Priority Substances List Assessment Report for Road Salts

2.0 Summary of Information Critical to Assessment of "Toxic" Under CEPA 1999 (Continued)

2.4.5 Off-site concentrations at patrol yards

A review of select case studies by Morin (2000) indicated that there have been high off-site concentrations of salts at several locations where road salts are stored. A few examples are presented in the following paragraphs.

Arp (2001) examined how the presence of a salt depot can be associated with elevated levels of salinity in surface waters. Sodium concentrations in areas surrounding the depot varied by type (pond and stream water), location and season. Sodium concentrations were highest at and near the salt depot, ranging from 3000 to more than 10 000 mg/L. Runoff and seepage from the depot were quickly diluted by flowing surface water next to the depot, with sodium concentrations decreasing to about 100-300 mg/L. There were also seasonal variations: salt concentrations near the salt depot were lowest during snowmelt and the rainy fall season. During summer, salt concentrations increased with reduced precipitation and increasing air temperature. Highest concentrations occurred during summer months. At this time, the high salt concentration envelope that surrounds the salt depot spread into the adjacent pond and caused elevated salt concentrations in the main drainage stream.

Exposed abrasive piles and salt reserves stored in sheds that offered poor protection from the elements resulted in the contamination of the municipal water supply for Heffley Creek, British Columbia (BC MoTH, 1993). Road salts and unprotected abrasive piles with a 3% salt mixture had been stored at this site since 1973 (AGRA, 1995). (Prior to 1975, concentrations of chloride had been less than 5 mg/L.) In 1993, the B.C. Ministry of Transportation and Highways received an official complaint from the Municipality of Heffley Creek that groundwater for one of its two wells had been contaminated by road salts stored at the Heffley Creek patrol yard. Water quality tests indicated that chloride concentrations at this well were approximately 1000 mg/L. By 1994, there were elevated chloride concentrations at both municipal wells located between 175 and 200 m from the patrol yard (Figure 7). Chloride concentrations eventually peaked at approximately 3200 mg/L in 1995. While chloride concentrations have decreased since 1996, well #1 still had chloride concentrations higher than 500 mg/L in 1998 (AGRA, 1999).

In 1996, complaints were registered with the Nova Scotia Departments of Transportation and Public Works and the Environment regarding salt contamination of residential wells in Bible Hill, Nova Scotia. Tests undertaken at three wells that supplied six duplexes indicated that chloride and sodium concentrations ranged from 130 to 640 mg/L (Rushton, 1999). The Bible Hill patrol yard, which is located relatively close to these residences, was determined to be the source of salt contamination. While road salts are stored in a salt dome, abrasive mixtures have traditionally been stored outside. The Nova Scotia Department of Transportation and Public Works indicated that groundwater contamination was typical of locations where mixed abrasive/salt piles are left uncovered (Rushton, 1999).

Figure 7 Chloride concentrations in Heffley Creek municipal wells (from AGRA, 1999)

A study by Ohno (1990) conducted on wetlands near sand/salt storage sites in Maine showed that concentrations of chloride and sodium at the affected sites were 2-3 orders of magnitude higher than those of control sites. The concentrations at the control sites were 6 mg sodium/L and 8 mg chloride/L, while the concentrations at the affected sites ranged from 16 to 8663 mg sodium/L and from 14 to 12 463 mg chloride/L.

Pinhook Bog, LaPorte County, Indiana, was adversely impacted by contamination from a salt storage pile from the late 1960s to the early 1980s (Wilcox, 1982). Chloride concentrations at control sites in 1980 and 1981 were 5-6 mg/L. This compares with a maximum single daily reading for salt-impacted locations of 1468 mg chloride/L in 1979, 982 mg chloride/L in 1980 and 570 mg chloride/L in 1981. The total chloride inputs to the bog over the 10-year period when salt storage occurred were estimated as follows: 2.3 million kilograms from the salt pile, 0.4 million kilograms from road salting and 0.012 million kilograms from direct precipitation.

2.5 Environmental fate and pathways

Road salts can enter and move through the environment in their salt form or as their dissociated ions. In aquatic systems, chloride salts are present in dissociated form as chloride ions and corresponding cations (sodium, potassium, calcium, magnesium). The circulation of chloride through the hydrological cycle is due mostly to physical rather than chemical processes. The chloride ions pass readily though soil, enter groundwater and eventually drain into surface waters. Because chloride ions are persistent and are entrained in the hydrological cycle, all chloride ions applied to roadways as road salts or released from patrol yards or disposal sites can be expected to be ultimately found in surface water. This section presents information on the fate and transportation of road salts in the environment and their potential pathways.

2.5.1 Fate/transport

Once road salts enter surface waters, they remain in solution unless their concentrations are very high and exceed their solubility, when crystallization and subsequent sedimentation of mineral salts could occur. When water containing road salts percolates through soil, positively charged ions (i.e., sodium, calcium, magnesium, potassium) are attracted to and bond with the inherently negatively charged soil surfaces. The extent of bonding depends on the cation exchange capacity and the number of negatively charged sites in the soil. The sodium ion has a high solubility and may, once dissolved, remain in solution; however, since it is readily adsorbed onto soil particles, it is less probable that it will reach groundwater and surface waters. Nevertheless, in cases with limited adsorption or where extreme leaching from the soil occurs, the sodium ion will follow water pathways and eventually find its way to groundwater and surface waters (MDOT, 1993).

There are no major removal mechanisms, such as volatilization, degradation (photodegradation, biodegradation), sorption (to particulates) or oxidation, that would remove the salts from surface waters. Dilution resulting from mixing with low-salinity water would decrease the salt concentrations in the aqueous phase. Hence, the salts will always be present in the aqueous phase rather than in the particulate phase (suspended or bottom sediments). In the case of benthic sediments, salts may accumulate in sediment interstitial water (pore water).

