1.Throw out the most relevant data.
2. Make up new data.
3. Wrongly assume ducks have a less harmful form of mercury than fish do.
4. Reduce risk by averaging out time, geography, and species.
5. Make finding data difficult.
6. Eschew epidemiological opportunities.
7. Conclusion.

1. Throw out the most relevant data.

“Approximately 1,000 tissue sample results were available for mercury, arsenic, and lead in fish and waterfowl collected from OU2 [Operable Unit 2]. However, as most of these results were collected prior to 2010, they are not representative of current conditions. Therefore, these tissue results were not used directly in the HHRA [Human Health Risk Assessment] . . .” (CBandI Federal Services, LLC, 2017, Section 4.1, p. 4-1, pdf page 73).

I could not find graphs or statistical analyses of mercury concentrations of fish or ducks over time in the Superfund site documents that I could download to indicate that, in fact, concentrations before 2010 were different from concentrations after 2010. In reporting mercury concentrations in fish from “Lahontan Reservoir”, CBandI Federal Services, LLC (2017, Exhibit 2-2, p. 2-12) and U.S. Environmental Protection Agency (2020, Table 4, p. 10) did not even bother to distinguish between pre-2010 and post-2010 data. The NDOW walleye data in the table of Mercury Concentrations in Fish From “Lahontan Reservoir” looks periodic with repeating highs and lows so that establishing a recent trend would require considerable post 2010 data. The U.S. Environmental Protection Agency (2020, Table 4, p. 10) has post-2010 data for only 4 of the many fish species in “Lahontan Reservoir”. It apparently has no post-2010 duck data.

Contrary to the quotation above, 1990-1992 tissue data were used to estimate current and perpetual animal tissue data and Attachment B-14 “Occurrence and Distribution of Chemicals of Potential Concern Bird Tissue – Carson River Mercury Site” (CBandI Federal Services, LLC, 2017, pdf page 1,341) does not list any bird data sets other than for 1990-1992. There are a few post-2010 fish data. Attachment B-14 “Occurrence and Distribution of Chemicals of Potential Concern Fish Tissue – Carson River Mercury Site” (pdf pages 1,339-1,340) indicates at least 5 and up to 25 more (i.e., collection years “2002, 2012”) of about 270 mercury analyses of fish could have been collected after 2010. Consequently, the intent of the above statement may be to discourage anyone from looking at any previously published data and to trust only CBandI Federal Services, LLC.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

2. Make up new data.

Continuing the quotation in #1 above:

“Approximately 1,000 tissue sample results . . . but they were used along with matching concentrations in soil, sediment, or surface water to estimate site-specific BAFs [bio-accumulation factors]. These BAFs were then applied to OU2 [Operable Unit 2] subarea-specific abiotic exposure point concentrations to estimate COPC [chemical of potential concern] concentrations in food items (fish, wild plants, small game, and waterfowl).” (CBandI Federal Services, LLC, 2017, p. 4-1, pdf page 73).

Instead of using measurements of mercury in a sample of ducks to determine the general mercury concentrations in ducks, CBandI Federal Services, LLC (2017) used mercury concentrations in pond bottom sediments. For the different subareas of Operable Unit 2, they averaged sediment mercury concentrations and “waterfowl” mercury concentrations and calculated bioaccumulation factors by dividing the average (or some other measure of central tendency) waterfowl mercury concentrations by the average (or some other measure of central tendency) sediment mercury concentration to arrive at a bioaccumulation factor. Some other sediment statistic was then multiplied by the bioaccumulation factor to estimate some statistic of the mercury concentrations in all waterfowl in each subarea. That waterfowl statistic was ultimately used to estimate the amount of mercury that would be ingested by someone who eats ducks from the subarea.

There is no information in CBandI Federal Services, LLC (2017) that suggests this method is likely to give accurate results.

The Human Health Risk Assessment in Appendix B of CBandI Federal Services, LLC (2017) stated: “There are no readily-available literature-based BAFs for the sediment to waterfowl pathway.” (CBandI Federal Services, LLC, 2017, p. B-6-24, pdf page 338).

The Screening Level Ecological Risk Assessment in Appendix C of CBandI Federal Services, LLC (2017) stated: “There are no readily-available literature-based BAFs for the sediment to waterfowl pathway.” (CBandI Federal Services, LLC, 2017, p. C-7-11, pdf page 1,591).

Tuttle and others (2000, p. 34) found that “Geometric-mean concentrations of mercury in sediment and food-chain organisms did not correlate with mean concentrations in bird eggs or livers” at Stillwater National Wildlife Refuge and “Carson Lake”.

The method of using bioaccumulation factors based on sediment mercury concentrations to estimate duck mercury concentrations could have been tested. Rowe and others (1991) have mercury concentrations in waterfowl from Stillwater Wildlife Management Area, “Carson Lake”, Fernley Wildlife Management Area, Humboldt Wildlife Management Area, Massie Slough, and Mahala Slough. The sloughs are between Fernley and Fallon. They also have bottom sediment concentrations of mercury for Stillwater Wildlife Management Area and Fernley Wildlife Management Area. Hoffman and others (1990) reported results for a few samples of bottom sediments from Stillwater and “Carson Lake” and for more samples of waterfowl from “Carson Lake” and Stillwater, Fernley, and Humboldt wildlife management areas. There may be bottom sediment data for Humboldt Wildlife Management Area and the sloughs elsewhere. A glance at the data suggests coots, at least, were collected at all sites.

There may well be similar data available for national wildlife refuges in other areas because the Stillwater studies were part of a nation-wide response to the biological effects of irrigation drainage at Kesterson National Wildlife Refuge (selenium was the killer there). The scientific approach would have been to calculate a regression of coot (+/- other species) mercury and sediment mercury using at least 3 areas, or more if available. The Environmental Protection Agency’s approach was to use a novel, non-validated method to make up data instead.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

The bioaccumulation factor method is rather circular. It goes from sediment to duck to find bioaccumulation factor and then, with the same bioaccumulation factor, from different sediment to final duck. The equation is:
sediment Hg (total) x BAF = duck Hg (total). The steps are as follows.

a) Calculate the bioaccumulation factor by selecting certain sediment and certain biological data. Use total mercury concentrations, rather than methylmercury or mercuric chloride concentrations, because that is all there is.

The data used in the BAF calculations are buried in Attachment B-13 tables for “BAF Backup Information – Sediment Data” and “BAF Backup Information – Waterfowl Tissue Data” in Attachment B-1a of Appendix B of CBandI Federal Services, LLC (2017, pdf pages 1,332-1,337). Of 88 analyses for Operable Unit 2, subarea D, CBandI Federal Services, LLC (2017) selected 48 1990-1992 analyses of mostly lake sediment in “SWMA” (the now defunct Stillwater Wildlife Management Area, which includes Stillwater National Wildlife Refuge and some areas to the west). The geometric mean concentration of mercury in sediment was 0.3098 ppm.

I identified the data that were used by calculating the arithmetic and geometric means of various blocks of data and comparing them to those in Table B6-4 “Sediment to Waterfowl Bioaccumulation Factors for Mercury, Arsenic, and Lead” in CBandI Federal Services, LLC (2017, pdf page 410). The number of analyses in the appropriate blocks were “SED FOD” in Table B6-4. Data for “Carson Lake” and various drains and canals were excluded from the means. Maybe the purpose of the selection is explained in some other document but the effect of the selection was to lower the sediment geometric mean.

CBandI Federal Services, LLC (2017) selected 90 of 107 1990-1992 analyses of stilts, coots, and ducks. The geometric mean concentration of mercury in ducks was 0.4224 ppm (wet weight basis).

Again, I identified the data used by calculating the arithmetic and geometric means of various blocks of data and comparing them to those in Table B6-4 “Sediment to Waterfowl Bioaccumulation Factors for Mercury, Arsenic, and Lead” in CBandI Federal Services, LLC (2017, pdf page 410). The number of analyses in the appropriate blocks were “Tissue FOD” in Table B6-4.

The inclusion of protected, migratory stilts raised the duck geometric mean above what is appropriate for ducks. If “Eurasian teal” is just the incorrect name for green-winged teal rather than a fabricated species (it’s not listed in my 1977 Audubon Society Field Guide to North American Birds), then the preponderance of teal data (45 analyses versus 39 for all other ducks and coots) lowered the geometric mean below representative levels. Green-winged teals at “Carson Lake and Pasture” had lower mercury concentrations than mallards and shovelers but higher concentrations than redheads (see reduce risk by averaging out species). Green-winged teals comprised 32% of the average state-wide hunter take in 1973-1992 (Hogan and Smucker, 1994, p. 70).

The selected duck data were only for “Stillwater WMA”. Analyses of the 13 birds shot at “Carson Lake” were excluded. Geometric means of “Stillwater WMA” and “Carson Lake” bird tissue concentrations were 0.42 and 1.77 mg/L mercury, respectively. Including the “Carson Lake” data and 4 “Canal/Ditch” analyses of ducks would have raised the geometric mean used in the bioaccumulation factor calculation from 0.42 ppm to 0.50 ppm (wet weight basis).

The bioaccumulation factor was calculated by dividing the 0.422 ppm geometric mean total mercury in waterfowl (wet weight basis) by the 0.31 ppm geometric mean total mercury in sediment concentration and rounding to 1.36 (CBandI Federal Services, LLC, 2017, Table B6-4 “Sediment to Waterfowl Bioaccumulation Factors for Mercury, Arsenic, and Lead”, pdf page 410).

0.422 ppm geometric mean Hg (total) for selected duck data / 0.31 ppm geometric mean Hg (total) for selected sediment data = 1.36 for biased bioaccumulation factor

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

b) Determine an “exposure point concentration” by using sediment mercury data that are different from those used in the bioaccumulation factor calculation. Select a statistic to represent this different data set.

The statistic selected for the exposure point concentration was identified in Table 3.5 “Medium-Specific Exposure Point Concentration Summary for Combined Surface Sediment – Carson River Mercury Site”(CBandI Federal Services, LLC, 2017, Attachment B1-a, pdf page 445). This table is a print of a spreadsheet that shows the “exposure point concentration” for mercury in “OU2D Surface Sediment” as 4.68 ppm. A cell to the right labels this as “95% KM-Cheby” statistic. Helpfully(?), the rightmost cell in the row provides the “Rationale” as “Test(1)”. Notes below the spreadsheet print explain Test(1) as “Kaplan-Meier method recommended by ProUCL due to multiple detection limits”.

