Guidelines for Canadian Drinking Water Quality: Guideline Technical Document - Haloacetic Acids

5.0 Exposure

Available data suggest that drinking water may be a significant source of exposure to HAAs, but there are few data available to determine the exposure from other media, such as food and air. Since HAAs are neither volatile nor absorbed significantly through the skin, exposure via dermal and inhalation routes is considered negligible.

5.1 Water

In general, levels of HAAs are highest in treated water from sources with high organic matter content, such as rivers and lakes, and low when the source water is groundwater. Within a single distribution system, however, HAA levels can vary greatly, depending on both water quality (e.g., HAA precursors, pH, temperature, ammonia and carbonate alkalinity) and treatment conditions (e.g., disinfectant dose, contact time, removal of NOM before the point of disinfectant application, prior addition of disinfectant) (Nikolaou et al., 1999; CDBP Task Group, 2000; IPCS, 2000).

Health Canada has conducted a series of studies to characterize the presence of CDBPs, including HAAs, in drinking water from treatment plants of various sizes using surface water and groundwater and using different disinfection processes. A 1995 survey investigated 53 sites, covering nine Canadian provinces, to determine the concentrations of HAA5 in drinking water for larger communities (10 000-100 000 people). Treatment plants in the study used one of three disinfection processes: chlorine-chlorine (n = 35), chlorine-chloramine (n = 10) and ozone- chloramine (n = 7). Samples were collected during winter and summer months for raw water, treatment plant water (after final disinfection) and treated water from the distribution system (5-10 km from the treatment plant) (Health Canada, 1995).

All HAAs were non-detectable (<0.01 µg/L) or at very low levels in raw water. DCA and TCA were the major HAAs present in treatment plants and distribution systems (winter and summer) for all treatment processes; concentrations ranged from 0.2 to 163.3 µg/L and from <0.1 to 473.1 µg/L, respectively. MCA, MBA and DBA were detected at concentrations ranging from 0.03 to 9.7 µg/L, from <0.01 to 9.2 µg/L and from <0.01 to 1.9 µg/L, respectively. Most treatment plants and distribution systems had DCA concentrations below 50 µg/L. Generally, mean DCA concentrations in treatment plants and distribution systems were higher for the chlorine-chlorine disinfection process, and mean summer concentrations in treatment plants and distribution systems were slightly higher than those in winter. Most treatment plants and distribution systems had TCA levels below 50 µg/L, although a few facilities using the chlorine-chlorine process had relatively high values (>100 µg/L). As with DCA, mean TCA concentrations were higher in summer than in winter for all processes. A comparison of TCA concentrations for the chlorine-chlorine process for both seasons indicated that there was a marked increase going from the treatment plant to the distribution system (Health Canada, 1995).

Table 3 shows the results of another Health Canada study (Aranda-Rodriguez et al., 2002; Health Canada, 2003), in which CDBPs (including HAA5) were surveyed in treated drinking water from small systems located in 27 communities (<10 000 people) within nine provinces. Sixteen of the 27 systems used chlorine only, while the balance used chlorine combined with flocculation and filtration processes. A majority of the locations (n = 23) used surface water, whereas two used groundwater and two used a combination of surface water and groundwater. Samples were collected in the warm water season (August to September 1999) and cold water season (January to March 2000) from five sites at each location: raw water, treatment plant (T) and within the distribution system at 0.1-6 km (D1, close to treatment facility), 0.75-16 km (D2, midpoint of system) and 1-23 km (D3, far system site).

No CDBPs were detected in raw water samples. In the treated water, THMs and HAAs accounted for 80% of the CDBPs. DCA and TCA were the most prevalent HAAs, and their concentrations for all locations ranged from <0.3 to 231 µg/L and from <0.1 to 257 µg/L, respectively. Concentrations of MCA, MBA and DBA ranged from <0.3 to 17.4 µg/L, from <0.4 to 18 µg/L and from <0.1 to 4.6 µg/L, respectively.

DCA and TCA concentrations in summer for small treatment plants and distribution systems significantly exceeded those in winter, whereas MCA concentrations in summer slightly exceeded those in winter. In summer, mean concentrations of MCA, DCA and TCA peaked in the treatment plant (T), at the D1 site and at the D2 site, respectively, indicating different formation-degradation patterns for these compounds in warm water conditions (Table 3). In winter, mean concentrations of MCA, DCA and TCA all peaked at the D2 site. Concentrations of MBA and DBA were relatively constant, regardless of the site or season.

