Page 3: Guidelines for Canadian Drinking Water Quality: Guideline Technical Document - Nitrate and Nitrite

• Previous Page
• Table of Contents
• Next Page

Part II. Science and Technical Considerations

4.0 Identity, use and sources in the environment

4.1 Identity

Nitrate (NO₃¯) and nitrite (NO₂¯) are ubiquitous and naturally occurring ions in the environment. Both are products of the oxidation of nitrogen, as part of the cycle required by all living systems for the production of complex organic molecules, such as proteins and enzymes (Environment Canada, 2003; IARC, 2010).

Nitrate and nitrite are chemically expressed in two different ways: in terms of the concentration of ions (i.e., mg NO₃¯/L or mg NO₂¯/L); or as the element nitrogen (N) [i.e., mg NO₃-N/L or mg NO₂-N/L]. More specifically, 1 mg NO₃¯/L equals 0.226 mg NO₃-N/L, and 1 mg NO₂¯/L equals 0.304 mg NO₂-N/L (Pfander et al., 1993; WHO, 2007). Thus, 10 mg NO₃-N/L is equivalent to approximately 45 mg NO₃¯/L, and 1 mg NO₂-N/L is equivalent to 3.29 mg NO₂¯/L. Unless stated otherwise, units of concentrations are reported as cited in the literature and conversions to concentrations of ions (nitrate or nitrite) are provided in brackets where relevant. To obtain the equivalent ion concentration, the given concentration is multiplied by the applicable conversion factor found in Table 1 below: :

Although nitrate is the more stable form of oxidized nitrogen, under anaerobic conditions and in the presence of a carbon source, it can be reduced by microbial action to nitrite, which is relatively unstable and moderately reactive. Under low oxygen conditions, the denitrification process further reduces nitrite to nitrogen gas (Appelo and Postma, 1996).

Nitrification is a two-step process during which ammonia is oxidized to nitrite, which further is oxidized to nitrate by ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB), respectively (U.S. EPA, 2002a; IARC, 2010); these bacteria have no impact on health. This nitrification process is described according to the following equations (U.S. EPA, 2002a):

The chemical equation in which ammonia and water react to produce nitrite, hydrogen ions and free electrons

NH₃ + O₂ → NO₂¯ + 3H⁺ + 2e‾

The chemical equation in which nitrite and water react to produce nitrate, hydrogen ions and free electrons

NO₂‾ + H₂O → NO₃¯ + 2H⁺ + 2e‾

In addition to bacterial nitrification, organic nitrogen sources, such as organic matter in the soil, manures and urea-based fertilizers, can be transformed to nitrate by mineralization and hydrolysis (Ward et al., 2005a; Cartes et al., 2009).

The Chemical Abstracts Service numbers for nitrate and nitrite are 14797-55-8 and 14797-68-0, respectively. Their molecular weights are 62.00 and 46.01, respectively (U.S. EPA, 2011).

4.2 Major uses and sources

Environmental nitrate and nitrite formation occurs both naturally and through anthropogenic processes. Naturally, nitrate and nitrite are products of the oxidation of nitrogen (which comprises approximately 78% of Earth's atmosphere) by microorganisms in plants, soil and water and, to a lesser extent, by lightning (WHO, 2007; IARC, 2010).

Anthropogenic processes are the most common sources of both nitrate and nitrite. These sources include agricultural activities (including inorganic potassium or ammonium nitrate fertilizer and organic nitrate livestock manures), wastewater treatment, nitrogenous waste products in human and other animal excreta, and discharges from industrial processes and motor vehicles (Environment Canada, 2003; WHO, 2007; Keeney and Hatfield, 2008). Nitrate and nitrite can be produced as a result of nitrification process in source water or distribution systems, which add ammonia as part of chloramine disinfection practices (Kirmeyer et al., 1995; U.S. EPA, 2006a; WHO, 2007).

In addition to their use as agricultural fertilizers, nitrate and nitrite salts have been used for centuries to cure and preserve meats and fish and in the manufacture of certain cheeses. Nitrate is also used in industrial applications as an oxidizing agent (e.g., production of explosives), and purified potassium nitrate is commonly used for glass making (WHO, 2007). Historically (during the 1930s), large doses of ammonium nitrate were used medically as a diuretic, until incidences of methaemoglobinaemia were reported (L'hirondel and L'hirondel, 2002).

In a joint Cooperation on Emission Inventories, Trends and Mapping between Canada and the United States, the total emission of nitrogen oxides from 1980 to 2010 was determined to be less than 2.5 million tonnes in Canada. Wet nitrate deposition for the periods of 1990-1994 and 1996-2000 remained relatively unchanged (U.S. EPA, 2010).

It has been estimated that aerial deposition of nitrate varies widely across Canada. Annual total deposition (dry plus wet) of nitrate at the Abbotsford aquifer (British Columbia) is estimated to be 192 mg/m² (1.92 kg/ha) (McGreer and Belzer, 1999). Wet deposition of nitrogen is greater in Eastern Canada, with a 10-year annual average for 1984-1994 of 3.44 kg N ha^-1 a^-1, which occurred east of the Manitoba-Ontario border, compared with 0.80 kg N ha^-1 a^-1 west of the border (Chambers et al., 2001).

4.3 Environmental fate

Nitrogen compounds are formed in the air by lightning or discharged into it by industrial processes and motor vehicles. Nitrate is present in air as nitric acid, inorganic and organic aerosols, organic gases and nitrate radicals (WHO, 2007). Atmospheric deposition is a source of nitrate in surface water in some catchments; in other areas, the majority of deposition occurs on the land, with subsequent transport of the nitrate ions from the terrestrial basin to the surface water (Environment Canada, 2003; WHO, 2007/).

Although ammonia, nitrite and nitrate can typically be found in surface water supplies as a result of natural processes, nitrate is found more commonly than nitrite in aqueous environments, as the nitrite ion is more unstable (Appelo and Postma, 1996; OEHHA, 1997; Dubrovsky et al., 2010).

The amount of rainwater, the depth of the water table, the presence of organic material and other physicochemical properties are important determinants of the fate of nitrate in soil (WHO, 2007). Most nitrate reduction in the soil occurs through plant uptake and utilization, whereas surplus nitrates readily leach into groundwater. The nitrate ion is negatively charged and does not adsorb to clay minerals or organic matter in soils unless they have a significant anion exchange capacity, which is uncommon in Canada. Generally, it is assumed that nitrate will not adsorb to soil particles and will have a high potential for mobility (Environment Canada, 2003; WHO, 2007).

Nitrate levels in groundwater are influenced by several factors, such as land use, soil type, geochemical conditions, aquifer type and groundwater age (Dubrovsky and Hamilton, 2010). They are usually higher than levels in surface water because of the minimal vegetation uptake and because the organic carbon needed for denitrification can be limited in groundwater (Burkart and Stoner, 2002; Dubrovsky et al., 2010). In groundwater, background nitrate concentrations are reported to vary from 4 to 9 mg NO₃¯/L (0.9 to 2 mg NO₃-N/L) (Burkart and Stoner, 2003; WHO, 2007; Dubrovsky et al., 2010). Background nitrite concentrations in groundwater are typically less than 0.01 mg NO₂-N/L (0.03 mg NO₂¯/L) and have been reported as not exceeding 0.3 mg NO₂¯/L (0.09 mg NO₂-N/L) (WHO, 2007; DeSimone, 2009). Nitrate levels in Canadian lakes and rivers rarely exceed 4 mg NO₃¯/L (0.9 NO₃-N/L), and large scale sampling in the U.S estimated a background nitrate concentration in streams of 0.24 mg NO₃-N/L (0.8 mg NO₃¯/L) (Environment Canada, 2003; Dubrovsky et al., 2010). The lower levels of nitrate in surface water are due to dilution of surface runoff, plant uptake and denitrification processes (Cohn et al., 1999).

Both agricultural and non-agricultural sources of nitrogen have the potential to elevate the levels of nitrate to several hundred milligrams per litre in groundwater and surface water (Wakida, 2005; Keeney and Hatfield, 2008; Dubrovsky et al., 2010). However, agricultural activities are most commonly associated with elevated nitrate concentrations in surface water and groundwater and have been the focus of recent scientific research. Incomplete nitrogen uptake by crops results in inorganic nitrogen remaining in the soil. Most residual soil nitrogen (RSN) is in the form of nitrate, which readily leaches from the soil into the groundwater or enters surface water through runoff and tile drainage (Environment Canada, 2003; Agriculture and Agri-Food Canada, 2010). Nitrate contamination can also occur when pastures are ploughed in the autumn and the ground is left fallow during the winter. The accumulated nitrate in the soil, resulting from the mineralization and nitrification processes, may leach into the groundwater (Power and Schepers, 1989; McLenaghen et al., 1996).

As a result of intensive use of nitrogen fertilizers in agriculture and resulting runoff, nitrate pollution of surface water and groundwater has generally increased over time (Dubrovsky et al., 2010; Lindsey and Rupert, 2012). An extensive study of the occurrence and distribution of nutrients in streams and groundwater in the U.S found that the median concentration of nitrate in streams in agricultural areas was 3.8 mg NO₃-N/L (16.8 mg NO₃¯/L), approximately six times greater than the background concentration. Similarly, nitrate concentrations in groundwater in agricultural areas were elevated with a median concentration of 3.1 mg NO₃-N/L (13.7 mg NO₃¯/L). It was also noted in this study that nitrate concentrations in surface and groundwater in urban areas were also statistically higher than background levels which was attributed to wastewater effluent from municipal or industrial facilities, fertilizers applied to lawns, golf courses and parks, septic systems, and atmospheric deposition (Dubrosky et al., 2010).

High nitrate levels in drinking water are most often associated with private shallow wells with depths less than 30 m in regions with permeable soils. Nitrate concentrations tend to decrease with well depth. The well types that are most often contaminated have a shallow, bored or dug construction within unconfined aquifers (Johnson and Kross, 1990; Fitzgerald et al., 2001; Ruckart et al., 2008). Dubrovsky et al. (2010) reported that shallow domestic wells near existing or former agricultural areas have the highest probability of elevated nitrate concentrations. More than twenty percent of 406 wells in this category exceeded a nitrate concentration of 10 mg NO₃-N/L (44.3 mg NO₃¯/L). The authors also noted that geochemical conditions have a strong control on the occurrence of nitrate in groundwater. A median nitrate concentration of 5.5 mg NO₃-N/L (24.4 mg NO₃¯/L) was reported for wells in well-oxygenated groundwater in agricultural areas, but it was less than 0.05 mg NO₃-N/L (0.22 mg NO₃¯/L) in less oxygenated water, despite similar nitrogen inputs and land use surfaces. High nitrate levels in drinking water often occur simultaneously with microbial contamination and poor water quality (OEHHA, 1997; Fitzgerald et al., 2001). Fitzgerald et al. (2001) reported a correlation between total coliform detection and higher nitrate concentration (> 10 mg/L) in private wells.

Ruckart et al. (2008) suggested that nitrate levels are fairly stable in groundwater from year to year, with most variability from well to well reflecting differences in land and fertilizer use near the well head, individual characteristics of the well, such as depth and aquifer geology, and maintenance of the well.

Nitrite is an intermediate product of both nitrification and denitrification processes and can be produced when either process is incomplete. However, in aqueous environments, nitrite persists only within a limited range of redox conditions (Appelo and Postma, 1998; Rivett et al., 2008). Although nitrite levels are typically low in surface and groundwater, its presence has been reported when water sources are in close proximity to high nitrogen inputs or when redox conditions result in partial denitrification (i.e., wastewater, manure application) (DeSimone, 2008; Debrovsky et al, 2010; Toccalino et al., 2010). Dubrovsky et al. (2010) observed nitrite concentrations greater than 1 mg NO₂-N/L (3.3 mg NO₂¯/L) in five streams impacted by wastewater effluent. Forrest et al. (2006) reported that nitrite concentrations up to 10 mg NO₂-N/L (32.9 mg NO₂¯/L) had been detected in shallow groundwater below a heavily-manured field. An analysis of drinking water systems in the U.S indicates that nitrite can be present in drinking water supplies. The median nitrite concentrations in groundwater and surface water systems in the US were 0.02 and 0.03 mg NO₂-N/L (0.07 and 0.1 mg NO₂¯/L), respectively. However, more than 635 surface and groundwater systems reported at least one detection greater 1 mg NO₂-N/L (3.3 mg NO₂¯/L) and an additional 1,353 systems reported detections above 0.5 mg NO₂-N/L (1.6 mg NO₂¯/L) (U.S. EPA, 2009c).