The exposure point concentration is more than 10 times higher than the geometric mean sediment concentration used in the bioaccumulation factor calculation. This is presumably an effort to include some sort of safety factor in the calculations. To determine if it’s valid requires deeper digging. The “Arithmetic Mean of Detects” in Table 3.5 is 3.61 ppm, although CBandI Federal Services, LLC (2017) refers to ppm as “mg/kg”. Attachment B-9 “Combined Surface Sediment and Saturated Surface Sediment, 95% UCL on the Mean (EPA ProUCL Software), Carson River Mercury Site, OU2” (CBandI Federal Services, LLC (2017, pdf pages 1127-1129 for “Mercury Surface OU2D”) indicates that the 3.61 ppm mean was calculated from 218 “detects” of 254 samples. The median was 0.535 ppm. These are not the sediment data used in the calculation of the bioaccumulation factor.

The data used to calculate the exposure point concentration are in Attachment B-15. The “Combined Surface Sediment Sample List” for OU2D starts on pdf page 1,413 (CBandI Federal Services, LLC, 2017). After entering the B-15 listed total mercury data for subarea OU2D in a spreadsheet, I had 254 analyses and a mean of 3.622 ppm and a median of 0.550 ppm for 218 analyses without the “u” flag, which I presume means below the detection limit. The mean and median of the 218 “detects” are pretty close to the 3.61 mean and 0.535 median in Attachment B-9. I may well have made a transcription error or 3. The geometric mean of the “detects” is 0.776 ppm. This is more than twice the geometric mean used in the bioaccumulation factor calculation.

The 254 analyses include the 48 analyses used to calculate the bioaccumulation factor. This again begs the question of why and how the 48 analyses were selected. Of course, if all 254 analyses had been used to calculate the bioaccumulation factor, then calculation of a mercury concentration in ducks using the bioaccumulation factor would have been a tautology.

Use of a 95% confidence limit based on some inferred data distribution must account for the rest of the 10-fold increase over the geometric mean of the 48 sediment analyses. The “95% KM Chebyshev UCL” may seem large compared to the geometric mean sediment concentration used in the bioaccumulation factor calculation but it is less than half of the 11 ppm 90th percentile. Maybe it provides an adequate margin of safety, maybe not.

There are other problems. Use of the sediment data in Attachment B-15 without regard to representativeness leads to a waterfowl mercury statistic strongly biased toward Stillwater National Wildlife Refuge and biased toward “West Dry Lake” within the refuge. The 254 analyses include 39 analyses with “Carson Lake” as the “subunit”, 13 with “Indian Lakes”, 2 with “seepage ponds”, 1 with Sheckler Reservoir, and 198 with “Stillwater WMA” as the “subunit”. Values which are flagged “u” are included in the following statistics. They appear to be half the reporting/detection limit, i.e., 0.02 ppm mercury is not flagged but 0.01 ppm is. Units are ppm.

• “Stillwater WMA”: n=198, median=0.31, geometric mean=0.35, range 0.007 to 30.4.
• “Carson Lake”: n=39, median=5.68, geometric mean=1.91, range 0.025 to 18.0.
• “Indian Lakes”: n=13, median=0.07, geometric mean=0.09, range 0.02 to 0.51.

For the full B-15 data set (including “u”), the median of 0.320 ppm is very close to that of the Stillwater WMA analyses and the geometric mean of 0.429 ppm is only 23% higher than that of “Stillwater WMA” alone. The high mercury concentrations in sediments of “Carson Lake” have been diluted into obscurity by the Stillwater data. Weighting by areas or bird populations would have been defensible, weighting by availability is not.

The generally low concentrations in sediments at “Indian Lakes” in the full data set are surprising. Hallock and others (1993, Figure 15, p. 53) reported 52.6 ppm and 97.8 ppm mercury in detritus there. While detritus is more organic matter than sand and silt particles, those high values beg the question of how the organic matter got such high values if the maximum mercury concentration for “Indian Lakes” sediment in Attachment B-15 is 1.04 ppm. Bioaccumulation factor of more than 50?

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

Within Stillwater National Wildlife Refuge, you can get about any statistic you want between 0.02 and 20 ppm mercury by choosing how many analyses to use from each of 5 lakes sampled multiple times by U.S Fish and Wildlife Service and U.S. Geological Survey scientists. As shown in the Histogram of Mercury Concentrations in Lake Sediment in Stillwater National Wildlife Refuge, different lakes have sediments with different, mostly well separated mercury concentrations. Attachment B-15 is heavily weighted with low mercury concentrations from “West Dry Lake”.

Data from CBandI Federal Services, LLC, 2017, Attachment B-15 “Combined Surface Sediment Sample List” (pdf pages 1413-1425 for OU2D).

Statistics, in ppm, for the 5 lakes in the histogram are as follows:

• “East Alkali Lake”: n=11, median=0.150, geometric mean=0.187, range 0.11 to 1.68.
• “Lead Lake”: n=25, median=12.6, geometric mean=14.8, range 7.03 to 30.4.
• “Stillwater Point Reservoir”: n=31, median=1.90, geometric mean=0.926, range 0.07 to 3.29.
• “Swan Check”: n=17, median=0.567, geometric mean=0.594, range 0.50 to 0.87.
• “West Dry Lake”: n=110, median=0.174, geometric mean=0.105, range 0.007 to 11.2.

“West Dry Lake” accounts for 57% of the 194 analyses with lake names.  I excluded 4 analyses for “South Lead Lake” due to their uncertain locations and possible overlap with “Lead Lake”.

To illustrate the effect of weighting by sample size, all but 11 analyses were randomly deleted for each lake in order to conform to the number of samples in “East Alkali Lake”. This was done 10 times and medians and geometric means were calculated for each trial. The median of the 10 resulting medians was 0.555 ppm (average median 0.567 ppm, range 0.540-0.653 ppm) and the geometric mean of the 10 geometric means was 0.686 ppm (average mean 0.692 ppm, range 0.563-0.914 ppm). Without the dominance of “West Dry Lake”, medians and geometric means were higher by about a factor of 2.

Weighting the statistics of each lake equally to calculate a more regional statistic may not be a good idea though. The ultimate goal is a mercury statistic for ducks (or “waterfowl”) so the best weighting scheme would be one that considers how much ducks feed at each lake. Nest counts, bird counts, or food abundance might be close enough. In the absence of that data, one could weight the data by the area of each lake. Tracing the lakes from 1:24,000-scale 7.5-minute topographic quadrangles, overlaying them on a grid, and counting the grid intersections in each lake is a quick way to estimate area without a GIS program. Giving “West Dry Lake” an area of 1.0 results in an area of 0.88 for “Swan Check”, 1.91 for “East Alkali Lake”, 2.15 for “Lead Lake”, and 5.35 for “Stillwater Point Reservoir”. The relatively small area of “West Dry Lake” makes it clear that weighting by number of analyses is unlikely to result in a statistic representative of the area of interest.

There was no sediment data for “Lead Lake” in the bioaccumulation factor calculation. The full data set shows that if there had been, the bioaccumulation factor would have turned out to be bioreduction, i.e., birds would have lower mercury concentrations than sediments do. Bioaccumulation of mercury is well documented and well accepted by biologists so bioreduction might not have looked good. Biologists, though, generally compare mercury concentrations in aquatic animals with concentrations in water when discussing bioaccumulation. In the glossary of Hoffman (1994, p. 31), bioaccumulation is defined as “The ability of an organism to accumulate a chemical to concentrations of one or more orders of magnitude greater than the concentration in water or food”.

exposure point concentration = 4.68 ppm Hg (total)

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

c) Calculate the concentration of total mercury in ducks once and for all by multiplying the exposure point concentration of total mercury in sediment by the bioaccumulation factor.

4.68 ppm Hg (total) sediment x 1.36 = 6.36 ppm Hg (total) duck

The concentration of mercury in ducks calculated by CBandI Federal Services, LLC (2017) is not 6.36 ppm (wet weight basis). 6.36 ppm could be considered harmful. Instead, Table B6-6 “Estimation of OU2 COPC Concentrations in Plants, Game, and Waterfowl, Bioaccumulation from Soil or Sediment” (CBandI Federal Services, LLC, 2017, pdf page 413) has the “Estimated Concentration in Food” for “Waterfowl” as “NE = not estimated (bioaccumulation assumed to be zero)” for elemental mercury, mercuric sulfide, and methylmercury. However, the concentration of mercuric chloride in waterfowl is 0.78 ppm (wet weight basis).

As discussed in Section 3, CBandI Federal Services, LLC (2017) thinks all the mercury in waterfowl is mercuric chloride, notwithstanding the fact that total mercury was used in the calculation of the bioaccumulation factor. The sediment “Surface EPC” that gives rise to this result is 0.56 ppm mercuric chloride, not 4.68 ppm total mercury. Whereas the broader data set and the use of a 95% upper confidence limit had increased the sediment mercury statistic from a geometric mean of 0.31 ppm by a factor of 15, the 4.68 ppm has been reduced by a factor of 8 to 0.56 ppm.

The conversion of the exposure point concentration from total mercury to mercuric chloride accounts for the factor of 8 decrease. Mercuric chloride was assumed to be 12% of total mercury (“Notes” in Table B6-1 “Estimation of Mercury Species Concentrations in OU2 Abiotic Media”, CBandI Federal Services, LLC, 2017, pdf page 407). Table B6-6 “Estimation of OU2 COPC Concentrations in Plants, Game, and Waterfowl, Bioaccumulation from Soil or Sediment”, (CBandI Federal Services, LLC, 2017, pdf page 413) also shows that the bioaccumulation factor has been rounded to 1.4. 0.56 ppm x 1.4 = 0.78 ppm. Converted to a dry weight basis by multiplying by 3.6 (see Hoffman and others, 1990, Table 7, p. 26), 0.78 ppm (wet) gives 2.81 ppm (dry).