Average DCA levels at D2 (midway in the distribution system) during summer (57.4 µg/L, Table 3) were higher than the average system concentrations of DCA in larger facilities (19.0 µg/L, chlorine-chlorine). A similar comparison for the cold water season revealed that average DCA concentrations were higher in the small systems (41.5 µg/L, Table 3) than in the larger systems (15.6 µg/L, chlorine-chlorine). Generally, a greater fraction of small systems had DCA values above 50 µg/L, and concentrations tended to increase in the distribution system after treatment (Table 3), whereas DCA concentrations appeared to level off to a greater extent in the distribution systems of larger facilities (chlorine-chlorine).

Table 3: HAA concentrations in small distribution systems (Health Canada, 2003)

		HAA concentrations (µg/L)
		Summer		Winter
Compound	Site	Mean	Range	Mean	Range
MCA	T	3.7	<0.3-17.4	1.6	<0.3-9.2
	D1	3.6	<0.3-16.6	2.0	<0.3-5.2
	D2	3.5	<0.3-12.4	2.4	<0.3-10.1
	D3	2.7	<0.3-8.1	1.8	<0.3-5.7
DCA	T	55.1	0.8-227	26.3	1.8-180
	D1	59.6	0.4-231	31.8	0.5-109
	D2	57.4	<0.3-195	41.5	0.5-188
	D3	43.0	<0.3-134	29.9	0.5-85.2
TCA	T	43.3	0.3-246	22.4	0.4-179
	D1	60.1	<0.1-257	32.8	<0.1-119
	D2	65.9	<0.1-198	42.6	<0.1-192
	D3	56.3	<0.1-230	34.4	<0.1-125
MBA	T	<0.4	<0.4-18	<0.4	<0.4
	D1	<0.4	<0.4-1.5	<0.4	<0.4
	D2	<0.4	<0.4-1.7	<0.4	<0.4-0.5
	D3	<0.4	<0.4-2.2	<0.4	<0.4-0.6
DBA	T	0.4	<0.1-2.8	0.3	<0.1-3.6
	D1	0.3	<0.1-3.4	0.4	<0.1-3.3
	D2	0.5	<0.1-4.6	0.5	<0.1-4.2
	D3	0.3	<0.1-3.7	0.4	<0.1-3.4

Average TCA levels were always higher in the distribution system than in the treatment plant, regardless of system size or season. A comparison of mean TCA values during summer in the distribution systems of small and large facilities indicated that smaller facilities (65.9 µg/L, D2, Table 3) had higher concentrations than large facilities (48.9 µg/L, chlorine-chlorine). In wintertime, there were higher concentrations in the distribution system of larger facilities (56.7 µg/L, chlorine-chlorine) than in the small facilities (42.6 µg/L, D2, Table 3).

The above Health Canada studies indicated that of the five HAAs, DCA and TCA were present in the highest concentrations. DCA and TCA levels ranged from <0.3 to 231 µg/L and from <0.1 to 473 µg/L, respectively, for both studies. Generally, concentrations of both compounds peaked in the distribution system (chlorine treatment) and decreased in the extremities of the system, were higher in summer than in winter and were higher in smaller facilities than in larger ones. Frequently, DCA peaked before TCA, indicating that the former may have a faster rate of formation and degradation. The remaining HAA5 compounds, MCA, MBA and DBA, were found at concentrations ranging from <0.01 to 18 µg/L. A comparison of HAA5 levels for the different disinfection processes indicated that levels were generally higher for plants using chlorine. Since HAA5 concentrations vary within and between distribution systems, depending on different factors, including water quality characteristics (e.g., HAA precursors, pH, season, temperature) and treatment conditions (e.g., disinfectant type, disinfectant dose, contact time), it is recommended that monitoring samples be taken at the water treatment plant and at points in the distribution system where historical data show the highest HAA concentrations.

The spatial variation in HAA5 concentrations in the distribution systems noted in these studies may be explained in part by differences between disinfectant residuals (chlorine versus chloramine) and the susceptibility of individual HAAs to microbial biodegradation. A U.S study (Williams et al., 1998) reported unexpectedly low HAA concentrations at the maximum residence time locations in distribution systems. Analysis of water quality parameters revealed that water at the maximum residence time locations had low levels of free chlorine and high heterotrophic plate counts. Others have previously identified specific bacteria and haloacid dehalogenase as being capable of degrading DCA (Uchiyama et al., 1992; Meusel and Rehm, 1993). Research on haloacid dehalogenase has shown it to have some degree of substrate selectivity, where MCA, DCA, MBA and DBA were degraded while TCA was not (Ploeg et al., 1991). Another factor that may affect the spatial variation of HAA5 in distribution systems is the pH. The rate of formation of TCA is significantly favoured by low pHs (<pH 7), whereas the rate of DCA formation is only slightly higher at low pHs (Miller and Uden, 1983).