Nitrite and nitrate can be formed as a result of nitrification of excess ammonia that occurs naturally in the source water and is not removed prior to disinfection or in systems that add ammonia as part of chloramination for secondary disinfection. Nitrification in the distribution system can increase nitrite levels 0.05-0.5 mg NO₂-N/L (0.16-1.6 mg NO₂¯/L) , although increases greater than 1 mg NO₂-N/L (3.3 mg NO₂¯/L) have been noted, particularly in stagnant parts of the distribution system (Wilczak et al., 1996; Zhang et al., 2009b).

5.0 Exposure

Canadians can be exposed to nitrates and nitrites through their presence in drinking water, food, air and soil. In addition, certain segments of the population may be exposed through the use of specific consumer products. The main route of exposure to nitrate/nitrite for the general population is via ingestion of food, followed by ingestion of drinking water. Approximately 5-8% of ingested nitrate is reduced by oral bacteria to nitrite (as reviewed in Walker, 1996; Mensinga et al., 2003). This nitrite, formed by the reduction of nitrate, represents approximately 80% of total exposure to nitrite, the remainder coming directly from exogenous sources.

5.1 Water

The nitrate concentration in surface water is generally below 18 mg/L (equivalent to 4 mg NO₃-N/L). However, recent data in many European countries have demonstrated that nitrate concentrations in surface waters have gradually increased over the last few decades, in some cases doubling over a 20-year period (WHO, 2007). Agricultural runoff, refuse dump runoff and contamination with human or other animal wastes are responsible for the progressive increase in nitrate concentrations in both surface waters and groundwaters (Liebscher et al., 1992; WHO, 2007). In most countries, nitrate concentrations in drinking water derived from surface water typically do not exceed 10 mg/L (2.3 mg NO₃-N/L).

Data collected through monitoring programs in several provinces over the years have characterized the occurrence of nitrate and its geographical distribution in Canadian drinking water. For example, in a 1982 survey of water supplies in Nova Scotia, detectable levels of nitrate (> 0.05 mg/L) were found in only 30% of community drinking water samples collected at 143 sites, with a maximum value of 2 mg/L recorded at one site (NSDH, 1982). In the summer of 1983, only six of the 59 (10.2%) municipal water supplies sampled in New Brunswick had nitrate levels greater than 4.4 mg/L, and only one sample (1.7%) had a nitrate concentration greater than 44 mg/L. In analyses conducted on 1,996 samples from registered and municipal drinking water supplies in Nova Scotia from 2000 to 2009, the percentage of samples with measurable concentrations of nitrate was 62%, with an average concentration of 5.8 mg/L; nitrate concentrations were above 45 mg/L in 19 samples (1%) (Nova Scotia Environment, 2010).

Data obtained from various jurisdictions at selected sampling sites provide some statistics on the occurrence of nitrate in Canadian drinking water systems over the 10-year period from 2000 to 2009. This unpublished information collected from the provinces and territories is a subset of data from specific monitoring and surveillance programs in each jurisdiction. Although these selected data may not accurately characterize the statistical distribution of nitrate concentrations in drinking water across Canada, they provide the average (and maximum) concentrations of nitrate when detected: Newfoundland and Labrador --1.8 mg/L (35.7 mg/L) (Newfoundland and Labrador Department of Environment and Conservation, 2010); Ontario -- 0.35 mg/L (18.8 mg/L) (Ontario Ministry of Environment, 2011); Yukon -- 0.6 mg/L (4.5 mg/L) (Yukon Environmental Health Services, 2010); Quebec -- 3.7 mg/L (93 mg/L) (Ministère du Développement durable, de l'Environnement et des Parcs, 2010); Nova Scotia -- 18.3 mg/L (207.8 mg/L) (consolidated data from registered and municipal drinking water supplies and private wells) (Nova Scotia of Environment, 2010) Saskatchewan -- 7.8 mg/L (93 mg/L) (Saskatchewan Water Security Agency, 2010); and Prince Edward Island -- 16.6 mg/L (289 mg/L) in (Prince Edward Island Department of Environment, Labour and Justice, 2010) ; Manitoba -- 2.5 mg/L ( 101 mg/L); more recent data (2009-2011) from Manitoba showed an average nitrate concentration in the treated water of 1.2 mg/L (maximum of 35.7 mg/L) (Manitoba Conservation and Water Stewardship, 2011).

In a national survey conducted by Health Canada in 2009 and 2010, 130 raw water samples and 130 treated water samples were analyzed for nitrate and nitrite. Nitrate was detected in 42.3 % of the raw water samples, at an average concentration of 3.75 mg/L (maximum of 23.9 mg/L) and in 41.5 % of the treated water samples, at an average concentration of 3.6 mg/L (maximum of 20.8 mg/L). None of the samples exceeded 45 mg/L nitrate. Nitrite was detected in 11.5 % of the raw water samples, at an average concentration of 0.05 mg/L (maximum of 0.3 mg/L) and in 6.9 % of the treated water samples at an average concentration of 0.05 mg/L (maximum of 0.3 mg/L) (Health Canada, 2012).

Generally, nitrate concentrations in well water are higher than those in surface water supplies (Liebscher et al., 1992). Nova Scotia analysed 1,996 samples taken from registered and municipal drinking water for nitrate from 2000 to 2009.  Of these, 471 samples were from a surface water source, 1,519 were from a groundwater source and the source of the 6 remaining samples were not provided. For the registered and municipal drinking systems from surface water sources, the average detected nitrate concentration was 0.8 mg/L, with the maximum value being 10.2 mg/L.  For the registered and municipal drinking systems with groundwater sources, the average detected nitrate concentration was 7.14 mg/L, with the maximum value being 141.8 mg/L.  Of these groundwater registered and municipal supplies, only 19 (1.2 %) of the samples were above 45 mg/L (Nova Scotia Environment, 2010).

Nova Scotia conducted a nitrate monitoring program which analysed 1,356 well water samples from 1999 to 2009. The average concentration of nitrate was 30.8 mg/L. The maximum annual nitrate concentration ranged from 113 to 207.8 mg/L. The result showed that the proportion of samples that exceeded 45 mg/L as nitrate for this period of time were in the range of 14.6 % to 24.4 % (Nova Scotia Environment, 2010).

In New Brunswick, nitrate concentrations in 20% of 300 well water samples collected in an agricultural area in 1984 exceeded 45 mg/L (Ecobichon et al., 1985). Very high concentrations of nitrate, up to 467 mg/L and 1063 mg/L, have been previously reported in selected groundwater samples in Ontario (Egboka, 1984) and Manitoba (Kjartanson, 1986), respectively. A study of the groundwater characteristics for seven watersheds located in intensive agricultural areas was conducted in Quebec. The analysis of the data demonstrated that there was a high probability for wells in agricultural areas, particularly surface wells, to be affected by nitrates when compared to wells in a control watershed.  Fifteen of the 59 wells (25.4%) monitored in the agricultural watershed had a nitrite-nitrate concentration (as nitrogen) above 1.5 mg/L; of these, 10 had a concentration above 3 mg/L, 8 of which were above 5 mg/L and 4 were above 10 mg/L. In comparison, of the 34 wells in the control watershed, 14 had a nitrite-nitrate concentration (as nitrogen) above 1.5 mg/L; 4 wells had a concentration above 3 mg/L; one well was above 5 mg/L and none had a concentration above 10 mg/L (Gouvernement du Québec, 2004).

Another study conducted in Quebec revealed that intensive potato culture on sandy soil may impact the groundwater nitrate concentration. Randomly selected samples showed low concentrations of nitrate: 14 (19.7%) of the 71 wells had a nitrate concentration equal to or greater than 3 mg NO₃-N/L and 4 of the wells (5.6%) had nitrate concentration greater than 10 mg NO₃-N/L. However, in the localized area with sandy soil (within 2 km of the potato fields), nitrate has been detected at concentrations equal to or greater than 3 mg NO₃-N/L in 41 (54.7%) of the 75 tested wells and 10 (13.3%) wells had concentration greater than 10 mg NO₃-N/L. According to the study, nitrate contamination seems to concentrate in sand point wells (Levallois et al., 1998).

In some cases, groundwater quality studies have focused on the water quality of domestic wells located on farms. A 1992 groundwater quality survey of 1,292 domestic wells located on farms in Ontario reported that 14% of the wells had nitrate concentrations above 10 mg NO₃-N/L (Goss et al., 1998). Similarly, Fitzgerald et al. (1997) reported that 6% of 816 farm wells sampled in 1995 and 1996 had nitrate concentrations greater than 10 mg NO₃-N/L. The mean concentration for all of the wells was 2.23 mg NO₃-N/L.

In Manitoba, 12.5% of raw well water samples analysed from 2002 to 2008 contained nitrate at concentrations higher than 45 mg/L, compared with 1.2% of surface water supplies. However, none of the supplies exceeded 45 mg/L of nitrate during  the period 2009-2011. (Manitoba Conservation and Water Stewardship, 2011).

In British Columbia, nitrate levels exceeded 45 mg/L in almost 60% of the 450 well water samples collected in the Fraser Valley. Similar to surface water supplies, average concentrations of nitrate in British Columbia groundwater appear to have gradually increased between 1975 and 1990 as a result of increased population and intensive agricultural use (Liebscher et al., 1992). In a more recent study of nitrate in a major aquifer in British Columbia, nitrate concentrations in domestic and municipal wells ranged from 4.1 to 113.7 mg/L; 10 of the 25 wells contained nitrate at levels above 45 mg/L (Wassenaar et al., 2005).

Although nitrate concentrations may be high in drinking water, nitrite levels are normally lower. The nitrite concentrations in drinking water are usually in the range of a few milligrams per litre or less (WHO, 2007). Chloramination may increase the potential for nitrite formation within drinking water distribution systems. Nitrite is not routinely monitored in all jurisdictions. However, where it is monitored, surveillance data obtained from provincial and territorial data sets demonstrate that nitrite is seldom found in Canadian drinking water samples. For example, in a survey conducted by Environment New Brunswick (1983), nitrite levels in municipal water supplies were below 0.03 mg/L, and the highest level reported was 0.3 mg/L. In another survey of groundwater sources in an agricultural area in 1984 (Ecobichon et al., 1985), nitrite levels exceeded 3.3 mg/L in only one well, whereas 20% of the 300 wells sampled had nitrate concentrations that exceeded 44 mg/L. This one high nitrite concentration was clearly attributable to contamination by surface water and manure runoff in the month of April. In Nova Scotia, 995 drinking water samples from registered and municipal drinking water systems collected in the period 2000-2009 were analyzed for nitrite. The calculated average concentration of nitrite was 0.15 mg/L, and the nitrite concentration was above 3.2 mg/L in only one sample (5 mg/L) (Nova Scotia Environment, 2010). Data from the remaining provinces and territories collected in the period from 2000 to 2009 indicate that no nitrite was detected at levels above 3.2 mg/L and that the mean nitrite concentration was below 0.1 mg/L.

5.2 Food

Nitrite and nitrate are found in many food commodities, either as natural components or as intentional additives. Vegetables and cured meats represent the main source of these compounds in diet, but they can also be found, to a lesser extent, in fish and dairy products. Nitrate and nitrite can be added as preservatives to some food items to protect them from the growth of Clostridium botulinum (which causes botulism) or to enhance their colour (characteristic pink colour of cured meat) (Food Safety Network, 2010).

Nitrate can be found at high concentrations, ranging from 200 to 2500 mg/kg, in vegetables and fruits (Van Duijvenboden and Matthijsen, 1989). Vegetables constitute a major source of nitrate, providing over 85% of the average daily human dietary intake (Gangolli et al., 1994). Many vegetables have been reported to contain high levels of nitrate, including lettuce, spinach, red beets, fennel, cabbage, parsley, carrots, celery, potatoes, cucumbers, radishes and leeks (Pennington, 1998). The concentration of nitrite in vegetables and fruits is lower than that of nitrate, at less than 10 mg/kg, and it rarely exceeds 100 mg/kg (WHO, 2007). However, vegetables that have been damaged, improperly stored, pickled or fermented may have nitrite levels up to 400 mg/kg (IARC, 2010).

