NC0000396_60% BOD Comments Revised FINAL_20170727

Water Resources
Environmental Quality
July 27, 2017
Ed Sullivan
Duke Energy
526 South Church Street
Mail Code EC13K
Charlotte, North Carolina 28202
Subject: Revised 60% Basis of Design Report Comments
Asheville Steam Electric Plant
Dear Mr. Sullivan:
ROY COOPER
Governor
MICHAEL S. REGAN
Secretary
S. JAY ZIMMERMAN
Director
On March 17, 2017, the North Carolina Department of Environmental Quality's Division of Water
Resources (Division) received the 60% Basis of Design (BOD) Report for the subject facility. Division
staff from the Asheville Regional Office reviewed the BOD Report and provided comments in
Attachment 1 of a letter dated July 18, 2017. After these comments were submitted, we discovered
several minor errors that require correction; generally, a few comments were omitted or repeated.
Therefore, we have attached revised comments (see Attachment 1) that shall supersede the previous
BOD Report comments. If you have any questions, please feel free to contact Ted Campbell at the
Asheville Regional Office at [phone] or Steve Lanter at [phone].
Sincerely,
Jon ' gaard, Section Chief,
Water Quality Regional Operations Section,
Division of Water Resources
Attachment: Attachment 1 Asheville Steam Electric Plant 60% Basis of Design Comments
cc: Landon Davidson — ARO Regional Office Supervisor
WQROS Central File Copy
State of North Carolina I Environmental Quality I Division of Water Resources
Water Quality Regional Operations Section
1636 Mail Service Center I Raleigh, North Carolina 27699-1636
919-707-9129

Attachment 1
Asheville Steam Electric Plant 60% Basis of Design Report Comments
1. The Division recognizes that the proposed interim action is intended to augment, and not replace,
corrective action required in 15A NCAC 02L .0106. The interim action activities (groundwater
extraction) should, within the BOD, be placed in the context of a corrective action or partial
corrective action under 02L .0106. The Basis of Design (BOD) Report should (a) state the
subsection of 02L .0106 (for example, (k), (1), and (or) (m)) that is being used to address the target
contamination if any. The expectation is that the proposed interim action, coupled with the future
corrective action, will lead to full compliance with 02L .0106 for all constituents and that adherence
to all relevant subsections of 02L will be demonstrated within the CAP.
2. The BOD Report should state whether adverse ecological impacts, if any, are expected because of
the proposed groundwater extraction system and the basis for this expectation.
3. The BOD Report should explain whether and how the interim action will address all constituents
of interest (COI), rather than a subset of COIs. If the proposed groundwater extraction system is
not intended to address all COIs, a section of the BOD Report should explain what additional
remediation efforts are contemplated and the conditions under which they would be implemented
in any additional corrective action. For example, Page 1-1 states that `B, Cl, Co, Fe, Mn, Se, SO4,
TDS, and V have been identified as constituents occurring at levels above 2L or IMAC. B, Cl,
5O4, and TDS impact, which have the potential to migrate beyond the compliance boundary and
off site, will be the focus of remedial efforts." While the Interim Action may focus on B, Cl, SO4,
and TDS, the BOD Report should acknowledge that Co, Fe, Mn, Se, and V must also be remediated
and that the CAP (not the BOD Report) will address these. The Division recognizes that
conservative, poorly-sorptive constituents like B, Cl, and SO4 will tend to move off site most
rapidly, but 02L .0106 compliance is expected for all constituents.
4. The bulk attenuation rate discussed on page 2-2 was computed using an inappropriately determined
slope value (m). Instead of using boron concentrations from monitor wells along a plume
longitudinal centerline to plot and compute the slope as described in the referenced EPA method
(EPA/540/S-02/500, page 10), the report plotted and used boron concentrations from monitor wells
that were side gradient and (or) screened across various, sometimes uncontaminated flow units.
The failure to properly apply and compute the bulk attenuation rate should be corrected in the BOD
Report. Comment 7 below discusses the need for additional monitor wells.