The assumption that mercury in bottom sediments of lakes is 12% mercuric chloride uses the pretense of 2 published studies which found the percentage to be 3.4% or 17% and of Environmental Protection Agency data reported by Hogan and Smucker (1994) (CBandI Federal Services, LLC, 2017, Table B4-1, pdf page 406). The first study was of tailings from mining placer gold with dredges near Redding, California. Mercury concentrations of 94 ppm in fine tailings and 194 ppm in coarse tailings are a long way from the geometric mean of 0.31 ppm in sediments in Stillwater National Wildlife Refuge. Rivers in the Sierra Nevada foothills also have different biological environments than marshes of the Carson Desert.

The other study was of a location in Brazil downstream of a gold mining region and of a mercury mine with mercury sulfide minerals that may make it “less representative of mercury speciation that is occurring at the CRMS” (CBandI Federal Services, LLC, 2017, p. B-4-3, pdf page 307).

Table B4-1 attributed a ratio of 4% to Hogan and Smucker (1994, Table B.3, p. 133). However, that report showed that 9 duplicate analyses by the Environmental Protection Agency of 34 samples from Operable Unit 1 analyzed by Oak Ridge Research Institute failed to yield even approximately close results for total mercury, “soluble mercury” (assumed to be mercuric chloride), or the ratio. To take 2 examples, the Environmental Protection Agency found total mercury of 991 ppm, soluble mercury of 15 ppm, and a ratio of 1.5% for sample MS 005-SL-12-A and total mercury 1,154 ppm, soluble mercury 112 ppm, and a ratio of 9.7% for sample MS012-SL-38-A. The comparable numbers determined by Oak Ridge Research Institute were 176 ppm, 88.5 ppm, and 50.3% for MS 005-SL-12-A and 780 ppm, 0 ppm, and 0% for MS012-SL-38-A (Hogan and Smucker, 1994, Tables B.2 and B.3, p. 132-133, pdf pages 154-155). Something is obviously wrong with these numbers but it was assumed that the Environment Protection Agency results were valid. Although Table B4-1 shows 4% for mercuric chloride, the mean, median, and geometric mean for the Hogan and Smucker (1994) data are 2.81%, 2.88%, and 3.04%. The factor of 50 difference between the high value of 9.7% and the low value of 0.2% indicate large uncertainties associated with this data.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

Adding to the numerical uncertainties are environmental differences, the Hogan and Smucker (1994) samples are from dry mill sites and tailings piles above Dayton, in Operable Unit 1, with very high total mercury concentrations (261-3,124 ppm) rather than from wet, vegetated sediments with low mercury concentrations like Stillwater Marsh. Further emphasizing the differences between dry soil and wet sediment, CBandI Federal Services, LLC (2017, p. B-4-3 to B-4-4, pdf pages 307-308) stated the following regarding results from the Carson River in subarea B:

“The finding that the sediment samples had an average methylmercury to total mercury ratio of 0.35 percent that was almost 4-fold greater than the average methylmercury to total mercury ratio of 0.091 in the soil samples is not unexpected, as methylation in the aquatic environment is greater than methylation in more upland areas (such as the bank soils).”

CBandI Federal Services (2017, p. B-4-4, pdf page 308) chose 10% for mercuric chloride speciation as “it is conservative (health protective) to use an upper-bound value for this form of mercury.” Except that they don’t know that it is really health protective because 3 flawed and irrelevant studies don’t provide a reliable upper bound on mercuric chloride speciation in sediments of Stillwater Marsh and “Carson Lake and Pasture”. 2% is added to 10% to account for mercuric oxide and mercuric sulfate “as this is the value of the one available result, rounded to one significant figure” (CBandI Federal Services, 2017, p. B-4-4, pdf page 308). It must be true that those species are converted to mercuric chloride in birds. 12% it is.

0.56 ppm Hg (chloride) sediment x 1.40 = 0.78 ppm Hg (chloride) duck (wet weight basis)

It has been a roundabout journey. The geometric mean of the 48 Stillwater WMA sediment analyses used in the bioaccumulation factor calculation, 0.31 ppm (median 0.32 ppm), was increased to 0.429 ppm (median 0.32 ppm) by including “Carson Lake”, “Indian Lakes”, and additional Stillwater WMA analyses in the full data set of 254 analyses. This was bumped up to a geometric mean of 0.776 ppm (median 0.535 ppm in Attachment B-9) by eliminating analyses below the detection limit. Then the geometric mean was eclipsed by the chosen 4.68 ppm “95% KM Chebyshev UCL”, a factor of 6 increase, in what may have been an effort to add a safety factor. A bit of furious back-pedaling then knocked the 4.68 ppm back down to 0.56 ppm, a factor of 8 decrease, by taking 12% of total mercury to get mercuric chloride. That’s a lot of smoke and mirrors to get a number that is pretty much the same as the 0.535 ppm median of the 218 “detects” of the 254 sediment analyses.

So that’s where the 0.78 ppm mercury concentration for all ducks in Operable Unit 2 comes from, now and forever more.

An alternative to this elaborate fabrication process is to find the median or other statistics of all the duck data (i.e., excluding avocets and stilts) in Attachment B-13 of (CBandI Federal Services, 2017) and in the 1987-1989 duck muscle data for “Carson Lake and Pasture” of Hoffman and others (1990) and Rowe and others (1991). Before calculating the comprehensive duck statistics, I converted the “Carson Lake and Pasture” concentrations to a wet weight basis using the given moisture contents of each analysis before calculating the statistics. The median of the 63 moisture contents was 73.512%, which would indicate a conversion factor of 3.78 ( = 1/(1-.73512) rather than 3.6 as generally used on these web pages and by Hoffman and others (1990, Table 7, p. 26). The results (wet weight basis) in units of ppm are:

• “Carson Lake” (wet): n=72, median=0.797, geometric mean=0.764, range 0.033-15.1.
• Stillwater WMA (wet): n=88, median=0.429, geometric mean=0.380, range 0.023-4.37.

The combined data set has a median of 0.515 ppm and a geometric mean of 0.520 ppm but, as discussed in “Geography” of step 4, mercury concentrations in ducks from the 2 areas are significantly different. For context, a duck concentration of 0.78 ppm mercury corresponds to the 58th percentile of the combined data. It lacks any appreciable margin of safety. Using the 90th percentile of the combined Stillwater – “Carson lake” mercury data for ducks, which is 3.16 ppm (wet weight basis), for the mercury dosage calculations could theoretically (i.e., assuming the data adequately represent mercury concentrations that vary by year, location, and species) protect duck hunters from all but the 10% of ducks with the highest mercury concentrations.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

Because all mercury in ducks is assumed to be mercuric chloride, 0.78 ppm on a wet weight basis (2.81 ppm dry weight basis) is the concentration of total mercury in waterfowl in Operable Unit 2, subarea D, that has been used to determine whether any human eating any ducks from the area is likely to consume a harmful amount of mercury. Summarizing the several flaws in the calculation of 0.78 ppm (wet weight basis) as the mercury concentration for all ducks in Operable Unit 2, subarea D before proceeding to calculate human dosage:

• There is no justification for using sediment concentrations of mercury to determine mercury concentrations in birds.
• Even if there were, a linear relation (i.e., duck = B x sediment) between mercury concentration in sediment and mercury concentration in duck has not been established, even theoretically. Other relations are possible. Mercury concentrations in ducks could increase slowly with mercury in sediments at first but then increase more rapidly as high concentrations of mercury become available (e.g., duck = e^(B x sediment)). Or, there could be a saturation effect where the rate of mercury increase in ducks slows at high mercury concentrations in sediments (e.g., duck = e^(A – B/sediment)).
• “Carson Lake”, “Indian Lakes”, and Stillwater National Wildlife Refuge should have been treated separately. They are far enough apart that ducks may feed mostly in 1 area rather than flying frequently between the areas.
• Because sediment mercury concentrations and numbers of samples differ markedly from lake to lake within Stillwater National Wildlife Refuge, some rational weighting method should have been used to combine data from different lakes rather than weighting all the analyses equally.
• Duck tissue data collected in 1987-1989 by Hoffman and others (1990) and Rowe and others (1991) were not used.
• Avocets and stilts should not have been used in the calculation of a mean (or other central tendency) mercury concentration for ducks. They are protected migratory birds. Their inclusion biases duck concentrations higher.
• Ducks at “Carson Lake” and Stillwater National Wildlife Refuge have different mercury concentrations and should have been treated separately. See “Geography” of section 4.
• Different species of ducks commonly have different concentrations of mercury and should have been treated separately so hunters could avoid those with the highest mercury concentrations. See “Species” of section 4. Both NDOW and the U.S. Environmental Protection Agency have treated fish species in the Carson River Mercury Superfund Site separately; why not bird species?
• Duck tissue data other than those collected in 1990-1992 were not used. Mercury concentrations in ducks change from year to year and samples from the period 1990-1992 are unlikely to capture the real range or the central tendency. See “Time” of section 4.
• Duck tissue concentrations change seasonally and the 1990-1992 data might be biased if the ducks were not all collected during the hunting season.

0.78 ppm (wet weight basis) is the number used to calculate the risk to humans of eating ducks from Operable Unit 2, subarea D. How much the eater weighs and how much duck is eaten are also used in the calculation of risk. It wasn’t just the mercury concentration in ducks that was made up. Data on how much duck duck hunters in Nevada eat was also made up. In the draft human health risk assessment and remedial investigation report for the Carson River Mercury Superfund Site, Hogan and Smucker (1994) had used data from NDOW to calculate yearly consumption of duck. That data was ignored. NDOW has a voluntary small game hunt questionnaire online (Small Game Regulations, p. 7) that could have provided relevant information but CBandI Federal Services, LLC (2017) evidently didn’t ask for it.

