Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 Environmental Issues Desalination has been used around the world on a municipal scale for many decades, yet it is still considered by many to be a ânewâ option for addressing water supply needs. Part of the hesitancy to accept this tech- nology comes from concerns over potential environmental impacts of desalination, which have not been fully quantified. The environmental issues surrounding desalination fall into four general categories, which are reviewed in this chapter: (1) impacts from the acquisition of source water, (2) impacts from the management of waste products and concen- trate from the desalination process, (3) issues with desalinated product waters, and (4) the impacts of greenhouse gas emissions from these en- ergy-intensive processes. Technologies and approaches to mitigate these impacts are also discussed. Environmental impact assessments for any project also include concerns that are not addressed in this chapter, such as environmental effects of plant construction, material use, potential releases to the air, disposal of used membranes, and socioeconomic con- siderations. These issues are discussed in a recent World Health Organization report âDesalination for Safe Water Supplyâ (WHO, 2007). Human use of any water supply will have some environmental impacts; ultimately, consideration of the potential impacts of desalination will need to be weighed against the impacts from other water supply alterna- tives. SOURCE WATER ACQUISITION Desalination technologies can provide high-quality water tailored to the userâs needs, and many otherwise unusable sources of water (e.g., oceans, estuaries, brackish aquifers, wastewater) can be treated by de- salination technologies. For each type of source water, there are distinct environmental considerations when that water is withdrawn. In coastal surface waters, issues of impingement and entrainment of marine organ- 108
Environmental Issues 109 isms are paramount. For inland aquifer systems, the renewability of the resource and land subsidence over time are significant issues. Marine Water Intake Issues: Impingement and Entrainment Pumps bringing large volumes of ocean or estuary water into desali- nation plants can cause impingement and entrainment. Impingement, defined as the pinning and trapping of fish or other larger organisms against the screens of the intake structures, can cause severe injury and death to organisms. Entrainment occurs when intake pipes take in small aquatic organisms, including plankton, fish eggs, and larvae, with the intake water. Organisms that are pulled into the system will die if they are subjected to high temperatures or are crushed by high-pressure mem- branes. Intakes for desalination plants co-located with power plants are regulated under Section 316B of the Clean Water Act, although states may choose to apply these regulations to stand-alone plants as well (see Box 5-1). Power plants have been well studied with regard to impingement and entrainment of organisms. Most desalination plants will take in far less water, roughly an order of magnitude lower than medium-sized power plants. However, very large stand-alone desalination plants might require comparable quantities of intake water if substantial volumes of water are needed for concentrate dilution. Intakes from a single large power plant are estimated to kill billions of juvenile-stage fishes each year and may affect recruitment of juvenile fish and invertebrates into the adult popula- tions (Brining et al., 1981). It has been estimated that the magnitude of loss from one large power plant is equivalent to the loss of biological productivity of thousands of acres of habitat (York and Foster, 2005). The decomposition of the dead organisms can reduce the oxygen in the water, causing an additional stress in the area. However, the population- level impacts of mortality due to entrainment of marine organisms may or may not be substantial because the normal mortality of larval organ- isms in the marine environment is generally very high. The impacts and the acceptability of this loss will likely vary from place to place. There are technologies and practices that can be applied to reduce the amount of impingement and entrainment associated with coastal desali- nation. To reduce the amount of entrainment, it is possible to reduce intake during the times when eggs and larvae are abundant in the water, and windows of operation can be set to minimize the entrainment of eggs and larvae of the species of concern. If intake pipes are located in deeper parts of a bay, there will also be fewer organisms that could be impinged or entrained (San Francisco Bay Conservation and Development Com- mission, 2005). Entrainment can also be reduced substantially by
110 Desalination: A National Perspective BOX 5-1 Environmental Regulatory Framework for Desalination Several national regulations serve as the legal framework to minimize envi- ronmental impacts from desalination processes. The most pertinent regulations are associated with the Clean Water Act, although the Safe Drinking Water Act and its Underground Injection Control program are also described here. Addi- tional environmental regulations that need to be considered in the permit process are described in Chapter 7. Clean Water Act Under Section 316(b) of the Clean Water Act (CWA; P.L. 92-500), the Envi- ronmental Protection Agency (EPA) has developed regulations that require that the location, design, construction, and capacity of cooling water intake structures reflect the best technology available for minimizing adverse environmental im- pacts. Phase 1, promulgated in 2001, addresses the intake structures of new power plants. Phase 2 addresses the intake structures of large existing electric generating plants and requires these plants to meet impingement and entrain- ment standards for reducing the number of organisms affected. As of July 2007, the Phase 2 regulations were suspended while the EPA addresses several is- sues remanded by the 2nd Circuit Court of Appeals.1 In the United States, effluent discharges are federally regulated by the CWA. The regulatory program includes the National Pollutant Discharge Elimination System (NPDES), established by Section 402 of the CWA, which sets limits on the amount of various pollutants that a point source (i.e., the desalination plant) can discharge into a surface water body in a specific time period. Effluent limits can be technology based or water quality based, but they are all performance standardsâthat is, the permittee is free to use any combination of process modi- fication, end-of-pipe treatment, or other strategies to meet them. Water quality standards can also vary depending on the specified use of the particular water body into which the concentrate is disposed. NPDES permits typically quantify areasâtermed mixing zonesâwhere surface waters may exceed water quality standards due to point source discharges. If state regulatory programs meet the EPA requirements, the programs can be delegated to be administered by the states; therefore, regulations may vary somewhat from state to state. Effluent limits for desalination plants may specify pH, metaphosphates, chlorides, dissolved oxygen, conductivity, copper, iron, radium, total dissolved solids (TDS), total nitrogen, sulfide, ammonia, turbidity, radionuclides, selenium, and others. EPA (or the delegated state regulatory pro- gram) specifies the monitoring requirements, frequency of testing, and reporting, and the monitored species can be regionally variable. Some states require whole effluent tests for desalination concentrate in addition to chemical-specific numeri- cal limits (Mickley et al., 1993). Whole effluent toxicity testing may include acute tests of 96 hoursâ duration using larval or juvenile fish and invertebrates, with survival as the end point, and chronic tests of 7 days in duration using early life stages of a fish and an invertebrate, considering metrics continued 1 For more information, see http://www.epa.gov/waterscience/316b/index.html.
Environmental Issues 111 such as growth. Local species may also be used instead of âstandardâ bioassay organisms if a bioassay has been developed for them and is approved by EPA. Safe Drinking Water Act The Safe Drinking Water Act (SDWA) regulates the levels of contaminants permitted in drinking water supplies and applies to every public water system in the United States. Based on data describing how often a particular contaminant occurs in the environment, how humans are exposed to it, and potential health effects of exposure, EPA sets a maximum contaminant level goal (MCLG), the level of a contaminant in drinking water below which there is no known or ex- pected health risk, including a margin of safety. These goals are not enforceable because they do not take available technology into consideration and are some- times set at levels that public water systems cannot meet. EPA proposes an enforceable standard in the form of a maximum contaminant level (MCL), which is the maximum amount of a contaminant allowed in water delivered to a user of a public water system. Every 5 years, EPA establishes a list of contaminants that are known or anticipated to occur in public water systems and may require future regulation under SDWA.2 EPA oversees deep-well injection of desalination concentrate through its un- derground injection control program within the SDWA. EPA has developed the following classification for injection wells (EPA, 2007b): â¢ Class I: wells that inject hazardous waste; â¢ Class II: wells associated with the oil and gas industry; â¢ Class III: wells that inject fluids for the extraction of minerals; â¢ Class IV: wells that inject hazardous or radioactive waste into a formation within one-quarter mile of a drinking water source; and â¢ Class V: all other injection wells not covered by Classes I-IV. These classifications each have associated standards and associated regula- tions. States, rather than EPA, generally enforce the program and issue permits. Subsurface injection of desalination concentrates are covered by the states un- der regulations for either Class I or V injection wells. reducing water intake volumes. Reducing the size of mesh of screens in intakes can reduce entrainment but will increase impingement. However, rotating screens and other types of technologies can minimize the intake of aquatic organisms. If intakes are placed below the surface through the use of beach wells or other subsurface intakes (see Chapter 4), the prob- lems of entrainment of marine organisms are largely eliminated. Technologies for reducing impingement and entrainment are discussed in detail in Chapter 4. Co-location of a desalination plant with an existing power plant takes advantage of existing intake structures (see also co-location in Chapter 7). Typically, a co-located desalination plant takes its source water from 2 http://www.epa.gov/safewater/sdwa/30th/factsheets/standard.html#4.
112 Desalination: A National Perspective the power plant discharge water; thus, as long as the power plant is oper- ating, the desalination facility does not increase the impacts from impingement and entrainment. However, should the power plant discon- tinue operating on an interim or permanent basis or if once-through cooling practices are phased out, water withdrawals would have to con- tinue to provide source water to the desalination plant. It is worth noting that many of the nationâs power plants were sited decades ago, before the adverse environmental impacts of their intake structures were understood and before many of the current federal environmental legislation and regulations were in place, and some of the existing power plant intakes are located in areas where they create considerable environmental dam- age. Thus, the potential source water impacts of co-located desalination facilities still need to be considered. Brackish Groundwater Source Issues Some inland and coastal communities utilize brackish groundwater as a source for desalination, and withdrawal of brackish groundwater creates a quite different set of environmental concerns, including the physical sustainability of the aquifer and the potential for subsidence. Following a brief overview of brackish water resources in the United States, the potential environmental impacts from brackish groundwater withdrawal are discussed in more detail. A brackish aquifer is a geologic deposit of water-bearing permeable rock or unconsolidated materials from which brackish groundwater can be usefully extracted using a well. The processes that generate brackish groundwater depend on the site-specific hydrogeology and geochemistry. In some cases, high levels of dissolved solids are derived from the pres- ence of connate water (i.e., seawater trapped at the time of original deposition), but in most inland brackish water systems these original sol- utes have long since been flushed away. In arid and semi-arid areas typical of the western United States, the major sources of salinity in groundwater are the evaporative concentration of solutes from precipita- tion and dissolution of minerals in the subsurface. In the humid east and other areas with higher groundwater recharge rates, major solutes in brackish waters originate from dissolution reactions of the water with minerals (e.g., halite [NaCl], gypsum [CaSO4], anhydrite [CaSO4 â¢2H2O], calcite [CaCO3], dolomite [CaMg(CO3)2]) present in the aquifer frame- work. Coastal aquifers form another class of natural brackish water created from mixing of groundwater that is discharging to the ocean. Under natu- ral conditions most groundwater in coastal areas discharges directly to the ocean (Figure 5-1). The processes of molecular diffusion and hydro-
Environmental Issues 113 dynamic dispersion (mixing by movement of fluids through a porous media) create a brackish zone of dispersion or a mixing zone. Coastal groundwater pumping can cause seawater intrusion that increases the thickness of the brackish water zone of dispersion. Brackish water from irrigation return flows can also be utilized as desalination source water, although the quantity and quality typically vary by season and region. In Colorado, some desalination plants use alluvial groundwater with elevated salinity as a result of agricultural land use in the drainage basin (e.g., Platte River). Development of this water source for desalination is site specific as to both quantity and quality. Brackish groundwater exists at less than 305 m (1,000 feet) across much of the United States (Feth, 1965) (see Figure 1-1). This groundwa- ter consists of highly variable concentrations of dissolved solids and ranges from slightly brackish to brines with salt concentrations many times the concentration of seawater. The distribution, volume, and water quality of brackish water aquifers in the United States are largely un- known. Some states, such as New Mexico (Huff, 2004a, 2004b; New Mexico Office of the State Engineer, 2004) and Texas (Brackish Groundwater Manual, 2003), have made brackish water inventories based on existing data. Huff (2004a) estimated that huge quantities of brackish water (16 trillion m3 or 13 billion acre-feet) exist in New Mexico relatively close to the surface, some fraction of which could be desalinated for human use. However, there have been no national as- sessments, and the current regional assessments exhibit inadequate detail FIGURE 5-1. Diagram showing typical discharge of potable groundwater to the ocean and zone of dispersion (mixing zone). SOURCE: USGS (2007a).
