ML23332A191

Exhibit 10 - FPL Year 3 Raasr Final

ML23332A191
Person / Time
Site:	Turkey Point
Issue date:	11/15/2021
From:	Miami Waterkeeper
To:	NRC/SECY/RAS
SECY RAS
References
RAS 56850, 50-250-SLR-2, 50-251-SLR-2
Download: ML23332A191 (0)

Text

EXHIBIT 10

Remedial Action Annual Status Report Turkey Point Clean Energy Center Year 3 November 15, 2021

FPL Turkey Point RAASR Year 3 November 2021 Executive Summary ES-1 EXECUTIVE

SUMMARY

The Florida Power & Light Company (FPL) has prepared this Remedial Action Annual Status Report (RAASR) to document the results of the Year 3 Recovery Well System (RWS) operation, in compliance with the monitoring and reporting objectives of the Miami-Dade County (MDC)

Consent Agreement (CA) and Florida Department of Environmental Protection (FDEP) Consent Order (CO). Pursuant to the MDC CA and FDEP CO, the RWS groundwater remediation system is designed to intercept, capture, contain, and retract hypersaline groundwater located to the west and north of FPLs property without creating adverse environmental impacts. The RWS consists of 10 extraction wells that remove hypersaline water from the Biscayne Aquifer and dispose of it in the Boulder Zone, more than 3,000 feet below the base of the aquifer, through an underground injection control (UIC) well system. FPL successfully initiated operations of the Turkey Point RWS on May 15, 2018.

FPL uses three primary tools to assess remediation progress: groundwater monitoring, continuous surface electromagnetic (CSEM) survey, and groundwater modeling (i.e., variable density flow and salt transport model). Data collected from groundwater monitoring wells from 2018 to Year 3 (October 1, 2020, to September 30, 2021) of remediation, in conjunction with the comparative 2018-2021 CSEM surveys and updated and recalibrated modeling results, were used collectively to assess changes in the extent of the hypersaline plume. The groundwater model was also used to estimate future reductions to the hypersaline plume based on 3 years of remediation model simulations. Data and modeling confirm that the objectives of the CA and the CO through Year 3 have been met. The following is a summary of the major findings of this evaluation:

The Year 3 CSEM results, compared to the 2018 baseline survey results, indicate the volumetric extent of the hypersaline plume has been reduced by 42% after only 3 years of RWS operation.

Since inception of the remediation system, approximately 18.45 billion gallons of hypersaline groundwater and 7.32 billion pounds of salt have been extracted from the Biscayne Aquifer. Approximately 5.92 billion gallons of hypersaline water and 2.32 billion pounds of salt were removed during this reporting period.

Florida Power & Light Company employs three types of data and associated analyses (i.e.,

monitoring, electromagnetic surveys, and modeling) to assess progress in meeting the objectives of the Miami-Dade County Consent Agreement and Florida Department of Environmental Protection Consent Order. Analyses of data through Year 3 of remediation demonstrate that the net westward migration of the hypersaline plume has been halted; hypersaline groundwater from the canal cooling system is being intercepted, captured, contained, and retracted by RWS operations. The CSEM data shows that the volume of hypersaline water in the compliance area has been reduced by 42% since remediation began in 2018.

FPL Turkey Point RAASR Year 3 November 2021 Executive Summary ES-2

• In total, 20 of 23 monitoring wells showed a statistically significant declining trend in one or more parameters (quarterly chloride, quarterly tritium, and weekly average automated salinity) and at least one or more parameters that were the lowest this reporting period compared to the baseline and Years 1 and 2.

• Based on CSEM data, the greatest reduction in percent hypersalinity volume is occurring in the lower portion of the aquifer, while the more significant reductions in groundwater monitoring well salinities are being measured in the upper portion of the aquifer as the plume shrinks from top to bottom.

• The Year 3 recalibrated V6 model forecast simulations for Years 5 and 10 show continuous improvement in hypersaline retraction, with complete retraction achieved in the upper two-thirds of the aquifer by Year 10. However, complete retraction in the southern portion of layer 9 and all of layers 10 and 11 are not achieved by Year 10 of the simulation. Further improvements to the model are needed to reliably represent the dynamics of the hypersaline plume responses to the RWS in the lower portions of the aquifer. It is expected that these differences will continue to be reduced as the model is informed by subsequent annual remediation results.

• Given the significant progress of remediation since initiation of the RWS, FPL does not propose any changes to the agencies approved remediation plan at this time.

In addition to the agency-approved remedial action plan described above, FPL has taken additional measures to help further the objectives of the CA and CO which include:

• Reactivating the UIC test production wells which extracted up to 3 mgd of hypersaline groundwater beneath the CCS in the reporting period. Reduction of hypersalinity and driving head beneath the CCS will increase the capture radius of the RWS to the west of the RWS, facilitating plume remediation.

• Increasing the CCS freshening allocation from 14 mgd to 30 mgd (34 mgd maximum monthly allocation) which will allow FPL to maintain seawater salinity levels during drought periods. By maintaining average annual salinities in the CCS at the CO target of 34 practical salinity units, the formation of hypersaline water in the CCS will be prevented.

It is important to note that the aquifer system is complex and subject to many external factors beyond the CCS and RWS; therefore, continued monitoring, model updates, and scientific data analyses are performed to improve our understanding of the impact of RWS operations in concert with these other factors. FPL will continue to monitor and evaluate progress in meeting the requirements of the CA and CO and make recommendations for modifications as needed.

FPL Turkey Point RAASR Year 3 November 2021 1 Introduction 1-1 1 INTRODUCTION

1.1 BACKGROUND

Florida Power & Light Company (FPL) submits this Year 3 Remedial Action Annual Status Report (RAASR) pursuant to paragraphs 17.b.ii and 17.d.v of the Miami-Dade County (MDC)

Department of Regulatory and Economic Resources (DERM) Consent Agreement (CA) and paragraphs 28, 29.c. and 33 of the Florida Department of Environmental Protection (FDEP)

Consent Order (CO). FPL entered into the CA on October 7, 2015, and the CO on June 20, 2016. FPL agreed to conduct specific actions, including the remediation of hypersaline groundwater adjacent to the FPL Turkey Point Power Plant (Turkey Point). The specific objectives of the CA are to demonstrate a statistically valid reduction in the salt mass and volumetric extent of hypersaline water in groundwater west and north of FPLs property without creating adverse environmental impacts and to reduce the rate of, and ultimately arrest, migration of hypersaline groundwater. The specific hypersaline groundwater remediation objectives of the CO are to halt the westward migration of the hypersaline plume from the Cooling Canal System (CCS) within 3 years and reduce the westward extent of the hypersaline plume to the L-31E canal within 10 years. Hypersaline groundwater, as defined in the CO and CA, is groundwater with a chloride concentration greater than 19,000 milligrams per liter (mg/L).

FPL initiated the evaluation and design of a recovery well system (RWS) as part of a Remedial Action Plan (RAP) to intercept, capture, and retract hypersaline groundwater west and north of the FPL property boundary in accordance with the requirements of the CA and CO. To design the RWS, FPL developed a groundwater flow and salt transport model, which was extensively reviewed by the South Florida Water Management District (SFWMD), FDEP, MDC, the United States Environmental Protection Agency (USEPA), and the University of Florida. The model and remediation design were ultimately approved by MDC on May 15, 2017. After obtaining all required environmental and well construction permits, FPL initiated the construction of the RWS, which included 10 groundwater recovery wells, a conveyance pipeline system, and a deep injection well (DIW) more than 3,000 feet (ft) below the base of the aquifer. The system was fully operable on May 15, 2018. FPL submitted an RWS startup report to MDC in October 2018 (FPL 2018a), and quarterly RWS status reports through May 2019 (FPL 2018b, 2019a, 2019b).

These reports provided information on the design and operation of the approved RWS.

Annual reports were then subsequently submitted with the first year of operation (Year 1) covering the period from May 15, 2018, through May 31, 2019 (FPL 2019c). In Year 2, collection of the Continuous Surface Electromagnetic (CSEM) survey data was delayed from the originally scheduled end of May 2020 timeframe until September 2020 due to restrictions on international travel and health risks associated with the COVID-19 pandemic. This resulted in the Year 2 RAASR having 16 months of data and being submitted in two parts: groundwater monitoring data from June 2019 to September 2020 (FPL 2020b) and the Year 2 CSEM survey and groundwater model (FPL 2021a). This reports timeframe encompasses October 1, 2020, to

FPL Turkey Point RAASR Year 3 November 2021 1 Introduction 1-2 September 30, 2021; and the 12-month period will be referred to as Year 3. This includes the CSEM survey conducted from June 19 to June 22 with groundwater monitoring trend analyses covering the year from October 2020 to September 2021. Groundwater model calibration covers data from August to October of each year using the May-June CSEM survey results and groundwater monitoring and plant operations data for the reporting period.

Data and accompanying analysis in the Year 1 reports indicated a statistically valid reduction in salt mass and a 22% reduction in the volumetric extent of hypersaline groundwater west and north of the FPL property. An additional 12 % in reductions were observed in the second year of operation which were documented in the Year 2 RAASR (FPL 2021a). In addition to capturing and reducing the plume extent, particle tracking using the V5 updated model demonstrated that the RWS creates a hydraulic barrier that intercepts and contains hypersaline groundwater located beneath the CCS from migrating west and north.

The Year 3 monitoring data show that the net number of monitoring wells with declining trends in chloride, salinity, and/or tritium have increased in all three depth intervals indicating positive signs of remediation vertically in the aquifer. In addition, the June 2021 CSEM survey shows a further reduction of 8% since the previous CSEM survey conducted in September 2020, in the horizontal and vertical extent of hypersaline plume. This is a positive indication that the remediation is meeting the objectives of the CA and CO.

1.2 SCOPE OF THE REMEDIAL ACTION ANNUAL STATUS REPORT This Year 3 RAASR report includes the following:

• Year 3 RWS operational summary, including analytical results from the RWS wells, salt mass and hypersaline groundwater removal and operation run times

• Data and assessment from monitoring wells in Year 3 and comparisons to baseline conditions and/or previous years

• Year 3 annual CSEM survey results with comparisons between the 2021 CSEM survey and 2018 baseline CESM survey

• An updated RWS groundwater model description and results as well as Year 5 and Year 10 remediation forecast results

• Appendices containing additional supporting information and data used in the report (Appendices A - H); and

• An evaluation of the RWS progress, after Year 3, in meeting the objectives of the CA (Appendix I)

FPL Turkey Point RAASR Year 3 November 2021 1 Introduction 1-3 Section 2 of this report provides an overview of RWS operations, and it includes a summary of automated and analytical data from the recovery wells and the calculation of total salt removed by the RWS.