An odd inconsistency in CBandI Federal Services, LLC (2017) reveals bigger problems with estimates of ingestion rates. According to Tables 4.25 and 4.26, Attachment B-1a, Appendix B, CBandI Federal Services, LLC (2017, pdf page 475, 476), recreational adults eat 1.6 times more duck than fish but recreational children eat 2.2 times more fish than duck. Looking deeper, the inconsistency in the eating habits of adults and children is a result of using different sources to make up the fish and duck ingestion rates. If ingestion rates had both been taken from data collected during the same 1987-1988 Nationwide Food Consumption Survey, the estimates would have resulted in both children and adults eating 3.8 times more fish than duck. The 2 relevant tables in U.S. Environmental Protection Agency (2011, Chapter 13. Intake of Home-Produced Foods) are Table 13-24. Consumer-Only Intake of Home-Caught Fish (g/kg-day)―West (pdf page 37) and Table 13-41. Consumer-Only Intake of Home-Produced Game (g/kg-day) with a line for the West Region (pdf page 54).

Neither of the sources used by CBandI Federal Services, LLC (2017) to estimate fish and duck ingestion rates are likely to provide rates relevant to Nevada hunters and anglers. For fish, CBandI Federal Services, LLC (2017) chose not to use the 1987-1988 Nationwide Food Consumption Survey.

“[I]t represents a 7-day survey period conducted during a 1987-1988 Nationwide Food Consumption Survey and does not account for variations in eating habitats during the rest of the year, and it represents home-caught fish that include both finfish and shellfish (Table 13A-1 in U.S. Environmental Protection Agency, 2011). It is unclear from the Nationwide Food Consumption Survey results what percentage of ‘home-caught fish’ represents fishing from stock ponds (compared with recreational fishing from rivers and lakes). Based on these uncertainties, a fish intake of 267 gpd is not recommended” (CBandI Federal Services, LLC, 2017, p. B-6-8, pdf page 322).

Instead, it chose to use a survey of recreational anglers in King County, Washington, in 2007 because it provides a “freshwater recreational intake rate for a western state” (CBandI Federal Services, LLC, 2017, p. B-6-8, pdf page 322). That survey had 128 “all respondents” and 81 “children of respondents” (U.S. Environmental Protection Agency, 2011, Chapter 10. Intake of Fish and Shellfish, Table 10-87 Consumption Rates (g/day) for Freshwater Recreational Anglers in King County, WA, pdf page 170). King County includes Seattle, Bellevue, and Redmond, but not Tacoma to the south, Bremerton to the west, or Everett to the North and continues east as far as Snoqualmie Pass on I-90. The results of the survey aren’t too far out of line with other surveys that could have been chosen from Chapter 10 but it seems an odd choice for Nevada.

CBandI Federal Services, LLC’s (2017) choice to estimate duck ingestion rates by Nevada hunters seems wildly inappropriate. Although Table 13-41. Consumer-Only Intake of Home-Produced Game (U.S Environmental Protection Agency, 2011, Chapter 13. Intake of Home-Produced Foods) does not define game, the term “game” implies deer, other large quadrupeds, pheasant, and other birds in addition to ducks. Use of “game” inflates the estimate of duck consumption and, hence, risk. Use of “consumer-only” rather than hunter-only deflates the estimate of duck consumption and, hence, risk. The combination of these 2 factors could fortuitously balance out or the resulting Nevada duck consumption could be much too high or much too low.

CBandI Federal Services, LLC (2017) selected the 95th percentile of the West Region home-produced game ingestion rate as the duck ingestion rate for recreational users of the Carson River Mercury Superfund Site. This may have been some standard margin of safety allowance but, in the end, it was for naught. The 95th percentile of the consumption rate for a consumer in the West from Table 13-41. Consumer-Only Intake of Home-Produced Game (g/kg-day) (U.S. Environmental Protection Agency, 2011, Chapter 13. Intake of Home-Produced Foods, pdf page 54) is 1.22 g/kg body weight – day. An adult weighing 80 kg (176 pounds) would eat 98 g/day, on average. An assumed post-cooking loss of 29.7% reduces that to 69 g/day (CBandI Federal Services, LLC, 2017, p. B-6-8, pdf page 322). For reference, the annual ingestion would amount to 25.2 kg (55.5 pounds). That would be a lot of duck but not so much deer.

An ingestion rate of 69 g/day would seem to be sufficient to calculate dosage but CBandI Federal Services, LLC (2017) decided that a daily consumption rate for western consumers was a hunting-day consumption rate for Nevada duck hunters.

“The exposure frequency for a sportsman hunting waterfowl is set equal to 10 days per year. Although the Nevada waterfowl hunting season is 107 days in duration (typically mid-October through late January), most hunters only participate 5 to 10 days out of this period (Neill, 2016)” (CBandI Federal Services, LLC, 2017, p. B-6-7, pdf page 321).

This seems to imply that 1 and only 1 duck with an edible weight of 69 g (2.4 ounces) is killed and eaten on each hunting day.

CBandI Federal Services, LLC (2017) multiplied 69 g/day by 10 hunting days per year to determine that Nevada duck hunters eat 690 g (1.5 pounds) of duck per year rather than 25.2 kg (55.5 pounds) (Attachment B-1a, Table 4.25 “Values Used for Daily Intake and Exposure Calculations – Current/Future Exposures to Food Items Carson River Mercury Site, OU2”, “Recreational User (Adult)”, pdf page 475).

Hogan and Smucker (1994, Table E.7, p. 85, pdf page 210) had previously estimated that Nevada duck hunters eat 12 g of duck on 350 days of the year for a total of 4,200 g (9.3 pounds) per year. By switching a daily rate to a hunting-day rate, CBandI Federal Services, LLC (2017) fabricated a consumption rate that is a factor of 6 less than that of Hogan and Smucker (1994) in spite of using a 95th percentile and a food item that is mostly something other than duck.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

3. Wrongly assume ducks have a less harmful form of mercury than fish do.

That methylmercury is the dominant form of mercury in fish is well documented (WHO and FAO, 2011, and references therein). Table B6-5 “Estimation of OU2 COPC Concentrations in Fish, Bioaccumulation from Surface Water” (CBandI Federal Services, LLC, 2017, pdf page 411) lists the estimated methylmercury concentration in game fish of OU2D (Operable Unit 2, subarea D, which includes “Carson Lake”) as 0.81 ppm (wet weight basis) but does not list other forms of mercury.

The form of mercury in ducks was assumed to be mercuric chloride rather than methylmercury. Table B6-6 “Estimation of OU2 COPC [chemical of potential concern] Concentrations in Plants, Game, and Waterfowl, Bioaccumulation from Soil or Sediment” (CBandI Federal Services, LLC, 2017, pdf page 413) shows an estimated 0.78 ppm (2.8 ppm dry weight basis) mercuric chloride in “waterfowl” of OU2D. Methylmercury is “NE = not estimated (bioaccumulation assumed to be zero)”.

Searching for “mercuric chloride” in the Carson River Mercury Superfund Site documents that I have been able to download turned up no evidence to support the claim that mercury in ducks only occurs as mercuric chloride.

Due to publisher restrictions on no-registration public access and to search engines’ inability to catalogue and produce results for scientific articles, the following published results cannot be considered conclusive but nonetheless offer no support for the choice of CBandI Federal Services, LLC, in its remedial investigation of the Carson River Mercury Superfund Site.

• Vermeer and others (1973) found that methylmercury in the muscle of 5 ducks sampled at “Clay Lake” in western Ontario “ranged from 69 to 99 percent of total mercury”.
• The percentages of methylmercury (“organic mercury”) in livers of 12 species of seabirds nesting on Gough Island in the south Atlantic varied widely from 2.6% to 92.6% of total mercury (Thompson and Furness, 1989).
• The average percentage of methylmercury in muscle of 9 species of seabirds was 66% in the study of Kmil and others (1996).
• Methylmercury concentrations in adult muscle of white-tailed eagles in Poland were 29% (n=3), 46% (n=8), or 63% (n=3) of total mercury depending on location. One adult osprey had muscle tissue with methylmercury 92% of total mercury (Kalisinska and others, 2014, Table 2, p. 862-863).
• In a literature review, Kalisinska and others (2014, p. 867) found that for white-tailed eagles in Norway, methylmercury percentages in muscle were 59% in one group with lower mercury concentrations and 65% in the other group.
• Methylmercury percentages of muscle and intestine in cormorant, great crested grebe, and Eurasian buzzard collected in the Czech Republic were 71% to 94% of total mercury (Houserova and others, 2007).
• The Utah Department of Health stated that “In duck tissue, the majority of mercury is methylmercury” in its evaluation of whether to issue health advisories for consumption of ducks shot near “Great Salt Lake” (Scholl and Ball, 2006, p. 5).
• Methylmercury was 61% of total mercury in northern shoveler livers, on average, and 78% of total mercury in cinnamon teal livers from the “Great Salt Lake” area (Cline and others, 2011, p. 146).
• In a study of the effects of mercury contamination on birds that visit San Francisco Bay, Ackerman and others (2014) wrote: “Elsewhere we have shown that methylmercury comprises 96 percent of mercury in bird eggs (Ackerman and others, 2013b) and 94 percent of mercury in prey fish (Ackerman and Eagles-Smith, 2010a; Ackerman and others, U.S. Geological Survey, unpub. data, 2013), and we provide a detailed assessment of the percentage of mercury in the methylmercury form in liver and kidneys in Eagles-Smith and others (2009b). Therefore, we analyzed the majority of tissue samples for total mercury only because nearly all mercury in most tissues (except liver, kidney, and invertebrates) is in the methylmercury form.” Further, their Figure 5 (p. 83) has graphs of methylmercury and total mercury in the livers, kidneys, blood, muscles, and feathers of avocets, black-necked stilts, Caspian terns and Forster’s terns from San Francisco Bay. Liver and kidney results plot on or slightly below the 1:1 line that indicates all the mercury is in the form of methylmercury. Muscle results are below but close to the line. Figure 86 (p. 159) shows that methylmercury concentrations in livers of the 4 species scatter symmetrically around the line for 90% of total mercury up to about 8 ppm total mercury but then decrease to about 60% of total mercury at 60 ppm. This is because “Demethylation of mercury to an inorganic form occurs in waterbird livers above a threshold dose that varies by species: Caspian and Forster’s terns 7.48 ppm dry, avocets and stilts 9.91 ppm dry” (Ackerman and others, 2014, p. 44).