Dissolved salts may alter the physical properties of water by changing its density. Density increases linearly with increasing salinity (Ruttner, 1963). The increases in the density resulting from salinity changes are large in comparison with density changes associated with temperature and have important implications for lakes, as inorganic salts can accumulate temporarily or permanently in deeper strata and prevent lake waters from mixing (Section 3.3.3).

The chemical and biochemical behaviour of major ions from four inorganic salts and the ferrocyanide anti-caking agent considered in this report are discussed below.

2.5.1.1 Chloride

Chloride is the principal contributing anion to salinity from application of road salts and is influential in general osmotic salinity balance (Wetzel, 1975; Hammer, 1977). Chloride is a highly soluble and mobile ion that does not volatilize or easily precipitate or adsorb onto surfaces of particulates. In freshwater ecosystems, chloride behaves conservatively (i.e., its concentration in water is not affected by chemical or biological reactions) (Wetzel, 1975; Pringle et al., 1981). Chloride levels in surface water will largely change only in response to the addition of chloride, dilution by precipitation or inflow, or concentration by evaporation of water. In the majority of inland surface waters, the concentrations of chloride and the respective cations rarely reach the solubility products of the respective chloride salts, and so precipitation rarely occurs.

2.5.1.2 Sodium and potassium

In surface waters, the monovalent cations sodium and potassium are relatively conservative in their chemical reactivity and low biotic requirements. Because of this, the spatial and temporal distributions of sodium and potassium in unimpacted freshwater systems are uniform, with little seasonal variation.

2.5.1.3 Calcium

Calcium is the most reactive ion of all the major cations contributing to salinity. Its concentrations in surface waters are affected by chemical and biological processes within the aquatic systems. It is required as a micronutrient for higher plants and is one of the basic inorganic elements of algae. Calcium concentrations in lakes with hard water undergo seasonal changes, with a decrease in calcium concentrations and total alkalinity in the summer, when biogenically induced decalcification removes calcium from the water column. Decreased concentrations of inorganic carbon in the epilimnion, resulting from increased rates of photosynthesis, are responsible for the precipitation of calcium with bicarbonate. Some of the precipitated calcium carbonate is resolubilized in the hypolimnion, and some is entrained in sediments. Decalcification of the trophogenic zone changes monovalent:divalent cation ratios, which has an effect on the distribution and dynamics of algae and larger aquatic plants in freshwater ecosystems (Wetzel, 1975).

2.5.1.4 Magnesium

In general, magnesium compounds are more soluble than other compounds. In hardwater systems, calcium carbonates precipitate before more soluble magnesium carbonates. Significant precipitation of magnesium carbonate and hydroxide occurs only at very high pH (>10). Magnesium is a micronutrient in enzymatic transformations of organisms. It is required by chlorophyll-bearing plants as the magnesium porphyrin of the chlorophyll molecules (Wetzel, 1975). However, the metabolic requirements for magnesium are small in comparison with its availability in freshwater systems.

2.5.1.5 Sodium ferrocyanide

Sodium ferrocyanide dissociates to yield sodium and ferrocyanide ions. In the ferrocyanide ion, the chemical bond between the cyanide group and the iron is very strong; in water, however, it can be photolysed to release cyanide ions (Hsu, 1984). Laboratory tests determined that a 15.5 mg/L solution of ferrocyanide would produce 3.8 mg cyanide/L upon exposure to sunlight for 30 minutes (U.S. EPA, 1971).

2.5.2 Environmental pathways

The transport pathway for road salts applied to roadways is illustrated in Figure 8. The following sections explain the different environmental compartments that can be affected by release of road salts via roadway applications, snow disposal sites and salt storage sites (patrol yards).

2.5.2.1 Roadway applications

Road salts applied to roadways can enter surface water from:

• salt-laden runoff from roadways into drainage ditches, watercourses or wetlands soon after application;
• salt-contaminated meltwater from roadway surfaces or from snow plowed to roadsides after application of salts; and
• the release of salts stored in surface soils (Scott, 1980).

Factors affecting the degree of roadway runoff into aquatic ecosystems include a) the length of major road treated and drained, b) the amount of salts applied prior to the thaw period, c) road drainage pattern and topography, d) level of discharge of the receiving stream, e) degree of urbanization, f) rate of rise and duration of temperatures above freezing and g) precipitation (Scott, 1981).

Road salts applied to roadways can enter soils and groundwater from meltwater runoff from roadways and wet and dry deposition of airborne salt.

Road salts applied to roadways can enter air from windborne vehicle spray or splash and windborne dry powder from residue on roads. Windborne vehicle spray or splash can be affected notably by vehicle speed, wind, road gradient and geometrical features of the road (McBean and Al-Nassri, 1987).

2.5.2.2 Snow disposal sites

Road salts released through snow disposal will enter surface water from:

• direct release when dumping into surface waters, releasing to untreated storm sewers or melting following disposal on banks of water bodies;
• indirect releases via wastewater treatment plants in the case of dumping to treated storm sewers or of surface disposal sites draining to a treated water system;
• overland runoff of stormwater and snowmelt water; and
• stormwater pipage, which drains part of the disposal area to a surface water body.

Road salts released through snow disposal will enter soil and groundwater from direct percolation of precipitation and meltwater from snow disposal sites.

Figure 8 Road salts transport pathway

Enlarge image

2.5.2.3 Storage sites/patrol yards

Road salts released from storage sites or patrol yards will enter surface water from:

• salt-laden leachate from unprotected material piles;
• overland runoff of stormwater that has dissolved salts and of snowmelt water; and
• vehicle washwater.

Road salts released from storage sites or patrol yards will enter groundwater from:

• groundwater recharge from salt-laden soil and surface water;
• releases from vehicle washwater where the effluent is discharged to a dry well; and
• entry through wells or around well casings.

