DERR-2024-007808

1

Running Head: Site-Specific Water Quality Standard for Selenium

Author to whom copyright and page proofs should be sent:

Kevin V. Brix
EcoTox
2001 NW Nye Street
Newport, Oregon 97365
Tel: [phone]
Fax [phone]
E-mail: [e-mail]
Word count: 5,335

2
DERIVATION OF A SITE-SPECIFIC WATER QUALITY STANDARD FOR 1
SELENIUM IN THE GREAT SALT LAKE, UTAH 2
3
4
5
Kevin V. Brix,† David K. DeForest, ‡ Rick D. Cardwell, § William J. Adams, ║ 6
7
8
9
10
11
12
†EcoTox, 2001 NW Nye Street, Newport, Oregon 97365 13
‡Parametrix, Inc., 5808 Lake Washington Blvd. NE, Suite 200, Kirkland, Washington 14
98033 15
§Parametrix, Inc., 1600 SW Western Blvd., Suite 165, Corvallis, Oregon, 97333 16
║Kennecott Utah Copper, 8315 West 3595 South, P.O. Box 6001, Magna, Utah 84044 17
18
19
20
21
3
Corresponding Author: [e-mail] 22
23
24
4
ABSTRACT 25
The purpose of this study was to develop a site-specific water quality standard for 26
selenium in the Great Salt Lake, Utah. The study examined the direct bioavailabilty and 27
toxicity of selenium, as selenate, to biota resident to the Great Salt Lake and the potential 28
for dietary selenium exposure to aquatic dependent birds that might consume resident 29
biota. Because of its high salinity, the lake has limited biological diversity with bacteria, 30
algae, diatoms, brine shrimp and brine flies being the only organisms present in the main 31
(hypersaline) portions of the Lake. To evaluate their sensitivity to selenium, a series of 32
acute and chronic toxicity studies were conducted on brine shrimp, Artemia franiciscana, 33
brine fly, Ephydra cinerea, and a hypersaline alga, Dunaliella viridis. The resulting 34
acute and chronic toxicity values indicated that resident species are more selenium 35
tolerant than many freshwater species. This is thought to result in part to the lake's high 36
ambient sulfate concentrations (>5,800 mg/L), as sulfate is known to reduce selenate 37
bioavailability. The acute and chronic test results were compared to selenium 38
concentrations expected to occur in a mining effluent discharge located at the south end 39
of the lake. Based on these comparisons, no appreciable risks to resident aquatic biota 40
were projected. Field and laboratory data collected on selenium bioaccumulation in brine 41
shrimp demonstrated a linear relationship between water and tissue selenium 42
concentrations. Applying a dietary selenium threshold of 5 mg/kg dw for aquatic birds to 43
this relationship resulted in an estimate of 27 µg/L Se in water as a safe concentration for 44
this exposure pathway and an appropriate site-specific water quality standard. 45
Key Words: Selenium Site Specific Water Quality Standard Great Salt Lake 46
5
INTRODUCTION 47
48
The Great Salt Lake (GSL) is the fourth largest terminal lake in the world [1] and 49
the largest hypersaline lake in North America [2]. In 1957, the Southern Pacific Railroad 50
Company constructed a rock-filled causeway across the lake, dividing it into two arms. 51
Although culverts link the two arms, they are insufficient to maintain mixing between 52
them. Consequently, the GSL essentially consists of two lakes, each with varying salinity 53
and dominant organisms. Approximately 92 percent of freshwater inputs enter the 54
southern arm [3], resulting in the northern arm being more saline (approximate salinity 55
330 g/L) than the southern arm (approximate salinity 100 g/L). 56
The food web of the southern arm of the GSL is relatively simple because few 57
organisms can tolerate its high salinity and low oxygen solubility [4, 5]. The aquatic food 58
web consists of at least four species of bacteria (mainly Halobacterium and Halococcus), up 59
to 20 species of algae (mainly Dunaliella viridis and D. salina), at least 17 diatom species, 60
brine shrimp (Artemia franciscana), and seven species of brine flies (Ephydra spp.). 61
Additionally, in areas near significant freshwater inputs where the salinity is less than <75 62
g/L, corixids (Trichocorixa verticalis), rotifers (Brachionus sp.) and two species of 63
copepods (Cletocampus albuquerquensis and Diaptomus connexus) have been observed [2, 64
5-8]. The abundance of these taxa fluctuates with season and salinity [1]. 65
Because of the high salinity, no fish occur in the lake except in freshwater 66
estuaries near the Bear, Jordan and Weber Rivers. This lack of aquatic predators, in turn, 67
can lead to extraordinarily high densities of brine shrimp and brine flies, which are an 68
important food source for resident and migratory birds. The lake and its surrounding 69
6
wetlands is an important stop-over point for migratory shorebirds and waterfowl. Greater 70
than 75% of the West's population of tundra swans (Cygnus columbianus), 50% of the 71
continent’s Wilson’s phalaropes (Phalaropus tricolor), 25% of the continent's northern 72
