NorthMet Comments by Daniel Pauly

Comments Overview:
On the following pages I have outlined issues identified with regard to the NorthMet Project proposed by PolyMet Mining, Inc. These comments address following general topics: 1) The SDEIS assessment of Tailings Basin mercury contamination conditions is fraught with systematic data integrity problems that include mathematical errors in key formulas, improper selection of data sets, and suspected sample collection errors. The errors and omissions in the Tailings Basin dataset have permeated the SDEIS and its supporting reports, resulting in incorrect fundamental conclusions as to current and future mercury discharges at the Tailings Basin and Waste Water Treatment Plant. In particular, the Tailings Basin and Waste Water Treatment Plant as proposed are likely to significantly increase total mercury, and especially methylmercury, loading in the Embarrass River watershed. In view of the data integrity issues and improper mercury discharge conclusions, the NorthMet SDEIS should review previously discarded Tailings Basin alternatives.
2)
3)
A summary of these comments is provided on the following pages, with detailed individual comments following thereafter.
Comments Summary:
SECTION 1: Errors & Omissions in Tailings Basin Mercury Data and Analysis
The NorthMet SDEIS assessment of current and future Tailings Basin mercury seepage is fraught with problems. These problems include mathematical errors in key calculations, improper selection of data sets, exclusion of probative conflicting information, and suspected sample collection/analysis errors. These flaws have resulted in SDEIS drafters reaching the erroneous conclusion that Tailings Basin mercury seepage will meet Great Lakes Initiative standards of 1.3 ng/L. Of particular note are errors in the calculation for average water quality at the Tailings Basin. As described in detail in my comments, the preparers of the SDEIS failed to appreciate that some of their data was in the units ng/L, and some of it was in the units ug/L (a ug/L is 1000 times greater than a ng/L). This failure resulted in large errors in a key SDEIS summary of mercury contamination, which incorrectly states that the majority of water seepage sites were below the Great Lakes initiative mercury level, when in fact most of these sites were actually above the Great Lakes Initiative level. These errors and others are discussed on the following pages, and have been magnified by overreliance on improper subsets of data. Specifically, the SDEIS repeatedly considered only data that would show favorable mercury levels, while ignoring data to the contrary. Specific instances included the following:
NorthMet SDEIS Comment Daniel Pauly Page 1
1.
The Tailings Basin has four active NPDES monitoring locations that are positioned to intercept Tailings Basin seepage. NorthMet documents state that these monitoring locations provide the best data for determining Tailings Basin discharges. Inexplicably, the SDEIS relies almost exclusively only on data from one of these discharge locations the one that appears to show seepage below 1.3 ng/L. All three of the other discharge locations, including one that is better positioned relative to the likely flow of NorthMet discharges, were essentially ignored. The result is an improperly skewed assessment that mercury levels will be lower than 1.3 ng/L. Correction of this error shows mercury discharges above 1.3 ng/L. Looking beyond NPDES monitoring sites, a comprehensive review of mercury seepage data collected over time at the Tailings Basin also shows that Tailings Basin seepage will exceed Great Lakes Initiative mercury levels of 1.3 ng/L. Despite the availability of this information, it is essentially ignored in the SDEIS. The SDEIS also relies upon a faulty mercury sequestration test that predicts remarkably low mercury levels in Tailings Basin seepage, while ignoring a superior test that does not predict significant sequestration. Specifically, the SDEIS gives great weight to a very flawed 8 hour experiment with NorthMet tailings in a flask, while never even mentioning in the SDEIS a much more comprehensive test prepared for the NorthMet site that showed mercury levels are likely to be significantly above Great Lakes Initiative standards. The SDEIS includes an Assessment of Existing Pond Water and Groundwater Quality at the Tailings Basin that shows mercury seepage at nearly six times the Great Lakes standard for mercury discharges. This result is inexplicably not discussed further in the SDEIS. Mercury assessment errors lead to faulty WWTP design and pilot testing, as well as a defective containment system, both of which will likely lead to significant methylmercury discharges into the Embarrass River watershed
2.
3.
4.
SECTION 2:
A combination of errors and faulty analysis cause the SDEIS to erroneously conclude that Tailings Basin mercury discharges will be below 1.3 ng/L. These errors result in numerous misstatements and faulty conclusions in the SDEIS. Two of these deficiencies lead to notable flaws: 1) failure to design and test a Waste Water Treatment Plant (WWTP) that will adequately determine viability of mercury removal; and 2) design of a discharge capture system that incorporates a wetland that will receive the majority of the mercury and sulfate from the Tailings Basin, and which will meaningful increase mercury contamination in the Embarrass River watershed. One of the most critical long-term components of the entire NorthMet project is the performance of the Tailings Basin, and the most critical component of Tailings Basin performance is the ability of the WWTP to remove toxins and contaminants. The operation of this plant represents the biggest long term postclosure expense at the NorthMet site, and will likely run for at least hundreds of years. It is essential NorthMet SDEIS Comment Daniel Pauly Page 2
that decision makers have a reasonable basis for concluding that mercury can be removed by the proposed WWTP, as well as the long term costs for such removal. The WWTP will use reverse osmosis to remove contamination from Tailings Basin seepage. The SDEIS discusses a pilot test for the WWTP that was conducted in 2012. Unfortunately, the WWTP pilot test never evaluated mercury removal using reverse osmosis. In fact, as the Pilot Test report itself states, the WWTP designers dont know how much mercury can be removed by the WWTP. The report states that maybe as little as 20 percent, or maybe as much as 99 percent, of mercury might be removed. One guestimate is given by a salesperson for PolyMets membrane manufacturer of about 70 percent for mercury removal. Even then, according to the SDEIS documents, removal of methylmercury does not appear possible using the planned WWTP design. In other words, removal of the most hazardous species of mercury, the one that bio accumulates in fish and humans, has not even been considered in the WWTP design. The challenge now is that the NorthMet site has an inadequate mercury assessment that lead to an improperly designed pilot test, which results in tremendous uncertainty about whether PolyMet will be able to adequately remove mercury from Tailings Basin seepage and WWTP inflows. Even more uncertain is what that cost of implementing an adequate removal system is likely to be. The SDEIS seeks to address these concerns with a proposed adaptive engineering approach for the WWTP design and operation. An adaptive engineering approach is inconsistent with best practices in the literature for removal of contaminants using reverse osmosis, because each location and system has unique problems and challenges. Due to the relatively complex design of the NorthMet project water flows, including significant seasonal variations of likely methylmercury flows, the WWTP will need to remove disparate contaminants from quite variable influent streams. By approaching critical issues and long term challenges with an adaptive engineering approach, the NorthMet project risks decades of uncertainty, contaminant release violations, and unforeseen costs because the analysis will be undertaken long after the mine is in operation, and some of analysis wont be undertaken until after the mine is closed. This approach fails to provide decision makers with adequate information to assess what the treatment costs will be. For example, it is possible that mercury removal alone could add significant costs to WWTP as currently envisioned, because the various species of mercury are removed by different types of technology. Inexplicably, the WWTP and Pilot Test never even consider mercury removal as an issue, and so no site-specific costs for mercury removal can be calculated. The second problem that has been overlooked relates to changes in the NorthMet proposed Tailings Basin since the DEIS was prepared. Specifically, the currently proposed NorthMet Tailings Basin has been modified since the DEIS to create a containment system outside of the Tailings Basin. The containment system will include an up to 160 acre wetland that will be receiving the mercury and sulfate laden waters from the combined LTV/NorthMet tailings. These seepage flows will increase relative to the current LTV tailings basin, and create a much larger wetland than currently exists. The NorthMet projects own model data shows in great detail that most mercury and sulfate will be delivered directly to this wetland, and not to the containment system drains.
Contemporary research, including important research recently conducted in Minnesota, shows that this wetland, with its mixture of mercury, sulfate, and organic matter, is a prime environment for methylation of mercury. In fact, recent research by government investigators in Minnesota in the last few years has shown shockingly high levels of methylmercury in wetlands, including methylmercury spikes at the toe of taconite tailing basins. The SDEIS never asks what will happen to that methylmercury. It is likely that some of it is going to be carried to the WWTP, but as the WWTP pilot test itself reports, there is no plan for removing it. What happens to the rest of the methylmercury, which might be the majority? Current research shows it will be absorbed by everything from mosquitos to earthworms, and will then travel up the food chain throughout the Embarrass River watershed. SECTION 3 Tailing Basin Alternatives were Prematurely and Improperly Eliminated As detailed in the attached comments, the present plan for the NorthMet Tailings Basin is to reuse LTVs 60 year old unlined tailings basin that contains decades of heavy mercury loadings. The existing LTV tailings basin has been drying out since 2001, and seepage volumes have declined. As noted in the SDEIS with reference to the No Action Alternative, the LTV tailings basin should continue to dry out in coming years, with a concurrent drop in contaminated discharges. The NorthMet recommended action is to essentially build another unlined Tailings Basin on top of the current unlined and inactive tailing basin, but this time with tailings that are likely to be higher in sulfates than the existing mercury-rich tailings. Water from the new Tailings Basin will seep through the old taconite tailings, and then be discharged out the bottom into a large wetland area. In the wetland mercury methylation is likely to occur, and the methylmercury will bio-accumulate in the wetland life forms, before being carried out into the Embarrass River watershed by everything from bugs to birds. Some of the mercury, whether methylated or not, will travel to the NorthMet proposal Waste Water Treatment Plant. However, the SDEIS has no plan to remove that mercury, and the SDEIS documents indicate removal may not even be possible with the reverse osmosis technology planned for the waste water treatment plant. In view of the information available, two alternatives, at a minimum, should have been considered more thoroughly to see if they could have avoided the problems of perching a new unlined basin on top of an old unlined basin: either (a) putting a liner between the basins; or (b) locating the new basin somewhere other than on top of an old unlined basin. Both of these alternatives offer significant improvements in regards to preventing release of mercury from the existing LTV tailings basin, avoid the interaction of seepage water between two different types of tailings, and allow for what will likely be a significantly less complex and less expensive waste water treatment facility. These two proposals would require a much more significant up-front capital cost, but would reduce long term costs and accompanying uncertainty. To the extent long term costs are an issue to project approval, consideration of these options should be undertaken to remove that uncertainty. Overall, it seems quite possible that either of these two options could offer the mutual benefit of lower costs and improved environmental protection.
These two alternatives were dismissed very early in the EIS process. From my review of the DEIS and SDEIS, I believe these alternatives were prematurely dismissed because DEIS and SDEIS preparers overlooked the fact that that mercury contamination would be an issue. The thorough review of these (and other) alternatives is obviously beyond the scope of these comments. However, at the end of these comments I provide a simplified matrix showing how these two alternatives compare in effectiveness and cost relative to the proposals in the SDEIS. Further note that viable implementation of either of these options is really only available now, before the NorthMet project proceeds, because neither can be retrofitted onto the project once it has gone forward. In other words, it will be too late in 10 or 20 years, once the Tailings Basin has been filled, to find out that there might have been a better and cheaper way. Should the adaptive engineering approach to the WWTP prove unworkable for long term water treatment, it will be too late, or at least unfathomably expensive, to segregate the NorthMet tailings and LTV tailings. Conclusion In conclusion, the errors and omissions in mercury analysis has led to subsequent failures to address the capture and removal of mercury before it can escape into local and regional waters. The long term economically-viable removal of mercury has not been assessed, but should be. In addition, prematurely dismissed alternatives should be evaluated before proceeding further with the NorthMet project. Failure to adequately assess the likely level of mercury contamination, and the expense to prevent it, will potentially result in long term damage to the St. Louis River watershed and enormous unforeseen future costs.
PAULY COMMENT SECTION 1