I didn’t find any articles that suggested mercury in birds is predominantly “inorganic mercury” or mercuric chloride.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

Elsewhere in the remedial investigation pdf file, CBandI Federal Services, LLC (2017) reported that 90% or more of the total mercury in birds was in the methylmercury form. The Final Carson River Mercury Site Operable Unit 2 Human Health Risk Assessment Report, which is included in CBandI Federal Services, LLC (2017) as Appendix B (starting on pdf page 283), has the following comments on mercury in birds.

“Average total mercury levels in the Lahontan Reservoir swallow blood (wet weight), liver (wet weight), and feather samples (dry weight) were 2.6, 4.0, and 2.0 mg/kg [ppm], respectively, compared with 0.74, 0.98, and 2.4 mg/kg [ppm], respectively, at the background location. Methylmercury accounted for greater than 90 percent of total mercury in blood, liver, and feathers of the fledgling swallows.” (p. B-3-9, pdf page 303)

“Average total mercury levels for cormorants samples collected at Lahontan Reservoir were 17.1 mg/kg [ppm](wet weight) for blood and 105 mg/kg [ppm](dry weight) for feathers, compared with 0.49 mg/kg [ppm](wet weight) for blood and 9.0 mg/kg [ppm](dry weight) for feathers from the background location. As was observed for swallows, methylmercury accounted for greater than 90 percent of total mercury in cormorant blood and feather samples.” (p. B-3-9, pdf page 303)

“In the US Geological Survey (USGS) study by Henny et al. (2002), methylmercury in adult double-crested cormorants, snowy egrets, and black-crowned night-herons contained very high concentrations of total mercury in their livers, with geometric mean concentrations of 134.8 mg/kg [ppm](wet weight), 43.7 and 13.5, respectively; and kidney concentrations of 69.4, 11.1, and 6.1, respectively” (p. B-3-9, pdf page 303).

This statement would make more sense if “methylmercury in” were deleted and “total mercury” were replaced by methylmercury. It’s as if the author initially wanted to write “methylmercury in birds was such and such” but then decided to include data for both kidneys and livers and realized it would be better to change the subject to “birds”.

Except for the extra “s” on “cormorants” and spelling out “Geological Survey”, the above statements also appear in the Screening Level Ecological Risk Assessment Report, which is included as Appendix C in CBandI Federal Services, LLC (2017) (starting on pdf page 1,539).

“Average total mercury levels in the Lahontan Reservoir swallow blood (wet weight), liver (wet weight), and feather samples (dry weight) were 2.6, 4.0, and 2.0 mg/kg, respectively, compared with 0.74, 0.98, and 2.4 mg/kg, respectively, at the background location. Methylmercury accounted for greater than 90 percent of total mercury in blood, liver, and feathers of the fledgling swallows.” (p. C-3-11, pdf page 1,561)

“Average total mercury levels for cormorant samples collected at Lahontan Reservoir were 17.1 mg/kg (wet weight) for blood and 105 mg/kg (dry weight) for feathers, compared with 0.49 mg/kg (wet weight) for blood and 9.0 mg/kg (dry weight) for feathers from the background location. As was observed for the swallows, methylmercury accounted for greater than 90 percent of total mercury in cormorant blood and feather samples.” (p. C-3-11, pdf page 1,561)

“In the USGS study by Henny et al., 2002, methylmercury in adult double-crested cormorants, snowy egrets, and black-crowned night-herons contained very high concentrations of total mercury in their livers, with geometric mean concentrations of 134.8 mg/kg (wet weight), 43.7 and 13.5, respectively; and kidney concentrations of 69.4, 11.1, and 6.1, respectively.” (p. C-3-12, pdf page 1,562)

CBandI Federal Services, LLC (2017) wrote that mercury in swallows, cormorants, snowy egrets, and black-crowned night herons is 90% or more methylmercury twice and then assumed that ducks have only the mercuric chloride form. Without the assumption that mercuric chloride in sediment is 12% of total mercury, the concentration of mercury in ducks of Operable Unit 2, subarea D, to be used in the human risk assessment would have been 8 times (1/0.12) higher. The 8 times higher mercury concentration in ducks would have resulted in a calculated dosage in duck hunters that is also 8 times higher.

The Environmental Protection Agency has determined that mercuric chloride is less toxic than methylmercury by a factor of 3. Table 6-1 “Non-Cancer Toxicity Data – Oral/Dermal, Carson River Mercury Site, OU2” (CBandI Federal Services, LLC, 2017, pdf page 480) shows the oral reference dose for mercuric chloride as 0.0003 milligram per kilogram body weight – day and that for methylmercury as 0.0001 milligram per kilogram body weight – day. For dried biological samples, the corresponding weights would be 0.0011 mg/kg body weight – day and 0.00036 mg/kg body weight – day. Routine consumption of amounts less than the reference dose are not considered harmful. That means it would be fine to eat a duck (total mercury assumed to be mercuric chloride) with 3 times the mercury concentration of a fish (total mercury assumed to be methylmercury) which has just below the reference dose, provided that the amounts of duck and fish eaten are the same.

Other organizations have also considered mercuric chloride as less toxic than methylmercury. For “inorganic mercury” in foods other than fish and shellfish, the World Health Organization determined a “provisional tolerable weekly intake” of 0.004 mg per week – kg of body weight (World Health Organization, 2021). This is based on the results of experiments in which rats were fed mercuric chloride (WHO and FAO, 2011, p. 672-673). Mercuric chloride affects the kidneys and immune system rather than the nervous system and does not cause developmental problems in fetuses. The World Health Organization’s corresponding “provisional tolerable weekly intake” of methylmercury in fish is 0.0016 mg per week – kg body weight (World Health Organization, 2021). The 2 values differ by a factor of 2.5. This is close to the factor of 3 difference in the U.S. Environmental Protection Agency’s reference doses. The “provisional tolerable weekly intakes” are probably for wet samples.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

The difference in reference doses is not the only difference in mercury risks for eating ducks or fish. The “remedial action objectives” for mercuric chloride and methylmercury in the proposed plan for Operable Unit 2 indicate that mercuric chloride is about 7 times less toxic than methylmercury (U.S. Environmental Protection Agency, 2021, p. 9).
_____”Remedial action objective” 1 is “Reduce the risk to adults and children practicing the traditional tribal lifestyle from consuming mercury-contaminated waterfowl and wild plants”, including reducing “the consumption of waterfowl and wild plants containing concentrations of mercuric chloride above 3.3 mg/kg [ppm]” (dry weight basis is 11.9 ppm).
_____”Remedial action objective” 2 is “Reduce human health risks from consumption of mercury-contaminated game fish. Under this RAO, EPA’s goal is to reduce game fish consumption levels to the EPA and FDA advisory of 0.46 mg/kg [ppm] for total mercury by the following exposure pathways”, which are all for fish (dry weight basis is 1.66 ppm). It was previously stated that “Mercury accumulates as methylmercury in game fish tissues” (p. 6).

The reason the oral reference doses for methylmercury and mercuric chloride differ by a factor of 3 but the Remedial Action Objectives differ by a factor of 7 is complicated. The first complication is different consumption rates.

Starting with fish, Table 2.6 of Aptim Federal Services, LLC (2018, pdf page 256) shows that only traditional tribal adults and youths foraging off-reservation have hazard quotients above 1 in Operable Unit 2, subarea D. Recreational users of “Carson Lake and Pasture” and Stillwater Marsh are ignored because they don’t. Consequently, the remedial action objectives apply only to traditional tribal foragers. The consumption data for traditional tribal youths and adults are buried in tables 4.23 for adults and 4.24 for youths in Attachment B-1a of Appendix B of CBandI Federal Services (2017, pdf pages 473, 474). Youths and adults are both assumed to eat 200 g of fish flesh per day on 37 days of the year but youths weigh 44.3 kg (98 pounds) and adults weigh 80 kg (176 pounds). Assuming that youths and adults consume the same amount of fish and ducks over the year does not make sense but that doesn’t matter for made up numbers.

With meal sizes, meal frequency, and body weights in hand, the maximum acceptable mercury concentration in food can be backed out of the intake equation in Tables 4.23 and 4.24 (CBandI Federal Services, 2017, pdf pages 473, 474). Using the methylmercury reference dose of 0.0001 mg/kg body weight – day as the intake, the “CFI” (concentration in food item) is 0.22 ppm for fish eaten by youths and 0.39 ppm for fish eaten by adults. The maximum acceptable fish mercury concentration for youths is lower because they weigh less. 0.22 ppm and 0.39 ppm are the “Methylmercury Performance Standards” that appear in Table 2.6. This validates the method I used to derive the remedial action objectives.

[Note: The intake equation and variables in Table 4.23, Attachment B-1a, Appendix B, CBandI Federal Services, LLC (2017, pdf page 473), and similar tables are misleading.
The equation is:
Intake (mg/kg/day) = (CFI x IR-FD x CF4 x ED x EF x FI) / (BW x AT)