114 Desalination: A National Perspective necessary for water resource management. Detailed site-specific evalua- tions, such as those conducted by the El Paso and Fort Bliss desalination facilities (see Box 5-2), will be necessary to assess the quantity and qual- ity of water available for a given desalination facility. Nevertheless, a national compilation of existing data and regional evaluations of flow and solute boundary conditions, thickness, extent, and hydraulic conduc- tivity of major brackish aquifer systems in the United States could provide the framework for a potentially greater utilization of brackish groundwater resources. Physical Sustainability Development of a brackish aquifer system for water supply demands an understanding the sustainability and renewability of the aquifer, in terms of both water quality and water quantity. The concept of physical sustainability of a natural resource has been defined in many ways. The United Nations Brundtland Commission (1987) popularized the term in the environmental sense when it defined sustainability as âthe ability to meet the needs of the present generation without compromising the abil- ity of future generations to meet their needs.â For the present report, a more conservative viewpoint is taken; a physically sustainable aquifer system is considered to be one in which recharge over human time frames approximately equals withdrawals and discharges from both an- thropogenic and natural processes (i.e., renewability). Groundwater withdrawals that exceed the recharge capacity of the aquifer are some- times referred to as groundwater mining. Under these circumstances, continued withdrawals may deplete the groundwater resource, create subsidence (discussed below), or affect the quality and quantity of adja- cent water bodies or aquifers. Because the hydrology of groundwater, lakes, streams, and wetlands are frequently interconnected, the removal of water from one source means less water for one or more of the other sources. In terms of water quality, sustainable aquifers are defined here as those having concentrations that do not change significantly beyond the natural variability over human time frames. Solute concentrations and their ionic ratios are naturally variable throughout all aquifer systems (Hem, 1986), and increased pumping can induce groundwater flow and, thus, solutes from adjacent, underlying, and overlying aquifers or surface waters. Induced groundwater flow will, in most cases, lead to changes in water quality due to chemical reactions and transformations within the aquifer matrix (e.g., ion exchange, dissolution, precipitation). Mineral precipitation and dissolution in the aquifer matrix can potentially alter the hydraulic conductivity of the aquifer over time (Johnson et al., 2005).
Environmental Issues 115 Box 5-2 Kay Bailey Hutchison Desalination Plant A desalination project was proposed for the El Paso, Texas, area after the 50-year Texas State-Wide Water Resource Management Plan for Far West Texas Region indicated that the projected future population growth of the El Paso area would experience water demands in excess of supplies. This proposed shortfall would occur in spite of the already large consumption decline associated with conservation and rate-structure change practices that resulted in a per cap- ita decline from 0.870 m3/day (230 gallons per day) in 1977 to 0.518 m3/day (137 gallons per day) in 2006. At the same time the nearby U.S. Army facility at Fort Bliss was expanding its mission and sought to increase its water supply. Looking at all possible water sources, El Paso Water Utilities (EPWU) studies indicated that desalinated water would be less costly than using reclaimed sewage or im- porting additional water but approximately twice as expensive as current groundwater and surface water supplies. A proposal was made for a cooperative effort between the city of El Paso and the U.S. Army to develop a 10,400 m3/day (27.5 million-gallon-per-day) desalination facility using brackish groundwater from the Hueco Bolson. The capital cost of the source water, desalination facility, and management of by-product concentrate was to be $87 million (2005 U.S. dollars). Previous pilot plant operation by EPWU in 1993-1994 suggested the likely success of using reverse osmosis technology with Hueco Bolson water. The source water would have TDS ranging between 1,200 and 1,500 mg/L TDS and produce a final blended water of 700 to 800 mg/L TDS, comparable to existing water quality. The desalination facility would require 17 wells for source water and 16 wells for blending. The addition of antiscalant and mineral acid to the source water is designed to inhibit mineral precipitation on the membranes, transmission pipes, and walls of the injection wells. A MODFLOW numerical model was constructed with cooperation by the U.S. Geological Survey and Fort Bliss, the International Boundary and Water Commission, and the City of Juarez Mexico Water Utility to evaluate the influ- ence of removing brackish water on the aquifer system. This peer-reviewed model found that the relatively low projected rate of withdrawal for the desalina- tion and blending from this large aquifer system would be environmentally benign and was sustainable through the projected life of the facility. Disposal of by- product fluids by well injection was determined to be cost-effective compared to both passive and enhanced evaporation procedures. Geophysical investigations and test wells indicated the Fusselman dolomite, a relatively high-permeability formation located 29 miles northeast of the desalination facility, would be hy- drologically suitable for disposal. A disposal permit was obtained from the State of Texas to inject desalination effluent into this aquifer. The injected water was lower in TDS than the native water in the Fusselman formation and only ex- ceeded drinking water standards for arsenic and selenium. SOURCE: Ed Archuleta, EPWU, personal communication, 2006. Ion exchange in certain instances may also change the physical proper- ties of the ion-exchanger clay mineral and result in loss of permeability of the aquifer (Civan, 2000). These impacts to water quality and quantity can be anticipated through hydrogeologic and geochemical analyses and the use of model
116 Desalination: A National Perspective simulations (see Box 5-3). The challenge to hydrogeologists is to identify and quantify all hydrologic boundaries and transport conditions under different scenarios of water supply development as well as any chemical reactions that add, subtract, or change solutes or ratios under varying flow conditions. Each well field is unique with respect to these properties and requires site-specific information. Subsidence When groundwater withdrawals exceed recharge rates, the hydraulic head (related to fluid pressure) will gradually be reduced and may result in land subsidence in areas with unconsolidated sediments (Figure 5-2). Subsidence results from the compression of the skeletal framework caused by reduced hydraulic head and rearrangement of grains in the aq- uifer matrix. Under equilibrium conditions, total stress (a function of the mass of water and rock) acting downward on a plane is balanced by a combination of the geologic framework acting upward (the skeletal framework resisting compression, or effective stress) and fluid pressure (hydraulic head). The reduction in hydraulic head associated with with- drawal of fluids increases the effective stress on the geologic framework and causes compaction and reduction in elevation of the land surface (Tergazi and Peck, 1967). Excessive development of a brackish water resource, particularly one that possesses thick sections of at-risk litholo- gies (i.e., clay, silt, organic material), may create the potential for subsidence. Land subsidence resulting from removal of groundwater has affected areas in 45 states (Figure 5-3) and ranges from regional lowering to ground failure and collapse (Galloway et al., 1999; NRC, 1991). Gallo- way et al. (1999) estimate that more than 80 percent of the identified subsidence in the United States has been caused by overexploitation of groundwater resources. Parts of Texas, California, and Nevada have ex- perienced tens of meters of surface decline (Leake, 2007). Regional lowering will increase the probability of flooding in coastal areas, while local settling may damage engineered structures such as buildings, roads, and utilities. Subsidence also has the potential to trigger earth- quakes and activate faults; for example, the Gulf Coast basin is a region where subsidence and fault activation are common around large, mature oil and gas fields (USGS, 2007b).
Environmental Issues 117 BOX 5-3 Tools for Addressing Brackish Groundwater Physical Renewability Several modeling tools are available to assess the potential impacts of a proposed brackish groundwater desalination facility on the physical renewability of an aquifer. Once appropriate site-specific information is obtained, numerical groundwater models, such as variations of MODFLOW (Harbaugh, 2005), can be used to test various scenarios of development and effects on water quantity. If density contrasts between water bodies are significant, then codes such as SEAWAT-2000 (Langevin et al., 2003) and SUTRA (Voss and Provost, 2003) may be more appropriate codes to use. Other models, such as PHRQPITZ (Plummer et al., 1988) and PHREEQC (Parkerhurst, 1995), address water quality impacts. MODFLOW-2005 simulates steady and nonsteady flow in an irregularly shaped flow system in which aquifer layers can be confined, unconfined, or a combination of confined and unconfined. Flow from external stresses, such as flow to wells, areal recharge, evapotranspiration, flow to drains, and flow through riverbeds, can be simulated. Specified head and specified flux boundaries can be simulated. In addition to simulating groundwater flow, MODFLOW-2005 incorpo- rates related capabilities such as solute transport and groundwater management. SEAWAT-2000 is the latest release of the SEAWAT computer program for simu- lation of three-dimensional, variable-density, transient groundwater flow in porous media. SUTRA is a model for saturated or unsaturated, variable-density ground- water flow with solute or energy transport. SUTRA version 2D3D.1 includes both two- and three-dimensional simulation capability. Water quality changes and aquifer transformation in brackish or brine condi- tions can be assessed using the model PHRQPITZ, which addresses chemical reactions (e.g., mineral precipitation, dissolution) within brackish aquifers. PHRQPITZ is a computer code that permits calculations of geochemical reac- tions in brines and other highly concentrated electrolyte solutions using the Pitzer virial-coefficient approach for activity-coefficient corrections. Reaction-modeling capabilities include calculation of (1) aqueous speciation and mineral-saturation index, (2) mineral solubility, (3) mixing of aqueous solutions, (4) irreversible reac- tions and mineral-water mass transfer, and (5) reaction path. PHREEQC (version 2), which adds the Pitzer coefficients for dealing with brackish water, is a computer program that is designed to perform a wide variety of low-temperature aqueous geochemical calculations. PHREEQC has capabili- ties for (1) speciation and saturation-index calculations, (2) batch-reaction and one-dimensional (1D) transport calculations, and (3) inverse modeling. Transport capabilities involve reversible reactions (e.g., aqueous, mineral, gas, solid- solution, surface-complexation, and ion-exchange equilibria) and irreversible reactions (e.g., kinetically controlled reactions, mixing of solutions, temperature changes). PHREEQC version 2 includes new capabilities to simulate dispersion (or diffusion) and stagnant zones in 1D-transport calculations and to model ki- netic reactions with user-defined rate expressions.3 3 For more information on these models, see http://water.usgs.gov/software/lists/ground_water/.