5. Page 2-3 states that "The following solution to the advective-dispersive equation (with a source
area) provides downgradient concentrations along the centerline of a plume (Domenico, 1987)".
The report attributed differences between computed and observed boron values to "differing levels
of dilution than anticipated in the model' and these were related to variability in hydraulic
conductivity. No mention was made to the fact that the selected equation is designed to predict
concentrations in porous media and along a plume centerline, and that the wells used to measure
boron are either bedrock wells (and unconnected to the centerline plume mass) or are side gradient
Page 1 of 11 July 27, 2017

wells not positioned along the plume centerline. The equation is not applicable to concentrations
observed in the monitor wells that currently exist. The BOD Report should be corrected.
6. The BOD Report does not need to discuss or address constituents such as Ba, Ca, Mg, Zn, etc. that
do not occur above 21-JIMAC (see the fourth bullet on page 2-5). The BOD should focus on B, Cl,
Co, Fe, Mn, Se, SO4, TDS, and V, and discuss the factors that have been shown (by actual site
data, F -T modeling, and geochemical modeling) to control their downgradient occurrence. The
expectation is that these factors be well understood and that the BOD or future corrective action
plan describe these controls in detail using data from the source area itself and how the proposed
corrective action will address them.
7. Current monitor wells are insufficient to map the boron plume or other COIs emanating from the
source area and migrating to the 12 -acre offsite parcel and beyond. A request for additional wells
was discussed during the June 30, 2017, meeting at the Mooresville Regional Office between Duke
Energy and the Division. Agreements reached during the meeting regarding the installation of
additional wells and installation schedule should be carried out as part of ongoing assessment,
accelerated remediation, and CAP efforts at the Asheville Plant.
8. The BOD Report appendices should contain a copy of the boring logs from all wells used to map
boron in the source area.
9. Boron isotope sampling was proposed (CSA Addendum, Synterra, 2016) to help evaluate whether
boron concentrations in supply wells west of the French Broad River are associated with coal ash
(to date, two wells contain unexplained, anomalous boron concentrations). Although alternate
water has been proposed for these wells, the area west of the river is still considered a future use
area for purposes of 2L .0106 compliance. The BOD Report should include a section describing
the boron isotope sampling effort to date. This section should include a map of receptor survey
supply wells and a list of the monitoring and supply wells that have been sampled for boron
isotopes. For supply wells that were not sampled, please state the reason (for example, well owner
declined, well owner was unable to be reached, well has been abandoned, etc.).
10. Section 8.1, System Performance Metrics, page 8-1. The BOD states that F -T simulations indicate
"noticeable" concentration reductions in the vicinity of EXT -01 within 5 years of system start up.
To help assess the groundwater extraction system (and F -T model) performance, the BOD should
provide a set of stacked concentration -time plots of boron for each modeled layer at the following
locations: CCR -101, CCR -102, MW -9, CB -8, EXT -01, MW -16, MW -17, A-1 (seep), B-1 (seep),
and at center, eastern edge of the 12 -acre parcel. The x-axis (time) should extend out to the time
it takes for concentrations to drop below the 2L standard of 700 ug/L at each location.
11. Page 8-1 includes a list of wells that are proposed to be sampled quarterly. This list should also
include new well MW-17OBR, MW -17A, CCR-101BR, CCR -102S, CCR -102D, CCR -103D,
CCR-103BR, and areas of wetness B-01, MH -17, PW -F, and C-01. If MW-17OBR does not
contain detectable concentrations of boron after 2 sample rounds, Duke Energy may contact the
Regional Office for a reduced sampling frequency for that well.
Page 2 of 11 July 27, 2017

12. On Page 8-1, daily water level measurements throughout the first quarter of operation are proposed
for a list of 5 bedrock wells (MW-8BR, MW -913R, CCRAO1BR, MW-16BR, and MW-18BR) and
the French Broad River. To assess the interconnectivity of wells and flow units in this area this
list should also include MW -17A, new well MW- 17OBR, new well MW -16D (or, if MW -16D is
dry, MW -16A), CCR -102D, CCR- 103BR, and MW -9D.
