as far as 5.6 km (3.5 miles) inland from the Pacific Ocean by the 1960s. Although OCWD prevented further intrusion through percolation of large amounts of imported water in the forebay area of the ground water basin, the need for a coastal barrier system was obvious.

OCWD began pilot studies in 1965 to determine the feasibility of using effluent from an advanced wastewater treatment (AWT) facility as injection water in a hydraulic barrier system to prevent the encroachment of saltwater into potable water supply aquifers. Construction of an AWT facility known as Water Factory 21 was started in 1972 in Fountain Valley, and injection of the treated municipal wastewater into the ground began in 1976.

Water Factory 21 receives activated sludge secondary effluent from the adjacent County Sanitation Districts of Orange County (CSDOC) and has a design capacity of 15 million gallons per day (mgd). Water Factory 21 has the following unit processes: lime clarification for removal of suspended solids, heavy metals, and dissolved minerals; air stripping for removal of ammonia and volatile organic compounds; recarbonation for pH control; mixed-media filtration for removal of suspended solids; activated carbon adsorption for removal of dissolved organic compounds; reverse osmosis (RO) for demineralization and removal of other constituents; and chlorination for disinfection and algae control. The current operation mode is shown in Figure 6.2.

Because California rules require that total dissolved solids cannot exceed 500 milligrams per liter (mg/l) prior to injection, RO is used to demineralize up to 5 mgd of the wastewater used for injection. The feed water to the RO plant is effluent from the mixed-media filters. Effluent from granular activated carbon adsorption columns is disinfected and blended with RO water. Activated carbon is regenerated on site in a multiple-hearth furnace. Solids from the settling basins are incinerated in a multiple-hearth furnace from which lime is recovered

and reused in the chemical clarifier. Brine from the RO process is pumped to the CSDOC facilities for ocean disposal. Reclaimed water produced at Water Factory 21 is injected into a series of 23 multicasing wells providing 81 individual injection points into 4 aquifers. The resulting seawater intrusion barrier is known as the Talbert injection barrier (Argo and Cline, 1985). A schematic of a typical injection well is shown in Figure 6.3. The wells are located at 183-m (600-ft) intervals in a city street approximately 5.6 km (3.5 miles) inland from the Pacific Ocean. Each well has the capacity to inject 450 gallons per minute (gpm). They vary in depth from 27 m (90 ft) to 130 m (430 ft). There are 7 extraction wells located between the injection wells and the coast. At the present time, the ground water is maintaining a positive hydraulic gradient toward the ocean, and the extraction wells are not in use. Prior to injection, the product water is blended 2:1 with deep well water from an aquifer not subject to contamination. The blended water is chlorinated in a blending reservoir before it is injected into the ground. Depending on conditions, the injected water flows toward the ocean forming a seawater barrier, inland to augment the potable ground water supply, or in both directions. On average, well over 50 percent of the injected water flows inland to augment the potable water supply.

The AWT processes at Water Factory 21 reliably produce a high-quality water. No total coliform organisms were detected in any of 161 samples of blended injection water tested during 1990 (Wesner, 1991). A virus monitoring program conducted from 1975 to 1982 demonstrated to the satisfaction of the state and county health agencies that Water Factory 21 effluent is essentially free of measurable levels of viruses (McCatry et al., 1982). The average turbidity of filter effluent was 0.20 nephelometric turbidity units (NTU) and did not exceed 1.0 NTU at any time during 1990. The average chemical oxygen demand (COD) and total organic carbon (TOC) concentrations for the year were 8 mg/l and 2.8 mg/l, respectively (Wesner, 1991). The effectiveness of the RO process in the removal of inorganic constituents at Water Factory 21 is indicated in Table 2-10 in Chapter 2. The concentrations of priority organic pollutants at various steps in the treatment train are presented in Table 2.11 in Chapter 2.

In 1992, the California Department of Health Services removed a restriction that required blending reclaimed water not of sewage origin prior to injection. Hence, OCWD is considering phasing out use of deep well water for blending and inject 100 percent reclaimed water. In addition, ground water studies indicate that approximately 25 mgd of injected water is needed to fully protect against seawater intrusion at the Talbert Gap, and consideration is being given to increasing the amount of water produced at Water Factory 21 by 5 to 10 mgd in future years.

FIGURE 6.3 Typical reclaimed water injection well.

Ground water is an integral component of southern California's water resources. Artificial recharge of aquifers is practiced to augment replenishment of ground water basins in several locations, including the Montebello Forebay area

of south-central Los Angeles County. Waters used to recharge via surface spreading include local stormwater runoff, imported surface water (Colorado River water and State Project water), and reclaimed municipal wastewater.

Imported surface water is not always available for recharge during the summer months when demands by domestic water systems are at their peak. In addition, the availability of imported surface water is likely to be severely limited in the future. When the Central Arizona Project is completed in the mid-1990s, California's allotment of imported Colorado River water could be reduced by as much as 600,000 acre-feet/year. Also, the state of California may limit deliveries from the State Project, which supplies water to southern California from the Feather River/Sacramento Delta in northern California.

Reclaimed water has been used as a source of ground water replenishment in the Montebello Forebay area since 1962. At that time, approximately 12,000 acre-feet/year of disinfected activated sludge secondary effluent from the Sanitation Districts of Los Angeles County (LACSD) Whittier Narrows Water Reclamation Plant (WRP) was spread in the Montebello Forebay area of the Central Groundwater Basin, which is the main body of ground water underlying the greater Los Angeles metropolitan area. The basin has an estimated usable storage capacity of 780,000 acre-feet. In 1973, the San Jose Creek WRP was placed in service and also supplied secondary effluent for recharge. In addition, effluent from the Pomona WRP that is not reused for other purposes is discharged into San Jose Creek, a tributary of the San Gabriel River, which ultimately becomes a source of recharge water in the Montebello Forebay. The use of effluent from the Pomona WRP is expected to decrease as the reclaimed water becomes more fully used for irrigation and industrial applications in the Pomona area.

The water reclamation plants were originally built as secondary treatment facilities; however, body contact recreational activities in the receiving waters dictated that additional public health protection measures be taken. In the late 1970s all three reclamation plants were upgraded to provide tertiary treatment via dual media filtration (for the Whittier Narrows and San Jose Creek WRPs) or activated carbon filtration (for the Pomona WRP), and chlorination/dechlorination (Nellor et al., 1984). The activated carbon filters at the Pomona WRP have since been converted to dual-media filters.

The Montebello Forebay ground water recharge project is a cooperative effort. LACSD collects and treats municipal wastewater and monitors the effluent quality. The replenishment program is operated by the Los Angeles County Department of Public Works (LADPW), while overall management of the ground water basin is administered by the Water Replenishment District of Southern California (WRDSC). LADPW constructed special spreading areas designed to increase the indigenous percolation capacity by modifying the San Gabriel River channel and constructing off-stream spreading basins, ranging in size from 4 acres to 20 acres, adjacent to the Rio Hondo and San Gabriel rivers. The Rio

FIGURE 6.4 Montebello Forebay ground water recharge facilities.

Hondo spreading basins have 427 acres available for spreading. The San Gabriel River spreading basins occupy 224 acres, which include approximately 133 acres in an unlined section of San Gabriel River. The locations of the spreading basins and water reclamation plants are shown in Figure 6.4.

Under normal operating conditions, the basins are rotated through a 21-day cycle consisting of (1) a 7-day flooding period during which the basins are filled to maintain a constant 1.2 m (4-ft) depth; (2) a 7-day draining period during which flow to the basins is terminated and the basins are allowed to drain; and (3) a 7-day drying period during which the basins are allowed to dry out thoroughly. This wetting/drying operation serves several purposes, including maintenance of aerobic conditions in the upper soil strata.

In the aftermath of the 1976-1977 drought, there was considerable pressure

to more fully use reclaimed water supplies in southern California, particularly for ground water recharge. However, concerns by the California Department of Health Services (DOHS) over potential health effects of using reclaimed water to replenish potable water supplies caused a moratorium on planned expansions. In an attempt to answer some of the health-related issues associated with ground water recharge, a Health Effects Study was initiated in 1978 (Nellor et al., 1984). The focus of the study, conducted by LACSD, was the Montebello Forebay ground water recharge project. At the time the study was conducted, the annual amount of reclaimed water spread and recharged averaged 26,500 acre-feet/year, which was 16 percent of the total inflow to the ground water basin, with no more than 32,700 acre-feet of reclaimed water spread in any given year. The percentage of reclaimed water in the ground water supply was estimated to range from 0 to 23 percent on an annual basis, and 0 to 11 percent on a long-term (1962 to 1977) basis.

The primary goal of the 5-year $1.4 million study was to develop a database which could be used to enable health and regulatory authorities to determine whether the use of reclaimed water for ground water replenishment in the Montebello Forebay should be maintained at the then-current level, cut back, or expanded. A wide range of research was undertaken, including (1) water quality characterizations of ground water, reclaimed water, and other recharge sources in terms of their microbiological and inorganic chemical content; (2) toxicological and chemical studies of ground water, reclaimed water, and other recharge sources to isolate and identify health-significant organic constituents; (3) percolation studies to evaluate the efficacy of soil in attenuating inorganic and organic chemicals in reclaimed water, (4) hydrogeological studies to determine the movement of reclaimed water through ground water and the relative contribution of reclaimed water to municipal water supplies; and (5) epidemiological studies of populations ingesting recovered water to determine if their health characteristics differed significantly from a demographically similar control population.

The results of the Health Effects Study indicated that the risks associated with the three sources of recharge water (i.e., imported water, stormwater, and reclaimed water), were not significantly different and the historical proportion of reclaimed water used for replenishment had no measurable impact on either ground water quality or the health of the population ingesting the water (Nellor et al., 1984). The epidemiological study findings are weakened somewhat by recognition that the minimum observed latency period for human cancers that have been linked to chemical agents is about 15 years. Because of the relatively short time period that ground water containing a substantial proportion of reclaimed water had been consumed, it is unlikely that examination of cancer mortality rates would have detected an effect, if present, of exposure to reclaimed water.