The NorthMet SDEIS assessment of Tailings Basin mercury contamination is fraught with systematic problems that include mathematical errors in key calculations, improper selection of data sets, exclusion of probative conflicting information, and suspected sample collection or analysis errors. These shortcomings have resulted in reaching the erroneous conclusion that Tailings Basin seepage will meet Great Lakes Initiative standards of 1.3 ng/L.
1.
The SDEIS contains major computational errors in mercury seepage from the Tailings Basin. Mercury is one of the most significant potential water contaminants from the NorthMet Project. Contamination of local surface and groundwater from the Tailings Basin is of particular concern. As noted in the SDEIS, the proposed NorthMet Tailings Basin will have significant levels of water seepage indefinitely, and the volume of discharge water will be increasing significantly compared to current conditions. The level of mercury in that seepage water is critical to evaluating the environmental impacts of the NorthMet project, as well as the financial viability of seeking to remove the mercury. Table 4.2.2-4 of the SDEIS, reproduced in relevant portion below, provides a summary of total mercury in the Embarrass River watershed near the Tailings Basin. In particular, Table 4.2.2-4 provides mean mercury levels for water seepage at 11 locations at the Tailings Basin. The accuracy of these numbers is paramount to understanding the potential for mercury contamination from the Tailings Basin.
According to Table 4.2.2-4, many of the most heavily sampled locations in the Tailings Basin have mercury levels are significantly below the Great Lakes Initiative. In fact, Table 4.2.2-4 shows five locations, highlighted in yellow for this comment, which have mercury levels below 1.0 ng/L. Table 4.2.2-4 was prepared using information in SDEIS appendix Barr 2006f, which I received from Ms. Lisa Fay of the Minnesota Department of Natural Resources. Barr 2006f allows for reconstruction of the mercury contamination calculations for each of these sample locations. In reviewing the data in Barr 2006f, I identified a fundamental mistake in the summary results: The calculations had a major error because the sample data was presented in two units: nanograms per liter (ng/L) and micrograms per liter (ug/l), yet the calculations ignored these different units. Reproduced below is a portion of page 113 of Barr 2006f, showing the table heading for columns of mercury discharge data for Cell 1E:
Data in the left hand column is presented in ng/L, while data in the right hand column is presented as ug/L. For reference, the data on the right can be expressed as <200 ng/L. Unfortunately, the SDEIS preparers interpreted this data as <.2 ng/L. What was being expressed in the right hand column is that the minimum detection level (mdl) for these samples was less than 200 ng/L. In other words, the analysis did not seek to measure low level mercury concentrations. The problem is that the SDEIS preparers thought that the data indicated a minimum detection level of 0.2 ng/L. In other words, the SDEIS preparers incorrectly thought this data indicated mercury levels were spectacularly low. The manner in which this error impacted the SDEIS results can be illustrated by looking at Cell 1E, which is reported in Table 4.2.2-4 to have a mean mercury level of 0.2 ng/L from a total of 25 samples. Specially, Cell 1E is reported in Table 4.2.2-4 of the SDEIS to have been sampled 25 times between 2001 and 2003, with merely 3 of these samples even having detectable levels of mercury, and none of those samples exceeded Great Lakes Standards. Barr 2006f shows exactly what those levels were for those 25 samples. For ease of review, the relevant portions of page 113 are reproduced below:
Pauly Comment Figure 1 Excerpt from Barr 2006F
Consistent with Table 4.2.2-4, this excerpt for Cell 1E shows 25 different sample results. In the left column three of the samples had mercury levels between 0.9 and 1 ng/L. In the right hand column the other 22 samples are shown. These right-hand samples are all assigned the value of <0.2, which likely meant that the sample was analyzed with a minimum detection level below 0.2. However, these samples on the right are expressed in ug/L, not in ng/L. In other words, the right hand side of the table indicates that the mercury detection levels for these samples
was 200 ng/L.1 The individual who calculated the data for SDEIS Table 4.2.2-4, and all the subsequent proofers and reviewers, never noticed that the units on Barr 2006f are not consistent from one column to the next. They incorrectly concluded that all of these data points are in the same units. Specifically, with regard to Cell 1E, Table 4.2.2-4 reports total mercury at the Cell 1E location is 0.2 ng/L. This number appears to have been reached by considering all of the numbers in the Barr 2006F report to be in ng/L, including those that are in ug/L. The preparer of the SDEIS table used half of each minimum detection limit (see footnote 4 to Table 4.2.2-4 of the SDEIS), but erroneously assumed the units should be nanograms. The preparer then averaged the readings with the three readings from the left column. In other words, it appears the preparer assumed that all the readings of <0.2 ug/L were actually <0.2 ng/L, converted those numbers to 0.1 ng/L per footnote 4 to Table 4.2.2-4, and calculated a mean as follows: ((0.9 + 0.9 + 1) + (0.1 x 22)) / 25 = 0.2 ng/L This formula produces a result that matches exactly with SDEIS Table 4.2.2-4, and I think is likely to be the erroneous formula used by the SDEIS preparer. This error should have been caught much earlier, because the possibility that 20 consecutive water samples from the Tailing Basin would all be beneath a .2 ng/L detection level is implausible, especially just after the LTV plant closed. Such results would be just a few percent of atmospheric levels, and a small fraction of other average mercury levels at sample locations in the Tailings Basin. It is impossible to know for certain what the correct mercury discharge number at Cell 1E was from the information presented, especially with almost all samples analyzed with very high minimum detection limits, but it is certainly higher than 0.2 ng/L. Cell 2E has the same errors as Cell 1E. In this case, the reported result was 0.35 ng/L as the calculated total mercury level. From the three detections that were actually measured, the average mercury level should be 1.8 ng/L, which is nearly 40 percent over the Great Lakes Initiative level and 500 percent higher than the levels erroneously reported in the SDEIS. Cell 2W states the most implausible Tailings Basin mercury data in the SDEIS: a purported <0.1 ng/L average from eight samples. What is most implausible about these results is that they come from 2001, which is the year the Tailings Basin discontinued operation as a basin for LTV taconite processing. The fact that such anomalous results would remain in a key dataset in the SDEIS after years of effort and review indicates a need for a comprehensive dataset review, as noted elsewhere in these comments.
1
That is a high minimum detection level, but could have been in use in 2001-2003, although 20 ng/L or 2 ng/L would have been more common detection levels.
The Emergency Basin mercury concentration calculations also have errors similar to Cell 1E. The correct level, based upon the available data, is probably closer to 1.8 ng/L, which again is nearly 40 percent over the Great Lakes Initiative level. The West Seep mercury concentration calculations also has errors similar to Cell 1E, and understates mercury concentration by a similarly large magnitude. 1.1 Another error in Table 4.2.2-4 is that most samples were improperly reported as nondetects for mercury.
Not only is the mean total mercury concentration incorrect in Table 4.2.2-4, but it is important to note that SDEIS Table 4.2.2-4 also dramatically overestimates the number of non-detect samples for mercury. The error is that the minimum detection levels were far above the actual Great Lakes Initiative levels, so a non-detect in many cases means only that the mercury levels were below the very high 200 ng/L, rather than a modern detection level of 0.5 ng/L. Nondetects for such high minimum detection levels should not be considered to be non-detects when the regulated target level for mercury is 1.3 ng/L. The problem is that someone reviewing Table 4.2.2-4 would mistakenly conclude that the number of samples with detectable levels of mercury was very, very low. This presentation of the data is profoundly misleading, because in fact almost all of those non-detects came from analysis that had very high minimum detection levels. A person reviewing Table 4.2.2-4 is likely to be misinformed about the data, and reach the erroneous conclusion that almost all of the samples at the ten sample locations were below a modern detection level of 0.5 ng/L. 1.2 Correcting these errors in the SDEIS mercury discharge calculations leads to a radically different view of likely NorthMet Tailings Basin mercury discharges
Once Table 4.2.2-4 is updated to correct these calculation errors, a radically different assessment of likely mercury discharges from the NorthMet project becomes evident. Rather than being below the Great Lakes Initiative standard of 1.3 ng/L, at 6 out of 11 seepage sites, probably 10 out of 11 seepage sites are above the Great Lakes Initiative standard. In the following pages I provide additional analysis that confirms this conclusion and make recommendations as to how to address these higher than assumed mercury discharge levels.
2.
The NorthMet Tailings Basin assessment improperly underestimates mercury discharges by disregarding data from three out of four of the NPDES monitored surface discharge locations at the Tailings Basin. In addition to the erroneous calculation with regard to water quality at numerous Tailings Basin locations, the SDEIS has improperly selected a subset of NPDES water sampling locations that has also resulted in a meaningfully inaccurate assessment of current conditions at the Tailings Basin. The Tailings Basin has four active seeps that have been monitored since the closing of the LTV plant site. In the PolyMet water quality analysis, which is an appendix to the SDEIS, PolyMets water quality scientists state that NPDES surface discharges from the Tailings Basin serve as the best proxy for concentrations of mercury seepage from the existing Tailings Basin. PolyMet 2013j at page 271. SD006 is in the southeast corner of the Tailings Basin, SD001, SD004, and SD005 are in the northwest corner of the Tailings Basin; and SD002 is in the north central portion of the Tailings Basin. There was apparently never an SD003, so there is no data. Also, SD005 has essentially been dry and not sampled for most of the last 13 years. That leaves four NPDES sample locations to review: SD001, SD002, SD004, and SD006 consistent with the opinion of PolyMets water model that these discharges are the best proxy for concentrations of mercury seepage from the existing Tailings Basin. Despite the fact that this seepage data is so critical to assessing mercury contamination, only Tailing Basin data from SD004 is used to support the determination that mercury levels will be below Great Lakes levels. SD0026, which is at the plant site and not the Tailings Basin, is also used, but it gives no indication of likely Tailings Basin discharges. In fact, it is not located in the Embarrass River watershed like the Tailings Basin. In order to evaluate whether this single sampling point should be the primary empirical assessment of groundwater conditions, I obtained NPDES sampling data from the Minnesota Pollution Control Agency for all four of the Tailings Basin NPDES surface discharges. I then evaluated total mercury in the Tailings Basin for each of these sample locations, but with more up to date information through 2006 to be consistent with most SDEIS data sets. Based upon this MPCA data, I calculated total mercury levels for each of the four NPDES locations, plus average of all samples regardless of which location was sampled (basically a weighted average), and further an average of the four sites (basically a weighted average). My spreadsheets with calculations are available upon request, but should be easily reproducible without them because the data is readily available from the MPCA. My results are provided graphically below, and are also summarized in the accompanying table. The graph indicates that rather than SD004 being a representative sample of conditions at the NorthMet Tailings basin, SD004 is an outlier the one location that isnt above the Great Lakes standard of 1.3 ng/L. All three of the other locations were above Great Lakes standards, as were averages weighted by sample number and location.