• “Intake (mg/kg/day)”. A unit that is in the divisor of a first divisor is a dividend of the first divisor, i.e., mg/kg/day = mg x day/kg. This is nonsensical. The proper notation is mg/kg – day, where the dash is not a subtraction sign.
• “CFI” is the concentration of a chemical in the food item. Its stated units are mg/kg. The result of dividing mg by kg is as follows.
  1 mg = 10exp-3 g and 1 kg = 10exp+3 g
  10exp-3 g / 10exp+3 g = 10exp-6 or 1/1,000,000, or 1 part per million, or 1 ppm.
  A ppm is a unitless fraction like a percent. A quantity of grams divided by grams is unitless; the grams cancel out. “mg/kg” is fictitious notation. That said, it is very widely used. Notably, the U.S. Geological Survey has used both mg/kg and micrograms per gram for ppm. The National Institute of Standards and Technology has disapproved of using the unit ppm, as well as ppb and ppt because million, billion, and trillion supposedly do not mean the same things in all the languages in which they appear (Thompson and Taylor, 2008, p. 20). The previously different meanings of billion and trillion in British and American English are well known but million has always meant the same in American English, British English, and French. As an example, Thompson and Taylor (2008) used ng/kg for ppt. Writing x 10-6, where -6 is a superscript, is an acceptable substitute for ppm but superscripts don’t always display properly in browsers.
• “IR-FD” is the ingestion rate of food in g/day.
• “CF4” is a conversion factor with a value of “1.0E-03” in units of kg/g. This is more fictitious notation, which follows from the fictitious notation for the concentration. A dividend in kg is needed to cancel out the kg in the divisor of CFI for the units in the equation to come out right. If the units of concentration were recognized as ppm, then calculations would use 78/1,000,000 or 0.0000078 to represent 78 ppm, for example. This is like using 0.05 as the multiplier rather than 5 to find 5% of something. But ppm is 1/1,000,000 and CF4 is 1/1,000. That’s fine. The mg in CFI is fictitious too so the grams in ingestion rate have to be converted to the milligrams in the intake. Dividing by 1,000 accomplishes that. Without the fictitious notation, 1/1,000,000 x 1,000 mg/g = 1/1,000, which is the same as CF4.
• “ED” is exposure duration in years. “AT” is averaging time in days.
  For the tables in CBandI Federal Services, LLC (2017), AT is always 365 x ED. Maybe it would be different in other risk estimate contexts. Because AT days = 365 x ED years, ED years / AT days = 1 year / 365 days. ED/AT is always the same, regardless of the assumed exposure duration. This is good because the exposure durations look like they have been made up too. The exposure duration for recreational adults is 26 years while that for recreational children is 6 years. This implies that adults move away after 26 years and kids leave after 6 years. They could become adults but the intake equation does not take into account the possibility that an adult might have prior exposure as a child or a child has subsequent exposure as an adult. There are no citations to support this. Tough luck for those who spend their entire lives in Churchill County or vicinity and hunt and fish there. Similarly odd, traditional tribal adults have an exposure duration of 70 years (i.e., they essentially never move and are capable of maintaining their hunter-gatherer lifestyle throughout their lifespans) but traditional tribal youths have an exposure duration of 10 years. The Environmental Protection Agency is counting on the tribal youths leaving. This begs the question of how different durations can even be justified. Luckily, it doesn’t matter. It all works out to 1/365.
• “EF” is the exposure frequency or how often duck or fish is eaten in days/year. Because ED/AT is 1 year/365 days, multiplying by EF cancels out both days and years. The intake equation can also be written for an average meal size per day. In that case, the variables EF, ED and AT can be eliminated. IR-FD says it all.
• “FI” is the fraction of the duck or fish that is eaten. In Table 4.23 and similar tables, it is always 1.0. This implies that the ingestion rate takes into account how much of the animal is edible and cooking losses. Hogan and Smucker (1994, p. 69-70) used FI = 0.5 and considered that conservatively high, i.e., people probably eat less than half the weight of a duck or fish.
• “BW” is the body weight of the person eating the duck or fish, in kg. In the U.S. Environmental Protection Agency’s risk assessments, it is the average weight of the population at risk. It changed from 70 kg in 1994 (Hogan and Smucker, 1994, Tables E.6, E.7, p. 185, pdf page 210) to 80 kg in 2017 (CBandI Federal Services, LLC, 2017, e.g., Table 4.25, pdf page 475).

Although misleading, the intake equation can be used as is to give mathematically appropriate results. It can also be simplified.]

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

Using the intake equation to back out thresholds of acceptable mercury concentrations in waterfowl for traditional tribal youths and adults from the reference dose assuming 40 g/day of duck are eaten on 37 days (for off-reservation foraging in OU2D) (CBandI Federal Services, 2017, Attachment B-1a of Appendix B, tables 4.23 and 4.24, pdf pages 473, 474), the results are 3.28 ppm mercury in ducks eaten by youths and 5.92 ppm mercury in ducks eaten by adults. Rounding 3.28 ppm gives Remedial Action Objective #1 of 3.3 ppm (wet weight basis) (U.S. Environmental Protection Agency, 2021, p. 9).

Inconveniently, these are not the numbers given in Table 2.5 Estimation of Numeric Performance Standards for Wild Plants and Waterfowl, CRMS OU2 FS (Aptim Federal Services, LLC, 2018, pdf page 254) for traditional tribal youths and adults in Operable Unit 2, subarea D. There, the “Mercuric Chloride Performance Standards” for “Waterfowl” for these populations are “NA”, or not applicable, because the hazard quotients are less than 1. That doesn’t matter because I used the reference dose as the concentration in the food item, not the actual concentration of ducks in the subarea. Conveniently, hazard quotients were greater than 1 for traditional tribal youths in subareas A, B, and C. For them, the waterfowl performance standard is 3.3 ppm. As I confirmed in my calculation, this is the maximum acceptable concentration of mercury in a food item based on the reference dose and is not a function of the estimated concentration in the food item from a particular area.

Circling back to the greater differences in the remedial action objectives for fish and ducks than in the reference doses for methylmercury and mercuric chloride, the complication due to consumption rate, as used in the intake equation, can be determined. The intake equation is linear so the factor of 5 difference between the ingestion rates of 200 g/day for fish and 40 g/day for ducks (CBandI Federal Services, LLC, 2017, Attachment B-1a of Appendix B, tables 4.23 and 4.24, pdf pages 473-474) carries through without change. Multiplying the factor of 5 difference in ingestion rates by the factor of 3 difference in reference doses implies that acceptable mercury concentrations in ducks are 15 times higher than those in fish.

Remedial Action Objective #1 of 3.28 ppm is not 15 times greater than Remedial Action Objective #2 of 0.46 ppm. The second complication is that the 0.22 ppm mercury concentration I calculated for fish (above) is not the Remedial Action Objective #2 in the proposed plan for Operable Unit 2 (U.S. Environmental Protection Agency, 2021, p. 6). In 2017, the U.S. Food and Drug Administration issued guidance that fish with mercury concentrations up to 0.46 ppm are okay to eat (Aptim Federal Services, LLC, 2018, p. 1-14, pdf page 44). This value is the value adopted as Remedial Action Objective #2. Dividing the 3.28 ppm performance standard for ducks by the 0.22 ppm performance standard for fish that I calculated does give 15.

The significance of the calculations above is that the assumption that traditional tribal members eat 5 times more fish than duck results in much higher acceptable mercury concentrations in ducks than in fish. This compounds the increase in maximum acceptable mercury concentration in ducks relative to fish caused by assuming the mercury in ducks is mercuric chloride.

To put some numbers on how this works, I used the intake equation to calculate the mercury concentration in ducks that would equal the mercuric chloride reference dose with the parameters specified by Table 4.25 of Attachment B-1a (CBandI Federal Services, LLC, 2017, pdf page 475). For a recreational adult weighing 80 kg (176 pounds), the ducks could have up to 12.7 ppm (wet weight basis, 45 ppm dry) mercury. Using the less fanciful duck ingestion data in Table E.7 of Hogan and Smucker (1994, pdf page 185) and the methylmercury reference dose, an 80 kg person could eat ducks with mercury concentrations up to 0.70 ppm (wet weight basis, 2.5 ppm dry). The numbers made up by Consistently Bogus and Imaginary Federal Services, LLC (2017) result in a threshold for a potentially harmful concentration of mercury in ducks that is about 18 times higher than a realistic threshold.

In light of the above, the fabricated mercury concentration in ducks of Operable Unit 2, subarea D, takes on particular significance. That number is 0.78 ppm (wet) and greater than the 0.70 ppm (wet) maximum acceptable concentration in food calculated with Hogan and Smucker’s (1994) parameters for recreational adults (except with 80 kg rather than 70 kg body weight) and the methylmercury reference dose. Switching to a higher mercuric chloride reference dose raises the maximum acceptable concentration to 2.09 ppm (wet) . That effectively eliminates the generalized mercury risk of eating ducks from “Carson Lake” and Stillwater Marsh.

A generalized risk assessment might not satisfy everyone. 18% of the duck analyses exceeded 2.09 ppm (wet) mercury (1990-1992 data for Stillwater WMA; 1987-1989 and 1990-1992 data for “Carson Lake and Pasture”). Hunting at Stillwater Marsh looks okay, only 10% of the 88 duck analyses there had mercury concentrations greater than 2.09 ppm (wet). 25% of the 81 analyses of ducks from “Carson Lake and Pasture” exceeded 2.09 ppm (wet) mercury. Switching from Hogan and Smucker’s (1994) ingestion rate of 4,200 g/year to CBandI Federal Services, LLC’s (2017) ingestion rate of 690 g/year takes care of that by raising the maximum acceptable duck concentration to 12.7 ppm (wet). Only 1 of the analyses from “Carson Lake and Pasture” and none from Stillwater Marsh exceeded 12.7 ppm (wet).

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

4. Reduce risk by averaging out time, geography, and species.

Reduce risk by averaging out time.
Reduce risk by averaging out geography.
Reduce risk by averaging out species.

Time

The answer is: “waterfowl” in OU2D have a mercuric chloride concentration of 0.78 ppm (wet weight basis), or 2.8 ppm (dry weight basis). Proposing a single number to represent mercury concentrations in ducks assumes they never change. The ecosystems of “Carson Lake and Pasture” and Stillwater Marsh will never change. The areas of wetlands will never change. Waterfowl food sources will never change.

Even the limited amount of data for mercury in birds indicates that mercury concentrations in ducks at “Carson Lake and Pasture” do change from year to year. Due to generally small sample sizes, much of the changes in observed mercury concentrations could be due to the non-representativeness of the samples. The changes can be dramatic. The drama may not be justified but there is no way to know without more data.

• Median mercury concentrations in coot livers from “Carson Lake” decreased from 9.20 ppm (n=5, range 4.90-25.0 ppm, dry) in June 1986 to 5.03 ppm (n=10, range 1.98-8.38, ppm dry) in July 1988.
• Median mercury concentrations in mallard livers from “Carson Lake” decreased from 4.40 ppm (n=10, range 0.42-17.0 ppm, dry) in October 1987 to 3.24 ppm (n=4, range 1.30-6.01, ppm dry) in July 1988.
• Median mercury concentrations in shoveler muscle from “Carson Lake” decreased from 21.5 ppm (n=10, range 2.10-55.7 ppm, dry) in October 1987 to 3.55 ppm (n=20, range 0.86-36.0, ppm dry) in October 1989.

Geographic location and timing within a year may affect year to year changes.