Based on the results of the Health Effects Study and recommendations of a state-sponsored Scientific Advisory Panel (State of California, 1987), authoriza-

tion was given by the Los Angeles Regional Water Quality Control Board (LARWQCB) and DOHS in 1987 to increase the annual quantity of reclaimed water used for replenishment by approximately 50 percent to 50,000 acre-feet/ year over a period of 3 years, contingent upon the evaluation of data generated by an expanded monitoring program. Other requirements limited the total quantity of reclaimed water spread in any year to 50 percent of the total inflow to the basin and stipulated that the reclaimed water must meet all drinking water maximum contaminant levels and action levels (i.e., concentrations of contaminants in drinking water at which adverse health effects would not be anticipated to occur, based on an annual running average). Approval also was contingent on demonstration that there was no measurable increase in organic chemical contaminants in the ground water as the result of using reclaimed water for recharge.

Since the initial authorization, three increments of 7,300 acre-feet/year have been implemented, increasing the quantity of reclaimed water for ground water recharge to 50,000 acre-feet/year, or approximately 30 percent of the total inflow to the Montebello Forebay. In 1991, the LARWQCB revised permit conditions to allow recharge of up to 60,000 acre-feet of reclaimed water in any one year as long as the running 3-year average does not exceed 150,000 acre-feet. This allowed for greater flexibility in spreading operations.

The Montebello Forebay ground water recharge project includes extensive sampling and analysis of reclaimed water from the Whittier Narrows, San Jose Creek, and Pomona WRPs with similar monitoring of six shallow monitoring wells within the confines of the spreading grounds, 20 production wells in and around the spreading grounds, and ground water both upgradient and downgradient of the spreading grounds. The results of this combined monitoring program indicate that there has been no degradation of the ground water quality in terms of total dissolved solids, nitrogen, trace organics, heavy metals, or microorganisms (Hartling, 1993). Sampling and analysis of reclaimed water from each of the WRPs indicate that the WRPs consistently produce reclaimed water that does not contain measurable levels of viruses, contains less than 2.2 total coliform organisms/100 ml, and has an average turbidity of less than 2 NTU. Tables 2.8 and 2.9 in Chapter 2 provide water quality data from the three water reclamation plants that provide reclaimed water for recharge in the Montebello Forebay.

In addition to providing a much-needed source of water for recharge, the use of reclaimed water is attractive from an economic standpoint. In 1992, WRDSC purchased reclaimed water from the Whittier Narrows WRP for $7per/acre-foot and reclaimed water from the San Jose Creek WRP for $11.56 per/acre-foot. Reclaimed water from the Pomona WRP that is not reused for other purposes, approximately 2,000 acre-feet, is captured for ground water recharge at no cost. The cost of the reclaimed water compares favorably to the seasonal storage rate of $130 per acre-foot for imported water purchased from the Metropolitan Water District of Southern California in 1992 (Hartling, 1993). The seasonal storage

rate is normally in effect during the wet, winter months and is expected to increase substantially in future years. The use of reclaimed water in place of imported water has saved an estimated $44 million since recharge using reclaimed water began in 1962 (Hartling, 1993).

Additional research conducted since the completion of the Health Effects Study has included an evaluation of the efficiency of the LACSD's full-scale carbon filters for removing mutagenicity as determined by the Salmonella microsome assay (Baird, 1987). This work indicates that average removals of 80 percent could be achieved based on a 10-minute empty bed contact time and that the effects of chlorine disinfection on mutagenic activity vary significantly. These results suggest that chlorine can oxidize (deactivate) some types of mutagens but also can react with available organic matter to create more mutagens in a given sample.

Ongoing research has focused on the development of a ground water tracer suitable for characterizing the movement of reclaimed water in ground water basins. The study has thus far evaluated a series of alkyl pyridone sulfonate (APS) compounds and several fluorocarbon compounds in the laboratory to measure the degree of adsorption of these compounds on soils, their ability to withstand biodegradation under aerobic and anaerobic conditions, and their ability to withstand photodecomposition. Volatility studies and biological assays have been conducted to determine the potential of the tracer compounds to elicit acute toxicity or mutagenicity. The laboratory phase of study has been completed, and the second phase of study will verify the laboratory results under actual field conditions (R. B. Baird, personal communication, 1992).

In 1993, research was initiated to provide comparative, supplemental data for the Health Effects Study findings. Similar toxicological and chemical procedures are being used to characterize any changes in reclaimed water or ground water quality that might have occurred since the Health Effects Study samples were originally collected for evaluation. Additionally, the researchers will use current techniques to learn more about the characteristics of compounds in mutagenic fractions, thereby providing a better understanding of the origins and health significance of these compounds as well as the alternatives available for their removal (Sloss, 1993).

The city of Phoenix and other municipalities in the Salt River Valley of Arizona are interested in renovating part of their treated municipal wastewater by SAT so that it can be used for unrestricted irrigation and stored underground for eventual potable use. There are two major sewage treatment plants in the Phoenix area: the 91st Avenue treatment plant (activated sludge, chlorination, capacity about 119,000 million gallons/day (mgd)) and the 23rd Avenue treatment plant (activated sludge, chlorination, capacity about 40 mgd. The feasibil-

FIGURE 6.5 Schematic of Flushing Meadows project showing infiltration basins and monitoring wells.

ity of SAT in the Phoenix area was studied with two experimental systems, a small test project installed in 1967 below the 91st Avenue treatment plant, and a larger, demonstration project installed in 1975 below the 23rd Avenue treatment plant. The latter could be part of a future operational project that would have a basin area of 119 acres and a projected capacity of about 73 × 109 gal/year. Both projects were in the normally dry Salt River bed.

The project below the 91st Avenue treatment plant, known as the Flushing Meadows project (Bouwer et al., 1974a,b, 1980), was an experimental project that consisted of six parallel, long, narrow infiltration basins of about 0.32 acres each (Figure 6.5) in the Salt River floodplain. The soil consisted of about 1 m (3 ft) of loamy sand underlain by sand and gravel layers. The ground water table was at a depth of around 3 m (9 ft). Monitoring wells 2 to 9 m (7 to 30 ft) deep were installed at various points between the basins and away from the basins. This configuration made it possible to sample renovated wastewater from the aquifer both while it was still below the basins and after it had moved laterally for some distance through the aquifer.

FIGURE 6.6 Sketch of the 23rd Avenue project with the 16-ha lagoon (left) that was split into 4 infiltration basins for the demonstration project, and the 32-ha lagoon (right) that can be split into 9 infiltration basins to increase the capacity of the system to that of plant outflow.

The second project was the 23rd Avenue project (Bouwer and Rice, 1984). This was a demonstration and possible future operational project on the north side of the Salt River bed. It consisted of an old 39.5-acre lagoon split lengthwise into four infiltration basins of 9.9 acres each (Figure 6.6). The soil lacked the loamy sand top layer of the Flushing Meadows project. Thus, the soil profile consisted mostly of sand and gravel layers. At this site, the ground water table was much deeper and ranged between 5 and 25 m (15 and 75 ft) depth (mostly around 15 m (45 ft)) for the study period. Monitoring wells to sample renovated wastewater were installed in the center of the project at depths of 18, 24, and 30 m (59, 79, and 98 ft), and on the north and south sides of the basin complex at depths of 222 m (728 ft) (Figure 6.6). In addition, a large production well (capacity about 2.6 mgd) was drilled in the center of the project with the casing perforated from 30 to 54 m (98 to 177 ft) depth.

The inundation schedule typically was 9 days flooding/12 days drying at the Flushing Meadows project and 14 days flooding/14 days drying at the 23rd Avenue project. Water depths in the basins were about 15 to 20 cm (6 to 8 inches). During flooding, infiltration rates typically were between 0.3 and 0.6

m/day (1 and 2 ft/day), yielding a total infiltration or hydraulic loading of between 60 and 120 m/year (197 and 394 ft/year) for Flushing Meadows and about 100 m/year (330 ft/year) for the 23rd Avenue project.

For the 23rd Avenue project, the effluent from the treatment plant initially flowed through a 79-acre lagoon before it entered the infiltration basins. Heavy growth of algae, particularly the unicellular Carteria klebsii, in the lagoon caused soil clogging in the infiltration basins, especially in the summer. In addition to forming a ''filter cake" on the bottom of the infiltration basins, the photosynthesis of the algae removed carbon dioxide from the wastewater, which raised the pH and, in turn, caused precipitation of calcium carbonate, which further aggravated the soil clogging. Because of this clogging, hydraulic loading rates initially averaged only 21 m/year (69 ft/year). Algae growth and resulting soil clogging were prevented by building a bypass canal around the lagoon (Figure 6.6), reducing the detention time of the effluent from a few days in the lagoon to about one-half hour in the canal. Also, the water depth in the basins was reduced from about 100 to 20 cm (39 to 7.9 inches). This increased the rate of turnover of the water in the basins and reduced growth of suspended algae. It also reduced compaction of the clogging layer, thus yielding higher infiltration rates (Bouwer and Rice, 1989). After the bypass canal was put into operation, hydraulic loading rates for the infiltration basins were almost 100 m/year (330 ft/yr), or approximately 5 times higher than they were before.

At a hydraulic loading rate of 100 m/year (330 ft/yr), 1 acre of infiltration basin can handle 330 × 325,900 = 10⁸ gal/year of wastewater. Thus, the 40 mgd of effluent from the 23rd Avenue wastewater treatment plant would require 144 acres of infiltration basins. Almost all of this area could be obtained by also converting the 79-acre lagoon east of the present infiltration system into infiltration basins. This would give a total basin area of about 118 acres, which could handle about 126 million gallons/year of effluent. The wells for pumping the renovated water from the aquifer would be located on the centerline through the project. At a capacity of 2.6 mgd per well, 12 wells would be needed to pump renovated water out of the aquifer at the rate at which it infiltrates as wastewater in the basins, thus creating an equilibrium situation (Bouwer and Chase, 1984).

For both recharge projects, most of the quality improvement of the waste-water occurred in the vadose zone (i.e., the zone between soil surface and the ground water table). The quality improvements are summarized below (for additional details, see H. Bouwer et al. (1974b, 1980), Gilbert et al. (1976), and E. J. Bouwer et al. (1984)):

The suspended solids content of the renovated water at the Flushing Meadows project was less than 1 mg/l, compared to a range of 10-70 mg/l for the secondary effluent (average was 32 mg/l in 1977). For the 23rd Avenue project, it averaged about 1 mg/l for the large production well. Most of these solids probably were fine aquifer particles that entered the well through the perforations in the casing. The suspended solids content of the secondary effluent at the 23rd Avenue project averaged about 11 mg/l.