118 Desalination: A National Perspective FIGURE 5-2. Land subsidence from groundwater withdrawal. SOURCE: Galloway et al. (1999). FIGURE 5-3. Areas where subsidence has been attributed to groundwater with- drawal. SOURCE: Galloway et al. (1999).
Environmental Issues 119 Potential of Brackish Groundwater to Meet Future Water Needs The volume of inland brackish groundwater in the United States is unknown; based on the relative shallow depths of brackish groundwater reported by Feth (1965) (see Figure 1-1) and surveys of several states in the southwestern United States (Huff, 2004a, 2004b), large quantities of brackish water appear to be available for development. However, knowl- edge of the brackish aquifer systems in the United States is insufficient to assess whether they can be developed without long-term groundwater mining or water quality deterioration of the primary or adjacent aquifers. Additional detailed studies will be needed to assess the potential impacts of increased groundwater withdrawals at any specific site. CONCENTRATE MANAGEMENT Desalination processes create waste products, including salt concen- trates, cleaning and conditioning reagents, and particulate matter, that must either be disposed of or reused (Malmrose et al., 2004). This sec- tion will focus on the waste products unique to seawater or brackish groundwater desalination and potential environmental impacts from common concentrate management practices. It is worth noting here that the limited research to date on the environmental effects of concentrate management practices has primarily been focused on seawater desalina- tion plants, because globally, seawater plants tend to be the largest desalination facilities. Currently, the majority of desalination facilities in the United States, both in the number of plants and in total capacity, use brackish source water (see Chapter 2). Chemical Constituents in Desalination Concentrate The largest component of the waste stream from a desalination facil- ity is concentrate that varies in salt concentration and in the ratios of particular species with the specific source water (see Table 2-1) and the desalination process used. The concentrate stream from a desalination process includes the constituents rejected from the feedwater stream, in a more concentrated form. Currently, depending on the source water qual- ity and constituents, the recovery of brackish water membrane desalination is between 50 and 90 percent, implying that 10 to 50 percent of the feedwater is converted to concentrate (concentrating the solutes by a factor of 2 to 10). For seawater desalination, the corresponding recov- ery range is 35 to 60 percent, implying that 40 to 65 percent of the feedwater is converted to concentrate (concentrating the solutes by a fac-
120 Desalination: A National Perspective tor of 1.5 to 2.5). In addition to salts, other naturally occurring elements from the source water are concentrated (e.g., selenium from some brack- ish groundwater sources), which, depending on the source water quality and discharge concentrations, could cause adverse impacts if discharged into sensitive environments. When desalination plants are co-located with power plants, desalination concentrate may also contain excess heat that may pose concerns for disposal into the environment. Desalination concentrate may also contain chemicals used in the desalination process (see Table 4-1). The nature and concentration of the chemicals in the concentrate vary and are site-specific, depending on which chemicals are used, the amounts used, how frequently they are used, and whether they are discharged in the concentrate or disposed of via sewage treatment plants or other disposal options. The nature and potential impacts of these chemical additives are discussed in detail below. Biocides, such as sodium hypochlorite for chlorination, may be added to minimize fouling in surface water (seawater and surface brack- ish water) desalination (Einav et al., 2002). Chlorine is quite toxic to marine biota if discharged to the environment without neutralization. Because EPA criteria for short- and long-term exposure are 13 and 7.5 Î¼g/L, respectively (EPA, 1985), and because there is potential for creat- ing environmentally persistent, bioaccumulative, and toxic halogenated organic compounds after chlorine addition, chlorine is considered a seri- ous concern (Lattemann and HÃ¶pner, 2003). Lattemann and HÃ¶pner (2003) report that chlorine has been found at concentrations up to 100 Î¼g/L within 1 km of desalination plant outfalls in Kuwait, which uses primarily thermal desalination plants, usually co-located with power plants. However, the high chlorine levels at the Kuwait plants would likely continue even if the desalination plant were no longer operating because of the biocide requirements of the power plants. Chlorination is not a major issue for most reverse osmosis (RO) plantsâthe vast major- ity of the desalination facilities in the United States. Chlorination in the RO process is typically intermittent, and the added chlorine is neutralized because membranes are sensitive to oxidizing chemicals. Coagulants, such as ferrous chloride and aluminum chloride, are of- ten added in seawater and surface brackish water pretreatment processes to remove suspended matter. These chemicals, released at 1-30 ppm in the concentrate, are of very low toxicity but can form precipitates and may cause increases in turbidity and discoloration in the case of iron compounds (Lattemann and HÃ¶pner, 2007). Filters in RO plants need to be backwashed every few days to clear the accumulation of solids. In the United States, this filter backwash is not permitted to be directly discharged to the environment, because it can cause both considerable discoloration in the water at the discharge site
Environmental Issues 121 and beach contamination. However, the practice may occur in other loca- tions. Antiscaling substances, antifoaming additives, oxygen scavengers, and anticorrosion chemicals may also be present in the discharge of sea- water concentrate (HÃ¶pner, 1999; Rachid and Abdelwahab, 2005). Antiscalants, which are added to inhibit the formation of scale precipi- tates and salt deposits, belong to different chemical groups. Commonly used antiscalants are polyphosphates, phosphonates, or carboxylic-rich polymers such as polyacrylic acid and polymaleic acid. Polyphosphates are increasingly being replaced by the other polymers that are more sta- ble at high operating temperatures and more resistant to chemical and biological breakdown. All antiscalants are nontoxic at the levels used for desalination (1-2 ppm in the concentrate). However, issues that may be of concern are nutrient levels in the discharge, which may be increased by polyphosphate, and the relatively slow degradability of the other an- tiscalant species (Lattemann and HÃ¶pner, 2003, 2007). Little is known about environmental effects of the polyphosphate scale control additive sodium hexametaphosphate, which was one of the first antiscalants to be used, although it is known to act as a nutrient. Polymer antiscalants such as polyacrylamide, however, are of low toxicity and low bioaccumula- tion and might actually reduce copper toxicity by binding to it (Lattemann and HÃ¶pner, 2003). Antifoaming additives include acylated polyglycols, fatty acids, and fatty acid esters, which are detergents (HÃ¶pner, 1999). Polyethylene gly- col (PEG) is nontoxic and is used in low concentrations, but it is persistent in the environment. PEG is specifically used in the multistage flash (MSF; see thermal processes in Chapter 4) process and usually only temporarily when natural surfactants cause seasonal foaming. Some MSF units with aggressive process designs may be more sensitive to foaming than others that may never require antifoaming agents (Lattemann and HÃ¶pner, 2003). Benzotriazole derivatives may be used as corrosion in- hibitors and may be released during the cleaning of MSF plants. Benzotriazole chemicals are persistent, but have low potential to bioac- cumulate and are toxic only at very high concentrations. Desalination concentrate from any source water may also contain metals from corrosion that may be toxic to organisms if they are dis- charged into the environment. Concentrations of metals in the desalination discharge will vary with the design and the specific materi- als used in the desalination plant. Chemicals of corrosion are unlikely to be a major concern for RO plants, although RO plants will discharge mi- nor amounts of iron, chromium, nickel, and molybdenum in their concentrate from stainless steel (Lattemann and HÃ¶pner, 2007). Chro- mium in its hexavalent form is known to be quite toxic, but there has been very little study of metal toxicity from RO plants.
122 Desalination: A National Perspective Corrosion-related chemicals are primarily an issue with MSF plants, because sections of the MSF process operate at higher, corrosive tem- peratures. Copper is estimated at 5-33 Î¼g/L in the blended discharge of co-located MSF facilities, which is above the EPA criteria of 4.8 and 3.1 Î¼g/L for short- and long-term exposure, respectively (EPA, 1985; Latte- mann and HÃ¶pner, 2003). While dissolved copper can form nontoxic complexes and bind to sediments, chronic effects are possible. Copper is an essential metal, but some organisms, such as algae and mollusks, are extremely sensitive to elevated levels. The most toxic form of copper is the free ion (Cu2+), which tends to be of low concentration in the water column because most of the copper binds to suspended material and ac- cumulates in the sediments. However, sediment-bound metals can be remobilized by a sudden change in environmental conditions and serve as a source of metal contamination to benthic and other marine organ- isms. Keeping the concentrate pH higher and adding chelators can reduce the concentration of free copper and, thus, the toxicity of the concentrate (Rinne, 1971). Nickel is less toxic than copper (EPA criteria are 74 and 8.2 Î¼g/L for short- and long-term exposure, respectively [EPA, 1986]), and although its concentration in the effluent may be similar to that of copper, it is not considered a serious risk. MSF plants built in the past 10 years have become much more corrosion-resistant, because titanium has increasingly been used for the tubes in the hotter sections of the plants. However, this trend may have halted or reversed due to the increased cost of titanium. Projects under construction now may once again rely primarily on copper alloy tubes (Lattemann and HÃ¶pner, 2003). Chemicals such as detergents are also used periodically to clean de- salination membranes to improve system performance, prevent membrane fouling, and extend membrane life (see Table 5-1). These may or may not be discharged with the concentrate. Acid is typically used to remove inorganic foulants (e.g., metal oxides or scale), and alkaline solu- tions and dispersants are used to remove biofilms, silt deposits, and organic foulants. Thus, membranes are cleaned with the use of low pH, high pH, oxidant, biocide (e.g., formaldehyde), and detergent solutions. If acidic detergents are used and not disposed separately, discharge of the concentrate into freshwater may have impacts on the pH of the receiving water, even if diluted. Seawater tends to be well buffered and is not likely to experience significant acidification. These chemicals can also be considerably toxic to freshwater and marine organisms. The nature and volume of cleaning solutions and waste depend on the system design, the size of the system, and the frequency of cleaning. Most plants clean their membranes every 3 months, or less frequently. Typically the volumes of cleaning solutions used (see Table 5-1) per cleaning are about 1.2 L/m2 (3 gallons per 100 square feet) of membrane (AWWA Membrane Re- siduals Management Subcommittee, 2004).