• At “Sprig Pond” at “Carson Lake”, median mercury concentrations in stilt livers increased from 7.80 ppm (n=5, range 6.00-12.0 ppm, dry) in July 1986 to 15.6 ppm (n=4, range 12.0-21.2, ppm dry) in August 1987.
• At the Islands Unit at “Carson Lake”, median mercury concentrations in stilt livers decreased from 5.70 ppm (n=5, range 3.60-6.50 ppm, dry) in July 1986 to 3.80 ppm (n=7, range 2.70-11.0, ppm dry) in April 1987.

Significant temporal changes in duck mercury concentrations at “Carson Lake and Pasture” are supported by data from other areas. Mercury concentrations in bird eggs collected in the San Francisco Bay area over the period 2003-2007 changed from year to year.

“Although statistically accounting for region and site effects, mercury concentrations in bird eggs differed among years in all species (avocets: F2,532=8.43, P<0.001; stilts: F2,199=8.79, P=0.001; Forster’s terns: F2,360=12.33, P<0.0001), and nest initiation dates in avocets and stilts (avocets: F1,532=10.28, P=0.001; stilts: F1,199=26.5, P<0.0001; Forster’s terns: F1,360=3.05, P=0.08). Mercury concentrations in bird eggs were higher in 2006 (avocets: 0.18±0.02; stilts: 0.54±0.06; Forster’s terns: 1.71±0.13), than in 2007 (avocets: 0.14±0.01; stilts: 0.40±0.04; Forster’s terns: 1.24±0.07) or 2005 (avocets: 0.13±0.01; stilts: 0.47±0.09; Forster’s terns: 1.11±0.08)” (Ackerman and others, 2014, p. 28).

Bird mercury concentrations also change from year to year at “Great Salt Lake”.

“Mercury concentrations were lowest in 2012, and generally similar in 2010 and 2011. On average, mercury concentrations in eggs were 53% and 60% lower in American avocets [n=336, p=0.0001], 35% and 30% lower in black-necked stilts [n=211, p=0.02], 46% higher and 14% lower in Forster’s terns [n=196, p=0.18, not significant], 21% and 15% lower in marsh wrens [n=274, p=0.001], 24% and 27% lower in white-faced ibis [n=321, p=0.003], and 34% and 36% lower in yellow-headed black birds [n=117, p=0.04] in 2012 compared to 2010 and 2011” (Ackerman and others, 2015, p. 16 and Table 5).

The Environmental Protection Agency’s approach could be justified if the annual variations in mercury concentrations in ducks oscillate about some central tendency without any significant decadal trends. In that case, high-concentration years would be balanced out by low-concentration years over the course of a duck hunter’s life. The problem is that there is not enough data to know what the decadal central tendency is. What looks like a high-concentration year now may turn out to be a low-concentration year over the next couple of decades. Of course, the claim that pre-2010 data is irrelevant merely cements the ignorance.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

Geography

Rather than trying to identify areas with the most contaminated ducks and identify species with the highest mercury concentrations, the Environmental Protection Agency’s approach is simply to not worry about the harm that may befall people who happen to hunt the riskiest species in the riskiest areas more often than not. The fact that there was, and may still be, a Greenhead Hunting Club at “Carson Lake” suggests that some duck hunters are likely to hunt at “Carson Lake” most of the time. Statewide though, duck hunters are only a tiny fraction of the population. Hogan and Smucker (1994, p. 70) reported only 3,200. On a population level, the number of duck hunters who hunt risky areas could be considered insignificant.

In Attachment B-13 “BAF Backup Information – Sediment Data” (CBandI Federal Services, LLC, 2017, pdf pages 1,335-1,337), “Carson Lake” sediments (n=15) have median and range of mercury concentrations of 9.50 ppm and 4.50-14.0 ppm. The comparable numbers for Stillwater WMA (n=48) are 0.32 ppm and 0.08-2.90 ppm. The Histogram of Mercury Concentrations in Bottom Sediment from “Carson Lake” and Stillwater WMA shows that the sediments at “Carson Lake” and Stillwater WMA are distinct populations. There isn’t even any overlap.

Data from CBandI Federal Services, LLC, 2017, Attachment B-13 “BAF Backup Information – Sediment Data” (pdf page 1335).

Limited corroboration comes from the few other available data. Hoffman and others (1990) collected single sediment samples from “Sprig Pond”, the Island Unit, the Big Water Unit, and “Carson Lake” Drain above “Carson Lake” in March 1987. The mercury concentrations of these “Carson Lake” samples were 9.0, 18.0, 3.8, and 1.7 ppm, respectively. Rowe and others (1991) collected samples from Stillwater National Wildlife Refuge and a few canals and drains upstream of it. Samples were split into coarser and finer size fractions that were analyzed separately. Although the bulks of the samples were likely in the finer fractions, the median of the coarser fraction was only 7% greater than the median of the finer fraction. Consequently, I averaged the analyses of the 2 fractions before calculating the sample statistics. The median and range of the mercury concentrations in the Stillwater samples (n=11) were 0.45 ppm and 0.05-3.75 ppm.

Given the stark differences in mercury sediment concentrations at “Carson Lake” and Stillwater WMA, the observations of Ackerman and others (2014, p. 62) regarding waterfowl in the San Francisco Bay area are pertinent.

“Among the most important findings from this study is the recognition that relatively fine scale foraging habitat is a critical driver of mercury exposure and risk in estuarine waterbird communities. . . . Additionally, we showed that waterbirds in the San Francisco Bay Estuary preferentially select some habitats in greater proportion than those habitats’ relative abundance in the environment and those habitats tended to have some of the highest mercury concentrations in prey items. This is an important observation because it highlights the need to quantitatively understand avian space use in heterogeneous environments to adequately characterize risk. Simply sampling habitats in proportion to their availability, randomly sampling from dominant habitats, or making broad assumptions about avian habitat use without quantitative support can be misleading and result in inadequately characterized mercury exposure profiles” (Ackerman and others, 2014, p. 62).

Geographic variations in mercury concentrations are also evident in waterfowl tissue. The Histogram of Mercury Concentrations in Waterfowl from “Carson Lake” and Stillwater WMA shows the non-normal, multi-modal distribution of the data. For those who may be interested, a Mann-Whitney test shows that the probability of the observed ranks for mercury concentrations in waterfowl from “Carson Lake” and Stillwater WMA being due to chance if “Carson Lake” and Stillwater WMA waterfowl mercury concentrations are not different populations is 0.0002 (1-tailed). “Carson Lake” waterfowl concentrations are significantly higher.

Data from CBandI Federal Services, LLC, 2017, Attachment B-13 “BAF Backup Information – Waterfowl Tissue Data” (total mercury in muscle, converted to dry weights) (pdf page 1332).

Averaging out geographic variations in mercury concentrations has resulted in a mischaracterization of risk.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

Species

Waterfowl species do not all have the same tissue levels of mercury when exposed to the same mercury-contaminated environment. Mann-Whitney tests indicate the probability that different ranks of mercury concentrations in shovelers and mallards is due to chance is 0.0025 and that different ranks of mercury concentrations in mallards and redheads is due to chance is 0.006. Conversely, the different ranks in mallards and green-winged teals could be due to chance. That shovelers have higher mercury concentrations than mallards, green-winged teals, and redheads and that mallards have higher mercury concentrations than redheads is made clear by the Histogram of Mercury Concentrations in Duck Species from “Carson Lake and Pasture”.

Data from Hoffman and others (1990) and Rowe and others (1991).

The histograms also suggest that populations of shovelers, mallards, and green-winged teals each have 2 or more modes. These modes could be due to timing or geography or both. The difference in mercury concentrations in shoveler muscle between October 1987 and October 1989 was mentioned above. Mallards sampled in early August 1987 (n=6, median=4.40 ppm) had higher mercury concentrations than those sampled in mid-October (n=10, median=1.36 ppm). The green-winged teals were all shot in October 1987. Geography rather than timing is a possible explanation for the wide range and modes of mercury concentrations in teals.

Although Ackerman and others (2015, p. 11) analyzed mercury concentrations in eggs at “Great Salt Lake” rather than in muscle, they found

“Mercury concentrations in eggs differed substantially among species (fig. 4), while accounting for the effects of region and year (ANOVA: species: F 32,2498=154.00, P<0.0001; region: F7,2498=11.37, P<0.0001; year: F2,2498=28.47, P<0.0001).”

Of 33 species sampled at “Great Salt Lake”,

“Caspian terns had mean egg mercury concentrations that were more than 131 times higher than Canada geese”.

Of water birds sampled at San Francisco Bay by Ackerman and others, (2014, p. 2, 19, and Figure 7, p. 85), Caspian terns (n=50) had liver mercury concentrations significantly higher (p less than 0.05) than those in surf scoters (n=158) and avocets (n=439). Concentrations for Forster’s terns (n=174) and black-necked stilts (n=217) were intermediate between Caspian terns and surf scoters but still significantly greater than those in avocets.

The 21.5 ppm (dry) median of shoveler muscle collected at “Carson Lake” in October 1987 (n=10) is more than 20 times higher than the 0.65 ppm (dry) median of redhead muscle collected in October 1987 (n=8). Failing to recognize species differences in mercury contamination results in a gross mischaracterization of health risks for duck hunters in the Carson River Mercury Superfund Site.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

The Utah Department of Health has taken a different approach to evaluating the risk of consuming mercury-contaminated ducks. In its September 2006 “Health Consultation”, the Department included a table of observed mercury concentrations in duck muscle for 10 species at 3 locations on the edge of “Great Salt Lake” – Bear River, Ogden Bay, and Farmington Bay (Scholl and Ball, 2006, p. 5). The mean mercury concentrations by species and location that exceeded a screening value of 0.3 ppm (wet weight basis, 1.1 ppm dry) were highlighted in bold font. Sample sizes for each species at each location were mostly in the range of 10-30 ducks. The mean was used as the measure of risk even though it is more sensitive to a few high values than a median or geometric mean.

Rather than averaging out geographic differences within the same species, the Department recommended consumption limits for a species if mercury concentrations for that species exceeded the screening value in any one area. For example, consumption limits were recommended for shovelers because the Ogden Bay and Farmington Bay means were above the screening value even though the Bear River mean was not. The advisory applied to all ducks of that species from “the Great Salt Lake marshes” (Scholl and Ball, 2006, p. 7-8). The Farmington Bay sampling area is about 57 km (34 miles) from the Bear River sampling area. It is 55 km (33 miles) from the northern edge of Stillwater National Wildlife Refuge to “Sprig Pond” on the west side of “Carson Lake”.