The following table1 shows a summary of the data used for the calculations:
Location
Data Used
Mean Mercury Level Ng/L2 1.8 1.4 1.2 1.6
SD001 SD002 SD0043 SD006
2001-2006 2001-2006 2001-2006 2001-2006
Included in primary SDEIS data sets for Mercury contamination? No No Yes No
In view of the importance of this NPDES data on assessing mercury contamination, it is critical to look at this data as thoroughly as possible. The selection of only one of four sample locations is not appropriate, especially when the selected sample location is the only one that shows mercury levels below the target level, while the other three sample locations are above the target level. Recommendation: The mercury modelling for the Tailings Basin mercury data should be recalculated to include data from all four surface discharge locations at the Tailings Basin site.
1 2
All data is available from the MPCA website, or can be received from comment author upon request. Mercury levels were calculated consistent with SDEIS protocols. Namely, any test for which minimum detection levels were above the Great Lakes standard of 1.3 ng/L were removed. Any test for which a non-detect was identified were calculated as 50 percent of the minimum detection level if the minimum detection level was less than 1.3 ng/L. This approach concurs with the methodologies used in the SDEIS tables. 3 SD003 does not appear to exist according to available records.
3.
A comprehensive review of Tailings Basin surface discharges confirms that mercury discharges are likely to be above Great Lakes standards, and also confirms that SD004 is not a representative of Tailings Basin seepage. Comment 2 explained how review of data from all four NPDES monitoring locations at the Tailings Basin indicates that seepage from the Tailings Basin is above Great Lakes Initiative standards. Other data is also available in the SDEIS to allow for an analysis of mercury at the Tailings Basin. This comment looks at whether a review of additional available surface water data will provide a different result. In particular, PolyMets own reports prepared for the original DEIS provide additional characterization of mercury at the Tailings Basin site. These locations, detailed below, were sampled for water quality, including mercury. Locations with discharges below Great Lakes Initiative mercury levels are shown in green. Locations with discharges above Great Lakes Initiative mercury levels are shown in red.
As can be seen in in this graph and the table below, seven out of eight of the sites/averages for NorthMet Tailings Basin seepage exceeded the Great Lakes mercury standard of 1.3 ng/L.
Comprehensive Tailings Basin Water Sampling Date Mean Mercury Level (ng/L) 2.445 2.4 3.1 2.4 3.4 Tailings Basin Operation Conditions Operation Closure Closure Closure Closure. Note, this location is actually quite far from the site. See Figure 4.2.2-15 of the SDEIS Closure Included in the SDEIS assessment of mercury levels? No No No No No
Sample Source LTV-DNR4 PM-86 PM-9 PM-10 PM-11
Sample Date 1997 2004, 20067 2004, 20068 2004, 20069 2004, 200610
SD00411 Composite
2005-08 1997- 2008
1.2 2.5
Yes
No matter how the samples are interpreted, it is clear that the SDEIS has a profound shortcoming in relying almost exclusively on data from SD004 as a supposedly representative sampling location. SD004 is not representative of water discharges at the Tailings Basin, but rather is an anomaly. The unfortunate decision by SDEIS preparers to select SD004 as foundation of the Tailings Basin mercury analysis results in flawed arguments, erroneous conclusions, and unexamined Project alternatives.
4
In researching the literature, the only known measurement of actual seep water mercury levels from the LTV Tailings Basin is from Berndt et. al 2003, which reported a 1997 test showing 2.44 ng/L of at least one sample. This measure is a potentially significant, because most pre-2001 total mercury tests appear to have been made with higher minimum detection levels, and are thus of no value to characterizing the Tailings Basin during operation. 5 Note that at the time of the LTV measurements in 1997, the Basin mercury levels were 3.48 ng/L, and the seep was 2.44 ng/L. These relative differences indicate that the tailings may lower the mercury level, but perhaps to an equilibrium level about twice the Great Lakes Initiative. 6 Note that PM-7, PM-8, PM-9 and PM-10 are all identified in PolyMets 2007 Summary and Interpretation of Surface Water Quality Monitoring Data, PolyMet Mining Company NPDES monitoring locations. Barr (2007i), at page 7. Based upon Figure 1 of that report, which provided a map of all monitoring stations, it appears that PM8, PM9, and PM10 correspond to NPDES sites SD0001, SD002, and SD006. To avoid potential double-counting of the same site in this single table, I am not including SD001, SD002, or SD006 in this table. 7 Barr 2006i, table B3 8 Barr 2006i, Table B3 9 Barr 2006i, Table B3 10 Barr 2006i, Table B3 11 For cross-reference purposes, SD004 appears to correspond to Seep 22 of various NorthMet reports, including Figure 4.2.2-11 of the SDEIS.
4.
Different conclusions on mercury concentrations at SD004 in the DEIS and SDEIS reveal the problem of over reliance on a narrow subset of discharge data: The over-reliance on just a small set of data points, especially from just one or two locations at the Tailings Basin, is problematic because a change in just a couple of samples can completely change the results, and lead to completely different actions. A good example of this statistical phenomena can be demonstrated by looking at how NPDES surface discharge SD004 lead to completely different conclusions between the DEIS and the SDEIS. Reproduced below are Table 4.1.31 from the original DEIS, and Table 4.2.2-4 from the SDEIS. These two tables show the same datasets for 11 sample locations at the Tailings Basin. Note that almost every location has very similar results between the DEIS and SDEIS (with the exception of some errors in the SDEIS, mentioned earlier). The one exception to this consistency is SD004, where the mean has gone from 1.9 ng/L from 15 samples in the DEIS, to a mean of 1.2 from 14 samples in the SDEIS. I highlighted in red the DEIS data for SD004, which is above the Great Lakes Initiative level; and highlighted in green the SDEIS data for SD004, which is below the Great Lakes Initiative level.
How can the averages be so radically different? It appears, from the limited data I was able to independently obtain from the Minnesota Pollution Control Agency, that one or two of the samples in the DEIS was considered to be erroneous, and corrected in the SDEIS. Im not in a position to say whether the DEIS data or the SDEIS data for surface discharge SD004 is most accurate, but I do point out how the SDEIS mercury data needs to be broadened out to minimize reliance on just one or two samples. Otherwise it is impossible to get a representative sample and make scientifically and statistically sound decisions. Therefore, in this comment I reiterate my strong recommendation to discontinue over reliance upon data from SD004.
5.
The SDEIS includes an Assessment of Existing Pond Water and Groundwater Quality at the Tailings Basin that shows mercury seepage at nearly six times the Great Lakes standard for mercury discharges, yet this result is inexplicably not recognized or discussed further. As noted elsewhere in these comments, the SDEIS reaches incorrect conclusions about mercury seepage from the Tailings Basin. Generally those incorrect conclusions can be traced to improper cherry picking of sample locations, as well as errors in calculations. There is one place in the SDEIS where a comprehensive summary of current data has been assembled to show the effect [of] passage through the existing LTVSMC tailings has on seepage water quality. This information is compiled into Table 4.2.2-23, and is found at page 4-110 of the SDEIS. I have replicated it with annotations on the following page of these comments. Notably, unlike some of the datasets used in the SDEIS, Table 4.2.2-23 contains some of the most current data available with regard to Tailings Basin discharges. As explained at page 4-111 of the SDEIS, the assessment summarized on Table 4.2.2-23 is useful to show the effect of passage of water seeping through the Tailings Basin. The first paragraph of Page 4-111 states as follows: Comparing existing pond water quality with water quality at the toe of the Tailings Basin helps define the effect passage through the existing LTVSMC tailings has on seepage water quality. Based on the parameters that were monitored in the Cell 2E pond, it appears that passage through the LTVSMC tailings reduces the average concentrations of arsenic, fluoride, and molybdenum, although it is difficult to determine to what extent these reductions are simply attributable to the effects of dilution. The concentrations of several other parameters, such as calcium, manganese, nickel, and TDS, increase as they seep from the tailings pond to the toe of the Tailings Basin. The preparers of the SDEIS commented on how calcium, TDS, and manganese all increased when passing through the Tailings Basin; and how fluoride, arsenic, and molybdenum go down. However, there is no mention or analysis whatsoever of the most significant data point: mercury has gone from 1.4 ng/L in pond water to 6.4 ng/L after passing through the Tailings Basin.2 I have highlighted these respective contaminants in yellow, green and red. These results should considered in the SDEIS, because they directly conflict with the conclusions that mercury is not a contaminant of concern at the NorthMet Tailings Basin.
Note that the underlying data shows some unusually high readings for mercury that may reflect sampling or computational errors. However, even when removing potentially erroneous sample results, the mercury levels are still above Great Lakes Initiative Standards at numerous locations and over numerous periods.
Supplemental Draft Environmental Impact Statement (SDEIS) NorthMet Mining Project and Land Exchange
Table 4.2.2-23 Existing Pond Water and Groundwater Quality at the Tailings Basin
Pond Water Quality (Cell 2E) Toe of Tailings Basin (GW-001,GW-006, GW-007, GW-008, GW-012 Surficial Aquifer) Groundwater Evaluation Criteria Detection Mean1 Range -62 of 62 83 21 to 211 250 61 of 61 18 1 to 30 2 47 of 61 1 <0.05 to 3 6.5 8.5 58 of 58 7 6.0 to 8.0 250 61 of 61 228 15 to 556 500 42 of 42 793 151 to 1,550 <10 to 29,000 <0.25 <0.25 to 7 <5 to 452 <0.1 to 1 <25 to 554 <0.1 to 2 <0.5 to 68 <0.1 to 18 <0.35 to 205 <25 to 31,000 <0.25 to 8 12 to 4,130 <0.25 to 153 <0.03 to 0.28 <0.1 to 47 <0.25 to 91 <0.5 to 5 <0.2 to 0.23 <0.1 to 1 <3 to 95 <5 to 40 <0.25 to 7 <25 to 531 <0.1 to 1 <0.5 to 3 <0.35 to 11 9 to 3670 <0.25 to 12 <1.0 <0.1 <3 to 51
Constituent
Units
