Location: Ministry Home > Publications > Technical Studies and Reports > 4022e2.pdf > HTML version

This is a HTML version of the original PDF document. The HTML version is being provided for reading purposes only and is not the official version of the document.

2. AQUATIC ENVIRONMENTAL EFFECTS OF THE FORMER DELORO MINE
SITE ON THE MOIRA RIVER SYSTEM

2.1 Water Quality

2.1.1 Water Quality Methods

Sources of Water Quality Data Used in This Study

Water quality data and station locations were obtained from the Ministry of the Environment (MOE)Provincial Water Quality Monitoring Network Program (PWQMNP) (MOE 1964 to 1999). A large proportion of the data was obtained electronically and placed in a Microsoft Access database. During 1996, the MOE upgraded their database. Some data gaps were identified for the period 1996 to present. Hard copy of some additional data were made available and manually entered into the Access database. The PWQMNP includes the routine analysis of several metals, nutrients and other parameters such as dissolved oxygen and turbidity. Analytical methods used for the PWQMNP are listed in Appendix I.

Data gaps exist within the PWQMNP. Not all sampling stations were sampled each year. These data gaps are identified by the notation "ND" in Table 2.1-1. There were some years where no data were available for a particular metal. The years where no measurements were recorded at any sampling station for each parameter are identified below (note: data for some of the following parameters and dates were available for stations sampled at and near the Deloro Mine Site by the Ontario Clean Water Agency):

• Arsenic 1998;
• Cobalt 1980 to 1984 and 1987;
• Hardness 1977 to 1987;
• Lead 1973 to 1975;
• Dissolved Oxygen 1994 to 1997; and
• Nickel 1980-1981 and 1983-1984 and 1987.

Silver and uranium were analyzed only once during the period of record. The remaining analytes in Table 2.1-1 (i.e., copper, pH, stream flow, suspended solids and turbidity) represent a continuous data set for most sampling stations for the sampling periods identified.

Table 2.1-1

Historical Provincial Water Quality Network Data Collection:Selected Relevant Parameters

Parameter	Sampling Period		Number of Samples Collected per Network Station (W) for Chemicals of Concern
			1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	22	23	24	DM1	DM7	DMS
Arsenic	1964	1999	382	495	4	446	32	502	8	8	462	493	498	419	439	47	472	11	492	ND	ND	453	36	481	78	74	60	60
Cobalt	1979	1999	65	96	ND	24	ND	78	15	8	26	27	26	18	26	ND	25	ND	27	6	5	26	ND	27	33	34	20	20
Copper	1972	1999	66	98	ND	53	ND	146	15	8	54	138	55	42	153	ND	54	ND	57	6	5	165	ND	191	33	34	20	20
Dissolved oxygen	1964	1998	157	220	ND	173	21	237	ND	ND	176	213	214	155	190	42	194	ND	204	1	2	251	36	274	7	ND	ND	ND
Hardness	1965	1999	26	77	ND	68	10	76	7	ND	65	85	85	60	99	16	75	ND	80	5	5	155	24	126	25	ND	ND	ND
Lead	1972	1999	65	97	ND	53	ND	146	15	8	54	138	56	42	154	ND	54	ND	57	6	5	148	ND	191	33	44	44	44
Nickel	1972	1999	66	98	ND	52	ND	81	15	8	54	58	56	42	89	ND	54	ND	57	6	5	127	ND	102	33	74	60	60
pH	1964	1999	66	128	1	70	13	186	6	ND	62	153	83	59	166	16	73	8	82	5	5	194	32	219	27	ND	ND	ND
Silver	1987	1987	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	1	ND	ND	ND	ND
Solids;suspended	1997	1999	ND	10	ND	5	ND	ND	10	4	12	11	11	7	11	ND	11	ND	11	ND	ND	11	ND	10	15	ND	ND	ND
Stream flow	1964	1989	ND	399	ND	ND	ND	ND	ND	ND	372	423	60	ND	ND	ND	ND	ND	ND	ND	ND	346	ND	8	ND	ND	ND	ND
Turbidity	1964	1999	133	247	ND	176	19	289	10	4	149	244	191	134	237	44	175	ND	182	5	5	312	39	309	33	ND	ND	ND
Uranium	1991	1991	ND	ND	ND	ND	ND	1	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND

ND: No Data.

DM: OCWA Station; 1993-1999.

Beginning in 1993, an intensive surface water monitoring program was initiated to collect daily and weekly water quality samples for aluminum, arsenic, nickel, lead, mercury, cobalt, copper, molybdenum and zinc in close proximity to the Deloro Mine Site. Results from station numbers DM1, DM7 and DM8 from the intensive study were added to the database for this study (see Table 2.1-1). Water quality data from the remaining stations DM2, DM3, DM4, DM5 and DM6 are not representative of the Moira River downstream of the Deloro Mine Site and were not considered in this study. The intensive program is conducted by the Ontario Clean Water Agency (OCWA) who operate the treatment plant at the Deloro Mine Site.

Some parameters measured in the PWQMNP were frequently less than detection limits[Table 2.1-2). The total number of samples collected for selected relevant parameters and the number of samples that yielded results below detection are identified in Table 2.1-2. There were no quality control samples included within the PWQMNP data set. It should be noted that detection limits declined significantly over the period of record (see later discussion).

Table 2.1-2

Total Number of Samples Collected and the Percent of Results Found Below Detection

Limits for Selected Relevant Parameters

See Figure 2.1-2 for changing detection limits over time.

The database for this study was augmented in 1999 and 2000 by the analysis of selected water quality samples for radioactivity. This was the only water quality analysis conducted specifically for this study. The samples were collected by MOE from stations at Stoco Lake, Moira Lake Narrows at Highway 62, Moira Lake Outlet, Moira River at Highway 7 and Malone. Samples were submitted to the Ministry of Labour, Radiation Protection Monitoring Service, evaporated and analyzed for the following parameters:

• Radium-226;
• Radium-228;
• Thorium-228;
• Uranium-238;
• Gross alpha activity; and
• Gross beta activity.

A description of the analysis for radionuclides is included in Appendix I.

Water quality station numbers, names and descriptions are provided in Table 2.1-3 and shown on Figure 2.1-1. The water quality sites have been numerically labelled in order of upstream/Malone to downstream/Belleville (i.e., Station Numbers 1-24). Station 21 was removed from the data set because no data have been collected to date.

Results found to be below detection were incorporated in the calculation of annual averages by using the detection limit as the value. Incorporating the value of the detection limits could bias the annual averages for arsenic, lead and nickel since 32% to 71% of results (Table 2.1-2) are at the limit of detection. Thus, the actual averages may be lower than the averages calculated using the detection limit because the actual concentrations may have been considerably less than the detection limit. In addition, detection limits have been decreasing intermittently since the PWQMNP was established for the Moira River system in 1964 (Figure 2.1-2). Therefore, decreasing annual average metal concentrations may correlate with increasing precision or decreasing detection limits of analytical instruments (Bowen 1994). Dealing with detection limits was further complicated by the fact that detection limits within a year fluctuate between stations and by month depending on the set-up of the analytical instrument when the samples were analyzed.

Table 2.1-3

Identification and Description of Water Quality Stations

Figure 2.1.1

(WATER QUALITY MONITORING NETWORK)SAMPLE LOCATIONS

Figure 2.1-2

Annual Average Water Quality Detection Limits for Selected Chemicals

Identification of Metals of Concern

Water quality data were screened to identify the metals of concern for this study. Metals of concern were identified by comparing concentrations to Provincial Water Quality Objectives (PWQOs) (MOE 1999). In some cases, concentrations in the Moira River system downstream of the Deloro Mine Site exceeded PWQOs, but did not exceed reference concentrations. These constituents were elevated throughout the study region (including reference sites) and were assumed to represent metals with high natural background concentrations. The high natural background constituents were aluminum, cadmium, chromium, iron, and zinc.

The metals of concern for this study, and their respective PWQOs in µg/L and mg/L are:

• Arsenic: 100 µg/L(0.1 mg/L);
• Cobalt: 0.9 µg/L(0.0009 mg/L);
• Copper: 5 µg/L(0.005 mg/L);
• Lead: 5 µg/L (0.005 mg/L);
• Nickel: 25 µg/L(0.025 mg/L); and
• Silver: 0.1µg/L (0.0001 mg/L).

These metals exceeded PWQOs at one or more stations downstream of the Deloro Mine Site, and were found at concentrations higher than at the reference sites at most or all of the stations downstream of the Deloro Mine Site. Silver was included despite the paucity of data because it is one of the more toxic metals to aquatic organisms and because it exceeded the PWQO at some stations.

Radionuclides were screened out as metals of concern in water. All uranium concentrations were less than the detection limit of 0.1 µg/L (Appendix I). Concentrations of radium-226 and thorium228 were less than detection limits at all sample sites (Appendix I). Gross alpha activity was less than the detection limit of 0.04 to 0.05 Bq/L at all sampling stations. Gross beta activity ranged from <0.04 to 0.10 Bq/L (Appendix I). These results confirm the lack of any significant transport of radionuclides in water from the Deloro Mine Site. No significant transport in water was expected because earlier studies of the extent of radioactive materials at the Deloro Mine Site found that the materials in the slag in the Industrial Area were immobile (SCIMUS 1999). Furthermore, SCIMUS found that radioactive contamination caused by tailings transported by erosion into the upper portion of Young's Creek became indistinguishable from background upstream of Highway 7.

The metals of concern are referred to as "metals" in the rest of this report. This includes arsenic, which is commonly referred to as a metal, although it is, in fact, a metalloid.

Temperature and Oxygen Profiles

Temperature and oxygen profiles were measured in each of the four study lakes (Moira, Stoco, Consecon and Round lakes) in late August, 1999. A WTW Microprocessor Pocket Oxygen Meter (Oxi 330/set) was used to measure dissolved oxygen. Profiles were taken at locations where previous data (MNR records) indicated maximum depths.

2.1.2 Results and Discussion

Metal Concentrations with Distance Downstream of the Deloro Mine Site

The concentration of the metals of concern declined with distance downstream of the Deloro Mine Site¹ (Figures 2.1-3 to 2.1-7). The data used to prepare these figures are found in Appendix I. The relative decline with distance downstream varied with each metal. For example, recent (1995-1999) mean annual concentrations declined with distance downstream as follows:

• Arsenic: 0.01 - 0.07 mg/L in the Highway 7 bridge-to-Moira Lake (immediatelydownstream) section to 0.002 to 0.005 mg/L in the Stoco Lake-to-Belleville (further downstream) section.
• Cobalt: 0.001-0.09 mg/L immediately downstream to 0.0002-0.001 mg/L in the further downstream section.
• Copper: 0.001-0.02 mg/L immediately downstream to 0.0003-0.002 mg/L in the further downstream section.
• Lead: 0.001-0.01 mg/L immediately downstream to 0.0008-0.005 mg/L in the further downstream section.
• Nickel: 0.003-0.02 mg/L immediately downstream to 0.0007-0.005 mg/L in the further downstream section.

The frequency of exceedances of the PWQOs also decreased with distance downstream. Recent mean annual arsenic and nickel concentrations did not exceed the PWQO at any of the stations downstream of station W2 (on the Moira River just downstream of the Deloro Mine Site, i.e., W6, W11, W15, W17 and W23) (Figures 2.1-3 and 2.1-7; Appendix I). Recent mean annual cobalt, copper and lead concentrations did not exceed the PWQO downstream of Moira Lake [Figures 2.1-4, 2.1-5 and 2.1-6; Appendix I).

Figure 2.1-3

Annual Mean Moira River Arsenic Levels (from Upstream to Downstream) (mg/L)

Figure 2.1-4

Annual Mean Moira River Cobalt Levels (from Upstream to Downstream) (mg/L)

Figure 2.1-5

Annual Mean Moira River Copper Levels (from Upstream to Downstream) (mg/L)

Figure 2.1-6

Annual Mean Moira River Lead Levels (from Upstream to Downstream) (mg/L)/L)

Figure 2.1-7

Annual Mean Moira River Nickel Levels (from Upstream to Downstream) (mg/L)

Changes in Metal Concentrations from 1965 to 1999

Arsenic concentrations in the Moira River have decreased substantially since 1965. Arsenic concentrations decreased by a factor of 10 times at station W2 (just downstream of the Deloro between the mid-1970s and mid-1980s (Figure 2.1-3). This decrease reflects the clean up activities conducted by the MOE at the Deloro Mine Site. These activities included covering of tailings, the installation and operation of a system of wells to capture contaminated groundwater and the construction and operation of a treatment plant to remove arsenic and other metals from groundwater.

Since the mid 1980s, mean annual arsenic concentrations in the river immediately downstream of the site at station W2 have varied from year to year with no apparent trend, although they were always at or below the PWQO (Figure 2.1-3). This variability may be partly in response to variation in river flow. During wetter years when average flows were higher, the concentration of metals, including arsenic, was lower. Conversely during drier years, there was less water available for dilution and the resulting concentrations of metals, including arsenic, were higher. For example, the mean annual flow in 1988 was low (2.62 m³/sec) near station W2. Arsenic concentrations at station W2 averaged 0.10 mg/L in 1988. In contrast, the mean annual flow in 1995 was 3.68 m³/sec. Arsenic concentrations at station W2 averaged 0.02 mg/L in 1995. Fluctuation in arsenic concentrations with flow also appear to have occurred upstream of the Deloro Mine Site. For example, arsenic concentrations peaked in the low-flow year 1988 atstation W1 near Malone (Figure 2.1-3).

Arsenic concentrations in Moira Lake and Stoco Lake have also declined over time and are nowabout a factor of 10 below concentrations recorded in the 1960s and 1970s (Station W6 and W15in Figure 2.1-3). Arsenic concentrations from Stoco Lake to Belleville have also declined by atleast a factor of 10 since the 1960s and 1970s (Appendix I). Variability in concentrations sincethe early 1980s may be related to river flow, as discussed above.

Cobalt concentrations have fluctuated over time, but no consistent trends were observed(Figure 2.1-4). Cobalt was not measured in the 1960s or 1970s; the record begins in 1988.Concentrations have remained consistently above the PWQO in the Deloro-to-Moira Lake sectionand at or near the PWQO below Moira Lake. Fluctuations appear to be related to flow; however,changes in analytical detection limits are another important factor.

Copper, lead and nickel concentrations appear to have declined with time; however, this may bean artifact of decreasing detection limits (Figures 2.1-2, 2.1-4, 2.1-5, and 2.1-6). Reference sitedata also show a decline over time (see station W1 in Figures 2.1-4, 2.1-5 and 2.1-6), lendingsupport to the assumption that declining detection limits, rather than any true decrease, areproducing the trend observed.

Seasonal Variation

Seasonal variation in metal concentrations was quite pronounced, particularly for arsenic. Peakconcentrations occurred from mid-summer through the fall when river flows were lowest. Since the mid-1980's, exceedances of the PWQO only occurred during this mid summer-fall period.This seasonal trend was evident in all of the years of record (1965-1999). It was morepronounced at the stations in the portion of the study area from the Highway 7 bridge to BendBay. However, the trend was evident at other stations. Examples of this seasonal variation aregiven in Figures 2.1-8 to 2.1-11.

Figure 2.1-8

1970 – Seasonal Variability in Arsenic Levels in Moira River at Hwy 7 Bridge (W2)

Figure 2.1-9

1980 – Seasonal Variability in Arsenic Levels in Moira River at Hwy 7 Bridge (W2)

Figure 2.1-10

1990 – Seasonal Variability in Arsenic Levels in Moira River at Hwy 7 Bridge (Site W2)

Figure 2.1-11

1999 – Seasonal Variability in Arsenic Levels in Moira River at Hwy 7 Bridge (Site W2)

The concentrations of the other metals of concern also varied seasonally; however, the late summer-early fall peaks were not as pronounced because concentrations were not as elevated (Appendix I). Lead was the only metal of concern that did not have a definite seasonal pattern.The concentrations of lead were the lowest of all of the metals of concern.

There were insufficient data to discuss spatial or temporal trends in silver concentrations. Silver exceeded the PWQO downstream of Stoco Lake during the one time that it was measured (1987).

Temperature and Dissolved Oxygen Profiles

Consecon Lake was the only lake with evidence of temperature stratification; however, oxygendepletion was observed in the deepest portions of Consecon, Moira and Stoco lakes (Appendix I).Dissolved oxygen concentrations fell sharply at depths below the thermocline in Consecon Lake, with anoxic conditions from 11 metres to the bottom of the profile at 16 metres (Appendix I). Although temperature was fairly uniform with depth in Moira Lake, dissolved oxygen concentrations declined below 7 metres and conditions just above the bottom at 9 metres wereanoxic (Appendix I). Temperature declined with depth in Stoco Lake, although a definite thermocline was not observed. The greatest change in temperature occurred within 1 metre above the bottom at 8.5 metres. Dissolved oxygen started to decline at 6 metres in Stoco Lake and conditions were anoxic below 8 metres. Temperature and oxygen conditions in Round Lake werefairly uniform with depth. Temperature declined by about 1°C between 1 meter and the bottom of the profile at 8 metres, while dissolved oxygen concentrations ranged from 7.2 to 8.7 mg/L (Appendix I). These results are similar to temperature and oxygen conditions observed inprevious monitoring by the MOE, the Ministry of Natural Resources, and the Quinte, Crowe Valley,Cataraqui, Napanee and Lower Trent Conservation Authorities.

The oxygen depletion observed in the deepest portion of three of the four lakes is likely related to oxygen consumption by decomposers in the organic-rich sediments of these lakes. This oxygendepletion will affect metal speciation. Effects on the benthic community and the sentinel fishspecies of each lake will depend on whether the oxygen depletion is transient within each season, whether it occurs every year, and the spatial extent of the depletion relative to the total area of the lake bottom. This study did not include a detailed examination of oxygen conditions. However,the fact that Round Lake appears to be the only lake without oxygen depletion was a cause forconcern with respect to interpretation of differences between reference and exposed populations and communities. Therefore, the overall conclusions for benthic invertebrates and sentinel fishwere carefully examined to ensure that they were the same with and without using Round Lake as a reference location.

References

Bowen, G. 1994. Internal Memoranda. Environmental Sciences and Standards Division. Ontario Ministry of the Environment.

Ontario Ministry of Environment and Energy (MOE). Provincial network water quality monitoring (database). 1964 to 1999.

Ontario Ministry of Environment and Energy (MOE). 1999. Water management policies: guidelines provincial water quality objectives. Queen's Printer for Ontario.

SCIMUS. 1999. Deloro Mine Rehabilitation Project: extent and character of radioactive materials. Prepared for Ontario Ministry of the Environment by SCIMUS Inc., Toronto, Ontario.

2.2 Sediments

2.2.1 Methods

Sediment quality data were obtained from the Ministry of Environment (MOE) In-PlacePollutants Program. Sediment grab samples were collected in the spring of 1999 from 66 locations as identified in Figure 2.2-1 and Table 2.2-1.

Most locations were selected as representative of the deeper areas of sediment accumulation within the lakes and Moira River. In addition, some grab samples were collected in near-shorebeach areas in the lakes. The metal concentrations from these beach locations plus data from the shallower river stations (<1 m) were considered in the exposure assessment conducted for the Preliminary Quantitative Risk Assessment (PQRA) reported in Chapter 3 of this study. The Moira River above the inlet of Moira Lake is generally fast flowing and characterized by arocky substrate. Areas of sediment accumulation are localized and where observed, consist offine-grained materials deposited in shallow backwater areas. In general, these areas of sediment accumulation are not attractive to recreational users because of the fine-mud nature of the bottomand the relatively shallow depth of the water. Sandy beach areas were not noted and only a fewcottages were observed in this section of the river during the sediment sampling field program. Ready public access to this section of the river is limited to road crossings at Highway 7 and theOld Marmora Road in the vicinity of Deloro. Therefore, there is limited opportunity for exposure by local residents to shallow-water sediments in this area of the Moira River.

Young's Creek, which joins the Moira River just below Highway 7, contains a significant volume of sediments that have accumulated behind a series of beaver dams. North of Highway 7, these sediments are known to include radioactive tailings as identified during the assessment of the Deloro Mine Site (SCIMUS 1999). The sediments south of Highway 7 were also expected to contain metals derived from the Deloro Mine Site; however, radionuclide content may be low because SCIMUS (1999) found that radioactivity declined to background levels upstream of Highway 7.

---

1 The Malone station (W1) is located upstream from Deloro Mine Site. Between Deloro and Moira Lake are (in descending elevation): Highway 7 Bridge (DM7/W2), Young's Creek (DM1) and Downstream Young's Creek (DM8). Station W6 is located in Moira Lake at the Highway 62 Bridge that crosses the lake. Downstream of Moira Lake are (in descending elevation): Stoco Lake inlet (W11), Stoco Lake outlet (W15), Stoco Lake Bridge (W17) and Belleville Highway 2 (W23).

---

Figure 2.2-1 MOIRA RIVER STUDY 1999 SEDIMENT QUALITY SAMPLE LOCATIONS

Table 2.2-1
Identification and Description of Sediment Core and Grab Stations

* Grab samples were also submitted for sediment toxicity testing.

Sediment core samples were collected in August of 1999 at 11 locations that overlap with 11 of the sediment grab stations (Figure 2.2-1 and Table 2.2-1). Sediment core samples were kept cool, but not frozen and shipped to Philip Analytical Services Corporation for analysis. Sediment grab samples were analyzed by the MOE for metals, major nutrients and particle sizes. MOE Standard Reference Methods for sediment samples were used to analyze metals (Appendix II). MOEidentifies their standard method procedures with an identification code for each analysis. Inductive coupled plasma (ICP) was used for the analysis of trace metals in sediments (Reference I.D. - E3062A). The hydride metals such as arsenic was analyzed with atomic absorption (AA) hydride generation (Reference I.D. - E3245A). Mercury was analyzed with cold vapour AA (Reference I.D. – E3059A). Finally, uranium concentrations in sediment were analyzed with an ICP mass spectrometer instrument. Through-out the chemical analysis, a 15-20% quality control (QC) program consisting of NIST (National Institute of Standards and Technology) standards, duplicates, blanks and controls was used. MOE verifies all QC samples to authorize the accuracy and precision of the analytical instruments and, thus, the reliability of results.

The top 0-5 cm portion of the sediment core samples was analyzed by Philip Analytical Services for metals using a sequential extraction (Tessier et al. 1979). The sequential extraction procedure partitions metals into five fractions that correspond to binding of metals to different substances within the sediments. The five fractions, in order of increasingly tight bonds to substrates are:

• Exchangeable metals:
  • Extracted with either magnesium chloride or sodium acetate.
  • This fraction represents metals that are most likely to desorb and re-enter the water column.
• Metals bound to carbonates:
  • Leached with sodium acetate adjusted to pH 5 with continuous agitation.
  • This fraction represents metals that desorb with changes in pH.
• Metalsbound to organic matter:
  • Extracted with nitric acid and peroxide plus ammonium acetate to prevent adsorption of extracted metals.
  • This fraction represents metals bound to various forms of organic matter such as living organisms, dead and decaying organisms, coatings on mineral particles, etc. Under oxidizing conditions, organic matter can be degraded, leading to release of soluble metals.
  • It should be noted that the organic fraction also includes sulphides, which also release metals if redox conditions are altered.
• Metals bound to iron and manganese:
  • Extracted with either sodium citrate or ammonium hydroxide.
  • Metals bound to iron or manganese oxides will be released under anoxic conditions.
• Residual metals:
  • Digested with hydrofluoric and perchloric acids.
  • This fraction represents metals bound in primary and secondary minerals, held within the crystal structure. These metals are not expected to be released into solution under conditions normally encountered in nature.

The information on the fractions of metals in the top five centimetres of sediment provided an indication of the relative bioavailability of metals in the Moira River system. For example, metals in the "exchangeable" and "carbonates" fractions are more readily released back into solution; therefore, these fractions can be assumed to be more available for uptake by aquatic biota. In contrast, metals in the "iron and manganese oxides" fraction are released only under anoxic conditions; therefore, metals in this fraction are not as available for uptake by aquatic biota in oxygenated conditions. Metals in the "residual" fraction can be assumed to be unavailable for uptake by aquatic biota. That the sequential extraction method used is not specific for sulphide-bound metals (instead, they are represented within the organic fraction). More recent studies have shown that the toxicity of some metals in sediments (notably cadmium, copper, lead, nickel, and zinc) can be reliably predicted using metal-sulphide relationships and/or pore-water metal concentrations (Ankley et al 1996). Acid-volatile sulphide concentrations were not measured in this study.

The 5-10, 10-20 and 20-30 cm portions of the cores were analyzed for total metals (Powell, pers. size. comm). The core sections were also analyzed for total organic carbon and particle size.