The total salt or dissolved solids content (TDS) of the water increased slightly as it moved through the SAT system (from 750 to 790 rag/l at the 23rd Avenue project). Evaporation from the basins (including from the soil during drying) should increase the TDS content by about 2 percent. The rest of the increase was probably due to mobilization of calcium carbonate due to a pH drop from 8 to 7 as the wastewater moved through the vadose zone.

At the Flushing Meadows project, nitrogen removal by SAT was about 30 percent at maximum hydraulic loading (100 to 120 m/year (330 to 390 ft/year)). This increased to 65 percent when the loading rate was reduced to about 70 m/ year (230 ft/year) by using 9-clay flooding and 12-day drying cycles and by reducing the water depths in the basins from 0.3 to 0.15 m (1 to 0.5 ft). The form and concentration of nitrogen in the renovated water sampled from the aquifer below the basins were slow to respond to the reduction in hydraulic loading (Bouwer et al., 1980). In the tenth year of operation (1977), the renovated water contained 2.8 mg/l of ammonium nitrogen, 6.25 rag/1 nitrate nitrogen, and 0.58 mg/l organic nitrogen, for a total nitrogen content of 9.6 mg/l. This was 65 percent less than the total nitrogen of the secondary treated wastewater, which averaged 27.4 mg/l (20.7 mg/l as ammonium) in that year.

At the 23rd Avenue project, the total nitrogen content in the treated waste-water averaged about 18 mg/l, of which 16 mg/l was as ammonium. The 2-week flooding and drying cycles must have been conducive to denitrification in the vadose zone because the total nitrogen content of the renovated water from the large center well averaged 5.6 mg/l of which 5.3 mg/l was as nitrate, 0.1 mg/l as ammonium, 0.1 mg/l as organic nitrogen, and 0.02 mg/l as nitrite. The nitrogen removal thus was about 70 percent. This removal was the same whether or not the secondary treated wastewater was chlorinated in the treatment plant, indicating that the low residual chlorine content of the treated wastewater by the time it

infiltrated into the ground apparently had no effect on the nitrogen transformations in the soil.

The flooding and drying sequence that maximizes denitrification in the vadose zone depends on various factors and must be evaluated for each particular system. Pertinent factors include the ammonium and carbon contents of the treated wastewater entering the soil, infiltration rates, cation exchange capacity of soil, exchangeable ammonium percentage, depth of oxygen penetration in the soil during drying, and temperature. The combined laboratory and field data from the Flushing Meadows experiments showed that to achieve high nitrogen removal percentages, the amount of ammonium nitrogen applied during flooding must be balanced against the amount of oxygen entering the soil during drying. Flooding periods must be long enough to develop anaerobic conditions in the soil. Infiltration rates must be controlled to the appropriate level for the particular wastewater, soil, and climate at a given site. Most of the nitrogen transformations in the Flushing Meadows studies occurred in the upper 50 cm (20 inches) of the vadose zone.

Phosphate removal increased with increasing distance of underground movement of the treated wastewater. After 3.m (9.8 ft) of downward movement through the vadose zone and 6 m (19.8 ft) mostly downward through the aquifer, phosphate removal at the Flushing Meadows project was about 40 percent at high hydraulic loading and 80 percent at reduced hydraulic loading. Additional lateral movement of 60 m (197 ft) through the aquifer increased the removal to 95 percent (i.e., to a concentration of 0.51 rag/l phosphate phosphorus versus 7.9 mg/l in the effluent). After ten years of operation and a total infiltration of 754 m (2470 ft) of secondary treated wastewater, there were no signs of a decrease in phosphate removal.

At the 23rd Avenue project, phosphate phosphorus concentrations in the last few years of the research averaged 5.5 rag/l for the secondary treated wastewater going into the ground and 0.37 rag/l for the renovated water pumped from the center well. The shallower wells showed a higher phosphate content, indicating that precipitation of phosphate again continued in the aquifer. For example, renovated water sampled from the 22-m-(72-ft-) deep north well showed phosphate phosphorus concentrations that averaged 1.5 rag/l, or about 4 times more than for the deeper center well. Most of the phosphate removal probably was due to precipitation of calcium phosphate.

Fluoride removal paralleled phosphate removal, indicating precipitation as calcium fluoride. At the Flushing Meadows project, fluoride concentrations in

1977 were 2.08 mg/l for the effluent, 1.66 mg/l for the water after it had moved 3 m (9.8 ft) through the vadose zone and 3 to 6 m (9.8 to 19.7 ft) through the aquifer, and 0.95 mg/l after it had moved an additional 30 m through the aquifer. At the 23rd Avenue project, fluoride concentrations averaged 1.22 mg/l in the secondary effluent and 0.7 mg/l in the renovated water from the center well.

Boron was not removed in the vadose zone and the aquifer of the Flushing Meadows and 23rd Avenue projects and was present at concentrations of 0.5 to 0.7 mg/l in both treated wastewater and renovated water. The lack of boron removal was due to insufficient mounts of clay in the vadose zone and aquifer.

At the Flushing Meadows project, movement of the secondary treated waste-water through 3 m (9.8 ft) of vadose zone and 6 m (19.7 ft) of aquifer reduced zinc from 193 to 35 μg/l, copper from 123 to 16 μg/l, cadmium from 7.7 to 7.2 μg/l, and lead from 82 to 66 μg/l (Bouwer et al., 1974b). Cadmium thus appeared to be the most mobile metal. More recent (about 1990) analyses showed much lower metal concentrations in the treated municipal wastewater. This is probably due to the use of better analytical equipment (atomic absorption with graphite furnace) and better control of industrial waste discharges into the sewer system. More recent concentrations were 36 μ/l for zinc, 8 μg/l for copper, 0.1 μg/l for cadmium, and 2 μg/1 for lead, which would give much lower metal concentrations in the water after SAT.

The secondary treated municipal wastewater at the Flushing Meadows project was not chlorinated and contained 105 to 10⁶ fecal coliforms per 100 ml. Most of these were removed in the top meter of the vadose zone. Some penetrated to the aquifer, however, especially when a new flooding period was started. The deeper penetration of fecal coliforms at the beginning of a flooding period was attributed to less straining of bacteria at the soil surface because the clogging layer had not yet developed. Also, since the activity of native soil bacteria at the end of a drying period can be expected to be lower because the input of nutrients was stopped, there probably was a less antagonistic environment for the fecal coliforms in the soil when flooding was resumed. Fecal coliform concentrations in the water after 3 m (9.8 ft) of travel through the vadose zone and 6 m (19.7 ft) through the aquifer were 10 to 500 per 100 ml when the renovated water consisted of water that had infiltrated at the beginning of a flooding period, and between 0 and 1 per 100 ml after continued flooding.

Additional lateral movement of about 100 m (330 ft) through the aquifer was necessary to produce renovated water in which fetal coliforms were undetectable in 100 ml at all times.

At the 23rd Avenue project, fecal coliform concentrations in the secondary sewage effluent entering the infiltration basins were 10,000 per 100 ml prior to November 1980 when the effluent was not yet chlorinated in the treatment plant and was first passed through the 79-acre lagoon. This concentration increased to 1.8 ×10⁶ per 100 ml when the nonchlorinated effluent was bypassed around the lagoon and flowed directly into the infiltration basins. It then decreased to 3,500 per 100 ml after the effluent was chlorinated in the treatment plant and still bypassed around the lagoon. The corresponding fetal coliform concentrations in the water pumped from the large center well from a depth of 30 to 54 m (98 to 177 ft) for these periods averaged 2.3, 22, and 0.27 per 100 ml, respectively, with ranges of 0 to 40, 0 to 160, and 0 to 3 per 100 ml, respectively. Higher fecal coliform concentrations were observed in the renovated water from the shallower wells, especially when the fecal coliform concentration of the infiltrating effluent was 1.8 × 10⁶/per 100 ml. At that time, water from the 18-m-(59-ft-) deep well showed coliform peaks after a new flooding period was started that regularly exceeded 1,000/per 100 ml and at one time even reached 17,000/per 100 ml. Thus, a considerable number of fetal coliforms passed through the vadose zone. However, chlorination of the effluent and resulting reduction of the fecal coliform concentration to 3,500/per 100 ml prior to infiltration, and additional movement of the water through the aquifer to the center well produced renovated water that was essentially free from fecal coliforms.

At the Flushing Meadows project, the viral concentrations of nonchlorinated secondary effluent averaged 2,118 plaque forming units (pfu) per 100 liter (average of six bimonthly samples taken for 1 year). Identified viruses included polio, echo, coxsackie, and reoviruses. No viruses could be detected in renovated water sampled after 3 m (9.8 ft) of movement through the vadose zone and 3 to 6 m (9.8 to 19.7 ft) of movement through the aquifer (Gilbert et al., 1976). At the 23rd Avenue project, vital concentrations in the renovated water from the center well averaged 1.3 pfu per 100 liter before chlorination of the secondary effluent, and 0 pfu per 100 liter after chlorination of the secondary effluent. The combined effects of chlorination and SAT thus apparently resulted in almost complete removal of the viruses, considering that the virus assays were done on 800 to 2,000 liter samples of the water after SAT.

At the Flushing Meadows project, the biochemical oxygen demand (BOD)

of the effluent after moving 3 m (9.8 ft) through the vadose zone and 6 m (19.7 ft) through the aquifer was essentially zero (too small to determine), indicating that almost all biodegradable carbon was mineralized. However, the renovated water still contained about 5 mg/l total organic carbon (TOC), as compared to 10 to 20 mg/l of TOC in the secondary effluent.

At the 23rd Avenue project, the TOC concentration of the secondary effluent averaged 12 rag/l where it entered the infiltration basins and 14 mg/l at the opposite ends of the basins. This increase probably was due to biological activity in the water as it moved through the basins. The renovated water from the 18-m (59-ft) well (intake about 5 m (16 ft) below the bottom of the vadose zone) had a TOC content of 3.2 mg/l and that from the center well (which pumped from 30to 54-m (98- to 177-ft) depth) had a TOC content of 1.9 mg/l, indicating further removal of organic carbon as the water moved through the aquifer. The TOC removal in the SAT system was the same before and after chlorination of the secondary treated wastewater, indicating that chlorination had no effect on the microbiological processes in the soil.

The concentration of organic carbon in the renovated water (1.9 mg/l) was higher than the 0.2 to 0.7 mg/l typically found in unpolluted ground water. The latter concentrations are mostly due to humic substances such as fulvic and humic acids (Thurman, 1979). The renovated sewage water from the SAT process thus could contain a number of synthetic organic compounds, some of which could be carcinogenic or otherwise toxic, or trihalomethane (THM) precursors.