Section 3 of the RAASR provides automated data and/or analytical samples from monitoring well sites (up to 28 wells) located within the areal extent of the hypersaline groundwater plume collected from October 1, 2020, to September 30, 2021. These data along with Year 1 and Year 2 data were used to evaluate changes and trends in groundwater quality from baseline conditions from March 2018 through September 2021 (43 months). These results were compared to the data collected prior to the startup of RWS to identify changes likely related to RWS operations.

Groundwater chloride contour maps for the shallow, middle, and deep monitoring well horizons augmented with CSEM data are generated for Year 3 and compared with similarly prepared 2018 baseline contour maps to identify changes in the extent of hypersalinity.

Section 4 of the RAASR includes the results of the Year 3 CSEM survey with comparisons to the baseline 2018 CSEM survey to document changes to the extent and volume of the hypersaline plume within the CO/CA compliance boundary that have occurred since RWS operations began.

Section 5 of the RAASR encompasses documentation of the updated, recalibrated Turkey Point groundwater flow and salt transport model with predictive model runs for Year 5 and Year 10 of plume remediation.

1.3 STATUS OF CONSENT AGREEMENT/CONSENT ORDER IMPLEMENTATION The RAASR focuses on the RWS and provides the information required by the CA and CO relevant to the RWS. Both the CA and CO include additional actions beyond the design and operation of the RWS, and a current status of those actions is included in Appendix A. At the time of this report, FPL has successfully completed many actions contained in the CA and the CO. The remaining actions, such as hypersaline plume extraction, salinity and nutrient management in the CCS, and monitoring, are ongoing, long-term activities. FPL continues to work cooperatively and effectively with regulators and interested parties to achieve the objectives of the CA and CO.

FPL has successfully completed multiple restoration and remediation activities outlined in the MDC CA and the FDEP CO and has made substantial progress in implementing and completing activities outlined in the CA and CO.

FPL Turkey Point RAASR Year 3 November 2021 2 Recovery Well System Year 3 Operation Summary 2-1 2 RECOVERY WELL SYSTEM YEAR 3 OPERATION

SUMMARY

2.1 HYPERSALINE EXTRACTION/DISPOSAL OPERATIONS FPL operates 10 recovery wells to extract up to 15 million gallons per day (mgd) of hypersaline groundwater, preferentially along the base of the Biscayne Aquifer. The extraction wells are cased to the lower high flow zone of the Biscayne Aquifer (FPL 2018a) allowing hypersaline water to be withdrawn along the base of the plume. As the extraction wells are pumped, hypersaline groundwater from beneath the CCS and from the plume west and north of the CCS flows laterally toward the points of withdrawal. As hypersaline water is removed, the plume shrinks both vertically and laterally with adjacent lower salinity groundwater replacing the area formerly containing hypersaline groundwater. The extraction of hypersaline groundwater from the lower extent of the Biscayne Aquifer along the western margin and north of the CCS accomplishes the objectives listed below:

• Reduces the salt mass and volumetric extent of hypersaline groundwater west and north of the CCS. The retraction of the hypersaline plume is accomplished primarily by direct extraction of hypersaline groundwater, which increases the natural seaward groundwater flow gradient eastward into the RWS capture zone, and secondarily by natural dilution and dispersion of hypersaline water with the lower salinity waters in the aquifer.

• Creates a hydraulic barrier that intercepts and contains the westward and northward migration of hypersaline groundwater from the CCS. RWS operations extend the hydraulic barrier effect of the interceptor ditch (ID) operation in the upper portion of the Biscayne Aquifer to the base of the aquifer.

• Decreases groundwater salinity and mass beneath the CCS, which reduces the driving force that contributed to lateral movement away from the CCS and which is a component of halting the westward migration of hypersaline groundwater from the CCS.

During the reporting period, FPLs groundwater remediation actions removed approximately 5.92 billion gallons of groundwater with an average chloride concentration of 27,000 mg/L that contained 2.32 billion pounds of salt. Since inception of the remediation system, approximately 18.45 billion gallons of groundwater with an average chloride concentration of 27,900 mg/L and 7.32 billion pounds of salt have been extracted from the Biscayne Aquifer.

FPL Turkey Point RAASR Year 3 November 2021 2 Recovery Well System Year 3 Operation Summary 2-2 The hypersaline groundwater is pumped from each recovery well into a collection system that consists of an approximately 9-mile-long pipeline that is routed to a DIW located near the center of the CCS for disposal.

The DIW is a 24-inch-diameter permitted Underground Injection Control (UIC) non-hazardous Class I industrial wastewater disposal well (Permit No.

0293962-004-UO/1I) constructed to a depth of 3,230 ft below ground surface into the regionally confined Boulder Zone. Near the end of Year 1, the permitted operating capacity of the DIW was increased from 15.59 mgd to 18.64 mgd (permit modification No. 0293962-005-UO/MM) to accommodate additional remediation flows.

The Consumptive Use Permit from SFWMD authorizes an RWS annual withdrawal allocation of 5,475 million gallons (15 mgd) and a maximum monthly allocation of 465 million gallons from RWS extraction wells 1 through 10. In early 2020, two UIC production test wells (UICPW-1 and UICPW-2), co-located with the DIW and constructed to the base of the Biscayne Aquifer in a similar manner as the recovery wells, were activated at a combined rate of approximately 3 mgd to remove hypersaline groundwater from beneath the CCS. This extracted hypersaline water is disposed in the DIW along with the RWS-extracted hypersaline water, utilizing the DIW UIC permits injection rate limit.

The groundwater extraction wells are controlled by a Supervisory Control and Data Acquisition (SCADA) system that controls the operation of all wells, has the capability to monitor and regulate individual well withdrawal rates, and maintains real-time-assigned total system extraction capacity in the event of individual well fluctuations. This system assists the operators in maintaining compliance with groundwater withdrawal and disposal permit limits. Flow pumped from each well is measured by totalizers; the combined flow down the DIW is also measured by a totalizer. All RWS and DIW flow meters were checked, calibrated, and certified in May 2021 as part an annual calibration process. During Year 3, individual wells were also temporarily shut down from a few hours to several days at a time for maintenance purposes (e.g.,

control system upgrades, annual calibration/fall-off test). Individual well shutdowns do not reduce total volumes extracted because the SCADA system allows for the remaining wells to make up the loss.

Overall, the RWS operated 98.1% from October 1, 2020, to September 30, 2021, with only 7 days in which the system was non-operational. During four of those days (June 19 - 22), the entire system was shutdown to reduce electrical noise for the CSEM survey.

Operational run times for each of the RWS wells are shown graphically on Figure 2.1-1. On average, individual extraction wells operated 92.5% of the time during the reporting period, a The RWS operated during 98.1% of the reporting period; there were only 7 days (1.9%) when the system was not operational. The individual wells collectively operated 92.5% of this reporting period, a slight increase over Year 2 (91.7%). Despite the limited outages, FPL continues to work towards improving system performance and optimizing individual well performance.

The deep injection well.

FPL Turkey Point RAASR Year 3 November 2021 2 Recovery Well System Year 3 Operation Summary 2-3 slight improvement in RWS operational time over the Year 2 performance. All wells had run times over 90%, except for RWS-1 which operated 78% of the time. RWS-1s lower run time was attributable to several issues including wellhead repairs, pump motor failures and long lead times for repairs. Nonetheless, FPL is continuing to work towards improving the operation of individual wells and further optimizing system performance.

2.2 RECOVERY WELL SYSTEM MONITORING RESULTS AND HYPERSALINE GROUNDWATER/SALT MASS REMOVED Automated flow, salinity, total dissolved solids (TDS), and water elevation data were continuously recorded from each RWS extraction well. Water quality samples were collected from each RWS well monthly and were analyzed for chloride along with field parameters.

Pursuant to execution of CA Amendment 2 on August 20, 2019, quarterly sampling of RWS nutrients was implemented in September 2019. All sampling/monitoring was conducted in accordance with the SFWMD-approved FPL Quality Assurance Project Plan (QAPP) (FPL 2013). Water quality data referenced in this RAASR are available in Microsoft Excel tables on the FPL Turkey Point Electronic Data Monitoring System (EDMS) database (https://www.ptn-combined-monitoring.com).

Automated data for all 10 RWS wells and the two UICPW wells are shown in Appendix B.

Analytic data are shown in Appendix C including the field parameters and additional analytes (nutrients), field sampling logs, data qualifiers, and quality assurance samples. Data usability summary (DUS) reports for the events are provided in Appendix D. Level 4 laboratory reports from the FPL Central Laboratory can be found on FPLs EDMS at https://www.ptn-combined-monitoring.com.

Table 2.2-1 shows a summary of the chloride values for all recovery wells. Chloride values in most of the recovery wells reflect hypersaline conditions, ranging between 22,000 mg/L and 32,000 mg/L in Year 3 (Figure 2.2-1). The only exception was RWS-1, which ranged between 15,000 mg/L and 20,000 mg/L and was less than 19,000 mg/L for 10 of the 12 months of the Year 3 reporting period. Although chloride values were generally within the same range as in Year 2, statistically significant declines in chloride concentrations were observed in all the RWS wells based on the Mann-Kendall trend analyses using the monthly data since May 2018.

This gradual reduction in salinity of the RWS wells was documented in early modeling of the RWS (Tetra Tech 2016). The design of the remediation system considers the fluid density of the plume, which is why the extraction wells are open to the base of the aquifer (i.e., dense hypersaline groundwater will naturally sink toward extraction points along the base of the aquifer). Accordingly, it is expected that the salinity levels of the extracted water from the RWS wells will remain elevated for an initial period while the thickness of the plume diminishes. As the vertical and lateral extent of the hypersaline plume diminishes over long-term operation of the RWS, larger portions of lower-salinity groundwater from above the extraction horizon mix with hypersaline water moving laterally along the base of the aquifer resulting in a gradual lowering of the extracted water salinity.