In 1998 and 1999, sediment samples from various stations in the Moira River system were collected and submitted to the Ministry of Labour, Radiation Protection Monitoring Service for radionuclide analysis. During the spring of 1999, 64 sediment samples were collected by MOE from the locations identified in Figure 2.2-1 for analysis of the following:

• Moisture;
• Radium-226;
• Radium-228;
• Thorium-228; and
• Uranium-238.

Samples were counted by gamma spectrometry using high purity Germanium detectors for 10,000 seconds each. Appendix II provides additional details on QA/QC, calibration of gamma detectors and determination of gamma emitters taken by the Radiation Protection Laboratory.

Field techniques were the standard MOE practices (Jaagumagi, pers. comm.). Sediment grab samples were sub-samples of three individual replicates that were mixed with a stainless steel spoon in a mixing bowl until an even colour and consistency were observed. The sediment core samples were processed similarly, except samples were composed of 4 to 5 individual replicates. The sediment grab samples taken from lakes were collected with a standard, full-size Ponar grab. Grab samples from river stations were collected with a 9" x 9" x12" stainless steel Ekman grab. Sediment core samples were collected with a stainless steel K-B Corer (Wildco) fitted with 2" plexiglas or a Benthos Gravity corer (custom-made) fitted with 3" plexi tubes. The sampling instruments, mixing bowls, and sampling handling gear were cleaned with lake or river water between each sample. Any residue was scrubbed off the device with a brush, and the device was then rinsed thoroughly.

Sample containers were provided by MOE and pre-cleaned using MOE standard practices (Powell, pers. comm.). There was no preservation required for sediment samples. At all times during the field-sampling program, field notes were maintained, and photos were taken at nearly every site sampled. All samples were labelled with the project name, station number, collection date and kept in a cooler for transport to a refrigerator at 4 degrees C until submission to the date and kept in a cooler for transport to a refrigerator at 4 degrees C until submission to the laboratory. No field or travel blanks were collected during the sediment sampling program. Sampling stations were located with a hand held GPS instrument, and samples were collected from areas were there was a suitable volume of soft sediments for analysis.

2.2.2 Identification of Metals of Concern

The same metals of concern determined for water quality were used for sediment quality; however, uranium was added to the list because of the potential for migration of uranium from the "red mud tailings" at the Deloro Mine Site. Therefore, the metals of concern for sediments were:

• Arsenic
• Cobalt
• Copper
• Lead
• Nickel
• Silver
• Uranium.

Uranium is both a metal and a radionuclide. The "daughter products" produced by the radioactive decay of uranium (such as radium-226) are not as mobile in the environment as uranium. This was confirmed by the very low levels of radium-226 found in sediments collected for this study.

With the exception of uranium, radionuclide concentrations in reference sediments as well as downstream sediments, were consistently low. Radium-226, radium-228 and thorium-228activities were less than or equal to detection limits at all stations, with the exception of one sample from Round Lake, one sample from Stoco Lake and Young's Creek (Appendix II). The Round Lake and Stoco Lake samples contained 0.02 Bq/g radium-226 (detection limit is 0.01 Bq/g). The Young's Creek station had 0.05 Bq/g radium-226.

2.2.3 Application of Ontario Sediment Quality Guidelines for Metals

The metal concentrations observed in the sediments from the Moira River system were compared to Provincial Sediment Quality Guidelines (PSQGs). Four out of the seven metals of concern have PSQGs (Table 2.2-2). The comparison with guidelines served as one indication of the potential for toxicity to aquatic biota. The MOE Guidelines (1993) state that if metal concentrations exceed the Lowest Effect Level (LEL), then there may be adverse effects on some aquatic biota. If metal concentrations exceed the Severe Effects Level (SEL), then there may be significant effects on aquatic biota.

As noted in Section 1, the PSQGs were developed for the protection of aquatic life. The PSQGs are used to indicate the need for additional biological testing and field assessment of the benthic invertebrate community. They are not meant to trigger remedial action. In the case of this study, sediment toxicity testing and benthic invertebrate community assessments were already part of the study design. Therefore, the sediment chemistry data were one component of a threecomponent assessment of the effects of the metal concentrations observed in the Moira River system.

Table 2.2-2
Provincial Sediment Quality Guidelines

Cobalt, uranium and silver do not have PSQGs. Screening Level Concentrations (SLCs) were derived for cobalt and silver according to the methods outlined in MOE (1993). An SLC for uranium could not be derived because the dataset was too small.

The site-specific development of SLCs is an effects-based approach that is used to develop sediment quality guidelines that are protective of benthic organisms. Guidelines are developed from the co-occurrence of benthic fauna and a wide range of chemical concentrations. The SLC is an estimate of the highest concentration of a metal that can be tolerated by 95% of the benthic fauna observed from the sampling locations.

The SLC was developed through a three-step process incorporating species from lake and river sites. From the list of benthic fauna recorded (Rivers: 117, Lakes: 68) in field investigations, a list of species (at least 10) that were present at 10 or more sites was compiled. For each species, the associated metal concentrations were found for each site where the species was present and the 90th percentile was determined. This represented the species screening level concentration the 90th percentile was determined. This represented the species screening level concentration (SSLC) that 90% of the population appeared to tolerate based on field data. The top 10% was removed from further analysis to produce a more conservative estimate and eliminate extreme sediment concentrations that may be a result of physical and/or chemical sediment characteristics. The SSLC was determined for each metal and species and plotted to determine the 5th percentile. This represented the screening level concentration (SLC) which 95% of the species found in the Moira River system and reference locations could tolerate.

Screening level concentrations were determined for lake and river communities combined. This approach was considered the most conservative. The SLC guidelines for lake and river sediments combined are presented in Table 2.2-3.

The SLCs for cobalt and silver were derived from a limited data set. Therefore, they are only applicable to the Moira River system and cannot be considered analogous to PSQGs.

Table 2.2-3
Screening Level Concentrations (SLC) for Lakes and Rivers (µg/g)

Interpretation of exceedances of the SLCs is the same as for exceedances of the PSQGs; i.e., they serve as one indicator of the potential for adverse effects on benthic invertebrates. They cannot be used as "triggers" for remedial action in the Moira River system.

Uranium concentrations in sediments downstream of the Deloro Mine Site were compared to reference concentrations at reference sites (upstream of the Deloro Mine Site, Black River, Skootamatta River, Salmon River, Round and Consecon lakes).

2.2.4 Results and Discussion

2.2.4.1 Sediment Grab Samples

Metal concentrations in all sediment grab samples from reference sites were consistently low compared to the Moira River system from the Deloro Mine Site to the Moira Lake outlet. This was true in the lakes (Consecon and Round lakes) and the rivers (Salmon, Black and Skootamatta rivers) (Figures 2.2-2 to 2.2-7). Arsenic concentrations were below LELs at all reference locations except in the Moira River system downstream of Malone (MR-3). Copper, lead and nickel concentrations exceeded LELs in Round Lake and the Skootamatta River. Reference concentrations of all the metals of concern were similar to concentrations reported from other Ontario reference locations (Mason and Dragun 1996).

Figure 2.2-2
Total Arsenic in Bottom Sediment Grab Samples (Excluding Near-Shore) – Moira River
Watershed and Reference Sites, 1999¹

Horizontal lines indicate Severe and Lowest Effect Level from Ontario Sediment Quality Guidelines. ¹Reference lake and mid-basin Moira Lake and Stoco Lake results are averages of two or more stations.

Figure 2.2-3
Total Cobalt in Bottom Sediment Grab Samples (Excluding Near-Shore) – Moira River
Watershed, 1999¹

Horizontal lines indicate Severe and Lowest Effect Level from Ontario Sediment Quality Guidelines. ¹Reference lake and mid-basin Moira Lake and Stoco Lake results are averages of two or more stations.

Figure 2.2-4
Total Copper in Bottom Sediment Grab Samples (Excluding Near-Shore) – Moira River
Watershed, 1999¹

Horizontal lines indicate Severe and Lowest Effect Level from Ontario Sediment Quality Guidelines. ¹Reference lake and mid-basin Moira Lake and Stoco Lake results are averages of two or more stations.

Figure 2.2-5
Total Lead in Bottom Sediment Grab Samples (Excluding Near-Shore) – Moira River
Watershed, 1999¹

Horizontal lines indicate Severe and Lowest Effect Level from Ontario Sediment Quality Guidelines. ¹Reference lake and mid-basin Moira Lake and Stoco Lake results are averages of two or more stations.

Figure 2.2-6
Total Uranium in Bottom Sediment Grab Samples (Excluding Near-Shore) – Moira River
Watershed, 1999¹

Horizontal lines indicate Severe and Lowest Effect Level from Ontario Sediment Quality Guidelines. ¹Reference lake and mid-basin Moira Lake and Stoco Lake results are averages of two or more stations.

Figure 2.2-7
Total Silver in Bottom Sediment Grab Samples (Excluding Near-Shore) – Moira River
Watershed, 1999¹

Horizontal lines indicate Severe and Lowest Effect Level from Ontario Sediment Quality Guidelines. ¹Reference lake and mid-basin Moira Lake and Stoco Lake results are averages of two or more stations.

Figure 2.2-8
Total Uranium in Bottom Sediment Grab Samples (Excluding Near-Shore) – Moira River
Watershed, 1999¹

¹Reference lake and mid-basin Moira Lake and Stoco Lake results are averages of two or more stations.

Metal concentrations from the three reference sites upstream of the Deloro Mine Site in the Moira River were similar to the other reference concentrations, except for arsenic, lead and uranium. Arsenic concentrations at the upstream station closest to Deloro (MR-3) were higher than at any of the other reference sites (Figure 2.2-2). Lead concentrations at all of the upstream stations were lower than lead concentrations found in the reference lakes and the Salmon River (Figure 2.2-5). Uranium concentrations at all of the upstream stations were lower than uranium concentrations found in Round Lake, the Salmon River and the Skootamatta River (Figure 2.2-8).

Metal concentrations in sediment grab samples followed a fairly consistent trend from upstream-to downstream in the Moira River system (Figures 2.2-2 to 2.2-8). A substantial increase usually occurred downstream of the Deloro Mine Site, followed by an additional increase downstream of the confluence with Young's Creek. The highest concentrations were observed in the one sediment sample collected from Young's Creek (YC-1). Metal concentrations then gradually declined, with the most substantial drop occurring after the Moira Lake outlet. Arsenic, cobalt and nickel concentrations remained elevated above reference values all the way to the farthest downstream station (MR-16: near Corbyville). Copper, lead and uranium concentrations fluctuated after the outlet of Moira Lake, but on the whole were more similar to reference concentrations than to the concentrations in the Deloro-to-Moira Lake section of the system.

Some of the fluctuations in metal concentrations from site-to-site may be related to differences in total organic carbon (TOC) or particle size, since both of these parameters are often correlated with total metal content in sediments (Håkanson 1992, Mahony et al. 1996). For example, Round Lake sediments (excluding beach samples) had higher TOC concentrations than in Consecon Lake (Appendix II, page II-3). Metal concentrations in Round Lake sediments were usually higher than in Consecon Lake (Figures 2.2-2 to 2.2-8), which may reflect a higher metal-bindingcapacity in Round Lake sediments because of the higher organic content. Moira Lake sediments had higher TOC concentrations than Stoco Lake sediments (Appendix II, page II-3). Therefore part of the explanation for lower metal concentrations in Stoco Lake may be a lower metal-binding capacity. The relationship between TOC and metal concentration was confirmed during capacity. The relationship between TOC and metal concentration was confirmed during statistical analyses in support of the interpretation of benthic invertebrate data (see Tables 2.3-5 and 2.3-13). Particle size did not differ consistently among sites (Appendix II, page II-7), and few statistical relationships were found between particle size and metal concentrations (see Section 2.3). The exception to this would be in the coarser-grained beach samples (see discussion below).

The SELs or SLCs for the metals of concern were usually exceeded at the stations from Highway 7 (Station MR-7) through to Moira Lake. The arsenic and nickel SELs were exceeded in the sediment grab samples from Highway 7 to the Moira Lake outlet (Figures 2.2-2, 2.2-4 and 2.2-6). Arsenic and nickel also exceeded the SEL at some locations in Stoco Lake. Copper exceeded the SEL at Young's Creek and in three stations within Moira Lake. The cobalt SLC was exceeded from Highway 7 to the Moira Lake outlet (Figure 2.2-3). Silver usually exceeded the SLC at stations from Highway 7 to mid-basin of Moira Lake (Figure 2.2-7).

The pattern of lead concentrations relative to the PSQG was different from all of the other metals of concern (Figure 2.2-5). The lead SEL was not exceeded at any location. Although lead concentrations downstream of the Deloro Mine Site were greater than upstream of the site, concentrations far downstream (e.g., at Corbyville) were similar to concentrations in Round Lake and the Skootamatta River. Therefore, lead in sediments may reflect other sources (such as local geology or inputs from leaded gasoline).

Uranium concentrations in the sediments of Young's Creek, mid-basin Moira Lake, Stoco Lake North, Stoco Lake outlet and Moira River near Roslin were higher than reference concentrations (Figure 2.2-8). This shows that some historical migration of uranium has occurred. The Young's Creek sediments have the highest concentrations, as would be expected given their proximity to the –red mud tailings - source at the Deloro Mine Site. The uranium concentrations in the lake basin sediments and in the Moira River near Roslin were only slightly elevated above concentrations found in Round Lake (one of the reference lakes), and may simply reflect the natural variability of uranium concentrations in the study area.

The results of the two methods used to measure uranium in the sediment grab samples were compared to ensure that the results were in reasonable agreement. Uranium was measured as a metal in µg/g using ICP-MS analysis, and as a radionuclide in Bq/g using gamma spectroscopy (Appendix II). The two results were compared by converting Bq/g to µg/g using the specific activity of ^natU. The composition of ^natU is 99.28% ²³⁸ U, 0.75% ²³⁵U and 0.0058% ²³⁴U (Eisenbud 1987). The specific activity of ^natU is 1.2 x 104Bq/g (Eisenbud 1987). All of the gamma spectroscopy results were less than the detection limit of 0.10 Bq/g. All but the Young's Creek samples also had very low results using ICP-MS analysis, ranging from 0.26 to 1.7 µg/g. These ICP-MS results would correspond to activities of < < 0.1 Bq/g; therefore, the two results appear to be comparable. Young's Creek sediments contained 19 µg/g, which converts to 0.23 Bq/g using the specific activity of ^natU. The discrepancy between 0.23 Bq/g and < 0.1 Bq/g is small, given laboratory precision when radionuclide activities are close to the detection limit.

The potential for effects on aquatic biota from the observed uranium concentrations cannot be assessed with the chemistry data alone because there are no sediment quality guidelines for uranium and because there are insufficient data to calculate Screening Level Concentrations. The benthic invertebrate community assessment and the monitoring of sentinel fish species (Sections 2.3 and 2.5) provide information on the status of aquatic biota exposed to elevated metal concentrations in water and sediment.

The potential for effects on aquatic biota from the radioactive emissions of uranium can be assessed using the measurement of uranium in activity units (Becquerels or Bq). As noted above, all uranium activities in sediments were less than the detection limit of 0.1 Bq/g. Thus, radioactivity levels in sediments downstream of the Deloro Mine Site were indistinguishable from levels upstream of the Deloro Mine Site and at other reference locations. These radioactivity levels are far below any levels that have been shown to cause effects in aquatic biota (by several orders of magnitude) (IAEA 1992).

The metal concentrations in river sediments collected in shallow water, at water depths of 1 m or less, were in general, less than the metal concentrations reported from sediments collected from deeper water (Table 2.2-4). This trend was most pronounced between the Deloro Mine Site and Moira Lake as is shown below. Metal concentrations below Stoco Lake were lower and the differences between the deep and shallow sediment grab samples were much less distinct.

Table 2.2-4
Comparison of Shallow (<1 m) and Deeper (>1 m)Sediment Metal Concentrations
in the Moira River

Note: All concentrations in µg/g.

There were also substantial differences in sediment metal concentrations between deeper lake samples and samples taken from near-shore beach areas (Table 2.2-5). The near-shore samples were taken from 1 m and 2 m depths. Differences in metal concentrations are likely attributable to the coarser nature of the substrates in near-shore areas, as well as lower TOC content.

Arsenic concentrations in the Moira River system sediments in 1967 were generally higher than concentrations observed in 1999 (Owen and Galloway 1969). For example, arsenic concentrations upstream of Bend Bay varied from 700 - 870 µg/g in 1967 and were 260 µg/g in 1999 (stations MR-9 and MR-10, Figure 2.2-2). Concentrations in Moira Lake sediments in 1967 varied from 50 µg/g (at the inlet) to 1000 µg/g (west basin). Concentrations in Moira Lake sediments in 1999 varied from 8.8 µg/g to 600 µg/g (Appendix II; Sediment Grab Data). Stoco Lake sediments contained 85 - 220 µg/g in 1967 and 2.1 - 130 µg/g in 1999 (Appendix II; Sediment Grab Data).

Table 2.2-5
Average Metal Concentrations in Near-Shore Area Samples Versus Deeper Samples in Lakes

* Result based on semi-quantitative method.

The only other study with comparable sediment grab data was in 1996 (Golder 1996). Concentrations at Moira River sites closer to the Deloro Mine Site were similar to these previous data. Samples taken from close to Highway 7 and in Young's Creek in 1996 produced arsenic concentrations ranging from 25-150 µg/g (Highway 7) to 1500 µg/g (Young's Creek) (Golder 1996). These 1996 results are similar to the 1999 data. Analytical methods were the same in the two studies.

2.2.4.2 Sediment Core Samples

The same overall pattern of metal concentrations with distance downstream of the Deloro Mine Site was evident in the sediment core data (Figures 2.2-9 to 2.2-15). Concentrations were usually highest in Young's Creek (YC-1) and declined substantially after Moira Lake (SL-2, SL-4, SL-5 and MR-13 (Moira River near Roslin).

The sediment profiles from this study show a variation in the relative concentration of metals at the surface (0-5 cm) compared to deeper sediments (Figures 2.2-9 to 2.2-15). At the station in Young's Creek (YC-1), the highest concentrations of arsenic and nickel were found in the upper 5 cm of the profile. This suggests the continuing contribution of these substances from the Deloro Mine Site, at least locally. At some of the lake stations (ML-3, ML-4, SL-5), the concentrations of some metals in the upper 5 cm were lower than in the underlying sediments (presumably reflecting the reduced metal inputs to this section of the Moira system since refining activities at the Deloro Mine Site were discontinued). However, the pattern at these stations was not consistent with metal or with depth of core. Furthermore, in other lake stations (ML-1, SL-2, SL-4) the concentrations in the upper 5 cm were higher than in the underlying sediments. This lack of consistency may reflect the effects of bioturbation (sediment disturbance caused by burrowing invertebrates), diffusion of metals both into the water and deeper into the sediments, and the recent deposition of sediments with lower metal concentrations.

There was a trend to decreasing concentrations with depth below 5 cm for some samples collected from Moira and Stoco lakes as well as from Young's Creek. This was most noticeable for the profiles developed for arsenic and cobalt at YC-1, and arsenic, copper, lead and nickel at ML-1, ML-4 and SL-2 (Figure 2.2-9 to 2.2-13). These results suggest that these areas have been relatively undisturbed. However, for some of the lake sampling stations, (e.g., ML-3 and SL-5), and the river stations, the pattern with depth was inconsistent. For these locations, bioturbation, diffusion, and sediment resuspension and deposition may have disturbed the profile of sediment accumulation.

The concentrations of arsenic and nickel exceed the SELs in the lower-most section of most of the core samples (Figure 2.2-9 and 2.2-13). This indicates that materials from the Deloro Mine Site are present to depths below 30 cm, the maximum depth of the sediment cores collected in this study. This contrasts with the results developed by Mudroch and Capobianco (1980) that show elevated metals in sediments were confined to depths of about 25 cm or less in the one core sample collected from Moira Lake.

Figure 2.2-9
Profile of Total Arsenic in Sediment Cores – Moira River Watershed, 1999

Concentration in Sediment (µg/g dry weight)

Values for 0-5 cm depth represent sum of results from sequential extraction procedure.

Profile of total arsenic in sediment cores; Moira River watershed, 1999.

Figure 2.2-10
Profile of Total Cobalt in Sediment Cores – Moira River Watershed, 1999

Values for 0-5 cm depth represent sum of results from sequential extraction procedure.

Profile of total cobalt in sediment cores; Moira River watershed, 1999.

Figure 2.2-11
Profile of Total Copper in Sediment Cores – Moira River Watershed, 1999

Concentration in Sediment (µg/g dry weight)

Values for 0-5 cm depth represent sum of results from sequential extraction procedure.

Figure 2.2-12
Profile of Total Lead in Sediment Cores – Moira River Watershed, 1999

Concentration in Sediment (µg/g dry weight)

Values for 0-5 cm depth represent sum of results from sequential extraction procedure.

Figure 2.2-13
Profile of Total Nickel in Sediment Cores – Moira River Watershed, 1999

Concentration in Sediment (µg/g dry weight)

Profile of total nickel in sediment cores; Moira River watershed, 1999.

Values for 0-5 cm depth represent sum of results from sequential extraction procedure.

Figure 2.2-14
Profile of Total Silver in Sediment Cores – Moira River Watershed, 1999

Concentration in Sediment (µg/g dry weight)

Values for 0-5 cm depth represent sum of results from sequential extraction procedure.

Figure 2.2-15
Profile of Total Uranium in Sediment Cores – Moira River Watershed, 1999

Concentration in Sediment (µg/g dry weight)

Values for 0-5 cm depth represent sum of results from sequential extraction procedure.

Mudroch and Capobianco (1980) cored the sediments in the west basin of Moira Lake to a depth of 92 cm and recovered a single core. This core was subsampled into 1 cm sections and analyzedfor arsenic, cobalt, nickel and copper concentrations. The profiles developed indicated elevated concentrations of arsenic to a depth of about 25 cm, cobalt and nickel to a depth of about 20 cm copper to a depth of about 15 cm. Peak concentrations were reported at the 5 cm depth of the profile for all four metals with a m arked decrease observed at 0-4 cm for cobalt, nickel andcopper. Arsenic concentrations remained relatively constant in the upper 5 cm of the profilereflecting the elevated concentrations of arsenic in the water column. The authors suggested that the lowest depth at which elevated metals concentrations are observed corresponds to the time around 1900 when the Deloro Mining and Smelting Co. began operations. They suggested that the reduction in concentrations observed in the upper 5 cm of the profile corresponds to about 1960 when refining activities ceased.

Other sediment core studies conducted in Moira Lake have reported maximum concentrations from varying sediment depths. Azcue and Nriagu (1993) reported that the highest arsenic concentrations were in the deeper sediment layers (> 20 cm) of Moira Lake. Cornett et al (1987) reported maximum arsenic concentrations at 10 cm and maximum nickel concentrations at 5 cm in Moira Lake. Cornett and Chant (1986) reported maximum arsenic and nickel concentrations in the 10-20 cm sediment layer in Moira Lake.

The range of concentrations found in previous sediment cores from Moira Lake appear similar to those found in this study. For example, the average arsenic concentration in the West and East Basins of Moira Lake was 623 µg/g and 723 µg/g, respectively (Azcue and Nriagu 1993) and Mudroch and Capobianco (1980) reported arsenic concentrations in Moira Lake ranging from 850 to 1237 &microg/g. In this study, arsenic ranged from 175 to just over 1000 µg/g in Moira Lake cores. It should be noted that Azcue and Nriagu performed sequential extractions on all sections of the sediment cores. Therefore, the total arsenic concentrations of Azcue and Nriagu cannot be directly compared to the results of this study, except in the 0-5 cm layer. This is because the hydrofluoric/perchloric acid digestion used in the sequential extraction procedure (for the "residual fraction") removes more metal from the sample materials than the nitric acid digestion used for the total metal analysis in this study. Mudroch and Capobianco determined arsenic by X-ray fluorescence, whereas in this study, arsenic was determined by atomic absorption hydride generation. Therefore, a direct comparison between the arsenic concentrations in the two studies is not possible.

In summary, the metal profiles in the sediment cores from this study illustrate a trend to decreasing concentrations with depth at some of the locations sampled for some, but not all, of the metals of concern. The sediment profile may have been disturbed by bioturbation, diffusion and sediment resuspension and deposition processes. The depth of metals originating from the Deloro Mine Site extends below the base of the 30 cm depth cored. The relative concentrations of arsenic in sediment observed in this study appear similar to concentrations reported in earlier studies for the same general locations; however, because different analytical methods were employed in the earlier studies, a direct comparison of the metal concentrations in sediment from the earlier studies cannot be made.