Data from 193 communities in Newfoundland and Labrador for the period 1999-2003 indicated that DCA and TCA were the main HAAs present in treated distributed water. Samples had the following concentrations of HAA5: TCA, <1-600 µg/L (average 66.2 µg/L); DCA, <1-499 µg/L (average 50.2 µg/L); MCA, <1-15 µg/L (average 1.1 µg/L); DBA, <1-13 µg/L (average 0.4 µg/L); and MBA, <1-4 µg/L (average 0.1 µg/L). Total HAA5 concentrations for all communities ranged from <1 to 1114 µg/L and averaged approximately 111 µg/L (Newfoundland and Labrador Department of Environment, 2003).

Monitoring data (1999-2003) for 178 Ontario communities similarly indicated that DCA and TCA were the main HAAs present in treated distributed water. Concentration ranges for HAA5 compounds were as follows: TCA, <0.05-141 µg/L; DCA, 0.2-95.9 µg/L; MCA, 0.5-30.5 µg/L; MBA, 0.05-26.6 µg/L; and DBA, 0.05-17.0 µg/L. Total average HAA5 concentrations (based on individual averages for each compound) for all communities ranged from approximately 1.2 to 142.8 µg/L (Ontario Ministry of Environment and Energy, 2003).

Monitoring data from 37 communities in Manitoba for 2000 indicated that DCA, MCA and TCA were the main HAAs present in treated plant water. Samples (n = 47) had the following concentrations: DCA, <0.5-210 µg/L (average 63 µg/L); MCA, <1-51 µg/L (average 7.9 µg/L); TCA, <0.5-35 µg/L (average 6.7 µg/L); DBA, <0.5-5.4 µg/L (average 0.9 µg/L); and MBA, <0.5-3.1 µg/L (average 0.9 µg/L). Total HAA5 concentrations for all communities ranged from 2.5 to 268 µg/L and averaged approximately 80 µg/L (Manitoba Department of Conservation, 2004).

Non-ingestion exposure (dermal and inhalation) to HAAs via showering and bathing was found to be insignificant, because these compounds are neither volatile nor lipophilic (Xu et al., 2002; Xu and Weisel, 2003).

5.1.1 Analysis of HAA5 data

In order to obtain a better understanding of how average HAA5 data for surface water varied in magnitude and if there were any major differences according to community size, an analysis of provincial and territorial monitoring data from 1990 to 2004* was carried out. Average HAA5 values for each location were calculated based on data (n = 1-24) provided for seasonal quarters (January-March, April-June, July-September, October-December) for the period 1999-2004*. The data used were not necessarily quarterly averages because of data scarcity; some locations had data for only one season, and others had data for many seasons. The location of sampling between sites also varied.

5.1.1.1 Communities with >5000 persons

Health Canada received HAA data from 135 water treatment plants (distribution systems) serving communities of greater than 5000 persons, representing a total population of approximately 19.3 million. The majority of these systems were located in Ontario, Quebec, Nova Scotia and Newfoundland and Labrador, with a few plants located in Alberta, British Columbia, Manitoba, Saskatchewan and the Yukon. Of these, 88% had average HAA5 concentrations below 80 µg/L, while 12% exceeded this level (Table 4). On average, DCA accounted for 46% of the total concentration of HAA5. However, for a significant percentage of systems (26%), DCA accounted for 50-59% of HAA5.

Table 4: Average total HAA concentrations above and below 80 µg/L in distribution systems serving communities with >5000 people, by province/territory

Province/ territory	No. of systems per province/ territory	No. of systems below 80 µg/ L HAA5	No. of systems above 80 µg/ L HAA5
Alberta	4	4	0
British Columbia	5	5	0
Manitoba	1	0	1
Newfoundland and Labrador	10	5	5
Nova Scotia	13	11	2
Ontario	74	71	3
Quebec	27	23	4
Saskatchewan	1	1	0
Yukon	1	1	0
Total no. of systems	135	119	16
%		88	12

5.1.1.2 Communities with <5000 persons

Health Canada received HAA data from 312 systems serving communities of fewer than 5000 persons, representing a total population of approximately 333 300. The systems were located in Ontario, Quebec, Nova Scotia and Newfoundland and Labrador. Of these, 56% of systems had average HAA5 concentrations below 80 µg/L, whereas 44% of systems exceeded this level. On average, DCA accounted for 42% of the HAA5. However, for a significant percentage of water treatment plants (25%), DCA accounted for 50-59% of HAA5.