Environmental Issues 123 TABLE 5-1 Typical nanofiltration and reverse osmosis cleaning formulations. Foulant Type Cleaning Solutions Inorganic saltsa 0.2% HCl 0.5% H3PO4 2% citric acid Metal oxides 2% citric acid 1% Na2S2O4 Inorganic colloids (silt) 0.1% NaOH, 0.05% Na dodecyl benzene sulfonate, pH 12 Silica (and metal sili- Ammonium bifluoride cates) 0.1% NaOH, 0.05% Na dodecyl bezene sulfonate, pH 12 Biofilms and organics Hypochlorite, hydrogen peroxide, 0.1% NaOH, 0.05% Na dodecyl bezene sulfonate, pH 12 1% sodium tripolyphosphate, 1% trisodium phos- phate, 1% sodium EDTA a Barium sulfate, calcium carbonate, calcium sulfate. SOURCE: Dow Chemical Co. (1994). Lattemann and HÃ¶pner (2003) discussed the toxicity and potential risk of several common membrane cleaning chemicals, considering typi- cal discharge concentrations and dilution, focusing on desalination plants in the Arabian Gulf. Sodium dodecyl benzene sulfonate (Na-DBS) was found to degrade relatively quickly (within a few weeks). There are no toxicity data for Na-DBS, but a similar detergent, sodium dodecylsulfate (Na-DSS), has moderate toxicity in fish, an invertebrate, and algae (LC50 or the concentration of salt lethal to 50 percent of the organisms of 1-10 mg/L). Na-DBS is designated as a hazardous substance under the Federal Water Pollution Control Act. In contrast, EDTA was found to degrade slowly (5 percent reduction in 3 weeks) but showed low bioaccumulation and low toxicity (acute effects at 10-100 mg/L). Sodium perborate, which was found to degrade at discharged concentrations, also has fairly low toxicity (LC50 of about 10-50 mg/L). Isothiazole at the levels used for membrane cleaning could be toxic in the immediate mixing area be- fore dilution. Routine cleanings tend to be more benign and utilize organic acids, particularly citric acid, and facilities should perform neu- tralization prior to blending or metered discharges. Aggressive cleanings with EDTA or other chemicals are less common (Lattemann and HÃ¶pner, 2003). Lattemann and HÅpner (2003) conclude that membrane-cleaning chemicals can be quite hazardous, since they include disinfectants (e.g.,
124 Desalination: A National Perspective formaldehyde, isothiozole), which interfere with cell membranes and are toxic at low concentrations. These process chemicals and the suspended solids can, in most cases, be removed from the concentrate or neutralized rather than being dis- charged directly into the environment. In fact, most facilities in the United States discharge their cleaning solutions separately to sewers, but approximately 17 percent of RO facilities in the United States discharge their cleaning solutions with the concentrate (Kenna and Zander, 2000, 2001). The volume of chemical additives required may also be reduced by using various filtration methods in the pretreatment process (Dudek and Associates, 2005). Removing or neutralizing added chemicals before discharging the concentrate reduces the environmental impact. In overall considerations of the contamination of the Arabian Gulf, where there are numerous MSF thermal and RO plants, Lattemann and HÃ¶pner (2003) considered copper and chlorine to be the most serious environmental threats from seawater desalination concentrate discharge. Chlorine in regions of Kuwait Bay is present at levels close to toxic concentrations for some phytoplankton, invertebrates, and vertebrates. HÃ¶pner and Lattemann (2002) estimated that 21 thermal desalination plants released a total of 2,708 kg of chlorine, 36 kg of copper, and 9,478 kg of antiscalants daily directly into the Red Sea. (Comparable chemical discharge data for RO facilities are not available.) Copper is less likely to be a problem in the United States, because RO plants do not generally generate copper in significant amounts. Although dilution of the concen- trate will reduce the environmental impacts of the higher salinity, it will not negate the effects of these other chemicals, because the overall load- ingânot the concentrationâis a stress on the receiving environment (HÃ¶pner, 1999). Lattemann and HÃ¶pner (2003) also emphasize that noth- ing is known about the potential interactive effects of the various chemicals that may individually be present at nontoxic levels. It is difficult to make broad assessments of the impacts of desalina- tion process chemicals because the treatment approaches vary widely from plant to plant. In addition, some plants will discharge more chemi- cals in the concentrate whereas others will send the chemicals to sewage treatment plants or to other disposal options. The chemicals in RO con- centrate tend to be less hazardous that those in thermal desalination concentrate, which can contain high levels of metals and chlorine. Nevertheless, membrane cleaning chemicals could have envi- ronmental impacts if they are discharged into the environment with the concentrate rather than being disposed of separately. Many of the pre- treatment chemicals (e.g., antiscalants, coagulants) used in RO desalination are of relatively low toxicity, but there remains a need for more information on the environmental effects of the various chemicals that are used in the desalination process, both individually and especially
Environmental Issues 125 in combination. Research should examine ways to reduce additives in concentrate discharge that can be harmful to the environment. Environmental Impacts from Concentrate Management Approaches As discussed in Chapter 4, concentrate management strategies may take place in a variety of ways, including surface water discharge (e.g., into oceans, seas, estuaries, lakes, rivers), wastewater discharge, injection into underlying aquifers, land application, evaporation ponds, and dis- posing of the salts in landfills after thermal evaporation. Various site- specific factors may limit the range of concentrate management options at a given site, such as state and local permitting, hydrological condi- tions, low season flows, capacity of local sewers, level of dissolved solids and toxic materials in the concentrate, availability of dilution wa- ter for land application, climate suitability, and costs of the different options (AWWA Subcommittee, 2004). These concentrate management approaches are implemented within the context of state and federal envi- ronmental regulations, as described in Box 5-1. These various concentrate management technologies and approaches and their relative use in U.S. desalination plants are described in Chapter 4 (see Figure 4- 15). Potential environmental impacts from coastal and inland concentrate management methods, and approaches to mitigate those effects, are dis- cussed in the following sections. Coastal Concentrate Discharge into Oceans, Seas, and Estuaries The salinity of seawater desalination concentrate can approach 2.5 times the salinity of seawater, and the impacts of this discharge on the marine environment will vary with discharge method, source water salin- ity and quality, site conditions, and ecosystem type. Without proper dilution, a plume of elevated salinity discharge may extend for a consid- erable distance beyond the mixing zone and can harm the ecosystem (Younos, 2005). If a desalination plant is co-located with a power plant or a wastewater treatment facility, the concentrate can be blended and diluted with the power plant discharge or treated wastewater effluent be- fore being returned back to the environment, reducing the potential for salinity stress in organisms in the receiving water. In the case of co- location with a power plant, however, the blended effluent will be of higher temperature and can have temperature stress effects on local or- ganisms. If concentrate is released in such a way as to maximize dispersion of the effluent, with diffusers and dense jets and nozzles ar-
126 Desalination: A National Perspective ranged in particular configurations, these effects could be reduced. In addition to the configuration of the outlet, the extent of mixing will also depend on local site conditions, including the bathymetry, waves, cur- rents, and depth of the water (Einav et al., 2002). Concentrate from brackish water desalination plants may have salini- ties that are similar to or lower than seawater, depending on the source water salinity and the recovery efficiency. However, depending on the source water composition, concentrate from brackish groundwater de- salination may also contain trace elements, such as arsenic or selenium, in elevated concentrations compared to seawater. Additionally, when concentrates originating from groundwater are discharged to seawater or brackish surface waters, major ion toxicity can result (Mickley et al., 2001). This toxicity occurs when certain ions are present in very different concentrations (higher or lower) relative to seawater diluted to the same salinity. Toxicity due to an âimbalanceâ of ions relative to seawater has been seen in mysid shrimp with respect to high calcium or fluoride or low potassium. This toxicity has sharp thresholds, but appropriate con- centrate dilution should eliminate it. The toxicity in brackish groundwater membrane concentrate from high levels of hydrogen sulfide or low levels of dissolved oxygen can be remedied by aeration. Different coastal and marine ecosystems are likely to vary in their sensitivities to concentrate discharge. HÃ¶pner and Windelberg (1996) ranked various marine ecosystems in terms of their perceived sensitivity to a seawater desalination plantâs discharge. This ranking ranged from the least sensitiveâhigh-energy oceanic coast and exposed rocky coastsâto the most sensitiveâcoral reef, salt marsh, and mangrove. However, the sensitivity scale they used was based on the sensitivity of different environments to oil spills. Thus, more study will be needed to determine the relative sensitivity of different types of estuarine and coastal environments to concentrate discharges. For example, although salt marshes and mangroves are very sensitive to the effects of oil spills because oil will persist in the sediments for a very long time, these estua- rine sites will probably have a higher tolerance to increased salinity, because estuaries normally experience fluctuations in salinity. While HÃ¶pner and Windelberg (1996) considered high-energy rocky coasts to be relatively insensitive (as they are to oil spills), rocky habitats with kelp beds along the California coast are considered critical sensitive eco- systems (Cooley et al., 2006). Studies of Effects of Concentrate on Specific Marine Organisms. There have been numerous papers discussing the potential for negative environmental impacts of effluents from desalination facilities, but there is a surprising paucity of useful experimental data, either from laboratory tests or from field monitoring, to assess these impacts. In sensitive organ-