FPL Turkey Point RAASR Year 3 November 2021 2 Recovery Well System Year 3 Operation Summary 2-4 Analytical and automated data indicate that all recovery wells have lower average chloride concentration and salinity values this reporting period compared to the first year of operation, and most show a gradual decline. The majority of changes since start-up are modest, i.e., RWS-2, RWS-3, RWS-7, and RWS-10 had both chloride and salinity reductions ranging between 5%

and 10 %, while RWS-4, RWS-5, RWS-6, RWS-8, and RWS-9 had one or both parameters lower by less than 5%. However, RWS-1, has exhibited greater reduction since the first year of operation as both average chloride and salinity values are now approximately 20% lower than at inception. RWS-1 is located approximately 0.8 miles north of the CCS where the plume is thin; and, as CSEM data shows, the plume has diminished significantly in this area since remediation began in 2018. Changes to current pumpage operations will be considered when an RWS well produces saline water that is consistently well below 19,000 mg/L chloride and when data indicate that CCS hypersaline groundwater within the capture radius of the RWS production well has been sufficiently remediated to warrant modifying pumpage rates.

Table B.1-1 in Appendix B shows the weekly volume of groundwater water pumped from each recovery well. From October 1, 2020, through September 30, 2021, approximately 5.17 billion gallons of water were extracted from the RWS and disposed of via the DIW.

An additional 0.75 billion gallons of hypersaline groundwater were extracted in Year 3 from the UICPW wells in the middle of the CCS (Table B.2-1), for a total of 5.92 billion gallons of hypersaline groundwater removed from the Biscayne aquifer from October 1, 2020, to September 30, 2021.

Table B.1-1 of Appendix B also shows the automated weekly TDS values and the associated amount of salt mass removed on a weekly basis for each recovery well, which is calculated in accordance with paragraph 29.f of the FDEP CO. Salinity data is provided alongside the TDS values for reference purposes because most people are familiar with salinity. The salt mass values were based on automated flow and TDS data, and the values were then summed for daily and weekly salt mass removal. The TDS value is calculated from specific conductance using a preprogrammed conversion factor of 0.64 (based on empirical data from monitoring wells TPGW-11D and TPGW-13D from 2010-2016). The equation for salt mass removal is as follows:

Salt mass removed (lbs/day) =

Flow gallons min x TDS g L x 1000 mg g x 3.7854 ( liters gallon) 453,592.37 (mg lbs) 1440 (

)

The total amount of salt mass removed varies since the pumping rates, run time, and salinity/TDS differ among wells and/or over time. In Year 3, approximately 2.00 billion pounds of salt was removed from the RWS wells (Table B.1-1) and 0.32 billion pounds for UICPW-1 Reporting Year Volume (billion gallons)

Salt (billion lbs)

Avg Salinity/

(in PSS-78 scale) 1 (12 months) 4.99 2.01 53.6 2 (16 months) 7.54 2.99 51.8 3 (12 months) 5.92 2.32 50.8 Total 18.45 7.32 Values shown are for total water and salt mass extracted each reporting period and total since startup through through September 2021.

FPL Turkey Point RAASR Year 3 November 2021 2 Recovery Well System Year 3 Operation Summary 2-5 and UICPW-2 (Table B.2-1), resulting in 2.32 billion pounds removed in the reporting year.

Combined with the 5 billion pounds removed in two previous reporting periods, 7.32 billion pounds of salt have been removed from the Biscayne Aquifer to date.

2.3 RECOVERY WELL SYSTEM DRAWDOWN ASSESSMENT In the 2019 RAASR (FPL 2019c), FPL determined that the drawdown solely from RWS pumping was approximately 0.11 ft, and the combined drawdown with the RWS and ID was approximately 0.25 ft at TPGW-15S which is located approximately 710 ft from RWS-3. The drawdown in the other depth intervals and at TPGW-1 was of similar magnitude. This amount of drawdown in the shallow portion of the Biscayne Aquifer is considered negligible; i.e., it is not considered harmful to wetlands or water resources. The SFWMD regulates drawdown impacts to wetlands and water resources. SFWMD water use rule criteria limit cumulative drawdowns beneath seasonally inundated wetlands to 1 ft during 1-in-10-year drought conditions and maximum authorized withdrawals (SFWMD 2015). Drawdowns that exceed this threshold are considered harmful to wetlands and water resources. Based on the measured drawdowns of the combined impacts of the RWS and ID operations, the combined withdrawals are negligible.

Subsequently, in the Year 2 assessment presented in the 2020 Part 1 RAASR (FPL 2020b), FPL confirmed the above findings of negligible drawdown of 0.10 ft at TPGW-1S and TPGW-15S from solely RWS operations as there was no time during the reporting period when the RWS was turned off and the ID pumps were operating.

For Year 3 of operation, and similar to the previous years, several periods were selected when the RWS wells near TPGW-1 and TPGW-15 were turned off to allow the groundwater to stabilize when there was little-to-no rainfall that could mask drawdown. Based on a review of RWS operations, there are two periods when this occurred that coincided with times when ID pumps were operational:

• March 12-24, 2021, when less than 0.1 inch of rain fell over the CCS, based on Next Generation Radar (NEXRAD) data

• May 20-31, 2021, when less than 0.1 inch of rain fell over the CCS The results support previous findings of a combined drawdown of approximately 0.25 ft at TPGW-1S and TPGW-15S when RWS and ID pumps are both operational. Additionally, the impact of the RWS operations on L-31E stage levels are not discernible (in the hundredths of a foot) when the pumps are turned on and off. The changes in water levels at the Turkey Point surface water canals TPSWC-1, TPSWC-2, and TPSWC-3 are within the range of normal Water table drawdown from combined RWS and ID operations continue to be negligible (approximately 0.25 ft) in Year 3, consistent with previous observations.

FPL Turkey Point RAASR Year 3 November 2021 2 Recovery Well System Year 3 Operation Summary 2-6 fluctuations due to typical minor meteorological influences (e.g., wind), and do not appear to be a result of RWS operations.

2.4 INTERCEPTOR DITCH OPERATIONS FPL has reviewed ID operations in conjunction with RWS operations on multiple occasions in accordance with paragraph 17.a.iii of the CA. FPL has presented these findings at various times, including in a meeting with DERM on May 16, 2016, in a letter to DERM dated May 23, 2016, in a presentation to DERM, FDEP, and SFWMD on May 19, 2017, in the RWS Start-Up Report (FPL 2018a), subsequent quarterly status reports (FPL 2018b, 2019a, 2019b), as well as in FPLs Annual Monitoring Reports (FPL 2012, 2016, 2017, 2018c, 2019d, 2020a, 2021b). Based on these evaluations, modifications to improve the ID function are not warranted at this time due to the following:

• Continued effectiveness of the ID in restricting westward migration of CCS groundwater into the upper portion of the Biscayne Aquifer, into wetlands west of the CCS, and into the L-31E canal

• Continued effectiveness in maintaining the freshwater lens thickness in the Biscayne Aquifer west of the CCS

• Demonstrated lack of harmful impacts to groundwater levels, wetlands, and other water resources in the area as further described in the reports referenced above.

FPL Turkey Point RAASR Year 3 November 2021 Tables 2-7 Table 2.2-1.

RWS Chloride Monitoring Results (mg/L).

Oct Nov Dec Jan Feb March April May June July Aug Sept Previous Year (10/1/19-9/30/20)

Current Reporting Period (10/1/20-9/30/21)

RWS-1 17700 17400 16600 16900 20000 18200 18100 19200 18100 15300 17000 16200 18492 17558 RWS-2 24300 24000 25700 26200 25000 24600 25900 24900 22800 22900 21700 23800 25750 24317 RWS-3 27300 26800 27500 28300 28000 26500 27100 27200 25400 24900 27900 26600 27625 26958 RWS-4 29400 29000 30400 30500 31100 28200 30700 27900 28000 27000 30700 30200 29983 29425 RWS-5 29800 29100 30200 31300 28700 28700 28200 29100 27500 26600 30400 30200 29908 29150 RWS-6 28800 28600 29700 30200 28500 27200 29300 27700 27300 26200 29600 28800 29433 28492 RWS-7 29000 28100 29400 30400 27900 27400 29200 26500 26300 25300 28400 28800 29108 28058 RWS-8 29900 29300 30800 32000 29100 29100 28400 28800 27600 26800 29700 29900 29667 29283 RWS-9 28500 27800 29700 30100 27400 28000 27900 26900 26700 26100 29200 28400 29258 28058 RWS-10 25800 25400 26500 27000 24500 25300 26100 24300 24400 23300 26700 25600 26300 25408 UICPW-1 30400 NA 32300 NA NA NA 31200 NA 28900 NA 30400 NA 31400 30640 UICPW-2 NA 29600 NA NA 31300 29500 NA 29100 NA 28900 NA 32400 32100 30133 Notes:

Previous year average for UICPW-1 based on three months of data and for UICPW-2 one month of data.

Key:

NA = not available/no pumping.

Average Sample ID 2020 2021

FPL Turkey Point RAASR Year 3 November 2021 Figures 2-8 Figure 2.1-1. Operation of RWS in Year 3 (pumping with more than 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> of flow in a day).

FPL Turkey Point RAASR Year 3 November 2021 Figures 2-9 Figure 2.2-1. RWS Chloride Results (mg/L).

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-1 3 GROUNDWATER MONITORING DATA 3.1 GROUNDWATER MONITORING Groundwater monitoring for the assessment of the RWS was performed on well clusters TPGW-1, TPGW-2, TPGW-4, TPGW-5, TPGW-7, TPGW-12, TPGW-15, TPGW-17, TPGW-18, and TPGW-19, and on historical individual wells TPGW-L3, TPGW-L5, TPGW-G21, and TPGW-G28 with samples collected for laboratory analysis in December 2020, March 2021, June 2021, and September 2021 (Figure 3.1-1). These are the same wells that were sampled as part of the March 2018 baseline and the Years 1 and 2 RAASRs. In this reporting year, an additional monitoring well installed by MDC, TPGW-22, was added to the monitoring network per the MDC letter dated September 16, 2020. The well horizons for this 3-cluster well, which were established by MDC in 2020, are screened at depths independent of consideration of the protocols previously established for the other FPL monitoring wells (e.g., geophysical and boring logs used to identify high flow zones, FPL/SFWMD professional geologist consultation). It is not clear if the screened intervals were placed in high-flow zones. Accordingly, the monitoring horizons may not be fully comparable with the zones established with the other wells currently in the network. Data on this well has been collected since February 16, 2021 Samples for all events were collected at discrete screen intervals from the well clusters (i.e.,

shallow, intermediate, and deep intervals), except for the historic L-and G-series, which are continuously screened wells where samples were collected at 18 ft and 58 ft below the top of casing. Samples from the groundwater clusters were collected using dedicated tubing and per the methods outlined in the QAPP (FPL 2013) and FDEP Standard Operating Procedures. To aid in the assessment of the RWS, field parameters (i.e., temperature, specific conductance, salinity, density) were measured, and samples from each of the monitoring wells were sent for laboratory analysis of TDS, chloride, and tritium.