Fresh meat normally contains low levels of nitrate and nitrite (Walker, 1990). However, meat and products that are cured contain much higher levels of nitrate and nitrite, depending on the amounts added as a preservative and on the curing process used (Gangolli et al., 1994). Meat products may contain nitrate at levels of < 2.7-945 mg/kg and nitrite at levels of < 0.2-64 mg/kg (ECETOC, 1988). Health Canada has limited the amount of nitrite and nitrate that can be added to meat products to 200 mg/kg (Food Safety Network, 2010).

Nitrate can also be found in dairy products at levels of < 3-27 mg/kg and nitrite at levels of < 0.2-1.7 mg/kg (ECETOC, 1988). Total nitrate exposure is negligible in breast milk; however, for bottle-fed infants consuming formula prepared with drinking water, this can be a substantial exposure pathway (OEHHA, 1997).

Levels of nitrate and nitrite in food commodities were measured as part of the total diet study conducted in Ottawa, Ontario, in 2000. Nitrite was found in negligible concentrations in cheese, cottage cheese, butter and margarine. The highest levels of nitrite found were in wieners and sausages, at 15.1 mg/kg, in luncheon meats and cold cuts, at 11.6 mg/kg, and in hot dogs, at 11.1 mg/kg. High levels of nitrate were also found in wieners and sausages, at 34.7 mg/kg, and in luncheon meats and cold cuts, at 41.2 mg/kg, but the highest levels of nitrate were found in soya-based infant formula, at 45.9 mg/kg, and in dinners prepared with meat, poultry and vegetables, at 43.7 mg/kg (Health Canada, 2003a).

A similar total diet study conducted in St. John's, Newfoundland and Labrador, in 2001 also measured nitrate and nitrite in food commodities. Nitrite was found in negligible concentrations in various food items, including cheese, cottage cheese, meat, poultry or eggs, and margarine. The highest levels of nitrite found were in luncheon meats and cold cuts, at 6.78 mg/kg, and in wieners and sausages, at 5.20 mg/kg. The highest levels of nitrate found were in frozen entrees, at 6.68 mg/kg, in processed cheese, at 5.11 mg/kg, and in luncheon meats and cold cuts, at 4.46 mg/kg. In this study, the level of nitrate in soy-based formula was found to be 1.86 mg/kg, much lower than in the Ottawa study (Health Canada, 2003b).

The total dietary intakes of nitrate and nitrite from the total diet studies performed in Ottawa (2000) and St. John's (2001) have not been calculated. However, average daily intakes from food in Canada have previously been estimated to be 44.3 mg for nitrate and 0.50 mg for nitrite, based on a survey of dietary habits (Choi, 1985). In the United States, the average adult daily intake from food has been estimated to be 40-100 mg for nitrate and 0.3-2.6 mg for nitrite (OEHHA, 1997). Other reported estimates of daily intake from many different countries are between 53 and 350 mg for nitrate and between 0 and 20 mg for nitrite (Pennington, 1998).

5.3 Air

In 1990, the annual average concentration of nitrate in ambient air was 0.88 µg/m³ for 34 communities in 50 sampling locations across Canada (Environment Canada, 1992). The average trend of aerosol nitrate concentrations measured at a station located in Nunavut from 1980 to 2007 was below 0.10 µg/m³. The highest levels of aerosol nitrate measured at this northern Canadian location were about 0.40-0.55 µg/m³ between the years 2000 and 2005 (Environment Canada, 2010). Levels of nitrite were measured in Edmonton, Alberta, monthly from November 1982 to October 1983.

Atmospheric nitrate concentrations were measured for several years of continuous sampling in a Pacific island network (Prospero and Savoie, 1989). The annual mean levels of nitrate aerosols measured for all the stations varied between 0.11 and 0.36 µg/m³. The lowest concentrations (mean 0.11 µg/m³) were constantly obtained at three South Pacific stations, where the effect of continental sources is minimal, whereas the highest nitrate concentrations (mean 0.36 µg/m3) occurred in the central North Pacific.

In the Netherlands, the mean monthly nitrate concentrations in the atmosphere were measured using a monitoring network between the summer of 1979 and the winter of 1986. The concentrations obtained ranged from 1.5 to 9.1 µg/m³ (Janssen et al., 1989).

5.4 Consumer products

Nitrate and nitrite exposure has been reported to occur from certain medications and volatile nitrite inhalants. Medications that have been reported in cases of nitrate or nitrite toxicity include quinine derivatives (anti-malarials), nitroglycerine, bismuth subnitrite (anti-diarrhoeal) and isosorbide dinitrate/tetranitrates (vasodilators). In addition, infants and children may be exposed to nitrate through silver nitrate application used in burn treatments (ATSDR, 2007). It has also been reported that household products containing amyl, butyl, isobutyl and cyclohexyl nitrites, such as air fresheners and other deodorizers, can be used as deliberate inhalants by adolescents and adults (U.S. EPA, 2007a).

5.5 Soil

Information on exposure of the general population to nitrate and nitrite in soil is not reported in the literature. Residual inorganic nitrogen levels in soil in Canada, predominantly in the form of nitrate, have been reported in the literature (Drury et al., 2007; Agriculture and Agri-Food Canada, 2010). However, as nitrate is highly soluble and weakly retained by soil, it readily leaches into groundwater or surface water (IARC, 2010). Therefore, the study of exposure to nitrate from environmental media has focused on its presence in groundwater and surface water. A national study on the potential risk of water contamination by excess nitrogen in soil found that in some agricultural areas in Canada, nitrate concentrations in drainage water may be greater than 10 mg NO₃-N/L as a result of excess nitrogen present in the soil (Agriculture and Agri-Food Canada, 2010).

6.0 Analytical methods

The U.S. Environmental Protection Agency (EPA) currently has three approved analytical methods (Method 300.0 revision 2.1, Method 300.1 revision 1.0, Method 352.1 and Method 353.2 revision 2.1) for the analysis of nitrate and nitrite in drinking water (U.S. EPA, 2007b). The following methods, developed by voluntary consensus standard organizations, are also available for the analysis of nitrate and nitrite. The cited methods in the 18th, 19th, 20th and 21st editions of the Standard Methods for Water and Wastewater as well as the online versions and selected ASTM International methods have been approved by the U.S. EPA (2007b, 2009a):

• Ion chromatography methods: SM 4110 B (APHA et al., 1992, 1995, 1998, 2005), SM 4110-B-00 (APHA et al., 2000), SM 4500-NO₃-H-00 (APHA et al., 2000), D4327-97 and D4327-03 (ASTM, 1997, 2003);
• Automated cadmium reduction methods: SM 4500-NO₃-F (APHA et al., 1992, 1995, 1998, 2005), online method SM 4500-NO₃-F-00 (APHA et al., 2000), D3867-99A (ASTM, 1999);
• Manual cadmium reduction methods: SM 4500-NO₃-E (APHA et al., 1992, 1995, 1998, 2005), D3867-99B (ASTM, 1999); and
• Automated hydrazine method: SM 4500-NO₃-H (APHA et al., 1992, 1995, 1998, 2005), SM 4500-NO₃-H-00 (APHA et al., 2000).

Additional methods using other analytical techniques have been developed for the analysis of either nitrate or nitrite. These methods have also been approved by the U.S. EPA (2009a):

• Ion selective electrode methods for analysis of nitrate: SM 4500-NO₃-D (APHA et al., 1992, 1995, 1998, 2005) and online method SM 4500-NO₃-D-00 (APHA et al., 2000);
• Spectrophotometric methods for analysis of nitrite: SM 4500-NO₂-B (APHA et al., 1992, 1995, 1998, 2005) and online method SM 4500-NO₂-00 (APHA et al., 2000).

EPA Method 300.0 revision 2.1 and EPA Method 300.1 revision 1.0 are based on ion chromatography and have method detection limits (MDLs) of 0.002 mg NO₃-N/L (equivalent to 0.009 mg NO₃¯/L) for nitrate and 0.004 mg NO₂-N/L (equivalent to 0.013 mg NO₂¯/L) for nitrite. The methods use injection of a small sample volume (2-3 mL) into an ion chromatograph for analysis of a variety of inorganic substances. The anions of interest are separated and measured using a system composed of a guard column, an analytical column, a suppressor device and a conductivity detector. Samples to be analysed for nitrate or nitrite individually should be cooled to 4°C and analysed within 48 hours. Samples for combined nitrate and nitrite analysis should be acidified using sulphuric acid to a pH less than 2 and analysed within 28 days (U.S. EPA, 1993).

APHA et al. (1992, 1995, 1998, 2005) and ASTM (1997) have two standard methods that are equivalent to EPA Method 300.0 revision 2.1: SM 4110 B and ASTM method D4327-97, respectively. These methods are based on ion chromatography with chemical suppression of eluent conductivity. A sample is passed through a series of ion exchangers where the anions are separated on the basis of their relative affinities for a low-capacity, strongly basic anion exchanger. The separated anions are then directed through a suppressor device and converted to their acid forms to be measured by conductivity. The MDLs for nitrate and nitrite using ASTM method D4327-97 are based on single U.S. EPA laboratory data and are 0.002 mg NO₃-N/L (equivalent to 0.009 mg NO₃¯/L) for nitrate and 0.004 mg NO₂-N/L (equivalent to 0.013 mg NO₂¯/L) for nitrite. The MDLs for nitrate and nitrite using SM 4110 B are a function of the sample size and conductivity scale used in the analysis; however, generally, minimum concentrations are near 0.1 mg-N/L (APHA et al., 1998). It should be noted, however, that the most recent version of APHA et al. (2005) methods, MDLs of 0.0027 mg NO₃-N/L (equivalent to 0.012 mg NO₃¯/L) for nitrate and 0.0037 mg NO₂-N/L (equivalent to 0.012 mg NO₂¯/L) are reported.

EPA Method 353.2 revision 2.1 uses an automated cadmium reduction with colorimetry method for the analysis of nitrite singly or nitrate and nitrite combined in drinking water. No MDLs are reported for this method. To use this method, a correction must be made for any nitrite present by analysing without the reduction step. A filtered sample is passed through a column of granulated copper-cadmium to reduce nitrate to nitrite. The nitrite is then formed into a coloured azo dye, which is measured colorimetrically. Samples must be preserved using sulphuric acid to a pH less than 2 and cooled to 4°C at the time of collection (U.S. EPA, 1993).

Standard Method SM 4500-NO₃-F (APHA et al., 1992, 1995, 1998, 2005) and ASTM method D3867-99A (ASTM, 1999) also use an automated cadmium reduction with colorimetry method and are equivalent to EPA Method 353.2 revision 2.1. The range reported for SM 4500-NO₃-F is 0.01-1.0 mg NO₃-N/L (equivalent to 0.04-4.4 mg NO₃¯/L), and this method is recommended particularly for levels of nitrate below 0.1 mg NO₃-N/L (equivalent to 0.4 mg  NO₃¯/L), where other methods might lack adequate sensitivity. No MDL was reported (APHA et al., 1998).

Two manual cadmium reduction methods have been approved for nitrate and nitrite analysis: Standard Method SM 4500-NO₃-E (APHA et al., 1995, 1998, 2005) and ASTM method D3867-99B (ASTM, 1999). In these methods, nitrate is reduced to nitrite in the presence of cadmium by manually adding a sample to a reduction column and measured using colorimetry after addition of a colour reagent. No detection limits are reported for these methods.

Standard Method SM 4500-NO₃-D has also been approved for the analysis of nitrate using an ion electrode method. The nitrate ion electrode is a selective sensor that responds to nitrate ion activity between 0.14 and 1400 mg NO₃-N/L (equivalent to 0.62 and 6200 mg NO₃¯/L). Standard Method SM 4500-NO₂-B is a colorimetric method for analysis of nitrite. This method determines the concentration of nitrite through the formation of an azo dye that is then measured using colorimetry. It is suitable for concentrations of nitrite between 5 and 1000 µg NO₂-N/L (equivalent to 16.4 and 3286 µg NO₂¯/L). No detection limits are reported for these methods (APHA et al., 1995).