General Parameters Mean Calcium mg/L 30 Chloride mg/L 23 Fluoride mg/L 5.2 pH s.u. 8.4 Sulfate mg/L 109 TDS mg/L 381 Metals Total Aluminum g/L -200 42 of 62 1,994 Antimony g/L -6 0 of 59 0 Arsenic g/L 5.0 10 30 of 59 2.0 Barium g/L -2,000 61 of 62 136 Beryllium g/L -0.08 4 of 59 0 Boron g/L 278 1,000 50 of 62 318 Cadmium g/L -4 10 of 59 0 Chromium g/L -100 28 of 59 6 Cobalt g/L 1.0 -54 of 59 2 Copper g/L 2.0 1,000 58 of 59 10 Iron g/L -300 55 of 62 5,259 Lead g/L --17 of 59 1 Manganese g/L 100 50 62 of 62 1,327 Mercury ng/L 1.4 2,000 39 of 51 6.40 Mercury, Methyl ng/L --6 of 50 0.06 Molybdenum g/L 113 -56 of 59 20 Nickel g/L 2.1 100 55 of 59 9 Selenium g/L -30 3 of 59 <1 Silver g/L 30 of 59 0 This table is -discussed at page 4-111 of2the SDEIS, and gives Thallium g/L -0.6 3 of 59 0 indications of passage of water through the LTV Zinc g/L -- the effect 2,000 17 of 59 12 tailings has on seepage water quality. The three parameters Dissolved/Filtered Metals Aluminum g/Lare highlighted -200 5 of that green are discussed at59 page 4-111 13 as Arsenic g/L -10 17 of 42 1 examples of declining levels; the three parameters highBoron g/L -1,000 21 of 27 300 lighted in yellow are discussed as examples of increasing Cadmium g/L -4 4 of 59 0 levels. in59 red is mercury Chromium g/L The one -- parameter 100 highlighted 10 of 1 contamination, which shows fold Copper g/L -1,000 a nearly 565 of 59 increase 2in Manganese g/L -- from 1.450 of 43 contamination to 6.4 ng/L. 43 This issue was 1,142 Nickel g/L -- not even discussed 100 51 of 59 3 inexplicably in the SDEIS, despite Selenium g/L -30 0 of 59 1 showing tailing water discharges of 5 times the Great Silver g/L -30 0 of 59 0 Lakes of 1.3 ng/L. is the biggest waterZinc g/L standard -2,000 Mercury 25 of 59 8
Sources: Barr 2013b; Barr 2006f.
1
Where non-detects occur, the mean was calculated using half the detection limit.
borne health threat in the St. Louis River watershed, and these results should have been evaluated in the SDEIS.
4.2.2 WATER RESOURCES
4-110
NOVEMBER 2013
6.
The SDEIS relies upon a faulty mercury sequestration test that predicts remarkably low mercury levels in Tailings Basin seepage, while ignoring a superior test that does not predict significant sequestration. At numerous points in the SDEIS and supporting documents, reference is made to a bench study conducted by NTS in 2006 that purports to show [T]he concentration of dissolved mercury in a treatment flask containing process water and NorthMet tailings decreased from 3.3 ng/L to 0.9 ng/L over an eight hour period. SDEIS at page 5-206, As discussed below, this test had obvious flaws and glaring mischaracterization of the collected data. In contrast, the SDEIS entirely fails to even mention that a second carefully designed mercury sequestration test was performed. This test concluded that there was no observable diminishment when mercury-laden water passed through the simulated LTV/NorthMet tailings combination. If anything, the test showed a slight increase in mercury levels. This thorough study, despite having tremendous potential to quantify likely mercury discharges from the combined Tailings Basin, is never so much as mentioned at any point in the SDEIS itself. This is a situation where a very low-quality test with favorable results was prominently presented in the SDEIS, while a high-quality test with unfavorable results was not even mentioned. In view of the fact that mercury is the most serious contaminant in the St. Louis River watershed to public health, the basis of this important research warrants further examination. The tests are also important because the level at which the Tailings Basin discharges mercury affects numerous other conclusions and design choices, in particular a large number of alternatives that should be considered if mercury will be leaching from the NorthMet Tailings at levels above 1.3 ng/L. Details of the tests and their results are provided below. 6.1 The SDEIS relies on an 8 hour mercury sequestration test in a single flask to argue that NorthMet Tailings will retain mercury for hundreds of years, yet this test was so poorly designed, executed, and analyzed that it risks becoming an opportunity to attack the credibility of the entire SDEIS preparation process.
The SDEIS and supporting documents assert that researchers had undertaken experiments that showed the NorthMet Tailings would capture mercury and prevent it from discharging from the Tailings Basin. This test is cited as one of the grounds for being able to conclude that mercury contamination would be below the Great Lakes Initiative standard. See page 5-206 of the SDEIS. The test report was not available on the SDEIS reference disk, but I was able to receive a copy of it from Ms. Lisa Fay at the Minnesota Department of Natural resources. As detailed below, the test was undertaken for just eight hours and apparently with just one sample. Amazingly, no effort was undertaken to try to replicate conditions in the Tailings Basin, and even the flawed results were mischaracterized to state conclusions that are clearly refuted by the test data. NorthMet SDEIS Comment Daniel Pauly Page 21
Briefly, here is a summary of the test and results: Two shake flasks (Jug C and Jug D) were filled with water that was spiked with mercury to obtain about 3.5 ng/L of mercury in the water. One flask had PolyMet tailings (Jug C) in addition to the mercury-spiked water, and was meant to test mercury absorption by tailings. The other flask just had the water (Jug D), and served as the control. The flasks were shaken for 8 hours, and samples removed and analyzed over time. The results are reproduced below:
From these results, NTS concluded that The results showed that mercury removal by adsorption to PolyMet tailings occurs rapidly and remains stable throughout the conditions of this experiment. NTS 2006 Report, page B-2, emphasis added. What is most peculiar to me about these test results is the conclusion that mercury sequestration remains stable. In fact, over the last 4 hours of the test, the mercury levels nearly doubled, and are on a trajectory to exceed the GLI Standard in approximately 4 more hours. Besides this obvious misstatement of the test results, a few other flaws are obvious just in the way the test was conducted: First, it looks like just one sample was tested for each of the control and the tailings. Second, there was no effort to replicate in situ conditions, such as redox conditions, oxygen levels, biological factors, etc. Third is the obvious shortcoming of using an 8 hour test on a centuries long problem. Another strange aspect of the results is that it NorthMet SDEIS Comment Daniel Pauly Page 22
reports testing was done on Jug C and Jug D. What might be shown in Jugs A and B? Were they analyzed? If so, what does their data show? 6.2 A much better test did exist, but the results are not even mentioned in the SDEIS. Those tests show that the PolyMet Tailings combined with NorthMet tailings provide no net mercury sequestration.
In July, 2007, some four months after the short NTS flask test mentioned above, PolyMet Mining Inc. did receive a report on the Tailings Basins ability to sequester mercury. This test, conducted over many months, sought to investigate potential chemical processes that can be expected between the tailings basins. SRK 2007c at page 15. The consultants who prepared the test worked with the Minnesota DNR3 to provide a column design reproduced below:
This design and experiment had many major improvements over the earlier NTS report: it sought to replicate travel through both the NorthMet tailings and the existing LTV tailings, it sought to replicate subterranean conditions, such as oxygen levels, and it ran for much longer (at least a year). This test concluded that [t]here were no clear increasing or decreasing mercury concentration trends along the flow path through the LTVSMC tailings. SRK 2007c at page 31. It is peculiar that this test was not disclosed in the SDEIS, while an 8 hour test under much less relevant conditions was included, and used as support for the conclusion that mercury discharges will be below the Great Lakes Initiative level. 6.3 The SRK test that showed no mercury sequestration should have been included in the SDEIS, because it is on its face far more probative than the NTS eight hour test.
SRK Consulting Report at Page 16.
The NorthMet SDEIS asserts that one of the fundamental long-term mechanisms for mercury control at the NorthMet Tailings Basin is removal of mercury by the pre-existing LTV taconite tailings. This mechanism is relied upon repeatedly in the SDEIS itself, as well as the supporting documents. This mechanism was also proposed in the original DEIS, but was objected to as lacking scientific integrity by the EPA, which requested that further analysis be provided to support this mercury removal mechanism. See page 11 of the EPA Detailed Comments to the NorthMet Project DEIS. Despite the EPAs express request for further support of this theory, the SDEIS fails to provide further support, and merely restates the previously challenged analysis. To correct this deficiency the SDEIS should address why preparers believe the 8 hour shake test is a better predictor of long term tailings basin mercury sequestration than the year-long column testing.
7.
The site-specific sampling of mercury at the Tailings Basin is consistent with prior taconite tailings seepage research At various points in the SDEIS4, reference is made to prior research by Minnesota Department of Natural Resources scientist Michael Berndt that looked at mercury releases from mining operations in Minnesota, including from stack emissions and tailings basin seepage. The SDEIS relies especially on a 2003 paper by Mr. Berndt entitled Mercury and Mining in Minnesota, submitted as a final report to the Minerals Coordinating Committee in June of 2003, and revised in October of 2013. SDEIS reference Berndt 2003. As discussed below, the interpretation that NorthMet Tailings Basin seepage will be above Great Lakes Initiative standards is consistent with the data and conclusions of the 2003 report. 7.1 Berndt 2003 included data from the existing LTV tailings basin, and the seepage had a total mercury concentration of 2.44 ng/L
Table 4 of Berndt 2003 is reproduced below. The total mercury level from the LTV tailings basin is highlighted in yellow, and is 2.44 ng/L. This measure of LTV tailings basin seepage was collected from an earlier MNDNR study5, and represents perhaps the only mercury measure made with a low detection level during operation of the LTV tailings basin.
4 5
See, in particular, SDEIS page 5-205 See page 48 of Berndt 2003.
This level of mercury contamination in LTV tailings seepage is consistent with some of the other data that is available. See, e.g., the chart entitled Tailings Basin Surface Discharges with regard to Comment 3 above.6 7.2 The composite SDEIS data for the Tailings Basin actually aligns closely to the finding of Berndt.
Comment 3, above, looked at a comprehensive set of mercury data for the Tailings Basin seepage. The results showed a composite average of 2.5 ng/L, which is very close to the LTV measure reported in Berndt of 2.44 ng/L at the LTV site. 7.3 The observed Tailings Basin mercury seepage is within the range expected by Berndt 2013, even though it is above the Great Lakes Initiative standard.
Table 4 of Berndt 2013 shows seeps and monitoring wells at various Minnesota taconite facilities ranged from .72 to 2.9 ng/L of total mercury. The composite measure of about 2.5 ng/L, reported above in Comment 3, fits within this observed range from Berndt. The levels are above Great Lakes Initiative standards, but still within ranges one would foresee from Berndt. 7.4 Even if the NorthMet Tailings Basin reduces the concentration of mercury in the seepage, that seepage can still exceed Great Lakes Initiative standards, and would still be in violation of the Great Lakes Initiative levels.
There is discussion in the SDEIS that the Tailings Basin mercury discharges will not be an issue because the discharge concentrations will still be less than the levels in either precipitation or the tailings basin pond water. This statement seems to apply that reduction in mercury levels in the Tailings Basin is sufficient, as opposed to actually meeting the Great Lakes Initiative requirements. Berndt 2003 accurately states that: It is important to note that just because an industry discharges water with a concentration that is less than that of the water it takes in does not mean it will meet water quality standard Berndt 2003 at page 13. In the present case that is exactly what seems likely to happen with the NorthMet Tailings Basin: water may come out of the Tailings Basin at a mercury level lower than it went in, but will still violate the Great Lakes Initiative.
6
Although this measure used in Berndt 2003 was for only one data point, it should be noted that Berndt 2003 used only a small number of samples for each measured site (see Table 4 from Berndt 2003, reproduced above).
7.5