Sequential Extraction Results

Metals are partitioned in sediments in many forms, as soluble free ions, soluble organic (lowmolecular- weight humic) and inorganic complexes, easily exchangeable ions, precipitates of metal hydroxides, precipitates with colloidal ferric and manganic oxyhydroxides, insoluble organic complexes, insoluble sulphides, and residual forms (Burton 1992). The residual fraction serves as the matrix vehicle and is associated with labile components (e.g., carbonates, amorphous aluminosilicates, organic matter) that are coated with iron/manganese oxides and organic matter (Burton 1992). This variable coating serves as an active sorption site for metals. Free metals (soluble free ions) are generally thought to possess the greatest toxicity to aquatic organisms; therefore, it is important to understand the relative proportion of metals among various fractions in sediments.

Predicting the partitioning of metals (and, thus, bioavailability) is difficult because of the number and complexity of processes that may simultaneously reduce and increase availability, depending upon the metal. The concentration of metal in porewater of sediments is dependent, to a large degree, on sorption/precipitation processes. The process depends on the metal and the environment. Adsorption is complicated, being related to the solid type, concentration,adsorption species, and surface property changes resulting from interactions such as coagulation (Burton 1992). In addition, sorption-site competition and reaction kinetics of constituents are unknown. High dissolved organic matter concentrations enhance the solubility and complexation of metals. High acid-volatile sulphide concentrations bind cationic metals, reducing bioavailability (Ankley et al. 1996, Berry et al. 1996). Bioturbation can result in the oxidation of acid-volatile sulphides, thereby releasing some cationic metals back into porewater and increasing bioavailability (Peterson et al 1996).

Mobilization of the metals of concern in this study has been associated with several processes, including lowered pH, reducing conditions, changes in the oxidation-reduction conditions of the sediments and the degradation of organic complexes (Burgess and Scott 1992). Given the conditions in the Moira River system, the latter three processes are the most likely to be active. In particular, the deeper depositional areas of the Moira River (such as Bend Bay) and the lake basins are likely to experience seasonal changes in redox conditions. These locations also have greater organic content in the sediment; therefore, these locations contain more organic substrates for metal sorption. However, they would also be more prone to the actions of decomposer organisms, which would degrade organic complexes (such as organo-arsenicals).

Because of the complexity of the processes governing bioavailability, the following results of the sequential extractions should not be interpreted as accurate representations of bioavailability of the metals of concern in the Moira River system. However, they are a reasonable first step in estimating the partitioning of the metals of concern among several important fractions known to be linked to processes that affect bioavailability.

The percent distribution of metals among chemical fractions in the 0-5 cm sediment layer varied with both metal and sample site.

Arsenic was found primarily in the residual fraction at all sites, followed by the iron and manganese oxide fraction (Table 2.2-6). These two fractions accounted for 70 to 88% of the arsenic in the surface sediment layer (0-5 cm). The high proportion in the residual fraction may indicate that most of the arsenic occurs as granular (tailings) material in the Moira River system.

Cobalt was more evenly distributed among the fractions, with approximately equal proportions in the oxide, organic/sulphide and residual fractions in Moira Lake (ML-1, ML-3, ML-4) (Table 2.2-6). For most of the river stations, the carbonate fraction was important, along with the oxide, organic matter and residual fractions. Cobalt at the Young's Creek site (YC-1) was found primarily in the iron and manganese oxide fraction. The variable speciation of cobalt may be due to leaching and runoff processes rather than deposition of mineralized material.

Table 2.2-6

% Distribution of Metals to the Analyzed Components of Sediments

Parameter	Description	ML-1	ML-3	ML-4	MR-1	MR-13	MR-7	MR-8	SL-2	SL-4	SL-5	YC-1
Ag	Exchangeable Metals	1.7%	5.4%	4.6%	1.9%	23.1%	8.1%	14.3%	23.1%	20.0%	21.4%	7.1%
Ag	Metals Bound to Carbonates	1.7%	5.4%	4.6%	1.9%	23.1%	8.1%	14.3%	23.1%	20.0%	21.4%	7.1%
Ag	Metals Bound to Iron and Manganese	1.7%	5.4%	4.6%	1.9%	23.1%	51.4%	19.0%	23.1%	20.0%	28.6%	7.1%
Ag	Metals Bound to Organic Matter	1.7%	5.4%	4.6%	1.9%	23.1%	27.0%	38.1%	23.1%	33.3%	21.4%	7.1%
Ag	Residual Metals	93.3%	78.6%	81.5%	92.4%	7.7%	5.4%	14.3%	7.7%	6.7%	7.1%	71.4%
Total (0-5cm)		100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
As	Exchangeable Metals	1.9%	6.5%	8.0%	2.9%	9.8%	3.3%	9.1%	8.7%	7.8%	10.7%	1.0%
As	Metals Bound to Carbonates	3.9%	4.4%	2.4%	14.9%	9.8%	7.7%	2.2%	8.7%	7.8%	10.7%	0.4%
As	Metals Bound to Iron and Manganese	22.4%	12.2%	15.7%	19.1%	15.9%	10.1%	4.6%	14.1%	14.6%	18.7%	0.6%
As	Metals Bound to Organic Matter	6.5%	7.4%	7.0%	5.3%	9.8%	11.2%	9.7%	2.2%	4.9%	2.7%	19.2%
As	Residual Metals	65.4%	69.5%	67.0%	57.8%	54.9%	67.7%	74.3%	66.3%	65.0%	57.3%	78.7%
Total (0-5cm)		100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
Co	Exchangeable Metals	3.7%	2.7%	2.4%	0.4%	1.0%	1.4%	0.8%	0.5%	0.7%	0.2%	1.9%
Co	Metals Bound to Carbonates	10.9%	8.7%	16.3%	5.1%	26.4%	35.3%	25.6%	19.2%	20.7%	11.8%	26.3%
Co	Metals Bound to Iron and Manganese	29.1%	27.5%	42.7%	22.1%	32.1%	14.3%	18.8%	30.7%	26.6%	29.1%	48.1%
Co	Metals Bound to Organic Matter	29.1%	27.5%	24.6%	44.1%	23.6%	15.9%	22.2%	26.3%	26.6%	34.1%	17.7%
Co	Residual Metals	27.2%	33.6%	14.0%	28.2%	16.9%	33.0%	32.6%	23.3%	25.4%	24.7%	6.1%
Total (0-5cm)		100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
Cu	Exchangeable Metals	1.2%	6.2%	1.4%	1.8%	1.1%	0.7%	1.6%	1.0%	2.4%	2.6%	5.4%
Cu	Metals Bound to Carbonates	8.3%	24.0%	10.6%	10.3%	11.0%	5.3%	8.1%	12.7%	9.3%	9.1%	10.1%
Cu	Metals Bound to Iron and Manganese	36.9%	42.1%	47.2%	44.3%	48.6%	56.2%	34.9%	58.3%	56.1%	51.0%	36.5%
Cu	Metals Bound to Organic Matter	50.7%	24.8%	37.3%	38.1%	19.9%	6.2%	15.5%	16.2%	17.8%	13.6%	40.9%
Cu	Residual Metals	2.9%	2.9%	3.5%	5.5%	19.5%	31.6%	39.9%	11.8%	14.3%	23.6%	7.0%
Total (0-5cm)		100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
Ni	Exchangeable Metals	32.0%	43.6%	33.0%	45.8%	31.7%	29.5%	35.9%	40.3%	39.4%	46.7%	29.7%
Ni	Metals Bound to Carbonates	10.9%	36.7%	10.3%	26.7%	16.8%	46.7%	38.6%	10.3%	33.7%	26.0%	22.8%
Ni	Metals Bound to Iron and Manganese	57.0%	19.6%	56.6%	27.4%	51.0%	23.5%	25.0%	49.2%	26.6%	26.7%	46.6%
Ni	Metals Bound to Organic Matter	0.0%	0.0%	0.0%	0.0%	0.3%	0.0%	0.1%	0.1%	0.1%	0.3%	0.1%
Ni	Residual Metals	0.1%	0.0%	0.0%	0.1%	0.2%	0.2%	0.5%	0.1%	0.1%	0.3%	0.8%
Total (0-5cm)		100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
Pb	Exchangeable Metals	34.2%	16.8%	36.3%	15.2%	16.4%	6.7%	10.2%	17.9%	16.3%	10.9%	5.9%
Pb	Metals Bound to Carbonates	3.2%	1.8%	4.5%	3.2%	16.8%	3.9%	11.0%	11.8%	8.3%	7.0%	10.9%
Pb	Metals Bound to Iron and Manganese	30.7%	20.7%	36.8%	21.0%	27.1%	19.6%	21.0%	29.7%	21.8%	27.4%	25.6%
Pb	Metals Bound to Organic Matter	19.9%	30.6%	15.9%	34.7%	21.7%	28.4%	27.6%	24.8%	28.9%	34.5%	34.2%
Pb	Residual Metals	12.1%	30.0%	6.5%	25.9%	18.1%	41.4%	30.2%	15.9%	24.7%	20.2%	23.4%
Total (0-5cm)		100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
U	Exchangeable Metals	1.3%	4.4%	2.4%	1.1%	3.7%	0.6%	1.1%	1.5%	1.7%	3.1%	3.8%
U	Metals Bound to Carbonates	5.7%	18.6%	9.8%	5.9%	3.7%	2.6%	3.2%	7.5%	5.9%	3.1%	8.6%
U	Metals Bound to Iron and Manganese	28.6%	39.0%	43.1%	36.5%	33.3%	55.3%	31.2%	43.6%	39.0%	33.8%	35.7%
U	Metals Bound to Organic Matter	58.5%	31.0%	39.4%	42.7%	24.1%	6.1%	17.7%	23.3%	26.3%	21.5%	43.7%
U	Residual Metals	5.9%	7.0%	5.3%	13.9%	35.2%	35.5%	46.8%	24.1%	27.1%	38.5%	8.2%
Total (0-5cm)		100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

Ag = silver; Ar = arsenic; Co = cobalt; Cu = copper; Ni = nickel; Pb = lead; U = Uranium.

Copper was consistently higher in the oxide fraction (37-58%), followed by the organic matter/sulphide or residual fraction, depending upon the sampling site (Table 2.2-6). As for cobalt, leaching and runoff processes may be the primary mechanisms for copper input to sediments in the system.

Nickel and lead distribution among the chemical fractions was variable, with no one fraction consistently dominating (Table 2.2-6). Nickel and lead were the only metals with a significant portion in the exchangeable fraction, ranging from 30-47% (nickel) and 6-36% (lead) (Table 2.2-6).

Silver was primarily associated with the oxides and organic matter/sulphide fractions when concentrations were high enough to obtain meaningful distributions (Table 2.2-6). The exception for silver was Young's Creek, where silver was found primarily in the residual fraction. TheYoung's Creek silver is probably associated with tailings material. The iron and manganese oxides at the river stations likely reflect runoff processes and then exchange with ligands and settling out in depositional areas of the river. settling out in depositional areas of the river.

Uranium was associated with the oxides and organic matter/sulphide fractions at all sites, although the residual fraction was also important at some of the river stations (Table 2.2-6). This may reflect a tailings origin, with resuspension at higher flows followed by attachment to ligands and deposition as oxides or with organic matter in depositional areas.

The data suggest that, with the exception of nickel and lead, the metals are primarily in the unavailable residual fraction or in the oxides fraction, where release of metals requires anoxic conditions. Partitioning into the organic matter/sulphides fraction may also be important in limiting the bioavailability of cobalt, copper, silver and uranium in the Moira River system. Therefore, bioaccumulation by benthic invertebrates and subsequent food chain transfer to fish may be limited by the low bioavailability of the metals of concern. In addition, toxicity to benthic invertebrates may also be reduced because of low bioavailability.

The distribution of metals among the chemical fractions is different from the distribution reported in two earlier studies of Moira Lake sediments (Cornett and Chant 1986; Azcue and Nriagu 1993). Comparisons can be made only for arsenic and nickel because the other metals were not measured in previous work.

The main difference in the speciation of arsenic in this study compared to previous studies is the much higher proportion in the residual fraction. Cornett and Chant (1986) reported that 7-10% of the arsenic from Moira Lake core samples was in the residual fraction; the dominant fraction was iron and manganese oxides (62-64%). Similarly, Azcue and Nriagu (1993) found that the oxide fraction was predominant (57%), although they reported a considerably higher percentage (30-37%) in the residual fraction. Both teams interpreted their results as indicating that significant remobilization of arsenic could occur into the overlying water of Moira Lake if sediments at the mud-water interface became anoxic.

The differences in arsenic speciation between earlier studies and this study may be caused by: (1) the fact that only surface sediments (0-5 cm) were analyzed for speciation in this study whereas the entire core sample was analyzed in the earlier studies; (2) differences in particle size or organic carbon content of the sediment samples in the two studies; (3) conversion of oxide-bound arsenic to the residual phases; or, (4) the changing nature of arsenic inputs to the system, causing a different fractional distribution of arsenic in the surface sediment. The first two explanations may be the most important sources of variability between this study and previous work). Characterization of the surface layer only will reflect different physical/chemical conditions and a different time period from an overall average of a 30-cm core. Furthermore, the particle size of the sediments in this study may be coarser than in previous studies (see Appendix II, although equivalent particle size data were not presented for the previous work). Organic carbon was measured using different methods in this study compared to previous studies; however, there do not appear to be major differences in organic carbon content among the three studies (see Appendix II). The core profile presented by Azcue and Nriagu (1993) shows that the oxide and residual fractions have made up over 90% of the arsenic over the time period represented by a 30 cm core, with periodic increases in the oxide fraction. The authors suggested that the core profile indicated that some of the oxide-bound arsenic is being converted to the residual phases. This process may have been continuing, producing the higher proportion of residual arsenic observed in this study. Azcue and Nriagu also comment that the fractional distribution of arsenic in pre-mining sediments would be in the order residual>organic>oxide>exchangeable>carbonate whereas the polluted sediment sequence would be > rxide>residual>organic>exchangeable>carbonate. The sequence in this study was residual>oxide>organic>exchangeable>carbonates. This more recent sequence may be indicative of a gradual change related to decreased arsenic loadings from the Deloro Mine Site (as illustrated by the water quality data).

The main difference in nickel speciation between this study and Cornett and Chant (1986) is the much higher proportion of nickel in the exchangeable fraction in this study. Cornett and Chant reported that nickel was found primarily in the oxide and organic fractions (37% and 19-42%, respectively), with a significant proportion in the carbonate fraction (17-35%). In this study, the oxide, exchangeable and carbonate fractions dominated (20-57%, 33-44% and 11-37%, respectively). None of the nickel was found in the organic fraction in this study.

The differences in nickel speciation between Cornett and Chant and this study could be caused by: (1) the fact that only surface sediments (0-5 cm) were analyzed for speciation in this study whereas the entire core sample was analyzed in the earlier studies; (2) differences in particle size or organic carbon content of the sediment samples in the two studies; or (3) changing nature of nickel inputs to the system, causing a different fractional distribution of nickel in the surface sediment. Cornett and Chant found that in the top 4 cm of the cores, < 20% of the nickel was found in the organic fraction. Deeper in the core, the organic fraction contributed 50-70% of the total nickel present. Therefore, the differences in the proportion of nickel found in the organic fraction between this study and Cornett and Chant may be because this study analyzed the 0-5 cm layer only. Examination of particle size data for this study indicates that the sediment samples may be coarser than in Cornett and Chant, although equivalent particle size data were not presented in the earlier study (see Appendix II). Differences in particle size would certainly contribute to significant differences in nickel speciation. Organic carbon was measured using different methods in this study compared to Cornett and Chant; however, there do not appear to be major differences in organic carbon content (see Appendix II). The nature of the nickel entering the Moira system may be changing, because earlier data indicated that nickel inputs were largely in the dissolved form (assumed because sediments in Moira Lake near the Moira River inflow did not have elevated nickel concentrations) (Cornett and Chant 1986). Data from this study show elevated nickel concentrations in sediments in all stations from Highway 7 to Moira Lake. Therefore, the proportion of nickel entering as particulate materials may have increased.This particulate nickel may be re-suspended and re-deposited several times en route to Moira Lake. Depending upon the nature of the particles and the strength of the adsorptive bond, the particulate nickel may be more or less exchangeable.

The importance of the exchangeable fraction in this study is noteworthy, since this fraction would be the most prone to contributing nickel back into the overlying water of Moira Lake Furthermore, the porewater concentration of nickel in surficial sediments would be expected to be relatively high. This may contribute to the toxicity of the surface sediments to benthic invertebrates.

References

Ankley, G.T., D.M. DiToro, D.J. Hansen and W.J. Berry. 1996. Technical basis for deriving sediment sediment quality criteria for metals. Environ. Toxicol. Chem. 15(12): 2056-2066.

Azcue, J.M. and J.O. Nriagu. 1993. Arsenic forms in mine-polluted sediments of Moira Lake, Ontario. Environment International 19: 405-415.

Berry, W.J., D.J. Hansen, J.D. Mahony, D.L. Robson, D.M. DiToro, B.P. Shipley, B. Rogers, J.M. Corbin and W.S. Boothamn. 1996. Predicting the toxicity of metal-spiked laboratory sediments using acid-volatile sulfide and interstitial water normalizations. Environ. Toxicol. Chem. 15(12): 2067-2079.

Burgess, R.M. and K.J. Scott. 1992. The significance of in-lace contaminated marine sediments on the water column: processes and effects. Pp129-165 In: Sediment Toxicity Assessment. G. A. Burton (ed). Lewis Publishers, Boca Raton, FL. 457pp.

Burton, G.A., Jr. 1992. Sediment collection and processing: factors affecting realism. Pp. 37-66 In: Sediment Toxicity Assessment. G. A. Burton (ed). Lewis Publishers, Boca Raton, FL. 457pp

CH2M Gore & Storrie. 1999. Deloro Mine Rehabilitation Project: Extent and Character of Radioactive Materials - Final Report, June, 1999. Submitted to Ontario Ministry of the Environment.

Cornett, R.H., B. Risto and L. Chant. 1986. Kinetics of arsenic and nickel in sediments in Moira Lake. Report to the National Uranium Tailings Program, Canada Centre of Mineral and Energy Technology, Ottawa, Ontario.

Cornett, R.J. and L. Chant. 1986. Speciation of arsenic and nickel in sediments of Moira Lake. Report to the National Uranium Tailings Program, Canada Centre of Mineral and Energy Technology, Ottawa, Ontario.

Eisenbud, M. 1987. Environmental Radioactivity. Third Edition. New York: Academic Press. (Harcourt Brace). 473pp.

Golder Associates Ltd. 1996. Investigation of subsurface and environmental conditions: proposed gas main crossings of Moira River and Young's Creek Highway 7. Prepared for Centra Gas, North York, Ontario.

HåKanson, Lars. 1992. Sediment variability. pp 19-35 In: Sediment Toxicity Assessment.
G.A. Burton (ed). Lewis Publishers, Boca Raton, FL. 457 pp.

Jaagumagi, R. Ontario Ministry of Environment and Energy (MOE). Personal communication. March 2000.

IAEA. 1992. Effects of ionizing radiation on plants and animals at levels implied by current radiation protection standards. Technical Report Series No. 332. International Atomic Energy Agency, Vienna.

Mahony, J.D. D.M. DiToro, A.M. Gonzalez, M. Curto, M. Dilg, L.D. DeRosa and L.A. Sparrow.1996. Partitioning of metals to sediment organic carbon. Environ. Toxicol. Chem. 15(12): 2187-2197.

Mason, S. and J. Dragun. 1996. Natural chemicals in sediments. Amherst Scientific Publishers, Amherst, Massachusetts.

MOE 1993. Guidelines for the protection and management of aquatic sediment quality in Ontario. Ministry of the Ontario Environment (MOE). P_p. 16-18

Mudroch, A. and J.A. Capobianco. 1980. Impact of past mining activities on aquatic sediments in Moira River basin, Ontario. J. Great Lakes Res. Internat. Assoc. Great Lakes Res. 6(2) 121-128.

Ontario Ministry of the Environment and Energy (MOE). 1993. Guidelines for the protection and management of aquatic sediment quality in Ontario. Queen's Printer for Ontario.

Peterson, G.S., G.T. Ankley and E.N. Leonard. 1996. Effect of bioturbation on metal-sulfide oxidation in surficial freshwater sediments. Environ. Toxicol. Chem. 15(12): 2147-2155

Powell, M. Ontario Ministry of Environment and Energy (MOE). Personal communication. March 2000.

SCIMUS. 1999. Deloro Mine Rehabilitation Project: extent and character of radioactive materials. Prepared for Ontario Ministry of the Environment by SCIMUS Inc., Toronto, Ontario.

Tessier, A., G.G.C. Campbell and M. Bisson 1979. Sequential extraction procedure for the speciation of particulate trace metals. Analytical Chemistry 51(7): 844-851.

2.3 Benthic Invertebrate Community

2.3.1 Methods

2.3.1.1 Sampling Design

Benthos samples were collected synoptically with bulk sediment chemistry samples and sediments collected for sediment bioassay testing. Because of this, benthic invertebrates were collected from depositional zones only from both lake and river habitats. For the Moira River, sites were typically located immediately upstream of a dam or weir within the low velocity head pond and sediment accumulation zone. Benthic invertebrate abundance data are provided in Appendix II.

Lake Habitat

The benthic invertebrate study for lake habitat focused on benthos communities exposed to elevated metals in Moira Lake, Stoco Lake and the Moira River at Bend Bay, and reference communities in Round Lake and Consecon Lake (Figure 2.2-1). Bend Bay was included in analyses of lake habitat because it exhibits characteristics that are more similar to lentic rather than lotic habitat. For Moira and Stoco lakes, five sites were sampled in each lake corresponding to locations of lake inflow, outflow and deep basins (Figure 2.2-1). Three sites were sampled in Round and Consecon lakes and were similarly located in the vicinity of the inflow, outflow and deep basin of each lake. A single site was sampled at Bend Bay. The general location and UTM coordinates of each sampling site are provided in Table 2.3-1.

River Habitat

Benthos samples were collected at 8 sites in the upper section of the Moira River (upstream of Moira Lake) (Figure 2.2-1). These sites were two upstream reference sites (MR-1, MR-3), and six sites (MR-5 to MR-10) located at increasing distances downstream of the former Deloro Mine Site. An additional site was located in Young's Creek due to concerns regarding high levels of mine-related metal concentrations and issues regarding future mitigative action. Benthos samples were also taken from single reference sites in the Black River and the Skootamatta River. Both rivers are tributaries of the Moira River and provided additional reference data for comparisons with exposed sites

Table 2.3-1
General Location and UTM Co-ordinates of Benthic Invertebrate Sampling Sites,
Lake Habitat, Moira River System, Spring 1999

(a) Universal Transverse Mercator (UTM).

(b) Site CL-D was replaced by site CL-2 due to potential anoxic conditions during summer stratification.

(c) D/S = downstream

In the lower Moira River (downstream of Moira and Stoco lakes), benthos samples were taken from four sites, including one located between Moira and Stoco lakes (MR-12) and three located downstream of Stoco Lake (MR-13 to MR-15) (Figure 2.2-1). Suitable reference sites were not available in this section of the Moira River; therefore, reference data were collected from three sites located on the adjacent Salmon River. Sites on the Salmon River were located at approximately the same latitude as the corresponding sites on the Moira River, providing a measure of natural downstream gradients. The general location and UTM coordinates of each sampling site are provided in Table 2.3-2.

Table 2.3-2
General Location and UTM Co-ordinates of Benthic Invertebrate Sampling Sites,
River Habitat, Moira River System, Spring 1999

U/S = upstream; D/S = downstream.

2.3.1.2 Sample Collection

The Ontario Ministry of the Environment (MOE) conducted field sampling from April 22 to May 12, 1999. Exact sampling locations (i.e., deposition zones) were finalized in the field to ensure that all sites were similar in habitat characteristics. Three replicate samples were taken at each lake and river site. Benthos samples were collected using a Ponar grab (0.05 m²) for lakes and deep river sites. An Ekman grab (0.05 m²) was used at all river sites that could be easily waded. Sampling dates for each site, as well as the range in grab fullness, are presented in Table 2.3-3. Each sample was sieved in the field using a #30 Tyler sieve (600-µm mesh) and preserved immediately in 10% buffered formalin. Benthic samples were shipped to Mr. William Morton, Guelph, Ontario for sorting and taxonomic identification. Chain-of-custody forms were used to document shipping rocedures and to ensure sample integrity.

Table 2.3-3
Fullness for Each Lake and River Benthos Site, Moira River System, Spring 1999

^(a)range from three replicates/site.

N/A - not available.

Supporting environmental data measured at each site included water depth, temperature and conductivity. Sediment grain size data and total organic content were also available from results of bulk sediment analyses (Appendix II). This information was used to evaluate similarity among sites and identify potential influences on benthos community structure other than mine-related metal contamination.