pintails (Anas acuta), the world’s largest nesting population of California gulls (Larus 73
californicus), and millions of other waterfowl use the lake during their annual migration 74
periods. 75
The eastern and southeastern shorelines of the lake's southern arm are bordered by 76
the Salt Lake City metropolitan area. Among the industries bordering the lake are the 77
smelting and refining facilities for a copper mine. The major constituent of this facility's 78
wastewater discharge is selenium, with concentrations as high as 300 µg/L Se having 79
been measured historically. Current selenium discharge levels are approximately 20-50 80
µg/L before dilution. The majority (>95%) of this Se is in the form of selenate and 81
unless otherwise noted, all discussion of Se in this paper is referring to selenate. The 82
effluent is considerably less saline (5 g/L) than the lake creating creating a small 83
estuarine zone in the immediate area of the discharge. 84
The outfall discharge has cut a channel 2-4 feet deep in the lake sediments 85
immediately offshore. Water depth surrounding the channeled area averages 86
approximately 8-18 inches. Lake sediments consist of well-compacted silty, sandy clays. 87
Sediments in the channeled area are less compacted and composed of finer material. 88
Dense stands of Phragmites sp. have established along the banks of the channel, 89
stabilizing it. Over time, deposition of fine sediments and organic material and continued 90
colonization by Phragmites has effectively extended the channel approximately 1,500 91
feet out into the lake (Figure 1). The water depth and velocity, along with the dense 92
7
Phragmites, effectively limits shorebird use in the channel proper, but they are routinely 93
observed to feed along the shorelines on either side of the channel. 94
Because of its unique water quality characteristics and biota, generic water quality 95
criteria do not apply to the GSL [9], and historically very little toxicity data has been 96
generated for the lake's resident species. Hence, the appropriate water quality standard 97
for Se in the GSL is unclear. Additionally, unlike most other metals and metalloids, the 98
diet typically represents the most important exposure pathway for Se, with top trophic 99
level consumers (e.g., fish and aquatic-dependent birds) being the most sensitive 100
environmental receptors in an aquatic system [10, 11]. Consequently, any site-specific 101
water quality standard for selenium must consider exposure via both water and dietary 102
pathways. This study was designed to evaluate potential exposure and effects from Se 103
discharges to the lake via both pathways through a series of laboratory and field studies 104
on resident species. Study results were then used to develop an appropriate site-specific 105
water quality discharge limit. 106
107
METHODS AND MATERIALS 108
109
Given that Se may cause either direct effects on aquatic biota resident to the lake 110
or the resident biota may accumulate Se to deleterious levels for organisms that consume 111
them, both pathways needed evaluation in order to propose an appropriate site-specific 112
water quality standard (Figure 2). 113
To evaluate the potential for direct effects on resident aquatic biota, we conducted 114
toxicity tests on brine shrimp, larvae of the brine fly, Ephydra cinerea, and the most 115
8
common alga in the southern arm of the lake, Dunaliella viridis. This species is a 116
principal food source for brine shrimp. Acute testing was conducted using brine shimp 117
and brine fly larvae and chronic testing was conducted using Dunaliella viridis and brine 118
shrimp. The chronic sensitivity of brine flies was not investigated because of their 119
extreme insensitivity when tested acutely. 120
To evaluate the potential for avian toxicity arising from the dietary pathway, Se 121
concentrations in brine shrimp were measured in specimens collected within and adjacent 122
to the mine discharge, as well as at background Se concentrations in the lake. These data 123
were then compared to appropriate dietary thresholds for aquatic dependent birds. 124
125
Toxicity Testing 126
127
Acute Testing 128
129
The methods used for conducting the acute tests were consistent with those 130
described in U.S. EPA [12], although parameters such as dilution water and test volume 131
were modified to meet species-specific requirements. Tests were static non-renewal 132
studies conducted at 25 ± 1 °C with five test concentrations and a control. Dilution water 133
for the acute tests was GSL water collected from the shoreline on the north side of 134
Antelope Island, a location well removed from anthropogenic inputs to the lake. 135
Conventional water quality parameters were measured in the dilution water prior to 136
testing (Table 1). 137
9
Reagent grade sodium selenate (CAS #13410-01-0) obtained from Sigma 138