Environmental Issues 127 isms, increased salinity can cause osmotic pressure changes in cells and may affect respiration and photosynthesis and reduce growth rates (Rinne, 1971). However, different species have different levels of toler- ance to increased salinity. Some organisms have very high salinity tolerances and are referred to as euryhaline, but others have only narrow tolerances and are called stenohaline. More euryhaline species are likely to be found in estuaries, where there is a natural variation in salinity every day with the tidal cycle, than in the ocean, where the salinity is more constant. In general, larvae and young individuals are more sensi- tive to environmental stresses than adults of the same species. The following section describes what is known from research about the po- tential impacts of seawater desalination concentrate on various marine biota. The effects of concentrated salt on marine organisms have been tested in some laboratory bioassays. Pillard et al. (1999) exposed adults of three estuarine species (sheepshead minnow, Cyprinodon variegatus; mysid shrimp, Mysidopsis bahia; and silversides, Menidia menidia) to water of different salinities and monitored their survival over a 48-hour period (Figure 5-4). Whereas the very tolerant sheepshead minnow toler- ated seawater of over 60 ppt, the mysids and silversides began to exhibit mortality at salinities of around 38-40 ppt. Based on this information, there have been conclusions that 38-40 ppt salinity represents a salinity tolerance threshold for marine organisms (Jenkins and Graham, 2006). While this estimate of salinity tolerance may indeed be correct, the data are not adequate to support it. These data reflect exposure over a very short time period and are focused only on mortality; thus, the data are far from adequate to establish what level of salinity is safe in the long term for marine organisms. Also, all three of these species are estuarine and very tolerant of changes in salinity. Latorre (2005) reported on effects of concentrate on the seagrass Posidonia oceanica, which is the most widespread species of seagrass in the Mediterranean and is subject to discharges from an RO seawater de- salination plant in Mazarron Bay, Spain. Mesocosm studies in the laboratory revealed that brief (15-day) exposures to 40 ppt salinity water caused 27 percent mortality and reduced the growth of surviving plants. At field sites near the RO plant, seagrass beds were degraded in areas where the salinity had increased to 39 ppt. Early life stages tend to be more sensitive than adults. For the Cali- fornia grunion (Leuresthes tenuis), Reynolds et al. (1976) determined that prolarvae (i.e., larvae with a yolk sac, up to about 4 days old) have an upper salinity tolerance (LC50) of 41 ppt after 24 hours of exposure and that 20- to 30-day-old larvae tolerate a maximum of 40 ppt for about
128 Desalination: A National Perspective FIGURE 5-4. Whole effluent toxicity (WET) salinity tests on marine animals. Data reflect survival after 48 hours of continuous exposure to various salinity levels. SOURCES: Pillard et al. (1999), Jenkins and Graham (2006). Reprinted with permission from Environmental Toxicology & Chemistry, Allen Press Publishing Services. 18 hours. Lasker et al. (1972) found that salinities greater than 40 ppt adversely affected eggs and larvae of the California Sargo (Anisotremus davidsoni). Testing larvae is a good approach, but to determine adequate safety it is necessary to examine long-term sublethal effects and whole life cycles of various local species to see to what degree their physiology, behavior, growth, reproduction, and other parameters would be affected by different salinities. Studies on invertebrates have also mostly focused on short-term survival at increasing salinities and did not examine long- term or life-cycle effects (Bay and Greenstein, 1993; Blazkowski and Moreira, 1986; Charmentier and Charmentier-Daures, 1994; Ferraris et al., 1994; Forster, 1998). A few studies have investigated sublethal effects of RO discharge water, but the effects and sensitivities differed from species to species. For example, Bay and Greenstein (1993) found that exposure of giant kelp (Macrocystis) blades to elevated salinities of 38.5 and 43 ppt did not affect either the spore generation rate or the length of the germination tube. A 14-day study of the California Sargo (Anisotremus davidsoni) found that the optimum salinity for juvenile feeding and growth was 33-
Environmental Issues 129 37 ppt, and adverse effects, such as reduced growth, were seen at 45 ppt (Brocksen and Cole, 1972). Iso et al. (1994) studied the effects of con- centrate on eggs and larvae of fish species and found that salinities below the incipient lethal level could produce sublethal effects, such as retarded development and growth. Field Studies of the Effects of Concentrate Discharge. Even after dilution, concentrate discharge will be of higher salinity than the receiv- ing water and will tend to sink and settle along the bottom of the marine environment. Thus, the benthic community will be particularly impacted, although the specific effects will be dependent on the local ecosystem, the composition of the discharge, and the degree of dilution at the point of contact (Mickley et al., 2001). Perez Talavera and Quesada Ruiz (2001) studied impacts of the discharges in the Canary Islands, specifi- cally examining effects on the seagrass Cymodocea nodosa and the alga Caulerpa prolifera. They found that although the initial dilution was high, salinities on the bottom remained elevated over a large area. The concentrate also elevated the turbidity of the water. Seagrasses were ab- sent in areas near the outfall; farther away they were present but in poor condition and covered with slime. At distances even farther from the dis- charges, the seagrasses were in good condition. A comprehensive study of the ecological impacts of concentrate dis- charge from a thermal desalination plant was performed by Chesher (1975) in Key West, Florida. This study examined chemical and physical properties of the discharge, historical analysis of sediments, and the abundance of benthic fauna over time. Laboratory bioassays were also performed on a number of species. Results indicated that the heated ef- fluent, which was highly contaminated with dissolved copper, had a significant impact on the biota near the discharge. Temperature and sa- linity of the effluent were such that the effluent settled at the bottom of the receiving basin, which reduced water circulation. An 18-month bio- logical study showed a marked reduction in biotic diversity. Gacia et al. (2007) studied the effect of concentrate on a shallow seagrass meadow (Posidonia oceanica) exposed to RO concentrate dis- charge for over 6 years and compared the results with two undisturbed reference zones. The concentrate contained both elevated salinity and nitrogen because of high nutrient levels in the source groundwaterânot a typical situation for desalination plants. Sea urchins and sea cucum- bersâspecies very sensitive to salinityâwere present in the reference transects but absent from the impacted transect. The seagrass proved to be sensitive to both elevated nutrients and salinity in the discharge; how- ever, the observed necrosis and reduced growth in the seagrass and the absence of sea urchins were attributed to the high salinity. There was no indication of extensive decline of the affected meadow, probably due to
130 Desalination: A National Perspective its very shallow condition, which results in fast dilution and dispersion of the concentrate plume. An intensive field study was performed to estimate the effects of concentrate discharge in Tampa Bay, Florida, using a study site on the island of Antigua under the assumption that it was similar to Tampa Bay (Blake et al., 1996; Hammond et al., 1998). The researchers redirected the Antigua seawater RO desalination plantâs discharge to a new site, which they studied before and during exposure to the concentrate. This study ultimately found no effect of salinity elevation on the tropical reef ecosystem (see Box 5-4 for detailed findings), and it is the most compre- hensive and thorough study that has been done to date, although it has not been published in the peer-reviewed literature. Despite the drawbacks of the study as described in Box 5-4, this extensive study should be a model for future studies in other locations. In contrast to the minor changes seen in the previous study, Pilar Ruso et al. (2007) found major impacts on benthic communities follow- ing the 2003 opening of a seawater RO plant in Alicante, Spain, which discharges 65,000 m3/day of 68 ppt TDS concentrate. The salinity at sta- tions farther from the discharge was 37.9 ppt, while near the discharge it was elevated to over 39 ppt. Investigators studied benthic infaunal com- munities along three transects perpendicular to the coast. At the beginning of the study, the most abundant groups of organisms at all sta- tions were polychaetes, nematodes, and bivalves. However, at 9 months from the beginning of the discharge, in the station closest to the dis- charge, the community became dominated by nematodes, and this dominance increased over time from 68 to 83 to 96 percent in subse- quent samplings. During the 2 years after commencement of the discharge, other stations in the transect receiving the discharge showed a similar change in the structure of the community to becoming dominated by nematodes, reflecting an extension of the area of influence of the dis- charge. The authors recommend increasing the mixing of the concentrate, for example by diffusers, to reduce the area affected by high salinity. In contrast, Raventos et al. (2006) found no significant impacts at- tributable to concentrate discharges from a small seawater desalination plant in the northwest Mediterranean on organisms present on the surface. Visual censuses were carried out 12 times before and 12 times after the plant began operating. No significant variations attributable to the discharges from the plant were found. The failure to observe impacts may be explained by the high natural variability and by the rapid dilution undergone by the concentrate upon leaving the discharge pipe, which had a diffuser with 43 perforations to facilitate rapid dilution of the discharge (salinity was back to normal 10 m away from the pipe).