Even if the Tailings Basin mercury concentrations are reduced relative to influent, the increased flow of water through the Tailings Basin will significantly increase total mercury discharges.
The NorthMet Tailings Basin is expected to have at least twice as much seepage water discharged during operations than is currently seeping from the Tailings Basin. Even if one assumes the concentration of mercury in the Tailings Basin seepage will not change, the total mass of mercury discharged will be significantly increased, likely by double or more. 7.6 Efforts to reduce atmospheric precipitation, especially from coal burning sources, will likely diminish background mercury loading dramatically in coming years, leaving the NorthMet Tailings Basin as a larger relative source of mercury to the Embarrass River and St. Louis River watersheds.
At the present time significant mercury loading into the Embarrass River and St. Louis River watersheds occurs as a result of atmospheric deposition, much of it from coal plants. In fact, an estimated 70 percent of atmospheric mercury deposition in Minnesota comes from anthropogenic sources, including burning coal and improper disposal of fluorescent lights. Efforts are underway in Minnesota, the broader United States, and outside the U.S. to reduce mercury emissions. Reductions in coal burning will likely lead to significant reductions even in the near term as increasing amounts of electricity generation switches from coal to natural gas. It is likely that atmospheric mercury deposition reductions will continue for many decades, and the amount of atmospheric deposition of mercury in the Embarrass River will decline. What may not decline, due to the tremendous amounts of mercury now held within the existing tailings basin at the NorthMet site, is the amount of mercury discharged from the Tailings Basin in coming centuries. Without adequate efforts to prevent mercury seepage and release at the Tailings Basin, the likely result will be a circumstance where in coming decades and centuries the NorthMet project becomes a larger and larger share of mercury contamination in the Embarrass River and St. Louis River watersheds.
8.
The Water Treatment Pilot Test data shows profound sampling irregularities that raise serious concerns about the adequacy of the protocols implemented and the accuracy of the test results obtained. In view of the relatively short timeframe available to review thousands of pages of documents, I have not spent extensive time reviewing sulfate levels at the NorthMet site, or sulfate removal from the WWTP and WWTP Pilot Test. However, I would be remiss in not pointing out that there appears to have been a very serious, repeated error in the collection and analysis of sulfate data from the WWTP Pilot Test. The evidence is quite strong that a serious sampling or analytical breakdown occurred during the Pilot Test as it relates to sulfate levels in the test influent, and perhaps other contaminants of concern. Figures 5 and 6 from the Pilot Test Report are reproduced below, and reveal the reasons for serious concern:
First, in looking at Figure 5, notice that surface discharge SD004 has relatively constant levels for the three measured parameters: Sulfate, Total Hardness, and TDS. In contrast, a new well installed in 2011 for the pilot test showed remarkable variation in sulfate levels, fluctuating almost exactly by 400 percent either at about 100 mg/l, or 400 mg/l. There is essentially nothing in between. TDS and Total Hardness fluctuate on the same days, but at different ratios that are closer to 200 percent. Such readings are inexplicable from a groundwater well.
Second, looking at Figure 6, the issue becomes becomes even more concerning, because on the same exact dates that sulfate, total hardness and TDS were changing, the iron and manganese concentrations remained all but constant from the same sources.
This combination of consistency in some parameters, but extreme volatility in other parameters, all from the same source and the same day, indicates likelihood that serious errors have been introduced into the data. There is no evident reason why such variation would occur naturally. If these variations were naturally occurring in the well water, they would expected to be more random in nature. Also, if groundwater dilution were to be a factor, such as from rainfall, then the dilution would be expected to be somewhat uniform across the different parameters. In other words, one would expect sulfate, iron, manganese and the other materials to have been uniformly diluted, other than perhaps small changes for solubility. It seems very possible that the errors occurred after sample collection, maybe during analysis of the samples, when possible errors in dilution or peak calculations occurred on different samples destined for different tests. The overwhelmingly strong indication of errors in the samples from these sources is of course very concerning. Due to the presentation of the data in a graph, it is possible to spot the errors. What concerns me even more is what other data from the Pilot Test, and the SDEIS, contains the same errors, but have gone undetected? To put it simply, what other data sets are 400 percent too high, 400 percent too low, or just right? NorthMet SDEIS Comment Daniel Pauly Page 29
Unfortunately this rather obvious error was not detected earlier, because it calls into question fundamental questions as to accuracy of any data sets in the WWTP Pilot Test. In combination with the improper confusion of ng/L and ug/L of mercury in the Tailings Basin (see Comment 1, above), there is reason to be concerned about the fundamental integrity of the very data that is being used to make decisions in the NorthMet project. As stated elsewhere in these comments, I strongly recommend that an independent third party be retained by PolyMet Mining, Inc., or the lead agencies, to audit the data. This will be a significant undertaking, but would allow regulatory bodies the ability to make decisions with confidence that at least the information used for the decision making was sound. I would be willing, at the request of the MDNR or other NorthMet project participants, to undertake a further initial review of other SDEIS datasets to assist a third party audit.
9.
The SDEIS improperly locates one of the Tailings Basin sample locations at Unnamed Creek, which raises further concern about data integrity in the SDEIS process. The most important sample locations for Tailings Basin seepage are arguably the NPDES permit discharges. Proper sampling of these locations is critical to assessing contamination from the Tailings Basin, including mercury, sulfates, etc. Reproduced below is a portion of SDEIS Figure 4.2.2-9. Inexplicably, it has the wrong location for NPDES discharge MN0054089-SD-001 (SD001). The label is placed at a different sample site on Unnamed Creek, but actually it should be in the west drainage ditch of the Tailings Basin.
This discrepancy should be investigated to determine when the error was introduced into the figure, and whether or not the error may have resulted in sampling of water from erroneous locations. Thus, is there any risk that this error, or a similar error, was present on other maps used by field technicians when collecting samples? If an individual had incorrect map data, they would have collected samples from the wrong location and subsequently mislabeled those samples. This could explain WWTP Pilot Test data issues if samples were collected in the incorrect location. I recommend a review of the field instructions and materials to evaluate the potential for erroneous sampling. Note, if technicians collected GPS coordinates at the time samples were collected, it should be possible confirm accuracy of the sample locations.
10.
In view of the identified serious errors in NorthMet data sets as they relate to the Tailings Basin site, a comprehensive audit of NorthMet data should be undertaken. I have had a very limited amount of time, and limited resources, to review NorthMet Project data, and my review has primarily been focused on just a small portion relating to Tailings Basin impacts. Despite this very narrow focus of my inquiry, I have identified the following serious flaws and errors in the NorthMet data set: 10.1 The Tailings Basin mercury contamination calculations confused nanograms and micrograms.
As noted above, the Tailings Basin mercury contamination calculations confused nanograms and micrograms. The result is a profound misstatement of mercury seepage levels at nearly half of the Tailings Basin sample locations. The error was of serious significance because it calls for a revision of Tailings Basin discharges from being below Great Lakes Initiative levels to above Great Lakes Initiative levels. See Comment 1 for more explanation of this problem. 10.2 The Waste Water Treatment Plant influent water has a serious error in either sample collection or data analysis.
The SDEIS relies upon a WWTP Pilot Test that shows serious errors in testing of the influent water. At a minimum these errors prevent analysis and operation of the Pilot Test system because the influent was not properly characterized. In a worst case scenario, other samples were also erroneously collected or analyzed, but those errors have not (and possibly cannot) be identified. See Comment 8 for more explanation of this problem. 10.3 One of the most important sample locations at the NorthMet site is wrongly located on a map.
The SDEIS contains at least one indication of a critical sample location being shown in an incorrect location. It is not clear whether other similar errors have been made, or whether this error impacted sample collection and analysis. See Comment 9 for more explanation of this problem. 10.4 Due to the critical nature of the NorthMet SDEIS, and the centuries-long impacts of the NorthMet Project, an audit is necessary.
A thorough review of the datasets and collection details should be undertaken to confirm integrity and accuracy. A third party competent to perform such a review should be retained, because an independent review is more likely to find serious errors than having the existing team of consultants review their own work. NorthMet SDEIS Comment Daniel Pauly Page 32
I would be willing, at the request of the MDNR or other NorthMet project participants, to undertake a further initial review of other SDEIS datasets to assist a third party audit.
11.
In view of the foregoing comments with regard to mercury contamination, statements made in the NorthMet SDEIS that the Tailings Basin is expected to discharge mercury at levels below Great Lakes Standards should be removed from the SDEIS. The following statements, among others, should be modified as below (or in accordance with alternative language that is accurate): 11.1 At ES-36, in the Executive Summary, edit as follows, or with other language to make the statement accurate and complete: Mercury is another constituent of concern, primarily because many of the lakes and rivers in the area are currently classified as impaired waters by the MPCA due to elevated mercury content in fish tissue. The NorthMet Project Proposed Action is located within the Lake Superior Basin and would be subject to the Great Lakes Initiative (GLI) mercury discharge standard of 1.3 nanograms per liter (ng/L). . . . The mercury concentration in seepage from the Tailings Basin is anticipated to be above below the GLI standard. . . . 11.2 At Page 5-8 of the EIS, edit as follows or with other language to make the statement accurate and complete, and follow up with review of Tailings Basin alternatives in view of this correction: There would also be mercury in the tailings, although about 92 percent of the mercury in the ore is predicted to remain in the ore concentrate. However, and the mercury concentration in seepage from the Tailings Basin is expected to be greater than less than the standard. 11.3 At page 5-206 of the EIS, edit as follows or with other language to make the statement accurate and complete, and follow up with review of Tailings Basin alternatives in view of this correction: Therefore, the total mercury concentration in seepage from the Tailings Basin is expected to be greater than less than the Great Lakes Initiative standard of 1.3 ng/L.