2.3.1.3 Sample Sorting and Taxonomic Identification

Benthos samples were initially treated with rose bengal stain for at least 24 h to facilitate sorting. Samples were then washed with water using a 425-µm sieve (US #40) to remove preservative and excess stain. Sorting was performed by placing the sample in a gridded pan and viewing it under a dissecting microscope (10x magnification). Entire samples were sorted when possible; however, sub-sampling was required for samples that were large in volume (typical of grab samples from depositional areas) (Table 2.3-4). When sub-sampling was required, the sample was evenly distributed in a 425-µm sieve and divided in half. Half of the sample was removed and archived, while the remaining half was either processed or sub-sampled further. Sorting typically started with a 1/8 or 1/4 sub-sample. Sorting efficiency was estimated by randomly selecting 10 samples (9.2%) and re-sorting the originally sorted material. Any specimens found in the re-sorting procedure were compared against the original data. Sorting efficiency of ≥90% was considered acceptable.

Invertebrates were identified to the lowest practical level (species when possible) and enumerated. A reference collection containing specimens of each taxon was maintained and identifications verified by recognized experts including:

• Annelida: Dr. David Barton, University of Waterloo, ON
  Kathryn Coates, Royal Ontario Museum, Toronto, ON
• Acarina: Dr. Ian Smith, Agric. Canada, Biosystatics Research Farm, Ottawa, ON
• Amphipoda: Dr.d Bousefield, National Museum, Ottawa, ON
• Isopoda: Famina Rafi, National Museum, Ottawa, ON
• Decapoda: Dr. Dave Barr, Royal Ontario Museum, Toronto, ON
• Chironomidae: Dr. Len Ferrington, Lawrence, KA
  Dr. Don Oliver, Agric. Canada, Biosystematics Research Farm, Ottawa, ON
• Ephemeroptera: Dr. Bob Waltz, Indiana Dept. Natural Resources, Indianapolis, IN
• Trichoptera: Dr. John Morse, Clemson University, Clemson, SC
  Pat Schaefter, Royal Ontario Museum, Toronto, ON
• Mollusca: Dr. Gerry Mackie, University of Guelph, Guelph, ON

Table 2.3-4
Percent of Benthos Sample Sorted for Taxonomic Identification,
Moira River System, Spring 1999

N/A - not available.

2.3.1.4 Sediment Toxicity Tests

Sediment grab samples collected at the benthic invertebrate sampling sites were tested for toxicity using a battery of three tests, including:

• Hyalella azteca (14-day) survival and growth (Protocol EPS 1/RM/33; Environment Canada 1997a);
• Chironomus riparius (10-day) survival and growth (Protocol EPS 1/RM/32; Environment Canada 1997b); and
• Fathead Minnow (Pimephales promelas) juvenile (21-day) survival and growth (Ontario Ministry of the Environment Laboratory Sediment Biological Testing Protocol; Bedard et al. 1995).

Toxicity tests were carried out by ESG International, Guelph, Ontario (Hyalella and Chironomus) and HydroQual Laboratories, Calgary, Alberta (Fathead minnow).

All of these tests were conducted within six weeks of sample collection. However, the Chironomus riparius test had to be re-run because of problems with the performance of the controls. The second run of the Chironomus riparius test was conducted 7 weeks after the sediments were collected.

2.3.1.5 Data Analysis

Data Preparation

The raw abundance data were checked for data entry errors by the taxonomist and corrections were made as necessary. To prepare the data for analysis, non-benthic and terrestrial taxa were deleted and the abundance of taxa per sample were converted to abundance per square metre.

General Approach

The alternatives for analysis of the benthos data included the gradient approach (e.g., regression analysis using sediment metal concentrations as the independent variable), or comparisons of means of site groups representing discrete levels of sediment metal concentrations (e.g., ANOVA). Since there was no recent information on sediment chemistry at the study design phase, the exact approach could not be identified prior to availability of the spring 1999 field data. To select the analytical approach, the spatial variation in sediment metal concentration was first examined for each habitat type (river and lake). The approach was then selected according to the following principles: if there were distinct groups of sites representing similar levels of metals in sediments (e.g., background, low, moderate, high), then comparison of these groups as "treatments" would be appropriate using ANOVA or ANCOVA. Alternatively, if the sites sampled represented a wide range of sediment metal concentrations without obvious grouping, then a gradient-type analysis, such as linear regression would be applicable.

Raw metal concentrations were used to select the appropriate statistical approach. The metals included in the analysis were those identified as metals of concern after screening for differences with reference concentrations and exceedances of Provincial Water Quality Objectives (Section 2.1). The sediment data were not normalized because at the present there is no general consensus regarding the best procedure to normalize sediment metal data. Additionally, initial examination of the sediment data did not suggest any obvious choices for normalization. There were no significant correlations between levels of any of six metals of concern (arsenic, cobaltcopper, lead, nickel, silver) at lake sites with sediment particle size (% sand, silt, clay), but copper, lead and silver were correlated with sediment total organic carbon (TOC). At river sites, copper was correlated with % silt, and lead was correlated with all sediment variables, but no other correlations were found. Uranium was not included in these analyses because it was significantly elevated only in the Young's Creek sample.

Sediment metal concentrations were generally similar within each lake, but varied among lakes (Figure 2.3-1 and Section 2.2). Sediments in the two reference lakes were characterized by low concentrations of all metals, though Round Lake sediments had higher concentrations compared to Consecon Lake. In particular, lead concentration was higher at two Round Lake sites than at the other four reference sites. Stoco Lake sediments had low levels of metals, and moderate to high levels were found in Moira Lake sediments. The sediment sample collected at the single site in Bend Bay had particularly elevated concentrations of four of the six metals of concern. Based on this information, the ANOVA/ANCOVA approach was adopted for analysis of the lake data set, using each lake as a treatment and the sampling site as the unit of replication.

In the river data set, there was no clear grouping of exposed river sites with respect to sediment metals (Figure 2.3-2 and Section 2.2). Below the former Deloro Mine Site, a gradual increasing trend was found in arsenic, cobalt, copper, nickel and lead from sites MR-5 to MR-7, followed by high but variable levels of each metal from MR-8 to MR-10. Silver concentrations were much lower than those of the other metals, but exhibited a similar spatial trend. In the lower Moira River, concentrations of most metals of concern were close to background, with the exception of site MR-13 (first site below Stoco Lake), where concentrations were similar to those at MR-5. Based on the large variation in metal concentrations among sites and the lack of site groups with similar levels of sediment metal concentrations, the gradient approach using linear regression was adopted for the analysis of the river benthic invertebrate data set.

Figure 2.3-1

Metal Concentrations in Sediment at Benthic Invertebrate Sampling Sites in Lake Habitat, Moira River System, Spring 1999

Figure 2.3-2

Metal Concentrations in Sediment at Benthic Invertebrate Sampling Sites in River Habitat, Moira River System, Spring 1999

Lake Habitat

Benthic community variables were compared among lakes using one-way ANOVA or ANCOVA, followed by Tukey's tests comparing the two reference lakes (Consecon and Round) and each exposed lake (Moira and Stoco) with each reference lake. Because only a single site was sampled in Bend Bay, it was excluded from the statistical tests, but was shown on graphs for each biological variable tested and was included in the Sediment Quality Triad (SQT) analysis (see below). Site CL-D (17.25-m deep site in Consecon Lake) was excluded from the lake data analysis because it was considerably deeper than all other sites (4 to 10 m) and would have been influenced by persistent anoxic conditions during summer stratification (Section 2.1).

A relatively large number of benthos variables were included in these comparisons, because there is considerable variation in the specific variables that have proven useful in studies of metal concentrations (Clements 1991). Community-level variables included total abundance, taxonomic richness, abundance of major taxonomic groups and the percentage of total abundance represented by those groups. Additionally, individual taxa (at the lowest level of identification) were selected for the analysis based on initial graphical examination of the data. Common taxa (i.e., those accounting for ≥1% of the total number of invertebrates in the lake data set) that appeared to vary in abundance in response to metal concentrations were included in the statistical analysis.

A number of additional biological variables were also included in the statistical analysis to allow assessment of trends in the abundance and percentage of metal-sensitive and metal-tolerant invertebrates. Abundances of metal-sensitive and metal-tolerant organisms were summed and were also expressed as the percentage of total abundance for statistical tests. Species-specific information on metal sensitivity (Klemm et al. 1990) and more general information on sensitivity of higher taxonomic groups (Mance 1987, Clements 1991) were used to derive these variables.

Because very few of the taxa in the study area were designated as metal-tolerant or sensitive by Klemm et al. (1990), it was necessary to rely on the more general information regarding metalsensitivity. Metal-tolerant invertebrates included Limnodrilus spp., Tubifex tubifex, and midges in the subfamilies Orthocladiinae and Tanypodinae. Tanypodinae midges were included in this group because they have been shown to comprise a larger fraction of the community at sites affected by mine-related metals (Winner et al. 1980, Clements 1991). Sensitive invertebrates included all mayflies (Ephemeroptera), and midges in the tribes Tanytarsini and Chironomini. Although not all Chironomini midges are metal-sensitive, Clements (1991) showed their general sensitivity to one metal (copper) in outdoor stream mesocosms. Additionally, more than half of the Chironomini genera in the present data set were either metal-sensitive, or had at least one sensitive species according to sensitivity designations by Klemm et al. (1990).

The sensitivities of benthic taxa to arsenic and metals were compared based on studies of rivers that contained arsenic in addition to heavy metals (Canfield et al. 1994, Poulton et al. 1995, Beltman et al. 1999, Mori et al. 1999). These studies reported similar impacts on benthos between sites containing arsenic and sites containing arsenic plus metals, and identified the same taxonomic groups as sensitive or tolerant. Therefore, in the absence of specific information on arsenic-sensitivity of benthic taxa, it was assumed that metal-sensitive taxa were also sensitive to arsenic, and to about the same degree.

Since there was a strong mean-variance relationship for all abundance variables, the abundance data were log(x+1) transformed before statistical analysis. Percentage-type variables were arcsintransformed. Richness variables were not transformed, as no violations of the assumptions of parametric tests were found.

The relationships between benthic community variables and physical variables were investigated to identify potentially confounding factors and to select covariates for statistical tests. First, the range in each physical variable was examined to evaluate whether the range was wide enough to influence invertebrate communities, or to act as a confounding factor. This examination eliminated water temperature, conductivity and sediment particle size. Then, Pearson correlation coefficients were generated between each biological variable (transformed as above) and the remaining physical variables (depth, TOC). Significant correlations were examined as scatterplots to determine whether they represented real trends (i.e., not resulting from one or a few atypical points). The physical variables that were correlated with biological variables were then used as covariates in the ANCOVA analyses.

River Habitat

The relationships between benthic invertebrate variables and metal concentrations in bottom sediments were evaluated using linear regression analysis. The dependent variables (i.e., biological variables) included those described above for the lake data analysis. Abundance variables were log(x+1) transformed before analysis. The independent variables included concentrations of the metals of concern (summarized as a single variable [PC1] by Principal Component Analysis [PCA], described below), water depth, % fine sediments (% silt + % clay) and sediment TOC. The single site sampled in Young's Creek was excluded from the analysis because of habitat differences (it is a very small stream compared with the other watercourses sampled) and because the high sediment metal concentrations at this site would act as outliers during data analysis or result in high leverage in regressions.

To reduce the number of independent variables to a minimum, a Pearson correlation matrix was generated to assess correlations between each biological variable and the three major habitat variables (water depth, % fine sediments, TOC). The habitat variables that were significantly correlated with the biological variables, if any, were then used in combination with sediment chemistry in multiple regressions against the biological variables. In the event that none of the habitat variables were correlated with a biological variable, sediment chemistry PC1 was used as the only independent variable. To ensure that the best set of independent variables was found (i.e., the combination that explains the maximum amount of the variation in the dependent variable), regressions were run using each independent variable by itself and in all possible combinations with the other independent variables and sediment chemistry PC1.

In addition to the above analysis, the three sites sampled in the lower Moira River (MR-13, MR-14, MR-15) were compared with the corresponding reference sites in the Salmon River (SR-1, SR-2, SR-3), as planned during the study design phase. Because the supporting data indicated that there was a strong declining trend in TOC with distance downstream in both rivers, the biological variables were compared between rivers using paired t-tests, which allowed matching of sites with regard to sediment TOC content.

Multivariate Analysis

The lake and river benthic data sets were summarized separately by ordination to assess the variation in community structure among sites and to generate community summaries for the Sediment Quality Triad (SQT) analysis. Non-metric Multidimensional Scaling (NMDS) was used for this analysis. NMDS is a non-parametric ordination method that allows reduction of a data set consisting of a large number of variables (taxa in this case) to typically two or three new dimensions (analogous to ordination axes) (Clarke 1993). NMDS is considered more useful than PCA to summarize biological data because: it does not assume linear responses along environmental gradients; it is less sensitive to outliers; it is appropriate for data encompassing long environmental gradients; and, it is not limited by the number of variables that can be included in the analysis (although some degree of data reduction is still recommended before running NMDS) (Paine 1998).

Before performing the ordinations, the lake and river data sets were reduced to exclude rare taxa, defined as those that collectively constituted <2% of the mean total number of invertebrates in each data set. This procedure reduced the number of taxa from 68 to 39 in the lake data set and from 117 to 65 in the river data set, while retaining 98% of the original number of invertebrates in each data set. The reduced abundance data were log(x+1) transformed before ordination to reduce the influence of dominant taxa on the ordination results. A Bray-Curtis distance matrix was generated from the transformed abundance data (as recommended by Clarke 1993) and was used as the input for the ordination. Two dimensions were selected for the ordination to facilitate the interpretation of results, after confirming that the stress of the two-dimensional configuration was reasonably low (<0.2) for both the lake and river data sets.

Ordination results were presented as two-dimensional scatter-plots of the sampling sites in ordination space. To facilitate recognition of trends in community structure with sediment metal concentrations, the sizes of symbols representing sites were scaled to reflect the relative degree of sediment arsenic concentration at each site. Correlations between ordination scores, physical variables and sediment chemistry were assessed using Spearman rank correlation analysis. Since NMDS does not provide an indication of the taxa associated with each dimension (only inter-site distances are used in the analysis), Spearman rank correlation coefficients were also generated between each dimension and abundances of the taxa in the reduced biological data set used for the ordination.

Sediment Quality Triad

The present study collected information describing benthic community structure, sediment chemistry and sediment toxicity. These components make up the basis of what has been referred to as the Sediment Quality Triad approach (SQT) proposed by Long and Chapman (1985), Chapman et al. (1987) and Chapman (1990). Chapman et al. (1991) argued that the SQT incorporates three essential components: 1) measures to determine the presence and degree of anthropogenic metal concentrations (i.e., bulk sediment chemistry); 2) measures that demonstrate that substances are present that can interfere with the normal functioning of at least some biological organisms tested in the laboratory (i.e., sediment toxicity tests); and, 3) measures of in situ alteration of resident biological communities (i.e., benthic community structure).

Although each component provides important information regarding the magnitude and spatial extent of potential environmental effects, the SQT attempts to integrate information from each component by examining the relationships or associations among biological, chemical and toxicological data. In essence, concordance among data sets provides the weight of evidence that observed biological effects are related to metals in the system (i.e., infers possible causation).

To relate the three data sets (benthos, sediment chemistry and toxicity) accumulated during the spring 1999 field survey, a variant of the two-step methods described by Green et al. (1993) was used. This method consists of first summarizing the pattern within each data set using ordination and then relating the within set patterns (ordination axes) to each other using correlation analysis. This approach allows description of within-set patterns based on the ordinations, which can provide important information for the interpretation of triad results. Separate SQT analyses were carried out for each habitat type (river and lake).

The biological data set was summarized using NMDS, as described above. This analysis generated two dimensions (ordination axes) which were both included in the SQT analysis. Total invertebrate abundance and taxonomic richness were also included as additional biological variables in the SQT because relationships between these basic community variables and sediment chemistry/toxicity are usually considered ecologically meaningful and are useful in evaluating the severity of biological effects. Benthos data and associated sediment chemistry data were available for 17 sites in each habitat type, excluding Site YC-1 from the river data and Site CL-D from the lake data for reasons discussed above.

Sediment chemistry and toxicity data sets were each summarized using Principal Component Analysis. The raw chemistry data were used for the chemistry PCA, including concentrations of five of the six metals of concern in sediment grab samples collected at the benthic invertebrate sites (arsenic, cobalt, copper, lead, nickel). The data for silver were not used because they were incomplete (i.e., no data for four sites). Exclusion of this metal from the analysis was unlikely to affect ordination results because silver concentrations were positively correlated with those of most other metals of concern.

Before ordination of the toxicity data, each test result was expressed as the % of the control test result (% control). This was necessary to standardize the data across sites, because two batches of tests were run, each with its own control sediment. In addition, the number of toxicity variables was reduced by omitting results for fathead minnow survival and weight. The fathead minnow data were invalid due to problems with dissolved oxygen concentrations during the test (Appendix III, Table 8). Therefore, the variables used in the toxicity PCA included Hyalella azteca survival and weight and Chironomus riparius survival and weight. The detailed toxicity test results are presented in Appendix III. Toxicity data were only available for 13 sites in each habitat type.

A Spearman rank correlation matrix was generated to evaluate the relationships among the three SQT components.

2.3.2 Results and Discussion

2.3.2.1 Lake Habitat

Variation in Habitat

Field measurements revealed differences among lakes in conductivity (Table 2.3-5), which largely reflected the transition from shield to limestone bedrock in the study area (generally from north to south, or in a downstream direction) and the source of water inflows to each lake. The two reference lakes were at the two extremes in conductivity. Round Lake is located on the Canadian Shield, where surface waters typically have low levels of dissolved ions. Consecon Lake is located on limestone bedrock, which usually results in elevated dissolved ion levels relative to shield lakes. In the Moira River system, the decline in conductivity from Bend Bay and Moira Lake to Stoco Lake reflected the inputs of water from the Black and Skootamatta rivers, which drain shield areas and account for the majority of the inflow to Stoco Lake (Owen and Galloway 1969). Overall, although conductivity varied among lakes, the variation observed was within the typical range for surface waters in central Ontario and was unlikely to appreciably influence benthic community structure.

Table 2.3-5
Variation in Field Water Quality Parameters, Depth and Sediment Particle Size Among
Lake Sites, Moira River System, Spring 1999

Oxygen depletion occurred at the sediment-water interface in Consecon, Stoco and Moira lakes (based on single profiles in each lake). This created concern regarding the suitability of Round Lake as a reference lake because Round Lake did not show oxygen depletion. Although the potential for anoxia was considered during the selection of reference sites (e.g., one deep Consecon Lake site was excluded from the data analysis), the available dissolved oxygen profile data were too limited to justify excluding Round Lake from the analysis.

Other physical variables also varied among sites and lakes. Temperature measured at the water surface was similar at most lake sites (Table 2.3-5). Vertical variation in temperature was small at the time of sampling except in Consecon Lake in mid-summer, where a thermocline was present between 9 and 11 metres (Section 2.1). Water depth varied substantially among sites, even without the deep site in Consecon Lake (CD-D; excluded from the data analysis) (Table 2.3-5). Bottom sediments consisted mostly of silt and clay at the majority of sites, with relatively little variation. The only exception was Site SL-3 in Stoco Lake, where the sediments were dominated by sand. Sediment TOC content varied widely among sites, sometimes reflecting particle size distribution (i.e., TOC was low at the sandy site), but also exhibiting noticeable among-lake differences. In particular, Round Lake sediments had high organic carbon content, and the greatest within-lake variability in TOC was found in Moira Lake.

The variation in the physical attributes of sampling sites was sufficient to account for part of the variation in benthic community structure in the study area. Thus, physical variation could have obscured biological differences due to metal concentrations among lakes. The physical variables of greatest concern with regard to benthic community structure were depth and sediment TOC, because both can profoundly influence community structure in lakes and both varied relatively widely from site to site. Furthermore, TOC is related to metal speciation in sediments. Therefore, these habitat variables were included as covariates in the statistical comparisons of lakes, where appropriate (see Section 2.3.1.5 for the procedure used to select covariates).

Benthic Community Analysis

There were no readily apparent differences in community composition between the reference lakes and the exposed lakes that would suggest a metal-related effect on the benthos. A total of 68 benthic taxa were identified in the samples collected at lake sites. Lake benthic communities consisted largely of chironomid midge larvae (Chironomidae), oligochaete worms (Oligochaeta), nematode worms (Nematoda), mussels (Bivalvia) and snails (Gastropoda), which together accounted for about 80% of total abundance (Figure 2.3-3). Percentages of these groups varied considerably within each lake, with the exception of Consecon Lake. In this lake, all three sites were strongly dominated by chironomids and other groups were nearly absent. Of the other lakes, the benthic fauna of Round Lake (a reference lake) appeared most variable at this coarse level of examination, largely due to varying abundances of bivalves and oligochaetes. Statistical comparisons of % Oligochaeta and % Chironomidae among lakes found a significant difference in % Oligochaeta between Consecon Lake and each exposed lake (Tables 2.3-6 and 2.3-7). Chironomidae percentage did not vary significantly among lakes.

The variation in total abundance was not influenced by metal concentrations. Total invertebrate abundance ranged between 1500 (ML-2, ML-3, CL-2) and 7000 (CL-1) individuals per square metre, based on site means (Figure 2.3-3). Total abundance varied widely within lakes and no significant differences were found among lakes (Table 2.3-6). The sites in Consecon Lake (the reference lake with the lowest sediment metal concentrations) encompassed nearly the entire range in total abundance observed during this survey.

The richness data suggested that the number of taxa might have been reduced by sediment metal concentrations in Moira Lake and possibly in Bend Bay (Figure 2.3-3). Mean richness (mean of the three replicate samples at a site) ranged from 3 (ML-3) to 19 (SL-3) and total richness (all taxa found at a site) varied from 5 (ML-3) to 28 (CL-1). Ranges in both mean and total richness were similar in Consecon, Round and Stoco lakes. However, the maximum richness (14 taxa) in Moira Lake and Bend Bay was only slightly greater than the minimum value (12) in the reference lakes. Total richness was significantly lower in Moira Lake relative to Consecon Lake, but not when compared with Round Lake, and mean richness did not vary significantly among lakes (Tables 2.3-6 and 2.3-7). Within Moira Lake, the pattern in richness (i.e., lowest values at ML-3 and ML-4) did not reflect the sediment metal concentrations (generally higher concentrations at ML-1 to ML-3 than at ML-4 and ML-5; Figure 2.3-1). Thus, the significantly lower total richness suggested a metal-related effect in Moira Lake, but the pattern in richness within this lake did not correspond with metal concentrations in the sediments.

Figure 2.3-3
Total Invertebrate Abundance, Taxonomic Richness and Proportions of Major
Invertebrate Groups at Lake Sites, Moira River System, Spring 1999

Total Abundance

Richness

Community Composition

Table 2.3-6
Results of Overall ANOVA and ANCOVA Tests Comparing Benthic Community Variables
Among Lakes, Moira River System, Spring 1999

^(a)Probability value for testing an overall difference among means.

^(b)Total degrees of freedom (d.f.) = 15, except for tests of Limnodrilus spp. abundance and %, where d.f.=14

^(c)Probability value for testing an overall difference among slopes and intercept means.

^(d)One outlier was removed.

^(e)No result; test not carried out.

Table 2.3-7
Results of Tukey's Tests Comparing Benthic Community Variables Among Lakes for Variables
with Significant ANOVA or ANCOVA Results, Moira River System, Spring 1999

^(a)P-values are from Tukey's pairwise comparisons (following a significant ANOVA/ANCOVA F-statistic).

^(b)% Difference is for exposed relative to the reference lake (or Round Lake relative to Consecon Lake) calculated using ANOVA/ANCOVA adjusted least squared means.

^(c)One outlier was removed.

Visual examination of oligochaete abundances suggested the potential for subtle effects of sediment metal concentrations. Oligochaete numbers were generally higher in the exposed lakes and were particularly high in Bend Bay, but there was no overall significant difference among lakes (Figure 2.3-4, Table 2.3-6). In particular, the family Tubificidae, which included two genera (Tubifex and Limnodrilus) with common metal-tolerant species, was more numerous at Moira River system sites than at most sites in reference lakes. Despite this apparent trend, the among-lake variation in Tubificidae abundance was not significant (P = 0.075; Table 2.3-6).