Chemical Company, St. Louis, Missouri was used to create stock solutions. For the brine 139
shrimp study, a 10 g/L stock solution was prepared by adding 23.9 g of sodium selenate 140
to 1 L of deionized water. The extreme insensitivity of brine fly larvae prevented 141
preparation of a single stock solution. Instead, the sodium selenate was added directly to 142
5 L batches of dilution water in order to achieve the desired nominal test concentrations. 143
The brine shrimp test was initiated with nauplii <24 hours old that were hatched 144
overnight at 25 °C in 25 g/L artificial seawater. Nauplii were not acclimated to the 145
dilution water salinity (82 g/L) prior to testing. This treatment reflects natural conditions 146
where cysts hatch in the relatively low salinity lens of water on the lake surface and then 147
drop down in the more saline water column. Nauplii were randomly introduced to 148
exposure chambers (600 mL beakers with 400 mL of test solution) for each of the five 149
treatments and control. Four replicates were conducted with each treatment. Preliminary 150
testing indicated brine shrimp required daily feeding to achieve acceptable control 151
survival and so were fed daily 2 mL of a 500,000 cells/mL stock of the marine algae 152
Platymonas sp. 153
Brine fly larvae were collected for testing from White Rock Bay on the north 154
shore of Antelope Island. Larvae were identified to species by Chadwick and Associates 155
in Littleton, Colorado. Test organisms were held in GSL water in 40 L aquaria at 12 °C 156
for eight weeks prior to testing. During holding 40 mL of 3.5 x 106 cells/mL solution of 157
Dunaliella viridis were added weekly to the aquaria. Forty-eight hours prior to testing, 158
larvae were acclimated to the test temperature of 25 °C. Four replicate 1 L beakers with 159
10
800 mL of test solution were tested at each Se concentration. Test organisms were not 160
fed during testing. 161
162
Chronic Testing 163
164
The methods for conducting the chronic brine shrimp life-cycle test were 165
previously described in Brix et al. [13]. Briefly, this 28-day test measured survival, 166
growth and reproduction of the parental generation, and survival and growth of the F1 167
generation. The test was conducted under intermittent flow-through conditions beginning 168
with brine shrimp nauplii <24 hours old. After 11 days, brine shrimp matured sexually 169
and began pairing for mating. At this time, they were thinned by collecting and weighing 170
(dry) a random subsample from each test concentration. Six adult pairs for each test 171
concentration were then monitored for reproduction until day 28 when surviving shrimp 172
were measured for dry weight. For each test concentration, randomly selected nauplii (F1 173
generation) from the pairs were subjected to the same conditions as the parental 174
generation for 11 days, with survival and dry weight being monitored for comparison 175
with the parental generation. 176
The experimental design for the algae toxicity test followed U.S. EPA [14] and 177
EU [15], except for the dilution media, which was GSL water passed through a 1 µm 178
filter. In this test, 1 x 104 cells of D. viridis from a culture in log-phase growth were 179
inoculated into 125 mL test flasks with 50 mL of test solution. Test flasks were placed 180
on a shaker table rotated at 100 cpm. Each test concentration consisted of 16 replicates 181
and every 24 hours, four of the replicates were terminated, whereupon water quality was 182
11
monitored and cell densities measured using a Hach 300 spectrophotometer. The 183
spectrophotometer was calibrated against known cell density stocks of D. viridis. 184
185
Analytical Chemistry 186
187
For all tests, water quality parameters (temperature, salinity, pH and dissolved 188
oxygen) were measured daily in one replicate of each treatment and control. Samples 189
from each concentration were collected for Se analysis at test initiation and termination 190
using the hydride generation method of Cutter [16]. The exception to this sampling 191
regime was the chronic brine shrimp study where samples were collected on a weekly 192
basis. 193
194
Data Analysis 195
196
For the acute brine fly and brine shrimp tests, statistical analyses were conducted 197
using the statistical computer package Toxis® [17] to estimate the LC50 and its 95% 198
confidence interval, as well as the no observed effect concentration (NOEC) and lowest 199
observed effect concentration (LOEC). The NOEC and LOEC were determined by 200
Steel’s many-one rank test and the LC50 was estimated by probit analysis. 201
The statistical evaluation of the chronic brine shrimp results included testing for 202
differences between the treatments and controls at reproductive pairing (Day 11), Day 21 203
and Day 28 by parametric or non-parametric methods depending on whether data met 204
