BPM006010_ATTACH_12_Method_of_Approach.pdf

Request for Proposal

Solicitation No. BPM006010

ASRAC Assessment and Remediation of Hazardous Substances

Arizona Department of Environmental Quality (ADEQ)

1110 W. Washington Street Phoenix, AZ 85007

Solicitation Attachment – Method of Approach | Revision No. 1 | Revised 11-22

Attachment 12 – Method of Approach

The Offeror shall provide a narrative response to each section that demonstrates their understanding of the Scope of Work requirements and describes the company’s overall method of approach for providing the Service(s) stated in this Solicitation.

Offeror shall provide a response either after each question or in a separate document entitled: “Attachment - Method of Approach.” If a separate document is used, please clearly designate each section and double-check to ensure that each question has been answered. If any information provided is considered CONFIDENTIAL by Offeror, the information must be specifically included in the Confidential Information Designation Attachment and related documentation to be assessed for confidential treatment under A.A.C. R2-7-103.

Responses shall be limited to one (1) page per question. Response to Question #5 may be up to two (2) pages, in font size 10 pt or greater.

Question 1: Describe your company’s quality control/quality assurance process for reviewing proposals and technical documents prior to submission to ADEQ. What deliverables can ADEQ review to verify QA was completed? If ADEQ finds a quality issue, who, other than the PM, is the point of contact for a remedy?

Offeror Response:

CDM Smith’s quality management system (QMS) is integrated into a program we call the project lifecycle that guides all aspects of planning and execution on every project. Our QMS conforms with the international QMS standard, International Organization for Standardization (ISO 9001:2015), which is an internationally recognized QMS standard and is intended to be applicable to all organizations, regardless of size, type, and product or service provided.

Project Deliverable Review CDM Smith implements formal and systematic reviews of proposals and project deliverables prior to submittal. Reviews are conducted by qualified personnel. This includes a technical specialist review process and a project technical leader (PTL) program that are designed to meet the need for review of deliverables. The deliverable preparer is excluded from completing this review. Reviews are conducted according to the project schedule and include:

• Independent checking that includes detailed internal checking of all calculations, reports, data, memoranda, drawings, and specifications is completed by a qualified person independent of the document preparer.

• Technical specialist reviews are independent reviews of deliverables for technical accuracy, accomplishment of objectives, and clarity of presentation by qualified reviewers. Technical specialist reviews also check for evidence that independent checking of equations and calculations, reference citations, and tables or figures have been performed. Technical specialist reviewers possess appropriate technical qualifications and are independent of the document being produced. The PTL is responsible for verifying technical specialist reviews are scheduled, conducted, and comments resolved.

• All engineering design projects, and all large non-design projects such as large environmental investigations, undergo a Technical Review Committee (TRC), which is a collaborative, independent review by senior personnel with technical backgrounds aligned with the project scope. Subjects addressed during the TRC process include adequacy and appropriateness of technical data, pertinent application of technology, and whether a sound, practical, and cost-effective approach has been made. The PM is responsible for ensuring that all TRC follow-up actions are completed and documented.

• At the appropriate time, a final review is completed to verify that deliverables meet contract requirements. This review is conducted by the PM or other nominated person other than the preparer.

• Subcontractors, subconsultants, and suppliers are expected to comply with the requirements of their respective agreements with CDM Smith. The PM is responsible for verifying this occurs.

Arizona Department of Environmental Quality (ADEQ)

• Deliverables that do not conform to client and/or CDM Smith requirements must be revised. After a nonconforming deliverable is revised, it is subjected to the same reviews and checks as the original deliverable.

• Corrective actions can be identified during a variety of activities and may arise from nonconformance found during review and checking, internal or external audits, lessons learned, or client complaints. All employees are empowered to identify quality problems or deficiencies and make suggestions for preventive actions.

CDM Smith maintains evidence of conformance with our QMS in project files, including independent review forms documenting completion of independent checking, technical specialist review, and quality assurance and/or editorial reviews as necessary for the project. Following TRC completion, meeting minutes and documentation of review comment incorporation are developed prior to deliverable submittal and available to ADEQ upon request.

