Health Canada
www.hc-sc.gc.ca
Common menu bar links
Institutional links
Priority Substances List Assessment Report for Ammonia in the Aquatic Environment
2.0 Summary of Information Critical to Assessment of "Toxic" Under CEPA 1999 (Continued)
2.3 Exposure characterization (continued)
2.3.2 Environmental concentrations
2.3.2.1 Air
Atmospheric levels of gaseous ammonia in urban areas around the world are on average about 20 µg/m3. Japanese researchers found ammonia concentrations up to 210 µg/m3 downwind of a heavily industrialized area of Tokyo. Non-urban sites can have a wide range of levels (0.2-2000 µg/m3), depending on their proximity to point sources (WHO, 1986). One such area in California was studied as a comparison with urban areas. Air near a large 600-animal dairy farm had an ammonia concentration of 560 µg/m3 (Luebs et al., 1973), and air in the region contained 190 µg/m3 on a routine basis. Concentrations of ammonia in the troposphere are heavily influenced by ground temperature and so exhibit strong seasonal variations. German researchers found winter concentrations of 1-2 µg/m3 at 1500 m in winter and 5 µg/m3 at 4000 m in summer (WHO, 1986). Levels of particulate NH4+ ions in the atmosphere above the oceans have been studied; concentrations were found to be between 0.01 and 0.12 µg/m3 (Servant and Delapart, 1983; Quinn et al., 1988). The authors concluded that the oceans are a source of ammonia for the atmosphere. The general background concentration of ammonium in particulates is around 1 µg/m3, with a measured average for an American urban area of 7.6 µg/m3. In general, atmospheric ammonia levels show a seasonal variation, with the lowest levels occurring during the summer and the highest during the winter in Europe, and a reversed pattern in Japan (WHO, 1986; Yamamoto et al., 1995). For the 1970s, rainfall in the continental United States had concentrations of ammonia ranging from 0.01 to 0.15 mg/L (NRC, 1979). In a mixed coniferous stand at Whitaker Forest in the western Sierra Nevada Mountains, California (1988-1990), seasonal ammonia 12-hour daytime averages of 1.11-1.56 µg/m3 were recorded, with the highest 12-hour daily averages reaching 3.75 µg/m3. When expressed on a molar basis, NH3 was the most abundant nitrogen air pollutant and represented almost 50% of the total nitrogen (Bytnerowicz and Riechers, 1995). In Edmonton, Alberta, the average concentration in rainfall was 0.41 mg NH4+/L during the summers of 1977 and 1978. In July 1978, an intensive sampling effort detected a distinct gradient of total ammonia, with low values (<0.06 mg/L) found in the foothills of the Rocky Mountains and high values (>0.4 mg/L) eastwards into the agricultural areas of the province (Klemm and Gray, 1982).
Table 3 provides some atmospheric concentrations of ammonia and ammonium from regions with different sources. There is considerable variation in atmospheric concentrations of total ammonia, even in unpolluted regions; however, agriculturally polluted regions, particularly in Europe and California, can have very high concentrations (up to 4000 µg/m3).
At the Hubbard Brook Experimental Watershed in New Hampshire, Fisher et al. (1968) detected 0.18-0.22 mg NH4+/L for the years 1965-1968.
Janzen et al. (1997) collected precipitation around Lethbridge, Alberta, and analysed it for ammonia and nitrate in an attempt to determine the nitrogen cycling in local soils. Analysis of precipitation collected over a 2-year period suggested that the annual nitrogen input, as nitrate and ammonium, amounted to 5.6 kg N/ha. The authors cited Peake and Wong (1992) to provide an average ratio of 19 µeq NH4+ to 14 µeq NO3- for rainfall around Lethbridge. Using this ratio with a mean total nitrogen input of 5.6 kg N/ha gives a mean annual input of 3.2 kg NH3/ha.
The Canadian Acid Aerosol Measurement Program was established to gain an understanding of the atmospheric behaviour of particulate acidity, which involved the measurement of gaseous ammonia (Brook et al., 1997). The mean concentrations are assumed to be representative of typical ammonia levels in the early to mid-1990s. Mean concentrations of 1.72 and 4.28 mg/m3 were observed at sites with industrialization and human populations (Windsor and Hamilton, Ontario, respectively). Intense agricultural activity also produced elevated mean concentrations of 1.63 mg/m3 (Egbert, Ontario). In non-industrialized rural settings, mean concentrations were lower, at 0.83 mg/m3 (Sutton, Quebec); over water, concentrations were 0.41 mg/m3 (Kejimkujik National Park, Nova Scotia).