The errors and omissions in the Tailings Basin mercury dataset, and the selective consideration of only favorable mercury test results, have (1) resulted in a defectively designed and tested Waste Water Treatment Plant, which should be corrected with a supplemental comprehensive pilot test, and (2) have resulted in design of a Tailings Basin containment system that fails to consider likely significant transport of methylmercury into the Embarrass River watershed.
12.
The underestimation of mercury contamination in Tailings Basin seepage water resulted in design and testing of a WWTP that has no demonstrated ability to economically remove mercury from the seepage water. In 2012 PolyMets consultants undertook a pilot test with the objective of demonstrating that the WWTP can collect sufficient information to demonstrate that a cluster of technogies, focused on reverse osmosis (RO), can demonstrate reliable satisfaction of water quality objectives, support the design of the WWTP, refine capital and operating costs, and support performance guarantees and system warrantees. As discussed below, the Pilot Test should have addressed all four of these objectives as it relates to mercury, but unfortunately the Pilot Test failed to provide even a rudimentary test and analysis of mercury removal. Only a very few sentences in the Final Pilot Testing Report (SDEIS Reference 2013g) address potential mercury removal. Page 40 to 41 provides the entire summary of the conclusions that could be reached from the pilot test, a review of scientific literature, and an inquiry to the membrane supplier. The entirety of the mercury removal discussion fits on four lines: Mercury removal by RO is highly variable and dependent upon speciation and membrane selection. For these reasons, its removal is difficult to quantify. However, mercury concentrations in the WWTP influent during operations were not estimated by the GoldSim model. PolyMet 2013g at page 40 to 41. Thats it. In other words, as to the question of mercury removal: We dont know, and we didnt even try to find out. It is inexcusable that a pilot test collecting seepage water from a 60 year old unlined tailings basin, known to discharge mercury laden water at two times the Great Lakes Standard, did not even explore mercury removal at a pilot plant that is supposed to clean water for hundreds of years! A deeper dig into the Final Pilot Testing Indicates even more significant issues, which I summarize below: 12.1 The Pilot Test Report Shows mercury removal is likely to be from 22 to 99.9 percent.
Besides the very brief conclusions presented above, the discussion of mercury is given a total of four sentences on page 39 of the Final Pilot Testing Report. I reproduce them below in their entirety: Mercury removal by RO membranes is highly dependent on the type of membrane used. Mercury rejections ranging from 22 to 99.9% have been reported. The chemical state of the mercury is also an important factor in mercury removal. Urgun-Demirtas et al. (Reference (19)) (sic), found that NorthMet SDEIS Comment Daniel Pauly Page 36
mercury in the colloidal or particulate form was easily removed but that free mercury was removed at a lesser rate. Rejection values for organic mercury by RO membranes could not be found in the peer reviewed literature, but one RO membrane vendor (DuPont) and the University of Nevada Cooperative Extension claim that methylmercury cannot be removed across a RO membrane. Paul Dilallo of GE indicated in a personal communication (Reference (8)) that the rejection for mercury is estimated to be approximately 70%. These four sentences are what we have to go on to figure out how well the WWTP might remove mercury. Here is what we know: 12.2 Mercury removal could be between 22 and 99.9 percent, but no one knows and no one has asked.
We know that mercury might be removed by the WWTP. Or it might not be removed. There are two ways to find out, either do Comprehensive Pilot Test based upon more realistic conditions, or wait to see what comes out of the WWTP in a couple decades when the maximum discharges start to occur. 12.3 Mercury contamination is highly dependent upon species, and PolyMets own vendor states that methylmercury cannot be removed across an RO membrane.
Methyl mercury is, far and away, the most dangerous form of mercury. It is the form that is most bio-available, and most likely to accumulate in the tissue of fish and infant humans. According to the best information assembled by PolyMet, the WWTP as designed will not remove methylmercury. Instead it will simply be going out with the discharge water, where it will be part of the makeup water directed into the Embarrass River. Most disturbing about this issue is the fact that if the WWTP selectively removes colloidal or particulate mercury (as at least one reference predicts), then what will be left will be the methylmercury, which wont be removed. The result is the potential for technical satisfaction of the Great Lakes Initiative standard of 1.3 ng/L of mercury in WWTP discharges, while having this 1.3 ng/L of mercury be disproportionately methylmercury. Under such a scenario the WWTP will be in compliance with regulatory standards while actually discharging the most toxic form of mercury in abundance. 12.4 The RO membranes that best remove mercury also have unacceptably low system recovery rates.
A Supplemental Comprehensive WWTP Pilot Test is also necessary because PolyMets own literature review shows that even if an RO membrane can remove mercury, it may not do it with adequate system recovery. Specifically, the Final Pilot Testing Report includes Table 29, which shows a summary of Metals Removal Literature Review Summary. NorthMet SDEIS Comment Daniel Pauly Page 37
Note that mercury removal is reported to be greater than 98 percent in Reference 16, but the system recovery is just 50 percent, which means that only about half the water is recovered as permeate, with a very high volume of high-mercury retentate. Even this is measured only using a pilot level test.
These issues, such as rejection rate and system recovery are non-trivial, and can dramatically impact long term performance and cost of mercury removal. For example, if system recover rates are low, a two-phase system might be needed, which would significantly increase capital and operating costs. Alternatively, it may be necessary to have different systems for mercury removal and sulfate removal if a single treatment system cannot be found that adequately removes both contaminants. In summary, we simply do not have enough information to evaluate a multi-century waste water treatment plant, either from a technical or financial point of view. A comprehensive pilot test should be performed, and it should be directed both to technical viability, as well as financial predictions. In view of the known seasonal variations in contaminant streams (see, e.g., report of Michael Berndt discussed below, showing large seasonal fluctuations in methyl mercury discharges at the toe of a taconite tailings basin), I recommend that such test run for at least from the start of one summer through the end of a following summer.
13.
The WWTP Pilot Test report also indicates serious flaws in the sulfate pilot testing. In an earlier comment I expressed serious concern over the integrity of the data set from the WWTP Pilot Test. The reported influent levels for sulfate concentrations show that error was introduced into sample collection and/or analysis process. The error seems to be on the order of 400 percent, but isnt clear if influent levels have been overstated by 400 percent or understated by 400 percent. In other words, no one really knows what the influent sulfate levels were. In addition, just as alarming, is the potential that the effluent results may have the same errors, or even other unidentified errors. This is not a trivial question, in particular because the final step of water treatment, VSEP process, had permeate level reported to be above 10 mg/l, and about 6 to 60 mg/l. See Figure 12 of the WWTP Pilot Test report. So, should those levels be 24 to 240 mg/L? Or maybe they should be about 1.5 to 15 mg/l. Preparers of the SDEIS should review sample records, analytical data, and related material to ascertain the accuracy and integrity of the data set.
14.
A Comprehensive WWTP Pilot Test should be conducted in a manner that provides meaningful information as to likely costs. As noted above, the WWTP Pilot test failed to even look at mercury as a substance to be removed from the NorthMet site, and also failed to even properly monitor sulfate levels in the influent. Other contaminants, such as aluminum, were also not properly analyzed during the Pilot Test. The need for a supplemental WWTP pilot test is clear from a technical viability standpoint, but is also necessary from a financial viability standpoint. The RO literature is clear that each site is different, and the costs of systems really cant be estimated with insufficient information about feed water and likely membrane performance. A good example of that information is a 2013 Report from the U.S. Department Of Energy entitled Reverse Osmosis Optimization (available from comment author upon request), which states as follows: The cost of optimizing an RO system is influenced by many parameters that are specific to the application and operation of the system, such as feed water quality, membrane type, system configuration, and purity requirements. Therefore, to determine the costs and financial benefits of optimization options, the financial analysis must take into account the site-specific nature of the technology. Reverse Osmosis Optimization, by Pacific Northwest National Laboratory, page 19 (emphasis added). These are non-trivial issues for which the SDEIS contains inadequate information, which is quite problematic for a WWTP that is expected to process 630,000 gallons per day of water for hundreds of years.
15.
The SDEIS incorrectly states that the Pilot-testing has indicated that treated effluent from the Plant Site would meet water quality standards for all regulated constituents. As discussed above, the WWTP Pilot Test did not determine whether mercury would be removed by the WWTP. The SDEIS should be updated to reflect this fact. A partial list of suggestions is provided below: 15.1 Table 5.2.2-28 should be modified to include mercury as a target. Table 5.2.2-28 provides the WWTF preliminary water quality targets. Mercury is not included but should have been included. After mercury is added as a target effluent, design of the WWTF should be evaluated for mercury removal, and a review should be made of SDEIS analysis and conclusions that presumed mercury did not need a WWTF water quality target. This review should include evaluation of alternatives that were prematurely eliminated. 15.2 Page 5-125 should be modified to clarify that mercury removal was not tested in the pilot plant. Suitable language could include: Table 5.2.2-28 presents the target WWTF effluent concentrations for the different mine phases. Pilot-testing of a WWTF with RO demonstrated that all of the target closure effluent concentrations could be achieved with the planned WWTF design, with the possible exception of mercury, for which no pilot testing has been undertaken and for which significant disagreements exist on viability of removal by RO processes. 15.3 Page 5-203 of the SDEIS should be amended to include statements from the Pilot Plant Test that removal of mercury using RO technology is uncertain. Potential language to consider is: It should be noted that the West Pit overflow would be treated by the WWTF using RO technology prior to discharge, and the RO process is known to remove mercury. However, as indicated in the Pilot Test Report, there are disagreements in the literature about how much mercury can be removed, and the most dangerous form of mercury (methylmercury) has been reported as unremovable using RO methods. 15.4 Table 5.2.2-52 should be modified to provide correct information of estimated mercury concentration of the combined inflows to the Plant Site WWTP Table 5.2.2-52 describes Mercury Concentration from Tailings Basin seepage water and Runoff (not interacting with tailings) to be 1.1 ng/L, which is below Great Lakes standard levels. In fact, mercury levels for these two sources will likely be higher than these estimated levels.
Correcting this error is essential because these two sources of inflows are expected to account for over 80 percent of the water inflows into the WWTP, and will impact feasibility and design of the WWTP, as well as long term costs to operate the WWTP. For the assistance of the SDEIS drafters, current table 5.2.2-52 is reproduced below with the erroneous unsupported information highlighted in yellow. Also provided below is alternative calculations that are based upon Comprehensive Tailings Basin Water Sampling data of Pauly Comment 3:
Table 5.2.2-52 Estimated Mercury Concentration of the Combined Inflows to the Plant Site WWTP
Stream Seepage Water Runoff (interacting with tailings) Runoff (not interacting with tailings) Tailings Basin pond dewatering Combined stream Flow Rate (gpm) 1,498 294 72 365 2,229 Mercury Concentration (ng/L) 1.1 1.1 3.5 2.0 1.3 Total Mercury Flow (ng/yr) 3.3E+09 6.4E+08 5.0E +08 1.5E+09 5.9E+09
When corrected, the numbers become approximately as follows: Recommended Table 5.2.2-52 Estimated Mercury Concentration of the Combined Inflows to the Plant Site WWTP
Stream Seepage Water Runoff (interacting with tailings) Runoff (not interacting with tailings) Tailings Basin pond dewatering Combined stream Flow Rate (gpm) 1,498 294 72 365 2,229 Mercury Concentration (ng/L) 2.5 2.5 3.5 2.0 2.5 Total Mercury Flow (ng/yr) 7.5E+09 1.5E+09 5.0E +08 1.5E+09 1.0E+10
It should be noted that the mercury discharge numbers in recommended Table 5.2.2.52 correspond quite closely with the only known high quality sample of Tailings Basin discharges when the Tailings Basin was last receiving tailings (Berndt 2003), and also align quite closely to current Tailings Basin discharges when a comprehensive sample
analysis is undertaken as opposed to the very narrowly focused sample sets proposed in the SDEIS (and discussed earlier).
16.
The SDEIS proposes a perimeter wetland that will receive sulfate and mercury, and has the potential to lead to significant methylmercury production and transport into the Embarrass River watershed. One of the most significant changes in the Tailings Basin since the DEIS was prepared is a proposal to construct a perimeter cutoff wall and drainage system around the north and west sides of the Tailings Basin. The objective of capturing discharges from the Tailings Basin has merit, but the proposed plan is seriously flawed because the cutoff wall and drain pipe are set back over 250 feet from the edge of the Tailings Basin, so as to form a large perimeter wetland of up to 160 acres in size. As discussed below, due to the nature of the soil in this wetland, the majority of mercury and sulfate laden seepage will be delivered into the wetland, as opposed to going directly to the drain pipe system. The potential extent of this perimeter wetland is shown in the figure below, which is from Barr 2013F, but has been highlighted in yellow to show the perimeter area between the Tailings Basin and drainage pipe.
Wetlands have been known for some time to be prime locations for methylation of mercury from its inorganic form to its far more hazardous organic methylmercury form. Recent research in Minnesota has shown that some of the highest methylmercury levels ever recorded have been observed in restored wetlands. In addition, recent Minnesota DNR research at tailings basins has shown that seepage water from tailings basins has particularly high methylmercury concentrations during seasons of heightened biological activity. As discussed below, the proposed Tailings Basin capture system has the potential to create very high levels of methylmercury. Once created, the methylmercury will be able to exit the capture
area either by 1) biological transport into the Embarrass River watershed, or 2) passing through the Waste Water Treatment Plant, which has not been designed or tested to remove it. These concerns are detailed below. 16.1 The unique nature of the soils at the NorthMet site result in delivery of tailings seepage to the wetlands, rather than to the drainage pipe.
The Containment System called for in the SDEIS is shown in cross section below, which is taken from Figure 4 of reference PolyMet 2013f. This figure is a useful one for evaluating containment system flow during mine operations because it is taken on the north end of the Tailings Basin where the greatest amount of discharge is predicted.
The black lines show the flow path of the seepage according to NorthMet Project Modeling. Notably, according to the model, 78 percent of the seepage water is delivered to the wetlands, while only 22 percent of it is delivered to the Containment System Drain Pipe. The reason for this flow path, according to SDEIS documents, is that the relatively low hydraulic conductivity of the soils forces the water to the surface, where wetlands will likely expand and proliferate. Within these large perimeter wetlands, the ideal conditions for mercury methylation are likely to occur: abundant mercury, sulfate, dissolved organic compounds (DOC), and water. The SDEIS entirely fails to consider what will happen when the mercury, sulfates, DOC and water combine in these perimeter wetlands. The SDEIS should be modified to consider the following possible impacts:
16.2
A recent Minnesota study showed concentrations of methylmercury in restored wetlands had some of the highest in published literature, suggesting creation of wetlands to receive a mercury and sulfate mixture is a serious concern for the Embarrass River ecosystem and downstream inhabitants.
In 2013 the USGS released a study on mercury methylation in northwestern Minnesota, and found stunningly high levels of methylmercury. Key portions of that report are reproduced below, with emphasis added: Compared to concentrations in stream sediment samples collected throughout the United States, Glacial Ridge National Wildlife Refuge wetland sediment samples contained typical total-mercury concentrations, but methylmercury concentrations were nearly twice as high. The maximum concentration measured in Glacial Ridge National Wildlife Refuge wetland water approached the highest published water methylmercury concentration in uncontaminated waters of which we are aware. However, the upper quartile of water methylmercury concentrations is similar to concentrations reported for some impoundments and wetlands in northwestern Minnesota and North Dakota. Methylmercury concentrations in sampled wetlands were much higher than those from typical lakes or flowing streams throughout the United States. The high concentrations of methylmercury measured in sampled wetlands indicate the potential for substantial methylmercury concentrations in aquatic biota and wildlife that consume those biota. These wetlands also are a methylmercury source for downstream lakes and rivers. The high concentrations of methylmercury in water, its bioaccumulation potential, and its known toxicity in aquatic birds and food webs highlight a need to assess methylmercury in the biota within these ecosystems. Better understanding of factors that control methylmercury production concentrations within aquatic food webs in ecosystems of the Glacial Ridge National Wildlife Refuge would enable resource managers to better understand and manage risk to wildlife.
Scientific investigations Report 2013-5068, Mercury in Wetlands at the Glacial Ridge National Wildlife Refuge, Northwestern Minnesota, 2007-2009. Emphasis added. Copy available from comment author upon request. 16.3 Recent DNR research also shows that taconite tailings basins can have particularly high methylmercury releases that coincide with times of greatest biological activity in wetlands This research should be considered in evaluating impact of wetlands positioned between the Tailings Basin and containment system drain pipe.
Another very serious issue as it relates to methylmercury releases is that recent MNDR research shows that discharges from a taconite tailings basin had a seasonal spike in methylmercury discharges during the summer months when most wetland biological activity occurs.
This research is compiled in Sulfate and Mercury Cycling in Five Wetlands and a Lake Receiving Sulfate from Taconite Mines in Northeastern Minnesota, Berndt and Bavin, 2011. Reproduced below is a portion of Figure 9 of Berndt and Bavin, showing the seasonal changes in total mercury and methylmercury from May to October.
May
to
October
The excerpt above shows how total mercury was essentially constant from May to October, but the amount of methylmercury increased significantly during summer months. It is not clear if more methylmercury is being produced during the summer months, or if it is simply being released in greater quantities, but this increase should be of great concern to anyone proposing a Tailings Basin perimeter wetland. Should this effect also hold true at the NorthMet Tailings Basin, levels of methylmercury must be contemplated, especially during summer months. A further concern raised in Berndt and Bavin with regard to the examined wetland, at a taconite tailings basin toe similar to the NorthMet Tailings Basin, is that the mechanism for the increases in methylmercury are not obvious nor necessarily fully understood. Berndt and Bavin at page 13. It may be that increases in sulfate reduction in summer months lead to accelerated methylmercury production. It appeared that the methylmercury increased across a backdrop of continuous amounts of sulfate reduction and DOC release. In evaluating the containment system as proposed, evaluation of increased methyl mercury loading should be evaluated. Inquiry should be made to assess how much of the methylmercury can be expected to bioaccumulate in organisms in the wetland and how much will go to the NorthMet SDEIS Comment Daniel Pauly Page 47
WWTP. Based on Figure 4, above, it would appear that most seepage from the toe of the Tailings Basin will be delivered to the wetlands. This is of great concern because summer is the time of greatest biological activity, and the time when greatest bio absorption can be inferred to occur. 16.4 This methylmercury is likely to be transported out of the containment system wetlands by biological activity. That which remains and is captured, will go to the WWTP where no plans have been made for it to be captured and removed.
One potential major impact of the newly proposed containment system is that in addition to creating a large wetland where methylmercury can accumulate and form, the wetland will provide a biological pathway for transport of the methylmercury to the Embarrass River watershed and then into the St. Louis River. Current research (available from the comment author upon request) shows that everything from mosquitos to earthworms can accumulate methyl mercury in high doses. These simple life forms will likely either simply fly out of the containment system (in the case of insects) or be consumed by other mobile organisms (such as great blue herons, egrets, pipers, gulls, ducks, raccoons, etc.) that will allow transport of the mercury up the food chain in the Embarrass River watershed, where it will eventually accumulate in fish and humans. Whereas this avenue of transport totally avoids the drainage pipe collection system and the WWTP, it is a problem for which no solution has been proposed. Note that with 78 percent of mercury being delivered to the wetlands, along with potential methyl mercury spikes that are known to occur at taconite tailings basins, the potential for this bioaccumulation is significant. It is also likely to continue for hundreds, if not thousands of years, and will be far greater than would occur with the lower water flows of the No Action Alternative or the other Tailings Basin alternatives discussed below.