APPENDIX C - UPDATED FLOW AND TRANSPORT MODEL
1. As discussed in Section 4.3, the western boundary of the flow model along the French Broad River
is treated as a specified head in the uppermost active layer and as a no flow boundary for all deeper
layer boundaries. Constructing the model in this way precludes, apriori, the possibility of flow
beneath or along deeper bedrock fractures that are known to occur in this area of the model. Wide,
often flowing fracture sets (with openings of up to 78 inches) were observed below 200 feet during
geophysical logging of supply wells near the river (Synterra, August 2016, CSA Supplement 1,
Appendix F). In addition, the western boundary of the model domain does not include a number
of supply wells west of and in close proximity to the river. The Regional Office requests a meeting
with Duke Energy and its modelers to discuss the model domain and model boundary conditions.
This meeting should occur before the remaining comments below are addressed or incorporated
into a revised model report (not expected in the BOD).
2. The time, direction, and distance of contaminant travel must be predicted assuming the source has
been excavated. The predictions should be presented in the CAP (not expected in the BOD) and
include the following:
a. A concentration -time plot - for each COI - corresponding to the following locations: (i)
nearest supply well, (ii) nearest future groundwater use area (please consult with Asheville
Regional Office prior to modeling), and (iii) nearest water of the state. In the plot margin,
the following information should be provided: the time in years it takes for the COI to reach
(i), (ii), and (iii), the time it takes for the COI to reach (i), (ii), and (iii) at its 2L/IMAC
concentration, the time it takes for the COI to reach (i), (ii), and (iii) at its maximum
concentration, and the time it takes for the COI to reach (i), (ii), and (iii) at a concentration
that is back below the 2L/IMAC concentration.
b. A map showing the maximum migration distance at any detectable concentration of each
COI.
c. A map showing the maximum migration distance at the 2L/IMAC concentration of each
COI.
3. The following additional figures are needed in the CAP (not expected in the BOD Report):
a. Observed potentiometric surface map of shallow system. Map should include the water
level of any surface water feature that was measured as part of the potentiometric
mapping effort.
Page 3 of 11 July 27, 2017

b. Observed potentiometric surface map of bedrock system. Map should include the
water level of any surface water feature that was measured as part of the potentiometric
mapping effort.
c. Simulated potentiometric maps for each modeled layer, during pre -pumping
conditions. Maps should show the simulated water level of key surface water features
(FBR upstream and downstream, outfalls to the FBR, surface water locations measured
in (a) above, etc.).
d. Simulated potentiometric maps for each modeled layer, after pumping has reached
steady state conditions. Maps should show the simulated water level of key surface
water features (FBR upstream and downstream, outfalls to the FBR, surface water
locations measured in (a) above, etc.).
e. Simulated versus observed residual heads map — for each modeled layer - showing
color coded ranges of error across the model domain (for example, residuals greater
than +10 ft, +3 to +10, 0 to +3, 0 to -3, -3 to -10, greater than -10).
f. Simulated boron isoconcentration map - for each modeled layer - showing the
following ranges: 0 to 300 ug/L, 301 to 700, 701 to 2000, 2001 and above.
g. Simulated versus observed residual boron concentrations map — for each modeled layer
— showing color coded ranges of error across the model domain (for example, residuals
greater than +1000 ug/L, +200 to +1000, 0 to +200, 0 to -200, -200 to -1000, greater
than -1000).
h. Simulated isoconcentration map of each COI — for each modeled layer - showing a
maximum of three concentration ranges per map (one range below the 02L/IMAC
standards and two ranges above the 021/IMAC standards).
4. For each figure, need a map scale and a legend that defines all features/elements that are color
coded or otherwise mapped.
5. Section 2.5, Sources and Sinks. The modeled drains shown in Figure 6 include a stand-alone,
isolated drain along I-26 that appears to be disconnected from all other streams that drain to the
river. Another disconnected drain is located in the 1982 basin. The report should describe how
water and contaminant mass can be removed from the flow system through these disconnected
features. It seems that the water in any such feature would either flow to the river or re -infiltrate
to the groundwater system.