2.3.2.2 Atmospheric deposition
Researchers at the Lethbridge Agricultural Research Station (Janzen et al., 1997) measured the rate of NH3 deposition to soil at nine sites throughout southern Alberta for up to 2 years.
Lowest average rates, typically about 4-6 kg N/ha per year, were observed at the two control sites at the Research Station, while the highest average rates (about 66 kg N/ha per year) were observed near a beef feedlot. The high rates are enough to significantly affect soil nitrogen fertility. The researchers also studied the relationship between distance from NH3 source and the rate of NH3 deposition. Soil collectors were set up at various distances downwind of the feedlot; deposition rates were highest close to the feedlot and then diminished with distance. Average background deposition rates (4.4 mg N/m2 per day) were not reached within 1 km (Figure 4).
Figure 4 Dispersion of NH3 downwind of a feedlot
This is similar to atmospheric dispersion of ammonia from other areas of the world where intensive livestock facilities exist. The air above a large dairy area in Chino, California, and from an area not close to known ammonia sources was sampled for ammonia. The livestock area was 150 km2 containing
143 000 dairy cows on 380 dairies (Luebs et al., 1973). Continuous simultaneous sampling of the air at the dairy area site and at the control site showed the nitrogen concentration to be 23 times greater within the dairy area, with concentrations of 80 µg/m3 at the dairy area site compared with 3-5 µg/m3 at the control site. Rainfall measurements showed that the rain over the dairy area contained roughly 3 times more distillable nitrogen than the control area. The ammonia concentrations ranged from 0.4 to 1.7 mg/L in the dairy area compared with a range from 0.2 to 0.6 mg/L in the control area. The rainfall in the dairy area added 1.59 kg N/ha to soils compared with 0.53 kg N/ha in the control area. At the fence line of the dairy area, concentrations of distillable nitrogen were 540 µg/m3; at 200 m, NH3 concentrations were roughly 50 µg/m3; and at 800 m downwind, the concentrations were 18 µg/m3 (Luebs et al., 1973).
In one U.S. study, a lake 2 km from a large cattle feedlot (90 000 head of cattle) was found to receive considerable quantities of ammonia from the air, sufficient to raise its total nitrogen concentration by 0.6 mg/L over a year. On average, the differences in atmospheric concentrations of NH3 between background sites and those closest to the feedlot (400 m) were 20-fold. The average deposition of ammonia closest to the feedlot was 145.6 kg NH3/ha per year, while at the background site it was 7.8 kg/ha per year (Hutchinson and Viets, 1969).
Ammonia flux density was determined above a large feedlot to be on average 1.4 ± 0.7 kg N/ha per hour in spring and summer in northeast Colorado. A feedlot surface had lower average values than this when wet, but higher values than this during drying. Total NH3 emissions equalled about half the rate of urinary nitrogen deposition, or about one-quarter of the rate of total nitrogen deposition (Hutchinson et al., 1982). Actual ammonia concentrations were fairly stable, being 361 ± 46 µg NH3 -N/m3. Drying events and periods of warm, calm weather generated much higher NH3 concentrations (970-1200 µg NH3 -N/m3).
Deposition of up to 66.4 kg NH3/ha per year was determined within 50 m of a poultry house containing 8000-12 000 chickens near Athens, Georgia. At 1.2 km from the poultry house, the ammonia trapped was at background deposition rates (15 kg/ha per year). Near a beef cattle feedlot, 26.5 kg/ha per year was trapped. At distances greater than 500--800 m, the concentrations dropped to background for the Athens area (Giddens, 1975).
Figure 5 Ammonia air emissions inventory for the Lower Fraser Valley, 1996-1997
Nitrogenous air pollutants were monitored during three summer seasons (1988-1990) in a mixed coniferous stand at Whitaker Forest in the western Sierra Nevada Mountains, California. NH4+ deposition fluxes to ponderosa pine (Pinus ponderosa Dougl. ex Laws.) branches during the three summer seasons ranged from 17 to 67 kg/m2 per year. During the 1990 summer season, NH4+ washed from branch surfaces provided 0.2 kg/ha per year. The estimated internal uptake of NH3 was 0.6 kg N/ha per year. The elevated levels of air pollutants and nitrogen deposition could adversely affect the natural ecosystems of the western Sierra Nevadas (Bytnerowicz and Riechers, 1995).