Rather than averaging out species differences, the Utah Department of Health issued consumption advisories for cinnamon teal, northern shovelers, and common goldeneyes but not for mallard, northern pintail, and other species. Recommended consumption limits were based on meal sizes of 227 g (8 ounces) for adults and 113 g (4 ounces) for children and on body weights of 70 kg (154 pounds) for adults and 16 kg (35 pounds) for children (Scholl and Ball, 2006, p. 7, 38). The Utah Division of Wildlife Resources published the following consumption advisory for northern shovelers and cinnamon teals.

“Adults should eat no more than two 8-ounce meals per month, and children, pregnant women and women who may become pregnant should eat no more than one 4-ounce meal per month of northern shoveler and cinnamon teal harvested from Great Salt Lake marshes” (Utah Division of Wildlife Resources, 2006, p. 2).

Rather than averaging out annual differences, the Utah Department of Health has modified its consumption advisories based on new annual data.

“Mercury levels in Northern Shoveler sampled in Bear River, Ogden Bay and Farmington Bay drainages in 2005-2006 are lower than concentrations found in sampling conducted in 2004-2005. The consumption advisory for this species should reflect this change” (Scholl and Ball, 2006, p. 9).

Mercury concentrations in ducks were apparently not measured again until 2022. On a “Utah Waterfowl Advisories” web page (deq.utah.gov/water-quality/utah-waterfowl-advisories), the Division of Water Quality rescinded the 2006 advisories for northern shoveler and cinnamon teal “as concentrations are below the advisory threshold of 0.3 microgram per gram [ppm]. The consumption advisory for the Common Goldeneye will remain in place due to inadequate data to recommend its removal.” Although maximum mercury concentrations in shovelers and teals were above 0.3 ppm, the means were well below that concentration. Goldeneyes had a maximum concentration of only 0.19 ppm mercury but because only 3 were collected, the Department of Water Quality evidently decided to err on the side of caution.

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

5. Make finding data difficult.

Hiding data or obfuscating data does not itself affect the U.S. Environmental Protection Agencies estimates of the risks from eating mercury-contaminated ducks from “Carson Lake” but it does increase the chances that no one will realize how bad they are. It’s hard to say if this was intentional or simply due to incompetence but the pdf file that contains CBandI Federal Services, LLC (2017) is the epitome of data that is hard to find.

• Create a huge file (96 MB) of at least 4 large documents without a single table of contents. The main document of CBandI Federal Services, LLVC (2017) is the remedial investigation report. It takes up the first 139 pages of the pdf file. Next is Appendix A – Draft Carson River Mercury Site, OU2 RI Statistical Evaluations. Appendix B is Final Carson River Mercury Site Operable Unit 2 Human Health Risk Assessment Report, Carson City, Nevada and starts on pdf page 282. Appendix C is Screening Level Ecological Risk Assessment Report and starts on pdf page 1,539.
• Put the data in a pdf file with 1,982 pages to make searching slow and scrolling hypersensitive.
• Put the data in documents without consecutive page numbers that have multiple, independently numbered sections and subsections and skip page numbering for some figures and tables.
• Include screen prints of work sheets with a misleading “page x of 361” numbering scheme when the page numbers never get to 361 but restart at unpredictable intervals and still use “page y of 361” numbering (e.g., page numbering in Attachment B-9 restarted after pages 1-46 and after 1-53, among others).
• List attachments as a single page in a table of contents when they are actually 10s of pages long (e.g., Attachments 1 and 2 had page numbers 16 and 17 but Attachment 1 started on pdf page 186 and Attachment 2 started on pdf page 202).
• In a list of attachments, don’t indicate how many pages they have (e.g., list of attachments for Appendix B, pdf page 287).
• Give attachments abstruse names (e.g., Appendix B-1a CRMS OU2 RAGS-D Tables 1 through 10).
• Include screen prints of work sheet pages with no table or page numbers and no reference in a contents list (e.g., Attachment B-5).
• Include different data sets in the same attachment so that it is impossible to determine where 1 data set ends and the next begins without scrolling through the whole thing (e.g., the 191 pages of Attachment B-15).
• Don’t sort work sheets by chemical before creating screen prints for the pdf file so that arsenic, lead, and mercury analyses are all mixed up (e.g., Attachment B-15).
• Create work sheet screen prints in an 11 x 17 format so that it is difficult to see data in the useful columns without scrolling back and forth (e.g., Attachment B-15).
• Include blank columns in work sheet screen prints so that it is difficult to see data in the useful columns without scrolling back and forth (e.g., Attachment B-15).
• Include columns in work sheet screen prints with superfluous columns for things that never change, like units (e.g., Attachment B-15).
• Put figures and tables together at the end of a long document so that multiple windows of the document must be kept open to have any hope of verifying statements in the text.
• Don’t put page numbers on different pages of the same table (e.g., Table 3.7).
• Don’t put tables in numerical order (e.g., Table 7.1 follows Table 10.7 and Table 9.1 follows Table 7.16).
• Skip table numbers (e.g., Appendix B has tables 4, 6, 8, and 9 but not 1, 2, 3, 5, or 7).
• Use the same table numbers for different data (e.g., Table 10.1 on pdf page 627 and on pdf page 696).
• Use ambiguous numbering schemes for different sets of tables (e.g., Table 3-1, Table 3.1, and Table 3).
• Don’t give a series of tables any numbers at all (e.g., pdf pages 823-887).
• Repeat key words and phrases frequently in different contexts so that searching is almost useless (e.g., “methylmercury”, “table 10.1”)
• Use lots of jargon and acronyms without context, particularly on work sheets (e.g., “COPC” and “exposure point”, which is not a point but an area).

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

6. Eschew epidemiological opportunities.

There is a great opportunity to collect epidemiological data on long-term exposure to duck hunting at “Carson Lake and Pasture” and Stillwater Marsh. The Greenhead Hunting Club was founded in 1912 to control hunting at “Carson Lake and Pasture” (U.S. Bureau of Reclamation, 2009, p. 4-5). Most of the members’ duck hunting was likely done at “Carson Lake and Pasture”. The area was opened to the broader public in 1979 but the club probably still controlled hunting access. It may still exist. If membership is commonly passed down within families, there is the possibility of doing multi-generational studies of mercury consumption.

The Canvasback Gun Club owns land within Stillwater National Wildlife Refuge which encompasses “Dutch Bill Lake” and nearby areas (U.S. Fish and Wildlife Service, 2008). The club acquired the land before the refuge was established in 1991 and possibly before the Stillwater Wildlife Management Area was created in 1948. It would be odd if its members didn’t do most of their duck hunting on the club’s land. If membership is commonly passed down within families, there is the possibility of doing multi-generational studies of mercury consumption.

Comparisons of the fates and mercury concentrations of those hunting at the Greenhead Hunting Club and the Canvasback Gun Club could indicate whether there are significant, persistent differences in the mercury concentrations of ducks at “Carson Lake and Pasture” and those at Stillwater Marsh. Importantly, identifying and tracking symptoms of mercury poisoning would be much better measures of the real threats to human health than mercury concentrations in a sample of ducks.

There is no indication that the Environmental Protection Agency pursued either of these opportunities. Nor that the agency conducted surveys within Churchill County to identify local duck hunting habits.

7. Conclusion.

Thanks to the factor of 3 reduction in health risk due to assuming ducks have mercury only in the form of mercuric chloride, the factor of 6 underestimation of the consumption of duck by duck hunters, and to averaging out the high mercury concentrations in some species of ducks at “Carson Lake” in some years, the U.S. Environmental Protection Agency has concluded that consumption of ducks shot at “Carson Lake and Pasture” or Stillwater Marsh is not a mercury hazard. “EPA found that mercury levels did not pose elevated health risks (i.e., HQs were less than 1) for on-site adult and child residents, (see footnote c in Table 1, below) adult recreational users[,] and agricultural workers exposed to mercury in all OU2 Subareas” (U.S. Environmental Protection Agency, 2021, p. 7). Footnote c relates to rural residents that might be exposed to mercury contaminated dirt: “Residents of existing homes in the floodplain may be at risk in areas that have not been sampled for mercury.”

Plugging the 0.78 ppm (wet) estimated mercuric chloride concentration of ducks and the annual duck consumption of 690 g (1.5 pounds) from Table 4.25 of Attachment B-1a (CBandI Federal Services, LLC, 2017, pdgf page 475) into the intake equation, a person weighing 80 kg (176 pounds) could safely eat ducks with mercury concentrations up to 12.7 ppm (45 ppm dry weight basis). Only 1 duck at “Carson Lake and Pasture” in the data I have seen had more than that. Using a more realistic annual duck consumption of 4,200 g (9.3 pounds) and the reference dose for methylmercury, the maximum acceptable mercury concentration in ducks would be 0.70 ppm (2.52 ppm dry weight basis).

The Environmental Protection Agency could be wrong but no worries. Mercury poisoning through consuming mercury-contaminated fish or birds is probably not that bad. Symptoms of low-dose poisoning by methylmercury include paresthesia (tingling or pins-and-needles feeling), ataxia (poor coordination or balance problems), decreased muscle strength, impaired tactile sensation, poor vision or hearing, memory loss, and reduced learning ability (Bernhoft, 2012; U.S. Department of Health and Human Services, 2022, and references therein). Except for fetuses, it seems to be no worse than getting drunk and comes without the side effects of vomiting and elevated cancer risk.

Given the belittling of the mercury risks to human health, it is no surprise that risks to the ecosystem were also found to be negligible. “EPA concluded that site contaminants posed insignificant risks to wildlife. Studies of impacts on birds that eat fish in Lahontan Reservoir were inconclusive. Because significant ecological risks were not identified, EPA will not address them” (U.S. Environmental Protection Agency, 2021, p. 7).

How EPA Ducked the Mercury Risk of Ducks at “Carson Lake”- top

---

Next

back to The Problem at “Carson Lake and Pasture”