In the event that ADEQ identifies quality concerns or problems with project deliverables, there are multiple points of contact to remedy the situation. Other than the PM and PTL—who are responsible for ensuring that the QMS and required reviews are completed—Principal-in-Charge Carlos Ortiz has the authority and responsibility to rectify any issues related to project quality which arise during execution of the work.

Additionally, CDM Smith has identified a quality assurance representative in the organization chart for work under the program, Robert Alexander, to oversee implementation of our quality program. Mr. Alexander has served in a quality assurance specialist and reviewer role and field auditor for our federal and public sector clients for nearly 20 years. The QA/QC representative works independently of the project managers and has authority to communicate with the project team to identify necessary remedies including correction and review of non-conforming deliverables, identification of root causes of quality issues, and implementation of corrective actions to mitigate identified issues.

Question 2: Describe the methods and criteria your company would propose for confirming that a soil vapor extraction system located at a former dry cleaner in downtown Phoenix could be turned off and demobilized from the site. Consider both vapor and direct exposure. How does the Arizona “Soil Rule” apply at closure?

Offeror Response:

Property in downtown Phoenix with a mature soil vapor extraction (SVE) remedy could be of significant interest for redevelopment. However, prior to turning off and demobilizing an SVE system, it would be necessary to collect representative soil and vapor samples to confirm sufficient remedial progress.

For soils, for an unrestricted closure, it would be necessary to demonstrate that the SVE system had achieved an endpoint that is compliant with Arizona Administrative Code (A.A.C.) Title 18, Chapter 7, Article 2 (the “Soil Rule”).

Specifically, representative soil samples would have to below background concentrations, pre-determined risk-based soil remediation levels (SRLs), or site-specific remediation standards (if established for the site) to demonstrate that direct exposure would not constitute unallowable risk. If the system had not achieved the standard SRLs, it may be worth evaluating whether site specific remediation standards would be appropriate.

It would also be necessary to collect indoor vapor samples from any potentially impacted buildings to confirm no unallowable exposure. Sampling would be conducted using Summa® cannister(s), over a period of at least 24 hours.

Sampling would be done with a survey to identity the presence of any potentially confounding chemicals unrelated to subsurface contamination. Sampling would also only be appropriate after SVE had been turned off to allow sufficient time for potential contaminant rebound. If analytical results indicated concentrations below appropriate EPA screening levels (SLs), and those results aligned with at least one previous set of data to show no seasonal effects, it would be appropriate to move forward with closure. If concentrations remained above SLs, a site-specific risk assessment would be necessary to determine if exposure exceeded allowable EPA risk standards.

Once compliant soil and soil vapor samples were collected then it would be appropriate to provide the data to DEQ and request a Letter of Completion (R18-7-209). Once the Letter of Completion was received, the SVE system could be turned off permanently and removed from the site. However, in some cases, even with appropriately dense SVE well spacing, sufficient pneumatic control, plenty of pore volume exchanges over the life of the SVE system, and aggressive system optimization, some systems can fail to achieve the criteria necessary for unrestricted use.

Arizona Department of Environmental Quality (ADEQ)

In those cases, to turn off an SVE system, it would be necessary to pursue a Declaration of Environmental Use Restriction (DEUR) under A.A.C. R18-7-208 and Arizona Revised Statutes (A.R.S.) § 49-152 and § 49-158. A DEUR allows for the cessation of active remediation but requires use of sufficient engineering and/or institutional controls (ECs/ICs) such as membranes, sub-slab depressurization systems (SSDS), pavement, and land use restrictions. Prior to removal of the SVE system, it would be appropriate to implement the planned controls and confirm effectiveness (e.g., follow-up sampling to show a SSDS was successfully preventing vapor intrusion). The DEUR then would form a restrictive covenant that runs with the property until released by DEQ.

CDM Smith has pursued this strategy of aggressive SVE optimization followed by long-term ECs/ICs, when necessary, at multiple sites in southern California, such as the Omega Chemical Superfund Site in Whittier, CA. We understand that is much more appropriate to make a case that no further active remediation is necessary when it has been demonstrated that any potential receptors are protected for the long term, and that further active remediation will not produce significant remedial progress.