Total inorganic nitrogen deposition in the most highly exposed forests in the Los Angeles Air Basin may be as high as 25-45 kg/ha per year. Nitrogen deposition in these highly exposed areas has led to nitrogen saturation of chaparral and mixed conifer stands. In nitrogen-saturated forests, high concentrations of nitrate are found in stream water, soil solution and foliage (Bytnerowicz and Fenn, 1996).
2.3.2.2.1 Case study: The Lower Fraser Valley
In the Lower Fraser Valley, deterioration of surface water, groundwater and air quality is a major environmental issue. There is particular concern about the potential for nitrate pollution of the Abbotsford/Sumas aquifer, which supplies both Canadian and U.S. drinking water in the area (Zebarth et al., 1997), as well as a decline in the visible air quality around Vancouver (Hoff et al., 1997; Pryor et al., 1997a,b,c). The source of the nitrates is manure applied to fields in winter, and the cause of the reduced air quality is ammonium sulphate particulates. The valley has many farms and livestock facilities that contribute to both direct volatilization of NH3 and local redeposition (Paul, 1997).
In order to conduct an ammonia modelling project, an NH3 air emissions inventory was constructed using the latest census data available (1996) from both Canadian and U.S. sources (Jennejohn et al., 1996; Barthelmie and Pryor, 1998). The ammonia emissions inventory for the Lower Fraser Valley in 1996-1997 is shown in Figure 5. Agriculture dominates NH3 emissions in the Lower Fraser Valley, with an estimated generation of 5260 tonnes/year, while within the Greater Vancouver Regional District 3511 tonnes/year are generated. Cattle contribute approximately half of the agricultural NH3 emissions, poultry are the next major source and mineral fertilizer use contributes significantly (Barthelmie and Pryor, 1998). Approximately 8800 tonnes of ammonia come from the Canadian part of the valley and 2400 tonnes from the U.S. portion.
Figure 6 Modelled total nitrogen deposition (kg/ha per year) for the Lower Fraser Valley
Since ammonia is chemically important to the production of atmospheric aerosols, understanding the concentrations and deposition of ammonia requires modelling the atmospheric chemistry of aerosols. The Inorganic and Secondary Organic Particle (ISOPART) model was used because it is a Lagrangian model with extensive chemistry and aerosol dynamics.
Atmospheric sampling relevant to the current application was undertaken (Hoff et al., 1997). In addition to airborne monitoring, intensive ground-based sampling took place in the summer of 1993 during the Regional Visibility Experimental Assessment in the Lower Fraser Valley (REVEAL) campaign. Aerosol samplers were deployed at seven locations in the Fraser Valley to collect 24-hour averaged fine particulate and gas concentrations (Pryor et al., 1997a,c).
A comparison of modelled near-surface and surface observed aerosol concentrations was performed for the afternoon of August 5, 1993. The areas of highest aerosol concentrations were observed east of Vancouver in the north-central valley in a band running from northwest to southeast. This pattern of aerosol concentrations was also predicted by ISOPART for this period (Pryor et al., 1997b), giving confidence to the aerosol components of the model such that gaseous ammonia concentrations can be derived.
As expected, for NH3 , the highest concentrations were in the central and eastern portions of the valley, in mainly agricultural areas (the source region); the urban area of Vancouver is a source of NOy (total nitrogen oxide compounds, including nitrate). Ammonium aerosol concentrations were high in a band from Greater Vancouver southeast towards the central valley, while nitrate concentrations were highest around Vancouver and in the central valley. It should be noted that the peak in NH4+ concentrations was associated with high sulphate concentrations.
Figure 7 Un-ionized ammonia concentrations upstream and downstream of Edmonton
Enlarge image
Figure 6 shows total modelled nitrogen deposition in kg N/ha per year. For NH3 , the highest deposition is over the central and eastern portions of the valley, shown as the concentration area on the right side of the figure (labelled NH3); this is the source region for ammonia. NOy deposition is primarily in Greater Vancouver and the highly urban portions of the domain, shown as the concentration area on the left side (labelled NO3-). Ammonium aerosol deposition is highest in a band from Greater Vancouver southeast towards the central valley, while nitrate deposition is highest east of Vancouver and over the downtown region. This pattern of nitrogen deposition is to be expected, as it represents two major and separate source regions. Ammonia is rapidly redeposited as a gas, but it is also relatively rapidly converted to the aerosol phase and both transported and deposited as NH4+ . Estimated maximal ammonia deposition was 105 kg NH3/ha per year in the rural portions of the Lower Fraser Valley. Most of the agricultural portion of the valley is subjected to a level of nitrogen deposition (primarily as ammonia) considerably greater than the critical load of 10 kg/ha per year (Figure 6). The grid numbers in Figure 6 correspond to the grid numbers in Figure 5.