A summary of the Year 3 quarterly chloride and tritium results is included in Tables 3.1-1 and 3.1-2, respectively, along with baseline results for comparison. These results help gauge the progress of remediation, and the chloride data support the calibration of CSEM survey and groundwater modeling updates. Time-series graphs showing quarterly chloride and tritium data from March 2018 (baseline) through September 2021 (end of Year 3 reporting period) are provided in Appendix E and show the extent of change since RWS start-up. Note that in the first year of monitoring, chloride samples were collected weekly for the first month of operation and monthly for the first quarter; data was presented in the 2019 RAASR (FPL 2019c). Chloride trend analyses conducted on monitoring data collected since RWS start-up were based on quarterly data (i.e., early weekly/monthly values were not included) to avoid sample frequency biases.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-2 In addition to analytical data, all the monitoring wells, except TPGW-L3, TPGW-L5, TPGW-G21, and TPGW-G28, are equipped with automated probes that record specific conductance, salinity, and water levels at 1-hour intervals. Depending on when the well was installed, automated data have been recorded since at least April 2018, with several well clusters (TPGW-1, TPGW-2, and TPGW-12) having data that extends back to 2010. For the newly added well site, TPGW-22, automated probes were deployed in February 2021. Appendix E shows times-series salinity graphs of each automated instrumented well.

Nearly all the analytical and automated data for Year 3 meet the data quality objectives of the QAPP (FPL 2013). All analytical monitoring well data are usable and exceed the QAPP completeness goal of 90%. Collectively, automated monitoring well water quality data and water level data are over 90% complete in Year 3, except for wells TPGW-22S, TPGW-22M, and TPGW-22D. This well cluster (TPGW-22) has had ongoing issues, mostly with obtaining valid automated water level readings. The automated water elevations have been uncharacteristically variable and do not match field readings which are more accurate. The water level readings at this station are not needed to assess the progress of remediation. FPL has swapped out probes and cables multiple times, reset reference levels, and is working with the manufacturer to isolate the cause of the oscillations. FPL will continue troubleshooting efforts to improve data completeness at TPGW-22. Additional details on qualified analytical and automated data can be found on FPLs EDMS at https://www.ptn-combined-monitoring.com.

3.2 2021 YEAR 3 WATER QUALITY CONDITIONS AND TRENDS To assess trends and influence of RWS operations, a multifactor data screening process was applied to analytical and automated groundwater monitoring data. This included comparing Year 3 data against period of record low values and conducting objective statistical trend analyses (i.e., linear regression and Mann-Kendall trend) for chloride, tritium and salinity.

Assessments for tritium help confirm chloride/salinity trends and indicate a potential precursor to a declining chloride and salinity trend as lower tritium may indicate a reduction in CCS-sourced water (Table 3.2-1).

In addition to identification of a new period of record low data values and statistical time series trend analyses, data were assessed to determine if chloride values declined and remained below 19,000 mg/L in monitoring wells that were hypersaline prior to RWS operation (Table 3.1-1).

Three stations met this criterion: TPGW-1S, TPGW-2S, and TPGW-15S; these are the same A multi-factored objective data assessment was used to identify meaningful changes in groundwater quality associated with the ongoing groundwater remediation. The assessment identified 20 of the 23 RWS monitoring wells exhibited statistically significant declining trends in either chlorides, salinity, and/or tritium since remediation began in May 2018, with three monitoring wells previously transitioning in Years 1 or 2 from being classified as hypersaline to saline with chloride concentrations staying below 19,000 mg/L through Year 3. In addition, the majority of the RWS monitoring wells reported lowest values on record during the reporting period.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-3 wells as noted in last years RAASR (FPL 2020b). Chloride concentrations in both TPGW-1S and TPGW-15S dropped below 19,000 mg/L in Year 1 of RWS operation; and they continue to decline with the lowest chloride concentrations of 5,850 mg/L and 3,440 mg/L, respectively, recorded during this reporting period. At TPGW-2S, chloride concentrations transitioned below 19,000 mg/L by March 2020, and the lowest concentration over the period of record was recorded in June 2021 (15,600 mg/L).

There are three other wells (TPGW-17S, TPGW-18M and TPGW-19M) where the chloride concentrations are hovering around 20,000 mg/L. In the case of TPGW-17S, chloride concentrations in March for the past two years have dropped below 19,000 mg/L. The data is showing that the transition from hypersaline to saline is occurring first along the upper edge of the hypersaline plume as expected, and it is anticipated to transition deeper as the thickness of fresher groundwater increases in the aquifer as more hypersaline groundwater is removed.

Year 3 data were also reviewed against period-of-record values to determine if new low values were observed compared to Years 1 and 2, baseline values, and/or to the entire period of record.

For well clusters TPGW-1, TPGW-2, and TPGW-12, and wells TPGW-L3 and TPGW-L5, the period of record dates to at least mid-2010, TPGW-15 dates to September 2015, TPGW-17 and TPGW-19 date to January 2018, and TPGW-18 dates to April 2018. The lowest value may not be indicative of a trend, but it strongly suggests that there are some positive changes occurring over a broad area and that reductions in concentrations are continuing. This assessment found the following:

• Chloride - 15 out of 23 (65%) wells had the lowest analytic quarterly chloride concentration ever recorded during Year 3. This same number of wells had the lowest analytic chloride concentration this reporting period (Year 3) compared to baseline and Years 1 and 2.

• Tritium - 14 out of 23 (61%) wells evaluated exhibited the lowest quarterly tritium value ever recorded in Year 3; and 16 out of 23 (70%) wells evaluated exhibited the lowest value in Year 3 compared to the baseline and Years 1 and 2. Reduction in tritium levels can be caused by: 1) replacement of CCS-sourced groundwater with surrounding less-saline, non-CCS sourced groundwater; 2) inflow of older CCS groundwater, containing lower levels of depleted tritium, farther away from the CCS that are being drawn eastward toward the RWS extraction wells; or 3) radioactive decay of in-place tritium that is no longer being replenished by younger tritiated water from beneath the CCS.

• Automated Salinity - 12 out of 21 (57%) wells with automated instrumentation recorded the lowest average weekly salinity value in Year 3 compared to the period of record; and 15 out of 21 (65%) wells exhibited the lowest average weekly value in Year 3 compared to the baseline and Years 1 and 2. Similar to the chloride results, the number of stations exhibiting the lowest salinity in Year 3 indicates progress in reducing the extent of the hypersaline plume.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-4 In an effort to assess if these lower values were reflective of a trend, analytical data from nine monitoring sites (23 individual monitoring wells) located within the area containing hypersaline groundwater from the CCS (well clusters TPGW-1, TPGW-2, TPGW-12, TPGW-15, TPGW-17, TPGW-18, and TPGW-19, and individual wells TPGW-L3 and TPGW-L5 at the 58-ft sample depth) were statistically analyzed using the Mann-Kendall trend analyses. These analyses were conducted with XLStat (Addinsoft Inc., Paris, France) using quarterly chloride and tritium data from March 2018 through September 2021 to determine whether there were statistically significant decreasing trends observed since start-up of the RWS. No trend or historical comparative analysis was conducted on TPGW-22 due to the short period of available monitoring data from March 2021 to September 2021.

Two key requirements for the appropriate application of the Mann-Kendall trend analysis are that (1) there are at least four data points in a time series, and (2) the time between samples in a data set are sufficiently large so that there is no correlation between measurements collected at different times. Both requirements were met for the analyses. Note that for the analytical data assessment, only quarterly analytical results were used in the Mann-Kendall analysis to compare equally spaced time periods and to avoid skewing the results by using RWS start-up data when more frequent sampling was conducted. For chloride and tritium analytic data, 15 quarterly samples were available for trend analysis from the March 2018 baseline sampling event through September 2021. Further reviews of trends were conducted on weekly average automated salinity data where there are more data points. Since the Mann-Kendall trend analysis recommends that the time between sampling points be sufficiently large to avoid correlations between measurements, automated salinity data from a predetermined repeating interval (Meals et al. 2011) was selected. An average weekly salinity value from midnight each Sunday from the first full week after March 1, 2018, through the first full week ending before September 30, 2021, was used for the Mann-Kendall automated salinity analysis for 21 wells with automated data. The results are shown in Table 3.2-1 and discussed below. Wells TPGW-L3-58 and TPGW-L5-58 are not included since they do not have automated probes.

Mann-Kendall analysis showed that 11 monitoring wells had a statistically significant declining trend for chloride, while 17 monitoring wells for tritium and 16 monitoring wells for salinity had statistically significant declining trends. Compared to the previous reporting period, a net of two additional wells this reporting period showed a declining trend in chloride, while three additional wells showed a decline in tritium, and one additional well showed a decline in salinity.

The fact that: 1) the findings in Year 1, Year 2, and Year 3 consistently show most wells having a continual declining trend and 2) nearly all wells continue to show lower chloride, salinity, and tritium concentrations each year, indicate positive progress in meeting the objectives of the CO and CA.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-5 In addition to the Mann-Kendall analysis, the analytic and weekly automated salinity time series data underwent a linear regression analysis (Statistix v. 10, Analytical Software Inc.,

Tallahassee, Florida) to further identify/confirm wells with statistically significant trends. The regression analysis identified statistically significant declining trends for the same wells as the Mann-Kendall analysis, plus one additional well for chloride and tritium (TPGW-15S), since the start of RWS operations. The regression analysis for automated salinity data confirmed statistically significant declining trends in 16 monitoring wells. Appendix E.1 shows time-series graphs for quarterly analytical chloride results and monthly automated salinity data for each well with linear regression statistics for chloride. Appendix E.2 shows time-series graphs for tritium.

Appendix E.3 shows the output summary from the Mann-Kendall Analysis for chloride, salinity, and tritium.