Question 3: Provide a table for a sampling and analysis plan for the implementation of an in-situ enhanced reductive dechlorination program for a period of 12 months. Assume the program includes two injection wells (IW-01 and IW-02), four performance monitoring wells (one well [PW-01] located 10-feet upgradient of the injection wells with the remaining three wells located 10- [PW-02], 20- [PW-03], and 30-feet [PW-04] downgradient of the injection wells), and one injection utilizing emulsified vegetable oil. Sample and analysis plan should include what parameters are to be collected for the design of the implementation and for remedy performance monitoring.

Offeror Response:

Designing an effective in-situ enhanced reductive dechlorination (ERD) requires an understanding of the groundwater hydrogeology, geochemistry, and VOC concentration profiles. During well installation, lithology will be recorded. Hydraulic tests will also be performed to understand groundwater hydraulics. Pre-injection groundwater monitoring will be performed to collect VOC and geochemical parameters to help understand baseline conditions in the aquifer.

In lower permeability aquifers, as many as two to three months may be required between well development and groundwater monitoring to obtain results that are representative. In some cases, a pH buffer capacity test may be required to ensure the pH in the aquifer will remain stable without added buffers. These data will be used to design the EVO injection, including injection volume, EVO dose, injection flow rates, and monitoring frequency. Effective injections of an EVO amendment into a VOC-impacted subsurface will result in significant geochemical and VOC changes. These changes require using multiple lines of evidence to confidently understand the ERD performance.

• Amendment Delivery Parameters – Although dissolved hydrogen is the most direct parameter for amendment, field methods are often unreliable. Therefore, TOC increases are anticipated in monitoring wells downgradient of the injection points upon injections of an EVO. However, because the emulsion of EVO breaks after a short time after injection, the oil will sorb to soil particles and aqueous TOC measurements will not be completely representative. Specific conductance is also a simple measurement method that can be effective. Lastly, tracers (aqueous or oil-soluble) can be blended with the injection water to estimate amendment delivery.

• Competing Electron Acceptors – Naturally occurring bacteria will compete with dechlorinating bacteria to utilize the injected amendment. These include bacteria that utilize dissolved oxygen (DO), nitrate, ferric iron, manganese, and sulfate. Therefore, monitoring the changes in concentrations of these competing parameters provide critical information to the geochemical state of the aquifer. Similarly, under strongly reducing conditions, methanogenesis often occurs where bacteria consume carbon dioxide, resulting in elevated methane concentrations.

• pH – A pH range between 6.5 and 8 is optimal for microbial growth and development. Naturally low (or high) pH aquifers and/or pH reductions due to EVO implementation will be measured to ensure optimal conditions for ERD.

• Oxidation-Reduction Potential (ORP) – An ORP value of less than -100 millivolts can be expected upon development of a highly-reducing environment.

Arizona Department of Environmental Quality (ADEQ)

• VOCs & Dissolved Gases – PCE, TCE, and their dechlorination by-products cis-1,2-DCE, VC, ethene and ethane are expected to decline significantly after substrate additions and bioaugmentation.

Monitoring Well ID

Pre-Injection Quarterly Monitoring (4 Quarters)

FP

VO

Cs

TO

C

DH

C

In or ga ni cs

DH

Gs

FP

VO

Cs

TO

C

DH

C

In or ga ni cs

DH

Gs

IW-01 (Injection well) X X X X X X X

IW-02 (injection well) X X X X X X X

PW-01 (Upgradient) X X X X X X X X X X X PW-02 (10' downgradient) X X X X X X X X X X X X PW-03 (20' downgradient) X X X X X X X X X X X PW-04 (30' downgradient) X X X X X X X X X X X

FP = Field parameters: groundwater elevation, dissolved oxygen (DO), oxidation-reduction potential (ORP), pH, temperature, specific conductance, turbidity, ferrous iron; VOCs = Volatile Organic Compounds; TOC = Total Organic Carbon; Inorganics = Alkalinity, anions (sulfate, nitrate, nitrite, and chloride), and total metals (iron, manganese); DHG = Dissolved Hydrocarbon Gases (ethane, ethene, and methane)

Question 4: Describe the multiple lines of evidence necessary to evaluate monitored natural attenuation (MNA) in a typical Arizona aquifer and your experience with the viability of MNA in the deep aquifers of Arizona (focusing on chlorinated solvents), including the limits on MNA in our hydrogeologic settings? Provide key references where appropriate (does not count against page limit). Describe why low-flow sampling would be preferable over passive sampling in MNA evaluation.