normality and homogeneity assumptions. If the data met the assumptions of normality 205
12
and homogeneity, an ANOVA was computed to determine whether any differences 206
existed among levels (concentrations or generations). If either of the assumptions could 207
not be met, the non-parametric Kruskal-Wallis test was used to test for differences. The 208
statistics were calculated using Statgraphics [18]. 209
For the chronic algae tests, statistical analyses for the NOEC and LOEC were 210
conducted using SPSS [19] in accordance with procedures described in EU [15]. Specific 211
growth rate and cumulative area under the curve were calculated for each replicate, as were 212
summary statistics for each time period. The statistical computer package Toxis® [17] was 213
used to estimate the EC50 value and the 95% confidence interval based on results from 214
specific growth rate and cumulative area under the growth curve calculations. 215
216
Field Bioaccumulation Study 217
218
In order to evaluate the potential for Se in the mining effluent to bioaccumulate in 219
aquatic organisms that might be fed upon by migratory shorebirds, a field program was 220
implemented to sample water and co-located brine shrimp at various locations relative to 221
the mining effluent discharge (Figure 1). Two sampling events (June and August) were 222
undertaken to characterize Se concentrations in water and biota. However, because brine 223
shrimp were not found at most stations in the discharge channel during the August 224
sampling event, only results for the June sampling event and a single sampling station 225
(station 7) with brine shrimp present during the August sampling event are presented. 226
Surface samples were collected because preliminary sampling efforts indicated 227
the majority of brine shrimp occurred in the upper water column. Water depth along the 228
13
sampling transect varied from 0.5 to 1.5 meters. Water samples were collected using a 229
battery-powered peristaltic pump using methods consistent with U.S. EPA [20]. Samples 230
were collected within the channel midway between the banks wherever possible. 231
When present, brine shrimp were collected at the same time and place as water 232
samples to evaluate the relationship between water and tissue Se concentrations. Brine 233
shrimp were collected using a dip net with a 15 x 30 cm basket constructed of 500 µm 234
Nitex™ screen. The dip net was slowly trawled through the water column approximately 235
15 cm below the water surface until the net contained sufficient specimens (5 g wet 236
weight) for analysis. 237
Total recoverable and dissolved Se were measured in water samples at the 238
Kennecott Environmental Laboratory using the hydride generation method of Cutter [16] 239
with an analytical detection limit of 2 µg/L Se. Total selenium was determined on the 240
tissue digestate by hydride generation – atomic fluorescence spectrometry. A total 241
reduction/oxidation digestion, converting all forms to selenium (IV) was accomplished 242
by boiling the digested sample in 4M HCl with potassium persulfate. The analytical 243
detection limit in tissues was 0.5 mg/kg dw. 244
245
RESULTS 246
247
Toxicity Testing 248
249
Well defined concentration-response relationships were observed for all of the 250
studies. For the acute brine fly study, an LC50 of 495 mg/L Se was estimated. The brine 251
14
shrimp LC50 of 78 mg/L indicated it was substantially more sensitive than the brine fly 252
(Table 3). In the chronic D. viridis study, EC50s of 45 and 32 mg/L were observed for 253
the specific growth and area under the curve endpoints (Table 4). The NOEC was 11 254
mg/L for both endpoints. A number of different endpoints were evaluated in the chronic 255
brine shrimp study. Day 11 growth of the parental generation and Day 21 reproduction 256
were comparable and the most sensitive endpoints evaluated. For both, the NOEC was 3 257
mg/L Se and the LOEC 8 mg/L Se (Table 5). Overall, these two endpoints for the brine 258
shrimp were also the most sensitive of any endpoint and species evaluated. 259
Water quality parameters were within expected ranges for all studies (Table1). 260
Measured dissolved oxygen concentrations (1.8-6.0) mg/L require a brief discussion, as 261
the values are lower than what is typically considered acceptable in toxicity tests. The 262
low dissolved oxygen values measured during testing are a result of the hypersalinity of 263
the test solutions, which limits oxygen solubility. Dissolved oxygen saturation at these 264
salinities ranges from 3.6 to 5.0 mg/L depending on salinity and test temperature 265
(supersaturated values were measured in the study with D. viridis as will typically occur 266
in algal assays). Hence, the measured values in the tests were typically >60% saturation, 267
