Loading...
HomeMy WebLinkAboutResponse to DEQ letter dated 8-12-2019 - Compiled•� DUKE �.ENERGY® October 7, 2019 Mr. Steve Lanter North Carolina Department of Environmental Quality 1636 Mail Service Center Raleigh, North Carolina 27699-1636 Paul Draovitch Senior Vice President Environmental, Health & Safety ano Operations Support 526 S. Church Street Mail Code: EC3XP Charlotte, NC 28202 (980) 373-0408 Subject: Responses to North Carolina Department of Environmental Quality Letter Dated August 12, 2019 — Interim Monitoring Plan Analysis for Dissolved Constituents and Metals Speciation Dear Mr. Lanter: Duke Energy is in receipt of the above -referenced letter from the North Carolina Department of Environmental Quality (NCDEQ) which requests additional groundwater monitoring well analysis for dissolved constituents along with metals speciation. Duke Energy and our geochemistry experts have carefully reviewed NCDEQ's request in the context of the extremely large database available for each site and the extensive geochemical modeling that has been performed in collaboration with the NCDEQ's geochemical expert. The results of this review and discussions with NCDEQ were factored into the Interim Monitoring Plans (IMPs), which were submitted to NCDEQ in late 2018 and updated in first quarter 2019. Generally, the most current NCDEQ approval for the IMPs is dated April 4, 2019. Below is the NCDEQ request from the August 12, 2019 letter in italic font along with the Duke Energy response in normal font. NCDEQ Comments: DEQ conducted a review of the optimized IMPs and related submittals to determine if the necessary data are being collected to support remedial design for corrective action. Upon review of recent submittals, DEQ staff have identified some discrepancies between the approved IMPs and the actual data being collected and analyzed. For example, Duke Energy is no longer collecting and analyzing dissolved samples at some wells for some COls included in the IMP sampling program. Further review of the current groundwater monitoring program indicates speciation ofgroundwater samples for metals has not been performed since 2015. Concentrations of total and dissolved constituents in combination with speciation analytes are necessary in the development of geochemical models that will be used to understand contaminant occurrence, oxidation state, and mobility undercurrent (pre -closure) conditions and under the site conditions that may occur during and after decanting, dewatering and source control measures have been implemented. To address data needs, DEQ requests the following action to begin with the next IMP Quarterly sampling event: • Revise the existing IMPs if needed for each facility to ensure that dissolved samples are collected and analyzed at wells along the geochemical model transects, selected background locations, and selected downgradient wells thatexhibitexceedances of the 15A NCAC 02L .0202 groundwater quality standards. • Collect total and dissolved samples and perform speciation for the following constituents and oxidation states: Fe(II, Ill), Mn(11,111, IV), As(111,V), Cr(III,VI), Se (-II,IV,VI) at wells along the geochemical model transects, selected background locations, and selected downgradient wells that exhibit exceedances of the 15A NCAC 02L .0202 groundwater quality standards. Duke Energy Response: Potential Discrepancies in IMP Sampling - Duke Energy reviewed the IMPS for the individual sites and did not identify any discrepancies between the approved IMPS and the actual data being collected, analyzed and submitted to the NCDEQ. In some cases, Duke Energy collects additional groundwater samples on a voluntary basis that may differ from the IMPS. These voluntary data are also submitted to the NDCEQ. If there are any specific discrepancies that NCDEQ has observed, please provide this information to Duke Energy so we can work with the NCDEQto resolve any issues. Duke Energy is aware that the IMPs can be confusing to follow due to wells that are sampled at different frequencies and for different constituent lists. Duke Energy plans to update the IMPS to make improvements so that Duke Energy and NCDEQ are both comfortable with them. Duke Energy plans to update the IMPS by end November 2019 so that the NCDEQ can review and approve the IMPs so they are ready for implementation at the start of 2020. Groundwater Sampling for Dissolved Constituents — The following summarizes Duke Energy's planned approach in response to the NCDEQ's letter regarding sample collection and analysis for dissolved constituents. This approach was verbally discussed with NCDEQ at a meeting at the Mooresville Regional Office on September 4, 2019. Duke Energy along with our consultants, SynTerra Corporation and Rosewater Geochemical Modelling, LLC., have evaluated existing and proposed future sampling events and have determined what additional or