Environmental Issues 131 BOX 5-4 Antigua Study of the Impacts of Concentrate Discharge The study at Antigua was conducted to investigate the potential impacts of seawater desalination concentrate discharge in Tampa Bay. Prior to the concen- trate discharge diversion to the study site, the researchers conducted baseline studies of the habitat. Six transects, extending out radially 10 m from the dis- charge site, and at 60-degree angles from one another, were established (see Figure 5-5). Transects extended both on- and offshore from the discharge site. The study area contained a diverse assemblage of organisms, including sea- grass (Thalassia), algae (Dictyota), hard corals (Porites) and soft corals (Pseudoterogorgia), as well as other invertebrates and fishes. Results indicated that the discharge water had elevated temperature (2-3oC warmer), elevated salinity (see below), and a reduced pH (0.2-0.3 units lower) compared to the am- bient water. Differences between the discharge and the ambient water were detectable within the study area but were rapidly dissipated by mixing. Dye injec- tion demonstrated that the plume rapidly dissipated and moved toward deeper water. The elevated salinity âsignalâ was detectable beyond the 10 m study area and distributed mainly down slope. Maximum bottom salinities, recorded in the immediate vicinity of the discharge opening, were 35-40 ppt in June and 34-38 ppt in October; but because the pipe discharge flowed upward and contacted the surface, surface salinities were higher (35-44 ppt in June, 34-43 ppt in October). Because of strong mixing, salinities at the 8-10 m transect positions averaged only 0.2 ppt above ambient, with salinity increases extending farther downslope than upslope. No significant changes were noted in biological communities along the plume. Studies of the seagrass beds indicated no changes in the number of new shoots, biomass, or growth rate over the survey period. All plants showed nu- merous parrotfish bites, which indicated that this fish frequented the study area in spite of the elevated salinity. A brown alga (Dictyota) showed variations in growth rate and a weak correlation with salinity. Tissues from plants living within the study area also showed a higher concentration of nitrogen than did plants sam- pled from outside the study area. Diatom numbers and types did not change from prediversion conditions. Benthic foraminifera distribution and abundance varied considerably and reflected a âstressed habitat,â but there were no differences that related to the presence of the concentrate discharge. Dominant benthic infauna were annelids and one snail species. There were significant differences in the infaunal assemblage at different times (March and October had more animals than June), but abundance and diversity did not appear to be affected by ele- vated salinity. Epifauna were collected on settling plates. Bryozoans and polychaete worms were the dominant forms, with hydroids, snails, clams, and sea urchins also settling. Variation in the groups that settled on the plates at the different sampling periods was attributed to biological factors (reproductive sea- son) rather than an elevated salinity. However, because there were no settling plate data prior to diversion, it is unknown whether or not increased salinity ex- cluded any species from settling. Coral heads in the transect area exposed to an average salinity elevation of 4.5 ppt showed no ill effects over the 6-month ob- servation period. There were no obvious effects of the discharge on either the macroinvertebrates or fishes in the different observation periods. However, no survey of fishes was done before the diverting the discharge. continued
132 Desalination: A National Perspective There are five notable limitations to this study: (1) The sampling periods were limited to only two postdiversion observations and extended only 6 months. Six months is too short a period to determine the community effects of variables such as rainfall, seasonal effects, nutrient cycles, and biological cycles. (2) The settling plate studies did not have âprediversionâ data, so a contrast between pre- and postdiversion settling cannot be made. No measurements of fish were per- formed prior to the discharge either. (3) Monitoring of a control site, where no salinity changes occurred, would have provided important baseline data. (4) Re- sults of this study have limited applicability to other systems, but many of the biological communities are similar to those in the Tampa Bay area. There is about a 50 percent overlap in species and abundant seagrass communities at both locations, though there was only a 10 percent similarity of benthic infauna. Clearly, however, the study could not have been done in Tampa Bay itself. (5) The study included chemical analyses of metals, but not of other chemical addi- tives in the discharged concentrate. The accumulation of these chemical additives in the sediments or organisms in the vicinity of the plume was not stud- ied. Nevertheless, this study is the most comprehensive of its kind. Its limitations strongly suggest the need for standard rules and protocols for future studies. FIGURE 5-5. Diagram of sampling area for Antigua study. SOURCE: Hammond et al. (1998). Another study that provides additional useful information is an in- vestigation of the ecological effects resulting from the application of granulated sea salt as a management technique to kill the invasive alga Caulerpa taxifolia (OâNeill et al., 2007). Although the salt treatment causes cell lysis and death of the target alga, nearby seagrasses (Zostera),
Environmental Issues 133 their epiphytes, and nearby infauna showed no consistent effects on abundance or diversity. From these studies, the effects of desalination concentrate discharge into oceans, seas, and estuaries appear to vary widely. The impacts de- pend on the site-specific environment, the organisms examined, the amount of dilution of the concentrate, and the use of diffuser technology. Modeling and Monitoring Coastal Concentrate Discharge. Com- puter simulation models can be created and utilized to predict the environmental changes that may occur as a result of concentrate dis- charge (see Box 5-5). Through these predictive models, technical evidence can be produced to support applications for the NPDES permit and other environmental permits that may be required for concentrate discharge. Through this process, a utility can minimize environmental impacts from concentrate discharge and also generate greater public ac- ceptance of the technology. Additional environmental ecosystem monitoring is needed to better demonstrate the environmental sustainability of desalination using sur- face water concentrate discharge. Hydrobiological monitoring programs using scientifically accepted methods that allow utilities to track changes to coastal ecosystems over the operating period should be strongly con- sidered. Pilot studies should be performed prior to construction of a plant, and baseline data should also be collected. After construction of the desalination facility, monitoring studies should be continued. Protocols should be developed for monitoring programs using stan- dardized procedures, including before-and-after studies, the length of transects to be examined, and the use of reference sites. A possible moni- toring approach might be to follow transects perpendicular and parallel to the shoreline at both the impacted and reference sites, to survey the abundance, diversity, and health of resident planktonic, benthic, and nek- tonic (e.g., fish) species. Utilities should begin monitoring at least 1 year prior to the start of the projected operating period to establish a baseline of ecosystem data with which to compare postproduction data. Monitor- ing should be continued over a period of several years after the plant comes online and should include samples taken during different seasons of the year. When the receiving environment includes seagrass beds or kelp forests, they should also be carefully evaluated. Settling plates should be included for surveying attached organisms. Because the con- centrate will tend to settle on the bottom, benthic community analysis (i.e., abundance, diversity, community structure) should be an important part of a monitoring program. There are well-established protocols and techniques for performing these analyses (Cao et al., 1996; Gaston and Young, 1992; Horne et al., 1999; Weis et al., 2004). The development of
134 Desalination: A National Perspective BOX 5-5 Modeling Environmental Impacts of Estuarine and Marine Desalination The regulatory approval and permitting process for seawater desalination requires the application of computer modeling to estimate the environmental ef- fects of concentrate discharge. Various models have been developed for surface water concentrate and blending discharge and must be developed on a site- specific basis. This is necessary to provide an estimate of salinity changes within the three dimensions (depth and area) of the water body. Site-specific variation includes seasonal fresh surface water flows, tidal changes, water depth, and wind speed and direction. Once the modeled salinity changes are determined, an evaluation can be performed on the marine species to estimate health effects on those species. These models are referred to as three-dimensional (3D) time- dependent hydrodynamic models. Most of these models descended from the Princeton Ocean Model (Blumberg and Mellor, 1987). Comprehensive environ- mental monitoring programs are necessary to collect background data prior to the desalination operation and after the desalination project startup. Data from these monitoring programs can be used to calibrate the environmental models for future plant expansions or for other desalination projects in the vicinity. One example of model application is the Tampa Bay Seawater Desalination project. In this case, a 3D hydrodynamic model for âfar-fieldâ effects was devel- oped and applied based on an advanced version of the Blumberg-Mellor Estuarine Coastal Ocean Model (ECOM-3D) (Luther et al., 1998). Modeling for prediction of ânear-fieldâ effects was performed using MIKE 3, a 3D nonhydro- static finite difference model developed and applied by the Danish Hydraulic Institute. monitoring protocols is very important to facilitate environmental impact assessment studies. The development of a monitoring framework is also an objective of the European research project MEDINA (Membrane- Based Desalination, an Integrated Approach) funded by the European Commission. Management of Inland Desalination Concentrates Brackish desalination facilities that are a significant distance from the coast have unique considerations when it comes to concentrate man- agement. Brackish desalination facilities operate at higher recovery efficiencies (50 to 90 percent) and therefore produce lower quantities of concentrate per equivalent volume of product water. Because the source water typically has significantly lower salinity than seawater, the concen- trate from brackish water desalination plants also has significantly lower salinities than seawater desalination plants, and inland facilities that use brackish groundwater as source water require fewer pretreatment chemi- cals. Nevertheless, despite these apparent advantages, finding cost-
Environmental Issues 135 effective approaches for concentrate management with minimal envi- ronmental impacts in inland locations remains a challenge. Alternatives to disposal are to treat the concentrates as potential economic products (e.g., selective salt recovery) that utilize the concentrates in some type of beneficial or commercial product (Ahmed et al., 2003; Jordahl, 2006). Environmental impacts of various approaches for inland concentrate management are discussed in the following section (see Chapter 4 for technical descriptions of the current concentrate management ap- proaches). Freshwater Rivers. Desalination concentrates are sometimes dis- charged to rivers and other inland surface water bodies in accordance with local, state, and national water quality regulations (Mickley, 2006). Additionally, 31 percent of municipal desalination facilities with greater than 95 m3/day discharge directly to sewers upgradient of the wastewater treatment plant (Mickley, 2006). Most wastewater treatment plants ulti- mately discharge treated effluent to rivers, and while the traditional treatment process would remove many organic compounds, suspended solids, and some metals, it will not remove salts. However, the salinity in the concentrate will be diluted by the volume of treated wastewater. Inland communities that desalinate river water (or alluvial aquifers) and blend the concentrate with treated wastewater prior to river discharge should see only minor net downstream increases in river salinity caused by evaporative concentration and other salinization processes. However, concentrate from a deep brackish groundwater source could significantly alter the water quality of the surface water body. Depending on the source water composition, brackish groundwater concentrate may add toxic trace and radioactive constituents leached from the subsurface, such as selenium, arsenic, uranium, or radium. The potential for major ion toxicity, due to âimbalanceâ of ions in the concentrate, discussed previ- ously, is also a concern when concentrate from brackish groundwater is discharged into freshwater ecosystems. Impacts of discharge of brackish water to surface water sources would be expected to be greater in freshwater than in estuaries or the ma- rine environment, where salt is a natural component of the ecosystem. Some freshwater organisms are only able to tolerate low levels of dis- solved solids. If salinity increases in the water body, a shift to more salinity-tolerant species can be expected. High salinity may interfere with the growth of aquatic vegetation. Salt may decrease the osmotic pressure, causing water to flow out of the plant, resulting in stunted growth. Water containing high salt concentrations may create brackish layers in receiv- ing lakes. Since saltwater is denser than freshwater, it tends to sink and form a layer at the bottom that does not mix with remainder of the lake water, leading to decreased dissolved oxygen levels.