The current U.S. EPA practical quantitation limit (PQL), based on the capability of laboratories to measure the concentrations of nitrate and nitrite within reasonable limits of precision and accuracy, is 0.4 mg N/L (U.S. EPA, 1991). The PQL for nitrate was determined using data from Water Supply studies conducted prior to the final regulation. Due to a lack of analytical performance data for nitrite, the PQL for nitrite was assigned the value determined for nitrate based on the use of similar analytical methods (U.S. EPA, 1991). Recently, as part of the U.S. EPA's 6-year review, an assessment of the analytical data for nitrate and nitrite from the Performance Evaluation Program was conducted. The U.S. EPA reported variable passing rates for laboratories analysing samples at the current PQL concentration of 0.4 mg NO₃-N/L (equivalent to 1.8 mg NO₃¯/L) and therefore have not recommended lowering the PQL. However, the data for nitrite indicated a passing rate of greater than 75% for laboratories analysing samples with concentrations of 0.4 mg NO₂-N/L (equivalent to 1.3 mg NO₂¯/L), suggesting a possible reduction in the PQL (U.S. EPA, 2009b).

7.0 Treatment technology

Nitrite is relatively unstable, is rapidly converted to nitrate in the presence of oxygen and is typically not found in high concentrations in source water. Generally, the concentration of nitrite in surface water and groundwater is far below 0.1 mg NO₂-N/L (equivalent to 0.3 mg NO₂¯/L; U.S. EPA, 2002b). Therefore, drinking water treatment methods focus on the treatment of nitrate, and treatment methods for nitrite are rarely reported (Department of National Health and Welfare, 1993). Many of the treatment technologies discussed below are, however, expected to be effective for both nitrite and nitrate. Nitrite is more prevalent in the distribution systems of municipal water treatment plants that practise chloramination for secondary disinfection. Nitrite and, to a lesser extent, nitrate concentrations in the distribution system may be elevated when nitrification occurs (Cunliffe, 1991; Kirmeyer et al., 1995, 2004; WHO, 2007; Zhang et al., 2009b).

Control options for addressing nitrate concentrations above 10 mg NO₃-N/L (equivalent to 45 mg NO₃¯/L) in source water used for drinking include blending of nitrate-rich water with water of low nitrate content, the removal of nitrate by treatment processes at the public water supply or household level and the selection of alternative low-nitrate sources. Control measures are also available for minimizing the occurrence of nitrite in distribution systems experiencing nitrification (Kirmeyer et al., 1995; Skadsen and Cohen, 2006).

Conventional water treatment processes (coagulation, sedimentation, filtration and chlorination) used at municipal water treatment plants are not effective for nitrate removal (Dahab, 1991; Kapoor and Viraraghavan, 1997; Beszedits and Walker, 1998; MWH, 2005; WHO, 2007). Nitrate is a stable and highly soluble ion with low potential for co-precipitation and adsorption. Effective technologies for the removal of nitrate from municipal water supplies include ion exchange, biological denitrification, reverse osmosis and electrodialysis (Dahab, 1991; Kapoor and Viraraghavan, 1997; Beszedits and Walker, 1998; Shrimali and Singh, 2001; MWH, 2005). The treatment processes that are capable of nitrate removal at the residential scale include reverse osmosis, distillation and ion exchange.

7.1 Municipal scale

Depending on the design and operation of the treatment plant, ion exchange, biological denitrification, reverse osmosis and electrodialysis processes are capable of removing over 80% of nitrate from water (Beszedits and Walker, 1998) to achieve effluent concentrations of nitrate as low as 3 mg NO₃-N/L(equivalent to 13 mg NO₃¯/L). Ion exchange, reverse osmosis and biological denitrification are the most commonly reported treatment technologies for the municipal-scale removal of nitrate in drinking water (Dahab, 1991; Green and Shelef, 1994; Clifford and Liu, 1995; Wachinski, 2006). Electrodialysis is less commonly reported, however, it is also effective for the reduction of nitrate in drinking water (Dahab, 1991; Hell et al., 1998). Detailed information on the effectiveness and operational considerations of the various treatment technologies for nitrate removal are available in reviews conducted by Dahab (1991), Clifford and Liu (1995), Kapoor and Viraraghavan (1997), Meyer et al. (2010) and Seidel et al. (2011).

The selection of an appropriate treatment process for a specific water supply will depend on many factors, including the characteristics of the raw water supply, the source and concentration of nitrite and nitrate, the operational conditions of the specific treatment method and the utility's treatment goals. Historically, nitrate treatment plants have been designed and operated to achieve nitrate concentrations slightly below 10 mg NO₃-N/L (equivalent to 45 mg NO₃¯/L), however, these technologies are capable of consistently achieving nitrate concentrations of 5 mg NO₃-N/L (equivalent to 22 mg NO₃¯/L). Treatment plants should strive to minimize nitrate levels in the treated water.

7.1.1 Ion exchange

Ion exchange is a physicochemical process in which there is an exchange of ions in the raw water with ions within the solid phase of a resin. Ion exchange is currently the most common nitrate removal process for municipal-scale treatment. Several studies have demonstrated that it is an effective treatment method for the removal of nitrate from drinking water (Lauch and Guter, 1986; Richard, 1989; Fletcher et al., 1991; Rogalla et al., 1991; Andrews and Harward, 1994; Clifford and Liu, 1995; Ruppenthal, 2004, 2007; Wang et al., 2007). A conventional ion exchange process involves the exchange of nitrate ions (anions) in the source water with chloride ions on the resin material (Clifford, 1999; Wachinski, 2006). As nitrate displaces chloride on the resin, the nitrate capacity of the resin is gradually exhausted resulting in effluent nitrate concentrations that increase with the volume of water that has been treated (nitrate breakthrough). Once the resin has reached its capacity (i.e., when the nitrate ion begins to appear in significant concentration in the column effluent) the resin must be regenerated using a sodium chloride (salt) solution to reverse the process. Regeneration results in a brine waste stream that contains high nitrate concentrations and must be disposed of appropriately.

Exchange resins exhibit a degree of selectivity for various ions, depending on the concentration of ions in solution and the type of resin selected. Strong base anion and nitrate-selective resins are typically used for nitrate removal. The ion exchange capacity and the selectivity of the resin are important considerations when selecting a resin. Traditional strong base anion exchange resins have a greater preference for sulphate ions than nitrate ions. Therefore, the effectiveness of these types of resins can be limited when the sulphate concentration in the source water is high. Clifford (1990, 2011) reported that the number of bed volumes that can be treated before nitrate breakthrough occurs can decrease significantly when sulphate is present. A decrease of greater than 120 bed volumes was reported when the sulphate concentration in the source water was 100 mg/L in comparison with a concentration of 40 mg/L. Chromatographic peaking can occur when a system is operated beyond nitrate breakthrough, causing the effluent nitrate concentration to be greater than the influent nitrate concentration due to sulphate ions displacing nitrate ions on the resin. Operating ion exchange columns in parallel at different stages of exhaustion can increase column run times and decrease chromatographic peaking (Clifford, 2011). In addition, nitrate-selective resins have been developed and may be a more suitable choice for source water with high sulphate concentrations (Guter, 1981, 1995; Liu and Clifford, 1996).

A common practice in the treatment of nitrate using ion exchange has been to conduct bypass blending, which involves diverting a portion of the influent flow around the treatment vessel and blending the diverted water with the treated effluent water (Clifford, 1999). As ion exchange resins can produce effluent water with minimal concentrations of nitrate, bypass blending has been used as a strategy to reduce treatment costs and achieve the required (regulated) concentration. Depending on the influent nitrate concentration, water treatment plants have typically bypassed between 10% and 50% of the influent water (Clifford, 1990, Clifford et al., 2011). Another approach that has been used to reduce costs during the operation of ion exchange treatment plants is to practise partial regeneration of the resin, which typically involves removal of only 50-60% of the exchanged nitrate on the resin (Clifford and Liu, 1995). This results in the presence of some nitrate in the treated water; however, this can be acceptable if the concentrations remain below the regulatory limit for nitrate or if bypass blending is conducted. Utilities need to give careful consideration to the level of nitrate breakthrough, the percentage of raw water that bypasses ion exchange treatment and the use of partial regeneration when determining the lowest achievable nitrate concentration using ion exchange treatment.

Many studies of full-scale ion exchange treatment plants for nitrate removal have been reported in the literature (Lauch and Guter, 1986; Richard, 1989; Dahab, 1991; Fletcher et al., 1991; Rogalla et al., 1991; Clifford and Liu, 1995; Ruppenthal, 2004, 2007; Wang et al., 2007). Treatment at these plants has generally been based on minimizing the use of regenerant and the volume of waste brine produced by bypassing a certain percentage of the raw water around the ion exchange units and operating these units to a predetermined level of nitrate breakthrough. A 3.8 ML/day full-scale ion exchange plant has been in operation since 1984 and has effectively reduced concentrations of nitrate in groundwater from 15 to below 8 mg NO₃-N/L (equivalent to 67 and 35 mg NO₃¯/L; Lauch and Guter, 1986). The process uses a strong base anion resin with a capacity of approximately 1.3 eq/L and three reaction vessels with 0.9 m of resin bed depth and empty bed contact times for each vessel of 2.54 minutes. The ion exchange units treat water to nitrate levels of 2-5 mg NO₃-N/L (equivalent to 9 and 22 mg NO₃¯/L) with run lengths of 260 bed volumes. An effluent concentration of 7 mg NO₃-N/L (equivalent to 31 mg NO₃¯/L) from the plant is achieved by treating 70% of the influent water with ion exchange and blending the remaining 30% of the influent water with the treated water. The resin is partially regenerated using a 6% sodium chloride solution with brine disposal to the municipal wastewater treatment plant (Lauch and Guter, 1986). A 4.5 ML/day full-scale ion exchange treatment plant with an average raw water nitrate concentration of 58 mg NO₃¯/L (13 mg NO₃-N/L ) was capable of achieving an effluent nitrate concentration of 45 mg NO₃¯/L  (equivalent to 10 mg NO₃-N/L ) using a nitrate-selective resin with three vessels in series followed by blending with raw water. The blended plant effluent water comprised 30% ion exchange treated water and 70% raw water with nitrate breakthrough occurring at approximately 250 bed volumes. The resin was regenerated using a 6% sodium chloride solution, with spent brine discharged to a river (Andrews and Harward, 1994). Rogalla et al. (1991) reported data from an ion exchange plant that achieved lower treated water nitrate concentrations without conducting bypass blending. A 160 m3/h plant treating surface water with nitrate concentrations ranging between 10 and 37 mg NO₃-N/L (equivalent to 44 and 164 NO₃¯/L) achieved average effluent concentrations of 2.3 mg NO₃-N/L (equivalent to 10.2 NO₃¯/L) without blending. The resin capacity was 1.2 eq/L and was exhausted after 400 bed volumes. The resin regenerant waste was sent to the municipal wastewater system (Rogalla et al., 1991).

Disposal of the resin regenerant is a major consideration for ion exchange treatment plants. Seidel et al. (2011) discuss several resin regenerant disposal options that are available to utilities including discharge to wastewater systems, waste volume reduction using drying beds, off-site approved land application and deep well injection. Since nitrate laden brine disposal is generally costly significant research into reducing the volume of the brine waste stream through recycling and/or treatment of the waste has been conducted. Several researchers have examined combining ion exchange and biological denitrification processes to reduce the concentration of nitrate in the regenerant brine (Liu and Clifford, 1996; Van der Hoek et al., 1998; Lehman et al., 2010). In this process, nitrate is removed from the source water using ion exchange followed by NaCl regeneration and subsequent biological denitrification of the spent brine to remove nitrate prior to reuse. Liu and Clifford (1996) conducted a pilot study examining the effectiveness of a sequencing batch reactor to biologically denitrify the spent regenerant brine. The denitrified brine was then filtered, sodium chloride was added and the brine was reused as regenerant. Results indicated that brine denitrification and reuse were feasible with conventional and nitrate-selective resins and were capable of reducing salt consumption and waste discharge by over 90%. More recently, pilot-scale testing demonstrated that ion exchange with brine denitrification using a fluidized bed reactor followed by brine reuse can be used successfully for nitrate treatment (Lehman et al., 2010).