The nature and concentration of trace organics in the secondary sewage treated wastewater and in the renovated water from the various wells of the 23rd Avenue project were determined by Stanford University's Environmental Engineering and Science Section, using gas chromatography and mass spectrometry. The studies were carried out for 2 months with nonchlorinated effluent, and then for 3 months with chlorinated effluent. As could be expected, the results showed a wide variety of organic compounds, including priority pollutants (many in concentrations on the order of micrograms per liter (see E. J. Bouwer et al., 1984; and H. Bouwer and Rice, 1984).

The chlorination had only a minor effect on the type and concentration of organic compounds in the treated municipal wastewater. Of the volatile organic compounds, 30 to 70 percent were lost by volatilization from the infiltration basins. Soil percolation removed 50 to 99 percent of the nonhalogenated organic compounds, probably mostly by microbial decomposition (Table 6.1). Concentrations of halogenated organic compounds decreased to a lesser extent with passage through the soil and aquifer (Table 6.2). Thus, halogenated organic compounds (including the aliphatic compounds chloroform, carbon tetrachloride, trichloroethylene, and 1,1,1-trichloroethane, and the aromatic di-and tri-

chlorobenzenes and chlorophenols) were more mobile and refractory in the underground environment than the nonhalogenated compounds, which included the aliphatic nonanes, hexanes, and octanes, and the aromatic xylenes, C₃-benzenes, styrene, phenanthrene, and diethyl phthalate.

In addition to the aliphatic and aromatic compounds mentioned, other compounds tentatively identified in organic extracts of the samples of treated municipal wastewater and renovated water using gas chromatography and mass spectrometry were fatty acids, resin acids, clofibric acid, alkylphenol poly-

ethoxylate carboxylic acids (APECs), trimethylbenzene sulfonic acid, steroids, n-alkanes, caffeine, diazinon, alkylphenol polyethoxylates (APEs), and trialkylphosphates. Several of the compounds were detected only in the secondary effluent and not in the renovated water. A few others—diazinon, clofibric acid, and tributyl-phosphate—decreased in concentration with soil passage, but were still detected in the renovated water. The APEs appeared to undergo rather complex transformations during filtration through the soil. They appeared to be completely removed with soil percolation when nonchlorinated effluent was infiltrated; when chlorinated effluent was used, however, two isomers were found following soil filtration, while others were removed.

These studies show that SAT is effective in reducing concentrations of a number of synthetic organic compounds in treated municipal wastewater, but that the renovated water still contains a wide spectrum of organic compounds, albeit at very low concentrations. Thus, while the renovated water is suitable for unrestricted irrigation and recreation, recycling it for drinking would require additional treatment, such as membrane filtration to remove the remaining organic compounds. The water would also have to be disinfected. Treatment of renovated municipal wastewater from an SAT system for potable use would, however, be much more effective and cheaper than treatment of municipal waste-water after conventional (primary and secondary) treatment (Semmens and Field 1980).

The results of the Phoenix studies show that the renovated water from the 23rd Avenue SAT projects meets the public health, agronomic, and aesthetic requirements for unrestricted irrigation, including parks, playgrounds, and vegetable crops that are consumed raw (U.S. Environmental Protection Agency, 1992). The water also meets the standards for lakes with primary contact recreation and for most industrial and other nonpotable uses. Potable use of the renovated water would require additional treatment, for example, reverse osmosis and disinfection. Such treatment, however, would be more effective and economical for renovated water from an SAT system than for effluent from a conventional sewage treatment plant (Bouwer, 1992).

In the Phoenix studies, secondary treated municipal wastewater was used because that was what the treatment plants provided. In general, however, the secondary (biological) treatment step is not necessary because SAT systems can handle relatively large amounts of organic carbon. Thus, where treated municipal wastewater is to be used for a rapid-infiltration system, primary treatment may suffice (Rice and Gilbert, 1978; Lance et al., 1980; Leach et al., 1980; Carlson et al., 1982; Rice and Bouwer, 1984). Some additional clarification or filtration of the primary effluent may, however, be desirable. The higher TOC content of the Filtered primary effluent actually may enhance denitrification, removal of recalcitrant organic compounds through secondary utilization and co-metabolism (McCarty et al., 1982), and removal of pathogens in the SAT system.

The city of El Paso, located in far west Texas, is an arid environment with an annual rainfall averaging about 20 cm (8 inches). Water supplies are scarce, and the water rights for the Rio Grande are fully appropriated. El Paso obtains about 10 percent of its water supply from the Rio Grande. The remaining 90

percent comes from ground water. Of this, approximately 25 percent of the water supply comes from the Canutillo well field located on the Mesilla Bolson to the west of the Franklin and Organ mountains, and the remaining 65 percent comes from the Hueco Bolson well field located to the east of the mountains (Figure 6.7). (A bolson is a broad and nearly flat mountain-rimmed sediment-filled desert basin with interior drainage.) Juarez, Mexico, which is located directly across the Rio Grande from El Paso, has roughly double the population of El Paso and also depends on ground water from the Hueco Bolson aquifer for its water supply. Because of these two major water users, the freshwater layer of the Hueco Bolson gradually is being depleted and replaced by the more saline ground water that surrounds the fresh ground water resource.

The total water use for El Paso is about 100 mgd. Of this, about 50 percent is returned to sewage plants for treatment. Thus, there is an opportunity to reuse some of this water to recharge the Hueco Bolson aquifer and extend the long-term use of the resource as a potable water supply. The Fred Hervey Water Reclamation Plant provides up to 10 mgd of reclaimed water for ground water recharge. This means that each 10-year period of operation of this facility can extend the resource lifetime of the aquifer by 1 year.

The El Paso recharge project is a 10-mgd direct-injection system that was selected over infiltration basins because of the area's deep water table (about 107 m (350 feet) below the surface). The overall recharge system consists of an advanced wastewater treatment plant, a pipeline system through the Hueco Bolson, and 10 injection wells. All sewage collected in the northeast of the city is pumped to the treatment plant. Following treatment, the wastewater is pumped to the injection system for injection across the freshwater section of the bolson between existing production wells. After injection, the water travels approximately 1.2 km (.75 mile) through the aquifer to production wells for municipal water supply. In addition, the reclaimed water is available directly for industrial cooling at a nearby plant and some is used to irrigate a city golf course.

Recharge of the Hueco Bolson aquifer occurs along the foothills of the mountains and plateaus where sediments are coarse grained and permeable. In some places, additional recharge comes from overlying alluvium. The thickness of the Hueco Bolson deposits ranges from 305 to 2,740 m (1,000 to 9,000 ft), with the deepest known freshwater at a depth of about 430 m (1,400 ft).

Before development of the Hueco Bolson as a ground water resource, the

floodplain of the Rio Grande served as the discharge zone for the ground water, and the Rio Grande was a gaining stream. The first well supplying ground water to El Paso was drilled around 1892. In 1901 the development of the Mesa well field north of Fort Bliss began, and 44 wells had been completed by 1917. In 1917, development of the Montana well field at a lower elevation near the eastern side of the city began. Subsequently, deep wells were drilled in the Mesa and Montana fields to meet increasing demands. A 1970 water level map shows two cones of depression in the water table east of the Franklin Mountains.

Development of ground water from the bolson deposits near the city has lowered the water level below the water level in the overlying Rio Grande alluvium. Instead of being an area of discharge, the alluvium has become an area of ground water recharge, and the Rio Grande is now a losing stream. Digital modeling suggests that recharge from the Rio Grande alluvium has exceeded the natural recharge rate since the 1950s and 1960s.

Municipal and industrial ground water pumpage has increased steadily since the early 1900s. From 1906 to 1975, about 1.80 million acre-feet was pumped from the Texas part of the northern Hueco Bolson aquifer, whereas from 1925 to 1975 about 570,000 acre-feet was pumped from the Juarez area in Mexico. In 1975, 72,000 acre-feet was pumped from the Texas part of the northern bolson, and about 40,000 acre-feet was pumped from the Juarez area. These pumpages have resulted in a water level decline of 18 to 21 m (60 to 70 ft) in the northern and southern part of El Paso and as much as 29 m (95 ft) in downtown El Paso and Juarez. Irrigation can be a big water user, varying from 10,000 acre-feet per year to as much as 150,000 acre-feet per year during a drought period (Charbeneau, 1982).

The El Paso recharge project involves injection of treated wastewater into the Hueco Bolson, which includes areas in Texas, New Mexico, and Mexico. The injection wells are constructed so that water is recharged from the top of the saturated zone to the point where TDS levels approach 1,000 mg/l. This is generally in the interval from 107 to 290 m (350 to 960 ft). The formation throughout the injection interval is fluvial in nature and contains gravel, silt, and clay lenses. Horizontal hydraulic conductivity of the interval averages about 7.6 m/day (25 ft/day) with the porosity averaging about 20 percent. The transmissivity and specific yield are about 1,225 m²/day (13,000 ft²/day and 10 percent. The formation below the injection interval is lacustrine in nature and has a much lower permeability (Knorr and Cliett, 1985).

The Fred Hervey Water Reclamation Plant is a wastewater treatment facility designed for recycling water used by the residents in the northeast area of El Paso back to the Hueco Bolson to help meet the water needs of El Paso. The influent characteristics indicate a moderately weak sewage that is primarily do-

mestic in origin and has a minimum industrial component. TO avoid the presence of highly toxic compounds in the source water, an extensive public awareness campaign and vigorous enforcement of the industrial waste discharge regulations have been developed and maintained. In addition, there is no chlorination in the wastewater collection system.

The 10-mgd advanced wastewater treatment plant has two 5-mgd parallel process trains which include screening, degritting, primary settling, equalization, a two-stage PACT^R system, lime treatment, sand filtration, ozone disinfection, granular activated carbon filtration, and storage. Generally, the first part of the plant flow scheme concentrates on the removal of organic pollutants, nitrogen, and suspended solids, while the second part provides removal of pathogens, phosphorus, radioactive material, and heavy metals, with additional removal of suspended solids, dissolved organic compounds, and color, taste, and odor problems. In the PACT^R process, the wastewater is aerated so that bacteria can feed on some pollutants while others are adsorbed on powdered activated carbon. This process removes dissolved pollutants including organics and ammonia. Then, in a clarifier, bacteria and powdered carbon are separated from the water, and methanol is added for denitrification. The lime process at a pH level of 11.1 provides high vital and heavy metal removal performance. This process serves as the first step in the disinfection process. Ozone is used as the final disinfecting process because of its advantages over chlorine in disinfection performance and to avoid the formation of halogen, a disinfection by-product (DBP). A very low (0.25 mg/l) chlorine dose is provided ahead of storage to prevent buildup of film or growths in the clear well during the 8-hour storage periods. Granular activated carbon (GAC) filtering is incorporated in the process train as a polishing process in the removal of residual organic compounds (Knorr and Cliett, 1985).