136 Desalination: A National Perspective Elevated salinity is stressful to many freshwater organisms. Sarma et al. (2005) studied freshwater crustaceans (anostracans) that inhabit ephemeral water bodies in which the water level decreases due to evapo- ration, increasing the salt concentration. They found that increased salinity resulted in decreased survivorship. Females showed several peaks of reproduction at 0 and 1 ppt salinity, whereas at 4 or 8 ppt there were fewer peaks. The highest reproductive rate was in 0 ppt of salt, while the lowest was at 8 ppt. Average lifespan, life expectancy, gross and net reproductive rates, generation time, and the rate of population increase were inversely related to the salt concentration. Freshwater spe- cies from environments that do not normally experience increased salinity would likely be much more susceptible than these anostracans. To minimize environmental effects, concentrate discharge to rivers (and sewers) needs to be coordinated with background water quality, the composition of the concentrate, discharge rates, blending characteristics, and local water quality standards. Sewer discharge of concentrate up- stream of the wastewater treatment plant should also be managed so as not to exceed the capacity of the treatment plant or to adversely impact its biological processes with excessive salinity. Evaporation Ponds. Potential environmental impacts from the use of evaporation ponds include leakage of the concentrate and degradation of underlying aquifer systems or adjacent freshwater resources. Engi- neered low-permeability barriers are used to reduce the likelihood of leakage from the pond. Other factors that affect environmental water quality include sufficient basin storage volume to prevent overflow in case of major precipitation events, and location of sites topographically above long-term flood reoccurrence intervals of nearby water sources. The elevated salinity and trace constituents in evaporation ponds may be problematic for breeding and migrating birds, as was seen with the sele- nium effects on birds at the Kesterson National Wildlife Reserve (Hannam et al., 2003; Hoffman et al., 1988; NRC, 1989). Land Application. As described in Chapter 4, the allowable salinity for land application depends on the tolerance of target vegetation, perco- lation rates, and the ability to meet the groundwater quality standards and is, therefore, more viable for lower salinity concentrate. In many cases, additional dilution water is needed for land application to be feasible. It may be possible to genetically engineer better salt-tolerant plants in the future and utilize these plants for animal fodder (Grattan et al., 2004; Grieve et al., 2004). However, if transpiration from the plants exceeds precipitation to the soil, over time any salts not taken up by the plants will accumulate in the soil. If the source water, and thus the concentrate, contains contaminants of concern such as arsenic, nitrate, or other harm-
Environmental Issues 137 ful trace metals, the potential environmental impacts could include up- take of these contaminants by the plants or leaching of these contaminants into the soils or groundwater. Currently, in arid and semi- arid environments (generally west of the 100 Meridian in the United States), land application is not a physically sustainable method for dis- posal of desalination concentrate, because it is likely to exacerbate an already a large worldwide problem of soil salinization (NRC, 1993). Injection Wells. Disposal of concentrates through injection wells is required to meet criteria established by the EPA for its Underground In- jection Control Program (EPA, 2007b; see Box 5-1) to ensure that well injections do not endanger aquifers supplying drinking water by allowing the injected concentrate to enter the aquifer and degrade the resource. It should be understood that the amount of aquifer storage in a typical con- fined aquifer injection environment is smallâabout 1 m3 per 10,000 m3 of aquifer materialâand thus, there will be displacement of current aqui- fer fluids. If disposal occurs in a depleted oil reservoir or an unconfined aquifer, storage would be much largerâabout 1 m3 per 10 m3 of aquifer material. To prevent adverse impacts to surrounding aquifers, the vol- ume, location, and solute composition of any displaced fluids and how they might influence the water quality of surrounding aquifers or surface waters should be well understood. This involves quantifying all flow boundaries and simulating groundwater flow dynamics using appropriate three-dimensional numerical transport and flow models (see Box 5-3). Concentrate injection in artesian aquifer systems, which are typical of most formations used for deep-well injection, locally causes increase in fluid pressure and vertical expansion of the aquifer framework, which may be expressed as a rise in land surface. This increase in fluid pressure can also trigger earthquakes in certain geologic environments. Deep in- jection wells have caused several large-magnitude earthquakes (5 or greater on the Richter scale) and several thousand smaller ones in areas that are structurally stressed, such as the Rocky Mountains in Colorado (Evans, 1966; Hsieh and Bredehoeft, 1981) and Rangely oil field, Colo- rado (De la Cruz and Raleigh, 1972). Thus, proposed injection sites need to consider the potential for this condition if the target formation is deep and in an area that has experienced tectonic activity in the relatively re- cent geologic past. Landfilling. One potential disposal option is to convert the concen- trate from a liquid to a solid (or a dense slurry) and then dispose of the waste material in a suitable landfill (see Thermal Evaporation in Chapter 4). It requires a great deal of energy, however, to remove and recover the liquids from the concentrate and then to transport the wastes to a landfill, and these requirements may have significant financial, social, and envi-
138 Desalination: A National Perspective ronmental ramifications (see Greenhouse Gases in this chapter). Because most landfills eventually leak, there are also potential future environ- mental impacts to groundwater near the landfill. WATER QUALITY ISSUES IN DESALINATED PRODUCT 4 WATERS Because desalination processes employ advanced water treatment techniques, it is commonly assumed that desalinated water is devoid of contaminants. In reality, although desalination technologies remove vari- ous constituents to a large extent, not all constituents are fully removed and some species are removed to a lesser extent than others. In RO, a small fraction of ions, especially monovalent ions such as sodium and chloride, and dissolved organic molecules (e.g., some pesticides or her- bicides) can pass through to the permeate water. Desalinated product water quality depends on the raw water quality, the treatment technology selected (e.g., RO, distillation, electrodialysis), and within membrane technologies, by the specific membranes employed and the implementa- tion of second-pass RO. Boron and bromide are two inorganic constituents associated with water quality concerns in RO desalination, and these challenges along with approaches to mitigate these concerns are described next. Boron occurs in the oceans at an average concentration of 4.5 mg/L (Weast et al., 1985). Although thermal desalination removes boron, the rejection of boron in RO desalination is dependent on the pH. Rejection increases with pH, although the single-pass RO process is operated at a low pH to avoid scaling. Single-pass RO desalination processes do not remove the majority of boron in the raw water at typical operating pH ranges; thus, boron (occurring as borate or boric acid) can be found at milligram-per-liter levels in the finished water. Implementation of a sec- ond pass through RO membranes with a pH adjustment to place boric acid in its negatively charged borate form can provide effective boron removal (Karry, 2006; Magara et al., 1998). Second-pass RO installation and operation, however, have significant cost implications and histori- cally are not routinely included in desalination projects. 4 Minor changes have been made to this section, a related conclusion at the end of the chapter, and an associated research recommendation in Chapter 8 after release of the prepublication version to incorporate data on boron toxicity and exposure levels from the 2000 IOM report, Dietary Reference Intakes for Vita- min A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc.
Environmental Issues 139 Although boron is recognized to have a beneficial role in some physiological processes in some species, higher exposure levels may cause adverse health effects (IOM, 2000). No human health effect data are available on adverse health effects from ingestion of large amounts of boron from food and water, although data are available on the human health effects of large and small doses of boric acid or borax. However, most of the boron toxicity data come from studies in laboratory animals. High-dose boron exposures had the greatest effect on developing fetuses and on testes and led to reduced fertility in experimental animals (IOM, 2000). The EPA concluded that there is inadequate data to assess the human carcinogenicity of boron (EPA, 2006). Some research on the en- vironmental effects of boron has shown that boron at milligram-per-liter levels also can adversely impact crops and grass species (Yermiyahu et al., 2007). Boron was considered in the EPAâs second Drinking Water Con- taminant Candidate List (CCL2).5 In the CCL2, the EPA defined a reference dose (or the level of lifetime exposure at which no adverse health effects are expected) of 0.2 mg/kg/day, conservatively estimated based on developmental effects in rats as well as applied uncertainty fac- tors based on the extrapolation of data from animals to humans (EPA, 2006). Using the same toxicity data, but slightly different assumptions about uncertainty factors, the Institute of Medicine (2000) recommended a modestly higher exposure level of 0.32 mg/kg/day (or 20 mg/day for adults; 17 mg/day for adolescents, and exposures ranging from 3 to 11 mg/day for children ages 1-13). Translating these exposure guidelines into recommended limits for boron concentrations in water requires assumptions regarding other pos- sible sources of boron. According to IOM (2000), airborne boron contributes very little to the daily exposure of the general population. For humans not taking dietary supplements, diet is the major source of boron intake followed by drinking water. In the U.S., the median intake of die- tary and supplemental boron was estimated to be approximately 1.0 to 1.5 mg/day for adults (IOM, 2000).6 Based on its calculated boron refer- ence dose and an assumption that 20 percent of total daily boron consumption would come from drinking water, the EPA developed a health reference level for drinking water as 1.4 mg/L boron. With differ- ent assumptions of the total amount of boron exposures from drinking 5 See http://epa.gov/safewater/ccl/ccl2.html#chemical. 6 Ninety-fifth percentile dietary intakes of boron in the U.S. are approximately 2.3 mg/day and 1.6 to 2.0 mg/day for men and women, respectively; 2.7 mg/day and 4.2 mg/day for vegetarian men and women, respectively (Rainey et al., 1999). The average intake of supplemental boron at the ninety-fifth percentile is approximately 0.4 mg/day for adults (IOM, 2000).
140 Desalination: A National Perspective water and other sources, this water quality guidance can vary. The State of California has adopted a notification level for boron at 1 mg/L (Cali- fornia Department of Public Health, 2007). The current World Health Organization (WHO, 2004) guideline for boron in drinking water is 0.5 mg/L, but this is due to be reconsidered under the rolling revision of the guidelines (WHO, 2007). A recent draft WHO report notes that the re- vised health-based guideline (anticipated in 2008) might be 1 mg/L or higher (WHO, 2007). Because boron is not likely to be found at levels of concern in sur- face waters and groundwater, the EPA also made a preliminary determination not to regulate boron with enforceable drinking water standards (see Box 5-1; EPA, 2007a). Instead, the EPA encouraged states with public water systems that have boron at concentrations higher than 1.4 mg/L to evaluate site-specific protective measures and to consider whether state-level guidance or regulation is appropriate. Although boron is not specifically regulated in product water in the United States, consumer expectations may pressure desalination planners to design future seawater plants to follow these current guidelines. Treatment to these levels will increase the cost of new seawater desalina- tion plants. Additional analysis of the human health effects of boron in drinking water, considering other sources of boron, are needed to support firm state-level water quality guidance for seawater desalination process design that is suitably protective of public health. If seawater desalina- tion becomes a significant source for drinking water supply in the United States, additional regulatory attention or national guidance may be needed. Because it is difficult for RO technologies to meet current boron guidelines in single-pass operations if there is boron in the feedwater, membranes and processes are being developed to reduce the level of bo- ron in the product water. In some areas, specific resins combined with a small-scale RO are used to reduce the amount of boron. Boron can also be removed by optimization of RO, such as via multistep desalination or by coprecipitation with hydroxides (Cotruvo, 2005; Hyung and Kim, 2006). A technique for boron removal through reacting seawater with fly ash and coal materials has also been developed (Vengosh et al., 2004). Future seawater desalination projects should consider boron treatment options early in their planning efforts when considering the various end uses of the water produced. Bromide is another water quality consideration for membrane desali- nation projects. Bromine is formed by the reaction between bromide and free chlorine, which is often used as a biocide to control biological growth in the intake and pretreatment systems for seawater desalination plants. Bromine in its uncharged form (HOBr) passes through RO mem- branes and is found in permeate water. Bromine participates in the
Environmental Issues 141 formation of disinfection by-products (e.g., bromoform, dichloro- bromomethane, dibromochloromethane) when it reacts with natural organic matter (Laine et al., 1993; Singer, 1999; Summers et al., 1994). These by-products may have adverse human health effects (Richardson, et al., 2007) and are regulated through the SDWA (see Box 5-1) Disin- fection By-product Rules. RO membranes have relatively low rejection capability for trihalomethane disinfection by-products; thus, some of these compounds pass through the membranes and reside in the perme- ate. Brominated by-products may also be formed if the desalination product water containing bromine is blended with water from other tradi- tional sources containing natural organic matter. Bromide can also adversely affect the stability of chloramine in finished waters (Duirk and Valentine, 2007). To minimize disinfectant by-product formation, chlorine is generally used only intermittently during pretreatment, but, in some cases of high organic loading, this may not be possible. At the Tampa Bay Seawater Desalination Plant, chlorine dioxide is utilized to control biological growth in the pretreatment process. The plant had previously used free chlorine but the disinfectant was changed due to elevated disinfection by- product formation. During the post-treatment process, monochloramines are formed as a secondary disinfectant to further reduce the disinfection by-product formation. Most utilities that have switched from a free chlo- rine residual to a monochloramine residual have done so primarily to reduce disinfection by-product formation to comply with current and fu- ture disinfection by-product regulations (Dyksen, 2007). During the formation of monochloramines, the ammonia can also combine with bro- mine to form bromamines (Bousher et al., 1989). Bromamines are not recognized by the EPA as an approved drinking water disinfectant. In summary, boron, bromide, and disinfection by-products can affect product water quality. All are controllable through treatment optimiza- tion, but that treatment could aversely affect the cost of desalination. GREENHOUSE GAS EMISSIONS Water resource management currently uses significant amounts of electrical and natural gas energy to capture, treat, and transport water. The California Energy Commission (2005a) estimates that capture, transportation, and treatment of water uses approximately 5 percent of the electrical energy consumed in the state. Because of climate, geology, topography, and long water conveyance routes, the energy use for cap- ture, transportation, and treatment in California is higher than the national average of 3.5 percent of electrical energy consumed (U.S. De- partment of Energy, 2006). Desalination is an energy-intensive process
142 Desalination: A National Perspective that would add more demand. A comparison of energy use for different water sources (Table 5-2) suggests that seawater RO requires about 10 times more energy than traditional treatment of surface water (Cohen et al., 2004). Concerns over anthropogenic climate change have spurred interest in the energy requirements of desalination. Although the percent- age of statewide energy use is likely too small for desalination, planners will need a clear understanding of the energy and climate implications of desalination relative to other water supply alternatives as the nation takes steps to address the issue of greenhouse gas emissions. Energy sources other than fossil hydrocarbons can provide energy for desalination and thus avoid or significantly reduce greenhouse gas emissions. Technologies such as nuclear (19.3 percent of electrical power in the United States), hydroelectric (6.5 percent), wind (<1 per- cent), and solar photovoltaic (<1 percent) are providing input to the electrical grid (Edison Electrical Institute, 2005) and are not associated with the generation of greenhouse gases. Other alternative energy sources such as biofuels (1.6 percent) are nearly neutral in terms of greenhouse gas emissions (Adler et al., 2007), and closed-loop geother- mal systems can be essentially greenhouse gas-free. As discussed in TABLE 5-2. Comparison of Energy Use for Different Water Sources in California. Energy Used per Cubic Meter Water Source of Water (kWh/m3) Pumping groundwater 120 ft 0.14 Pumping groundwater 200 ft 0.24 Treatment of surface water 0.36 Brackish water desalination ~0.3 to 1.4 Water recycling (no conveyance) ~0.3 to 1.0 Conveyance of water (examples): Colorado River Aqueduct to San Diego 1.6 San Francisco Bay Delta to San Diego 2.6 Seawater desalination (no conveyance) ~3.4 to 4.5 NOTE: Numbers reflect cited case-study examples and are not statewide averages. SOURCE: Cohen et al. (2004), reprinted with permission from the National Re- sources Defense Council.