A summary of the findings from the above screening process is provided in Figure 3.2-1 (see also callout box to the right) with 20 of the 23 wells showing statistically significant declining trends. In Year 3, nine monitoring wells (TPGW-1S, TPGW-1M, TPGW-2S, TPGW-17S, TPGW-17M, TPGW-17D, TPGW-18S, TPGW-19S, and TPGW-19M) had statistically significant declining chloride and salinity trends from the start of RWS operations to the end of September 2021 for both Mann Kendall and linear regression analysis. Eight of those well (excludes TPGW-1M) had statistically significant declining chloride, salinity trends, and tritium trends since startup.

Ten other groundwater monitoring wells (TPGW-1D, TPGW-2D, TPGW-12M, TPGW-12D, TPGW-15S, TPGW-15M, TPGW-18M, TPGW-18D, TPGW-19D, and TPGW-L5-58) showed statistically significant declining trends in chloride or salinity while TPGW-2M only had a declining trend in tritium. The remaining three wells showed no statistically significant declining trends (TPGW-12S, TPGW-15D and TPGW-L3-58), but they exhibited at least one positive remediation factor, such as lowest recorded value in Year 3 compared to Years 1 and 2. (Note:

TPGW-12S is not classified as a hypersaline well, but hypersaline groundwater occurs at depths below this well so it is tracked along with co-located deeper hypersaline monitor wells TPGW-12M and TPGW-12D to assess remediation progress.)

Summary of Wells Influenced by RWS Operations.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-6 Although the majority of the wells showed declining chloride, salinity, and or tritium trends, three wells (TPGW-12S, TPGW-15D, and TPGW-18D) showed an increase in chloride or salinity, but not for both.

These three wells also had increasing trends in tritium. Groundwater at monitoring well TPGW-12S, which is located near Biscayne Bay north of the plant, had chloride concentrations around 19,000 mg/L during the reporting period. Tritium values during the same period ranged from 22 picocuries per liter (pCi/L) to 95 pCi/L, indicating that this well has little to no influence from a CCS groundwater pathway. Due to this wells proximity to Biscayne Bay and the previously reported effect of Biscayne Bay water quality at this location (FPL 2017), it is suspected that the trends at TPGW-12S are dominated by Biscayne Bay.

Furthermore, the lowest tritium concentration and the lowest weekly salinity value were reported at TPGW-12S during this reporting period despite the apparent increasing trend in chloride.

Monitoring well TPGW-15D is situated east of the RWS line of extraction wells and between the CCS and RWS-3. As a result, operations of the extraction well pull higher concentrated hypersaline groundwater from beneath the CCS toward and into RWS-3, causing the salinity and tritium levels in the deep monitoring horizon at TPGW-15D to increase. At the same time, the RWS well is reducing the vertical extent of the plume, resulting in lowering of salinity in the upper portion of the Biscayne Aquifer at this site (TPGW-15S). The increasing salinity and tritium trends were first observed in Year 2 operation of wells TPGW-15D and TPGW-15M.

However, in Year 3, salinity in the intermediate depth well TPGW-15M transitioned from increasing trends in Year 2 to no trend in Year 3 along with a declining trend in chloride. This is a possible precursor to gradual reductions in saltwater at intermediate depths at this location.

The significance of the increasing trend in tritium in monitoring well TPGW-L3-58 since 2018 is questionable at this time. During this reporting period, chloride levels reached an 11-year period of record low levels in March and June, and show no increasing trend in salinity this year or during Year 2. Last year, tritium at this station reached a 10-year period of record low and there was also no trend in the data. The Year 3 trend was heavily influenced by two period of record high tritium results in June 2021 and September 2021, which are considered anomalous. Future data will provide additional insights as to the significance of the two out of normal tritium values.

Lastly, monitoring well TPGW-18D showed an increasing salinity trend, but the chloride data showed a declining trend similar to Year 2 findings. The trend for salinity was based on an abbreviated automated data set with usable data starting in October 2018, and it does not reflect potentially higher values (based on review of chloride and field salinity values) prior to RWS Mann-Kendall trend analysis showing an increasing number of wells with statistically significant declining trends at all three depth horizons over time.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-7 start-up and the first part of RWS operation. Thus, the reported salinity trend may not be fully representative of what is happening at this well because the factors for chloride and tritium all indicate positive changes since start of RWS operation in May 2018.

In total, 20 of 23 wells showed a declining trend in one or more parameters; and 20 of 23 wells also showed at least one or more parameters that were the lowest this reporting period compared to the baseline and Years 1 and 2. The continuation of the establishment of new low values year after year indicate continued progress in plume remediation. Wells TPGW-15M, TPGW-17M, TPGW-17D, and TPGW-19S went from no trend at the end of Year 2 to a declining trend at the end of Year 3 for chloride. A similar finding was made for salinity at wells TPGW-1M, TPGW-1D and TPGW-19S and tritium at TPGW-19D. With each year of RWS operation, there is an increase in the net number of monitoring wells that show declining trends in chloride, salinity, and/or tritium in all three depth intervals.

3.3 CHLORIDE CONCENTRATION CONTOUR MAPS As requested by MDC, plan view chloride concentration contour maps were created for the shallow, middle, and deep monitoring horizons using chloride measurements from up to 22 monitoring well sites and nine CSEM chloride measurement sites for the June 2021 CSEM survey and for comparisons between chloride contour locations in 2018 and 2021. The contours (isochlors) were objectively generated by Earth Volumetric Studio, a program developed by C Tech Development Corporation using kriging algorithms. On these maps, the 90th percent confidence interval was quantified by the kriging software and graphically shown in green around the location of the monitoring point. The green area shows the distance around a measured point where the estimated chloride values can be expected to be within 10% of the measured range of chloride values. The uncertainty in the estimated chloride values increases rapidly with distance away from the 90th percent confidence area. Isochlors were generated using the kriging software and contoured for chloride levels of 1,000, 4,000, 9,000, 14,000, 19,000, and 24,000 mg/L (Figures 3.4.1 to 3.4-3). These figures were modified to clip or blank isochlors that trend into areas not supported by monitoring data or outside of the remediation compliance area east and south of the CCS.

To reduce some of the uncertainty in spatial data gaps between monitoring wells in the area between the CCS and Tallahassee Road, which covers the western extent of the hypersaline plume and CSEM survey data, chloride measurements from the CSEM survey were added for mapping purposes at nine different areas at shallow, middle, and deep layers. The CSEM chloride values were added at midpoints between monitoring wells to place the CSEM points where the monitoring well confidence levels are lowest. The added CSEM locations and their spatial relationship to the monitoring wells are shown graphically in Figures 3.4-1 through 3.4-6.

With each year of RWS operation, the net number of monitoring wells with declining trends in chloride, salinity, and/or tritium have increased in all three depth intervals indicating positive signs of remediation vertically in the aquifer.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-8 The 2021 chloride contour maps for the shallow, middle, and deep flow zones are shown on Figures 3.4-1 through 3.4-3 while Figures 3.4-4 through 3.4-6 show comparative positions of the 19,000 mg/L chloride contour for the 2018 baseline condition and the 2021 Year 3 condition along with associated bands representing the 95% confidence intervals identified in the 2018 (17,000 to 21,000 mg/L) and 2021 (16,000 to 22,000 mg/L) CSEM surveys. Comparison of the 2018 and 2021 maps show the 19,000 mg/L contour line is being retracted closer to the CCS to varying degrees and locations for all three depth horizons, which is supported by other data findings reported in this Year 3 RAASR. However, for the intermediate and deep horizons, most of the locations of the 2018 and 2021 19,000 mg/L chloride isochlor lines fall within the overlap of the 95% confidence bands for each of the surveys, meaning there is uncertainty regarding the differences in these locations. Any definitive conclusions in specific areas, however, are constrained in accuracy by the spatial distances between the existing monitoring wells, the degree that chloride concentrations change spatially, inconsistencies between the CSEM and laboratory determination of chloride concentration, the sheer size of the study area, and the assumptions of hydraulic continuity among all monitoring wells in each layer.

3.4 GROUNDWATER LEVEL TRENDS Groundwater levels in the area vary seasonally; levels are generally higher during the wet season and lower during the dry season. However, the groundwater levels can also vary daily and rise within hours of a rainfall event and, in some wells, change hourly with tides. Despite these complicating factors, groundwater contouring can provide broad insights into regional gradients, flow directions, and flow rates. Figures 3.3-1 and 3.3-2 show groundwater elevation contour maps generated from daily average automated water level data for two separate days (April 1, 2021, representing dry season conditions, and September 24, 2021, representing wet season conditions) collected from shallow monitoring wells TPGW-1S, TPGW-2S, TPGW-12S, TPGW-15S, TPGW-17S, TPGW-18S, and TPGW-19S. A single, field water elevation measurement was used for April 1, 2021, at TPGW-22S (the automated water elevation data was not available, so a field measurement was recorded on that day). The contours were developed using manual linear interpolation contouring methods and best professional judgment and informed by the above-referenced monitoring wells and additional wells (TPGW-10, TPGW-13 and TPGW-21) which are part of other monitoring efforts.

The representative groundwater contour maps for the dry (Figure 3.3-1) and wet (Figure 3.3-2) seasons indicate a generally eastward flow direction, with a slightly steeper gradient during the wet season relative to the dry season. These maps are based on measured water levels and are not adjusted for freshwater head equivalents, so care must be taken to interpret the results.

Because of the variable fluid densities in the Biscayne Aquifer, modeling tools are needed to more accurately represent groundwater flow rates, direction, and gradients.

Regionally, the groundwater levels during the dry season were higher this year in April 1, 2021, compared to the dry season in March 31, 2020. The wet season groundwater elevations at inland areas west of Tallahassee Road were higher this year in September 2021 compared to the wet season in September 2019. Continuous eastward groundwater gradients with stages equal or above sea level are generally considered helpful in reducing saltwater intrusion and aid in plume remediation.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-9 Table 3.1-1.

Monitoring Well Baseline and Year 3 Quarterly (Dec 2020 to Sept 2021)

Chloride Concentration Data.