continued data are necessary for development of the geochemical models (e.g. collection of both total and dissolved [0.45-micron filtered] groundwater monitoring well samples). The following approach has been developed in response to the NCDEQ request: • New wells (i.e., those installed after December 2018) will be sampled for a minimum of four quarters and analyzed for total and dissolved (0.45- micron filtered) constituents for the full CAMA parameter list (note there are other constituents that Duke Energy voluntarily samples for to provide data to our geochemical modelling team). The four rounds of sampling assists in the evaluation of constituent stability and provides a baseline dataset. • Existing wells (i.e., those with historical IMP data including both total and dissolved samples) will be sampled on a site -wide basis for total and dissolved (0.45- micron filtered) constituents either the first or second quarter of each year to monitor potential changes in groundwater conditions, and to accommodate any future geochemical modeling efforts, as needed. This 2 population of wells will include all wells along geochemical transects (i.e., those included in or associated with the PHREEQCII-D advection models). Speciation Sampling —This request was reviewed by Duke Energy's geochemistry experts and the following response is provided. Additional technical details in support of this response is provided in Attachment A to this letter. Speciation sampling was completed during 2015 as part of the groundwater assessment plan. Duke Energy does not believe that additional speciation sampling will be informative beyond the extensive data already being collected at the Duke Energy coal ash sites and site -specific geochemical modeling performed for each site. The need for speciation sampling can be addressed again on a targeted basis should it be required to better understand changes observed in groundwater concentrations. This approach was shared with the NCDEQ and their geochemical expert during our meeting at the Mooresville Regional Office on September 4, 2019. If you have any questions on the enclosed information, please contact Ed Sullivan at 980-373-3719. Sincerely, PatA Draovitch Senior Vice President Environmental, Health & Safety Attachment: Geochemical Modeling Discussion on Metals Speciation — September 2019 cc: Ed Sullivan — Duke Energy Scott Davies —Duke Energy Kathy Webb — SynTerra Corp. Jim Wells - Duke Energy Matt Hanchey — Duke Energy 1 PHREEQC- United States Geologic Survey (USGS) equilibrium modeling program with the original acronym pH-REdox-EQuilibrium written in C programming language ATTACHMENT A Technical Response to the North Carolina Department of Environmental Quality's 8/12/19 Response for Metals Speciation Sampling and Analysis Prepared by: SynTerra Corporation and Rosewater Geochemical Modeling, LLC. 9/17/19 On August 12, 2019, the North Carolina Department of Environmental Quality (NCDEQ) sent a letter to Duke Energy requesting, in part, groundwater sampling and analysis for metals speciation, specifically for: Fe(II, III), Mn(II, III, IV), As(III, V), Cr(III, VI), Se( -II, IV, VI). In this Attachment, metals speciation is discussed in terms of how results are used in the development of the geochemical model as well as the limitations inherent in sample collection and analysis and the equilibrium modeling approach. The structure of the current geochemical model - using the aqueous speciation and surface complexation modeling program, PHREEQC — was developed assuming that the groundwater system being modeled is in equilibrium (Note: the second "E" in PHREEQC denotes equilibrium and the model assumes equilibrium conditions are present). Often, geochemical systems are not present at equilibrium but are along a reaction path towards an equilibrium state (Zhu, 2009: Stumm and Morgan, 1996; Tinnacher and Honeyman, 2010). This summary describes how potential non -equilibrium conditions can be accounted for using PHREEQC predictive models. Additionally, collecting reliable field samples to preserve the oxidation/reduction (redox) conditions and constituents of interest (COI) speciation is profoundly difficult because of the potential exposure of the samples to atmospheric conditions during sample collection and processing. Therefore, redox speciation is best evaluated from expert judgement using a combination of; 1) the geochemical models to see the constituents and likely species present and note the thermodynamic driving forces and direction of change of the system, and; 2) observations of changes in constituent concentrations and geochemical conditions over time. Therefore, collection of total and dissolved constituent concentrations periodically is important and is supported by Duke Energy as described in the letter response for this attachment. If there are observed changes in constituent concentrations, the