Table 5: Average total HAA concentrations above and below 80 µg/L in distribution systems serving communities with <5000 people, by province

Province/ territory	No. of systems per province	No. of systems below 80 µg/ L HAA5	No. of systems above 80 µg/ L HAA5
Newfoundland and Labrador	220	108	112
Nova Scotia	38	22	16
Ontario	32	31	1
Quebec	27	16	11
Total no. of systems	312	174	138
%		56	44

5.2 Air

Stack gases of municipal waste incinerators have been reported to contain 0.37-3.7 µg TCA/m³ and 3.2-7.8 µg MCA/m³ (Mowrer and Nordin, 1987).

Air samples taken in rural Scotland and the Netherlands contained DCA and TCA concentrations of #0.0007 µg/m³ (Heal et al., 2003; Peters, 2003), whereas atmospheric particulate measurements of MCA, DCA and TCA from Athens, Greece, ranged from 0.01 to 2.01 ng/m³, from 0.0006 to 0.46 ng/m³ and from 0.0009 to 0.125 ng/m³, respectively (Bakeas et al., 2003).

The detection of chlorinated acetic acids in rain water is an indication of their presence in the atmosphere. Reimann et al., (1996) reported the following levels in rainwater: MCA, 0.05- 9 µg/L; DCA, 0.05-4 µg/L; and TCA, 0.01-1 µg/L. Rainwater in Germany contained 1.35 µg DCA/L and 0.1-20 µg TCA/L (IARC, 1995). Sidebottom and Franklin (1996) reported that TCA concentrations in rainwater in remote areas (Antarctic, Arctic and sub-Arctic regions) generally ranged from 10 to 100 ng/L.

No information was available on exposure to MBA or DBA in air (U.S. EPA, 2003b). No Canadian data were available.

5.3 Food

It is speculated that MCA, DCA and TCA may be found in meat and other food products

(U.S. EPA, 2003b). This would result from the use of chlorine in food production and processing, including the disinfection of chicken; processing of seafoods, poultry and red meats; sanitizing of equipment and containers; and oxidizing and bleaching in the flour industry (U.S. EPA, 1994).

MCA and TCA can be taken up from cooking water (Raymer et al., 2001). In addition, there is evidence that TCA may be taken up by plants via the roots or by leaves via uptake from the air (Schroll et al., 1994; Sutinen et al., 1995). TCA was found in food (vegetables and fruits) at concentrations of 0.01-0.19 mg/kg following irrigation (Demint et al., 1975). Reimann et al. (1996) examined the concentrations of MCA, DCA and TCA in a limited number of samples of several vegetables, fruits, grain and beer. MCA concentrations ranged from <0.7 to 5.3 µg/kg in vegetables, from 1.7 to 13.2 µg/kg in grains, from 2.3 to 11.8 µg/kg in flours/breads and from 0.2 to 2.6 µg/L in beer. DCA concentrations ranged from <0.9 to 3.5 µg/kg in vegetables, from <0.6 to 11.1 µg/kg in grains, from 0.8 to 19.8 µg/kg in flours/breads and from 1.5 to 15.2 µg/L in beer. TCA concentrations ranged from <0.2 to 5.9 µg/kg in vegetables and from <1.6 to 4.1 µg/kg in grains. TCA was below the detection limit of 1.5 µg/kg in breads and was not analysed in the flours or beer. None of these compounds was detected in fruits or tomatoes.

No information on concentrations of MBA or DBA in food was located (U.S. EPA, 2003b).

5.4 Contribution of drinking water to total exposure

The data for HAA5 in water suggest that drinking water has the potential to be a significant source of these compounds. While data for air and food indicate that these media are also potential sources of HAA5, the data are insufficient to quantify their relative contributions with reasonable certainty. On this basis, a default value of 20% can be used to describe the contribution of drinking water to total daily intake. The U.S. Environmental Protection Agency (U.S. EPA, 2003b) came to a similar conclusion with respect to lack of data for air and food and selected a default relative source contribution of 20% for MCA and TCA. No relative source contribution value was assigned specifically to DCA and DBA because it is classified as a carcinogen.