Other innovative solutions to managing high concentrations of nitrate in source water have been implemented. Jones et al. (2007) reported on a combination of ion exchange, blending two different source waters and source water storage with biological denitrification to manage nitrate concentrations in the treated water. In spring, when the groundwater nitrate concentrations are 13.9 mg NO₃-N/L (equivalent to 62 mg NO₃¯/L), the well water at the first plant is supplemented with low-nitrate surface water from a nearby gravel pit, prior to filtration and disinfection. A second treatment plant uses water from an infiltration gallery located next to a surface water supply that has seasonal nitrate concentrations as high as 18 mg NO₃-N/L (equivalent to 80 mg NO₃¯/L). The corresponding peak nitrate concentration, seen in the water from the infiltration galleries, was 10 mg NO₃-N/L (equivalent to 45 mg NO₃¯/L). This lower nitrate concentration was achieved through bank filtration and biological denitrification of stored pond water that infiltrated into the galleries.

The major operational considerations when using ion exchange treatment include nitrate breakthrough and chromatographic peaking, disposal of the resin regenerant and increased corrosivity of the treated water (Clifford, 1990; Dahab, 1991; Clifford et al., 2011). The replacement of nitrate, sulphate and bicarbonate ions with chloride ions can cause mineral imbalances in the water that could increase the corrosive nature of the treated water (Schock and Lytle, 2011). This is principally due to an increase in the chloride concentrations and their possible influence on the calcium carbonate balance (i.e., hardness and buffering capacity). In some cases, post-treatment corrosion control measures must be taken to ensure that corrosion problems do not occur in the distribution system following treatment using ion exchange (Schock and Lytle, 2011). An additional consideration for utilities using strong base anion exchange resins is the potential for the release of nitrosamines from the resin. Kemper et al. (2009) found that new resin or resin that is exposed to disinfectants (chlorine and chloramines) may release nitrosamines due to shedding of manufacturing impurities (Kemper et al., 2009).

Several modified ion exchange treatment processes that result in low-brine usage have recently been reported. These processes are proprietary and include magnetic ion exchange (MIEX ® ), continuous ion exchange separation (ISEP ® ), and Basin Water and Envirogen systems (Seidel et al., 2011).

Magnetic ion exchange systems (MIEX ® ) use a resin comprised of small magnetic particles that are fluidized in reactor vessels which are placed in series. The system operates in a continuous countercurrent flow mode enabling the spent resin to be continuously regenerated. The magnetized resin particles settle to the bottom of the reactor quickly and can then be removed, regenerated and then continuously returned to the top of the reactor.  This process results in lower waste brine volumes and eliminates the risk of chromatographic peaking of nitrate in the treated water. Full-scale magnetic ion exchange treatment systems have been reported for the removal of dissolved organic carbon (Warton et al., 2007) and more recently for nitrate removal (Seidel et al., 2011). A 0.1 ML/day magnetic ion exchange plant has effectively achieved an effluent nitrate goal of less than 8 mg NO₃-N/L (equivalent to 35 mg NO₃¯/L) from an average influent nitrate concentration in groundwater of 14 mg NO₃-N/L (equivalent to 62 mg NO₃¯/L; Seidel et al., 2009). The process has a design regeneration rate of 125 BV (30 litres of resin are regenerated per 3700 L of water treated). The waste brine is stored in a tank and periodically disposed of using land application.

A continuous ion exchange separation system (ISEP ® ) has also been developed for removal of nitrate from drinking water. This process uses ion exchange columns placed in a carousel that rotates around a feed valve that is capable of delivering influent water, partially treated water, rinse water and regenerant to the various columns. The system can operate continuously with vessels in various phases of operation; feed, wash, rinse and regeneration. This process has been reported to produce more consistent treated water quality, eliminate chromatographic peaking and improve water recovery. A full-scale continuous ion exchange system was reported by Seidel et al. (2011). The 2.7 ML/day treatment systems uses 30 ion exchange vessels to reduce nitrate concentrations in groundwater ranging from 9 to 45 mg NO₃-N/L (equivalent to 40 and 200 mg NO₃¯/L) down to a treated water goal of 4 mg NO₃-N/L (equivalent to 18 mg NO₃¯/L). The resin is regenerated continuously and the system produces approximately 2800 L/hr of waste brine that is discharged to the municipal wastewater system.

A similar continuous ion exchange treatment system has been developed by Basin Water (now Envirogen Technologies Inc.) that uses stationary multiple ion exchange beds and valves operated in a staggered design. A 7.6 ML/day system is reported to reduce nitrate concentrations from 10 to 13 mg NO₃-N/L (equivalent to 45 and 56 mg NO₃¯/L) to 7.5 mg NO₃-N/L (equivalent to 33 mg NO₃¯/L).

Another modified ion exchange process, referred to as the CARIX process, uses a mixed bed with a weakly acidic and a strongly basic exchanger material for the removal of a variety inorganic minerals including calcium, magnesium, sulphate and nitrate. Although this process is not typically used for nitrate removal, it has been implemented successfully for nitrate removal at full-scale plants (Höll and Hagen, 2002). In this process, nitrate ions are exchanged for bicarbonate by the anion exchange material, using carbon dioxide as the chemical regenerant for the exchangers. A 120 m³/h mixed bed ion exchange plant reduced a nitrate concentration of 9 mg NO₃-N/L (equivalent to 40 mg NO₃¯/L) to below 6 mg NO₃-N/L (equivalent to 27 mg NO₃¯/L) in the treated water. The plant used three ion exchange vessels in parallel with filter diameters of 3.2 m and 2.5 m of exchange material. The vessels treated only five bed volumes of water prior to regeneration (Höll and Hagen, 2002).

7.1.2 Reverse osmosis/nanofiltration

Reverse osmosis and, to a lesser extent, nanofiltration can be effective technologies for producing water with low nitrate concentrations (Cevaal et al., 1995; Paynor and Fabiani, 1995; Beszedits and Walker, 1998; Santafe-Moros et al., 2005). These processes are based on forcing water across a membrane under pressure while the ionic species, such as nitrate, are retained in the waste stream. Reverse osmosis is typically used for nitrate removal when high concentrations of other dissolved solids need to be removed. In general, when utilities are considering reverse osmosis systems primarily for nitrate removal, the systems must demonstrate high nitrate rejection, high water flux and a high recovery rate for the systems to be economically viable (Dahab, 1991; Duranceau, 2001; MWH, 2005). Reverse osmosis treatment systems typically require prefiltration for particle removal and often include other pretreatment steps, such as the addition of anti-scaling agents, prechlorination/dechlorination and softening. Post-treatment typically includes pH adjustment, addition of corrosion inhibitors and disinfection (Cevaal et al., 1995). Concentrate disposal must also be considered in the design and operation of reverse osmosis plants.

Data from a full-scale reverse osmosis plant for nitrate removal indicate that a nitrate concentration of 13.7 mg NO₃-N/L (equivalent to 61 mg NO₃¯/L) can be lowered to 0.58 mg NO₃-N/L (equivalent to 2.6 mg NO₃¯/L) using spiral wound polyamide thin-film composite membranes. The plant has a capacity of 630 m³/h of permeate which is supplemented with 209 m³/h of blended water with a final nitrate concentration goal of less than 8 mg NO₃-N/L (equivalent to 35 mg NO₃¯/L). A two- stage unit was used to achieve a recovery of 80% at a feed pressure of 170 psi (1172 kPa). Pretreatment includes filtration and addition of acid and anti-scalant, and post-treatment consists of degassing (carbon dioxide) and addition of caustic, chlorine and zinc orthophosphate (Cevaal et al., 1995). Schoeman and Steyn (2003) reported data from a small reverse osmosis plant that is capable of rejecting high concentrations of nitrate in the source water. The 55 m³/d plant operates with a 50% water recovery and a pressure of 1375 kPa. A nitrate rejection of between 96% and 98% was achieved in the plant with permeate concentrations lowered to less than 5 mg NO₃-N/L (equivalent to 22 mg NO₃¯/L) from a feed concentration that varied between 42 and 53 mg NO₃-N/L (equivalent to 186 and 235 mg NO₃¯/L).

More recently, pilot-scale testing has been conducted on the use of ultra low pressure reverse osmosis (ULPRO) membranes for the removal of nitrogen species. ULPRO membranes have the advantage of requiring lower operating pressures and therefore have lower operating costs. Drewes et al. (2008) demonstrated that ULPRO membranes are capable of rejecting over 90% of nitrate in the feed water. The concentration of nitrate in the treated water was consistently below 1 mg NO₃-N/L (equivalent to 5 mg NO₃¯/L), whereas feed water nitrate concentrations varied between 3 and 11 mg NO₃-N/L (equivalent to 13 and 49 mg NO₃¯/L). The membrane system was operated at a recovery of 82% and a feed pressure of 130 psi (900 kPa). The authors noted that membrane compaction and fouling increased the feed pressure required up to 160 psi (1103 kPa) following 1 month of operation.

Bench- and pilot-scale testing has also been conducted to evaluate the effectiveness of nanofiltration membranes for nitrate removal. It has been demonstrated that, in general, membranes with smaller pore sizes are required to achieve a rejection of greater than 75% of nitrate (Van der Bruggen et al., 2001). A laboratory-scale study evaluated the effectiveness of four nanofiltration membranes for pesticide and nitrate removal. It was found that one membrane with small pore sizes was capable of rejecting 76% of nitrate from 10 mg NO₃-N/L (equivalent to 45 NO₃¯/L) in the influent water to achieve 2.5 mg NO₃-N/L (equivalent to 11 mg NO₃¯/L) in the treated water (Van der Bruggen et al., 2001). Pilot plant testing of nanofiltration membranes found that only one of the four thin-film composite polyamide membranes tested was capable of effectively rejecting nitrate. Greater than 90% rejection of nitrate was observed with an influent concentration of 100 mg NO₃¯/L (equivalent to 22.5 mg NO₃-N/L) (Santafe-Moros et al., 2005). Other research studies have found that nanofiltration membranes were not effective in removing nitrate from water (Bohdziewicz et al., 1999; Drewes et al., 2008). Therefore, testing of the nanofiltration membrane selected for nitrate removal will be an important step for utilities considering this treatment process.

Considerations when using reverse osmosis treatment include disposal of the reject water and possible increased corrosivity of the treated water (Schock and Lytle, 2011). Reverse osmosis rejects a significant portion of the influent water as contaminant-rich brine (Taylor and Wiesner, 1999), and the concentrate discharge must be disposed of appropriately. The removal of contaminants can cause mineral imbalances that could increase the corrosive nature of the treated water (Schock and Lytle, 2010). In most cases, post-treatment corrosion control measures need to be taken.

7.1.3 Electrodialysis/electrodialysis reversal

Electrodialysis is a membrane process that uses an electric potential for removing charged species from water. This electrochemical process removes cations and anions from the source water by forcing them through cation or anion exchange membranes in a stack of cation-anion membrane pairs under the influence of a DC voltage (AWWA, 1995). Nitrate is removed by moving from the influent water through the anion exchange membrane where it is then rejected by the cation exchange membrane and removed in the waste stream (Hell et al., 1998). Anion exchange membranes that selectively remove nitrate have also been developed (Chebi and Hamano, 1995). The electrodialysis reversal process is based on reversing the polarity of the electrodes several times every hour of operation to alter the direction of ion movement, which greatly helps to reduce membrane scaling. In general, electrodialysis and electrodialysis reversal treatment systems produce less reject water and have lower power consumption than other membrane processes (Kapoor and Viraraghaven, 1997).

A full-scale electrodialysis treatment plant demonstrated that a raw water nitrate concentration of 36 mg NO₃-N/L could be reduced to 9 mg NO₃-N/L (equivalent to 40 mg NO₃¯/L) using three parallel membranes stacks each with a hydraulic capacity of 48 m³/h. The concentrate was sent for treatment to the municipal wastewater treatment plant (Hell et al., 1998). Pilot-scale testing of an electrodialysis process using nitrate-selective membranes demonstrated a nitrate rejection of approximately 80% to achieve a nitrate concentration of 3 mg NO₃-N/L (equivalent to 13 mg NO₃¯/L) in the effluent. The system operated at 1 m³/h using two stacks of membranes and a recovery rate of 95% (Chebi and Hamano, 1995).