6. Section 2.7, Modeled Constituents of Interest. The report identifies Sb, As, B, Cr, Co, Fe, Mn,
SO4, Se, Tl, TDS, V, and ph as constituents of interest across the Asheville site based on their
presence above 2L/IMAC in pore water. Cl was also included based on its presence above
2L/IMAC in downgradient groundwater. However, only B, Cl, SO4, and TDS were modeled. All
constituents of interest must adhere to 2L .0106 and be addressed by corrective actions, whether
Page 4 of 11 July 27, 2017

they are migrating as a "plume" or are being solubilized from an otherwise relatively stable solid
phase as geochemically altered groundwater moves through the subsurface. Any constituent not
addressed by the F -T model should be evaluated for 02L .0106 compliance using another method
such as geochemical modeling or other approach designed to understand and predict solubility and
mobility. This should be addressed and presented in the CAP (it is not expected in the BOD
Report).
7. Section 4.3, Flow Model Boundary Conditions. Lake Julian and the French Broad River form
much of the east and west model boundaries and are treated as specified head boundaries in the
uppermost active model layer and as no -flow boundaries in the deeper layers. The report should
state the depth of incision (ft below land surface) of Lake Julian and the French Broad River, and
to what model layer each waterbody extends.
8. Section 4.4, Flow Model Sources and Sinks. The report states that creeks and drains exert a
significant control on the hydrology in the model and that they were simulated using the DRAIN
feature in MODFLOW with a high conductance value of 500 ft2/d/ft. Based on field observations,
some of the drains to the French Broad River may be wet (flowing) or dry (including on September
28, 2016), depending on various factors. The report should present the observed stage and flow
measurements that were made in support of drain calibration along with the corresponding
simulated values.
9. Section 4.5, Flow Model Calibration Targets. The Appendix states that 79 water level
measurements made in the 3rd and 4'h quarters of 2016 were used as calibration targets. Most of
the measurements were made in Q4, but in cases where only Q3 levels were available, those were
used. Appendix C should present a table of the 79 wells and associated WL measurements made
during Q3 and (or) Q4 to allow the reviewer to assess how well Q3 water levels align with Q4
water levels across the site.
10. For any well not used as target well for head calibration, the report should state why. All existing
wells should be used to assess the quality of calibration unless an obvious construction or
development issue precludes a valid water level measurement.
11. Section 4.6, Transport Model Parameters. Appendix C states that "The final basin recharge rates
used during sluicing in the transient flow model range from 12 to 24 inches per year. These rates
are much smaller than the rate of water inflow to the basins with the sluiced ash". The report
should state what accounts for this difference and how the much smaller modeled rate affected the
F -T model calibration and flow and concentration results.
a. The report states that the ash basins were subdivided during the transport calibration
process and different concentrations were assigned to different zones at different times.
The report should present these input concentration zones and amounts, and how the
amounts changed over time.
b. The effective porosity was set to a value of 0.001 in all the bedrock layers. The report
should explain the difference between primary porosity and secondary porosity, and
Page 5 of 11 July 27, 2017

how the selected effective porosity of 0.001 for all bedrock layer cells (regardless of
fracture characteristics such as openness and connectivity) would affect the
contaminant transport calibration and results.
12. Section 4.9, Transport Model Calibration Targets. 70 wells were used for COI calibration. For
any well not used as target well for COI concentration calibration, the report should state why.
13. Section 5.1, Flow Model Residual Analysis. The report states that each model layer was
subdivided into hydraulic conductivity zones and that all calibration was done by trial and error.