Date Baseline Chloride (mg/L)

Year 3 Chloride (mg/L) 03/2018 12/2020 03/2021 06/2021 09/2021 TPGW-1S 19400 5960 6030 12600 5850 TPGW-1M 27700 27000 25200 24900 25100 TPGW-1D 28500 28200 28200 27400 28000 TPGW-2S 24800 17100 15800 15600 16700 TPGW-2M 29500 30100 27500 28300 29800 TPGW-2D 31300 31000 28300 30500 30800 TPGW-4S 2280 932 1420 2490 1930 TPGW-4M 15100 13100 15100 14200 15900 TPGW-4D 14800 14400 15900 14900 16400 TPGW-5S 164 148 156 154 153 TPGW-5M 11700 13000 13400 10300 10800 TPGW-5D 13100 16200 15900 13000 14000 TPGW-7S 37.0 37.6 37.4 37.6 34.6 TPGW-7M 40.0 50.5 39.3 40.2 49.3 TPGW-7D 3970 4360 5000 5590 4980 TPGW-12S 16500 18900 18600 19100 19300 TPGW-12M 20900 22300 21900 21400 22300 TPGW-12D 24000 26400 26400 26600 26700 TPGW-15S 20100 3440 6920 13000 4970 TPGW-15M 30000 26800 24400 27500 27300 TPGW-15D 28800 29600 26200 29000 29700 TPGW-17S 24900 22800 18700 21800 21400 TPGW-17M 29300 29200 25800 27400 26900 TPGW-17D 28600 29200 24300 27400 28000 TPGW-18S 14200 4990 3360 3130 2810 TPGW-18M 25200 23700 23200 20500 23500 TPGW-18D 26400 23800 24100 21000 23900 TPGW-19S 1830 1020 1070 309 1350 TPGW-19M 26000 20900 20300 20100 20800 TPGW-19D 26800 24000 24500 23400 24600 TPGW-22S NA NA 15300 13900 16600 TPGW-22M NA NA 21200 20800 22800 TPGW-22D NA NA 21000 20600 22400

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-10 Date Baseline Chloride (mg/L)

Year 3 Chloride (mg/L) 03/2018 12/2020 03/2021 06/2021 09/2021 TPGW-L3-18 2030 62.2 163 806 142 TPGW-L3-58 31400 32200 29700 28500 31700 TPGW-L5-18 1290 68.8 186 654 87.7 TPGW-L5-58 29500 29600 26100 25900 30300 TPGW-G21-18 49.2 31.5 47.7 47.5 37.7 TPGW-G21-58 7210 7610 8140 6150 7600 TPGW-G28-18 693 494 471 437 436 TPGW-G28-58 14200 13800 14300 14200 16300 Notes:

Laboratory results are reported with 3 digits although only the first 2 are significant figures.

Key:

NA = not available. TPGW-22 was added to the CA sampling requirements in March 2021.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-11 Table 3.1-2.

Monitoring Well Baseline and Year 3 Quarterly (Dec 2020 to Sept 2021)

Tritium Concentration Data.

Date Baseline Tritium (pCi/L)

Year 3 Tritium (pCi/L) 03/2018 12/2020 03/2021 06/2021 09/2021 TPGW-1S 954 127 145 347 139 TPGW-1M 2173 1926 1768 2460 2415 TPGW-1D 2307 1788 1912 1945 1887 TPGW-2S 2166 1253 1140 1292 962 TPGW-2M 3130 2539 2404 2781 2486 TPGW-2D 3123 2391 2512 2522 2472 TPGW-4S 17.4

-1.5 14.4 23.0 22.1 TPGW-4M 342 295 300 292 314 TPGW-4D 403 317 366 368 357 TPGW-5S 10.9

-6.0 3.5 4.3 17.0 TPGW-5M 271 227 214 214 180 TPGW-5D 362 349 302 305 301 TPGW-7S 6.6

-11.3 4.4

-6.6 3.0 TPGW-7M 5.2

-5.8

-10.6 17.0 5.5 TPGW-7D 20.3 24.4 38.1 32.3 31.4 TPGW-12S 46.4 22.1 39.8 52.9 95.2 TPGW-12M 931 447 295 238 424 TPGW-12D 1344 1158 1056 1156 1110 TPGW-15S 1555 110 205 617 116 TPGW-15M 2605 4412 4473 4323 4049 TPGW-15D 2509 3064 3097 3108 3066 TPGW-17S 1482 828 721 707 634 TPGW-17M 2518 1737 1679 1537 1429 TPGW-17D 2272 1895 1871 1857 1729 TPGW-18S 550

-0.60 49.0 22.2 25.3 TPGW-18M 1568 1273 1206 1259 1189 TPGW-18D 1600 1296 1319 1288 1252 TPGW-19S 42.9 32.6 40.7 17.1 79.0 TPGW-19M 864 540 548 543 494 TPGW-19D 1082 899 842 829 857 TPGW-22S NA NA 310 347 382 TPGW-22M NA NA 640 686 612 TPGW-22D NA NA 834 864 818

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-12 Date Baseline Tritium (pCi/L)

Year 3 Tritium (pCi/L) 03/2018 12/2020 03/2021 06/2021 09/2021 TPGW-L3-18 108 183 74.2 86.8 61.2 TPGW-L3-58 3014 3371 2934 4614 4358 TPGW-L5-18 86.7 57.4 61.9 50.5 42.6 TPGW-L5-58 2640 2110 2069 2036 1995 TPGW-G21-18 8.5 2.5

-1.4 8.6

-8.8 TPGW-G21-58 40.0 27.8 37.3 53.5 50.6 TPGW-G28-18 7.3

-9.0

-15.8 4.4 20.4 TPGW-G28-58 333 302 314 308 322 Key:

NA = not available. TPGW-22 was added to the CA sampling requirements in March 2021.

FPL Turkey Point RAASR Year 3 November 2021

3. Groundwater Monitoring Data 3-13 Table 3.2-1.

Assessment of Analytical Chloride, Tritium, and Automated Salinity Data from Monitoring Wells.

Period of Record low in Year 3?

Year 3 minimum <

Year 1, 2 & baseline minimum chloride?

Statistically significant declining linear regression?

Mann-Kendall chloride trend Period of Record low in Year 3?

Year 3 minimum <

Year 1, 2 & baseline minimum tritium?

Statistically significant declining linear regression?

Mann-Kendall tritium trend Period of Record low in Year 3?

Year 3 minimum <

Year 1, 2 & baseline minimum weekly salinity?

Statistically significant declining linear regression?

Mann-Kendall average weekly automated salinity trend Period of Review:

Start1 - September 30, 20212 October 2020 -

September 2021 vs.

March 2018 -

September 2020 March 1, 2018 -

September 30, 2021 March 1, 2018 -

September 30, 2021 Start1 - September 30, 20212 October 2020 -

September 2021 vs.

March 2018 -

September 2020 March 1, 2018 -

September 30, 2021 March 1, 2018 - June 30, 20212 Weekly Average Start1 to September 30, 2021 October 2020 -

September 2021 vs.

March 2018 -

September 2020 Weekly March 1, 2018 - September 30, 2021 March 1, 2018 -

September 30, 20213 TPGW-1S Yes Yes Yes Decrease Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-1M Yes Yes Yes Decrease Yes Yes No No Trend Yes Yes Yes Decrease TPGW-1D No No No No trend No No Yes Decrease Yes Yes Yes Decrease TPGW-2S Yes Yes Yes Decrease Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-2M Yes Yes No No trend No No Yes Decrease No No No No Trend TPGW-2D Yes Yes No No trend No No Yes Decrease No Yes Yes Decrease TPGW-12S No No No Increase No Yes No No Trend No Yes No No Trend TPGW-12M No No No No trend No Yes Yes Decrease Yes Yes Yes Decrease TPGW-12D No No No No trend Yes Yes Yes Decrease No Yes Yes Decrease TPGW-15S Yes Yes Yes No trend Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-15M No No Yes Decrease No No No Increase No No No No Trend TPGW-15D Yes Yes No No trend No No No Increase No No No Increase TPGW-17S No No Yes Decrease No No Yes Decrease No No Yes Decrease TPGW-17M Yes Yes Yes Decrease Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-17D Yes Yes Yes Decrease Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-18S Yes Yes Yes Decrease Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-18M Yes Yes No No trend Yes Yes Yes Decrease No No Yes Decrease TPGW-18D Yes Yes Yes Decrease Yes Yes Yes Decrease No No No Increase TPGW-19S Yes Yes Yes Decrease Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-19M No No Yes Decrease Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-19D No No No No trend Yes Yes Yes Decrease Yes Yes Yes Decrease TPGW-L3-58 Yes Yes No No trend No No No Increase TPGW-L5-58 Yes Yes No No trend Yes Yes Yes Decrease NOTES:

Wells with cells shaded in gray indicate station has had chloride values >19,000 mg/L.

Wells highlighted in blue have transitioned from hypersaline to saline in Year 1 or year 3 of RWS operation, and chloride concentrations have stayed below 19,000 mg/L.

Text highlighted in green are indications of postive RWS influence.

TPGW-22 came online in February 16, 2021 so insufficient data for comparative and trend analysis.

KEY:

1 Startup period varies: TPGW-1 to -12 started reporting around mid-2010; TPGW-15 started to report in September 2015; TPGW-17 and -19 started on January 10, 2018. TPGW-18 came online April 15, 2018.

2 Period included is first full week in March 2018 and first full week before the end of September 2021.

3 Trend based on abbreviated dataset from October 2018 through September 2021 and therefore does not reflect conditions during baseline or the first part of Year 1 operation.

- No analysis since stations do not have automated instrumentation.

Assessment:

Automated Salinity Data Analytical Quarterly Chloride Data Analytical Quarterly Tritium Data

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-14 Figure 3.1-1. Groundwater Monitoring Wells used in the assessment of the RWS system.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-15 Figure 3.2-1. Summary of Monitoring Well Influences from RWS Operations in Year 3.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-16 Figure 3.2-2. Chloride Values at TPGW-1S, TPGW-2S, and TPGW-15S.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-17 Figure 3.3-1. Dry Season Water Level Contour Map (April 1, 2021).

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-18 Figure 3.3-2. Wet Season Water Level Contour Map (September 24, 2021).

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-19 Figure 3.4-1. Groundwater Chloride Contour Map based on 2021 Shallow Monitoring Well Data and CSEM Horizon Chloride Values.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-20 Figure 3.4-2. Groundwater Chloride Contour Map based on 2021 Middle Monitoring Well Data and CSEM Horizon Chloride Values.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-21 Figure 3.4-3. Groundwater Chloride Contour Map based on 2021 Deep Monitoring Well Data and CSEM Horizon Chloride Values.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-22 Figure 3.4-4. Comparison of the 2018 Baseline and 2021 Year 3 Inland Extent of Hypersaline Groundwater (19,000 mg/L Chloride Isochlor) based on Shallow Horizon Monitoring Well Data.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-23 Figure 3.4-5. Comparison of the 2018 Baseline and 2021 Year 3 Inland Extent of Hypersaline Groundwater (19,000 mg/L Chloride Isochlor) based on Middle Horizon Monitoring Well Data.