Environmental Issues 143 Chapter 4, thermal desalination plants can utilize low-grade or waste heat resources and substantially reduce their prime energy demands. Commercial applications of alternative energy sources to power de- salination remain somewhat limited. A 125,000 m3/day membrane desalination facility in Perth, Australia (Water Corporation, 2007), that began operation in 2007 is the first example of using alternative energy to power desalination at a large scale. The Perth wind farm is not a dedi- cated stand-alone power source; rather it feeds into the power grid from which the desalination plant contracts to withdraw its electrical power. Waste heat from Japanese nuclear plants has been used to generate boiler water for the plantsâ own use, but no dedicated nuclear power plants have yet been developed for the purpose of powering water desalination (IAEA, 2005; Minato and Hirai, 2003; Pankratz, 2005). A review of the potential for alternative energy for desalination (European Commission, 1998) and discussions of alternative energy for remote offgrid areas (GarcÃa-RodrÃguez, 2003; Tzen and Morris, 2003) suggest that several alternative energy sources hold promise. A variety of alternative energy sources have been proposed for various locations, de- pending on local conditions. These include photovoltaic (Richards and SchÃ¤fer, 2003) and heat-driven processes, such as direct solar evapora- tion (Trieb et al., 2003), closed geothermal (Bourounia et al., 1999; Karytsasa et al., 2004), ocean thermal energy conversion, and salinity- gradient solar ponds (Lu et al., 2001). Solar-powered desalination cou- pled with water reuse is a centerpiece of Masdar, an initiative in the United Arab Emirates to build the worldâs first carbon-neutral city. Pro- posed mechanical-driven alternative energies for desalination include wind power (Liu et al., 2002), wave power, tides, and hydrostatic head. Thus, there are numerous alternative energy technologies available, and these technologies may be able to provide the right quality of energy for desalination while reducing overall greenhouse gas emissions. More re- search, however, is needed to analyze the alternatives for coupling desalination with alternative energy sources in both inland and coastal areas. Climate Change and Desalination There seems to be no question that climate change will significantly impact the water resources sector and, as such, will indirectly impact de- salinization. A rise in sea level over tens of years may have adverse impacts on coastal aquifers from increased seawater intrusion. Direct impacts of rising ocean levels may over the lifetime of the project have some minor effect on desalination structures built adjacent to coastlines because current sea-level rise is approximately 2 mm/year (United Na-
144 Desalination: A National Perspective tions Intergovernmental Panel on Climate Change, 2007). Furthermore, storms associated with climate warming may be of either higher fre- quency or higher intensity. Depending on the location of the intake, the temperature of the water may increase slightly, requiring small changes to the desalination process. Although these direct impacts to desalination structures and processes appear to be small, they should be clearly under- stood prior to the design of a major desalination facility. CONCLUSIONS AND RECOMMENDATIONS Knowledge of the potential environmental impacts of desalination processes is essential to water supply planners when considering desali- nation among many water supply alternatives. All components of the water-use cycle should be considered, including source water impacts, the likely greenhouse gas emissions from the energy requirements of the desalination process, potential impacts from concentrate management approaches, and environmental health considerations in the product wa- ter. Ideally, these considerations should be compared against equally rigorous environmental impact analyses of water supply alternatives. The role of science and engineering is to clearly articulate the environmental impacts in a transparent manner so that society can make an informed decision after comparing the full economic costsâincluding environ- mental costsâand benefits among the various water supply alternatives (as discussed in Chapter 6). Because of the limited amount of long-term research, there is presently a considerable amount of uncertainty about the environ- mental impacts of desalination and, consequently, concern over its potential effects. A variety of environmental impacts are possible with desalination. Seawater desalination can cause impingement and entrain- ment of marine organisms and create ecological impacts from concentrate discharge. Desalination of inland brackish groundwater sources could lead to groundwater mining and subsidence, and improper concentrate management practices can negatively affect drinking water aquifers and freshwater biota. Site-specific information necessary to make detailed environmental conclusions on the ecological impacts of both source water withdrawal and concentrate management associated with desalination is lacking. The limited studies to date suggest that the environmental impacts may be less detrimental than many other types of water supply, but definitive conclusions cannot be made until more re- search is done.
Environmental Issues 145 Site-specific assessments of the impacts of source water with- drawals and concentrate management should be conducted and the results synthesized in a national assessment of potential impacts. Adequate understanding of impacts of source water withdrawals and concentrate management results only from site-specific assessments. The ecological effects of concentrate discharge into the ocean appear to vary widely and depend on the site-specific environment, the organisms ex- amined, the amount of dilution of the concentrate, and the use of diffuser technology. The ecological impacts of surface water intakes (i.e., im- pingement and entrainment) have been well studied for power plants but not for desalination plants, and these impacts will likely vary from place to place. General information on potential impacts from groundwater withdrawal and injection are available from decades of hydrogeologic studies for other purposes, but site-specific analyses are necessary to un- derstand the impacts from a proposed facility. Once a number of rigorous site-specific studies are conducted, this information should be synthe- sized to develop an overarching assessment of the possible range of impacts from both seawater and brackish water desalination in the United States. A characterization of the volume, hydraulic properties, flow boundary conditions, and solute chemistry of the nationâs brackish groundwater resources and a characterization of the spatial distribution, thickness, and hydraulic properties of aquifer systems suitable for con- centrate injection, relying heavily upon existing data, would assist the financial and environmental planning process for inland desalination fa- cilities. Longer-term, laboratory-based assays of the sublethal effects of concentrate discharge should be conducted. Except for a few short- term lethality studies that do not give insight into long-term effects, re- search on the impacts of concentrate discharges on organisms in receiving waters has been minimal. Longer-term laboratory-based bio- logical assays, running from weeks to months in duration, should evaluate impacts of concentrate on development, growth, and reproduc- tion using a variety of different organisms, including those native to areas where desalination plants are proposed. These results should be put into a risk assessment framework. Monitoring and assessment protocols should be developed for evaluating the potential ecological impacts of surface water concen- trate discharge. Adequate site-specific studies on potential biological and ecological effects are necessary prior to the development of desalina- tion facilities because biological communities in different geographic areas will have differential sensitivity. For large desalination facilities, environmental data should be collected for at least 1 year in the area of
146 Desalination: A National Perspective the proposed facility before a desalination plant comes online so that suf- ficient baseline data on the ecosystem are available with which to compare postoperating conditions. Once a plant is in operation, monitor- ing of the ecological communities (especially the benthic community) receiving the concentrate should be performed periodically and com- pared to reference sites. Water quality guidance, based on an analysis of the human health effects of boron in drinking water and considering other sources of exposure, is needed to support decisions for desalination process design. There are concerns about boron in product water from seawater desalination because the boron levels after single-pass RO commonly exceed current WHO health guidelines and the EPA health reference level. A range of water quality levels (0.5 to 1.4 mg/L) have been proposed as protective of public health based on different assump- tions in the calculations. The EPA has decided not to develop an MCL or health-based MCLG for boron because of its lack of occurrence in most groundwater and surface water and has encouraged affected states to is- sue guidance or regulations as appropriate. Therefore, most U.S. utilities lack clear guidance on what boron levels in drinking water are suitably protective of public health. Boron can be removed through treatment op- timization, but that treatment could aversely affect the cost of seawater desalination. Further research and applications of technology should be car- ried out on how to mitigate environmental impacts of desalination and reduce potential risks relative to other water supply alternatives. For example, intake and outfall structures could be designed to minimize impingement and entrainment and encourage improved dispersion of the concentrate in coastal discharges. Research could also explore beneficial reuse of the desalination by-products and develop technologies that re- duce the volume of this discharge. There are numerous alternative energy technologies available, and these technologies may be able to provide the right quality of energy for desalination while reducing overall green- house gas emissions; however, research is needed to analyze the alternatives for coupling desalination with alternative energy sources in both inland and coastal areas. Additional research investments should be able to clarify the potential risks of desalination and develop approaches to substantially mitigate the environmental impacts. Nevertheless, de- salination efforts do not need to be halted until this research is done and uncertainties removed.