Road salts released from storage sites or patrol yards will enter soil from:

• spillage of salts at storage sheds during stockpiling or transferring of salts or blended abrasives/salts;
• spillage of salts or blended abrasives/salts during loading and unloading of maintenance vehicles;
• spillage of salts in the remainder of the yard as maintenance vehicles drive through;
• wind-driven erosion of salts from piles and from uncovered winter maintenance vehicles;
• salt-laden leachate from unprotected salt and blended abrasive/salt piles;
• precipitation or snowmelt waters, which dissolve salts and infiltrate into the soil; and
• release from vehicle washwater discharging onto the soil.
  
  Top of Page

2.6 Chloride concentrations in surface waters across Canada

This section presents information on the environmental concentrations of constituents of road salts in Canadian surface waters and presents estimates of chloride concentrations in surface waters across Canada based on mass balance modelling methods. Information in this section is summarized in Mayer et al. (1999) and Evans and Frick (2001) and is based on a review of numerous agency reports, scientific publications and monitoring surveys.

The surface water ecosystems covered include running (lotic) waters, such as rivers and creeks, and standing (lentic) waters, which include lakes, ponds and wetlands. Both the pelagic and the benthic compartments of these ecosystems are discussed. Also presented is a review of background concentrations and concentrations for water bodies that have been impacted by the release of road salts. Examples of water bodies that have been impacted by the storage of road salts were included in the section on patrol yards (see Section 2.4). The data suggest that the surface waters most sensitive to road salt impacts are low-dilution environments such as wetlands, small urban lakes and ponds with long residence times and small streams draining large urbanized areas.

2.6.1 Mapping of chloride concentrations in watersheds across Canada

Concentrations of chloride in water across Canada were mapped by watershed. Observed chloride concentrations in Canadian watersheds are provided in Figure 9. The concentrations of chloride in surface water are based on mean values obtained from federal and provincial water quality monitoring stations in Canada. The methods and database used in generating the map are provided in Mayer et al. (1999). Basically, monitoring data from near the mouth of each watershed are combined to provide estimates of the concentration for the entire watershed. The amount of data are varied, ranging from many stations and a monthly sampling interval to intermittent spot samples. Accordingly, data in Figure 9 should be used as an indication of the relative variations in the concentrations of chloride across Canada.

Watersheds containing major urban areas and areas of intense road salt loadings in southern Quebec, Ontario and the Maritime provinces have elevated chloride concentrations. Note that these concentrations represent averages over a watershed and, as such, do not indicate actual exposure concentrations of a specific watercourse. The map also indicates that some watersheds in the Prairie provinces have elevated ambient concentrations. These concentrations are related to the arid climate of the Prairies and are associated with the saline lakes and rivers situated in this region.

Road salts are not the only input that influences the concentrations in Figure 9. Effluents from various industries (e.g., pulp and paper mills) and releases from private septic systems and municipal wastewater treatment plants can influence chloride concentrations. It has been estimated for the lower Great Lakes that road salts contribute approximately 20% of the chloride load, with natural and other anthropogenic sources contributing the balance (Jones et al., 1986).

2.6.2 Environmental concentrations: Lakes and rivers

In the absence of anthropogenic sources, surface waters acquire their characteristics by dissolution and chemical reactions with solids, liquids and gases with which they come into contact during various phases of the hydrological cycle (Stumm and Morgan, 1981). For instance, stream water chemistry can be related to the chemistry of the source rock and to equilibria controlling the formation of solid phases (Feth et al., 1964; Garrels and Mackenzie, 1967). Watershed bedrock and surficial geology play a dominant role in determining the salt concentrations in surface waters. In addition, the climate and the proximity of the surface waters to the sea will be important in regulating the background salt concentrations. Thus, the background salt concentrations (salinity) of surface waters vary.

Surface waters range from fresh waters, with salinities lower than 500 mg/L, to saline waters, with salinities equal to or greater than 3000 mg/L (Hammer, 1986). While enclosed lakes and ponds may become saline, streams and rivers will rarely have natural salinity high enough to be classified as saline.

For the purpose of this Assessment Report, surface water quality is discussed by geographic region rather than by hydrogeological region in order to better address the effects of human factors on surface water quality. Both naturally occurring concentrations and those impacted by human activity are presented. While anthropogenic sources such as road salts, sewage and industrial effluents can all increase salinity, the main emphasis for data on impacted areas relates to the releases of road salts.

2.6.2.1 Atlantic region (Newfoundland, Nova Scotia, New Brunswick and Prince Edward Island)

The median concentrations of chloride and calcium from recent surveys (1985 and later) were compiled by Jeffries (1997). The data set includes 38 lakes from Labrador, 63 lakes from Newfoundland, 150 lakes from Nova Scotia and 166 lakes from New Brunswick. The median chloride and calcium concentrations in lakes in largely unimpacted areas ranged between 0.3 and 4.5 mg/L and between 0.82 and 1.10 mg/L, respectively.

Comprehensive studies by Kerekes et al. (1989), Kerekes and Freedman (1989) and Freedman et al. (1989) at Kejimkujik National Park, Nova Scotia, investigated the water quality of several lakes and rivers in the park. These studies concluded that sea salt influence is significant in Nova Scotia. The concentration ranges were 3.6-5.4 mg/L for chloride, 2.3-3.4 mg/L for sodium and 0.36-0.96 mg/L for calcium. The levels of these ions were comparable with those measured in other surface waters in the Atlantic region (Jeffries, 1997).

Figure 9 Observed chloride concentrations in Canadian watersheds (from Mayer et al., 1999)

Studies on the Chain Lakes near Halifax indicate that winter de-icing activities can have an impact on water quality. Thirmurthi and Tan (1978) explained that, from June to September 1975, the chloride concentration was approximately 19 mg/L. Concentrations began to increase in November and eventually peaked in May 1976, at 43 mg chloride/L. The chloride concentration then dropped rapidly through the rest of May and June, decreasing to 19 mg/L in July. The increase in chloride concentrations for both lakes represents an approximate addition of 19 440 kg of chloride to First Chain Lake and 7920 kg to Second Chain Lake. The Chain Lakes continued to be studied through the early 1980s (Hart, 1985, 1988). Chloride concentrations in both lakes increased gradually from 1976 to reach a maximum in the mid-1980s of approximately 120 mg/L in First Chain Lake and about 170 mg/L in Second Chain Lake. Highest chloride concentrations were observed during spring runoff.