The chironomid abundance data also provided only limited evidence of effects in the exposed lakes. A significant difference was found in total chironomid abundance between Consecon Lake and Moira Lake (the lakes with the lowest and highest metal concentrations), but not in any of the other comparisons (Figure 2.3-4, Tables 2.3-6 and 2.3-7). Differences among lakes and Bend Bay in the percentage of major chironomid groups were consistent with their relative sensitivities. The order of sensitivity for these groups is Orthocladiinae < Tanypodinae < Chironomini < Tanytarsini (in order of increasing sensitivity; based on information summarized by Clements 1991). In the lake data set, the highly sensitive Tanytarsini were nearly absent from all lakes except for Consecon Lake, which had the lowest metal concentrations. The percentage of the moderately sensitive Chironomini was lower in the exposed lakes than in the reference lakes. The percentage of the moderately tolerant Tanypodinae was higher in the exposed lakes (Figure 2.3-4). The percentages of these groups relative to total chironomids varied significantly among lakes in the overall ANOVA (Table 2.3-6). Significant pairwise differences were found in % Tanytarsini between Consecon Lake and each of the other lakes, and in % Tanypodinae and % Chironomini between Round Lake and the two exposed lakes (Table 2.3-7).

Visual examination of the abundances of common invertebrate groups and individual taxa also provided limited evidence of metal-related effects. Of the 27 taxa examined, a handful exhibited spatial patterns consistent with metal concentration. In most cases, these patterns were consistent with the reported sensitivity of these organisms. Taxa that appeared sensitive to metals included total mayflies (Ephemeroptera), Hyalella azteca, Chironomus, Tanytarsini and Tribelos (Figure 2.3-5). These invertebrates occurred in low numbers or were absent in Moira Lake and Bend Bay, occurred in variable numbers in Stoco Lake and were more abundant in at least one of the reference lakes. Clements (1991) and Mance (1987) have identified mayflies, crustaceans in general and Tanytarsini midges as sensitive. The two chironomid genera Chironomus and Tribelos are members of the Tribe Chironomini, which has been shown to be moderately sensitive to copper (Clements 1991).

Figure 2.3-4
Abundances of Oligochaeta and Chironomidae at Lake Sites,
Moira River System, Spring 1999

Oligochaeta

Chironomidae

Figure 2.3-5
Abundances of Ephemeroptera, Hyalella azteca, Chironomus, Tanytarsini and Tribelos at
Lake Sites, Moira River System, Spring 1999

Very few taxa showed the opposite pattern (e.g., Limnodrilus, Cladopelma Dero digitata, Figure 2.3-6). Clear examples of tolerance, exhibited in an increase in abundance at polluted sites, were not apparent in the lake data set, but D.digitata was very abundant in Bend Bay. Of the taxa showing some tolerance, Limnodrilus has at least one known tolerant species (L. hoffmeisteri) (Clements 1991), but Cladopelma is a member of the generally sensitive Chironomini and there is no information on the metal-tolerance of D. digitata (a naidid worm).

Figure 2.3-6
Abundances of Limnodrilus, Dero digitata, and Cladopelma at Lake Sites

Statistical comparisons of the abundances of these organisms were only possible for Chiromonus, Tanytarsini, Cladopelma and Limnodrilus, because the remaining taxa were present at fewer than three sites in most lakes. All but Cladopelma varied significantly among lakes (Table 2.3-6), but significant pairwise differences between the exposed lakes and both reference lakes were not found (Table 2.3-7).

The combined abundance and percentage of metal-sensitive invertebrates was reduced in the exposed lakes relative to the reference lakes (Figure 2.3-7), and the differences between each reference lake and both exposed lakes were statistically significant (Table 2.3-7). The trends for metal-tolerant taxa were less pronounced, largely because Consecon Lake communities included a moderate proportion of tolerant taxa. As a result, both the abundance and percentage of metaltolerant taxa differed significantly between the two reference lakes, and the exposed lakes were significantly different only from Round Lake (Table 2.3-7).

Figure 2.3-7
Abundances and Percentages of Metal-sensitive and Metal-tolerant Invertebrates at
Lake Sites, Moira River System, Spring 1999

In summary, examination of the lake benthos data and univariate analysis detected some trends in benthic community structure that were consistent with the effects of elevated metal concentrations, but these effects were subtle. Total abundance was not affected by metal concentrations in the exposed lakes, and richness was only reduced in Moira Lake. Despite the trends apparent upon visual examination of the data, statistical tests rarely detected significant differences between the exposed lakes and both reference lakes, which was largely the function of low statistical power in the pairwise tests. However, the abundance and percentage of metalsensitive invertebrates were significantly reduced in both exposed lakes and were also reduced in Bend Bay relative to both reference lakes. In addition, the among-lake variation in abundances of a number of individual taxa was consistent with the relative metal-sensitivities of the taxa.

Results of the benthic community survey indirectly confirmed the suitability of Round Lake as a reference lake. Differences in dissolved oxygen profiles between Round Lake and the other three lakes (Appendix I) were of some concern during the analysis, because oxygen depletion occurred at the sediment-water interface in Consecon, Stoco and Moira lakes, but not in Round Lake. Based on the statistical analysis (Table 2.3-7), the number of biological variables that differed significantly between the study lakes and either reference lake were about equal and the corresponding % differences from each reference lake mean were similar. Thus, the study lakes differed about equally from each reference lake (despite the occasional significant difference between the reference lakes), thereby confirming the suitability of Round Lake as a reference lake.

Statistical Power

The power of the non-significant statistical tests was generally low (Table 2.3-8), as may be expected based on the low number of sites (replicates) in each lake. Considering power ≥0.80 as sufficient, only one of the non-significant tests (% Chironomidae) had sufficient power to detect a 50% difference in means and two had sufficient power to detect a 100% difference.

Multivariate Analysis

Ordination of the lake benthos data showed separation of Consecon Lake communities from those of other lakes, but Round Lake sites were widely separated from each other (Figure 2.3-8). The stress of the final configuration was 0.17, which represents "fair" fit of the results to the input data in qualitative terms (Clarke 1993). There was a general trend from reference to exposed sites along Dimension 1, as illustrated for arsenic in Figure 2.3-8 (larger symbols represent sites with higher arsenic concentrations). Arsenic was used to represent sediment chemistry because it was one of the most toxic substances released from the former Deloro Mine Site, and its concentration in sediment was correlated with those of all other metals of concern, with the exception of lead.

Table 2.3-8
Statistical Power to Detect Differences of 20, 50 and 100% (ES, effect size) for Comparisons
That Did Not Detect Significant Differences,
Lake Habitat, Moira River System, Spring 1999

Note: Power was based on ANOVA/ANCOVA model including all four lakes.

Figure 2.3-8
Ordination Plot of Lake Sites, Moira River System, Spring 1999
(Areas of symbols are proportional to arsenic concentration in sediment)

Seven of the eleven taxa with significant correlations to Dimension 1 were designated as metalsensitive in this study or belong to a generally metal-sensitive group, two were metal-tolerant and sensitivity of the remaining two taxa is unknown (Table 2.3-9). All metal-sensitive taxa and both tolerant taxa were negatively correlated with Dimension 1. There was no consistency in the metal-sensitivity of taxa correlated to Dimension 2.

Table 2.3-9
Spearman Rank Correlations Between NMDS Dimensions, Physical Factors and
Abundances of Common Invertebrates at Lake Sites, Moira River System, Spring 1999
(n=17; significant correlations [p<0.05] are identified by bolded coefficients)

^(a) T = metal-tolerant; S = metal-sensitive.

The communities of sites RL-2 and RL-3 did not "fit" the general trend along the metal concentration gradient (Figure 2.3-8). Sediments at one of these sites (RL-2) contained lead at a concentration (140 µg/g) that exceeded the Ontario lowest effect level guideline (31 µg/g), which may account for the position of this site on the ordination plot. However, there is no obvious explanation for the unusual community at RL-3, where sediment metal concentrations were similar to those at RL-1. At RL-3, the major difference in community structure was contributed by the dominance of bivalves, which accounted for >50% of total abundance.

Correlation analysis between environmental variables and NMDS dimensions found that sediment chemistry PC1 was significantly positively correlated with Dimension 1 (Table 2.3-9). Other environmental variables were not significantly correlated with either dimension. Thus, the positive correlations between sediment metal concentrations and Dimension 1 (which is generally negatively correlated with abundances of metal-sensitive taxa) were consistent with the expected trend in community structure with increasing sediment concentrations. Overall, results of the ordination support the interpretation that there was a shift in benthic community structure in lakes in response to sediment metal concentrations.

Sediment Quality Triad (SQT)

This section attempts to integrate information from each SQT component by examining the relationships among the biological, chemical and toxicological data. Concordance among these data sets was evaluated using correlation analysis of major patterns in each data set, as summarized by separate ordinations.

The biological data set was summarized as Dimensions 1 and 2 generated by NMDS. Two additional major community descriptors, total abundance and richness, were also included in the triad analysis to allow greater depth of interpretation. Only PC1 was used to represent the chemistry data set, because four of the metals (i.e., arsenic, cobalt, copper and nickel) had high loadings on this component and PC1 explained most of the variance in the chemistry data set (Table 2.3-10). Although silver was not included in the chemistry PCA, it was strongly correlated with PC1. Chemistry PC2 largely represented lead, which was not necessarily contributed by the former Deloro Mine Site, as indicated by elevated lead concentrations in Round Lake (a reference lake) sediments.

The first two PCs were used to represent the toxicity data set, because they both accounted for large proportions of the overall variance (Table 2.3-10). None of the four tests on lake sediments detected appreciable toxicity, with the possible exception of the reduction in H. azteca weight at two sites in Moira Lake and reductions of C. riparius survival in Consecon and Moira lakes (Table 2.3-11).

Table 2.3-10
Summary of the Sediment Chemistry and Toxicity PCAs,
Lake Habitat, Moira River System, Spring 1999

^(a) Pearson correlation between silver concentrations and PC scores; silver was not included in the PCA due to missing data.

Table 2.3-11
Sediment Toxicity Data Set, Lake Habitat,
Moira River System, Spring 1999

Significant correlations among the SQT components were limited to those between the biological and chemical data (Table 2.3-12). Richness of the benthic community declined and site scores on benthos Dimension 1 increased with increasing metal concentrations (Chemistry PC1), suggesting a metal-related shift in benthic community structure. The additional significant correlations were inter-correlations within the biological data set. The lack of significant correlations between the toxicity data and the other data sets is not surprising, in light of the general absence of toxicity of the sediments tested in the laboratory.

Table 2.3-12
Spearman Rank Correlations Among SQT Components,
Lake Habitat, Moira River System, Spring 1999
(Significant correlations [P < 0.05] are identified by bolded coefficients; n=17, except for
correlations with toxicity variables, where n=13)

The SQT analysis produced generally similar results to other analyses; i.e., there is some indication of a relationship between metal concentrations and the benthic invertebrate community, but this relationship is not pronounced.

2.3.2.2 River Habitat

Variation in Habitat

Field measurements revealed large differences in water depth and sediment TOC among sites and low to moderate variation in sediment particle size (Table 2.3-13). Conductivity was within a reasonably narrow range at most sites (150 to 250 µS/cm), with the exception of two sites where the minimum (50 µS/cm; SKR-1) and maximum (330 µS/cm; MR-13) values were measured. Water temperature was similar at all sites in the Moira River (10 to 15°C), but was higher in the reference rivers (17 to 20°C). Since the reference rivers were sampled a few days later than the Moira River, variation in air temperature is the most likely source of this difference.

Table 2.3-13
Variation in Field Water Quality Parameters, Depth and Sediment
Particle Size Among River Sites

The variation in the physical attributes of sampling sites was sufficient to account for part of the variation in benthic community structure. As in the lake data set, the physical variables of greatest concern were water depth and sediment TOC, because both varied widely. The greater depth at two sites (3.5 m; MR-10 and MR-12) may complicate the data analysis, because all other sites were up to 1.5 m deep (the single Young's Creek site was also deeper than most sites, but was not included in the analysis). Initial examination of the benthos data suggested that depth was a significant factor at the two deep sites in the Moira River. Total abundance, richness and abundances of a number of individual taxa were low at these two deep sites relative to others, despite widely differing metal concentrations in the sediments (Figure 2.3-2). A similar effect was not apparent for sediment TOC, which had a strongly bimodal distribution (TOC was >10% at five sites; >5% at 12 sites and intermediate at only one site).

Habitat variation was taken into account during data analysis by using the important habitat variables (depth, TOC, sediment particle size) as independent variables in the regression analysis. Diagnostic tests were run and scatter-plots were checked for each variable to see if the deep sites or those with high TOC were acting as outliers, and that outliers were removed if warranted.

Benthic Community Analysis

The benthic communities of the river sites had higher species richness than lake communities. The river data set had a total of 117 taxa, compared to 68 taxa in the lake data set. Chironomid midges and oligochaete worms dominated the river benthic communities, though mollusks and amphipods occasionally accounted for a significant fraction of the total abundance at a site (Figure 2.3-9). At the level of major group, community composition was highly variable among the reference sites and among exposed sites.

Total invertebrate abundance was not influenced by mine-related sediment metal concentrations (Figure 2.3-9). Regression analysis detected no significant relationship between sediment metal concentrations and total abundance (Table 2.3-14). Similarly, abundance at the three lower Moira River sites did not differ significantly from those in the Salmon River (Table 2.3-15). The variation in total abundance within the Moira River appeared related to habitat attributes rather than sediment metal concentrations. Total abundance varied moderately, but without a noticeable trend from Sites MR-1 to MR-8 (Figure 2.3-9), despite a sharp increasing trend in sediment metal concentrations from Site MR-5 to MR-8 (Figure 2.3-2). Habitat characteristics of these sites were similar (Table 2.3-13). Conspicuously low total abundances were found at sites MR-9, MR-10 and MR-12. Two of these sites (MR-10 and MR-12) were more than twice as deep as all other sites. The greater depth most likely accounted for the low abundances at these sites since metal concentrations differed widely between them (Figure 2.3-2; note that these sites were excluded from the regression analysis because of this depth-effect). The remaining site in this reach with low total abundance (MR-9) had elevated metal concentrations (as did MR-7 and MR-8), but also had much higher sediment TOC than all upstream sites. Since high TOC did not limit abundance at other river sites (e.g., SR-1 and MR-13, both with low metal concentrations), there is no obvious explanation for the low abundance at MR-9.

Figure 2.3-9
Total Invertebrate Abundance, Taxonomic Richness and Proportions of Major
Invertebrate Groups at River Sites, Moira River System, Spring 1999

Total Abundance

Richness

Community Composition

Table 2.3-14
Results of Linear Regressions Between Biological Variables
and Environmental Variables at River Sites

^(a)One outlier was removed.

^(b) PC1 refers to scores on sediment chemistry Principal Component 1 (Table 2.3-17).

Table 2.3-15
Results of Paired t-tests Comparing Biological Variables
Between the Salmon River and the Lower Moira River

^(a)% Difference is for Moira River relative to the Salmon River.

The very low abundance and richness at the single site in Young's Creek (YC-1, not included in the statistical analysis) most likely resulted from high metal levels originating from the Deloro Mine Site. Effects in Young's Creek can be assessed qualitatively, by comparing the benthic fauna of YC-1 with that of MR-10. Both sites were deep (3-3.5 m) and had high sediment TOC (14-15 %), but differed widely in sediment metal concentrations in grab samples as shown below:

The corresponding values in total abundance and richness of the benthic community were:

Clearly, the Young's Creek site supported a very poor fauna, even compared to another site with generally elevated metal concentrations (except for uranium). In terms of the entire river benthos data set, YC-1 had by far the lowest values for both abundance and richness, also suggesting a severe effect in Young's Creek (the ranges without YC-1 were 2,153 to 24,523 organisms/m² and 15 to 47 taxa). Moreover, toxicity of YC-1 sediment was pronounced with 100% mortality in Hyalella azteca and 86% mortality in Chironomus riparius (Appendix III). Therefore, based on a qualitative assessment, the benthos of YC-1 was severely affected by metals originating from the former Deloro Mine Site.

The trend in taxonomic richness along the Moira River generally reflected that in total abundance, with the exception that richness increased in a downstream direction in the upper Moira River (Figure 2.3-9). There was a significant positive relationship between sediment metal concentrations and richness (Table 2.3-14; significant regressions between PC1 and both richness variables), which was contrary to expectations. This result does not necessarily mean that increased metal concentrations resulted in increased numbers of taxa because the metals were largely in the more unavailable forms, such as the residual, oxides or organic fractions (Table 2.2-3). The lower Moira-Salmon River comparisons of richness did not detect significant differences (Table 2.3-15). Since richness is typically one of the most sensitive indicators of metal-related effects, the results of this study suggest that there is no indication of adverse effects. The low richness in Young's Creek was likely related to sediment metal concentrations; however, the sediments in Young's Creek were highly organic. Therefore, low oxygen concentrations may also have been a factor.

Abundances of chironomid midges and oligochaete worms also provided no consistent evidence of metal-related effects on the benthos of the Moira River. Both were highly variable among reference sites (Figure 2.3-10) and numbers did not clearly reflect habitat variation. Regression analysis found no significant relationships between sediment metal  concentrations and numbers in either group, when expressed as abundance or percentage of total abundance (Table 2.3-14). The lower Moira-Salmon River comparisons also did not detect significant differences (Table 2.3-15).

The oligochaete fauna of the study area was strongly dominated by tubificid worms (Figure 2.3-10). Hence, results of the regression analysis and lower river comparisons for this group were nearly identical to those for total oligochaetes (Tables 2.3-13 and 2.3-14).

Within the Chironomidae, sediment TOC and, in one instance, water depth influenced abundances and percentages in all major subfamilies and tribes, except the Tanypodinae (Table 2.3-14). Both sediment metal concentrations and TOC significantly influenced the Orthocladiinae (Table 2.3-14). The direction of the relationship with sediment concentration (positive) was consistent with the tendency of the Orthocladiinae to increase in abundance at locations with increased metal concentrations (Clements 1991). Comparisons of the lower Moira River with the Salmon River found only one significant difference, also in Orthocladiinae abundance (Table 2.3-15). Mean abundance of this group was higher in the Moira River that the reference river. Since the Orthocladiinae represented a small component of the chironomid fauna, accounting for <20% of total chironomids at all sites, the effect of sediment metal concentrations on chironomids can only be considered minor. In contrast, both the chironomid and oligochaete fauna of Young's Creek were nearly eliminated by the high metal concentrations originating at the former Deloro Mine Site.

Visual examination of the abundances of common invertebrate groups and individual taxa also did not provide consistent evidence of effects. Of the 39 taxa examined, only five exhibited spatial patterns consistent with effects in the Moira River, beginning at Site MR-5 (Figure 2.3-11). These invertebrates occurred in low numbers or were absent at the reference sites, and increased in abundance below the former Deloro Mine Site. This pattern was not consistent with the reported metal-sensitivity of at least one of these organisms: Caenis is a member of a highly metal-sensitive invertebrate group (mayflies) and has also been identified by individual studies as sensitive to certain metals (Clements 1991).

Figure 2.3-10
Abundances of Oligochaeta and Chironomidae at River Sites,
Moira River System, Spring 1999

Oligochaeta

Chironomidae

Figure 2.3-11
Abundances of Aulodrilus pluriseta , Caenis, Dicrotendipes, Conchapelopia and Cricotopus at
River Sites, Moira River System, Spring 1999

Regression analysis detected significant positive relationships between sediment metal concentrations and abundances of four of the above five taxa (A. pluriseta , Dicrotendipes, Conchapelopia, Caenis; Table 2.3-14), but no differences were found between lower Moira River sites and the Salmon River reference sites (Table 2.3-15). Abundance of Cricotopus was negatively related to sediment TOC and was significantly higher in the lower Moira River than in the Salmon River despite similar TOC levels in both rivers.

The combined abundances and percentages of metal-sensitive and metal-tolerant invertebrates (Figure 2.3-12) were not significantly related to sediment metal concentrations or habitat variables (Table 2.3-14), and the lower Moira/Salmon River comparisons yielded non-significant results (Table 2.3-15)

Overall, examination of the river benthos data set using univariate methods did not provide a consistent indication of adverse effects on benthic community below the former Deloro Mine Site. Total abundance, richness and abundances of major invertebrate groups and metal-sensitive invertebrates were not adversely affected by elevated metal concentrations in bottom sediments, but rather tended to reflect habitat conditions. Abundances of a few individual taxa were significantly related to sediment metal concentrations, but in all cases abundances increased with increasing metal levels.

Multivariate Analysis

Ordination of the river benthos data did not show clear separation of communities along the metal concentrations gradient encompassed by the river data set, but also did not preclude a slight shift in community structure in response to sediment metal concentrations. The stress of the final configuration was 0.19, which represents a relatively poor fit to the input data; however, this level of stress does not preclude a useful description of patterns in the input data (Clarke 1993). Upper Moira River sites in similar habitat and with elevated sediment metal levels (MR-5 to MR-8) tended to cluster together, but the Black River reference site (BR-1) was also close to this group (Figure 2.3-13). Salmon River sites and lower Moira River sites each formed unique, but relatively heterogeneous groups. Ordering of sites along Dimension 1 suggested some correspondence between benthos community structure and sediment metal concentrations. The one conspicuous outlier, Site MR-10 was one of the 3.5 m deep sites, which supported unusual benthic communities relative to all other sites.

Figure 2.3-12
Abundances and Percentages of Metal-sensitive and Metal-tolerant Invertebrates at River
Sites, Moira River System, Spring 1999

Figure 2.3-13
Ordination Plot of River Sites, Moira River System, Spring 1999
(Areas of symbols are proportional to arsenic concentration in sediment; upper Moira River
sites in similar habitat are identified by light-shaded symbols and site labels in italics, all other sites are shown as black circles)

There was no consistency with regard to metal-sensitivity in the groups of taxa that were significantly correlated with either biological dimension (Table 2.3-16). Similarly, there were no significant correlations between environmental variables and either dimension. Overall, results of the ordination suggest the potential for a slight shift in benthic community structure in response to sediment metal concentrations, but the type of community shift does not lend itself to a simple interpretation based on metal-sensitivity of individual benthic taxa.

Table 2.3-16
Spearman Rank Correlations between NMDS Dimensions, Physical Factors and
Abundances of Common Invertebrates at River Sites
(n=17; Significant correlations [p<0.05] are identified by bolded coefficients)

Sediment Quality Triad

The biological data set was summarized as Dimensions 1 and 2 generated by NMDS. Total abundance and richness were also included in the triad analysis. Only PC1 was used to represent the chemistry data set, because the four metals of greatest concern (i.e., arsenic, cobalt, copper, and nickel) had high loadings on this component and PC1 explained most of the variance in the chemistry data set (Table 2.3-17). As in the case of the lake data set, silver was not included in the chemistry PCA, but was strongly correlated with PC1. The first two PCs were used to represent the toxicity data set (Table 2.3-18), because they both accounted for large proportions of the overall variance (Table 2.3-17). The final correlation analysis was run after deleting the two deep sites (MR-10 and MR-12) to remove the additional variation contributed by the depth-effect described earlier.

Table 2.3-17
Summary of the Sediment Chemistry and Toxicity PCAs,
River Habitat, Moira River System, Spring 1999

^(a)Pearson correlation between silver concentrations and PC scores; silver was not included in the PCA due to missing data.

Table 2.3-18
Sediment Toxicity Data Set, River Habitat,
Moira River System, Spring 1999

The toxicity test results did not show a consistent relationship between effects on growth or survival and metal concentrations. For example, Hyalella growth and survival were substantially lower than controls in the Skootamatta River reference sample (SKR-1), the Moira River below Young's Creek sample (MR-8) with high metal concentrations and the Moira River at Latta sample (MR-14) with low metal concentrations. The Black River reference sample (BR-1) produced lower survival in Chironomus riparius, yet the Moira River below Young's Creek (MR-8) sample did not differ from controls. The only indication of any possible relationship between metal concentrations and toxicity test results was for the Moira River below Young's Creek sample (MR-8), where three of the four measures were lower than the control.

There were no significant correlations among the SQT components for the river data set (Table 2.3-19). In light of the high variability among reference sites and the general lack of statistically significant trends in the benthos with sediment metal concentrations in the univariate analysis, this result is not surprising. The absence of adverse effects at river sites is also consistent with the generally low bioavailability of sediment-bound metals in the Moira River system (Section 2.2). Metals released from the Deloro Mine Site are primarily in the more unavailable sediment fractions, with the exception of nickel and lead.

Table 2.3-19
Spearman Rank Correlations Among SQT Components, River Habitat,
Moira River System, Spring 1999
(n=17, except for correlations with toxicity variables, where n=13; No significant
correlations)

The only previous benthic survey in the Moira River system was carried out in 1969 (Owen and Galloway 1969). The authors reported adverse effects on the benthos of erosional habitats within a <10 km reach below the Deloro Mine Site. These effects consisted of reductions in the diversity and abundance of the bottom fauna and were attributed to arsenic toxicity. Concentrations of the metals of concern have declined substantially in river water since the 1960s, to levels that are below the PWQO in three seasons in a typical year, with the exception of cobalt, which usually exceeds the objective year-round (Section 2.1). Therefore, the lack of effects on depositional benthos in the Moira River in 1999 is consistent with the declining trends over time in both waterborne and sediment-bound metals in the upper Moira River.