The basic objective of the El Paso recharge project was to increase potable water supplies with the lowest practicable risk. Two basic criteria for design were to provide (1) maximum recovery of recharged water to minimize the costs and (2) adequate aquifer residence time to provide the opportunity for additional purification in the aquifer. The El Paso Water Utilities ground water recharge project has been operating since 1985 the recharge to Hueco Bolson aquifer. Ground water movement studies indicate that a downgradient spacing of 1,200 feet between injection and production wells would provide a residence time of 2 years. The design aquifer residence time of 2 years was chosen based on viral inactivation periods, which are on the order of months. The actual spacing in the field has the injection wells located approximately 1.2 km (.75 mile) upgradient from the existing production wells. This spacing should give residence times of at least 6 years. In addition, where future production wells are drilled to fill in a

5 well per section pattern, injection well to production well spacing will still exceed 370 km (1,200 ft). Actual monitoring of breakthrough at observation and recovery wells has proved to be difficult because of the similarity between recharge and formation water qualities. Subsequent field monitoring suggests that the minimum residence time is closer to 7 years (P. Buszka, personal communication, 1993). The produced water is chlorinated before reuse, with no other treatment. To date, there have been no operational difficulties (J. Balliew, personal communication, 1993).

The injection rate of 700 gpm for the wells is one-half to two-thirds of the capacity of production wells in the area, allowing for a decrease in well efficiency with time. The wells are periodically pumped for back-flushing and cleaning. There have been no long-term performance problems (J. Balliew, personal communication, 1993).

The quality of the water produced by the Fred Hervey Water Reclamation Plant is tested continually. Table 6.3 shows that the water produced is comparable to the water currently in the Hueco Bolson aquifer and that it meets state and federal regulatory levels for safe drinking water.

An ongoing U.S. Geological Survey study is focusing on the fate of disinfection by-products and on tracing injection waters. Results from monitoring have suggested that tribromomethane concentrations decrease with distance from injection wells. Associated with a decrease in tribromomethane concentrations is an increase in dibromomethane concentrations, although the mechanisms responsible for this have not been identified (P. Buszka, personal communication, 1993).

A breakdown of construction costs for the treatment plant and recharge system is shown in Table 6.4. Although this table is based on a $27 million bid cost, the final cost was $33 million.

Project managers selected the northernmost well fields serving El Paso for recharge because they offered the greater potential benefits to the city than the other well fields in the aquifer. The city of El Paso's Public Service Board owns most of the water in this locality and would receive the most benefit from an artificial recharge program. Further, the location of the recharge and production

wells suggests that virtually all of the recharged water will be recovered by the city because gradients are generally toward the city's production wells.

The discharge permit from the Texas Water Commission requires the monitoting of 23 variables, with 30-day average values to be used on most variables. Monitoring frequency also is specified. The permit limits are the same as the drinking water standard. The permit also requires less than 10 mg/l nitrate (N) and less than 5 NTU turbidity in each 8 hour batch of water.

El Paso is in an arid environment with a limited water supply and a water resource that is being depleted. Water management is a necessity. Ground water recharge operations are extending the lifetime of the Hueco Bolson aquifer. Water reclaimed from sewage is being injected with the intent of recovery for potable reuse with only chlorination. Monitoring suggests that injection waters are only now reaching production wells after seven years of production. No

significant problems have been identified during the relatively short operational history of this project.

Urbanization of Nassau and Suffolk counties on Long Island, with the attendant construction of highways, houses, shopping centers, industrial areas, streets, and sidewalks in previously undeveloped or agricultural areas, has caused a twofold water management problem: (1) disposal of stormwater runoff from impervious areas and (2) reduction in land-surface area available for infiltration of precipitation to naturally recharge the ground water reservoir. To obviate the need for costly trunk storm sewers to convey runoff to streams or coastal waters and to minimize the loss of recharge, shallow stormwater collection basins have been built to contain stormwater runoff and allow it to infiltrate to the water table. These stormwater infiltration basins have been used since 1935. Their number increased as development progressed on Long Island—from 14 basins in 1950, to more than 700 in 1960, to more than 2,100 in 1969, to more than 3,000 by 1986 (Seabum and Aronson, 1974; Ku and Simmons, 1986). The basins are located mainly in eastern Nassau County and western Suffolk County in the inner pan of Long Island.

In addition to facilitating the disposal of stormwater runoff in developed areas, the basins provide recharge to the ground water reservoir, which is the sole source of water for the 2.7 million inhabitants of Nassau and Suffolk counties. More than 10 percent of the 3,070 square-kin (1,200 square-mile area) in Nassau and Suffolk counties drains to recharge basins (Ku and Simmons, 1986). The collection and routing of stormwater from impervious areas to these basins have countered the loss of land-surface area available for infiltration of precipitation. In areas drained to the basins, ground water recharge from precipitation is probably equal to or slightly greater than recharge under predevelopment conditions (Seaburn and Aronson, 1974). Recharge provided by the stormwater basins partially replenishes the ground water withdrawn for use by Long Island residents and thereby helps retard seawater intrusion into the aquifers and the drying up of streams.

The Long Island recharge basins are a system Of unlined pits of various sizes and shapes excavated in moderately to highly permeable surficial sand and gravel deposits. They range in size from 0.1 to 30 acres and average between 1 and 2 acres. Most extend 3.1 to 4.6 m (10 to 15 ft) below land surface, but some are as deep as 12 m (40 ft) (Seaburn and Aronson, 1974). For the most pan, the basins are built in areas where the water table is sufficiently deep to remain below the floor of the basin most of the time.

Most of the basins are constructed with overflow structures that are not more than 3.1 m (10 ft) above the floor of the basin. The required capacity of the basin below the overflow altitude is estimated by multiplying the volume of water equivalent to 13 cm (5 inches) of rainfall on the total area drained to the basin by a factor ranging from 30 to 100 percent based on drainage area conditions such as land slope and percentage of paved area. A factor of 30 percent is used in most residential areas, whereas for industrial areas with a higher proportion of impervious surface the factor is as much as 100 percent. In calculating the design capacity, infiltration into the floor and sides of the basin during inflow is not considered, thereby providing a safety factor.

Various construction features are used to ensure operational efficiency. These include multilevel basin floor, retention basins, wells, and scarification of basin floors. The lower levels of a basin floor act as a settling area for inflowing sediment and trash and allow higher infiltration rates in the higher-level floor areas receiving relatively sediment-free water as overflow. Retention basins serve the same purpose as basins with integral settling areas except that they are separate basins connected by pipes or channels to adjacent or nearby basins. Where deposits of low hydraulic conductivity immediately underlie the basin

floor, wells often are installed beneath the basin floor to provide access to deeper-lying more permeable strata. Scarification of basin floors is also used to expose more permeable underlying deposits.

The surficial deposits into which the recharge basins are excavated consist mainly of unconsolidated outwash deposits and ground and terminal moraine deposits of the Wisconsin Glaciation. The outwash deposits are well-sorted sand and gravel of moderate to high permeability. The morainal deposits consist of poorly sorted till of moderate to low permeability. The surficial deposits overlie unconsolidated interbedded sand, gravel, silt, and clay of Cretaceous age. The Cretaceous deposits have high to low permeabilities, depending on lithology. The water supply for Nassau and Suffolk counties is obtained totally from various aquifers composed of both the glacial and the cretaceous sediments.

Water recharged by the Long Island basins is stormwater runoff from residential, industrial, and commercial areas and from highways. As such, it is highly variable in both quantity and quality. Quantity varies depending on the intensity and duration of the storm as well as the percentage of impervious area within the drainage basin. Quality varies depending on the land use activities within the drainage area, length of time of the runoff event, and time since the last runoff event. The stormwater runoff is piped directly to the recharge basin and receives no treatment except for the settling of sediment.

Most of the basins in which stormwater is impounded are commonly dry within a day or so after a storm, but some hold water perennially (Aronson and Seabum, 1974). Of the basins existing on Long Island in 1969, 9 percent were found to contain water more than 5 days after a 2.5 cm (1-inch) rainfall on the drainage area (Aronson and Seabum, 1974). Containment of water in these basins over prolonged periods occurs for one or more of these reasons: (1) they intersect either the regional or a perched water table, (2) they are excavated in materials of low permeability, or (3) sediment and debris accumulating on the basin floor reduce infiltration rates.

A study of water-containing basins revealed that those that drain commercial and industrial areas had the highest percentage, 28 percent, of clogging caused by sediment and debris. Only 7 percent of residential area basins and 9 percent of highway basins were clogged and thus contained water. The high percentage of water-containing commercial and industrial basins is probably

largely related to the large influxes of asphalt, grease, oil, tar, and rubber particles in storm runoff from adjacent parking fields (Aronson and Seabum, 1974). Although basins that drain highways might be expected to receive similar materials, the small percentage of highway basins that contain water probably reflects a combination of the more regular maintenance of these basins and the use of special structures incorporated into the basin floors.

The 2,124 recharge basins operating on Long Island in 1969 recharged an estimated annual total of 68,100 acre-feet of water (Seabum and Aronson, 1974). Infiltration rates were studied at three typical basins in Westbury, Syosset, and Deer Park. At Westbury infiltration rates computed from data collected during 63 storms ranged from 9.1 to 52 cm/hr (0.3 to 1.7 ft/hour) and averaged 27 cm/ hour (0.9 ft/hour). At Syosset, rates during 22 storms ranged from 9.1 to 55 cm/ hour (0.3 to 1.8 ft/hour) and averaged 24 cm/hour (0.8 ft/hour). At Deer Park, rates during 24 storms ranged from 3.0 to 15 cm/hour (0.1 to 0.5 ft/hour) and averaged 6.0 cm/hour (0.2 ft/hour) (Seabum and Aronson, 1974). These rates were observed under a wide range of meteorological conditions.

Five recharge basins representing different land use areas were monitored during 46 storms to determine quality of stormwater runoff and precipitation and quality of ground water immediately beneath the basins 1 or 2 days after the storm. Samples were analyzed to identify standard inorganic constituents, heavy metals, organic compounds, and bacteria. Conclusions from this study included (Ku and Simmons, 1986):

• Most of the load of heavy metals in the stormwater was removed during infiltration through the unsaturated zone, but nitrogen and chloride were not removed.