Section 2.1 acknowledges the "very large degree of heterogeneity in the bedrock due to the
distribution and characteristics of fractures and fracture zones". It is noted that measured hydraulic
conductivities vary over four orders of magnitude from near zero to 125 ft/d in bedrock, and
numerous locations within the model domain do not have wells for slug testing.
a. The report should explain whether and how the distribution of simulated hydraulic
conductivity "zones" was field verified with corroborating slug test data. For example,
significant adjustments were made to 3-dimensional hydraulic conductivities during
the transient pump test analysis to get simulated heads to match observed heads (page
18 to 20). Because bedrock boron concentrations are so difficult to accurately simulate
across time and space given the presence of known and unknown subsurface
heterogeneities, the report should point out on a map which zones were not verified by
slug testing or similar field data, and should present the results of sensitivity analyses
that show how an incorrect shape, size, or magnitude would affect the required
contaminant transport predictions discussed in comment 2 above.
b. Known and unknown (that is, as yet unmeasured) stratigraphic heterogeneities also
contribute to modeling uncertainties, as acknowledged in Section 3.2 which states that
the flow model was challenging to run due in part to "layers that outcrop in the model".
The report should explain why subsurface heterogeneities create conceptual and
modeling uncertainties for which the CAP reviewer must account (for example, in
many areas of the model domain heterogeneities which can affect boron concentrations
exist but are unknown and un -simulated).
14. Quantitative sensitivity analyses are expected in the CAP for all flow model variables, including
recharge, draintriver conductance, and hydraulic conductivity zone shapes and magnitudes.
15. Quantitative sensitivity analyses are expected in the CAP for all transport model variables,
including source concentration, sorption coefficient, effective porosity, and dispersivity. The
report should discuss the limitations of using a single sorption coefficient (per COI) for all
layers given that sorption would be expected to vary across space and with depthtflow systems.
APPENDIX D - UPDATED GEOCHEMICAL MODELING REPORT
The purpose of the geochemical modeling report is to explain current concentrations and predict fixture
(post -remediation) concentrations of constituents of interest whose mobility is controlled largely by
geochemical conditions. Each of the following comments should be incorporated into a revised
Page 6 of 11 July 27, 2017

geochemical modeling report that supports the accelerated remediation and (or) the corrective action
plans for the Asheville Plant.
1. The report would benefit from the development of conceptual geochemical models for the two
Northwest and West Transects that start in the ash basin and extend downgradient from the source.
The geochemical models should explain the interactions between the groundwater and the solid
phases in the aquifer that produce changes in pH, Eh, and groundwater concentrations. These
processes could then be incorporated into PHREEQC geochemical models to demonstrate the
ability to simulate current site conditions along flowpaths. The validated models could then be used
to estimate how conditions might change in response to the proposed accelerated remedial
alternatives. This would provide support for the qualitative conclusions developed in the report
from evaluations of Asheville site data and data from other similar facilities.
2. Page 1-1, first paragraph. "Boron, chloride, cobalt, iron, manganese, selenium, sulfate, TDS, and
vanadium are the primary focus within the focused area of interest. This report highlights five
select constituents from that list: boron, chloride, iron, manganese, and sulfate. These select
constituents were chosen because of their elevated mobility and concentrations in the area of
interest at the Asheville Plant." Given their elevated levels, why were cobalt, selenium, and TDS
not also evaluated?
3. Page 2-1, last paragraph. In the general observations of pH/Eh/DO, it would be useful to point out
that some of the highest pH values and lowest Eh/DO values are associated with samples of the
ash pore water. This represents the source area of potential groundwater contaminants and needs
to be considered in developing the conceptual geochemical model for this ash basin.
4. Page 2-3. In the discussion of groundwater concentrations of select constituents, it would be
beneficial to relate concentrations not only to pH, but also to locations relative to the ash basin
(background, ash porewater, downgradient near and far from a basin).
5. Page 2-3, third bullet.
a. "Chloride and sulfate concentrations are highly variable but generally increase with pH
(Figure 2-4)." The increasing trend with pH is not apparent in the figure, especially for
sulfate, which has its highest concentrations at the lowest pH values.
b. "With a majority of the B and Mn species forming anionic complexes..." In the pH range
shown in Figure 2-5, the dominant boron species is B(OH)3, which is a neutral species,
and the dominant manganese species is likely Mn2+, which is a cation, not an anion. Note
that the last bullet on this page states that Mn is dominantly Mn(II) present as a cation. The
B and Mn Pourbaix diagrams shown later in the report do not show anions as the dominant
species in the pH range of interest.