2.3.2.3 General Summary

Lake Benthic Invertebrate Communities

Analysis of the lake benthos data detected some trends in benthic community structure that were consistent with the effects of elevated metal concentrations. However, despite the trends apparent upon visual examination of the data, statistical tests rarely detected significant differences between each of the exposed lakes (Moira and Stoco) and both reference lakes (Consecon and Round). This was largely the function of low statistical power in the pair-wise tests.

Total abundance was not affected by elevated metal concentrations and there were no readily apparent differences in benthic community composition between the reference lakes and the exposed lakes. In contrast, the richness data suggested a slight effect in Moira Lake and potentially in Bend Bay. The maximum richness in Moira Lake and richness at the single site sampled in Bend Bay were only slightly greater than the minimum value in the reference lakes, and total richness was significantly lower in Moira Lake relative to Consecon Lake. However, the pattern in richness within Moira Lake did not correspond with metal concentrations in the sediments.

Examination of the abundances of individual taxa and major invertebrate groups also suggested subtle effects of sediment metal concentrations, but low statistical power limited the ability to detect significant differences. For example, Oligochaete numbers were generally higher in the exposed lakes and were particularly high in Bend Bay, but there was no overall significant difference among lakes. A significant difference was found in total chironomid abundance between Consecon Lake and Moira Lake (with the lowest and highest metal concentrations, respectively), but not in any of the other comparisons. Additionally, differences among lakes in the percentages of major chironomid groups (which differ in metal sensitivity) were also consistent with metal concentrations in the sediments.

The strongest indication of metal-related effects on the benthic community was obtained by grouping taxa with known metal-sensitivity into sensitive and tolerant categories. The combined abundance and percentage of metal-sensitive invertebrates was significantly lower in Moira and Stoco lakes relative to both reference lakes. The trends for metal-tolerant taxa were less pronounced, largely because Consecon Lake communities included a moderate proportion of tolerant taxa.

Ordination of the lake benthos data showed separation of Consecon Lake communities from those of other lakes, but Round Lake sites were widely separated from each another. This analysis detected a gradual shift in community structure with increasing metal concentrations, although two of the three Round Lake sites detracted from this trend. Based on correlations of abundances of individual taxa with ordination axes, the change in community structure along the metal gradient appeared consistent with the metal sensitivity of common invertebrates in the lake communities.

Sediment Quality Triad analysis detected a significant correlation between benthic community structure and sediment chemistry. No significant correlations were found with sediment toxicity, as was expected, because lake sediments were not toxic to Hyalella azteca orChironomus riparius in the laboratory.

The lack of any major effects on species composition or abundance may be related to the relative bioavailability of the metals in Moira Lake and Stoco Lake. Sequential extraction of the upper 05 cm of sediment cores from the lakes showed that the majority of metals are in the more unavailable forms (such as iron and manganese oxides and metal sulphides).

River Benthic Invertebrate Communities

Analysis of the data generated by the river benthic survey did not provide an indication of effects from metals of concern originating at the former Deloro Mine Site, with the exception of the site located in Young's Creek which appeared severely impacted. Variation in total invertebrate abundance was relatively large within the Moira River and was related to habitat attributes rather than sediment metal concentrations. The spatial pattern in richness generally reflected that in total abundance, with the exception that richness increased with distance downstream in the upper Moira River. As a result, there was a significant positive relationship between sediment metal concentrations and richness, which cannot be considered as evidence of a metal-related effect. A detailed study of sediment chemistry completed during this survey revealed that the metals of concern were largely in the more unavailable forms, such as the residual, oxides or organic/sulphide fractions.

Examination of a large number of biological variables also did not provide an indication of metalrelated effects downstream from the former Deloro Mine Site. Occasional significant relationships were found between biological variables and sediment metal concentrations, largely in abundances of minor taxa. A number of these were also inconsistent with known metal sensitivity of the taxon involved. The combined abundance and percentages of metal-sensitive and metal-tolerant invertebrates was not significantly related to sediment metal concentrations or habitat variables. The lower Moira/Salmon River comparisons yielded non-significant results for all but two of the 25 biological variables tested.

Ordination of the river benthos data implied the potential for a slight shift in benthic community structure in response to sediment metal concentrations, but the type of community shift did not lend itself to a simple interpretation based on metal-sensitivity of individual benthic taxa. There was no consistency with regard to metal-sensitivity in the groups of taxa that were significantly correlated with either biological ordination axis. Similarly,there were no significant correlations between environmental variables and ordination axes.

The laboratory toxicity tests did not reveal a consistent relationship between effects on growth or survival of Hyalella azteca or Chironomus riparius and metal concentrations in sediments.

Sediment Quality Triad analysis did not find significant correlations between benthic community structure, sediment chemistry or sediment toxicity. The apparent lack of effects on the benthic community of the upper Moira River below the former Deloro Mine Site was most likely a function of low bioavailability of sediment-bound metals. The lack of effects in the lower Moira River were consistent with generally low metal concentrations resulting from settling of metalladen sediments in Moira and Stoco Lakes.

The lack of any significant effects on the benthic invertebrate community in the Moira River may be because the metals are not available for uptake by benthic organisms. Sequential extraction of the upper 0-5 cm of sediment cores from river depositional areas showed that most of the metals are in the more unavailable fractions (such as the residual, iron and manganese oxide, or organic/sulphide fractions).

References

Bedard, D., A. Hayton and D. Persaud. 1995. Ontario Ministry of the Environment Laboratory Sediment Biological Testing Protocol. Resources Branch, Ontario Ministry of the Environment, Queen's Printer for Ontario. ISBN 0-7729-9924-4.

Beltman, D.J., W.H. Clements, J. Lipton and D. Cacela. 1999. Benthic invertebrate metals exposure, accumulation, and community-level effects downstream from a hard-rock mine site. Environmental Toxicology and Chemistry, 18(2): 299-307.

Canfield, T.J., N.E. Kemble, W.G. Brumbaugh, F.J. Dwyer, C.G. Ingersoll and J.F. Fairchild. 1994. Use of benthic invertebrate community structure and the sediment quality triad to evaluate metal-contaminated sediment in the upper Clark Fork River, Montana. Environmental Toxicology and Chemistry, 13(12): 1999-2012.

Chapman, P.M. 1990. The sediment quality triad approach to determining pollution-induced degradation. Sci. Total Envir. 97/98: 815-825.

Chapman, P.M., B.A.. Power, R.N. Dexter and H.B. Anderson. 1991. Evaluation of effects associated with an oil platform, using the sediment quality triad. Environ. Toxicol. Chem. 10: 407-424.

Chapman, P.M., R.N. Dexter and E.R. Long. 1987. Synoptic measures of sediment contamination, toxicity and infaunal community composition (the Sediment Quality Triad) in San Francisco Bay. Mar. Ecol. Prog. Ser. 37:75-96.

Clarke. K.R. 1993. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology. 18: 117-143.

Clements, W.H. 1991. Community responses of stream organisms to heavy metals: a review of observational and experimental approaches. In Newman, M.C. and A.W. McIntosh. 1991. Metal Ecotoxicology, concepts and applications. Lewis Publishers, Chelsea Michigan.

Environment Canada. 1997a. Biological test method: test for survival and growth in sediment using the freshwater amphipod Hyalella azteca. Environment Canada, Conservation and Protection, Ottawa, Ontario. Report EPS 1/RM/33.

Environment Canada. 1997b. Biological test method: test for survival and growth in sediment using the larvae of freshwater midges (Chironomus tentans or Chironomus riparius). Environment Canada, Conservation and Protection, Ottawa, Ontario. Report EPS 1/RM/32.

Green, R.H., J.M. Boyd and J.S. Macdonald. 1993. Relating sets of variables in environmental studies: the sediment quality triad as a paradigm. Environmentrics, 4(4): 439-457.

Klemm, D.J., P.A. Lewis, R. Fulk and J.M. Lazorchak. 1990. Macroinvertebrate field and laboratory methods for evaluating the biological integrity of surface waters. U.S. Environmental Protection Agency, Office of Research and Development, Environmental Monitoring Systems Laboratory - Cincinnati, Ohio. EPA/600/4-90/030.

Long, E.T., and P.M. Chapman. 1985. A sediment quality triad: measures of sediment contamination, toxicity and infaunal composition in Puget Sound. Marine Pollution Bulletin. 16: 405-415.

Mance, G. 1987. Pollution threat of heavy metals in aquatic environments. Elsevier, New York.

Mori, C., A. Orsini and C. Migon. 1999. The impact of arsenic and antimony contamination on the benthic invertebrates in a minor Corsican river. Hydrobiologia, 392(1): 73-80.

Owen. G.E., and D.L. Galloway. 1969. Biological survey of the Moira River, 1969. Ontario Ministry of the Environment.

Paine, M. 1998. Volume II: CANMET/MMSL - INTEMIN Manual on statistical analysis of environmental data. Prepared by Paine, Ledge and Associates.

Poulton, B.C., D.P. Monda, D.P. Woodward, M.L Wildhaber and W.G. Brumbaugh. 1995. Relations between benthic community structure and metals concentrations in aquatic macroinvertebrates: Clark Fork River, Montana. Journal of Freshwater Ecology, 19(3): 277-293.

Winner, R.W., M.W. Boesel and P. Farrell. 1980. Insect community structure as an index of heavy-metal pollution in lotic systems. Can. J. Fish. Aquat. Sci. 37: 647-655.

2.4 Metal Concentrations in the Two Fish Sentinel Species in the Moira River System

2.4.1 Methods

The original study design included analysis of white sucker liver and benthic invertebrate samples for metals. The spring 1999 sediment and benthic invertebrate sampling program showed that there were insufficient numbers of benthic invertebrates at the reference and exposed sites to allow analysis of metals in invertebrate tissues. Furthermore, Golder had expressed some concerns regarding the reliance on white sucker liver samples to represent metals uptake. Golder also suggested substituting analysis of longnose dace for analysis of benthic invertebrates. Golder was asked to prepare recommendations regarding alternatives for tissue analysis. These recommendations were presented to the Steering Committee and approved. The methods outlined below represent the approved changes to the study design for examination of metal concentrations in biota from the Moira River system.

There were two main objectives for the analysis of metals, arsenic and radionuclides in fish tissue:

1. To obtain levels in fish flesh for use in the human health risk assessment; and
2. To obtain exposure information for the sentinel species (i.e., to assess the degree of metals uptake in white sucker and longnose dace.)

The first objective was met by the sport fish sampling conducted by the MOE. Sport fish monitoring data are summarized in Chapter 3 as part of the supporting information for the PQRA. The second objective was met by using the study design outlined below. The detailed rationale, study design and sampling protocols are in Appendix IV.

he main objective for the analysis of metals in white sucker and longnose dace was to obtain exposure information (i.e., to assess the degree of metals uptake). This was necessary in order to establish that these two species did, in fact, have elevated metal concentrations in the exposed study sites compared to the reference sites. This was a required first step in the evaluation of potential links between metal exposure in the sentinel species and effects.

Although logical, the application of a tissue residue approach to exposure characterization is complex. First, there should be an understanding of the mechanisms or modes of action for the chemical of concern, such that appropriate tissues are sampled. For example, in order to define robust residue-based effects relationships for cationic metals, it may be necessary to focus upon tissues or sites of action (e.g., gill) that are not routinely sampled (Wood et al. 1997). Alternatively, even though such relationships may not allow determination of mechanistic processes, it might be possible to derive useful correlations between some internal chemical residue for metals and observed effects (Wood et al. 1997). Therefore, the challenge for the Moira River Study was to be sure that the appropriate tissue was selected in order to allow a reasonably confident analysis of relationships between tissue residue levels and observed effects.

Selection of Appropriate Tissues for Analysis

Appropriate tissues vary with the metal. For example, the appropriate tissue for arsenic may be different than that for copper or zinc. Existing water quality data for the Moira River system indicated that the primary metal of concern was arsenic.

The majority of studies linking tissue level with effects use whole body analysis of arsenic (Jarvinen and Ankely 1999). For example, McGeachy and Dixon (1990, 1992) reported whole body levels of sodium arsenate along with survival or growth endpoints for rainbow trout. Cockell and Hilton (1988) reported "Carcass" residues of arsenic trioxide and survival and growth endpoints. Sorensen (1976) reported whole body, liver, gut, and muscle residues of sodium arsenate (As⁺⁵) in green sunfish along with survival endpoints. Gilderhus (1966) reported whole body levels of sodium arsenite (As⁺³) in bluegill, along with survival and growth endpoints.

The information for the other metals of concern confirmed that appropriate tissues varied with the particular metal. Copper accumulates to a greater extent in liver than in most other tissues, and several authors have made linkages between liver residues and effects on growth and survival (Jarvinen and Ankley 1999). However, linkages between whole body residues of copper and effects are also available in the literature (Jarvinen and Ankely 1999). Nickel appears to accumulate primarily in gill tissue and in kidney, although liver, muscle and whole body residues are also significant (Jarvinen and Ankley 1999). The only linkage between nickel tissue residue and effect is for survival. The only data on the linkage between silver tissue residues and effects are for whole body, gill and "internal organs" data (Coleman and Cearley 1974). Lead appears to accumulate preferentially in kidney and spleen, although liver and gill tissues also can contain significant amounts (Jarvinen and Ankley 1999). Linkages between tissue residues and effects have been made for liver, kidney, gill, spleen, and whole body data. There are no data linking tissue residues of cobalt and effects.

On the basis of the existing literature linking tissue residues with effects, whole body analyses were conducted in both white sucker and longnose dace. This allowed for comparison of results with a longer list of literature values and a more comprehensive database linking residues with effects than for liver analysis only. White sucker were collected from lakes and larger sections of the Moira River including:

• Moira River at Bend Bay (MR-11);
• Moira River downstream of Stoco Lake (MR-17, MR-14, MR-15 at thomasburg, Latta and Highway 37, respectively);
• Moira Lake (ML);
• Stoco Lake (SL);
• Consecon Lake (CL); and
• Round Lake (RL).

Longnose dace were collected from the following river sites:

• A reference site upstream of the Malone bridge (F-1 near MR-1)
• A reference site at the Malone bridge (F-2 near MR-3); and
• An exposed site near the Old Marmora Road (F-3 near MR-7).

These sites are different from those specified in the original study design (Appendix IV) because the actual longnose dace sample locations changed once the field crew was in the field. The collection of these fish species at these locations corresponded to the requirements of the fish population survey, which was conducted at the same time.

Ten specimens of white sucker and 10 composite samples of longnose dace were taken and analyzed for metals (including uranium). The sample size of 10 has been shown to be adequate to distinguish "exposed" from reference sites in other studies (radionuclides and metals) (Swanson 1985, U.S. EPA 1993). This sample size will distinguish differences of approximately 30-35% for uranium-series radionuclides (Swanson 1985) and approximately 10-20% for metals (assuming a mean: variance ratio of 0.5-1.0) (U.S. EPA 1993).

The 10 samples of white sucker were individual fish (gutted). Each of the 10 samples of longnose dace consisted of a minimum of 5 individuals (preferably 10 if obtainable).

Two size-classes of white sucker were selected for analysis (5 specimens from each size class). The first size-class represented the younger, faster-growing fish (with potentially greater metal uptake). The first size-class was 10-30 cm (fork length). The second size class was > 30 cm.

A detailed Technical Procedure for sampling fish for metals and radionuclides is included in Appendix IV. The Technical Procedure for producing homogenates of the fish samples is also included in Appendix IV.

Metal concentrations were compared among sampling sites using one-way ANOVA, followed by Tukey's tests. The metals examined for differences with site were: arsenic, cobalt, copper, lead, nickel, silver and uranium.

Bioaccumulation factors (BAFs) were calculated for water-to-fish and sediment-to-fish metal transfer. The average water concentration in 1999 for each fish sampling station was used for the water-to-fish BAFs. The sediment metal concentration from sediment grab samples in 1999 was used for the sediment-to-fish BAFs. In the case of the lakes, the data from all water quality and sediment quality stations in each lake were combined, since fish are mobile and will experience lake-wide water and sediment quality.

2.4.2 Results and Discussion

White Sucker

Arsenic was the most elevated metal in whole body white sucker samples compared to reference concentrations. Arsenic concentrations in whole body white sucker samples were significantly higher (P < 0.05) than reference locations at Bend Bay (MR-11), Moira Lake (ML), Stoco Lake (SL) and the Moira River at Latta (MR-14) (Figure 2.4-1). The highest overall concentrations were in Bend Bay samples, ranging from 0.26 to 1.09 µg/g. In contrast, reference fish arsenic concentrations ranged from 0.01 to 0.05 µg/g.

Figure 2.4-1
Average (with Standard Error) of Arsenic Concentration in
White Sucker Tissue Samples Taken at Each Location

Fish collected from Moira Lake, (ML), Consecon Lake (CL) and Round
Lake (RL) were collected from several locations within the lake.

Cobalt concentrations in whole body white sucker samples were significantly higher (P < 0.05) at Bend Bay (MR-11), Moira Lake (ML) and Latta (MR-14) (Figure 2.4-2). The highest concentrations were at Bend Bay, ranging from 0.04 to 0.2 µg/g. Reference concentrations ranged from 0.01 to 0.07 µg/g.

Figure 2.4-2
Average (with Standard Error) of Cobalt Concentration in
White Sucker Tissue Samples Taken at Each Location

Site
Fish collected from Moira Lake, (ML), Consecon Lake (CL) and Round
Lake (RL) were collected from several locations within the lake.

Copper concentrations in whole body white sucker samples were significantly higher (i>P /i><0.05) at Bend Bay (MR-11) and Thomasburg (MR-17) (Figure 2.4-3). Bend Bay concentrations ranged from 0.57 to 1.1 µg/g. The highest overall concentrations were in fish from the Thomasburg site, ranging from 0.98 to 1.9 µg/g. Reference concentrations ranged from 0.41 to 0.77 µg/g.

Figure 2.4-3
Average (with Standard Error) of Copper Concentration in
White Sucker Tissue Samples Taken at Each Location

Site
Fish collected from Moira Lake, (ML), Consecon Lake (CL) and Round
Lake (RL) were collected from several locations within the lake.

Lead and nickel concentrations in whole body white sucker samples were not elevated above reference concentrations at any of the Moira River system sites (Appendix IV).

Silver concentrations in whole body white sucker samples were relatively uniform across all reference and Moira River system sites, with the exception of Bend Bay (Figure 2.4-4). Concentrations at Bend Bay ranged from 0.004 to 0.01 µg/g. Concentrations at all other sites ranged from 0.001 to 0.005 µg/g.

Figure 2.4-4
Average (with Standard Error) of Silver Concentration in
White Sucker Tissue Samples Taken at Each Location

Site
Fish collected from Moira Lake, (ML), Consecon Lake (CL) and Round
Lake (RL) were collected from several locations within the lake.

Uranium concentrations in whole body white sucker samples were significantly higher (i>P /i><0.05) in Stoco Lake than at reference locations (Figure 2.4-5). The significant difference was produced by only a small actual difference. Concentrations in Stoco Lake white sucker ranged from 0.001 to 0.004 µg/g whereas at the other sites, concentrations ranged from 0.0005 to 0.003 µg/g.

Figure 2.4-5
Average (with Standard Error) of Uranium Concentration in
White Sucker Tissue Samples Taken at Each Location

Site
Fish collected from Moira Lake, (ML), Consecon Lake (CL) and Round
Lake (RL) were collected from several locations within the lake.

There was no relationship between fish size and metal concentrations in whole body samples. Smaller white sucker (<30 cm length) did not have significantly lower or higher metal concentrations than larger white sucker (> 30 cm length) (Appendix IV).

There are few previous data on metal concentrations in white sucker from the Moira River system. Azcue and Dixon (1994) reported that arsenic concentrations in 4 composite samples of white sucker from Moira Lake ranged from 0.099 to 0.166 µg/g. Moira Lake white sucker samples in this study ranged from 0.20 to 0.76 µg/g.

Arsenic concentrations in whole white sucker from the Moira River system are similar to concentrations found in northern Saskatchewan where uranium mining activity has resulted in elevated arsenic concentrations because arsenic occurs in the uranium ore (Table 2.4-1). Collins Bay in Wollaston Lake, Saskatchewan, is the site of an open-pit uranium mine that extended out into the bay. Arsenic concentrations in lake whitefish from this location averaged 0.22-0.33 µg/g. Arsenic concentrations in whole white sucker from Bend Bay ranged from 0.26 to 1.09 µg/g. Thor Lake, N.W.T. and South McMahon Lake were The sites of baseline studies prior to mining development. Arsenic concentrations in lake whitefish and white sucker from these locations are very similar to the reference concentrations recorded in this study (Table 2.4-1).

Table 2.4-1
Comparison of Arsenic and Metal Concentrations Measured
in Fish Tissue at Some Canadian Sites

Copper concentrations in whole white sucker from the Moira River system are somewhat lower that those reported from ten lakes in the Sudbury region, and much lower than concentrations measured in several fish species from Baie du Doré and Toronto Harbour (Table 2.4-1). The copper concentrations found in this study appear to be most similar to those measured in fish from Wollaston Lake and South McMahon Lake, Saskatchewan and Thor Lake, N.W.T. (Table 2.4-1).

Lead concentrations in white sucker in this study are much lower than concentrattions measured in several fish species from Baie du Doré and Toronto Harbour (Table 2.4-1). Concentrations were similar to those found in several fish species in the Great Lakes (Table 2.4-1). Lead does not usually accumulate in significant quantities in edible portions of fish, except in cases of extreme pollution (Moore and Ramamoorthy 1984).

Nickel concentrations measured in white sucker in this study are much lower than in lake whitefish from Collins Bay of Wollaston Lake, Saskatchewan (Table 2.4-1). Nickel is a significant component of the uranium ore mined at Collins Bay. The nickel concentrations found in this study are similar to baseline concentrations measured in South McMahon Lake, Saskatchewan and Thor Lake, N.W.T. (Table 2.4-1).

Most literature on cobalt uptake in fish is for the marine environment (e.g., Eisler and LaRoche 1972, Pentreath 1973, Udea and Nakahara 1983), describes the results of whole-lake experiments tracing cobalt-60 in food chains (Klaverkamp et al. 1983), or describes uptake in laboratory studies (Feldt and Melzer 1978, Baudin et al. 1990). There are few studies of cobalt uptake in wild fish. The average cobalt concentrations measured in carp (Cyprinus carpio) from the Danube River and Danube Canal was 0.18 µg/g (± 0.08) (Rehwoldt et al. 1976). This mean falls within the range of cobalt concentrations measured in white suckers from Bend Bay in this study (0.04-0.2 µg/g). The Danube River receives significant effluent input from a wide variety of industries. The Moira River has received input from the Deloro Mine Site for decades. Despite this, cobalt concentrations in fish in these two locations are relatively low.

Uranium concentrations in whole white sucker from the Moira Lake system were similar to reference (unaffected by mining) concentrations measured in northern Saskatchewan and Thor Lake, N.W.T. (Table 2.4-1). They were much lower than concentrations reported in white sucker in Beaverlodge Lake, Saskatchewan (Table 2.4-1). Beaverlodge Lake received tailings effluent from the Eldorado Nuclear uranium mine for almost 30 years.

Bioaccumulation factors (BAFs) for uptake in whole body white sucker from water were highest for copper (averaging 937) and lowest for uranium (averaging 5.6) (Appendix IV). Arsenic BAFs were relatively low (averaging 22).

Bioaccumulation from sediment was also highest for copper (averaging 0.01) and lowest for lead and uranium (averaging 0.0005) (Appendix IV). Sediment BAFs were an order of magnitude higher at Bend Bay for arsenic, cobalt, nickel and silver than at any of the other sites.

Bioaccumulation factors from water-to-sucker did not vary substantially with site. There were no consistent trends (such as higher BAFs at the exposed sites versus the reference sites).

Sediment-to-fish transfer was higher at the Bend Bay site for some of the metals (arsenic, cobalt, nickel and silver). There was a similar range of sediment-based BAFs at all other sample sites.

Longnose Dace

Arsenic, silver and nickel were the only metals that had significantly higher concentrations (P < 0.05) in longnose dace at the site downstream of the Deloro Mine Site compared to the two upstream sites (Figures 2.4-6 to 2.4-8). Arsenic was the most elevated relative to the reference data. Arsenic ranged from 0.87 to 1.9 µg/g in the downstream dace samples, whereas the upstream reference dace concentrations were from 0.05 to 0.14 µg/g. Nickel concentrations ranged from 0.28 to 0.64 µg/g at the downstream site. Upstream nickel concentrations ranged from 0.16 to 0.48 µg/g. Downstream silver concentrations were from 0.093 to 0.07 µg/g whereas reference concentrations were from 0.005 to 0.01 µg/g.