Offeror Response:

CDM Smith has implemented remedies which include MNA, as well as completed robust MNA evaluations during remedial investigations, at sites including the Bountiful/Woods Cross Superfund Site and 700 South 1600 East PCE Plume Superfund Site in Utah which are located in the Basin and Range physiographic province with similar characteristics to many deep Arizona aquifers. MNA may be considered as a remedy option at sites impacted with chlorinated solvents after thorough site characterization is performed and a conceptual site model is developed. According to EPA guidance (EPA 1999), lines of evidence necessary to evaluate MNA fall into three tiers.

1. Historical groundwater data at strategically placed monitoring points should demonstrate a decreasing trend of contaminant mass and/or concentration over time. This evaluation is conducted using statistical trend analyses for contaminant concentrations and evaluation of degradation product formation (PCE, TCE degradation, cis-1,2-DCE, vinyl chloride, ethene) (ADEQ 2014) to confirm decreasing trends are due to degradation rather than migration.

2. Hydrogeologic and geochemical data may also be used indirectly to understand the type of attenuation processes and rate of attenuation. For chlorinated solvents, geochemical processes of interest are dissolved oxygen, nitrate reduction, iron reduction and formation of ferrous iron, sulfate reduction and formation of sulfide, and methanogenesis. Measurement of total or dissolved organic carbon in groundwater is useful for understanding whether sufficient carbon is present to maintain degradation reactions.

3. If data from the first two tiers are inconclusive, additional studies may supplement the data to show directly that an attenuation process is occurring. Specific studies at chlorinated solvent sites include compound-specific isotope analysis (CSIA) to provide evidence of destructive processes occurring and estimate degradation rates, molecular biological tools to document presence and populations of chlorinated solvent-degrading microbes, and evaluation of aquifer materials for the presence of iron minerals capable of abiotic chlorinated solvent degradation. These techniques were developed based on work by CDM Smith engineers and scientists (Schaefer et al. 2018).

Arizona Department of Environmental Quality (ADEQ)

Assessing the viability of MNA in Arizona aquifers requires understanding how geochemical and hydrogeologic properties influence degradation of chlorinated compounds. Several deep aquifer characteristics of the Basin and Range physiographic region can affect MNA progress and are generally aerobic/oxygenated with low organic carbon content, limiting the ability of many microorganisms to degrade chlorinated solvents. While aerobic cometabolism can be a significant degradation mechanism tor TCE, cis-1,2-DCE, and vinyl chloride at some sites, PCE is recalcitrant under aerobic conditions (Dolinova et al 2017). Potential for aerobic cometabolism can be evaluated using molecular biological tools. Aquifers are also high in sulfate due to oxidizing conditions and hydrothermal or ore deposits in most areas, presence of gypsum and anhydrite evaporites at depth (Robertson 1991). These high levels of sulfate sometimes serve as competing electron acceptors to available electron donors (dissolved hydrogen), and limit intrinsic degradation of chlorinated solvents under ambient conditions. CDM Smith has extensive experience evaluating and implementing MNA at chlorinated solvent sites in the Basin and Range physiographic province. We know that MNA for PCE sites are particularly challenging and would rarely be a viable standalone remediation alternative. MNA may be viable for sites where TCE is the primary contaminant and aerobic cometabolism can be demonstrated. If significant iron mineral quantities are present and abiotic dechlorination is occurring, MNA could be viable. MNA may also be viable when implemented with other remedial alternatives: in-situ bioremediation/chemical reduction, if attenuation processes can be documented post-remediation. We developed the MNA monitoring and implementation approach at the Bountiful/Woods Cross Site following successful bioremediation implementation to substantially reduce TCE concentrations.