• The median number of indicator bacteria in stormwater ranged from 10⁸ to 10¹⁰ MPN (Most Probable Number) per 100 milliliter. Virtually no bacteria were detected in ground water beneath the recharge basins, indicating complete removal during percolation of stormwater through the unsaturated zone.

• Concentrations of pesticides in basin-bottom soils generally were much higher than those in stormwater, suggesting that pesticides are probably sorbed or filtered out in the soil layer.

• Use of recharge basins on Long Island to dispose of stormwater runoff

and to recharge ground water does not appear to have significant adverse effects caused by chemical and microbiological constituents.

There is no pre-recharge treatment of stormwater other than the settling of solids that is accomplished by the initial construction of settling areas on the basin floors or by separate retention basins. Thus the cost of the recharge operations is limited to land acquisition, basin construction, and basin maintenance. Basin maintenance consists mainly of collecting and removing bulk debris and cutting and removing grass on the basin floor. Scarification of basin floors is done as required. to break or loosen material on the basin floor or to remove a thin layer of clogging material.

The water recharged via these basins is not directly recovered. Rather, it becomes mixed with the water naturally recharging the ground water reservoir and is recovered by numerous public and private supply wells distributed throughout Nassau and Suffolk counties.

Most of the basins on Long Island are owned and maintained by local or state governmental agencies. Developers are required to construct storm sewers and recharge basins of adequate size for their project. On completion, recharge basin ownership and maintenance become the responsibility of the local government.

Ground water recharge of stormwater has proved to be a viable means of locally disposing of storm runoff in a rapidly urbanizing part of Long Island and, at the same time, countering the reduction of natural recharge to the ground water system caused by the increase in impervious areas. The Long Island aquifer system has been designated by EPA as the ''sole-source aquifer" for water supply in Nassau and Suffolk counties, a 3,070 square km (1,200 square mile) area having a population of 2.7 million. Studies have not shown any significant adverse impact on ground water quality, even after several decades of operation for some of the early basins. The stormwater receives no treatment other than the settling of sediment before it infiltrates, which for most basins occurs within hours to a few days following the storm event. Studies also have shown that in areas served by drainage basins, ground water recharge is equal to or slightly exceeds that occurring under predevelopment conditions. The storage and flow accretions to the aquifer system realized through the recharge of stormwater appear to far outweigh any potentially detrimental water quality impact.

Drainage wells to dispose of excess surface water in the city of Orlando and surrounding areas have been in operation since 1904 (Kimrey and Fayard, 1984). The drainage wells alleviate flooding and control lake levels in this area of closed drainage basin lakes and low topographic relief. Although drainage wells are used in other parts of Florida also, the Orlando area has the highest concentration of wells (Kimrey and Fayard, 1984). These wens are used to facilitate drainage in this internally drained karst environment and also to recharge the Floridan aquifer system, the primary source of water for the Orlando area and most of Florida.

In 1990, there were about 310 drainage wells within the greater Orlando area (Figure 6.8), an area of about 230 square km (90 square miles) (Bradner, 1991). These wells inject surface water by gravity into the Upper Floridan aquifer. More than half of the wells are 30 cm (12 inches) or more in diameter, but they range from 10 to 61 cm (4 to 24 inches) in diameter. The drainage wells range in depth from about 37 to 320 m (120 to 1,050 feet); median depth is about 120 m (400 ft). They are eased to or near the top of the Upper Floridan aquifer and then finished as an open hole in the aquifer.

The Orlando area is underlain by about 15 m (50 ft) of sand and silt that constitute the surficial aquifer. The surficial aquifer is, in turn, underlain by about 46 m (150 ft) of sandy clay, silt, and shell that constitute the intermediate confining unit. Beneath the intermediate confining unit lies the Floridan aquifer system, which is made up of about 460 m (1,500 ft) of limestone and dolomite. The Floridan aquifer system has been subdivided into three units—the Upper Floridan aquifer (91 to 120 km (300 to 400 ft) thick), the middle semiconfining unit of less permeable limestone (91 to 180 m (300 to 600 ft) thick), and the Lower Floridan aquifer (120 to 180 m (400 to 600 ft) thick). These aquifers have a lot of secondary porosity (karstic) and are very permeable; within the Orlando area, transmissivity of the Upper Floridan aquifer ranges from 4,700 to 37,000 square m per day (50,000 to 400,000 square feet per day), and transmissivity for the Lower Floridan aquifer ranges from 9,200 to 57,000 square m per day (100,000 to 600,000 square feet per day) (Tibbals, 1990). Water supply for the Orlando area is obtained mainly from the Lower Floridan aquifer, but some is pumped from the Upper Floridan aquifer as well. Virtually all of the drainage

FIGURE 6.9 Generalized hydrogeologic section in the Orlando area Bradner, 1991.

wells are completed in the Upper Floridan aquifer. Figure 6.9 shows a generalized hydrogeologic section beneath the Orlando area.

The water table in the surficial aquifer is higher in altitude than is the potentiometric surface of the Upper Floridan aquifer. However, the low-permeability materials in the intermediate confining unit impede the downward flow of ground water from the surficial aquifer to the Upper Floridan aquifer. The drainage wells "short circuit" the intermediate confining unit and provide for direct input of surface water to the Upper Floridan aquifer.

About 50 percent of the 310 drainage wells receive stormwater runoff directly from streets or other impervious areas, 45 percent receive lake or wetland overflow, and 5 percent receive air-conditioning return water or are unused at present but have received industrial effluent or sewage in the past. The water receives no treatment prior to injection. Grates, either directly on the wellhead or on the intake chamber leading to the well, screen coarse debris carried by the water. Water flowing directly from storm sewers is available intermittently during storm events. Many lake-overflow wells receive water continuously.

The cavity fiddled, highly permeable nature of the limestone of the Upper Floridan Aquifer allows the drainage wells to function more or less unattended. However, some wells are reported to have been completely filled by debris and

sand, and many wells must be cleaned periodically to maintain their effectiveness.

The potentiometric surface of the Upper Floridan aquifer normally is well below land surface throughout most of the area, so sufficient head is available for the drainage wells to accept large quantifies of water under gravity feed. However, during very wet years, such as 1959 and 1960, heads in the Upper Floridan aquifer increased to the point that the capacity of the drainage wells to recharge surface water to the aquifer decreased. In fact, some drainage wells actually flowed, and pressure injection pumps had to be installed to continue to use the wells for recharge until the potentiometric surface once again declined to below land surface (Kimrey, 1978).

Limited quantitative data are available on acceptance rates of drainage wells, but the range is reported as a few hundred to several thousand gallons per minute. An acceptance rate as high as 9,500 gpm was reported for a drainage well in west Orlando (Stringfield, 1933). Collectively, the 310 drainage wells in the Orlando area are reported to recharge an estimated 23 mgd of surface water to the Upper Floridan aquifer, but this is probably a low estimate (Bradner, 1991). A lake overflow well gaged from November 1987 through December 1988 averaged an inflow rate of 2.1 mgd, whereas a stormwater well reportedly accepted an average of 9,000 gpm during 1988 (Bradner, 1991).

Quality of inflow to the drainage wells has been summarized by Bradner (1991). According to that review, stormwater runoff contains high concentrations of total organic carbon, organic nitrogen, iron, lead, sulfate, and zinc, while concentrations of most anions and cations are lower in stormwater runoff than in water from the Upper Floridan aquifer (Wanielista et al., 1981; German, 1989). Concentrations of total organic carbon can range from 18 to 284 mg/l in runoff to Lake Eola in downtown Orlando (Wanielista et al., 1981). German (1989) found that inflow to the drainage wells frequently had detectable concentrations of many pesticides, with diazinon being detected in 77 percent of the samples collected and malathion being detected in 50 percent of the samples.

Studies of loads of nutrients and organic compounds entering the Upper Floridan aquifer at nine drainage well sites indicate that approximately 45,000 kg (100,000 pounds) of total nitrogen enters the Upper Floridan aquifer each year through drainage wells in central Florida (German, 1989). Studies of stormwater runoff also have reported sporadic detections of phthalates, com-

pounds widely used in the plastics industries, and polycyclic aromatic hydrocarbons, such as fluoranthene, pyrene, anthracene, chrysene, and benzo-a-pyrene, commonly associated with petroleum products (Wanielista et al., 1981; German, 1989).

The quality of water in the Upper Floridan aquifer in the Orlando area also has been studied. Schiner and German (1983) concluded that drainage wells and upper-producing-zone supply wells yielded water very similar in chemical characteristics, particularly major dissolved constituents. Water in the upper producing zone of the Floridan aquifer is primarily a calcium and magnesium-bicarbonate type. Bicarbonate generally accounts for more than 75 percent of the ions, and calcium and magnesium account for more than 85 percent of the cations. But in several supply wells, and several drainage wells, more than 25 percent of the anions consisted of sulfate plus chloride, and more than 15 percent of the cations consisted of sodium plus potassium. Water from the lower producing zone (also a calcium and magnesium-bicarbonate type water) was more consistent within its chemical type. This consistency may be because most samples from the lower producing zone were clustered in a small part of the study area or it may be because the zone is deeper and more isolated from surface influences.

The study also noted that water from drainage wells generally has slightly higher concentrations of most constituents than water from supply wells. The primary differences in water quality between drainage wells and supply wells were for total nitrogen, total phosphorus, total recoverable iron, and total coliform. The comparisons are shown in Table 6.5.

For some supply and drainage wells, color, hydrogen sulfide, iron, and manganese in these studies exceeded the National-Secondary Drinking Water Regulations, with the frequency of exceedance greater for drainage wells than for supply wells. Concentrations of metals and pesticides did not exceed the limit specified in Florida standards for potable ground water. Pesticide did not appear to be present in significant amounts.

Overall, the quality of water from the group of supply wells in the Orlando area is about the same as the quality of water from wells in adjacent areas where

no drainage wells exist. Water quality for drainage wells that receive street runoff was about the same as water quality for drainage wells that receive lake overflow, except for bacteria colony counts. Bacteria counts were considerably lower in wells that receive lake overflow than in those that receive direct street runoff.