In view of the data integrity issues and improper mercury discharge conclusions, the NorthMet SDEIS review should consider previously discarded Tailings Basin alternatives, and conduct a new WWTP pilot test.
17.
The failure to appreciate the potential for groundwater mercury impacts from the NorthMet Tailings Basin resulted in alternatives being improperly dismissed in the DEIS and SDEIS that should have been considered. NEPA and MEPA require consideration of alternatives, as summarized below in one of PolyMets ERM reports: NEPA and MEPA have slightly different requirements for considering alternatives. NEPA describes the consideration of alternatives as the heart of the environmental impact statement. NEPA requires the consideration of all reasonable alternatives and a discussion of alternatives which were eliminated from further study. The decision maker must consider all reasonable alternatives and cannot consider alternatives not discussed in the EIS. MEPA (Minnesota Rules, part 4410.2300, subpart G) states that an alternative may be excluded if it would not meet the underlying need for or purpose of the Project; it would likely not have any significant environmental benefit compared to the Project as proposed; or another alternative, of any type, that will be analyzed in the DEIS would likely have similar environmental benefits but substantially less adverse economic, employment, or sociological impacts. (ERM Report, Emphasis added). In the present case, as discussed in related comments, the failure of the DEIS and SDEIS process to consider mercury impacts from the NorthMet Project has resulted in potentially superior alternatives being prematurely dismissed.
18.
The failure of the SDEIS to consider non-compliant mercury discharges resulted in improper elimination of alternative tailing basin covers during the ERM process. In December, 2010 the ERM presented recommended alternatives for the NorthMet Mine Tailings Basin Cover Options, including both an evaluation and recommendation. In that 12/2010 report the ERM stated, correctly, that: The Tailings Basin cover system is a very important environmental component of the overall NorthMet Project . . . . At least some seepage is expected to occur long term (e.g. centuries) from the Tailings Basin after closure, so the quality of that seepage is of critical importance and will largely determine the need for long-term operation, maintenance, and/or treatment. (ERM at Page 1, emphasis added) The report then states, incorrectly, that: Based on the analysis of the DEIS, the current PolyMet Proposal (i.e. referred to as the Tailings Basin Alternative in the DEIS) was predicted to generally meet groundwater standards with the primary exception of sulfate. . . . [but] no longterm operation, maintenance, or treatment would be needed for the Tailings Basin as currently modeled for any other parameter because they are predicted to meet groundwater standards. (ERM at page 17). The ERM report subsequently evaluated five different alternatives on the basis of sulfate loads, with no mention or consideration of potential for mercury release from the Tailings Basin. In fact, this key planning document discusses sulfate 50 times on its 22 pages, but not once does the word mercury even appear. This is a prime example of why the NorthMet Project should be reevaluated to consider alternatives that will satisfy the NEPA and MEPA screening requirement for mercury. In this case, the NorthMet Project has identified the criticality of the Tailings Basin cover to groundwater seepage, but SDEIS entirely disregards likely mercury impacts, even though mercury release is the single biggest threat of public health from the NorthMet Project.
19.
Deficiencies in Tailings Basin mercury characterization have resulted in failure to evaluate two viable, and potentially superior, Tailings Basin alternatives. In particular, TBM-3 and TBM-21 should be evaluated as better performing alternatives During the DEIS process, two specific Tailings Basin alternatives were proposed that have the potential to meet all environmental requirements and also be less expensive than the proposed NorthMet alternative. The first of these alternatives, TBM-3, would have included a full liner between new and existing tailings. The second of these alternatives, TBM-21, would have called for an alternative location of a new basin. Both of these mitigation measures were deemed to address all four of the issues identified in the Tailings Basin mitigation process (See DEIS at pages 3-56 to 3-58), but were dismissed as being economically inferior to the recommended proposal. Specifically, it should be noted that each design would meet the purpose and need, would be technically feasible, and would be regulatorily feasible. When TBM-3 and TBM-21 were dismissed, the economic viability was being compared to the original DEIS proposal, which did not include the containment system or WWTP as proposed, both of which increase the cost of the presently proposed design relative to the prior DEIS proposal. In addition, the now apparent need to also remove mercury from Tailings Basin seepage, as well to prevent it from entering wetlands surrounding the Tailings Basin, will add additional cost to the proposed Tailings Basin plan. The failure to properly conduct a WWTP pilot test adds even greater uncertainty to the potential costs of the proposed Tailings Basin alternative. In aggregate, these changes in the Tailings Basin options since the DEIS eliminated TBM-3 and TBM-21 indicate that the dismissed alternatives should be reconsidered. On the following page I have a brief summary of distinctions between the proposed NorthMet Tailings Basin Alternative compared to TBM-3 and TBM-21
Tailings Basin Alternatives in view of Mercury Contamination Metric Proposed Action Compared to no action option Tailings basin mercury discharges . Likely to increase as the current LTV tailings basin is drying out and discharges greater mercury levels when the tailings basin receives new water and tailings TBM-3 - Segregated Tailings Basin Compared to proposed action Lower than Proposed Action because a lined NorthMet basin can reduce mercury release from the existing LTV tailings and will also not have water directed through the tailings. The liner will also improve capture of NorthMet seepage, reducing releases to wetlands. Lower than Proposed Action because a lined NorthMet basin can prevent mercury from entering adjacent wetlands. Also, the liner will reduce mercury release from the existing LTV tailings because less water be directed through the tailings. Lower by avoiding discharge to perimeter wetlands of the Proposed Action Higher cost will increase with regard to the segregation system, but decrease in regard to the capture system. Lower than Proposed Action because any treatment plant will potentially be smaller, less complex, need to treat less water, and likely run for a shorter duration. Lower Same TBM-21: Distinct Tailings Basin Compared to Proposed action Lower than Proposed Action because the NorthMet basin will not be in fluid communication with the LTV basin. Mercury releases from the current tailings basin will continue to decline, and discharges from a new basin can be captured Lower than Proposed Action because the NorthMet basin the effects of a combined discharge of sulfate and mercury can be diminished, either with a liner or a capture system that doesnt create a wetland. Lower by avoiding discharge to perimeter wetlands of the Proposed Action Higher
Mercury methylation
Higher
Potential methylmercury production Cost Installation
Increased Lower
Cost Long term operation, maintenance and support.
Higher system will require significant long term water treatment of a combined discharge stream from both the LTV and NorthMet tailings. Higher No Change
Potential for non-modeled risks, including combinatorial reactions in the mixed waste stream. Land use disruption
Lower than Proposed Action because any treatment plan will likely be smaller, less complex, need to treat less water, and likely run for a shorter duration. Lower More