Because geochemical parameters are critical to understanding attenuation, low-flow sampling is preferred over passive sampling techniques. Passive samplers have limited volume and are often insufficient to collect all parameters necessary for a robust MNA evaluation. Many passive samplers are diffusion or sorption-based and may be limited to VOC analyses, thus preventing analyzing for MNA parameters. Previous studies by CDM Smith have shown little correlation with anion and metals data when using low-flow vs. passive sampling approaches. Lastly, low-flow sampling results in more accurate measurement of field parameters such as dissolved oxygen and ORP, which are important for MNA evaluation.

• Robertson, F. 1991. Geochemistry of Ground Water in Alluvial Basins of Arizona and Adjacent Parts of Nevada, New Mexico, and California. U.S.

Geological Survey Professional Paper 1406-C. https://pubs.usgs.gov/publication/pp1406C

• Arizona Department of Environmental Quality (ADEQ). 2014. Site Investigation Guidance Manual.

https://static.azdeq.gov/vrp/site_investigation_guidance.pdf

• United States Environmental Protection Agency (EPA). 1999. Use of Monitored Natural Attenuation at Superfund, RCRA Corrective Action, and Underground Storage Tank Sites. OSWER Directive 9200.4-17P. https://www.epa.gov/sites/default/files/2014-02/documents/d9200.4-17.pdf

• Schaefer, C.E., P. Ho, E. Berns, and C. Werth. 2018. “Mechanisms for Abiotic Dechlorination of Trichloroethene by Ferrous Minerals Under Oxic and Anoxic Conditions in Natural Sediments.” Environ. Sci. Technol. 52(23):13,747–13,755. https://pubs.acs.org/doi/10.1021/acs.est.8b04108

Question 5 (RESPONSE to this Question may be up to two (2) pages): Describe the in-situ technologies you have used and provide a critical evaluation of successes and failures. What would useful data quality objectives be for the various in-situ technologies you have used (plume control, source remediation, mass reduction)? What site characterization data and steps are needed to fully evaluate the applicability of specific in-situ technologies?

Offeror Response:

CDM Smith has designed, implemented, and evaluated over 100 in-situ remedial systems, including bench-scale tests performed in our Denver, CO and Bellevue, WA treatability laboratories. In-situ treatment technologies performed have included bioremediation, chemical oxidation, chemical reduction, and thermal remediation. In addition, these in-situ treatment technologies have been implemented using a variety of delivery mechanisms, including well injections, DPT injections, groundwater recirculation, permeability enhancement (via DPT or packers), and subsurface trenching / barriers.

Depending on site conditions and remedy objectives, different treatment technology and delivery mechanism combinations will be more suitable for a site, and multiple treatment or delivery strategies may be implemented.

https://pubs.usgs.gov/publication/pp1406C https://static.azdeq.gov/vrp/site_investigation_guidance.pdf https://www.epa.gov/sites/default/files/2014-02/documents/d9200.4-17.pdf https://pubs.acs.org/doi/10.1021/acs.est.8b04108

Arizona Department of Environmental Quality (ADEQ)

One of the main takeaways from our experience is the importance in understanding how characterization, remedial design, and implementation/performance monitoring all are best used continually to optimize design and implementation, as critical data are often collected during the implementation phase that may require revised conceptual site model (CSM) assumptions. Our technical experts have used this experience to author nationally recognized guidance on in-situ remediation such as the Interstate Technology Regulatory Council (ITRC) guidance called “Optimizing In-situ Remediation Performance & Injection Strategies”.

One tool that we have found essential to CSM development, injection design, and performance monitoring is 3D visualization models. These 3D models can quickly and cost-effectively provide a strong understanding of hydrogeologic conditions and when combined with vertically discrete contaminant data are extremely useful to understand relationships between contaminant concentrations and hydrogeological changes. Additionally, the 3D model is very effective for client, regulator, and/or stakeholder discussions.

Data quality objectives (DQOs) and appropriate remedial goals (interim or final) are essential for successful in-situ remediation. Development of well-thought-out DQOs is a critical step in our planning process that allows us to understand what data are really needed and how they will be used both to evaluate remedy performance and how they may be used to update the CSM / optimize remedy design.

During a pilot study, the primary DQOs that we want to answer for in-situ remediation systems are: 1) amendment distribution, 2) establishment of conditions suitable for dechlorination, 3) VOC degradation, 4) injection performance, and

5) longevity of dechlorinating conditions. These 4 DQOs are all critical to understanding appropriate injection point spacing, vertical treatment intervals, amendment treatment volumes, and potential reinjection requirement needs. These same DQOs are carried out during the full-scale implementation, with additional emphasis on injection performance (e.g., fouling) and treatment longevity.