Figure 2.4-6
Average (with Standard Error) of Arsenic Concentration in
Longnose Dace Tissue Samples Taken at Each Location

Figure 2.4-7
Average (with Standard Error) of Silver Concentration in
Longnose Dace Tissue Samples Taken at Each Location

Figure 2.4-8
Average (with Standard Error) of Nickel Concentration in
Longnose Dace Tissue Samples Taken at Each Location

Arsenic concentrations in longnose dace from the downstream site were about an order of magnitude higher than concentrations measured in rock bass, spottail shiner, bluntnose minnow,golden shiner and emerald shiner sampled from Moira Lake in 1991 (Azcue and Dixon 1994). The concentrations from the downstream site were similar to concentrations in creek chub fromMoira Lake in 1991 (Azcue and Dixon 1994).

Since methods were very similar, the most likely explanation for the differences between theresults of this study and the results reported by Azcue and Dixon is that there are markeddifferences among fish species in the rates of arsenic uptake and elimination (Foley et al. 1978). Azcue and Dixon observed inter-species differences, even among closely related species. For example, mean arsenic concentrations in spottail shiner, golden shiner and emerald shiner were 0.03 µg/g, 0.167 µg/g and 0.04 µg/g, respectively. Highest arsenic concentrations were found in creek chub (2.36 µg/g) and lowest concentrations were found in northern pike (0.025 µg/g). Azcue and Dixon did not find differences between benthic-feeding fish and pelagic feeders. They suggested that the absence of elevated arsenic concentrations in benthic feeders was related to the very low percentage of arsenic in the most biologically-available form (the exchangeable fraction) in Moira Lake sediments. Although this study did not compare benthic and pelagic fish, it also found that arsenic was primarily in the less biologically-available fractions.

Bioaccumulation factors for water-to-fish transfer of metals to longnose dace were highest for opper (averaging 1921) and lowest for lead (averaging 3). Arsenic and silver showed intermediate BAFs, averaging 70 and 92, respectively. Cobalt and nickel BAFs averaged 600 and 208, respectively. Uranium BAFs averaged 4.

Bioaccumulation factors for sediment-to-fish transfer of metals to longnose dace were highest for copper (averaging 0.03) and lowest for lead and uranium (averaging 0.0006 and 0.0009 respectively) (Appendix IV).

There was some variability in BAFs among the three sample sites; however, the exposed site (downstream of Deloro) did not consistently produce higher BAFs

Comparison of BAFs in This Study with Other Data

The BAFs found in this study generally fall within ranges reported in other studies for arsenic, copper and uranium uptake in wild fish (Moore and Ramamoorthy 1984; Swanson 1985, 1989).

Long-term laboratory exposures to arsenic in the diet produced BAFs that were similar to those observed for sediment-to-fish in this study. For example, exposure of rainbow trout for 56 days to arsenic in the diet produced BAFs ranging from 0.004 to 0.02 (Cockell and Hilton 1988). The BAFs for sediment-to-fish in the Moira River system ranged from 0.0007 to 0.02.

The BAFs for cobalt in this study were greater than those reported by Harrison et al (1990) for fathead minnows and lake trout 63 days after cobalt was added to an experimental lake. The BAFs for fathead minnow and lake trout in were 190 and 58, respectively (Harrison et al. 1990). The BAFs for longnose dace in this study ranged from 33-967 (Appendix IV). The BAFs for white sucker in this study ranged from 40-120 (Appendix IV).

Long-term laboratory exposures to copper in the diet also produced somewhat similar BAFs to those observed for sediment-to-fish in this study. The BAFs for rainbow trout exposed to copper sulphate in the diet ranged from 0.005 to 0.5, depending upon tissue (Lanno et al. 1985). The BAFs for copper transfer from sediment-to-fish in the Moira River system ranged from 0.002 to 0.04.

The BAFs for nickel and silver from this study are higher than BAFs calculated from data presented in laboratory studies (data from comparable field studies could not be located). Laboratory exposures produced nickel BAFs for water-to-fish ranging from 0.8 to 4 for rainbow trout and 5 to 129 for carp (Calamari et al. 1982; Sreedevi et al. 1992). The nickel BAFs for in this study ranged from 61 to 208. Laboratory exposures produced silver BAFs for water-to-fish ranging from 10 to 17 in largemouth bass (Coleman and Cearley 1974). Silver BAFs in this study ranged from 12-202.

The BAFs for lead from this study are lower than BAFs calculated from laboratory studies. Laboratory exposure produced BAFs for water-to-fish ranging from 13-1083 for rainbow trout and 3 to 295 for brook trout (Holcombe et al. 1976; Hodson et al. 1978). The BAFs in this study ranged from 2 to 15.

While interesting in their own right, the main purpose of the whole body metal concentration data in this report is to confirm exposure in the sentinel fish specimens collected from the "exposed" sites. The data clearly show that that the fish collected for this study have been exposed to elevated metal concentrations.

Relationship Between Metal Concentrations in Fish and Effects

On the basis of the whole body concentrations observed in fish from the Moira River system, the literature indicates that effects would not be expected. Existing information compiled by Jarvinen and Ankley (1999) linking tissue residues and effects in laboratory exposures shows that effects occur at whole body concentrations that are considerably higher than observed in this study (Table 2.4-2). In contrast to the whole body concentrations in Table 2.4-2, the highest arsenic concentrations found in white sucker and longnose dace in this study were 0.26 to 1.09 µg/g and 0.87 to 1.9 µg/g, respectively. oncentrations of other metals in fish from this study were even lower than arsenic, with the exception of copper where the highest concentrations in white sucker were from 0.57 to 1.1 µg/g.

This study went beyond measurement of metal concentrations in fish to a study of fish health parameters in the two sentinel species (Section 2.5). The results of the fish health study confirmed the prediction that no effects would be found.

Summary

The measurement of metal concentrations in whole body fish samples confirmed the presence of elevated concentrations in fish downstream of the Deloro Mine Site. Thus, the white sucker and longnose dace studied for effects did, indeed, have greater exposure to the metals of concern than reference fish.

While elevated, metal concentrations in longnose sucker from the Moira River system were not as high as in fish from other mining areas, and were much lower than areas affected by a broad spectrum of contaminant sources (e.g., Toronto Harbour). The exception was arsenic concentrations, which were similar to concentrations found in fish near uranium mines.

Metal concentrations in longnose dace were higher than concentrations in some other small fish species from the Moira River sampled in 1991. These differences are likely due to differences among fish species in the rates of arsenic uptake and elimination.

Table 2.4-2
Whole Body Metal Concentrations and Effects in Various Fish Species
(Jarvinen and Ankley 1999)

References

Azcue, J.M. and D.G. Dixon. 1994. Effects of past mining activities on the arsenic concentration in fish from Moira Lake, Ontario. J. Great Lakes Res. 20(4): 717-724.

Baudin, J.P., A.F. Fritsch and J. Georges. 1990. Influence of labelled food type on the accumulation and retention of ⁶⁰Co by a freshwater fish, Cyprinus carpio L. Water, Air and Soil Pollution. 51:261-270.

Bradley, R.W. and J.R. Morris. 1986. Heavy metals in fish from a series of metal-contaminated lakes near Sudbury, Ontario. Water, Air and Soil Pollution 27: 341-354.

Brown, J.R. and L.Y. Chow. 1977. Heavy metal concentrations in Ontario fish. Bull. Environ. Contam. Toxicol. 17(2): 190-195.

Calamari, D., G.F. Gaggino and G. Pacchetti. 1982. Toxicokinetics of low levels of Cd, Cr, Ni and their mixture in long-term treatment on Salmo gairdneri Richardson. Chemosphere11:59-70.

Cockell, K.A. and J.W. Hilton. 1988. Preliminary investigations on the comparative chronic toxicity of four dietary arsenicals to juvenile rainbow trout. Aquat. Toxicol. 12: 73-82.

Coleman, R.L. and J.I. Cearley. 1974. Silver toxicity and accumulation in largemouth bass and bluegill. Bull. Environ. Contam. Toxicol. 12:53-61.

Eisler, R. and G. LaRoche. 1972. Elemental composition of the estuarine teleost Fundulusheteroclitus (L.). J. Exp. Mar. Biol. Ecol. 9:29-42.

Feldt, W. and M. Melzer. 1978. Konzentrationsfatoren der Elemente Kobalt, Mangan, Eisen, Zink und Silber für Fische. Arch. Fisch. Wiss.29:105-112.

Foley, R.E., F.R. Spotila, J.P. Giesy and S.H. Wall. 1978. Arsenic concentrations in water and fish from Chautauqua Lake, New York. Environ. Biol. Fish. 3: 361-367.

Gilderhus, PA. 1966. Some effects of sublethal concentrations of sodium arsenite on bluegills and the aquatic environment. Trans. Am. Fish. Soc 95: 289-296.

Harrison, S.E., J.F. Klaverkamp, and R. H. Hesselin. 1990. Fates of metal radiotracers added to a whole lake: Accumulation in fathead minnow (Pimephales promelas) and lake trout (Salvelinus namaycush). Water, Air and Soil Pollution. 52:277-293.

Hodson, P.V., B.R. Blunt and D.H. Spry. 1978. Chronic toxicity of water-borne and dietary lead to rainbow trout in (Salmo gairdneri) in Lake Ontario water. Water Res. 12-869-878.

Hodson, P.V., D.M. Whittle, P.T.S. Wong, U. Borgmann, R. L. Thomas, Y.K. Chau, J.O. Nriagu, and D.J. Hallett. 1984. Lead contamination of the Great Lakes and its potential effects on aquatic biota. IN: J.O. Nriagu and M.S. Simmons (eds.) Toxic Contaminants in the Great Lakes. John Wiley and Sons, Inc.

Holcombe, G.W., D.A. Beniot, E.N. Leonard and J.M. McKim. 1976. Long-term effects of lead exposure on three generations of brook trout (Salvelinus fontinalis). J. Fish. Res.Board Can. 33:1731-1741.

Jarvinen, A.W. and G.T. Ankley. 1999. Linkage of effects to tissue residues: development of a comprehensive database for aquatic organisms exposed to inorganic and organic chemicals. SETAC Technical Publication Series. SETAC Press. Society of Environmental Toxicology and Chemistry, Pensacola, FL.

Kilgour, B.W., C.B. Portt and D.G. Dixon. 1999. Moira River Impact Study: Detailed design to determine the impact of the former Deloro Mine Site on the Moira River System. Final Report submitted to Ontario Ministry of Environment.

Klaverkamp, J.F., M.A. Turner, S.E. Harrison, and R. H. Hesslein. Fates of metal radiotracers added to a whole lake: accumulation in slimy sculpin (Cottus cognatus) and white sucker (Catostomus commersoni). The Science of the Total Environment. 28:119-128.

Lanno, R.P., S.J. Slinger and J.W. Hilton. 1985. Maximum tolerable and toxicity levels of dietary copper in rainbow trout (Salmo gairdneri Richardson). Aquaculture 49: 257-268.

McGeachy, SM and DG Dixon. 1992. Whole-body arsenic concentrations in rainbow trout during acute exposure to arsenate. Ecotoxicol. Environ. Saf. 24: 301-308.

McGeachy, SM and DG Dixon. 1990. Effect of temperature on the chronic toxicity of arsenate to rainbow trout. Can. J. Fish Aquat. Sci. 47: 2228-2234.

Moore, J.W. and S. Ramamoorthy. 1984. Heavy metals in natural waters: Applied monitoring and impact assessment. Springer-Verlag, New York.

Pentreath, R.J. 1973. The accumulation from sea water of ⁶⁵Zn,⁵⁴Mn, ⁵⁸Co and ⁵⁹Fe by the thornback ray, Raja clavata L. J. Exp. Mar. Biol. Ecol. 12: 327-334.

Rehwoldt, R. 1976. Distribution of selected metals in tissue samples of carp, Cyprinus carpio Bulletin of Environmental Contamination and Toxicology. 15(3): 374-377.

Sorensen, EMB. 1976. Thermal effects of the accumulation of arsenic in green sunfish. Arch Environ. Contam. Toxicol. 4: 8-17.

Sreedevi, P., A. Suresh, B. Sivaramakrishna, B. Prabhavathi and K. Radhakrishnaiah. 1992. Bioaccumulation of nickel in the organs of the freshwater fish, Cyprinus carpio , and the freshwater mussel, Lamellidens marginalis, under lethal and sublethal nickel stress. Chemosphere 24: 29-36.

Swanson, S.M. 1985. Food-chain transfer of U-series radionuclides in a northern Saskatchewan aquatic system. Health Physics 49(5): 747-770.

Swanson, S.M. 1989. Results of the study of effects on the Rabbit Lake minewater spill, November 6-7, 1989. Saskatchewan Research Council Report  No. E-2130-2-E-89. Prepared for Saskatchewan Environment, Mines Pollution Control Branch, Regina, Saskatchewan

U.S. EPA. 1993. Guidance for assessing chemical contaminant data for use in fish advisories. Environmental Protection Agency Report EPA 823-R-93-002.

Udea, T. and M. Nakahara. 1983. Accumulation of Co by marine fish. Bulletin of the Japanese Society of Scientific Fisheries. 49(4):651-654.

Wood, C.M., WS Adams, GT Ankley, DR DiBona, SN Luoma, RC Playle, WC Stubblefield, HL Bergman, RJ Erickson, JS Mattice and CE Schlekat. 1997. Environmental toxicology of metals. Pp. 31-51 In Bergman, HL, Dorward-King EJ (eds.) Reassessment of metals criteria for aquatic life protection. SETAC Press, Pensacola, FL.

2.5 Fish Populations

A sentinel fish survey was the recommended approach for assessing the potential effects of minerelated metal concentrations on fish populations within the Moira River system (Kilgour et al. 1999). Sentinel species monitoring is a common and effective approach used to evaluate the effect of stressors (e.g., metals) on wild fish populations. The sentinel species approach evaluates the performance (e.g., growth, condition, reproductive parameters) of a sentinel fish species inhabiting a particular site of interest (e.g., downstream of the Deloro Mine Site), relative to reference and/or historical performance data. The underlying premise of the approach is that the status of the sentinel species is a reflection of the overall condition of the aquatic environment in which the fish resides (Munkittrick 1992).

The receiving environment within the Moira River system consists of both river and lake habitat. Consequently, separate sentinel species surveys were required because physical/chemical conditions and fish communities differ greatly between the river and the lake habitats.

2.5.1 Methods

2.5.1.1 Sentinel Species and Sampling Locations

Lake Habitat

White sucker (Catostomus commersoni) was chosen as the sentinel species for lake habitat of the Moira River system. White sucker was selected as a suitable sentinel species because:

• It occurs in both Moira Lake and Stoco Lake (Kilgour et al. 1999);
• It is a benthic feeder and utilizes deeper areas of lakes where metal-containing sediments accumulate (i.e., increased exposure);
• Gonadal development in the fall is sufficiently advanced to measure reproductive parameters such as gonad size and fecundity; and
• It has proven to be an effective sentinel species in numerous other studies (e.g., McFarlane and Franzin 1978, Swanson 1982, Trippel and Harvey 1987, Munkittrick and Dixon 1988, McMaster et al. 1991, Miller et al. 1992, Hodson et al. 1992, Gibbons et al. 1998).

The sampling locations for the sentinel survey were Moira Lake and Stoco Lake (exposed to mine-related metal concentrations), and Round Lake and Consecon Lake (reference locations) (Figure 2.5-1). However, capture success of white sucker in Stoco Lake during the fall collection trip was limited (see Section 2.5.2.1), and the Moira River at Bend Bay was included as an alternate exposure site. Photos of each lake and of Bend Bay are provided in Appendix V.

River Habitat

As suggested by Kilgour et al. (1999), and confirmed during a reconnaissance field survey, longnose dace (Rhinichthys cataractae) was selected as the sentinel species for the river habitat of the Moira River system. Longnose dace was a suitable sentinel species because:

• It is abundant and widely distributed in the upper reaches of the Moira River near Malone and the Deloro Mine Site;
• It is a riverine species which completes its life cycle within the Moira River (i.e.,  maximum exposure);
• It exhibits limited potential for large-scale movement among reference and exposure sites relative to larger fish species; and,
• Gonadal development in the fall is sufficiently advanced to measure reproductive parameters.

Longnose dace were collected from two reference sites upstream of the Deloro Mine Site and an exposure site immediately downstream of the mine (Figure 2.5-1). The general locations and UTM coordinates of each sampling site are provided in Table 2.5-1. The sites differ somewhat from sites used for sediment and invertebrate sampling. This is because the longnose dace sites had to be good longnose dace habitat, rather than the depositional areas that were chosen for sediment and invertebrate sampling.

Each site was dominated by riffle/run habitat with bedrock and cobble/boulder substrates (see photos - Appendix V). During the study period, all sites were influenced by low-water levels. This was particularly noticeable at reference site F-2. Rainfall near the end of the study period, and immediately prior to sampling the exposure site (F-3), resulted in a considerable increase in water level and flow. However, it is unlikely that whole-organism characteristics of longnose dace from site F-3 would have been affected by such a recent change in habitat conditions.

Figure 2.5-1 MOIRA RIVER STUDYWHITE SUCKER AND LONGNOSE DACESAMPLING LOCATIONS

Table 2.5-1
General Location and UTM Co-ordinates of Sampling Sites for
Longnose Dace, Moira River, Fall 1999

^a) Universal Transverse Mercator (UTM).

2.5.1.2 Fish Collections

Sentinel fish collections on the Moira River system were conducted during October 3-22, 1999 A minimum of 20 mature males and 20 mature females was the target sample size for each sample site. White sucker were collected with the assistance of the Ontario Ministry of the Environment (MOE). Sucker were collected using the MOE electrofishing boat (Smith-Root SR-12), as well as the Golder Associates Ltd. portable boat-electrofishing unit (Smith-Root GPP portable). Electrofishing effort was concentrated on the margin/backwater habitat of each lake, although passes through deeper pelagic zones were also made. Trap nets and gill nets (15.4 m x 0.9 m x 9 or 10.25 cm mesh size, 2-3 h sets) were also used to sample habitats that are difficult to sample using electrofishing gear. Neither netting technique was successful in capturing white sucker. Longnose dace were collected using a backpack electrofisher (Smith-Root Type VII) and dip nets. Collection of white sucker and longnose dace was conducted following detailed methods outlined in Golder Technical Procedure 8.1-3 "Fish Inventory Methods" (Appendix V). All fish species captured were identified and enumerated.

Each fish was anaesthetized prior to being processed by placing it in a 0.1 g/L solution of tricaine methanesulfonate (MS-222) for 30-60 seconds. Fish were removed from the anaesthetic and measured for total length, fork length, and fresh body weight. Fish were sacrificed by a sharp blow to the head and dissected to measure carcass (i.e., eviscerated) weight, gonad weight and liver weight. External and internal pathology examinations were conducted for each fish. Other information recorded included: gender, stage of sexual maturity, and results of a visual assessment of gut contents. All fish were processed according to methods described in Golder Technical Procedure 8.16-0 "Fish Health Assessment- Metals" (Appendix IV).

A sample of ovarian tissue from mature female sucker was removed from the mid-section of the right ovary, weighed (approx. 0.75-1.0 g) and preserved in Gillson's solution for fecundity analyses. For longnose dace, both ovaries were weighed and preserved in Gillson's solution. The total number of eggs per sample was counted, and these results were used to estimate the total number of eggs per fish (total fecundity). As well, the average diameter of 30 individual eggs/female was determined as an estimate of egg size.

Scales and pectoral fin rays were removed from white sucker for ageing (i.e., annulus count). Longnose dace were aged using otoliths and scales. Otoliths from longnose dace were aged using the "crack-and-burn" procedure. Briefly, the otolith is cut along the transverse axis and held in an alcohol flame to char the cut surface and enhance the visual identification of growth annuli. For both species, scales were used as secondary ageing structures and read, when required, from acetate imprints. Ageing structures were prepared and read by Northshore Environmental Services, Thunder Bay, Ontario. Ages were estimated following procedures outlined in Mackay et al. (1990).

2.5.1.3 Data Analyses

Statistical analysis of sentinel fish species data was done using SYSTAT® statistical software (Wilkinson 1990). Analysis of Variance (ANOVA) was used to compare fork length, body weight, egg diameter and age estimates among sites. Estimates of size-at-age (fork length vs. age), condition (carcass weight vs. fork length), gonad size, fecundity and liver size were evaluated using ANCOVA. With the exception of size-at-age, carcass (i.e., eviscerated) weight was used as a covariate to adjust for any differences in body size. Using carcass weight instead of body weight eliminated possible confounding effects of altered organ weight (e.g., gonad weight, liver weight) on the interpretation of variables related to body weight. An assumption of the ANCOVA model is that the slopes of the regression lines are equal between lakes/sites. Therefore, differences in slopes were tested prior to conducting the ANCOVA. Generally, ANCOVA is fairly robust even when slopes are not equal, so slopes were considered different when p<0.01 (Paine 1998). Following significant ANOVA or ANCOVA (intercepts) results, Tukey's procedure was used to identify specific differences among sites. Data were log₁₀ transformed where appropriate and sexes were analyzed separately.

Outliers

Data were initially screened for potential outliers by visual examination of scatterplots and box plots. The procedure for removing outliers was based on the evaluation of Studentized Residuals (SR). Observations that were more than four standard deviations (i.e., SR>4) from the cell mean were removed and the analysis run again. If any new outliers (SR>4) occurred, they were also removed and the analysis was redone. No further outliers were deleted after this point. Adopting SR>4 as a cut-off is considered conservative, as greater than 99% of Studentized Residuals were expected to have lower values (Grubbs 1971).

Power Analysis

Power analysis was used to evaluate the possibility of false negative results i.e., concluding that no differences in fish performance exists when in fact they do. Peterman (1990) has argued that the consequences of false negatives may be greater than the dangers of false positives (i.e., concluding a difference when none exist), and that environmental impact assessments should have adequate statistical power to detect meaningful differences. In other words, it is necessary to determine whether a comparison that did not show a statistical difference actually had sufficient power (i.e., statistical strength) to detect a given difference, if one existed.

Statistical power is a function of sample size, variability and the magnitude of difference (i.e., effect size) one wishes to detect. The effect size is not easily defined. The Environmental Effects Monitoring program for the pulp and paper industry have set an effect size of ±25% difference in gonad weight (Environment Canada 1997). For the purposes of this study, parameter-specific sample sizes were estimated for an effect size of 20, 30, and 50% (i.e., differences between sites). The mean squared error (MSE) term from the ANOVA or ANCOVA statistical model provided the estimate of among-site-variance for each fish parameter. The probability of making a Type I error (i.e., a) was set at a=0.05.

Because the study design for white sucker and longnose dace consisted of more than two sites, simple power equations comparing two samples could not be used. Cohen (1988) provides comprehensive methods for power analyses for more than two samples, and for a variety of statistical tests (e.g., ANOVA, ANCOVA). Power analyses were conducted using G*Power software (Faul and Erdfelder 1992), that performs computations based on methods described by Cohen (1988).

Statistical comparisons were considered to have sufficient power (P, probability of detecting an effect size) when P>0.80 (Environment Canada 1997).

2.5.2 Results and Discussion

2.5.2.1 White Sucker

Capture success of mature white sucker was variable among sites (Table 2.5-2). In general, capture success of female white sucker was higher than male sucker, although in Stoco Lake and Bend Bay, both sexes were difficult to collect in high numbers. Relative to the reference lakes, the greatest capture success of adult sucker was in Moira Lake – one of the sites exposed to metals from the Deloro Mine Site (see Sections 2.1, 2.2). A summary of incidental fish species captured in each lake during the collection of white sucker is provided in Appendix V.

Table 2.5-2
Total Number of Female, Male and Immature White Sucker
Captured from Lake Habitat, Moira River System, Fall 1999

Due to limited sample sizes, statistical analyses of fish parameters were restricted to data from Consecon, Round and Moira lakes. Emphasis was placed on analyses of female sucker because of larger sampler sizes. Analyses of male sucker were done to see if they supported results of female sucker, although definitive conclusions were not possible due to low sample sizes. Data from Bend Bay (males and females) and Stoco Lake (males only) were included in all graphs of fish parameters to provide some perspective on how they compared with exposure/reference sites.

Fish Measurements

Based on internal/external pathology examinations of white sucker, the incidence of abnormalities ranged from 95-100% per site (Table 2.5-3). Most observed abnormalities were associated with the gills (marginate lamellae, fish lice infestation) or related to internal parasitism. There was no clear evidence indicating that the prevalence of abnormalities was different between reference and exposed populations of white sucker.