NorthMet Comments by Daniel Pauly

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

NorthMet Comments by Daniel Pauly

Загружено:

Авторское право:

Доступные форматы

Comments Overview:

NorthMet SDEIS Comment Daniel Pauly Page 1

NorthMet SDEIS Comment Daniel Pauly Page 3

NorthMet SDEIS Comment Daniel Pauly Page 4

NorthMet SDEIS Comment Daniel Pauly Page 5

PAULY COMMENT SECTION 1

NorthMet SDEIS Comment Daniel Pauly Page 6

NorthMet SDEIS Comment Daniel Pauly Page 8

Pauly Comment Figure 1 Excerpt from Barr 2006F

NorthMet SDEIS Comment Daniel Pauly Page 9

NorthMet SDEIS Comment Daniel Pauly Page 10

NorthMet SDEIS Comment Daniel Pauly Page 11

NorthMet SDEIS Comment Daniel Pauly Page 12

Mean Mercury Level Ng/L2 1.8 1.4 1.2 1.6

SD001 SD002 SD0043 SD006

2001-2006 2001-2006 2001-2006 2001-2006

Included in primary SDEIS data sets for Mercury contamination? No No Yes No

NorthMet SDEIS Comment Daniel Pauly Page 13

NorthMet SDEIS Comment Daniel Pauly Page 14

NorthMet SDEIS Comment Daniel Pauly Page 15

Sample Source LTV-DNR4 PM-86 PM-9 PM-10 PM-11

2005-08 1997- 2008

NorthMet SDEIS Comment Daniel Pauly Page 16

NorthMet SDEIS Comment Daniel Pauly Page 17

NorthMet SDEIS Comment Daniel Pauly Page 18

NorthMet SDEIS Comment Daniel Pauly Page 19

4.2.2 WATER RESOURCES

SRK Consulting Report at Page 16.

NorthMet SDEIS Comment Daniel Pauly Page 23

NorthMet SDEIS Comment Daniel Pauly Page 24

See, in particular, SDEIS page 5-205 See page 48 of Berndt 2003.

NorthMet SDEIS Comment Daniel Pauly Page 25

NorthMet SDEIS Comment Daniel Pauly Page 26

NorthMet SDEIS Comment Daniel Pauly Page 27

NorthMet SDEIS Comment Daniel Pauly Page 28

NorthMet SDEIS Comment Daniel Pauly Page 30

NorthMet SDEIS Comment Daniel Pauly Page 31

NorthMet SDEIS Comment Daniel Pauly Page 33

NorthMet SDEIS Comment Daniel Pauly Page 34

PAULY COMMENT SECTION 2

NorthMet SDEIS Comment Daniel Pauly Page 35

NorthMet SDEIS Comment Daniel Pauly Page 38

NorthMet SDEIS Comment Daniel Pauly Page 39

NorthMet SDEIS Comment Daniel Pauly Page 40

NorthMet SDEIS Comment Daniel Pauly Page 41

NorthMet SDEIS Comment Daniel Pauly Page 42

NorthMet SDEIS Comment Daniel Pauly Page 43

NorthMet SDEIS Comment Daniel Pauly Page 44

NorthMet SDEIS Comment Daniel Pauly Page 45

NorthMet SDEIS Comment Daniel Pauly Page 46

NorthMet SDEIS Comment Daniel Pauly Page 48

PAULY COMMENT SECTION 3

NorthMet SDEIS Comment Daniel Pauly Page 49

NorthMet SDEIS Comment Daniel Pauly Page 50

NorthMet SDEIS Comment Daniel Pauly Page 51

NorthMet SDEIS Comment Daniel Pauly Page 52

Potential methylmercury production Cost Installation

Cost Long term operation, maintenance and support.

Вам также может понравиться