Road salts have impacted Chocolate Lake in Nova Scotia. The lake is small (82.7 ha) and shallow (mean depth 3.9 m, maximum depth 12.2 m), with an estimated volume of 350 000 m³ (Kelly et al., 1976). The watershed is large (900 ha), approximately 11 times as great as the lake surface area. Chloride concentrations were highly elevated in the lake between April and August 1975. Average summer chloride concentrations were 207.5 mg/L compared with an estimated background value of 15-20 mg/L for non-impacted lakes. Chloride concentrations varied with depth, ranging from 199 to 224 mg/L at the surface, from 189 to 217 mg/L at 6.1 m, and from 225 to 330 mg/L at 12.1 m. Sodium concentrations ranged from 108 to 125 mg/L, from 102 to 125 mg/L and from 116 to 183 mg/L at the same depths. The gradient in salt concentrations and the absolute concentration of salt in deep waters were large enough to prevent complete vernal vertical mixing of the water column; deep waters became anoxic in summer. The primary source of chloride was attributed to highway runoff containing road salts. Kelly et al. (1976) estimated that 58.4 tonnes of road salt (representing 35 409 kg of chloride and 22 737 kg of sodium) were applied to the Chocolate Lake drainage basin during the winter of 1974-75.

A study on the water quality of 50 lakes in the Halifax/Dartmouth metropolitan area noted that most lakes experienced a very pronounced increase in the concentrations of sodium and chloride between 1980 and 1991 (Keizer et al., 1993). Of the lakes sampled in 1980, 12 had chloride concentrations greater than 50 mg/L and 3 had concentrations greater than 100 mg/L (the maximum concentration was 125 mg/L). In 1991, 22 of the 50 lakes sampled had chloride concentrations greater than 50 mg/L, 12 had concentrations greater than 100 mg/L and 4 had concentrations greater than 150 mg/L (the maximum concentration was 197 mg/L). The concentrations of sodium in these lakes showed similar trends. Of the lakes sampled in 1980, 9 lakes had sodium concentrations greater than 40 mg/L and 3 had concentrations greater than 60 mg/L (the maximum concentration of sodium was 73 mg/L). In 1991, 21 lakes had sodium concentrations greater than 40 mg/L, 12 had concentrations greater than 60 mg/L and 7 had concentrations greater than 80 mg/L (the maximum concentration was 114 mg/L). It is noteworthy that sampling was at the surface in the middle of the lakes and may not be characteristic of bays, where higher concentrations may occur. Furthermore, since denser, salt-laden water generally sinks to the bottom, these samples may be a low estimate of the overall lake concentrations.

Keizer et al. (1993) found that lakes in more developed watersheds had higher chloride and sodium concentrations than lakes in rural areas. In addition, salt concentrations from lakes in less developed areas did not increase markedly between the two sampling periods, unlike lakes in urban areas. For example, 19 of the rural lakes sampled had chloride concentrations below 25 mg/L in both 1980 and 1991; 23 of the lakes sampled had sodium concentrations below 20 mg/L in 1980 and 1991. Three of the lakes with the lowest concentrations had chloride concentrations of 5.1, 3.9 and 5 mg/L in 1980 and 4.5, 3.4 and 5 mg/L in 1991. Sodium concentrations were 3.4, 2.5 and 3.2 mg/L in 1980 and 2.9, 2.2 and 2.9 in 1991.

In general, the chloride concentrations in unimpacted water bodies in the Atlantic region do not exceed a concentration of 20 mg/L. Lower concentrations (<10 mg/L) are generally observed in inland water bodies, while the higher concentrations occur in coastal areas.

2.6.2.2 Central region (Ontario and Quebec)

The concentrations of chloride, sodium and calcium in lakes and rivers in the Canadian Shield, which covers a large area of Quebec and Ontario, including the upper Great Lakes, are very low (e.g., 1-10 mg chloride/L) (Jeffries, 1997). According to Weiler and Chawla (1969), the average annual concentrations of chloride in lakes Erie and Ontario were 24.6 and 27.5 mg/L, respectively, while the corresponding chloride concentration in Lake Superior was only 1.3 mg/L. Earlier studies, which investigated water quality trends in the Great Lakes, reported a rapid increase in chloride concentrations related to human activities (Kramer, 1964; Beeton, 1965; Dobson, 1967). The largest increase was observed for Lake Ontario. Later data (Williams et al., 1998), however, showed a decline in chloride concentrations from about 22 mg/L in 1977 to about 16 mg/L in recent years (1985-1993). The decline in the lower Great Lakes can probably be ascribed to lower loadings from industrial and domestic sources, resulting from improved control/treatment of industrial and domestic effluents. The tributaries to Lake St. Clair (e.g., Sydenham and Thames rivers) have somewhat higher chloride concentrations (about 30-40 mg/L) than other southern Ontario rivers because they drain the Paleozoic sedimentary rock adjacent to Michigan Basin, which contains saline formations such as salt (sodium chloride) and anhydrite (calcium sulphate) bearing strata. These strata also contribute to salt (sodium chloride and calcium sulphate) inputs to Lake St. Clair and western Lake Erie.