6. Page 2-3, fourth bullet. "Below pH 5, sorption of the neutrally charged 1-1313O3 or anionic H213O3
complexes likely reduces the aqueous concentration (Figure 2-5)." This figure shows that at pH
and below, most of the dissolved boron concentrations are similar to those at higher pH values
(near 1,000 ug/L). It is not apparent that boron adsorption is having an effect on dissolved
concentration. The pKal for H313O3 is 9.2 meaning that at a pH of 5 the concentration of H2BO3
Page 7 of 11 July 27, 2017

is about 10,000 times less than that of H3BO3. Consequently, anion adsorption is probably not an
important attenuation process for boron at acidic pH values.
7. Page 2-5. How do groundwater flow directions compare to the two chosen transects? In other
words, is a flowpath being followed from the ash basin to downgradient locations? If not, then
comparing two locations along a transect is not useful in identifying geochemical reactions in the
aquifer.
8. Page 2-6, first paragraph. "In general, along transects at the Asheville Plant, pH and Eh increases
slightly (Figure 2-6 and Figure 2-7)." The observation that the pH increases along the transects is
apparently based on comparing pH values in the saprolite beneath the ash basin with those
downgradient. The opposite conclusion would be reached comparing ash pore water pH with
downgradient pH values. In these cases, the pH decreases away from the basin. Because the ash
pore water better represents the contaminant source than the saprolite, comparisons to ash pore
water characteristics better represent contaminant flow paths.
9. Page 3-2, first paragraph, last sentence. "...conditions that could contribute to sustained
concentrations include dispersion and a lack of sorption or precipitation onto mineral surfaces."
How does dispersion contribute to sustained concentrations? Dispersion along a flowpath usually
results in lower concentrations.
10. Page 3-2, second paragraph, second sentence. "While boron is an anion and will be more attracted
to mineral surfaces at lower pH values..." As shown in Figure 3-1, boron exists in groundwater at
this site primarily as the neutral species H313O3. The anionic species of boron is not important
from an adsorption standpoint because of its low concentration in groundwater.
11. Page 3-6, Figure 3-6. This figure plots boron adsorption onto HFO and HAO versus the aqueous
anionic boron species H213O3. Why wasn't the dominant aqueous boron species H3BO3 used in
this figure? Was total aqueous boron used to calculate boron Kds or just the 11213O3 species?
12. Page 4-1, second paragraph. The mineral ferrihydrite is incorrectly spelled as ferrihydrate. Also
hematite rarely, if ever, precipitates from groundwater; thus, it does not control the aqueous iron
concentration.
13. Pages 4-1, last paragraph and 4-3, first paragraph. "The higher Fe concentration (in CB -06) can be
attributed to the more acidic conditions (pH 3.6) present during the January, 2016 sampling event."
"The groundwater fluctuations in the alluvial floodplain around the time of pH decrease were
significant enough to cause soil aeration resulting in sulfide oxidation." Where is monitoring well
CB -06? Will groundwater extraction associated with remediation likely lower the water table a
similar amount, also potentially resulting in the production of acidic conditions in the aquifer?
14. Page 4-2, Figure 4-1. Comparing the locations of the global Eh values on this figure with the
locations on Figure 3-1 (page 3-1) for boron, it appears that the Eh values for the iron figure are
significantly lower than shown on the boron figure. For example, Figure 4-1 shows one Eh value
near the lower water stability boundary, whereas Figure 3-1 does not show any values near this
boundary.
Page 8 of 11 July 27, 2017

15. Page 4-3, second paragraph. "However, if cobalt concentrations are not sufficiently high, hematite
(a - Fe2O3) is expected to be the dominant phase (Figure 4-4)." As mentioned in comment 410,
hematite will not typically form in aquifers. Ferrihydrite is the more common precipitate that limits
dissolved iron. Hematite should not be included as a potential iron -forming mineral on the Pourbaix
diagrams, including Figures 4-4 and 4-5.