Table 2.5-3
Number of Adult White Sucker with Specific External/Internal Abnormalities,
Lake Habitat, Moira River System, Fall 1999

^(a) Fish lice, parasitic crustacean (Copepoda).

^(b) Unknown parasites of intestinal tract, liver and heart; white cysts - possibly a trematode.

^(c) An individual fish may exhibit more than one type of abnormality.

Variations in age, body size, size-at-age and organ metrics for white sucker from each study lake are presented in Figures 2.5-2 to 2.5-5. Results of overall ANOVA and ANCOVA statistical comparisons among lake populations of white sucker are presented in Table 2.5-4.

Figure 2.5-2
Mean Age, Fork Length and Body Weight of Female and Male White Sucker, Moira River System, Fall 1999 ^(a),

Female

Male

^(a) Values represent mean + S.E. Open bars are reference populations, solid bars are exposed populations.

Figure 2.5-3
Condition and Size-at-Age of Female and Male White Sucker, Moira River System, Fall 1999^(a)

Female

Male

^(a) Values represent mean + S.E. Open bars are reference populations, solid bars are exposed populations.

Figure 2.5-4
Relative Liver and Gonad Weights of Female and Male White Sucker, Moira River System, Fall 1999 ^(a)

Female

Male

^(a) Values represent mean + S.E. Open bars are reference populations, solid bars are exposed populations.

Figure 2.5-5
Fecundity and Mean Egg Size of Mature Female White Sucker, Moira River System, Fall 1999 ^(a)

^(a) Values represent mean + S.E. Open bars are reference populations, solid bars are exposed populations.

Results of Tukey's pairwise testscomparing the performance of white sucker populations from Consecon, Round and Moira lakes are presented in Table 2.5-5. The percent difference was included in Table 2.5-5 to provide a better understanding of the magnitude of difference in individual fish characteristics.

Table 2.5-4

Results of Overall ANOVA and ANCOVA Tests Comparing Whole-Organism
Characteristics of White Sucker Among Reference and Exposed Lakes, Moira River
System, Fall 1999

^(a) Probability value for testing an overall difference among means.
^(b) Probability value for testing an overall difference among slopes and intercepts means.

Table 2.5-5

Summary of Statistical Analyses Comparing the Performance of White Sucker Populations
from Lake Habitats, Moira River System, Fall 1999

^(a) P-values are from Tukey's pairwisecomparisons (following a significant ANOVA/ANCOVA F-statistic).
^(b) % Difference is for exposed relative to the reference site (or Consecon Lake relative to Round Lake) calculated using ANOVA/ANCOVA adjusted least squared means.
^(c) P-value from ANOVA/ANCOVA model testing for differences in intercepts among all sites.

Comparison of Reference Populations (Consecon Lake vs Round Lake)

Consecon Lake and Round Lake were chosen as reference lakes based on similarities in water chemistry, trophic level, cottage development, and general fish community composition (see Section 1.4). However, there were significant differences in whole-organism characteristics of white sucker populations from these lakes.

There was no difference in mean age of male and female sucker between lakes, although statistical power for male sucker was poor (Table 2.5-6). Sucker from Consecon Lake were longer, heavier, and had increased growth (as estimated by size-at-age) relative to sucker from Round Lake (Figures 2.5-2, 2.5-3). Condition of sucker from Consecon Lake was also higher, indicating they were heavier at any given length (i.e., fatter) (Figure 2.5-3). Liver weight was lower in female sucker from Consecon, although no difference was observed in male sucker (Figure 2.5-4), despite sufficient power to detect a difference of at least 30% (Table 2.5-6).

Table 2.5-6

Statistical Power to Detect a Difference of 20, 30 and 50% (ES, effect size) for Those
Comparisons of White Sucker That Were Not Statistically Different

* Power based on ANOVA/ANCOVA model including all three sites.

Testis weight was greater in sucker from Consecon Lake, although there was no difference in ovary weight or egg size (i.e., diameter) (Figures 2.5-4, 2.5-5). Interestingly, fecundity was approximately 30% higher at Consecon Lake (Figure 2.5-5) suggesting that individual eggs from sucker of Round Lake were heavier or denser.

In general, white sucker from Consecon Lake showed evidence of increased energy expenditure and storage relative to sucker from Round Lake. This is commonly a response to increased food availability and can occur for several reasons, such as relaxed competition or increased food productivity (Gibbons and Munkittrick 1994). The reason for the observed difference in fish performance is uncertain, although benthic community structure does not appear to be markedly different between lakes (see Section 2.3). Oxygen depletion occurs in the deeper areas of Consecon Lake, but does not occur in Round Lake. Oxygen depletion would be expected to decrease available habitat (and potentially release more metals from sediments), creating more unfavourable conditions. However, the white sucker from Consecon Lake showed no evidence of decreased growth or reproductive capacity relative to Round Lake; in fact, the opposite was true. The results are important in that they emphasize the degree of reference variability in the performance of white sucker and provide a baseline from which to evaluate the health status of white sucker from Moira Lake.

Exposed (Moira Lake) vs. Reference Populations

The response of the white sucker population from Moira Lake was compared to populations from Consecon Lake and Round Lake. Based on differences observed between the reference lakes, it was not surprising that the relative response of exposed fish from Moira Lake also differed.

Female sucker from Moira Lake were older relative to sucker from both reference lakes (Figure 2.5-2). There was no difference in male age, although there was insufficient power to detect even a 50% difference in age (Table 2.5-6).

Both male and female sucker from Moira Lake were generally longer, heavier and had increased growth (i.e., size-at-age) relative to sucker from Round Lake; whereas, they were shorter, lighter and had decreased growth relative to sucker from Consecon Lake (Figures 2.5-2, 2.5-3). There was no difference in condition between exposed and reference sucker (Figure 2.5-3).

Liver size of females from Moira Lake was greater than observed at Consecon Lake, but similar to females from Round Lake (Figure 2.5-4). There was no difference in male liver size.

Ovary weight was greater at Moira Lake relative to either reference population; however, testis weight was intermediate between Round and Consecon lakes (Figure 2.5-4). Fecundity was higher in sucker from Moira Lake relative to Round Lake, but similar to estimates obtained from Consecon Lake (Figure 2.5-5). There were no differences in mean egg diameter among sites (Figure 2.5-5).

The evaluation of the health status of white sucker from Moira Lake depends on what reference population is used for comparison (Table 2.5-7). If the population from Round Lake was selected then one would conclude that sucker from Moira Lake were older and showed increased energy expenditure (e.g., increased size, growth, reproductive effort), but no change in energy storage (condition, liver size). Conversely, relative to Consecon Lake, sucker from Moira Lake were older, had decreased energy expenditure (size, growth), and no change (condition) or an increase in energy storage (female liver size). In many instances, whole-organism characteristics of sucker from Moira Lake were intermediate between Consecon and Round lakes.

Table 2.5-7

Summary of the Whole-Organism Response of White Sucker from Moira Lake,
Moira River System, Fall 1999

+ signifies a increase relative to reference data, - signifies a decrease, O signifies no change.

na - not applicable.

^(a)comparisons of male age did not have sufficient power to detect differences <50%.

If one accepts that white sucker populations from both Consecon Lake and Round Lake represent natural reference variability in fish performance, then the only characteristic of exposed white sucker that is consistently different (same direction) from both reference conditions is an older age structure in females. Whether this difference is related to metals is unknown, but an increase in mean adult age often results from increased juvenile and/or early life-stage mortality or decreased recruitment (Munkittrick and Dixon 1989). Conversely, mean age at the reference lakes may be lower due to increased adult mortality (e.g., recreational angling). A more thorough evaluation of age structure (not just limited to adult age classes) would provide better resolution to assess possible differences in age structure.

Whole body arsenic concentrations in white sucker were significantly higher in Bend Bay, Moira Lake and Stoco Lake (see Section 2.4). Other elevated metal concentrations in white sucker were silver in Bend Bay; cobalt in Bend Bay and Moira Lake; copper in Bend Bay; and, uranium in Stoco Lake. These data confirmed that white sucker populations in Bend Bay, Moira Lake and Stoco Lake were, indeed, exposed to metals of concern.

It is possible that fish that are chronically exposed to elevated metal concentrations in the Moira River system have become tolerant. Laboratory studies have shown that rainbow trout (Oncorhynchus mykiss) gradually increased their tolerance to arsenic during exposure to 0.22 of the lethal concentration (Dixon and Sprague 1981). Rainbow trout exposed to 10-20 mg/kg As in their diet (an order of magnitude above observed fish tissue concentrations in the Moira River system) did not differ in growth from controls (Oladimeji et al. 1984). Mortality of rainbow trout exposed for 28 days to 100 and 1000 µg/L As did not increase (Spehar et al. 1980).

Field studies specifically documenting the effects of elevated arsenic and metal concentrations on wild white sucker populations are rare. Three studies are summarized and compared to this study in Table 2.5-8. Metals of concern and the concentrations of those metals differ among the studies. Copper is the only metal in common among all studies and concentrations of copper are similar. The Moira River system has the greatest number of metals of concern and is the only study with elevated arsenic. Hardness and pH (when reported) are relatively similar. Lake morphometry differs significantly among the studies and important variables such as relative abundance of the food supply also vary (e.g., there was no evidence of limitation in food supply in Hamell Lake, but food supply was definitely limited in Beaverlodge Lake and food quality differed among Manitouwadge lakes). Observed effects in white sucker populations are different in each study, illustrating how difficult it is to relate specific responses in natural populations to specific toxicants.

Table 2.5-8

Comparison of Effects Observed in White Sucker Populations at Sites with Elevated Metal Concentrations

* Mean annual concentrations in Moira Lake.

Stoco Lake and Bend Bay (other exposed populations)

Although Stoco Lake and Bend Bay were not included in statistical analyses, data from these sites were included in Figures 2.5-2 to 2.5-5 to evaluate how they compared with Moira Lake and both reference lakes.

In general, white sucker from Bend Bay mirrored the response of sucker from Moira Lake. The response of male sucker from Stoco Lake was more variable (e.g., age, length, weight), although the overall response was not markedly different from other exposed lakes. As with Moira Lake, most parameters for sucker from Stoco Lake and Bend Bay were within the range of reference variation defined by Consecon Lake and Round Lake.

2.5.2.2 Longnose Dace

The abundance of female, male and immature longnose dace captured at the exposure site (F-3) and reference sites (F-1, F-2) is presented in Table 2.5-9. With the exception of male dace from the upper reference site (n=15), capture success of adult dace was good at each site. Relative abundance of longnose dace, as estimated by catch-per-unit-effort (CPUE), was not substantially different among reference sites (1.6 vs 1.8 adult dace/100 seconds electrofished). Unfortunately, CPUE could not be estimated for site F-3 because weather conditions caused the timer display on the electrofisher to malfunction. A summary of incidental fish species captured at each site during the collection of longnose dace is provided in Appendix V.

Table 2.5-9

 

Total Number of Female, Male and Immature Longnose Dace from River Habitat,
Moira River System, Fall 1999

F-I = reference site upstream of Malone; F-2 = reference site in Malone; F-3 = exposure site downstream of the Deloro Mine Site (old Marmora Road).

Fish Measurements

Results of internal/external pathology examinations of longnose dace indicated that abnormalities were found in 65-100% of dace per site (Table 2.5-10). The presence of external black spots and small, white cysts within the body cavity were the most prevalent abnormality. Both abnormalities are probably parasitic in origin. Raised black spots in the skin are commonly a sign of infestation by a trematode (Digenea) that uses fish as an intermediate host as part of its life cycle (i.e., black spot disease) (Post 1987). The parasite associated with the internal white cysts was not identified; however, small larvae were found inside the cysts that may be indicative of trematode infestation. In general, there was no evidence suggesting that the prevalence of abnormalities/parasitism was different between reference and exposed populations of longnose dace.

Table 2.5-10

Number of Longnose Dace with Specific External/Internal Abnormalities,
River Habitat, Moira River System, Fall 1999

^(a) Digenean trematode.

^(b) probable trematode infestation.

^(a) An individual fish may exhibit more than one type of abnormality.

F-I = reference site upstream of Malone; F-2 = reference site in Malone; F-3 = exposure site
downstream of the Deloro Mine Site (old Marmora Road).

Variations in age, body size, size-at-age and organ metrics for longnose dace from each study site are presented in Figures 2.5-6 to 2.5-9. Results of overall ANOVA and ANCOVA statistical comparisons among populations of longnose dace are presented in Table 2.5-11. Results of specific Tukey's pairwise tests comparing the performance of dace populations between sites F-1, F-2 and F-3 are presented in Table 2.5-12. The percent difference was included in Table 2.5-12 to highlight the magnitude of difference in individual fish characteristics.

Figure 2.5-6
Mean Age, Fork Length and Body Weight of Female and Male Longnose Dace,
Moira River System, Fall 1999 ^(a)

Female

Male

^(a) Values represent mean ± S.E. Open bars are reference populations, solid bars are exposed populations.

Figure 2.5-7
Condition and Size-at-Age of Female and Male Longnose Dace, Moira River System, Fall 1999 ^(a)

Female

Male

^(a) Values represent mean ± S.E. Open bars are reference populations, solid bars are exposed populations.

Figure 2.5-8
Relative Liver and Gonad Weights of Female and Male Longnose Dace, Moira River System, Fall 1999 ^(a)

Female

Male

^(a) Values represent mean ± S.E. Open bars are reference populations, solid bars are exposed populations.

Figure 2.5-9
Fecundity and Mean Egg Size of Mature Female Longnose Dace, Moira River System, Fall 1999^(a)

^(a) Values represent mean ± S.E. Open bars are reference populations, solid bars are exposed populations.

Table 2.5-11
Results of Overall ANOVA and ANCOVA Tests Comparing Whole-Organism Characteristics
of Longnose Dace Among Reference and Exposed River Sites, Moira River System, Fall 1999

^(a) probability value for testing an overall difference among means.
^(b) probability value for testing an overall difference among slopes and intercepts means.

Table 2.5-12

Summary of Statistical Analyses Comparing the Performance of Longnose Dace
Populations from River Habitats, Moira River System, Fall 1999

^(a) P-values are from Tukey's pairwise comparisons (following a significant ANOVA/ANCOVA F-statistic).

^(b) % Difference is for exposed relative to the reference site (or F-2 relative F-1) calculated using
ANOVA/ANCOVA adjusted least squared means.

^(c) P-value from ANOVA/ANCOVA model testing for differences in intercepts among all sites.

Comparison of Reference Populations (F-1 vs. F-2)

Reference sites located upstream and within the village of Malone were selected based on similarities in habitat characteristics (Photos V.6, V.7, Appendix V). Few differences in whole-organism characteristics of longnose dace were observed between these sites; this result was expected given the similarity in habitat.

Mean age of male and female dace was not significantly different between reference sites (Figure 2.5-6), although statistical power for both sexes was limited (Table 2.5-13) and unable to detect a difference of a factor of <1.38. Female dace from site F-2 were longer, heavier and had increased growth relative to females from F-1 (Figure 2.5-6, 2.5-7). Although overall tests of male length and weight were borderline (all sites included, Table 2.5-11), Tukey's pairwise comparisons between reference sites did not indicate significant differences (length, P=0.93; weight, P=0.53). As with age, the power of the comparison of male weight was unable to detect a difference of a factor of <1.38 (Table 2.5-13). Condition was statistically different between sites; however, the magnitude of difference was only 6-7% (Figure 2.5-7). No differences were observed in gonad weight, fecundity, egg size, and liver size (Figure 2.5-8, 2.5-9).

Table 2.5-13

Statistical Power to Detect a Difference of 20, 30 and 50% (ES, effect size) for Those
Comparisons of Longnose Dace That Were Not Statistically Different

* power based on ANOVA/ANCOVA model including all three sites.

In general, female dace from site F-2 showed some evidence of increased energy expenditure, although there was no concomitant increase in gonad size, fecundity or estimates of energy storage (e.g., condition, liver size). As well, no whole-organism differences were observed in male dace between sites, although statistical power was relatively poor.

Exposed (F-3) vs. Reference Populations

The response of exposed longnose dace (site F-3) was defined relative to populations from upstream reference sites F-1 and F-2.

With the exception of a 10% increase in egg diameter at the exposed site (Figure 2.5-9), there were no significant differences in whole-organism characteristics between adult dace collected downstream of the mine site relative to dace from reference site F-2 (Figures 2.5-6 to 2.5-9). Not surprisingly, when dace from the exposure site were compared to dace from site F-1, the response mirrored the response of dace from site F-2 relative to site F-1 (i.e., reference site comparison) (Figures 2.5-6 to 2.5-9). Consequently, there is little evidence that whole-organism characteristics of longnose dace collected downstream of the Deloro Mine Site are different from dace collected from reference sites not influenced by the Deloro Mine Site.

Longnose dace are commonly found in riffles and higher velocity areas consisting of coarse substrates and boulders. As such, they are not in direct contact with fine sediments that accumulate higher concentrations of metals downstream of the mine (Section 2.2.). However, the observed increase in tissue burden of metals (particularly arsenic, copper, nickel, mercury and silver) downstream of the Deloro Mine Site (Section 2.4) provides assurance that longnose dace are exposed to increased metal concentrations.

Despite the exposure (and uptake) of these heavy metals, whole-organism characteristics of longnose dace did not reflect the increase in metal concentrations. It is possible that concentrations were not high enough to cause individual or population level changes in resident fishes. Alternatively, longnose dace may not be sensitive to metals. Longnose dace has been successfully used to monitor the effects of bleached kraft mill effluent and municipal sewage wastewater (Golder 2000), but it is unknown whether they have been previously used to investigate the effects of metals. Finally, it is also possible that several stressors are impacting the same site in opposite directions, or the carrying capacity of the system was reduced and the population has had sufficient time to equilibrate with the new system (Gibbons and Munkittrick 1994). Examining the relative population sizes between reference and exposed sites can identify reduced carrying capacity. Unfortunately, CPUE data for site F-3 were not available. Identifying whether there are conflicting stressors is more difficult, experimentally and philosophically: if there is no detectable change in growth, survival or reproduction, has there been an impact?

The weight of evidence from the sentinel fish species data suggests that there is no readily apparent cause-effect relationship between metal concentrations and whole-organism or population responses. Additional study would help confirm whether there is, indeed, a difference in white sucker age structure; however, obtaining additional data on white sucker population characteristics is unlikely to change recommendations for clean up at the Deloro Mine Site. In other words, the sentinel fish data do not appear to justify a major incremental clean up effort at the Deloro Mine Site over and above what would be planned based upon meeting PWQOs.

References

Cohen, J. 1988. Statistical power analysis for the behavioural sciences. 2 ^nd Ed. L. Erlbaum Assoc., Hillsdale, NJ. 567 p.

Dixon, D.G. and J.B. Sprague. 1981. Acclimation-induced changes in toxicity of arsenic and cyanide to rainbow trout, Salmo gairdneri Richardson. J. Fish. Biol. 18: 579-589.

Environment Canada. 1997. Fish survey expert working group final report. EEM Technical Management Committee. EEM/1997/6.

Faul, F. and E. Erdfelder. 1992. GPOWER. A priori, post-hoc and compromise power analyses for MS-DOS [Computer program]. Bonn, Germany. Bonn University, Dept. of Psychology.

Gagnon, M.M., J.J. Dodson, P.V. Hodson, G. Van Der Kraak and J.H. Carey. 1994. Seasonal effects of bleached kraft mill effluent on reproductive parameters of white sucker (Catostomus commersoni) populations of the St. Maurice River, Quebec, Canada. Can. J. Fish. Aquat. Sci. 51: 337-347.

Gibbons, W.N. and K.R. Munkittrick. 1994. A sentinel monitoring framework for identifying fish population responses to industrial discharges. J. Aquat. Ecosyst. Health 3:227-237.

Gibbons, W.N., K.R. Munkittrick and W.D. Taylor. 1998. Monitoring aquatic environments receiving industrial effluents using small fish species 2: Comparison between trout-perch (Percopsis omiscomaycus) and white sucker (Catostomus commersoni) downstream of a pulp mill. Environ. Toxicol. Chem. 17:2238-2245.

Golder Associates Ltd. 2000. Final Report: Weyerhaeuser Canada Ltd., Grande Prairie Environmental Effects Monitoring Cycle 2. Prepared for Weyerhaeuser Canada Ltd., Grande Prairie, AB.

Grubbs, F.E. 1971. Procedures for detecting outlying observations in samples. Technometrics 11:1-22.

Hodson, P.V., M. Mc Whirter, K. Ralph, B. Gray, D. Thivierge, J.H. Carey, G. Van Der Kraak, D.M. Whittle and M. Levesque. 1992. Effects of bleached kraft mill effluent on fish in the St. Maurice River, Quebec. Environ. Toxicol. 11:1635-1652.

Kilgour, B.W., C.B. Portt and D.G. Dixon. 1999. Moira River impact study: detailed study design to determine the impact of the former Deloro Mine Site on the Moira River system. Prepared for the Ontario Ministry of Environment, Toronto, ON by Water Systems Analysts (WSA), Guelph, ON, C. Porrt and Associates, Guelph, ON and University of Waterloo, Waterloo, ON. 41 pp.

MacKay, W.C., G.R. Ash and H.J. Norris. (eds.). 1990. Fish ageing methods for Alberta. R.L. & L. Environmental Services Ltd. in assoc. with Alberta Fish and Wildl. Div. and Univ. of Alberta, Edmonton. 113 p.

McFarlane, G.A. and W.G. Franzin. 1978. Elevated heavy metals: a stress on a population of white suckers, Catostomus commersoni, in Hamell Lake, Saskatchewan. J. Fish. Res. Board. Can. 35:963-970.

McMaster, M.E., G.J. Van Der Kraak, C.B. Portt, K.R. Munkittrick, P.K. Sibley, I.R. Smith and D.G. Dixon. 1991. Changes in hepatic mixed-function oxygenase (MFO) activity, plasma steroid levels and age at maturity of white sucker (Catostomus commersoni) population exposed to bleached kraft pulp mill effluent. Aquat. Toxicol. 21:199-218.

Miller, P.A., K.R. Munkittrick and D.G. Dixon. 1992. Relationship between concentrations of copper and zinc in water, sediment, benthic invertebrates and tissues of white sucker (Catostomus commersoni) at metal-contaminated lakes. Can. J. Fish. Aquat. Sci. 49: 978-984.

Munkittrick, K.R. 1992. A review and evaluation of study design considerations for site-specifically assessing the health of fish populations. J. Aquat. Ecosyst. Health 1:283-293.

Munkittrick, K.R. and D.G. Dixon. 1988. Growth, fecundity and energy stores of white sucker (Catostomus commersoni) from lakes containing elevated levels of copper and zinc. Can. J. Fish. Aquat. Sci. 45:1355-1365

Munkittrick and Dixon. 1989. Use of white sucker (Catostomus commersoni) populations to assess the health of aquatic ecosystems exposed to low-level contaminated stress. Can. J. Fish. Aquat. Sci. 46: 1455-1462.

Oladimeji. A.A., U. Qadri and A.S.W. de Freitas. 1984. Long-term effects of arsenic accumulation in rainbow trout. Bull. Environ. Contam. Toxicol. 32:732-741.

Paine, M.D. 1998. Volume II: CANMET/INTEMIN manual on statistical analysis of environmental data. Prepared for Natural Resources Canada, Canada Centre for Mineral and Energy Technology (CANMET), Ottawa, ON., by Paine, Ledge and Associates (PLA), North Vancouver, B.C. 228 pp.

Peterman, R.M. 1990. Statistical power analysis can improve fisheries research and management. Can. J. Fish. Aquat. Sci. 47: 2-15.

Post, G. 1987. Textbook of Fish Health, 2nd Edition. T.F.H. Publications Inc. Neptune City, NJ, USA. 288 pp.

Spehar, R.L., J.T. Fiandt, R.L. Anderson and D.L. DeFoe. 1980. Comparative toxicity of arsenic compounds and their accumulation in invertebrates and fish. Arch. Environm. Contam. Toxicol. 9:53-63.

Swanson. S.M. 1982. Levels and effects of radionuclides in aquatic fauna of the Beaverlodge area, Saskatchewan. Saskatchewan Research Council Report No. C-806-5-E-82. Prepared for Environment Canada, Environmental Protection Service, Regina, Saskatchewan.

Trippel, E.A. and H.H. Harvey. 1987. Reproductive responses of five white sucker (Catostomus commersoni) populations in relation to lake acidity. Can. J. Fish. Aquat. Sci. 44:10181023.

Wilkinson, L. 1990. SYSTAT: the system for statistics. SYSTAT Inc., Evanston, IL, USA.