as is customary for toxicity tests. For comparison, the southern arm of the GSL has a 268
dissolved oxygen saturation of 2.0 mg/L, which is lower than noted above because the 269
lake is at an elevation of 4,200 feet [8], whereas the tests were performed in a laboratory 270
at sea level. Hence, the dissolved oxygen concentrations in these tests are characteristic 271
of what these organisms normally encounter in the environment. Selenium test 272
concentrations stayed relatively constant for all tests with coefficients of variation in test 273
concentrations ≤20% for all treatments in all studies. 274
15
275
Field Bioaccumulation Study 276
277
In the June sampling event, surface water Se concentrations generally decreased 278
with distance from the outfall. Near the mouth of the outfall (Station 1) concentrations 279
were as high as 120 µg/L Se, but declined relatively rapidly to background concentrations 280
(2 µg/L Se) at Station 5 and beyond. Total recoverable and dissolved Se were essentially 281
equivalent at all stations. This is expected for Se discharges in the form of selenate, as it 282
does not readily adsorb to suspended solids [21, 22]. 283
Consistent with water concentrations, Se in brine shrimp from the June sampling 284
event was highest near the outfall mouth, with concentrations as high as 15 mg/kg dw 285
(Table 6). Also consistent with waterborne Se data, brine shrimp tissue concentrations 286
dropped relatively rapidly to background (2-3 mg/kg dw) beginning at Station 4. The 287
single station (station 7) sampled in August also resulted in background Se concentrations 288
in brine shrimp tissue. 289
290
DISCUSSION 291
292
Toxicity Studies 293
294
The current U.S. EPA acute and chronic water quality criteria for Se in freshwater 295
systems are 20 and 5 µg/L [23]. However, U.S. EPA has recently proposed a revised 296
criterion in which the acute criterion for selenate is 185 µg/L and the chronic criterion is 297
16
based on a tissue residue concentration in fish [24]. In comparison, the lowest acute and 298
chronic toxicity values measured for biota resident to the GSL were one to two orders of 299
magnitude higher than the proposed acute criterion. However, as discussed below, a 300
close examination of the data indicates resident biota are actually average in sensitivity 301
relative to other freshwater species that have been tested. We make these comparisons 302
not as an argument that the freshwater water quality criteria is appropriate for the GSL, 303
but simply to understand why biota resident to the GSL may appear to be relatively 304
insensitive to Se. 305
The primary factor causing GSL biota to appear relatively insensitive is the effect 306
of ambient sulfate concentrations on selenate bioavailability. It is well recognized that 307
sulfate reduces selenate bioavailability to a variety of organisms, including algae, 308
bacteria, midges, daphnids and brine shrimp [25-30]. Brix et al. [31] quantified this 309
relationship by summarizing available data and conducting additional studies with 310
amphipods, daphnids and fish. They then developed a log-linear relationship similar to 311
that derived for hardness and divalent cationic metals to normalize for selenate 312
bioavailability as a function of ambient sulfate concentrations. This relationship is 313
important when evaluating the toxicity data in this study because the ambient sulfate 314
concentration in the GSL is 5,800 mg/L, high enough to significantly reduce selenate 315
bioavailability. 316
When the high ambient sulfate concentration of the GSL is considered, the 317
relative acute sensitivity of brine shrimp and brine flies is comparable to many freshwater 318
species. When the acute data from this study are plotted with available acute data and all 319
data normalized for ambient sulfate concentrations, brine shrimp and brine flies rank at 320
17
the 29th and 63rd percentiles of the species sensitivity distribution (Figure 3). The acute 321
brine shrimp data derived in this study are largely consistent with a previous study by 322
Forsythe et al. [25], who estimated 96hour LC50s of 1.4 and 82 mg/L Se at ambient 323
sulfate concentrations of 50 and 14,000 mg/L, respectively. 324
Similar to results for the acute studies, the effect levels from the chronic D. viridis 325
study are considerably higher than observed for other algal species that have been tested 326
with selenate, although the amount of data available are relatively limited. For example, 327
selenate chronic values for the freshwater green algae Selenastrum capricornutum and 328
Scenedesmus obliquus are in the range of 0.1-0.3 mg/L Se [32], compared with 14 mg/L 329
Se obtained for D. viridis in this study. While selenate toxicity to algae is also sulfate 330
dependent [30], the D. viridis study was conducted in an artificial media with a sulfate 331