Bench-studies are not always needed depending on the ultimate remedial goals because mechanisms are well understood for most available amendments. However, when using newer amendments or a combination of amendments, or if unique contaminants or geochemical conditions are present, bench studies can be useful to understand how to overcome these challenges.

In addition to the need for effective CSMs and remedy optimization as discussed above, our experience designing and implementing in-situ remediation has resulted in many lessons through successes and failures observed during implementation. Some examples of failures include overestimating the radial and vertical distribution of standard injections. Under moderately permeable conditions, injections via wells can be effective. However, under lower permeability or highly heterogeneous conditions, utilizing more robust injection strategies are often more effective. These include vertically discrete injections via DPT, vertically discrete permeability enhancement, and recirculation. Although DPT injections yield relatively small amendment distribution due to low volumes, often times they can be cost-effective as multiple points can be implemented simultaneously. In aquifers where DPT is not possible and standard well injections cannot be implemented, permeability enhancement may be considered.

Permeability enhancement (or fracturing) can be implemented to reliably distribute amendments 20-30 feet from the injection point, which significantly reduces the number of injection points required. Additionally, permeability enhancement is unique because solids (i.e., sand, ZVI) can be injected with this technique – and this can increase the treatment

Arizona Department of Environmental Quality (ADEQ) longevity and the enhanced permeability for years to decades. Another way to increase distribution is to recirculate impacted groundwater and blend in amendments before reinjection. By performing recirculation, the hydraulic gradient (and velocity) can be increased, thus increasing the effective amendment delivery distance. This method also allows for a higher loading of amendment to be blended in the treatment area.

In all cases, potable water should be used to “chase” the amendment injections to flush the well screen and sand pack so biological and mineral fouling are minimized. Failure to do so will cause excessive fouling of wells and require well rehabilitation. One other delivery enhancement method is to blend shear thinning fluids with water before injection. These shear thinning fluids (i.e., xanthan gum, guar) will increase the viscosity slightly. However, when injected in lower permeability units with less pore space for the injected water to travel, the injection velocity will increase and viscosity will decrease. In higher permeability units the injection velocity will decrease and viscosity will increase. This results in a more uniform injection distribution into both high and low permeability units, a critical factor because VOC mass is often diffused in low permeability units.

Using the proper amendment (or combination of amendments) is also critical depending on the delivery mechanism (examples described above) and the application (e.g., plume control, source remediation, mass reduction). For example, highly reactive amendments are very effective to establish reducing conditions quickly but are not always appropriate.

Some fast-release amendments metabolize so quickly after injected that poorly buffered aquifers will see pH decrease below 6. This will significantly slow bacterial growth and dechlorination. Also, some fast-release amendments promote biological growth so quickly when blended with groundwater (during recirculation) an cause substantial fouling in wells.

With the recent addition of injectable colloidal amendments (i.e., ZVI, activated carbon) similar issues have been observed. Colloidal ZVI is reactive enough that it will often not be effective after one to two years. Blending colloidal ZVI with carbon amendment (e.g., EVO) can elongate the reactive period, but since EVO often becomes depleted after 1-2 years there are limitations on the how long this can be effective. Blending activated carbon can help to sequester VOCs and act as barriers, but if the VOCs are not destroyed (via ZVI, EVO, or other amendments) before the activated carbon’s adsorptive capacity is spent, the barrier will need to be replenished over time.

ISCO is also frequently used for in-situ remediation. Common oxidants include modified Fenton’s (peroxide), permanganate, and persulfate. All oxidants are relatively short-lived and multiple injection events are generally required.

Because of this, ISCO is generally considered for mass removal as opposed to plume control. In our experience, persulfate has the most advantages, although activation is required - heat, high pH, and iron activation are the 3 most commonly applied. In iron rich environments, natural activation is possible.

Bench studies are often necessary to understand how much of the oxidant will be consumed by natural organic matter.