Results of the Schiner and German (1983) study indicate that drainage well recharge has not caused widespread contamination of the Floridan aquifer. Bacterial contamination found in some drainage wells appears highly localized, and water from drainage wells would generally be acceptable for public supply use as long as bacteria are not present. Another study (Bradner, 1991) of 11 supply wells in urban Orlando, where the highest density of drainage wells exists, found calcium, potassium, sodium, chloride, and ammonia in significantly higher concentrations than is samples from hydrogeologically similar areas elsewhere. Significant differences in other constituents were not indicated.

Specific capacity data under pumping conditions are available for 21 drainage wells. At pumping rates ranging from 240 to 460 gpm, the wells reportedly had specific capacities that ranged from 27 to 1,900 gpm/feet, with the median being 310 gpm/feet. The 23 mgd of recharge from the wells in the Orlando area has created a mound in the potentiometric surface of the Upper Floridan aquifer of 1.2 m (4 ft) (Tibbals, 1990).

The stormwater used in the Orlando drainage wells does not receive any pre-recharge treatment except for that provided by detention in lakes for wens receiving lake overflow, and so pre-recharge treatment costs are limited. In addition, operating costs axe low because the system operates without attention except for the infrequent need to clean out accumulated debris and sediment from the wens in order to maintain their efficiency.

Most drainage wells are owned by municipalities or the Florida Department of Transportation. They are regulated as Class V injection wells under the Safe Drinking Water Act. No new drainage wells are currently being permitted.

Drainage wells are the most economical way of disposing of stormwater in the internally drained karst environment of the Orlando area. The drainage wells emplace, by gravity injection, 23 mgd of recharge to the Upper Floridan aquifer, which helps to balance the 51 mgd of municipal ground-water pumpage in the Orlando area. Some drainage wells accept urban stormwater runoff directly

from street drains, whereas others accept overflow from lakes into which stormwater has drained. Therefore, they introduce contaminants directly to one of the aquifers used for public supply in the area. Although some water quality effects have been noted, widespread contamination has not occurred, even though usage of drainage wells began in 1904. However, at least one plume of contaminated water in the vicinity of a drainage well is known. Spills of chemicals and/ or fuels caused by accidents along transportation routes are possible and pose a risk of introducing highly concentrated contaminants into the aquifer.

Alternate means of stormwater disposal in this karst area would require extensive mink sewers and pumping at considerable cost. Moreover, loss of the recharge provided by the drainage wells would result in a reduction of head in the Floridan aquifer system and the possibility of vertical encroachment of deeper-lying saltwater into supply wells. The current level of risk of severely contaminating the potable source aquifer is accepted, but no new drainage wells are being permitted.

Where water is scarce, municipal wastewater can serve as an unconventional source of supply that can be integrated into the regional water supply system. In Israel, the increased demands for high-quality water and the shortage of natural water sources have resulted in the development of strategies to improve the quality of secondary effluent to make it suitable for nonpotable uses, especially unrestricted agricultural reuse. The best example of this approach is the Dan Region Wastewater Reclamation Project,^* which provides for the collection, treatment, recharge, and reuse of the wastewater from the largest metropolitan area of the country, including Tel Aviv-Jaffa and several other neighboring municipalities. The project serves a total population of about 1.3 million with an average municipal wastewater flow of 72 million gallons per day.

The recharge-recovery method developed and practiced in the Dan Region project relies on the soil-aquifer treatment (SAT) concept. Partially treated effluent percolates through the unsaturated sod zone (fine sand) until it reaches the ground water. Itmoves radially in the aquifer until it reaches recovery wells designed to pump the recharge water for supply (Figure 6.10). Depths for ground water range from 15 to 45 m (49 to 150 ft) for the various sites. Distances between recharge basins and recovery wells range from 320 to 1,500 m (1050 to 4900 ft). If the recovery wells are adequately spaced, the recharge and recovery facilities can be operated to confine the recharged effluent between the recharge

FIGURE 6.10 SAT system in the Dan Region project, Israel.

areas and the recovery wells. This underground subbasin is dedicated to the treatment and storage of effluent and represents only a small percentage of the regional aquifer. The recharge water, which can be traced and monitored by means of observation wells, is usually of high quality. It is generally appropriate for industrial uses, unrestricted agricultural uses (including irrigation of vegetables to be eaten raw and livestock watering), nonpotable municipal uses, and recreational uses. Accidental drinking of recharge water does not present a significant health hazard because of its high microbiological quality.

Recharge is done with spreading basins to take advantage of the purification capacity of both the unsaturated zone and the aquifer. Operation of the recharge basins is intermittent; flooding periods are alternated with adequate drying periods to maintain high infiltration rates and to allow oxygen penetration into the soil to enhance the purification capacity of the system. Although most of the purification takes place during vertical flow through the upper soil layer and the whole unsaturated zone, additional purification (mainly breakdown of slowly biodegradable organics) is gained during horizontal flow in the aquifer, and aerobic properties that were lost in sections where anoxic conditions prevail below the recharge basins are regained.

Stage one of the Dan Region project has been in operation since 1977. For the first 2 years, the wastewater underwent biological treatment in oxidation

ponds with recirculation and chemical treatment by the high lime-magnesium process; the water then moved to polishing ponds for partial free ammonia stripping and natural recarbonation. The combination of oxidation ponds and lime treatment is roughly equivalent to secondary treatment, although the effluent quality is different in some respects. In October 1989, chemical treatment was discontinued and the oxidation pond effluent was conveyed to the polishing ponds.

Stage two of the project has been in operation since 1978. About 60 percent of this wastewater is conveyed to a mechanical-biological treatment plant, where it undergoes primary treatment and secondary treatment by activated sludge with nitrification-denitrification. The remainder goes through oxidation and polishing ponds, parallel to the mechanical-biological treatment plant. The long detention times in the polishing ponds produce a high-quality effluent with relatively low algae content.

The recharge sites are located in areas of rolling sand dunes near the Mediterranean coast underlain by a calcareous sandstone aquifer—one of the three main potable water supplies of the country. The climate of the zone is typically Mediterranean. Summers are warm and dry, and winters are mild with rainy spells. The average annual precipitation is 500 to 600 mm (20 to 24 in). The average temperatures usually range between 20 and 30ºC in summer and between 10 and 20ºC in winter.

The recharge operation in the second stage is carried out at two sites located south of the treatment plant. One recharge site consists of four basins covering a net area of about 59 acres; each basin is divided into four subbasins. The depth of the unsaturated zone below the recharge basins varies between 27 and 36 m (89 to 120 ft). A ring of recovery wells spaced 300 to 400 m apart surrounds the recharge areas on the northern, western, and southern sides; they are located between 320 and 1,100 m (1,050 to 3,600 ft) from the nearest recharge basin. At some locations, two separate wells are drilled to different subaquifers or to different layers of the same subaquifer. A monitoring network of 20 observation wells was established between the recharge basins and the recovery wells; they are located between 20 and 570 m (65 to 1,870 ft) from the recharge basins.

The second recharge site consists of three basins covering a net area of about 44 acres; each basin is divided into three subbasins. The depth of the unsaturated zone below the recharge basins varies between 40 and 43 m (131 to 141 ft), and it is similar to the first recharge site. A ring of recovery wells surrounds the recharge areas located between 350 and 1,500 m (1,150 to 4,900 ft) from the nearest recharge basins. A monitoring network of 12 observation wells was established between the recharge basins and the recovery wells; they are located between 20 and 300 m (65 and 980 ft) from the recharge basins.

The spreading basins are flooded intermittently to maintain high infiltration rates and to enhance effluent purification during percolation. The water depths in the basins are generally below 0.6 m (2.0 ft). A short recharge cycle is

employed, usually consisting of 1 day flooding and 2 to 3 days drying, to ensure that aerobic conditions predominate in the unsaturated zone and in the aquifer.

The clogging layer on the basin bottom may reach a thickness of about 1.5 cm (0.6 in). The clogging layer is ''scratched" about once a week or every 2 weeks, especially in the winter when drying is slow, and "shaved off" completely every 1 or 2 months to restore infiltration capacity. There is no accumulation of clogging material deeper in the soil.

The major purification processes occurring in the soil-aquifer system are slow-sand filtration, chemical precipitation, adsorption, ion exchange, biological degradation, nitrification, denitrification, and disinfection. To illustrate the purification effect of soil-aquifer treatment (SAT), quality data were evaluated for the mechanical-biological plant effluent (RE) before recharge and SAT and for the reclaimed water (RW) after SAT (Tables 6.6., 6.7, 6.8).

The relatively high removal efficiency obtained for a variety of parameters confirms that SAT is an integral part of the municipal wastewater treatment process in the Dan Region project. The removal of suspended solids (mostly organics) and of biochemical oxygen demand was virtually complete. The average total and filtered chemical oxygen demand were reduced from 69 and 46 mg/ l respectively, to 12.5 mg/l in each case. Average total organic carbon and dissolved organic carbon were reduced from 20 and 13 mg/l respectively, to 3.3 mg/l. Ultraviolet absorbance was reduced significantly. The concentration of detergents was reduced from 0.5 to 0.08 mg/l, and that of phenols from 8 to less than 2 µg/l. Because of the efficient and reliable removal of organics, the soil-aquifer system can be regarded as a biological treatment unit.

Total and filtered nitrogen were reduced from 15.7 and 14.4 mg/l, respectively, to 7.8 mg/l. Ammonia was reduced from 7.6 mg/l as nitrogen to less than 0.95 mg/l. While most nitrogen in the recharge water is found in the unoxidized forms of ammonia and organic nitrogen, the residual nitrogen in the well water consists essentially of nitrates. Thus complete nitrification and partial denitrification occur in the soft-aquifer system. Phosphorous removal efficiencies were 25 percent in the oxidation and polishing ponds and 74 percent in the mechanical-biological treatment plant. The remaining phosphorous was removed efficiently by SAT from 3.4 mg/l in the recharge water to 0.02 mg/l in the recovered water, a concentration is similar to that in the natural ground water. Coliform bacteria, E. coli, S. faecalis, and enteroviruses were not detected in the recovered water.

The salinity of the recovered water is acceptable for unrestricted irrigation of all crops. The sodium adsorption ratio (SAR) in the recovered water is similar to that of the recharge water (about 4.6), which is acceptable for unrestricted

irrigation. The sodium concentration in the recovered water is similar to that of the recharge water (about 200 mg/l), so the cation exchange process was exhausted in the area surrounding the recharge basins. Adsorption of sodium and release of calcium and magnesium are still occurring in areas further away from the recharge basins. Potassium, which is found in relatively low concentrations in the recharge water, is still removed (30 percent) by SAT. Boron in the recharge water (0.5 mg/l) is only slightly removed (10 percent) by SAT. The concentrations of trace elements in the water after SAT are below the recommended maximum limits for irrigation water used continuously on all soils.