concentration of only 195 mg/L. Normalizing this value to 50 mg/L sulfate (generally 332
comparable to standard freshwater algal test media) only lowers the estimated chronic 333
value for D. viridis to 6.3 mg/L. Hence, D. viridis appears to be substantially less 334
sensitive than freshwater green algae that have been previously tested. 335
In the chronic brine shrimp study, growth of the parental generation on Day 11 336
and reproduction on Day 21 were the two most sensitive endpoints, with both endpoints 337
having a NOEC of 3 mg/L and LOEC of 8 mg/L Se. Hence, the chronic value for this 338
study is the geometric mean of the NOEC and LOEC, 5 mg/L. Published data on the 339
chronic sensitivity of other invertebrate species to selenate are limited to an LOEC of 340
>0.7 mg/L for the amphipod Hyalella azteca [33]. 341
Overall, the sensitivity of resident biota was relatively well characterized by the 342
studies performed. One shortcoming was the lack of testing of the corixid, Trichocorixa 343
18
verticalis, which has sporadically been observed in the discharge channel perimeter. 344
Although no standard toxicity testing with this species has been conducted, Thomas et al. 345
[34] did assess the short-term (48 hours) bioaccumulation of Se in T. verticalis by 346
exposing organisms to Se concentrations as high as 1 mg/L with no effect on survival. 347
Hence, the 48-hour LC50 for this species is >1 mg/L Se. 348
An overall assessment of the selenium toxicity data generated in this study 349
indicates brine shrimp is the GSL's most sensitive species resident, with a chronic value 350
of 5 mg/L Se. In comparison, selenium concentrations in the mine effluent typically 351
range from 20-50 µg/L Se, approximately two orders of magnitude lower than those 352
predicted to cause chronic effects. Accordingly, the direct effects of Se on resident biota 353
are not the critical exposure pathway in deriving a site-specific water quality discharge 354
limit for the GSL. 355
356
Bioaccumulation Study 357
358
When tissue Se in brine shrimp is plotted as a function of co-located waterborne 359
Se concentrations, a relatively strong relationship is observed (r2 = 0.92) (Figure 4). 360
These data demonstrate an inverse relationship between waterborne exposure 361
concentration and corresponding bioaccumulation factor that is frequently observed for 362
metals [35, 36]. Using this relationship, the site-specific waterborne Se concentration 363
that results in the dietary threshold for aquatic dependent birds can be estimated. For this 364
assessment, we used a conservative avian dietary Se threshold of 5 mg/kg dw [37, 38]. A 365
dietary threshold for birds was used because they are considered to be more sensitive that 366
19
other wildlife species. Using the equation for the linear regression model in Figure 3, the 367
water Se concentration resulting in a brine shrimp selenium concentration of 5 mg/kg dw 368
can be back-calculated using the following equation: 369
Slope
Intercept - ThresholdDietary StandardQuality Water Specific-Site = (3) 370
Where: Dietary Threshold = 5 mg/kg dw 371
Intercept = 2.2802 372
Slope = 0.1002 373
Using this equation, a waterborne Se concentration of 27 µg/L is the maximum 374
concentration that will not result in brine shrimp Se concentrations equal to or greater 375
than the avian dietary threshold of 5 mg/kg dw. Given that this value is more than two 376
orders of magnitude lower than the lowest effect level observed for direct Se toxicity to 377
resident aquatic biota, Se bioaccumulation in brine shrimp and subsequent dietary 378
toxicity to aquatic birds clearly represents the most critical exposure pathway. 379
Consequently, a site-specific water quality discharge limit of 27 µg/L Se appears 380
protective for aquatic species and sensitive wildlife for this site. 381
382
CONCLUSION 383
384
The GSL is a unique ecosystem in the United States for which there are no water 385
quality criteria and existing freshwater or marine national water quality criteria are 386
inappropriate due to the unique water quality characteristics and biota of the Lake. Using 387
a risk-based approach, we evaluated critical exposure pathways for Se released into this 388
environment with the objective of setting a site-specific water quality discharge limit. 389
20
Resident aquatic biota were found to be comparable in sensitivity to other species that 390
have been tested, but naturally high ambient sulfate concentrations significantly reduce 391
Se bioavailability in this environment. Field bioaccumulation data collected from the 392
study site indicate that waterborne Se concentrations as high as 27 µg/L will not result in 393
an exceedance of the Se dietary threshold for aquatic birds that feed on resident biota. 394
Therefore, 27 µg/L Se appears to be an appropriate site-specific water quality discharge 395