2.3.2.3 Surface water
Natural waters typically contain little total ammonia, usually in concentrations below 0.1 mg/L. Assuming temperatures of 20°C (typical of times when risk is highest and which are the focus of the risk scenarios) and pHs in the 7-8 range, natural NH3 levels are in the 0.0004-0.004 mg/L range. Higher concentrations may be an indication of anthropogenic input and organic pollution (CCREM, 1987). This tendency is shown in Figure 7 for waters above and below Edmonton. Above Edmonton, un-ionized ammonia was almost non-detectable (based on the detection limit for total ammonia); at 113 km downstream, un-ionized ammonia ranged as high as 0.026 mg/L and was consistently detected (Tchir, 1998). The highest concentration of un-ionized ammonia in Canadian municipal effluents was 0.68 mg/L, detected in effluents from the Annacis Island facility, Vancouver (Servizi et al., 1978).
Data on ammonia concentrations in fresh waters were collected from federal and provincial monitoring agencies and were examined in order to identify hot spots. Detection limits for unionized ammonia must be calculated from the detection limit for total ammonia at a specific pH and temperature. As these two parameters change, the detection limit for un-ionized ammonia will fluctuate; therefore, non-detectable limits for unionized ammonia are simply noted as "<detectable." A general analysis of water quality was received from the Ontario Ministry of Environment and Energy (OMEE, 1997). The average un-ionized ammonia concentration was 0.007 mg/L, with a median value of 0.0004 mg/L and range of <detectable-5.6 mg/L.
Only total ammonia concentrations were reported from across British Columbia (Ryan, 1998; Swain, 1998). They were generally very low, indicating perhaps the large dilution capacity of the rivers and lakes in the province. For the federal government monitoring sites, the average total ammonia concentration was 0.009 mg/L, with a median of 0.005 mg/L and a range of <0.002-0.48 mg/L. From the provincial monitoring sites, the average total ammonia concentration was 0.02 mg/L, with a median of 0.001 mg/L and a range of <0.002-8.4 mg/L.
The Northwest Territories and Nunavut also had extremely low ammonia concentrations in rivers, as would be expected for those territories. The average total ammonia concentration was 0.03 mg/L, with a median of 0.01 mg/L and a range of <0.002-0.68 mg/L (Halliwell, 1998).
There were 1225 samples from 66 sites for the three Prairie provinces from Environment Canada, primarily from interprovincial river sites in 1994 and 1995. Ammonia, temperature and pH measurements were taken so that un-ionized ammonia concentrations could be calculated (Chu, 1997). The average un-ionized ammonia concentration was 0.002 mg/L, with a median of 0.0006 mg/L and a range of <detectable-0.16 mg/L.
Alberta Environmental Protection provided detailed sampling data, but no temperature or pH values (Tchir, 1998). Many of the Alberta data indicated that cities and major industrial centres are elevating ammonia concentrations in the province's streams. The average total ammonia concentration was 0.23 mg/L, with a median of 0.03 mg/L and a range of <0.002-126 mg/L.
Manitoba Department of the Environment has taken water samples from 44 sites. The average un-ionized ammonia concentration was 0.002 mg/L, with a median of 0.0004 mg/L and a range of <detectable-0.21 mg/L (Williamson, 1998).
The City of Winnipeg supplied water quality monitoring data that it collects below each of its sewage treatment plants and downstream of Winnipeg at Lockport Dam (Ross, 1998). At the Fort Garry Bridge below the South End facility, the average un-ionized ammonia concentration was 0.012 mg/L; the median was 0.006 mg/L, with a range of <detectable-0.13 mg/L. At the Main Street Bridge, where the Assiniboine River joins the Red River, the average un-ionized ammonia concentration was 0.006 mg/L, the median was 0.003 mg/L and the range was <detectable-0.04 mg/L. At the North Perimeter Bridge on the Red River, the average un-ionized ammonia concentration was 0.017 mg/L; the median was 0.007 mg/L, with a range of <detectable-0.17 mg/L. At the Lockport Dam on the Red River, the average un-ionized ammonia concentration was 0.017 mg/L, the median was 0.01 mg/L and the range was <detectable-0.14 mg/L.