In urbanized areas, application of de-icing salts contributes significantly to chloride inputs. Long-term monitoring locations in the Greater Toronto Area show a gradual increase in chloride concentrations. In Highland Creek, a highly degraded urban watercourse, the statistical trend analysis shows an increase in chloride concentrations from 150 mg/L in 1972 (presumably already impacted) to over 250 mg/L by 1995 (Bowen and Hinton, 1998). In Duffin Creek, the median chloride concentrations increased from 10-20 mg/L in the 1960s to 30-40 mg/L in the early 1990s (Bowen and Hinton, 1998). In urban streams and rivers, such as the Don River in Toronto, concentrations of chloride increased to more than 1000 mg/L in winter (Scott, 1980; Schroeder and Solomon, 1998), with increases appearing to coincide with thaws. The autumn baseline values in the Don River were 100-150 mg chloride/L. Similar baseline concentrations of chloride (50-100 mg/L) were reported in another Toronto watercourse, Black Creek (Scott, 1980). Chloride concentrations 50 times higher than baseline values were measured during the thaw periods in this creek. Crowther and Hynes (1977) reported high winter concentrations of chloride (1770 mg/L), sodium (9550 mg/L) and calcium (4890 mg/L) in Laurel Creek, which passes through urban Waterloo. Rodgers (1999) observed high concentrations of chloride (4355 mg/L) in Red Hill Creek in Hamilton, Ontario.

The Ontario Ministry of Transportation has monitored water quality in a number of streams in Ontario, including Toronto-area watersheds, between 1990 and 1996. Results for four watercourses in urban areas are shown in Table 9. Highest concentrations were recorded during the winter months, indicating peak inputs of road salts. Chloride concentrations frequently exceeded 250 mg/L and often exceeded 500 mg/L (Table 9).

Little Round Lake in central Ontario was affected by cultural disturbances that included urbanization, septic tanks, highway road salt runoff and seepage from a salt storage depot (Smol et al., 1983). This small lake (7.4 ha, maximum depth 16.8 m) became meromictic. Meromictic conditions were attributed by Smol et al. (1983) to input of road salts from runoff and storage. Salt concentrations in the monimolimnion or deep layer were 58.4 mg sodium/L and 103.7 mg chloride/L, well in excess of that explainable by the natural geology of the region. Road salt additions apparently had resulted in the formation of meromixis in the 30 years prior to the study in 1981.

The Ministère des Transports du Québec (1980,1999) investigated the impacts of road salt on Lac à la Truite, near Sainte-Agathe-des-Monts. The lake drainage area, estimated at 728 ha, was affected by a 7-km stretch of highway (approximately 30 lane-kilometres). At certain places, the highway was located as close as 250 m from the lake. The lake has a surface area of 48.6 ha, a mean depth of 21.5 m and an estimated volume of 486 000 m³. In 1972, the average chloride concentration for the lake was 12 mg/L. This increased through the 1970s to reach a maximum concentration of 150 mg/L in 1979. This corresponds to an addition of 67 068 kg of chloride to the lake. To prevent further increases, abrasives (containing some salt to prevent caking) were used instead of road salts. Chloride concentrations fell through the 1980s to reach 45 mg/L in 1990 and have remained at that level since.

Table 9 Chloride concentrations in various streams in the Toronto Remedial Action Plan watershed for 1990-96 (from Toronto and Region Conservation Authority, 1998)

Enlarge image

In the United States, significant increases in chloride concentrations observed in four Adirondack streams were attributed to winter salt application (Demers and Sage, 1990). There was a significant difference in chloride concentrations between samples upstream and downstream from roadways. Demers (1992) reported that chloride concentrations were as much as 66 times higher in downstream samples. The overall mean chloride level was 0.61 mg/L upstream and 5.23 mg/L downstream. The terrain characteristics of Adirondack streams (New Hampshire) are similar to those of streams in the Eastern Townships of Quebec. Chloride concentrations of 11 000 mg/L (Hawkins and Judd, 1972) were measured in winter of 1969 in Meadowbrook, which drains about 10 km² of suburbs in Syracuse, New York (Hawkins and Judd, 1972). Also in the United States, chloride concentrations of about 400 mg/L were measured in the bottom waters of Irondequoit Bay, a small bay on Lake Ontario near Rochester, New York (Bubeck et al., 1971, 1995).

In summary, the chloride concentrations of unimpacted water bodies in the Shield region of Central Canada are among the lowest in the country and vary within the range <1-5 mg/L. Higher background concentrations of up to 10^-30 mg/L are observed in water bodies situated within the St. Lawrence Lowlands region, which includes the basin of the lower Great Lakes and the St. Lawrence River. A marked increase in chloride concentrations above background levels is particularly noticeable in many small urban lakes and watercourses in this region, as it is the most densely populated area of Canada. Seasonal and spatial trends indicate the important contribution of road salts to these increases.

2.6.2.3 Prairie region (Manitoba, Saskatchewan and Alberta)

Most naturally saline lakes in Canada are located within two regions. The first region encompasses several endorheic drainage basins on the Canadian Prairies (Hammer, 1984). The area where saline lakes are most numerous stretches over southern Alberta and Saskatchewan. The Alberta-Saskatchewan saline lake region is largely underlain by Cretaceous bedrock, composed mainly of shales, silts and sandstones (Hammer, 1984). Of the Prairie provinces, Saskatchewan has by far the greatest number and volume of saline lakes (Hammer, 1986). The relative proportions of cations in most Prairie lakes are as follows: Na > Mg > Ca. Of the lakes investigated, there are only two known meromictic lakes on the Canadian Prairies, Waldsea Lake and Deadmoose Lake (Hammer, 1984). Numerous potholes (shallow lakes or ponds), which range in salinity from fresh to saline, can be found in the southern Prairies.

There are fewer naturally saline lakes (about 30) in Alberta (Hammer, 1984). Most of them are concentrated in the Provost and Hanna regions, although they occur as far north as Edmonton. Even fewer saline lakes are in Manitoba. Barica (1978) investigated a group of over 100 small potholes, which varied in salinity from fresh to moderately saline, near Erickson in southwestern Manitoba. The concentration ranges were 1-448 mg/L for chlorine, 0.8-1075 mg/L for sodium and 27-380 mg/L for calcium. Twenty-three sodium chloride-dominated sites along the western shore of Lake Winnipegosis were studied by McKillop et al. (1992). The concentration ranges for chloride, sodium and calcium were 861-33 750 mg/L, 587-21 313 mg/L and 59-1400 mg/L, respectively.