ISCO failures have most commonly occurred because concentrations rebound to baseline, sometimes after multiple injections. This is because oxidant only treats what is in the available groundwater and does not treat contaminants diffused into pore spaces. Because the oxidants react quickly, sites with significant contaminant mass diffused in pore spaces are not directly treated and “back diffuse” into groundwater after oxidation occurs. Additionally, permanganate often reacts with subsurface materials to form manganese oxide precipitates. These further isolate the diffused contaminants from available groundwater reducing treatment effectiveness. Modified Fenton’s is strongly reactive and can cause bubbling in the subsurface. Contaminants are often trapped with these bubbles and are transported to shallower portions of the aquifer instead of being oxidized.

Overall, our 20+ years of experience with in-situ applications and investment in both R&D and participation in guidance document development has given us extensive expertise in in-situ remediation design and implementation.

Question 6: ADEQ is focusing more on statistical techniques to both monitor long term performance of remedies, including MNA and optimize remediation and monitoring programs. Describe how the Air Force Center for Environmental Excellence (AFCEE) MAROS process can be used for these types of programs. What steps and metrics would you use to optimize a monitoring well network on a large PCE plume over a ten year period that includes both MNA and pumping from existing production wells for municipal or agricultural use?

Arizona Department of Environmental Quality (ADEQ)

Offeror Response:

The MAROS process involves four steps: data preparation, plume stability analysis, monitoring network evaluation, and monitoring optimization. To apply MAROS to a large chlorinated solvent plume, CDM Smith would import and organize the site data, perform statistical and graphical analysis of the plume behavior, assess the adequacy and efficiency of the current monitoring network, and generate recommendations for improving the monitoring network based on the cost and benefit trade-offs. Results from the log-linear and Mann-Kendall trend tests in MAROS serve as metrics of plume stability and MNA progress.

Outcomes of the trend tests can be used to identify the optimal monitoring frequency for key wells within the plume, in many cases reducing the frequency of sampling required to obtain data necessary for trend monitoring. The difference between the measured concentration at a well and the estimated concentration at that location (based on an inverse-distance-weighting interpolation of measured concentrations from all other wells) is used as a metric to identify redundant wells in the monitoring network, which can be removed from future monitoring events.

As a supplement or alternative to MAROS, statistical analysis with generalized additive models (GAMs) would serve as an option with:

1. Greater power to detect temporal trends. The trend tests in MAROS are limited to simple linear and monotonic trends, so data collection and analysis must begin anew when site dynamics violate those assumptions (see figure). In contrast, GAMs can analyze the entirety of datasets that reflect a complex site management history, resulting in less time and fewer samples required to justify closure.

2. An ability to leverage more data. Site information (e.g., rainfall, season, plume centroid, biogeochemical conditions, influence of nearby pumping wells) can be encoded as additional input to the GAM to control for confounding factors or evaluate potential correlations. This broader statistical approach allows for power analyses with simulated datasets where the GAM’s ability to detect changes in CoC concentration is tested in the context of a proposed reduction to sampling efforts.

In the 12 years since MAROS was last updated, GAMs have been used by thousands of environmental science research articles (for example, Mellor and Cey 2015, Kyte et al. 2023) , and software implementations have matured to the point where sophisticated GAMs can be specified with a few lines of R code.

To optimize the monitoring well network described in the question, CDM Smith would use a GAM to identify the most informative wells and the optimal sampling frequency. The GAM would predict the spatial and temporal variation of PCE concentrations as a function of multiple explanatory variables (see item 2 above).

GAM output would include metrics assessing the contribution of each variable to plume dynamics, as well as evaluating the uncertainty of model predictions. Based on the GAM results, one would then select wells that have the highest influence on model predictions, and eliminate or reduce the sampling of wells that provide redundant information.

References:

Mellor, A.F.P., and Cey, E.E. 2015. Using generalized additive mixed models to assess spatial, temporal, and hydrologic controls on bacteria and nitrate in a vulnerable agricultural aquifer. Journal of Contaminant Hydrology Volume 182, p. 104-116.

Kyte, E., Cey, E.E., Hrapovic, L., and Hao, X. 2023. Nitrate in shallow groundwater after more than four decades of manure application. Journal of Contaminant Hydrology Volume 256, Article 104200