limit for the protection of all exposure pathways at this site. 396
397
REFERENCES 398
399
1. Stephens DE. 1990. Changes in lake levels, salinity and the biological community 400
of Great Salt Lake (Utah, USA), 1847-1987. Hydrobiol., 197:139-146. 401
402
2. Felix EA ,Rushforth SR. 1979. The algal flora of the Great Salt Lake, Utah, 403
U.S.A. Nova Hedwigia, 31:163-195. 404
405
3. Arnow T.1980. Water budget and water-surface fluctuations in the Great Salt 406
Lake, In Gwynn JW, Editor Great Salt Lake: A Scientific, Historical, and 407
Economic Overview. Utah Geological and Mineral Survey: Salt Lake City, Utah. 408
p. 255-264. 409
410
4. Post FJ. 1980. Biology of the North Arm. Utah Geo. Min. Sur. Bull., 116:313-411
321. 412
413
5. Wurtsbaugh WA ,Berry TS. 1990. Cascading effects of decreased salinity on the 414
plankton, chemistry, and physics of the Great Salt Lake (Utah). Can. J. Fish. 415
Aquat. Sci., 47:100-109. 416
417
6. Felix EA ,Rushforth SR.1980. Biology of the south arm of the Great Salt Lake, In 418
Gwynn JW, Editor Great Salt Lake: A Scientific, Historical, and Economic 419
Overview. Utah Geological and Mineral Survey: Salt Lake City, Utah. p. pp. 305-420
312. 421
422
7. Jorgenson EC. 1956. Ephydra of Utah. 1956, University of Utah: Salt Lake City, 423
Utah. 424
425
21
8. Post FJ. 1977. The microbial ecology of the Great Salt Lake. Microb. Ecol., 426
3:143-165. 427
428
9. Stephan CE, Mount DI, Hansen DJ, Gentile JH, Chapman GA, Brungs WA. 1985. 429
Guidelines for deriving numerical national water quality criteria for the protection 430
of aquatic organisms and their uses. 1985, U.S. Environmental Protection 431
Agency, Environmental Research Laboratory: Duluth. p. 98 pp. 432
433
10. DeForest DK, Brix KV, Adams WJ. 1999. Critical review of proposed residue-434
based selenium toxicity thresholds for freshwater fish. Hum. Ecol. Risk Assess., 435
5:1187-1228. 436
437
11. Fairbrother A, Brix KV, Toll JE, McKay S, Adams WJ. 1999. Egg selenium 438
concentrations as predictors of avian toxicity. Hum. Ecol. Risk Assess., 5:1229-439
1253. 440
441
12. USEPA. 1993. Methods for measuring the acute toxicity of effluents and 442
receiving waters to freshwater organisms. 1993, U.S. Environmental Protection 443
Agency: Cincinnati, Ohio. 444
445
13. Brix KV, Cardwell RD, Adams WJ. 2002. Chronic toxicity of arsenic to the Great 446
Salt Lake brine shrimp, Artemia franciscana. Ecotoxicol. Environ. Saf.:In 447
Review. 448
449
14. USEPA. 1994. Short-term methods for estimating the chronic toxicity of effluents 450
and receiving waters to freshwater organisms. 1994, U.S. Environmental 451
Protection Agency: Cincinnati, Ohio. 452
453
15. EU. 1992. European Union Protocol C.3. Official Journal of the European 454
Communities, L383A:179-186. 455
456
16. Cutter GA. 1986. Speciation of selenium and arsenic in natural waters and 457
sediments. Vol. 1: selenium speciation. 1986, Electric Power Research Institute: 458
Palo Alto, California. 459
460
17. EcoAnalysis. 1994. TOXIS. 1994: Ojai, California. 461
462
18. STSC I. 1991. Statgraphics. 1991: Rockville, Maryland. 463
464
19. Norusis MJ. 1994. SPSS for Windows. 1994: Chicago, Illinois. 465
466
20. USEPA. 1995. Method 1669: Sampling ambient water for trace metals at EPA 467
water quality criteria levels. 1995, U.S. Environmental Protection Agency, Office 468
of Water: Washington, D.C. 469
470
22
21. Measures CI ,Burton JD. 1978. Behaviour and speciation of dissolved selenium in 471
estuarine waters. Nature, 273:293-295. 472
473
22. Neal RH ,Sposito G. 1989. Selenate adsorption on alluvial soils. Soil Sci. Soc. Am. 474
J., 53:70-74. 475
476
23. USEPA. 1987. Ambient water quality criteria for selenium - 1987. 1987, U.S. 477
Environmental Protection Agency, Office of Water: Washington, D.C. p. 121 pp. 478
479
24. USEPA. 2002. Draft aquatic life water quality criteria for selenium. 2002, U.S. 480
Environmental Protection Agency: Washington, D.C. p. 159 pp. + appendices. 481
482
25. Forsythe BL ,Klaine SJ. 1994. The interaction of sulfate and selenate (Se6+) 483
effects on brine shrimp, Artemia spp. Chemosphere, 29:789-800. 484
485
26. Hansen LD, Maier KJ, Knight AW. 1993. The effect of sulfate on the 486
bioconcentration of selenate by Chironomus decorus and Daphnia magna. Arch. 487
Environ. Contam. Toxicol., 25:72-78. 488
489
27. Maier KJ ,Knight AW. 1993. Comparative acute toxicity and bioconcentration of 490
selenium by the midge Chironomus decorus exposed to selenate, selenite, and 491
seleno-DL-methionine. Arch. Environ. Contam. Toxicol., 25:365-370. 492
493
28. Kumar HD ,Prakash G. 1971. Toxicity of selenium to the blue-green algae, 494
Anacystis nidulans and Anabaena variabilis. Ann. Bot., 35:697-705. 495
496
29. Ogle RS ,Knight AW. 1996. Selenium bioaccumulation in aquatic systems: 1. 497
Effects of sulfate on the uptake and toxicity of selenate in Daphnia magna. Arch. 498