FPL Turkey Point RAASR Year 3 November 2021 Figures 3-24 Figure 3.4-6. Comparison of the 2018 Baseline and 2021 Year 3 Inland Extent of Hypersaline Groundwater (19,000 mg/L Chloride Isochlor) based on Deep Horizon Monitoring Well Data.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-1 4 CONTINUOUS SURFACE ELECTROMAGNETIC SURVEY

4.1 INTRODUCTION

Pursuant to the FDEP CO requirements of paragraph 29(a), MDC CA paragraph 17.d.iii, and as requested in item (3b) in a letter provided to FPL by MDC dated May 15, 2017, FPL conducted the 2018 baseline CSEM Survey from March 31 through April 6, 2018, using airborne transient electromagnetic (TEM) methods (described in ENERCON 2016). The purpose of the 2018 baseline survey was to map the hypersaline plume west and north of the FPL property adjacent to Turkey Point.

Paragraph 17(d)(iii) of the MDC CA, as amended on August 20, 2019, and paragraph 29(b) of the FDEP CO required a CSEM survey be conducted 30 days after the first year of RWS operation, which was initiated on May 15, 2018. The first-year CSEM survey was conducted from May 24 through 26, 2019, and the results were presented in the November 15, 2019, RAASR. Due to restrictions on international travel and health risks associated with the COVID-19 pandemic, collection of the Year 2 CSEM survey data was delayed from the originally scheduled May 2020 timeframe until September 26-27, 2020. The Year 3 CSEM survey was conducted June 18-22, 2021. Pursuant to paragraph 20.c. of the FDEP CO, FPL is to implement a remediation project that will halt the westward migration of hypersaline water from the CCS within 3 years and reduce the westward extent of the hypersaline plume to the L-31E canal within 10 years without adverse environmental impacts. Paragraph 20.c.iii., states for determining compliance, the westward migration of the hypersaline plume shall be deemed halted if the third CSEM survey shows no net increase in hypersaline water volume and no net westward movement in the leading edge of the hypersaline plume.

Information on data collection, data analysis, error assessment, three-dimensional (3D) mapping of the distribution of hypersaline chloride concentrations within the Biscayne Aquifer, and comparisons of the 2021 results with those of the 2018 baseline CSEM survey are provided in the following sections. Plan and profile color-flood maps of bulk resistivity and CSEM-derived chloride concentrations for the 2018 and 2021 surveys as well as 3D chloride views are provided in Appendices F (2018 baseline) and G (September 2021). The Year 3 CSEM data demonstrate there is no net westward movement in the leading edge of the hypersaline plume. The data also demonstrate a statistically significant reduction of 42% in the volume of the hypersaline water in 2021 as compared to 2018, with approximately 8% of that total reduction occurring during the 8-month period between the 2020 Year 2 and 2021 Year 3 surveys. In addition, there is no CCS-Year 3 Continuous Surface Electromagnetic Survey results were compared against the 2018 baseline survey and indicate net westward migration of the hypersaline plume has been halted and the volumetric extent has been reduced by 42% after 3 years of RWS extraction.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-2 sourced hypersaline water in layers 1 through 4 (the upper 15 feet of the aquifer) west of the L-31E canal and north of the FPL property; and there has been a statistically significant net retraction of the western extent of the CCS-sourced hypersaline plume in layers 5 through 14.

4.2 APPROACH AND METHODS To collect TEM data, an electrical current is sent through a large loop of wire consisting of multiple turns which generates an electromagnetic (EM) field. The EM field switches off and on at rapid rates. When the EM field is generated, it passes into the ground where it dissipates and decays with time, traveling deeper and spreading wider into the subsurface. The rate of dissipation is dependent on the electrical properties of the subsurface, which is controlled by the material composition of the geology including the amount of mineralogical clay, the water content, the presence of dissolved solids, and the percentage of void space. At the moment the EM field is turned off, a secondary EM field, which also begins to decay, is generated within the subsurface. The decaying secondary EM field generates a current in a receiver coil. This current is measured at several different moments in time (each moment in a time band is called a time gate). From the induced current, the time rate of decay of the magnetic field, B, is determined, thus: dB/dt. When compiled in time, these measurements constitute a sounding at that location. Short time measurements present data on near-surface conditions while longer timed measurements collect data from greater depths below land surface. Therefore, data on the decay of the magnetic field over multiple progressively longer time bands break up each sounding into sequential depth layers. By maintaining a consistent elevation of the transmitter/receiver (i.e., minor flight elevation variations are adjusted during post processing) and by using consistent time gates, the thickness of each individual earth layer derived from the field data is constant across the surveyed area relative to land surface; layer thickness is thinnest near the surface and deeper layers average data over progressively greater thicknesses.

The CSEM survey area encompasses approximately 30 square miles of mostly wetlands located to the west and north of the CCS. Figure 4.2-1 presents Turkey Point, the CCS, the survey area, monitoring well sites used to correlate chloride concentrations, and 2021 CSEM survey flight lines. The 2021 CSEM survey was performed using the same airborne platform and EM technique used for the 2018, 2019, and 2020 surveys. A helicopter-borne TEM system, developed and implemented by SkyTEM Canada, Inc. (SkyTEM), provided nearly continuous (i.e., one sounding every 6 feet along each flight line) EM survey data within the coverage area.

The geophysical data are collected using TEM sounding equipment suspended from an airborne platform flown along prescribed flight lines (transects) over the target area. In this application, the individual transects primarily run from west to east with north-to-south tie lines (as shown in Figure 4.2-1) and cover the entire region of interest.

The CSEM survey measures bulk resistivity of the ground. For water-saturated materials, bulk resistivity, or its inversebulk conductivity, is principally determined by pore fluid conductivity and porosity. When porewater chloride ion content is high, bulk conductivity and fluid conductivity have a nearly 1:1 relationship. This allows the measurement of fluid conductivity from bulk resistivity or conductivity values obtained from geophysical surveys. Consequently, the high electrical conductivity of saline groundwater makes it an excellent target for electrical geophysical methods. Due to lithologic effects, the relationship between bulk electrical

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-3 properties and fluid conductivity must be calibrated with local water quality data. ENERCON established a relationship for the Biscayne Aquifer near the CCS during performance of the proof-of-concept 2016 CSEM survey as reported in PTN Cooling Canal System, Electromagnetic Conductance Geophysical Survey, Draft Final Report, Florida Power and Light Turkey Point Power Plant (ENERCON 2016). ENERCON supplemented those data with the 2018 baseline CSEM survey and 2018 water quality data as presented in Appendix G of the Recovery Well System Startup Report (FPL 2018a). The relationship between bulk CSEM resistivity and laboratory chloride content was developed for the 2021 data sets as described in Section 4.2.3 of this report. The process conducted by ENERCON to assess the vertical and horizontal extents of a hypersaline plume in the Biscayne Aquifer in the vicinity of the CCS follows the USGS method previously conducted for the Biscayne Aquifer (Prinos et al. 2014).

4.2.1 Data Processing Airborne EM data acquisition was conducted by SkyTEM during June 18-22, 2021. Raw flight data were transmitted upon landing to Aqua Geo Frameworks, Inc. (AGF) for verification and quality control. In the event flight data was found to fall outside quality specifications, the flight segment would be re-flown and verified before proceeding to the next flight segment. This field-based quality control protocol helped ensure the field data were complete and verified prior to completing the airborne data acquisition phase of the survey.

Multi-zoned groundwater monitoring wells (Figure 4.2-1 and Table 4-2.2) were sampled during June 2021, and samples were analyzed for dissolved chloride using procedures and methods described in the approved Turkey Point Quality Assurance Project Plan (FPL 2013). In addition, continuous borehole induction logging was conducted by USGS from the deepest monitoring well at each monitoring well site during March 2021. These data were used in the calibration and conversion of the CSEM data to chloride concentrations.

Following data acquisition by SkyTEM, the field data were delivered directly to AGF for post-processing. The AGF-ENERCON team conducted the data processing, interpretations, method calibration, data correlations with monitoring well induction logs and water quality and prepared the survey reports. At each sounding along a flight line, the theoretical field response of a layered earth model was calculated and compared to the actual field data and adjacent data points. The resistivities of the model layers were adjusted until the differences between the calculated (model) response and the observed field response were minimized. This spatial averaging produced laterally constrained inversions (LCIs) and spatially constrained inversions (SCIs) of the data collected. AGF produced LCIs and SCIs of data collected with the SkyTEM system during daily flight operations to verify and confirm the functionality of the SkyTEM system and eliminate drift or calibration concerns. These inversions were conducted using the Aarhus University Geophysics Workbench software that allowed for editing of the CSEM data to remove EM couplings (noise) from power lines and pipelines. AGF also provided integration of continuous borehole induction log data. The inversions were then combined into a 3D electrical resistivity model of the area which generates calibrated 3D resistivity estimates for the survey area.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-4 Table 4.2-1 lists the thicknesses of the 14 CSEM layers that account for the estimated thickness of the Biscayne Aquifer in the area based on USGS maps (Fish and Stewart 1991) and used for the 3D inversions. Layer thicknesses increase with depth as CSEM resolution decreases.

Layer 1 has a thickness of about 3 ft, while Layer 14, with a bottom depth of approximately 100 ft, has a thickness of about 13 ft. The data in this CSEM survey were inverted first to two-dimensional (2D) resistivity sections, then to 3D resistivity versus depth data. Plan and profile views of the CSEM resistivity model for the 2021 survey are presented in Appendix G, Attachments 1 and 2.

4.2.2 Quality Control of 3D CSEM Data Inversion 4.2.2.1 Magnetic Field Noise The raw field data acquired along flight lines are filtered and processed to improve data quality and reliability. The data are converted to a uniform transmitter coil height above the ground using the helicopter altimeter data, and a geographic global positioning system location is determined for each data point. An analysis is made of background EM interference (noise) that originates from sources such as thunderstorms and power lines and data points that are too noisy (where the signal is obscured by excessive background interference to a degree the data are unreliable). Those data are blanked and not included in the data inversion. The data are also examined for spikes that occur over pipelines and other conductive objects. The spikes are also blanked.