geochemical models can be utilized to determine if the change in concentration is consistent with the predicted chemical speciation. Further field data collection may occur if the model simulations cannot explain the observed data. The geochemical models use an equilibrium -based aqueous speciation and surface complexation model (PHREEQC) to simulate the geochemical behavior of multiple constituents related to the coal ash impoundments. The geochemical model can be used to evaluate the behavior of COls in terms of the likely chemical species, dominant attenuation mechanisms, and overall potential for groundwater migration for current, and estimated future conditions at each site. Site -specific data have been evaluated to identify trends in geochemical behavior for each COI, including: • Aqueous concentration of constituents in both ash pore water and groundwater; Total digested and extractable solid phase concentrations of constituents for each flow zone (e.g., ash, shallow/surficial, deep/transition zone, and bedrock for the Piedmont physiographic region); Solid phase mineralogy/lithology data; and Leaching potential from source material or coal ash using the United States Environmental Protection Agency (USEPA) Method 1313 and Method 1316 analysis included in the USEPA-Leaching Environmental Assessment Framework (LEAF) to evaluate the leaching potential of inorganic constituents over a range of pH values (pH 3 to pH 11) and liquid -to -solid ratios (ranging from 0.5 to 10 mL/g-dry ash). Model simulations are performed using both the batch and the 1-D advective transport modes available in PHREEQC. As discussed in detail below, the inherent assumption of equilibrium in the PHREEQC simulations may lead to differences between predicted and observed oxidation/reduction conditions. An example of this disequilibrium is provided in this attachment. Despite this potential limitation, the model provides an assessment of the equilibrium conditions of the system that can be used in conjunction with observations from actual site field data to assess deviations from equilibrium. This concept and approach was discussed between Duke Energy and our Geochemical Modelling Team along with staff from the North Carolina Department of Environmental Quality (NCDEQ) and Mr. William Deutsch (external reviewer for NCDEQ) in meetings dating back to 2015 and have acknowledged agreement with this approach. The question of evaluating speciation can be complicated, but it is also not always necessary to predict the concentrations of a specific COI. As an equilibrium -based model PHREEQC does not consider reaction kinetics. Therefore, all chemical reactions are assumed to reach equilibrium, including oxidation-reduction and mineral precipitation -dissolution reactions, which can be kinetically limited in many cases. For example, naturally occurring pyrite (FeS2) is commonly found in surface water streams. When in contact with atmospheric 02(g) which can dissolve in water to form dissolved oxygen (02(aq)), an equilibrium model would predict that essentially all the pyrite would oxidize to form ferric iron (Fe(III)) and sulfate (SO4 2), leaving no pyrite in the system. In a surface water where pyrite minerals are observed, this reaction is clearly not reaching the predicted equilibrium state. The equilibrium model provides information on the expected equilibrium endpoint of the system but does not provide information on how quickly it will reach that state. The aqueous complexation, sorption/desorption, oxidation/reduction, and precipitation/dissolution reactions considered in the PHREEQC model all occur at different rates, and therefore, will take different times to reach equilibrium. Aqueous complexation and sorption reactions are commonly observed to reach equilibrium on time scales of minutes to hours. Therefore, equilibrium -based models are appropriate in many cases, particularly where groundwater flow rates are slow and the system has longer time to reach equilibrium on a local level. In contrast, oxidation-reduction reactions may occur at slower rates because they involve complete electron transfer during the chemical interactions and potentially structural rearrangement of the species [Example: addition of oxygen to hexavalent chromium (Cr04 2) during oxidation of trivalent chromium (Cr")]. The structural rearrangement of mineral phases is a potential rate limiting step that will cause a system not to reach equilibrium in relevant time scales. For example, in a PHREEQC simulation initially containing aqueous ferric iron (Fe(IIII), the system will predict the mineral hematite (a-Fe2O3) to be the most thermodynamically stable, and therefore, predominant phase instead of the more soluble minerals ferrihydrite (a- Fe(OH)3)(s)) and goethite (a-FeOOH). However, under the conditions in the coal ash impoundments, ferrihydrite is often found to be the dominant