Environ. Contam. Toxicol., 30:274-279. 499
500
30. Williams MJ, Ogle RS, Knight AW, Burau RG. 1994. Effects of sulfate on 501
selenate uptake and toxicity in the green alga Selenastrum capricornutum. Arch. 502
Environ. Contam. Toxicol., 27:449-453. 503
504
31. Brix KV, Volosin JS, Adams WJ, Reash RJ, Carlton RG, McIntyre DO. 2001. 505
Effects of sulfate on the acute toxicity of selenate to freshwater organisms. 506
Environ. Toxicol. Chem., 20:1037-1045. 507
508
32. Vocke RW, Sears KL, O'Toole JJ, Wildman RB. 1980. Growth response of 509
selected freshwater algae to trace elements and scrubber ash slurry generated by 510
coal-fired power plants. Wat. Res., 14:141-150. 511
512
33. Brasher AM ,Ogle RS. 1993. Comparative toxicity of selenite and selenate to the 513
amphipod Hyalella azteca. Arch. Environ. Contam. Toxicol., 24:182-186. 514
515
23
34. Thomas BV, Knight AW, Maier KJ. 1999. Selenium bioaccumulation by the 516
water boatman Trichocorixa reticulata (Guerin-Meneville). Arch. Environ. 517
Contam. Toxicol., 36:295-300. 518
519
35. Adams WJ, Conard B, Ethier G, Brix KV, Paquin PR, DiToro DM. 2000. The 520
challenges of hazard identification and classification of insoluble metals and 521
metal substances for the aquatic environment. Hum. Ecol. Risk Assess., 6:1019-522
1038. 523
524
36. McGeer JC, Brix KV, Skeaff JM, DeForest DK, Brigham SI, Adams WJ, Green 525
AS. 2002. The inverse relationship between bioconcentration factor and exposure 526
concentration for metals: implications for hazard assessment of metals in the 527
aquatic environment. Environ. Toxicol. Chem., Submitted. 528
529
37. Heinz GH, Hoffman DJ, Gold LG. 1989. Impaired reproduction of mallards fed 530
an organic form of selenium. J. Wildl. Manage., 53:418-428. 531
532
38. Skorupa JP, Morman SP, Sefchick-Edwards JS. 1996. Guidelines for interpreting 533
selenium exposure of biota associated with nonmarine aquatic habitats. 1996, U.S. 534
Fish and Wildlife Service, National Irrigation Water Quality Program: 535
Sacramento, California. p. 74 pp. 536
537
24
Table 1. Dilution water quality during the acute and chronic toxicity tests. 538
Parameter Range
Temperature (°C) 25 ±1
pH 7.9 – 8.4
Dissolved Oxygen (mg/L) 1.8 – 6.0
Salinity (g/L) 80 – 102
Total Organic Carbon (mg/L) 35 – 49
Dissolved Organic Carbon (mg/L) 34.8 – 40
Total Suspended Solids (mg/L) 5-18
Sulfate (mg/L) 5,800
539
25
Table 2. Composition of artificial water used for testing Dunaliella viridis. 540
Salt Concentration (mg/L)
NaCl 100,000
MgCl2 6 H2O 1,500
MgSO4 7 H2O 500
KCl 200
CaCl2 2 H2O 400
KNO3 1,000
NaHCO3 43
H3BO3 2.86
MnCl2 4 H2O 1.81
ZnSO4 7 H2O 0.222
Na2MoO4 2 H2O 0.39
CuSO4 5 H2O 0.079
Co(NO3)2 6 H2O 0.049
FeCl3 6 H2O 2.44
KH2PO4 35
26
Table 3. Summary of acute toxicity test results (all values are mg/L Se). 541
Species Se Form LC50 (95% C.L.) NOEC LOEC
Artemia franciscana Selenate 78 (71-86) 51 71
Ephydra cinerea Selenate 490 (445-542) 369 691
542
543
544
27
Table 4. Summary of chronic Dunaliella viridis toxicity test (all values are mg/L Se). 545
Evaluation Specific Growth Cumulative Area Under Growth Curve
EC50 (95% C.I.) 45 (36-71) 32 (28-36)
NOEC 11 11
LOEC 18 18
Chronic Value 14 14
546
28
Table 5. Summary of chronic A. franciscana toxicity test results. 547
Endpoint Evaluation mg/L Se
Survival – parental Day 11 NOEC 38
LOEC 74
Survival – parental Day 21 NOEC 74
LOEC >74
Survival – parental Day 28 NOEC 74
LOEC >74
Growth – parental Day 11 NOEC 3
LOEC 8
Growth – parental Day 28 NOEC 15
LOEC 38
Reproduction - parental Day 21 NOEC 3
LOEC 8
Reproduction - parental Day 28 NOEC 15
LOEC 38
Survival - F1 NOEC 15
LOEC 38
Growth - F1 NOEC 15
LOEC 38
Final Chronic Value 5
548
29
Table 6. Summary of Co-located Selenium Data in Surface Water and Brine Shrimp. 549
Sample Date Station
Total Se
(µg/L)
Dissolved Se
(µg/L)
Tissue Se
(mg/kg dw) BAF
6/21/98 1 120 121 15.5 129
6/21/98 2 117 116 15.4 132
6/21/98 3 85 81 7.82 92
6/21/98 4 30 30 3.36 112
6/21/98 5 2 2 2.75 1375
6/21/98 6 2 2 2.86 1430
6/21/98 7 2 2 3.14 1570
8/27/98 7 1 1 3.38 3380
550
30
Figure 1 - Map of Study Area and Sampling Locations.

Figure 2 – Conceptual Model.

Figure 3 - Species Sensitivity Distribution for Selenate (Sulfate-Normalized).

Figure 4 - Relationship Between Water and Brine Shrimp Selenium Concentrations.
31

32

Algae
Brine Shrimp
Effluent Surface Water
Aquatic Birds
Brine Flies
Waterborne Toxicity
Dietary Toxicity
Waterborne Toxicity
Dietary Toxicity
Waterborne Toxicity
33

Ephydra cinerea
Artemia franciscana
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
10 100 1000 10000 100000 1000000 10000000
Selenium (µg/L)
Pe
r
c
e
n
t

S
p
e
c
i
e
s

A
f
f
e
c
t
e
d
Raw data
Logistic
34

0
2
4
6
8
10
12
14
16
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Water Se (ppb)
Ti
s
s
u
e

S
e

(
p
p
m

d
w
)
y = 0.1002x + 2.2802
R2 = 0.9198