Much lower concentrations of chloride and calcium are reported for 27 and 193 lakes in northern portions of Saskatchewan and Alberta, respectively (Jeffries, 1997). The median concentrations of chloride were 0.2 and 0.5 mg/L, respectively. The calcium concentrations of Saskatchewan and Alberta lakes were 2.0 and 14.0 mg/L, respectively. The chloride and calcium concentrations of 26 lakes in Manitoba were 2.0 and 7.5 mg/L, respectively (Jeffries, 1997).

In summary, many of the saline water bodies in Canada are located within the Interior Plains, which cover the southern portions of the Prairie provinces. In general, dissolved salt concentrations, including chloride, are higher in this geographical region because of the underlying geology. However, low concentrations of chloride (<5 mg/L) are reported for unimpacted water bodies situated in the northern region of these provinces, outside of the Interior Plains region.

2.6.2.4 Pacific region (British Columbia)

The second region in Canada where a large number of naturally saline lakes are located is the Southern Interior Plateau (Fraser Plateau) of British Columbia, in the rain shadow of the Coast Mountains. Saline lakes in British Columbia are small and shallow (Topping and Scudder, 1977). The limnology of four of these lakes was described by Northcote and Halsey (1969), who showed not only that these lakes differed in total dissolved solids concentrations among themselves, but also that there were differences within the lakes (surface to near-bottom). Concentrations ranging from 5.1 to 800 mg/L were reported. The meromixis in three of these lakes was maintained by chemical density gradients, while in the fourth, the mixing was inhibited by morphometric features. In non-saline regions, the reported median chloride and calcium concentrations of six acid-sensitive lakes located in the southwest of the province were 2.5 and 3.7 mg/L, respectively (Phippen et al., 1996; Jeffries, 1997).

The impact of road salts on water quality was monitored for the Serpentine River in British Columbia's Lower Fraser Valley. Electronic data recorders monitored water temperature, pH, water level and conductivity every 15 minutes. Three-fold increases in conductivity were observed over 10- to 20-hour periods during thaw periods following a cold period when roads were sanded and salted (Whitfield and Wade, 1992). Furthermore, periods of elevated conductivity were followed by increased water levels, which are characteristic of snowmelt. Results of this study indicate that organisms living in certain streams during winter months may be subject to fluctuations in salt concentrations.

In summary, there are several lakes located in the Southern Interior Plateau of British Columbia whose chloride concentrations are higher than 100 mg/L, and several of these lakes are naturally saline. The chloride concentrations of unimpacted water bodies in the remaining parts of British Columbia are generally less than 5 mg/L.

2.6.2.5 Yukon, Northwest Territories and Nunavut

Few data were identified from the Yukon, Northwest Territories and Nunavut.

2.6.3 Environmental concentrations: Benthic sediments

Because inorganic chloride salts used for road maintenance are highly soluble and interactions with sediment particulates are minimal, inorganic salts are expected to accumulate in sediment pore water rather than in the solid phase. The study by Mayer et al. (1999) of an urban pond (Rouge River Pond) shows that the sediment pore water, which is in equilibrium with the overlying water, enriched in inorganic salts, may attain high salt concentrations (see Section 2.6.5). High concentrations of salts in sediment pore water can result not only in osmotic stress and direct toxic effects on benthic biota but also in complexation of metals with chloride, augmenting the concentrations of dissolved heavy metals (e.g., cadmium), which are toxic to benthic organisms.

2.6.4 Environmental concentrations: Wetlands

Wetlands are an important landscape feature in Canada. Their most distinguishing features are presence of standing water, unique wetland soils and vegetation adapted to or tolerant of saturated soils (Mitsch and Gosselink, 1986). Canada has a great variety of wetlands; detailed descriptions of various types of wetlands can be found in Mitsch and Gosselink (1986).

There is only limited information on the chemical characteristics of coastal and inland freshwater wetlands with respect to chloride, for much of the work on wetlands is concerned with nutrient cycling, acidification and geochemistry of metals. A 1987 study of wetlands within Kejimkujik National Park (Wood and Rubec, 1989) assessed the chemical characteristics of wetlands. The authors reported significant differences in peat chemistry between the bogs and fens, in particular in sodium and calcium concentrations. The concentrations of calcium, sodium and chloride in peat from bogs were 0.17, 0.8 and 0.15 mg/L, respectively. The concentrations of calcium, sodium and chloride in peat from fens were 0.32, 1.5 and 0.12 mg/L. There was no difference in the surface water chemistry between bog and fen. The concentrations of calcium, sodium and chloride in water of bogs were 6 x 10^-4, 3.0 x 10^-3 and 3.5 x 10^-3 mg/L, respectively, and the concentrations of calcium, sodium and chloride in water of fens were 6 x 10^-4, 2.8 x 10^-3 and 3.9 x 10^-3 mg/L, respectively.

Bourbonniere (1985, 1998) investigated the geochemistry of wetlands in Nova Scotia and Ontario. He measured the major ion chemistry in pore water of the ombrotrophic bog in Barrington County in Nova Scotia. The measured concentrations were 9.6-24 mg/L for chlorine, 5.2-12.8 mg/L for sodium and 0.47-1.5 mg/L for calcium. The concentrations of calcium, sodium and chlorine in Beverly Swamp, Ontario, were 58.1-95.3 mg/L, 5.1-20.0 mg/L and 8.7-41.2 mg/L, respectively. The water composition of potholes described in Section 2.6.3.3 may be considered to be characteristic of the prairie wetlands. No Canadian studies were found that address the impact of road salt application on the chemistry of wetlands.

Two studies were presented in the patrol yard section (Section 2.4) that focus on the impact of salt depots on wetlands.