Since November 1989, the recovered water from the Dan Region project has been transferred by a 100-kin-long (62.5-mile-long) conveyance main called the Third Line to a distribution net for irrigated areas in the southern part of Israel. At the head of the conveyance system, the recovered water is disinfected by chlorination. Along the Third Line, there are four open operational reservoirs, each with a capacity of 13 to 26 million gallons. The reservoirs are used to compensate for changes in water pressure in the main pipe and to facilitate control of the system during peak demand periods. The retention time of water in the reservoirs is not constant and depends on operational restrictions. The monthly consumption of the recovered water conveyed by the Third Line during 1991 has ranged from 2.6 to 29 million gallons. During 2 years of operation, about 44,000 million gallons reclaimed water from the Dan Region project was reused for unrestricted irrigation in the southern part of Israel.

The soil-aquifer treatment (SAT) system as applied in the Dan Region Project is an efficient, low cost process ($0.03/m³ for operation and maintenance only) for water reuse. Although the concentrations of several toxic substances in the recovered water are below the maximum permissible limits for drinking water, and turbidity is reduced by SAT, the recovered water is used only for nonpotable purposes. Overall, the recharge activities in the Dan Region illustrate how an unconventional source of water can be managed to increase the supply available for a variety of nonpotable uses. The very high quality of recovered water obtained after SAT makes the water suitable for agricultural uses (including unrestricted irrigation of vegetables to be eaten raw and livestock watering), industrial uses, nonpotable municipal uses (lawn irrigation and toilet flushing), and recreational uses. The main advantages of incorporating SAT are that it provides seasonal and multiyear underground storage, it is reliable, it provides a safety barrier, and it could lead to improved public acceptance of water reuse.

Argo, D. G., and N. M. Cline. 1985. Groundwater Recharge Operations at Water Factory 21, Orange County, California. In Artificial Recharge of Groundwater, T. Asano ed. Boston, Mass.: Butterworth.

Aronson, D. A., and G. E. Seabum. 1974. Appraisal of operating efficiency of recharge basins on Long Island, New York, in 1969. U.S. Geol. Surv. Water Supply Paper 2001-D, 22 pp.

Baird, R. B. 1987. GC-Negative ion CIMS and Ames mutagenicity assays of resins in advanced wastewater treatment facilities. In Advances in Sampling and Analysis of Organic Pollutants from water, I. H. Suffer and M. Malaiyandi, eds. Vol. 2, American Chemical Society, Advances in Chemistry, Washington, D.C.

Bouwer, E. J., P. L. McCarty, H. Bouwer, and R. C. Rice. 1984. Organic contaminant behavior during rapid infiltration of secondary wastewater at the Phoenix 23rd Avenue Project. Water Res. 18:463-472.

Bouwer, H. 1992. Agricultural and municipal use of wastewater. Water Sci. Technol. 26:1583-1591.

Bouwer, H., R. C. Rice, and E. D. Escarcega. 1974a. High-rate land treatment. L Infiltration and hydraulic aspects of the Flushing Meadows Project. J. Water Pollut. Contr. Fed. 46:835-843.

Bouwer, H., J. C. Lance, and M. S. Riggs. 1974b. High-rate land treatment. II. Water quality and economic aspects of the Flushing Meadows Project. J. Water Pollut. Contr. Fed. 46:844-859.

Bouwer, H., R. C. Rice, J. C. Lance, and R. G. Gilbert. 1980. Rapid-infiltration research—The Flushing Meadows Project, Arizona. J. Water Pollut. Contr. Fed. 52:2457-2470.

Bouwer, H., and W. L. Chase, Jr. 1984. Water reuse in Phoenix, Arizona. Pp. 337-353 in Proceedings, Water Reuse Symposium III: Future of Water Reuse, San Diego, California, August 26-31. American Water Works Association.

Bouwer, H., and R. C. Rice. 1984. Renovation of wastewater at the 23rd Avenue rapid-infiltration project. J. Water Pollut. Contr. Fed. 56:76-83.

Bouwer, H. and Rice, R. C. 1989. Effect of water depth in groundwater recharge basins on infiltration. J. Irrig. and Drain. Engr., 115. 556-567.

Bradner, L. A. 1991. Water quality in the upper Floridan aquifer in the vicinity of drainage wells, Orlando, Florida-U.S. Geol. Surv. Water Resour. Invest. Rep. 90-4175, 57 pp.

Carlson, R. R., K. D. Lindstedt, E R. Bennett, and R. B. Hartman. 1982. Rapid infiltration treatment of primary and secondary effluents. J. Water Pollut. Contr. Fed. 54:270-280.

Charbeneau, R. J. 1982. Groundwater resources of the Texas Rio Grande basin. Natur. Resour. J. 22(4):957-970.

German, E. R. 1989. Quantity and quality of stormwater runoff recharged w the Floridan aquifer system through two drainage wells in the Orlando, Florida, area. U.S. Geol. Surv. Water Supply Paper 2344, 51 pp.

Gilbert, R. G., C. P. Serba, R. C. Rice, H. Bouwer, C. Wallis, and J. L. Melnick. 1976. Virus and bacteria removal from wastewater by land treatment. Appl. Environ. Microbiol. 32:333-338.

Hartling, E. C. 1993. Impacts of the Montebello Forebay Groundwater Recharge Project. Bull. Calif. Water Pollut. Contr. Assoc. 29(3):14-26.

Kimrey, J. O. 1978. Preliminary appraisal of the geohydrologic aspects of drainage wells, Orlando area, central Florida. U.S. Geol. Surv. Water Resour. Invest. Rep. 78-37, 24 pp.

Kimrey, J. O., and L. D. Fayard. 1984. Geohydrologic reconnaissance of drainage wells in Florida. U.S. Geol. Surv. Water Resour. Invest. Rep. 84-4021, 67 pp.

Knorr, D. B. and T. Cliett. 1985. Proposed groundwater recharge at El Paso, Texas. In Artificial Recharge of Groundwater, T. Asano ed. Boston, Mass.: Butterworth.

Ku, H. F. H., and D. L. Simmons. 1986. Effect of urban stormwater runoff on ground water beneath recharge basins on Long Island, New York. U.S. Geol. Surv. Water Resour. Invest. Rep. 85-4088, 67 pp.

Lance, J. C., R. C. Rice, and R. G. Gilbert. 1980. Renovation of wastewater by soil columns flooded with primary effluent. J. Water Pollut. Contr. Fed. 52:381-388.

Leach, L. E., C. G. Enfield, and C. C. Harlin, Jr. 1980. Summary of Long-term Rapid Infiltration System Studies. EPA-600/2080-165. U.S. Environmental Protection Agency. Ada, Okla.

McCarty, P. L., M. Reinhard, N. L. Goodman, J. W: Graydon, G. D. Hopkins, K. E. Mortel-roans, and D. G. Argo. 1982. Advanced Treatment for Wastewater Reclamation at Water Factory 21. Techn. Paper No. 267. Department of Civil Engineering, Stanford University. Stanford, Calif.

McCarty, P. L., B. E. Rittman, and E. J. Bouwer. 1986. Microbiological processes affecting chemcial transformations in groundwater. Pp. 89-116, in Groundwater Pollution Microbiology, G. Bitton and C. P. Gerba, eds. New York: John Wiley.

Nellor, M. H., R. B. Baird, and J. R. Smyth. 1984. Health Effects Study—Final Report. NTIS No. PB-84191-568 County Sanitation Districts of Los Angeles County. Whittier, Calif.

Orange County Water District. 1991. Groundwater Management Plan. Orange County Water District. Fountain Valley, Calif.

Rice, R. C., and H. Bouwer. 1984. Soil-aquifer treatment using primary effluent. J. Water Pollut. Contr. Fed. 56:84-88.

Rice, R. C., and R. G. Gilbert. 1978. Land treatment of primary sewage effluent: Water and energy conservation. Pp. 33-36 in Hydrology and Water Resources in Arizona and the Southwest. Tucson, Adz: University of Arizona Press.

Schiner, G. R., and E. R. German. 1983. Effects of recharge from drainage wells on quality of water in the Floridan aquifer in the Orlando area, central Florida. U.S. Geol. Surv. Water Resour. Invest. Rep. 82-4094, 124 pp.

Seabum, G. E., and D. A. Aronson. 1974. Influence of recharge basins on the hydrology of Nassau and Suffolk counties, Long Island, New York. U.S. Geol. Surv. Water Supply Paper 2031, 66 PP.

Semmens, M. J., and T. K. Field. 1980. Coagulation: Experiences in organics removal. J. Am. Water Works Assoc. 72:476-483.

Sloss, E. M. 1993. Epidemiological assessment of groundwater recharge with reclaimed water in Los Angeles County. Proposal No. 93-019, submitted to the Water Replenishment District of Southern California by RAND, Santa Monica, Calif.

State of California. 1987. Report of the Scientific Advisory Panel on Groundwater Recharge with Reclaimed Wastewater. Prepared for the California State Water Resources Control Board, Department of Water Resources, and Department of Health Services. Sacramento, Calif.

Stringfield, V. T. 1933. Ground-water investigations in Florida. Flor. Geol. Surv. Geol. Bull. 11, 33 PP.

Thurman, E. M. 1979. Isolation, Characterization, and Geochemical Significance of Humic Substances from Groundwater. Ph.D. dissertation. University of Colorado, Boulder, Col.

Tibbals, C. H. 1990. Hydrology of the Floridan aquifer system in east central Florida. U.S. Geol. Surv. Prof. Paper 1403-E, 98 pp.

U.S. Environmental Protection Agency. 1992. Guidelines for Water Reuse. Technology Transfer Manual EPA/625/R-92/004. U.S. Environmental Protection Agency. Washington, D.C. 247 pp.

Wanielista, M. P., Y. A. Yousef, and S. J. Taylor. 1981. Stormwater Management to Improve Lake Water Quality. Submitted to Municipal Environmental Research Laboratory, Edison, N.J. Orlando, Fla.: University of Central Florida. 225 pp.

Wesner, G. M. 1991. Annual Report, Orange County Water District Wastewater Reclamation and Recharge Project, Calendar Year 1990. Prepared for Orange County Water District, Fountain Valley, Calif.