Elmidaoui et al. (2002) reported results from a 24 m³/d pilot-scale electrodialysis reversal plant that successfully reduced the nitrate concentration from 16 mg NO₃-N/L (equivalent to 71 mg NO₃¯/L) down to 2 mg NO₃-N/L (equivalent to 9 mg NO₃¯/L) at an 80% recovery rate. The plant was equipped with two membrane stacks in series with a total available membrane area of 500 cm² and an automatic polarity reversal every 20 minutes. A voltage of 24 V and a power consumption of 0.43 kWh/m³ were used to achieve these results. Seidel et al. (2011) also reported data from a full-scale electrodialysis reversal treatment system that was capable of removing over 93% of nitrate to achieve a treated water nitrate concentration of 0.97 mg NO₃-N/L (equivalent to 4.3 mg NO₃¯/L). The system had three stages and operated at 90% water recovery. Additional laboratory- and pilot-scale studies have been conducted on electrodialysis treatment systems to evaluate system optimization, new membrane performance and effects of variable source water (Salem et al., 1995; Elhannouni et al., 2000; Sahli Menkouchi et al., 2006, 2008).

The main considerations for systems using electrodialysis and electrodialysis reversal for nitrate removal are the operational complexity of the system, disposal of the reject water and the need for pH adjustment of the treated water (Kapoor and Viraraghaven, 1997).

7.1.4 Biological denitrification

Biological denitrification treatment processes are based on the removal of nitrate in source water through the biological reduction of nitrate to nitrogen gas (denitrification) in an anoxic environment. The denitrification process used for potable water treatment requires the addition of an electron donor to the source water so that the nitrate can be biologically reduced to nitrogen. There are two main types of biological denitrification systems that are used for potable water treatment. Heterotrophic denitrification uses organic compounds, such as ethanol or acetic acid, as both the electron donor and carbon source. Autotrophic denitrification uses an inorganic compound such as hydrogen or sulphur as the electron donor and inorganic carbon such as carbon dioxide as the carbon source. Biological denitrification systems can be designed as fixed bed reactors, fluidized bed reactors, membrane bioreactors and membrane biofilm reactors. In general, biological denitrification treatment systems require post-treatment to remove biomass and biodegradable organic materials that are present in the reactor effluent. Typical post-treatment includes aeration, filtration, activated carbon and disinfection (MWH, 2005; Meyer et al., 2010).

Biological denitrification processes have been used in Europe for many years for the removal of nitrate from drinking water (Richard, 1989; Dahab, 1991; Rogalla et al., 1991; Dordelmann, 2009) and have more recently been considered in North America (Meyer et al., 2010). However, there is currently limited full-scale experience with biological denitrification in North America. Meyer et al. (2010) reported that the most common and effective biological denitrification systems are heterotrophic two-stage, fixed bed, upflow systems and autotrophic hydrogen-based membrane biofilm reactors. Design and operational considerations for biological denitrification plants include electron donor and nutrient dosing, dissolved oxygen, pH and temperature control as well as biofilm management (Meyer et al., 2010). The most important operational parameters identified are optimization of the nitrate surface-loading rates and substrate and nutrient dosing. Under dosing of the electron donor can result in insufficient nitrate removal or the formation of nitrite due to incomplete denitrification. Overdosing of electron donors can result in excess biodegradable matter in the reactor effluent (Dahab, 1991; Dordelmann, 2009; Meyer et al., 2010).

Several authors have reported on full-scale biological denitrification drinking water treatment plants encompassing a wide variety of reactor configurations, donor types and denitrification mechanisms (Richard, 1989; Dahab, 1991; Rogalla et al., 1991; Dordelmann, 2009; Meyer et al., 2010). In general, it is reported that biological denitrification can reduce nitrate concentrations in source water as high as 100 mg NO₃-N/L (equivalent to 443 mg NO₃¯/L) to concentrations approaching 1 mg NO₃-N/L (equivalent to 4 mg NO₃¯/L; Dahab, 1991).

Meyer et al. (2010) reported on five heterotrophic fixed bed denitrification plants using ethanol as the electron donor that were capable of reducing influent nitrate concentrations in the range of 10-15 mg NO₃-N/L (equivalent to 45-66 mg NO₃¯/L) down to 2-6 mg NO₃-N/L (equivalent to 9-27 mg NO₃¯/L). The operating conditions of one of the most efficient plants included four fixed bed heterotrophic bioreactors in series. At a flow rate of 300 m³/h and a maximum nitrate loading rate of 1.5 kg-N/m³-d, the plant is capable of removing over 90% of nitrate down to a concentration of less than 2 mg NO₃-N/L (equivalent to 9 mg NO₃¯/L). Post-treatment includes a two-stage filtration process--aerobic filtration with oxygen addition and activated carbon filtration, to remove excess gas, biomass and carbon sources--followed by disinfection (Mateju et al., 1992). Similarly, Dahab (1991) reported on several heterotrophic biological denitrification plants capable of achieving nitrate removal efficiencies ranging from 65% to 95% with influent nitrate concentrations between 14 and 20 mg NO₃-N/L (equivalent to 62 and 89 mg NO₃¯/L). Full-scale heterotrophic fluidized bed treatment plants have also been reported in the literature. Mateju et al. (1992) reported data from a fluidized- bed treatment plant using methanol as the electron donor and phosphate nutrient addition. The plant achieves a reduction in nitrate from 23 mg NO₃-N/L (equivalent to 102 mg NO₃¯/L) down to 6 mg NO₃-N/L (equivalent to 27 mg NO₃¯/L). The authors noted, however, that elevated nitrite concentrations were observed intermittently in the effluent water. Dahab (1991) reported a 63% nitrate removal efficiency of an influent nitrate concentration of 15 mg NO₃-N/L (equivalent to 66 mg NO₃¯/L) using a fluidized bed configuration.

Although less commonly used, autotrophic denitrification has also been implemented at full-scale treatment plants. A 50 m³/h plant with an influent nitrate concentration of 18 mg NO₃-N/L (equivalent to 80 mg NO₃¯/L) is capable of achieving greater than 90% removal of nitrate to effluent concentrations of less than 1 mg NO₃-N/L (equivalent to 4 mg NO₃¯/L) at a nitrate loading rate of 0.25 kg-N/m³ -d. The plant consists of nine fixed bed bioreactors fed raw water that has been supersaturated with hydrogen and dosed with phosphate and carbon dioxide and post-treatment of aeration, two-layer filtration and ultraviolet  disinfection (Gros et al., 1986). A 35 m³/h full-scale autotrophic denitrification plant using a sulphur/limestone reactor was reported by Mateju et al. (1992). The process consists of vacuum deaeration of the raw water followed by upflow filtration in a bioreactor containing sulphur and limestone followed by aeration and artificial recharge. The plant achieves greater than 90% removal of nitrate with raw water containing 16-18 mg NO₃-N/L (equivalent to 71-80 mg NO₃¯/L).

Biological denitrification processes can be integrated with membrane technology using membrane bioreactors and membrane biofilm reactors. These combined technologies allow for retention of the biomass and in some cases the electron donors and nutrients so that post-treatment may not be as extensive as with conventional biological denitrification. There are many types of membrane systems that have been researched for this application, including extractive membrane bioreactors, ion exchange membrane bioreactors, gas transfer membrane bioreactors and pressure-driven membrane bioreactors (Velizarov et al., 2002; Matos et al., 2005; Xia et al., 2009; Meyer et al., 2010). It was reported that most of the research conducted on these systems has been at the bench or pilot scale, with limited full-scale applications (McAdam and Judd, 2006).

A 400 m³/d full-scale membrane bioreactor with ethanol, phosphoric acid and powdered activated carbon addition and hollow fibre ultrafiltration membranes is capable of reducing a median source water nitrate concentration of 12 mg NO₃-N/L (equivalent to 53 mg NO₃¯/L) to less than 2 mg NO₃-N/L (equivalent to 9 mg NO₃¯/L) at a nitrate loading rate of 0.1 kg NO₃-N/m³-d. The treated water from the membranes has a low organic carbon concentration, no nitrites and good biological stability (Urbain et al., 1996). Meyer et al. (2010) reported that a pilot-scale hydrogen-based autotrophic membrane biofilm reactor with hollow fibre membranes was capable of completely removing nitrate at a concentration of 19.6 mg NO₃-N/L (equivalent to 87 mg NO₃¯/L) at a surface loading rate of 3.0 g-N/m²-d. The authors noted that when the influent nitrate concentration was increased to 32.1 mg NO₃-N/L (142 mg NO₃¯/L), incomplete denitrification produced an effluent nitrite concentration of 2.5 mg NO₂-N/L (equivalent to 8 mg NO₂¯/L).

As nitrite is an intermediate compound in the reduction of nitrate to nitrogen gas, utilities need to ensure that their systems are optimized so that the biological process is complete and nitrite is not present in the treated water.

7.1.5 Emerging treatment technologies

Several drinking water treatment technologies for nitrate are being developed but are still primarily in the experimental stage or do not have published information on the effectiveness of large-scale applications. Some of the emerging technologies include the following:

• Chemical denitrification: Kapoor and Viraraghavan (1997) and Shrimali and Singh (2001) reviewed and provided general information on several laboratory studies that have been conducted on the use of metals for the chemical reduction of nitrate to other nitrogen species. Chi et al. (2005) demonstrated that a 50% reduction in nitrate from an initial concentration of 1500 mg NO₃-N/L (equivalent to 6, 650 mg NO₃¯/L) could be achieved using metallic iron when water was acidified to a pH of 5. Luk and Au-Yeung (2002) reported a maximum nitrate removal of 62% to achieve treated water concentrations of 8.3 mg NO₃-N/L (equivalent to 37 mg NO₃¯/L) using 300 mg/L of aluminum powder and a pH of 10.7. Seidel et al. (2008) conducted pilot-scale testing of sulphur-modified iron for the chemical reduction of nitrate. Results indicated that the highest nitrate removal, from approximately 15 mg NO₃-N/L to 10 mg NO₃-N/L (equivalent to 66 to 45 mg NO₃¯/L), occurred at a pH of 6.0 and an empty bed contact time of 30 minutes. The authors noted that a treated water goal of 8 mg NO₃-N/L (35 mg NO₃¯/L) was not achieved consistently during the pilot testing.
• Catalytic denitrification: Research studies have also examined the chemical denitrification of nitrate in the presence of catalyst (Reddy and Lin, 2000; Chen et al., 2003). Reddy and Lin (2000) conducted laboratory tests of catalytic denitrification using three catalysts: palladium, platinum and rhodium. Rhodium was the most effective catalyst for nitrate removal. The results demonstrated that addition of 0.5 g of rhodium per litre of water could decrease nitrate concentrations from 9 mg NO₃-N/L to 3 mg NO₃-N/L (equivalent to 40 to 13 mg NO₃¯/L) at a redox potential of −400 mV. Chen et al. (2003) found that a 4:1 palladium-copper combined catalyst maximized nitrate reduction to nitrogen gas. An initial nitrate concentration of 22.6 mg NO₃-N/L (equivalent to 100 mg NO₃¯/L) was reduced to less than 1 mg NO₃-N/L (equivalent to 5 mg NO₃¯/L) after 20 minutes of reaction time.
• Polyelectrolyte-enhanced ultrafiltration: Zhu et al. (2006) demonstrated greater than 90% removal of 60 mg NO₃-N/L (equivalent to 266 mg NO₃¯/L) using polyelectrolyte-enhanced ultrafiltration. The percentage of nitrate removed depended on the types of chelating polymers and the ultrafiltration membrane that was used in the study.