Two concentrations are available from the Lake Ontario sampling surveys done in 1992 and 1993. One is 0.96 mg/L, measured in Hamilton Harbour in 1992, and the other is 0.39 mg/L, measured in an area of east Toronto known as The Beaches, also taken in 1992 (Charlton, 1997).
Water from the centre of Hamilton Harbour has been analysed for ammonia since at least 1986; the results show routinely high concentrations of total ammonia, which builds up in the winter and degrades throughout the summer. This ammonia concentration process is the result of three municipal WWTPs depositing their effluents in the harbour, reduced water exchange in the harbour and the reduction in nitrifying bacteria in the winter. Environment Canada undertook a weekly survey in 1998 (January 6 - September 9) to determine the extent of ammonia concentrations throughout the harbour (Charlton and Milne, 1999).
At 1 m depth, the average un-ionized ammonia concentration was 0.023 mg/L; the median was 0.016 mg/L, with a range of 0.001-0.114 mg/L. These values all declined through the water column, so that at 19 m depth the average was 0.004 mg/L, with a median of 0.003 mg/L and a range of <detectable-0.012 mg/L.
The province of Quebec provided water quality data, including total ammonia, temperature and pH, for the years 1988-1998 (Dupont, 1998). Based on data for 16 372 samples, the average un-ionized ammonia concentration was 0.001 mg/L, the median was less than a detectable concentration (based on a total ammonia detectable limit of 0.002 mg/L) and the range was <detectable-0.69 mg/L. Many streams and rivers on the south shore of the St. Lawrence River have very high pH values in summer; this, combined with high summer temperatures, generates high un-ionized ammonia concentrations, even when there are relatively low concentrations of total ammonia. It appears that many of the streams with high average total ammonia concentrations are just north or south of Montréal or east of the Québec area on the south shore.
2.3.2.4 Soil runoff
Timmons and Holt (1977) determined the quantities and chemical composition of runoff from native (undisturbed by humans) prairie soils in Minnesota. Over 5 years, they determined that runoff from snowmelt accounted for 80% of the average annual ammonia in runoff. Rainfall caused appreciable runoff only in 1 year (37%). Dissolved ammonia losses ranged from 0.02 to 0.28 kg NH3 -N/ha in snowmelt, with rainfall- derived runoff containing 0.03 kg/ha in that year. The average loss of ammonia from native land was 0.13 kg/ha.
In a controlled deforestation study, Likens et al. (1970) showed that complete deforestation of a watershed in the eastern forests of New Hampshire had no effect on the runoff of ammonium. In watersheds that were not cut, the concentration of ammonium over 3 years in runoff ranged from 0.02 to 0.12 mg/L; in the cut watershed, the concentration of ammonium in the runoff ranged from 0.05 to 0.14 mg/L.
Data from 32 forested stream catchments in the Muskoka-Haliburton area of central Ontario, collected over 8 years, were used to develop regression models of long-term NH4+ export. There was a weak correlation between stream chemistry (including NH4+) and discharge for any site. Retention (defined as the fraction of annual deposition retained by the catchment) was very high (>0.87) for ammonium in all catchments. Deposition of NH4+ for the area was 4.794 kg NH4+/ha per year (Dillon et al., 1991).
Animal husbandry can significantly elevate the runoff of ammonia from land. Cooke (1996) studied the variations in nitrogen runoff from various land types in Alberta. Under forested land, neither nitrate nor NH4+ concentration was high in surface runoff. Under cropland, nitrate dominated, its concentration approaching 50 times the NH4+ concentration. Under agricultural land with cattle grazing (25-100 head), runoff delivered 95% of NH4+ to streams. Only 2% of the nitrogen in cropland streams was ammonia, 43% of nitrogen in forest streams was ammonia and 89% of nitrogen in streams draining cattle operations was ammonia.
Peak NH4+ concentrations were 27 mg/L below cow-calf operations, while spring concentrations in the forested streams were below 1 mg/L. Flow-weighted mean concentrations of 1-2.3 mg/L for NH4+ and 0.15-0.2 mg/L for nitrate were detected below cow-calf operations in the spring.