2.6.5 Environmental concentrations: Urban lakes and ponds

Natural small urban lakes and ponds that receive road runoff are susceptible to changes in ion concentrations. There are only limited data on the chloride concentrations in such bodies. In the absence of such data, ponds that are part of an urban stormwater management system can be used as a reasonable surrogate. Numerous ponds have been constructed and remain a widely used form of watershed management in urban areas. Although these ponds are engineered water bodies designed primarily for hydraulic flow management and stormwater treatment, they have developed aquatic ecology similar to natural small urban ponds. Like natural urban aquatic systems, these ponds are wetland habitats for many species of flora and fauna. They are likely candidates for showing maximum chloride concentrations, since they represent the type of water body that receives the maximum exposure to chloride loadings from winter maintenance activities.

A study of Lake Wabukayne, a small human-made lake in Mississauga, Ontario, reported chloride concentrations between 200 and 2000 mg/L and temporary meromixis during the winter and early spring (Free and Mulamoottil, 1983). Chloride concentrations of 1100-2000 mg/L (Vickers, 2000) were measured in the Harding Pond in Richmond Hill in the Greater Toronto Area. Seasonal surveys of the stormwater retention ponds in the Greater Toronto Area (Mayer et al., 1996; Mayer, 2000) showed higher concentrations of chloride during the winter months. Chloride concentrations of 380 and 800 mg/L, respectively, were measured in the Heritage and Unionville ponds, which are situated in residential settings (Mayer, 2000), while chloride concentrations as high as 5910 mg/L were measured in the Col. S. Smith Reservoir, receiving runoff from a multi-lane highway (Queen Elizabeth Way) in February of 1999.

Both a mass balance study and a sediment study were performed on the Rouge River Pond, situated in the Rouge River Valley, Scarborough, Ontario, near Highway 401 and Port Union Road. This pond is an urban stormwater management facility designed to treat highway stormwater by removing 70% of suspended solids. The mass balance study was conducted to measure the pollutant removal and the dynamics of several water quality constituents between 1995 and 1998 (Liang, 1998). The initial studies during 1995, 1996 and 1997 focused on balances during the ice-free season. A chloride imbalance related to permanent meromictic conditions in the pond was shown, with meromixis likely due to a combination of the pond design and high road salt loadings. The salt concentrations calculated as sodium chloride in the bottom layer of the pond were of the order of 5000 mg/L. Elevated concentrations of metals and ammonia were also measured in this bottom layer. The sediment study (Mayer et al., 1999) investigated the chemistry and toxicity of sediment pore water and assessed the solid-phase metal chemistry. A geochemical model was formulated to assess the impact of salts on metal speciation. The investigation revealed that sediment pore water, which is in equilibrium with the overlying water, enriched in inorganic salts, may itself attain high salt concentrations. At the sediment-water interface, concentrations of chloride and sodium greater than 3000 mg/L and 2000 mg/L, respectively, were measured in the pore water. A gradual decrease with depth in sediment was observed for both ions; however, even at a depth of 40 cm below the sediment-water interface, the chloride concentrations were about 1500 mg/L.

A study by Bishop et al. (2000) focused on the presence of contaminants in six stormwater retention ponds in the Greater Toronto Area and nine stormwater retention ponds in Guelph, Ontario. Of the 15 ponds in this study, 4 ponds in Guelph were sampled between 19 and 21 times each to determine the concentration of chloride in stormwater. These samples were taken at different times during the months of September, October and November 1997 and April, May, June, July, August, September and November 1998. Samples were also taken as close to the ponds' outfalls as possible (Struger, 2000). Mean concentrations of chloride in stormwater samples from these four ponds were between 120 and 282 mg/L; the maximum concentrations observed at the four ponds were between 416 and 1230 mg/L (Struger, 2000). While samples were not taken from December to March, the average chloride concentrations of samples for November, April and May were higher than the average chloride concentrations of samples for the other months at each of the ponds (336 vs. 153 mg/L; 183 vs. 79 mg/L; 469 vs. 82 mg/L; 444 vs. 188 mg/L).

The samples were taken close to the ponds' outfalls because these concentrations are probably indicative of the concentrations entering receiving environments. The stormwater retention ponds treat suspended solids and other attached constituents, but do not remove dissolved salts from runoff (OME, 1991; Bishop et al., 2000). Depending on the volume of runoff entering these ponds, they are typically designed to retain runoff for a period ranging between 24 and 72 hours; retention time may be as low as 10 hours during peak periods of large events.

Watson (2000) sampled the quality of water in 89 ponds located near roadways in southern Ontario. Multiple samples were taken during spring and summer of 2000, and the ponds were categorized according to the number of lanes of the nearest road. The mean concentration of chloride in ponds located near two-lane roads was 95 mg/L (range 0-368 mg/L). The mean concentration of chloride in ponds located near roads with more than two lanes but fewer than six lanes was 124 mg/L (range 0-620 mg/L). The mean concentration of chloride in ponds located near roads with six lanes or more was 952 mg/L (range 49-3950 mg/L). Watson (2000) does not provide an indication of background concentrations, but indicates that the main source of chlorides would be from road salts.

2.6.6 Chloride concentrations based on mass balance calculations

2.6.6.1 Calculation of watershed chloride concentrations resulting from the use of road salts

A mass balance model was used to estimate chloride concentrations in Canadian watersheds from the use of road salts. The calculated chloride concentrations were compared with the observed concentrations presented in Section 2.6.1. More details of this comparison and the analysis are presented in Mayer et al. (1999).

While there are some simplifications inherent to the model used (e.g., assumption of similar road salt application and volume of runoff from year to year), the model provides a reasonable estimate of the potential chloride concentrations resulting from road salt use. The model is also useful for identifying the relative concentrations of chlorides in the different watersheds. Sub-basin watersheds are used in the calculations; a more accurate comparison of calculated chloride concentrations would have been obtained if finer-scale resolution watershed maps had been used.