7.1.6 Nitrification in the distribution system

Nitrite and nitrate can be formed in the distribution system as a result of nitrification of excess ammonia that occurs naturally in the source water and is not removed prior to disinfection or in systems that add ammonia as part of chloramination for secondary disinfection. The potential increase of nitrite in the distribution system is significant, as, in some cases, the increase in nitrite due to nitrification may lead to a concentration that exceeds the guideline value of  1.0 mg NO₂-N/L (equivalent to 3 mg NO₂¯/L). In the case where a water treatment plant removes nitrate and nitrite to levels that just meet the guideline values and the water system uses chloramines for disinfection, there is a potential for both nitrate and nitrite values to increase above the guideline values in the distribution system during a nitrification event (U.S. EPA, 2002b). However, when nitrite concentrations increase due to nitrification, the primary concern for utilities is that nitrite consumes chlorine and decomposes chloramines which results in an increase in microbial counts, including an increase in the potential presence of coliform bacteria in the distribution system (Smith, 2006).

Nitrification is a sequential microbiological process where ammonia is oxidized to form nitrite and then nitrite is oxidized to form nitrate. Two groups of chemolithotrophic nitrifying organisms, ammonia-oxidizing bacteria (ammonia to nitrite) and nitrite-oxidizing bacteria (nitrite to nitrate), carry out this process (Kirmeyer et al., 1995, 2004; U.S. EPA, 2002b). Nitrification can have adverse impacts on water quality, including increasing nitrite and nitrate levels, increasing bacterial regrowth, and lowering chloramine residuals, pH and dissolved oxygen (Kirmeyer et al., 1995; Odell et al., 1996; Wilczak et al., 1996; U.S. EPA, 2002b; Kirmeyer et al., 2004; Zhang et al., 2009b). Studies have also reported possible links between corrosion problems and nitrification (Edwards and Dudi, 2004; Zhang et al., 2009a, 2010).

Nitrification in distribution systems where chloramine is used as a secondary disinfectant, can increase nitrite levels on the order of 0.05-0.5 mg NO₂-N/L (equivalent to 0.16-1.6 mg NO₂¯/L), although increases greater than 1 mg NO₂-N/L (equivalent to 3 mg NO₂¯/L) have been noted, particularly in stagnant parts of the distribution system (Wilczak et al., 1996; Zhang et al., 2009b). It was noted that maximum nitrite concentrations occur in areas of the distribution system with the longest detention time, such as at the extremities of the system or dead ends (Kirmeyer et al., 1995; Harrington et al., 2002). The U.S. EPA (2002b) reported that increases in nitrite levels during nitrification episodes more frequently range between 0.015 and 0.1 mg NO₂-N/L (equivalent to 0.048 and 0.32 mg NO₂¯/L). Generally, increases in nitrate concentrations during nitrification are small; however, increases of more than 1 mg NO₃-N/L (equivalent to 5 mg NO₃¯/L) have been reported (Cunliffe, 1991; Kirmeyer et al., 1995). In some cases, increases in nitrate concentrations have been observed with no corresponding increase in nitrite concentrations, indicating a nitrification episode with complete nitrification (Kirmeyer et al., 1995; Wilczak et al., 1996).

Factors contributing to nitrification in the distribution system include warm water temperatures, pH, a low Cl₂:NH₃-N ratio and the concurrent increase of free ammonia concentrations and chloramine residual. The optimum temperature for nitrifiers (nitrifying bacteria) to grow ranges between 20°C and 30°C (Baribeau, 2006). However, regrowth and nitrification can occur at temperatures as low as 5°C or even less in systems with long detention times (Pintar et al., 2000). Kors et al. (1998) discussed a case of nitrification under extreme cold-water conditions (below 4°C). An increase in temperature will increase the chloramine decomposition rate, which will subsequently promote nitrification, since more free ammonia will be released (Baribeau, 2006).

A number of distribution system parameters such as detention time, reservoir design and operation, presence of dead end mains, sediment and tuberculation in piping, biofilm, and the absence of sunlight can affect nitrification (Skadsen, 1993; Kirmeyer et al., 1995; 2004; U.S. EPA, 2002b; Zhang et al., 2009b; Baribeau, 2010). Harrington et al. (2002) and the U.S. EPA (2002b) noted that increases in nitrite up to 1 mg NO₂-N/L due to nitrification could theoretically occur in any system in which the total ammonia concentration entering the distribution system is greater than 1 mg-N/L. In theory, utilities using a CL₂:NH₃-N ratio of 3:1 could see increases of nitrite greater than 1 mg NO₂-N/L if a chloramine dose of 3 mg/L as Cl₂ is used in the treatment plant. A discussion of the optimal free ammonia and chlorine to ammonia-nitrogen ratio to minimize nitrification is provided in the guideline technical document for ammonia (Health Canada, 2013).

The formation of nitrite in combination with a decrease in chloramine residual and free ammonia concentrations in the distribution system may be used as an indicator of nitrification (Kirmeyer et al., 1995). Utilities that are chloraminating should monitor for nitrite and nitrate in the distribution system in addition to ammonia, total chlorine residual, HPC and other nitrification indicators. A site-specific evaluation is generally necessary to establish a nitrification monitoring program. The program should identify system-specific alert and action levels, which can be used to determine the appropriate level of action to address nitrification. The monitoring frequency of the parameters depends on the location and the purpose of the data. Changes in the trend of  priority nitrification parameters (such as total chlorine residual, nitrite and nitrate) in the distribution system should trigger more frequent monitoring of other parameters such as free ammonia.

Some studies have proposed that a nitrite level of 0.05 mg NO₂-N/L (equivalent to 0.16 mg NO₂¯/L) may be used as a critical threshold indicator in nitrification. It is proposed that once this level is reached, severe nitrification may be occurring, and control measures need to be implemented (Kirmeyer et al., 1995; Harrington et al., 2002; Pintar et al., 2005). However, Pintar et al. (2005) noted that full-scale data indicated that a nitrite level of 0.05 mg NO₂-N/L (equivalent to 0.16 mg NO₂¯/L) is too high to be used as a predictor of nitrification. Smith (2006) further suggested that a nitrite concentration of 0.015 mg NO₂-N/L (equivalent to 0.048 mg NO₂¯/L) should be used as an action level for utilities to address nitrification in the distribution system.

There are many preventive and control measures that can be taken to address nitrification. Any strategy should also ensure that other Guidelines for Canadian Drinking Water Quality are not exceeded.  Detailed information on nitrification control in chloraminated systems is available in reports and reviews by Kirmeyer et al. (1995), Skadsen and Cohen (2006) and Zhang et al. (2009b). Preventive methods include control of water quality parameters (pH, free ammonia entering the distribution system, organic matter) and operating parameters (Cl₂:NH₃-N weight ratio and chloramine residual), corrosion control programs, distribution system pipe flushing, establishing booster chlorination or chloramination stations, temporary/seasonal free chlorination (breakpoint chlorination), and  chlorite addition. Corrective methods are similar to the preventive methods and include distribution system pipe flushing, temporary/seasonal free chlorination (breakpoint chlorination), reservoir cycling or cleaning and chlorite addition. However, the addition of chlorite is considered to be controversial as its presence can lead to the formation of chlorate (Skadsen and Cohen, 2006). Utilities wishing to use chlorite addition as a control strategy should ensure that Health Canada's guidelines for chlorite and chlorate are not exceeded (Health Canada, 2008).

The different measures used to control the nitrification episodes vary in their effectiveness and their ability to provide long-term improvements in nitrification problems. For these reasons, comprehensive strategies aimed at the prevention of nitrification episodes are recommended over strategies aimed at controlling nitrification as they occur.

7.1.7 Formation of nitrate or nitrite from other treatment technologies

In some cases, other treatment methods may form nitrate or nitrite at the treatment plant. The principal by-products of UV photolysis of N-nitrosodimethylamine (NDMA) are dimethylamine (DMA) and nitrite (Bolton and Stefan, 2000; Mitch et al., 2003). When UV/hydrogen peroxide are applied, nitrate is the major degradation product (Bolton and Stefan, 2000).

The formation of nitrite has also been observed during disinfection using low-pressure UV light in source water containing nitrate. Lu et al. (2009) found that at a pH of 9.5, low-pressure UV treatment of water with an initial nitrate concentration of 10 NO₃-N/L produced up to 0.1 mg NO₂-N/L.

7.2 Residential scale

Municipal treatment of drinking water is designed to reduce contaminants to levels at or below guideline values. As a result, the use of residential-scale treatment devices on municipally treated water is generally not necessary but primarily based on individual choice. In cases where an individual household obtains its drinking water from a private well, a private residential drinking water treatment device may be an option for reducing nitrate and nitrite concentrations in drinking water. For most influent concentrations of nitrate in source water, residential treatment devices can remove nitrate from drinking water to concentrations below 45 mg NO₃¯/L (equivalent to 10 mg NO₃-N/L). It is important to note that the removal efficiency will also depend on the effectiveness of the treatment device selected.

Before a treatment device is installed, the water should be tested to determine general water chemistry and verify the presence and concentrations of nitrate and nitrite in the source water. It should be noted that bacterial contamination of a well water supply can occur in conjunction with nitrate contamination. Therefore, the bacterial and chemical aspects of the water quality should be considered prior to selecting a water treatment device. Periodic testing by an accredited laboratory should be conducted on both the water entering the treatment device and the finished water to verify that the treatment device is effective. Devices can lose removal capacity through use and time and need to be maintained and/or replaced. Consumers should verify the expected longevity of the components in their treatment device as per the manufacturer's recommendations. Residential drinking water treatment processes can be routinely monitored to ensure that treatment units are performing optimally.

Health Canada does not recommend specific brands of drinking water treatment devices, but it strongly recommends that consumers use devices that have been certified by an accredited certification body as meeting the appropriate NSF International (NSF)/American National Standards Institute (ANSI) drinking water treatment unit standards. These standards have been designed to safeguard drinking water by helping to ensure the material safety and performance of products that come into contact with drinking water. Certified devices for the reduction of nitrate and nitrite from drinking water in residential systems generally rely on reverse osmosis (RO) or on ion exchange, although devices that rely on distillation treatment processes may also be available.

Certification organizations provide assurance that a product conforms to applicable standards and must be accredited by the Standards Council of Canada (SCC). In Canada, the following organizations have been accredited by the SCC to certify drinking water devices and materials as meeting NSF/ANSI standards (SCC, 2011):

• Canadian Standards Association International (www.csa-international.org);
• NSF International (www.nsf.org);
• Water Quality Association (www.wqa.org);
• Underwriters Laboratories Inc. (www.ul.com);
• Quality Auditing Institute (www.qai.org);
• International Association of Plumbing & Mechanical Officials (www.iapmo.org).
• An up-to-date list of accredited certification organizations can be obtained from the SCC (www.scc.ca).

The NSF/ANSI standards for nitrate and nitrite removal currently require testing of a device for the reduction of 30 mg-N/L as nitrogen (27 mg NO₃-N/L plus 3 mg NO₂-N/L) to 10 mg-N/L of nitrogen, in which no more than 1 mg/L can be in the form of nitrite-nitrogen (NSF/ANSI 2009a, 2009b, 2009c). Distillation systems should only be installed at the point of use as the water they have treated may be corrosive to internal plumbing components.

Available data suggests that residential reverse osmosis units can achieve lower treated water nitrate concentrations, particularly when the source water nitrate concentration is below 17 mg NO₃-N/L (equivalent to 75 mg NO₃¯/L; U.S. EPA, 2002b, 2006b, 2007c). The U.S. EPA has reported on several studies evaluating POU as a compliance option for the removal of various contaminants. The reports assessed the effectiveness of residential reverse osmosis treatment devices for the removal of various contaminants, including nitrate. One study indicated that one type of reverse osmosis device was consistently capable of removing an average influent nitrate concentration of 10 mg NO₃-N/L (equivalent to 45 mg NO₃¯/L) down to below 1 mg NO₃-N/L (equivalent to 5 NO₃¯/L) in the treated water (U.S. EPA, 2007c). Additional information on nitrate removal using several types of residential reverse osmosis devices also indicated that influent nitrate concentrations up to 11 mg NO₃-N/L (equivalent to 49 mg NO₃¯/L)could be lowered to less than 1 mg NO₃-N/L (equivalent to 5 mg NO₃¯/L; U.S. EPA, 2006b). The study concluded that reverse osmosis devices can have varying removal efficiencies which was attributed to the different types of membranes that were used in the units.