16. Page 4-3, last paragraph. "For ferrous iron minerals, Ferroselite is predicted using the components
in Table A-12 with the addition of 1 x 10 - 6 mol/L total Fe (Figure 4-4)." Provide the formula for
the mineral ferroselite (FeSe2) to show that it is an iron -selenium mineral.
17. Page 4-5, only paragraph. "For both transects associated with the 1964 ash basin at Asheville
(Figure 4-6), iron exhibits a decrease in concentration with distance from the ash basin." Discuss
the possible geochemical reasons for this decrease in concentrations. What processes are
occurring? Can they be related to pH/Eh changes and mineral control on dissolved iron? What iron
minerals have saturation indices close to zero for the water samples? Why are the transition zone
samples so high in iron at the downgradient end of the West Transect?
18. Page 5-1, second paragraph.
a. "Under the Eh and pH conditions of the groundwater at seven sites evaluated for global
consideration, Mn(II) is the dominant oxidation state (Figure 5-1)." Figure 5-1 does
not show any global data.
b. The titles for Figures 5-1 and 5-2 appear to be reversed.
19. Page 5-1, last paragraph. "It is noteworthy that the Pourbaix diagrams leading to these conclusions
are based on a system with a total manganese concentration of 0.5 ppm (9 x 10-6 mol/L)." Why
weren't actual groundwater Mn concentrations used to create the Pourbaix diagrams for
manganese? The Figure 5-2 plot of monitoring well locations would be more useful if actual well
data were used for all the sampling parameters.
20. Page 5-3. Transect Model Analysis. Suggest calculating manganese mineral saturation indices with
PHREEQC using actual groundwater data for each sampling location to determine if any Mn
minerals might be in equilibrium with the groundwater and potentially limiting dissolved
concentrations at some locations.
21. Page 6-1, last paragraph. "It is noteworthy that the Na+ and Cl- concentrations were fixed at
mmol/L each when generating the Pourbaix diagrams (Figure 6-1)...". How does this chloride
concentration compare with actual groundwater levels? Why weren't actual levels used in the
plots?
22. Page 6-2, first paragraph. "Chloride concentrations at the Asheville site may be related to the
recently decommissioned, constructed wetlands systems in the northwest portion of the 1964 basin
formerly used to treat FGD water." Why would decommissioning of the wetlands systems cause
chloride levels to increase in groundwater? Are there any data showing elevated chloride in these
wetlands?
Page 9 of 11 July 27, 2017

23. Page 7-1, second paragraph. "Thus is can be expected that Se(VI) will be present in the site
groundwater under high reducing conditions." Se(VI) will be present under highly oxidizing
conditions, not reducing conditions.
24. Page 7-2, Figure 7-1. The first note for this figure needs to be corrected to specify the appropriate
diagrams.
25. Page 8-1, third paragraph. "When precipitation of mineral phases is allowed in the GWB
simulations, the mineral barite (BaSO4(s)) is predicted to be a dominant mineral phase under
oxidizing conditions." It is highly unlikely that barite would limit dissolved sulfate concentrations.
Its presence on the Pourbaix diagram is an artifact of the choice of an artificially high barium
concentration (0.5 mg/L) in the groundwater.
26. Page 8-3, last bullet. "Thus, in groundwaters with sufficiently high concentrations of Ba+2,
formation of barite could limit the aqueous concentration of sulfate (Figure 8-5)." Because the
barium source to groundwater is almost always much less that the sulfate source, barite formation
usually limits the barium dissolved concentration, not the sulfate concentration. Gypsum is the
more common mineral that limits dissolved sulfate.
27. Page 8-6, only paragraph. "Similar to chloride, sulfate concentrations in GW -3 increased from the
January (50 mg/L) to the April (3 10 mg/L) 2016 sampling events." Provide reasonable scenarios
for this increase in sulfate concentration.