The results are presented in Figure 10. The Island of Montréal has the highest modelled concentration of chloride in runoff. Figure 10 also indicates that watersheds on the north shore of Lake Ontario, notably those largely influenced by municipalities in the Toronto area, also have high chloride concentrations related to road salts. The watersheds in urban areas, particularly those in southern Ontario, southern Quebec, the Maritime provinces and larger urban centres in the Prairie provinces, have the highest modelled concentrations of chloride in runoff.

The results are presented in Figure 10. The Island of Montréal has the highest modelled concentration of chloride in runoff. Figure 10 also indicates that watersheds on the north shore of Lake Ontario, notably those largely influenced by municipalities in the Toronto area, also have high chloride concentrations related to road salts. The watersheds in urban areas, particularly those in southern Ontario, southern Quebec, the Maritime provinces and larger urban centres in the Prairie provinces, have the highest modelled concentrations of chloride in runoff.

It is noteworthy that the concentrations presented in Figure 10 are modelled averages for watersheds. Accordingly, the concentrations do not indicate actual exposure concentrations for specific watercourses or water bodies. Data in Figure 10 can be used to identify areas that face the greatest risk of adverse effects from the use of road salts.

Figure 10 Estimated road salt chloride concentrations by watershed, calculated from average annual road salt loadings and average annual runoff (from Mayer et al., 1999)

2.6.6.2 Calculation of chloride concentrations in roadway runoff

The previous section assessed chloride concentrations caused by road salts for entire watersheds. At a watershed scale, modelled concentrations are diluted by runoff from areas other than just the roadway. This section estimates the concentration of chlorides at a roadway scale, using a mass balance approach.

For these estimates, it was assumed that the mass of salt used per two-lane-kilometre of provincially maintained road was dissolved and diluted by the volume of average annual runoff that would accumulate on the roadway surface (i.e., in 7400 m², based on an assumed width of a two-lane kilometre of 7.4 m).

Figure 11 presents estimated concentrations of chloride and the length of provincial roads associated with these concentrations. Concentrations of chloride range between 30 and 31 000 mg/L. Based on these estimates, the majority of the concentrations (85%) range between 1000 and 10 000 mg/L. These estimates are of the concentrations of chloride in runoff from roadways; various application factors can be used to account for dilution in receiving waters. The accuracy of results depends on the quality of data used to calculate concentrations. One limitation is that estimates of average annual runoff were derived using a map at a scale of 1:7 500 000. More detailed maps (i.e., 1:50 000) might result in slightly different results. Furthermore, estimates of average annual runoff were used. Different values for runoff may have been calculated if runoff associated with winter months or spring thaw had been used.

Figure 11 Chloride concentrations in runoff for provincial roads in Canada

The concentrations of chloride observed in a recent study (Mayer et al., 1998) of highway runoff from three sites in southern Ontario are consistent with the above estimated chloride concentrations. The study reported chloride concentrations in runoff from roadways at three sites: a) the Burlington Skyway Bridge (a four-lane road with 92 000 cars per day), b) Highway 2 east of Brantford (a two-lane road with 31 100 cars per day) and c) Plains Road in Burlington (a two-lane roadway with 15 460 cars per day). Highest chloride concentrations (up to 19 135 mg/L) were observed in winter runoff from the Skyway Bridge, although high concentrations were also observed at Plains Road. The study, which investigated runoff toxicity, showed that salt-laden runoff has the potential to adversely affect aquatic organisms in low-dilution aquatic systems. The high chloride concentrations reported in this study are representative of concentrations at point of discharge and represent the worst-case scenario, which may be found along roadside ditches or small wetlands adjacent to major roadways.

2.6.7 Summary

Data show a broad range of background chloride concentrations in Canadian surface waters. Differences stem from the variance in bedrock and surficial geology, climate and proximity to the sea. Impacts of high use of road salts are evident from the data, particularly in highly urbanized areas. Water bodies most sensitive to the releases of road salts are low-dilution environments, such as small urban lakes and ponds with long residence times, urban stormwater ponds with short residence times, streams draining large urbanized areas and wetlands adjacent to major roadways.

In terms of natural concentrations across Canada, the following regional trends have evolved:

• In general, the chloride concentrations in unimpacted water bodies in the Atlantic provinces do not exceed about 20 mg/L. Lower concentrations (<10 mg/L) are generally observed in inland water bodies.
• Chloride concentrations of unimpacted water bodies in the Canadian Shield region of Central Canada are among the lowest in the country and vary within the range <1-5 mg/L. A marked increase in chloride concentrations is noticeable in many small urban lakes and streams.
• Many of the naturally saline water bodies in Canada are in the southern portions of the Prairie
  provinces. However, naturally low concentrations of chloride (<5 mg/L) are reported for unimpacted water bodies in the northern region of these provinces.
• There are several naturally saline lakes located in the Southern Interior Plateau of British Columbia (chloride concentrations higher than 100 mg/L). The chloride concentrations of unimpacted water bodies are generally less than 5 mg/L.

A spatial analysis of chloride concentrations was done based on two approaches. The first aggregated monitoring data to a watershed scale; the second used a mass balance analysis to model chloride concentrations at a watershed level. Mass balance techniques were also used to compute average annual concentrations in runoff from specific highways across Canada. The calculated concentrations range up to 31 000 mg/L, but mainly lie between 1000 and 10 000 mg/L. These estimates are consistent with measured undiluted runoff from roadways.

Elevated chloride concentrations were also measured in ponds and wetlands adjacent to roadways. Concentrations up to 4300 mg/L were observed in watercourses, and concentrations between 150 and 300 mg/L were observed in rural lakes that had been impacted by the use of road salts. While highest concentrations are usually associated with winter or spring thaws, high concentrations can also be measured in the summer, as a result of the travel time of the ions to surface waters and the reduced water flows in the summer.