Although devices that are certified to NSF/ANSI Standard 58 verify only that a final concentration of less than 10 mg NO₃-N/L (equivalent to 45 mg NO₃¯/L) is achieved, select residential reverse osmosis devices are capable of achieving lower treated water concentrations. Accredited certification organizations report that 30 to 50% of the reverse osmosis devices certified for nitrate reduction are capable of achieving treated water concentrations below 5 mg NO₃-N/L (equivalent to 22 mg NO₃¯/L; NSF, 2012; WQA, 2012). Consumers may refer to the manufacturer's claims in its literature to obtain more information on the amount of nitrate and nitrite that a treatment device may remove, as well as operational and maintenance requirements. Routine testing monitoring of nitrate concentrations in the treated water can be conducted to determine if lower concentrations are being achieved.

Reverse osmosis systems are intended for point-of-use installation, as larger quantities of influent (incoming) water are needed to obtain the required volume of treated water, which is generally not practical for residential-scale point-of-entry systems. Reverse osmosis systems should only be installed at the point of use as the water they have treated may be corrosive to internal plumbing components. A consumer may need to pre-treat the influent water to reduce fouling and extend the service life of the membrane.

Ion exchange may also be a feasible technology for nitrate removal. Ion exchange technology is typically designed and constructed for residential use by drinking water treatment system providers or dealers. Health Canada strongly recommends that homeowners ensure that these systems are constructed using materials certified to NSF/ANSI Standard 61 (NSF/ANSI, 2009d). Ion exchange using a standard type 1 or type 2 polystyrene strong-base anion exchange resin can result in treated water nitrate concentrations that are higher than the source water nitrate concentrations (chromatographic peaking). Chromatographic peaking occurs because competing ions (such as sulphate) displace nitrate ions on the resin, typically when the ion exchange system is operated beyond nitrate breakthrough. The use of a nitrate-selective ion-exchange resin will prevent chromatographic peaking and is strongly recommended for residential nitrate removal. If a nitrate-selective resin is not available, homeowners whose source water contains sulphate should consider the use of an alternative treatment. It is important to routinely monitor the nitrate concentration in the water treated by ion exchange to ensure that the system is effectively removing nitrate and that chromatographic peaking is not occurring.

8.0 Kinetics and metabolism

Kinetic and metabolism studies are complicated by endogenously synthesized nitrite, independent of dietary or drinking water sources.

8.1 Absorption

After ingestion in food or water, nitrate and nitrite are rapidly and almost completely absorbed in the small intestine of humans and transferred to blood (bioavailability at least 92%); less than 2% of dietary nitrate intake reaches the terminal ileum (Mensinga et al., 2003). The fasting plasma nitrate concentration is between 0.25 and 2.7 mg/L (L'hirondel and L'hirondel, 2002). After human oral exposure to sodium nitrate at 470 µmol/kg bw, the plasma nitrate levels rise rapidly, within 5 minutes, reaching a maximum after about 40 minutes (Cortas and Wakid, 1991). The return to the pre-exposure plasma nitrate concentration is independent of the ingested quantity of nitrate; it may be achieved after 24-48 hours (L'hirondel and L'hirondel, 2002).

In a randomized open, three-way crossover study, seven women and two men received a single oral dose of 0.06 or 0.12 mmol of sodium nitrite per millimole haemoglobin and after 7 days received 0.12 mmol of sodium nitrite per millimole haemoglobin intravenously (Kortboyer et al., 1997). Gastrointestinal absorption was rapid, with peak plasma concentrations observed 15-30 minutes after dosing. Under fasting conditions, 90-95% of the oral dose of sodium nitrite was absorbed in the gastrointestinal tract. The bioavailability of sodium nitrite was 73-110% after the lower oral dose. Before absorption takes place, extensive nitrite metabolism in the gastrointestinal tract may result in a large proportion of nitrite being transformed to other nitrogen-containing species (see Section 9.3).

Nitrates can be absorbed by inhalation (e.g., from cigarette smoke and car exhausts). However, in quantitative terms, absorption through the oral route is of greater importance (Lundberg et al., 2004). No information is available on dermal absorption of nitrate or nitrite.

8.2 Distribution

Absorbed nitrate is rapidly transported via the blood and selectively recirculated by the salivary glands. In humans, peak nitrate levels in serum, saliva and urine are achieved within 1-3 hours, with less than 1% reaching faeces (Bartholomew and Hill, 1984). In humans and most laboratory animals, except the rat, plasma nitrate is selectively and dose-dependently secreted by the salivary gland via an active transport mechanism shared with iodide and thiocyanate, increasing nitrate concentrations up to 10 times that in plasma; approximately 25% of ingested nitrate is recirculated into saliva and secreted via this mechanism (Spiegelhalder et al., 1976; Walker, 1996; Lundberg et al., 2004). The rise in salivary nitrate concentration is fast. The onset can be seen as early as 10 minutes after ingestion, reaching a maximum concentration between 20 and 180 minutes after ingestion; conversely, the decrease in salivary nitrate concentration is slow and may be terminated only after 24-48 hours (L'hirondel and L'hirondel, 2002).

After intravenous injection of labelled nitrate to a healthy volunteer, labelled nitrate was distributed throughout the body and accumulated linearly with time in the abdomen, supporting entero-salivary recirculation of nitrate/nitrite due to swallowing (Witter et al., 1979). In mice and rabbits, intravenous or intratracheal injection of labelled nitrite resulted in a homogeneous distribution of labelled nitrite to numerous organs, including liver, kidneys and bladder (Parks et al., 1981).

Plasma nitrite levels are normally lower than nitrate levels due to lower exposure and rapid reoxidation of nitrite to nitrate by oxygenated haemoglobin in the blood (Parks et al., 1981; Walker, 1996; Lundberg et al., 2004). In dogs and rats, nitrite is almost absent, except in saliva (Fritsch et al., 1985; Cortas and Wakid, 1991).

In human breast milk, levels of nitrate rise rapidly after parturition, peaking on days 2-5 postpartum and at concentrations higher than those in plasma (L'hirondel and L'hirondel, 2002). However, 1 hour after ingestion of a nitrate-containing meal, nitrate concentrations in human and canine milk rose, but did not exceed plasma levels (Green et al., 1982). There is evidence of placental transfer of nitrite to the foetus. Nitrite was found in foetal blood following maternal sodium nitrite dosing, oral or injected, of rats (2.5-5.0 mg/kg bw/day, equivalent to 1.7-3.3 mg nitrite/kg bw/day; Shuval and Gruener, 1972) and mice (0.5 mg/kg bw/day, equivalent to 0.3 mg nitrite/kg bw/day; Globus and Samuel, 1978). In men with normal semen, nitrate and nitrite concentrations were significantly higher in semen than in plasma, supporting a role for nitric oxide in sperm function (L'hirondel and L'hirondel, 2002).

8.3 Metabolism

Of the approximately 25% of exogenous nitrate actively recirculated by the salivary ducts, about 20% (representing 5-8% of ingested nitrate exposure) of it is reduced by oral bacteria to nitrite (as reviewed in Walker, 1996; Mensinga et al., 2003). This nitrite, formed by the reduction of nitrate, represents approximately 80% of total exposure to nitrite, the remainder coming directly from exogenous sources. Swallowing saliva exposes the stomach to the nitrite formed in the oral cavity. Microbial conversion of nitrate to nitrite is influenced by bacterial infection, nutritional status and age (Eisenbrand et al., 1980; Forman et al., 1985). In a healthy fasting adult stomach, a pH of about 1-3 is virtually sterile; hence, it is considered to be low for microbial growth and consequently low for microbial conversion of nitrate to nitrite (Mensinga et al., 2003). However, human variability in gastric pH (e.g., in hypochlorhydric patients) results in numerous individuals with a pH greater than 5, resulting in microbial growth and subsequently nitrate reduction to nitrite (Ruddell et al., 1976). In addition, antacids or other medication can decrease gastric acidity, consequently increasing the susceptibility to reduction of nitrate to nitrite (L'hirondel and L'hirondel, 2002) (see Section 9.3 for more information on endogenous nitrite formation). Further, the acidic environment of the stomach reduces salivary nitrite to nitrous acid and subsequently to nitrogen oxides, including nitric oxide (Lundberg et al., 2004, 2008).

In addition to nitrate reduction to nitrite, nitrite is reoxidized via a coupled reaction with oxyhaemoglobin, producing methaemoglobin and nitrate; nitrite appears in a dynamic equilibrium with nitrate, with nitrate being the normal state (Walker, 1999; Lundberg et al., 2004). Lastly, nitrate reductase results in the reduction of nitrite to ammonia (Lundberg et al., 2004, 2008).

8.3.1 Endogenous nitrate formation

There is an endogenous synthesis of nitrate, which amounts in normal healthy humans to an average of 1 mmol/day, corresponding to 62 mg of nitrate per day or 14 mg of nitrate-nitrogen per day (Mensinga et al., 2003; WHO, 2007). In mammals, the primary pathway of endogenous nitrate formation is the L-arginine-nitric oxide synthase (NOS) pathway, which is constitutively active in numerous cell types throughout the body. The amino acid L-arginine is converted by nitric oxide synthetase to nitric oxide and citrulline, followed by oxidation of the nitric oxide to nitrous anhydride and then reaction of nitrous anhydride with water to yield nitrite. Nitrite is rapidly oxidized to nitrate through reaction with hemoglobin (Addiscott and Benjamin, 2004; WHO, 2007; EFSA, 2008; IARC, 2010). Thus, when nitrate intake is low and there are no additional exogenous sources, such as during gastrointestinal infections, endogenous production is more important than exogenous sources (Mensinga et al., 2003).

8.4 Excretion

Nitrate is found in all body fluids, whereas nitrite concentrations in the body are low, as nitrite is readily oxidized to nitrate. After oral exposure, nitrate is found at highest concentrations in urine, but also in milk, gastric fluid, endotracheal secretion, saliva and sweat (L'hirondel and L'hirondel, 2002). In humans, independent of dose, approximately 65-70% of orally administered nitrate is rapidly excreted in urine and less than 1% is excreted in faeces; the remainder is excreted in sweat or is degraded in saliva or digestive secretions by bacteria. Excretion is maximal about 5 hours post-exposure and is essentially complete after 24 hours (Bartholomew and Hill, 1984). In infants under normal conditions, approximately 100% of nitrate is excreted in urine (Turek et al., 1980). Excretion follows first-order kinetics, and the elimination half-life is approximately 5 hours (Green et al., 1982).

Nitrate (approximately 25%) is actively transported by the sodium/iodide symporter (NIS) to saliva and breast milk; about 3% of nitrate also appears in urine as urea and ammonia in humans (Wagner et al., 1983; Walker, 1999). Mean nitrate clearance is estimated to be 25.8 mL/min (Cortas and Wakid, 1991).

The average plasma half-life of nitrite is 30 minutes in humans and less than an hour in most species; consequently, nitrite is not normally detected in body tissues and fluids after oral administration (Kortboyer et al., 1997). Elimination of nitrite from the stomach occurs through two competing pathways: absorption and reaction with amines, causing formation of nitrosamines (see Section 9.3 for details).

8.5 Physiologically-based pharmacokinetic models

Because nitrite is formed endogenously from nitrate in humans and is more potent than nitrate in terms of methaemoglobinemia, the amount of nitrite formed from ingested nitrate is important for the risk assessment of human exposure to nitrate. A model of the toxicokinetics of nitrate and nitrite was built to incorporate the uptake of nitrate from food and water, endogenous synthesis of nitrate, secretion of nitrate from blood into saliva, conversion of nitrate to nitrite by bacteria present in saliva and reconversion of nitrite to nitrate in blood (Zeilmaker et al., 1996). The model was validated with toxicokinetic data from volunteers (Wagner et al., 1983). The model estimates that: 1) the average adult synthesizes 120 mg of nitrate per day; 2) 32-60% of the oral dose of nitrate is secreted from blood into saliva; 3) 13-22% of salivary nitrate is converted into nitrite; and 4) 7-9% of nitrate is converted to nitrite in humans. The model further estimates that the average adult forms 0.27-0.36 mg of nitrite per kilogram body weight per day after single and repeated (once every 24 hours) doses of nitrate and that 31-41% originates from endogenously synthesized nitrate.

• Previous Page
• Table of Contents
• Next Page