28. Page 9-1, first paragraph. "As discussed above for both the flow transect models and the global
model, the pH and Eh of the system are the primary factors which could increase or decrease the
mobility of a constituent. Removal of ash from the basin has the potential to increase the redox
potential of the system through the introduction of oxygen. It is unclear how ash removal could
influence the pH. However, based on observations in wells surrounding the 1982 ash basin before,
during and after its dewatering and ash removal efforts, the pH is not expected to have any
significant change post remedial activities and the Eh is expected to fluctuate with only a slight
increase overall." Comments on these assumptions:
a. Figure 2-6 shows that the measured ash pore water pH (7 to 8) is much more alkaline
than that of groundwater in the underlying saprolite (pH about 4). It is likely that the
ash pore water is neutralizing the groundwater in the aquifer immediately beneath the
ash, thereby producing the more neutral pH values (5 to 6) in transition zones
downgradient from the ash basin (Figure 2-6). Removing the ash may reduce this
neutralizing effect and lower the pH in the aquifer to significantly more acidic
conditions beneath and downgradient of the ash basin. This possibility should be
considered in predicting the mobility of aquifer constituents of interest that may be
mobilized by removing the ash.
b. Page 4-1 discussed the occurrence of acidic groundwater (pH 3.7) measured in CB -06
and attributed the low pH values to exposure of soil sulfide minerals to molecular
oxygen that oxidized the sulfide to sulfuric acid. If sulfide minerals are present beneath
the ash basins because of low Eh conditions caused by the ash, wouldn't removal of
the ash expose those sulfide minerals to molecular oxygen producing sulfuric acid and
low pH conditions?
Page 10 of 11 July 27, 2017

c. The presence of low Eh conditions (50 to 100 mV) in the ash porewater was shown in
Figure 2-7. It is possible that removing the ash will have a significant effect on Eh in
the aquifer immediately beneath the ash basin and should be considered in estimating
future geochemical conditions and constituent mobility.
29. Page 9-1, paragraph 2. "A groundwater extraction well system will result in enhanced groundwater
removal which will generate a hydraulic gradient and bring in waters from finther up the flow path
at a faster rate." The gradient generated by groundwater removal will be the result of lowering the
water table at the extraction wells. How much will the water table be lowered? Exposing the
aquifer in the drawdown area to soil gas with molecular oxygen may result in a significant increase
in Eh. What will be the effect on constituent concentrations of dewatering a portion of the aquifer
beneath and adjacent to the ash basin?
30. Page 10-1, first bullet. "Boron, chloride, manganese, and sulfate are relatively non-reactive species
under the Eh and pH conditions in the area proposed for accelerated remediation." Sulfate may not
be non-reactive in the area of accelerated remediation if the ash basin has created subsurface
reducing conditions that have allowed for the formation of sulfide minerals. If sulfide minerals are
present in the remediation area and those minerals are oxidized by lowering the water table or
bringing in oxygenated groundwater from upgradient, then the sulfate concentration may increase
considerably and the pH may become highly acidic. If acidic conditions are generated, manganese
and iron minerals will become more soluble resulting in higher Mn and Fe groundwater
concentrations.
31. Page 10-2, second bullet. "TDS at the Asheville site is dominated by sulfate along the west transect
and chloride along the northwest transect." It is not clear why TDS is mentioned in the Summary
section. TDS is mentioned in the Introduction (page 1-1) as a component of primary focus, but is
not discussed in the main body of the report. Clarify the importance of TDS to this discussion.
32. Page 10-2, first paragraph. "The two primary remedial alternatives in the area of interest at the
Asheville site are coal ash removal from the 1964 basin and installation of a groundwater extraction
system northwest of the 1964 basin. Given these proposed alternatives, pH is not expected to have
any significant change post remedial activities while Eh is expected to show a slight increase
overall." Comments on these conclusions:
a. Add a caveat that, if sulfide minerals are present in the remediation zone, oxidation
may have a large impact on pH.
b. What is meant by Eh is expected to show a slight increase? How big an increase might
be expected? What is the basis for estimating only a slight Eh increase?
Page 11 of 11 July 27, 2017