iron mineral because of the additional energy and time required to form goethite or hematite. Therefore, the model simulations are written using a feature in PHREEQC to not allow goethite or hematite to form so that constraint provides a better description of the actual system. This minimizes the influence of the equilibrium assumptions for mineral precipitation/dissolution. A similar approach cannot be taken for the redox speciation. The different oxidation states of iron, manganese, arsenic, chromium, and selenium must be coupled in the model so that the distribution K between each oxidation state can be simulated based on the total concentration in the model. Therefore, the redox speciation is included as part of the PHREEQC simulations and is evaluated with expert judgement in comparison with observations of total and dissolved constituent concentrations over time. Conrlusion The above discussion shows the geochemical modeling approach used is very useful to understand site groundwater conditions, but the model results do not comprehensively provide all the answers. Expert judgment is required in many cases to interpret the results. As long as the PHREEQC model simulations are combined with expert judgment (based on existing empirical evidence and extensive literature review) are acceptably predicting the concentrations of redox active constituents, there is no need for additional speciation sampling. Speciation information was useful in original model development and site characterization, but the use of this data in our future models is not required unless site measurements cannot be adequately explained with the current approach. Therefore, additional speciation sampling is not proposed. Suaalemental Information Example of redox diseguilibrium using data from Sutton (Powell, 2016) To evaluate the potential for redox disequilibrium in the field samples, the expected redox potential (i.e., pe and EH) for each groundwater sample was calculated based on the Fe(II) and Fe(III) ratios. Using the reaction shown in Equation 1 for Fe(II) oxidation to Fe(III) and the equilibrium expression shown in Equation 2. Fe++ + 2H2O H Fe(OH)Z + 2H+ + e- log — 18.78 Equation 1 [Fe (OH)z ] [H+] 2 [e ] [Fe++] = equilibrium constant (K) = log —18.78 Equation 2 where [e] is the electron concentration in the system which is more commonly noted as pe (pe =-log[e- ]). This reaction has been written in terms of Fe++ oxidation to Fe(OH)z+ which are the expected aqueous species at the pH values under consideration. While linking this explicitly to the ferrihydrite concentration in the system could provide a more realistic estimation of the redox potential due to the low solubility of Fe(III) species, Fe(OH)z+ was used as the aqueous species because the exact ferrihydrite concentration is not known (note that extractable iron concentrations are available but the fraction of a specific iron mineral is unknown). Iron speciation is used for this calculation because the dissolved iron concentrations are relatively high compared to other redox active species and it is likely that iron would be a significant redox buffer in these systems. Also, this iron couple is known to be more likely at or near equilibrium than many other couples. Taking the log of the Equation 2, the expected pe value based on the ratio of Fe(II) to Fe(III) can be calculated as (Equation 3): Equation 3 [Fe(OH)Z log [Fe++] ] — 2pH + 18.78 = pe 41 The calculated and measured EH values are shown in Figure 1. From these data, it is clear that the expected EH values based on the iron redox couple are greater than the measured values in many cases and lower in some. Thus, iron is either 1) not the dominant redox buffer in this system causing the measured EH values, or 2) iron speciation is not present under equilibrium conditions. Soo 700 •� 600 -- 500 2 W 400 L 300 i 100 IC 4 140 Zoo 300 400 500 500 700 800 Estimated EH (mV) based on Fe(II)/Fe(III) couple Figure 1: Estimated EH values based on the Fe(II)/Fe(III) redox couple compared with the measured values in groundwater samples from the L.V. Sutton site (Powell, 2016). The solid black line represents the location where measured and estimated values would plot if in perfect agreement. 4 References Powell, B. A. (2016). Analysis of geochemical phenomena controlling mobility of ions from coal ash basins at the Duke Energy L. V. Sutton Energy Complex. Wilmington, NC. Summ, W.; Morgan, J.J., Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters. 3" ed.; Wiley-Interscience: Chicago, 1996; P 1040. Tinnacher, R.M.; Honeyman, B.D., Theoretical analysis of kinetic effects on the quantitative comparison of K-d values and contaminant retardation factors. Journal of Contaminant Hydrology 2010, 118, (1-2), 1-12. Zhu, C., Geochemical Modeling of Reaction Paths and Geochemical Reaction Networks. Reviews in Mineralogy and Geochemistry 2009, 70, 533-569. 61