A provincial stream survey in Alberta found that nutrient concentrations tend to be higher in streams that drain intensively farmed land than in streams that drain less intensively farmed land. Typical seasonal patterns were apparent: 1) highest concentrations were generally measured during spring runoff, 2) concentrations declined as flows subsided, and 3) later in spring and in summer, increases in nutrient levels (especially particulates) usually coincided with sudden increases in rainfall. These sudden concentration increases were more apparent in streams that drain land farmed with medium and low intensity, because rain-induced runoff occurred in these drainage basins, whereas none occurred in basins with high agricultural intensity (Anderson et al., 1998b).
Application of manure to fields can be a cost-effective means of disposal of animal wastes and a cost-effective fertilizer; however, at some times of the year, application can be problematic for nearby watercourses. In Quebec, Gangbazo et al. (1995) determined that fall application of manure, as a fertilizer, created significant quantities of ammonia in runoff. The fall application of 360 kg manure-N/ha to corn increased ammonia in runoff from 1.9 to 3.4 kg N/ha. The runoff concentrations were elevated for at least 3 years. For surface application to forage, only the fall application of 110 kg manure-N/ha caused excessive ammonia in field runoff. Ammonia was elevated for 2 years over controls.
In Manitoba, Green (1996) studied the spring runoff of ammonia from hog manure surface-applied in the winter. Mean ammonia concentrations were considerably higher in runoff than in field pools. Meltwater from control fields contained 0.19-0.26 mg ammonia/L, while that from manured fields contained 8.5 mg/L. Concentrations of total ammonia in local rivers were relatively high, both upstream (0.32 mg/L) and downstream (0.34-0.52 mg/L) of the study site. There was no apparent impact on local watercourses from the application, despite the fact that substantial quantities of ammonia were leaving the site in runoff.
As part of a eutrophication study in Iowa, Jones et al. (1976) made detailed measurements of the concentrations of nutrients in runoff from 48 small and large watersheds. They also conducted an inventory of the animal densities in the watersheds, the types of animal holding facilities in each and the land use in each watershed. In watersheds of over 100 ha, ammonia in stream water was significantly correlated only with the animal units/ha in the watershed. The researchers determined that NH3 -N was increased by 0.77 ± 0.23 mg/L for each animal unit/ha within the watershed. They also determined whether animal placement within the watersheds influenced NH3 -N losses. The number of feedlot animal units/ha with drainage to streams or tile intakes was the only significant variable in the analyses. Jones et al. (1976) estimated that 0.96 ± 0.18 kg NH3 -N/ha were associated with each feedlot animal unit/ha with drainage to streams or tile intakes. Concentrations of ammonia in feedlot runoff averaged 6.5 mg/L, while runoff from soybean fields, cornfields and pastures was in the 0.75-1.0 mg/L range. There was no ammonia detected in tile runoff from fields.
Intensive dairy operations conducted in close proximity to streams have the potential to contaminate local watercourses with high levels of ammonia, especially if they have steep slopes to drainage. Daniel et al. (1982) showed this with a survey of three dairies in Wisconsin and an urban construction site. The runoff from an intensive dairy operation on a steep slope and in close proximity to a stream contained 5 mg ammonia/L, while runoff from dairies either far removed from streams or on flat land contained around 1 mg ammonia/L. Runoff from the construction site contained around 0.2 mg/L.
2.3.2.5 Soil
There are few data on naturally occurring concentrations of ammonia in Canadian soils. In general, natural ammonia levels in soil are very low (<1 mg/kg) due to the rapid conversion of ammonium to nitrite by Nitrosomonas species and then to nitrate by Nitrobacter species in the temperature range 0-35ºC (Henry, 1995). In some areas of Canada, such as the Lower Fraser Valley, conditions may exist in winter where ammonia can build up in soil due to the application of manure to fields that are not frozen but are too cold for Nitrosomonas species to grow.
2.3.2.6 Groundwater
There are few data on concentrations of ammonia in Canadian groundwater. Ammonia contamination of groundwater is not usually an issue, as it is readily converted to positively charged ammonium ions that bind tightly to negatively charged cation exchange sites in soil. Ammonium is not sufficiently mobile in soil to create widespread groundwater contamination problems (Feth, 1966; Liebhardt et al., 1979; Olson, 1997). In rare instances, nitrogen fertilizers, livestock wastes and septic tanks may contribute significant amounts of ammonia to shallow groundwater, especially those underlying poorly drained soils (Gilliam et al., 1974; Rajagopal, 1978), those underlying feedlots and those in areas of groundwater recharge.