Figure 4.2-2 shows the locations of the decoupled and removed data (red lines) along the CSEM flight lines and the data used in the inversion (blue lines) in the 2021 project area. A noisy area near the RWS appeared during the 2019 survey. The noise was presumed to be associated with power delivery and operation of the RWS electric pump motors. As recommended in the 2019 report, the RWS was temporarily shutdown during acquisition of the 2020 and 2021 CSEM data.

The result was a quieter EM setting for the 2020 and 2021 surveys that resulted in less filtering and interpolation than in 2019.

4.2.2.2 Resistivity Model Verification Borehole induction logs were conducted by the USGS at each deep well within the TPGW-series monitoring sites located within the survey area. The induction logs were acquired with a single frequency EM logging tool that measures the bulk resistivity of the earth materials and pore fluids up to approximately 1 meter (m) outside the well bore. Details regarding the borehole induction logging and the result for each site are published in the 2021 Annual Monitoring Continuous surface EM mapping resistivity measurements are in close agreement with USGS induction (resistivity) logs collected from monitoring wells within the survey boundaries. Combined with the good calibration with chloride samples from multi-layered monitoring wells, this provides an independent check on the efficacy of the EM data processing method in estimating groundwater salinity concentrations within the study area.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-5 Report (FPL 2021b). The induction logs provide a continuous record of EM electrical resistivity with depth at each well where the induction log data were obtained.

The layer inversions (i.e., resistivity model) from the CSEM data were compared to the induction log data to verify that the parameters chosen in the CSEM inversion software were producing layer resistivities that are in close agreement with the borehole induction logs. Not all wells are located on flight lines; but several wells are close to or within a few hundred feet of a flight line, and they were used in the verification. CSEM resistivity sections were compared to induction logs obtained at wells TPGW-1, TPGW-2, TPGW-4, TPGW-5, TPGW-6, TPGW-12, TPGW-15, TPGW-17, TPGW-18, TPGW-19 and TPGW-22. Direct comparisons between induction log resistivity and the 3D CSEM resistivity inversions are shown graphically on the 2D resistivity profiles provided in Attachment 1 of Appendix G. Where a TPGW monitoring well induction log was performed near a CSEM flight line profile, the induction log resistivity is shown on the profile using the same color scale as the CSEM resistivity. The 3D CSEM resistivity inversions compare very well with the borehole induction logs, indicating that the 3D inversion has produced estimates of the variation of bulk resistivity versus depth comparable to values obtained in observation wells.

Another comparison between the CSEM-obtained earth resistivities and borehole induction logs is shown in Figure 4.2-3. This figure shows the relationship between resistivities from borehole induction logs and average CSEM resistivities from within 175 m of the borehole and in the CSEM layers corresponding closest to the borehole depths. The coefficient of determination (R2) for this relationship is 0.9, indicating a very strong statistical agreement between resistivities from borehole induction logs and CSEM resistivities.

4.2.3 Conversion of CSEM Resistivity to Estimated Chloride Concentrations of Ground Water Quarterly water quality data from the TPGW monitoring wells were used to develop an equation for conversion of CSEM resistivity to equivalent groundwater chloride concentration (chlorinity). The calculations utilized the relationship established between the June 2021 laboratory samples for the TPGW wells (Table 4.2-2) and CSEM resistivity. Normal seawater has a salinity of about 35 practical salinity units (PSU) and will have a chlorinity of about 19,000 mg/L. The CO, paragraph 8, and CA, paragraph 9, delineate 19,000 mg/L of chloride to be the boundary between normal salinity seawater or brackish waters and hypersaline groundwater.

Chloride concentrations > 19,000 mg/L equate to standard sea water salinity > 34.32 PSU.

The calibration of the CSEM data was conducted using a two-step approach, as presented in Fitterman and Prinos (2011) and Fitterman et al. (2012). First, a mathematical relationship was established between CSEM resistivity and the resistivity of groundwater samples from discrete depth intervals in the TPGW monitoring wells (water resistivity is the inverse of specific conductance). The mean values of the CSEM resistivities within the 175-m radius (574 ft) of each corresponding TPGW monitoring well were selected to develop a statistical range in bulk resistivities for the CSEM model layer that was at an equivalent depth to the screened intervals in the TPGW wells (FPL RAASR 2 Part 2, Appendix I, 2021).

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-6 The 2021 CSEM data are plotted on a log-log plot with the mean CSEM resistivity on the x-axis and groundwater laboratory sample resistivity on the y-axis. A regression equation is fitted to the plot to produce a power function of the form:

(1) = 0.10867 ( )1.1853 with R2 = 0.84, p < 0.001, r = 0.92 (Figure 4.2-4). R2 is the percent of the variance in the dependent variable (water resistivity) explained by the variance of the independent variable (CSEM resistivity); the p value measures the probability that the observed relationship is due to random variation and r is a measure of the correlation between groundwater resistivity and CSEM resistivity, with 0.92 indicating a very strong, nearly perfect, correlation. This is an expected strong relationship, as the groundwater samples and the CSEM resistivity soundings are taken at similar locations and depth intervals.

The second step in the calibration process is to mathematically relate chloride to water resistivity. As chloride concentration increases, water resistivity decreases. In groundwater influenced by seawater, the dominant and most conductive ions are chloride and sodium. The chloride ion comprises 55% of the total dissolved solids of sea water, so it is expected that there will be a statistically strong relationship between water resistivity and chloride ions. Again, a log-log plot is constructed with water resistivity of well samples on the x-axis and chloride concentration of well samples on the y-axis.

A regression equation is fitted to the data and has the form:

(2) = 2788 ( )1.1833 with R2 = 0.99, p < 0.001, r > 0.99 (Figure 4.2-5). Equations (1) and (2) are combined to form an equation that defines chloride concentration as a function of CSEM resistivity. This equation is then used to convert CSEM 3D inversion resistivity to chloride concentration.

4.2.3.1 CSEM System Response Shift In the 2019 RAASR report, the AGF-ENERCON team recommended using the regression equations derived from the 2016 and 2018 data in the two-step process to convert CSEM-apparent resistivity values to estimated porewater chloride content for all subsequent annual CSEM surveys. The intent was to use the 2016/2018 data as a baseline for year-to-year comparisons. An inherent assumption of this approach is that the CSEM instrument response to a given bulk earth resistivity does not change from survey to survey. The SkyTEM system firmware and software are frequently updated to reduce noise, improve signal to noise ratios, and provide better resolution of the resistivity of very shallow layers. While the CSEM instruments are calibrated at a test site, there may be small changes from year to year in the instrument response to a given earth resistivity.

In a review by MDC (Arcadis 2020) of the 2019 CSEM survey, it was recommended that a procedure be implemented to evaluate and reduce the effects of any year-to-year variations in the instrument response to a given earth resistivity. To determine if there has been any change in instrument response, an analysis was conducted of the change in the AEM bulk resistivity value

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-7 that correlates to an estimated porewater chloride value of 19,000 mg/L from 2018 to 2021 on the basis of the regression equations developed for each survey year. The results indicate that there has been a small and consistent change in instrument response through time (Table 4.2-3).

To eliminate the effects of this shift, and to comply with the MDC recommendation, the AEM survey results in each survey year are calibrated against water samples from that year. This procedure produces a unique AEM resistivity to chloride relationship for each survey, eliminating the consequences of any shift in instrument response that may occur from year to year.

Accordingly, the AGF-ENERCON team used the 2021 calibration between 2021 CSEM resistivity and 2021 water quality data to derive estimated porewater chloride values from CSEM resistivity data. This procedure normalizes the instrument response from year to year, produces an independent AEM resistivity-to-chloride relationship, and allows a valid 2021 survey year to 2018 baseline comparison of the distribution of porewaters with chloride content >19,000 mg/L.

This procedure also was used for the 2020 survey to address the recommendation provided by MDC (Arcadis 2020) to initiate a procedure to address any potential year-to-year variation in the response of the SkyTEM system. As the AEM-derived chloride distribution for each year is independently calculated, the instrument response drift does not impact the volume calculations or AEM-derived chloride distribution.

4.2.3.2 Impact of Monitoring Site TPGW-22 on the Calibration Data Set During 2021, a new groundwater monitoring site, TPGW-22, installed by MDC in 2019 (Figure 4.2-1), was included in the groundwater remediation monitoring network. The site contains three monitoring wells with screened intervals corresponding to the upper, middle, and lower flow zones within the Biscayne Aquifer. Water quality samples were collected from TPGW-22 in June 2021. The chloride and water resistivity data for TPGW-22 were initially included in the calibration procedure for deriving the relationship between AEM bulk resistivity and estimated porewater chloride content. TPGW-22 data were not used in the 2018, 2019 and 2020 calibration data sets, as the well site was not part of the groundwater remediation monitoring network when those surveys were completed.

It was noted that including the TPGW-22 data in the calibration data set results in a downward shift in the regression line between AEM resistivity and measured porewater resistivity. The location of each of the three TPGW-22 data points on the regression plot was below the regression trend line (Figure 4.2-6) causing a relatively small downward shift in the resulting Introduction of monitoring data not included in the original 2018 baseline monitoring network results in shifts in the CSEM resistivity versus porewater resistivity regression equation. Such shifts in the regression affects the hypersaline voxel volume, creating discontinuities between hypersalinity volumes calculated before and after any new monitoring station is added. To resolve this, annual volumetric hypersalinity calculations are based on regressions calculated using year-specific porewater chloride data collected from the original 2018 baseline monitoring network.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-8 trend line and regression equation. The effect of the small shift is to slightly increase the AEM-derived chloride concentration at any given point relative to concentrations obtained without the addition of the TPGW-22 data. In aggregate, these relatively small increases result in a slightly larger AEM estimated volume of hypersaline groundwater within the compliance area in Year 3 than was calculated for Year 2 based on regressions made without TPGW-22 data (Figure 4.2-7).

There was no new volumetrically significant source of hypersaline water documented during the reporting period, given the operation of the RWS, the freshening of the CCS, and the measured salinity levels in the monitoring well network. The calculated increase in the volume of hypersaline water that occurred when the TPGW-22 data was added to the Year 3 monitoring network is associated with the shift in the regression equation due to addition of new data points that were not included in prior year regressions. This explains why the calculated hypersaline plume volume with TPGW-22 data included does not agree with the trend in calculated volumes from 2018 through 2020 (Figure 4.2-7).

To understand the shift in the regression trend line and the resulting effect on hypersaline volumes, it is important to note that the AEM-to-chloride conversion relies primarily on the following assumptions: (1) chloride is the dominant conductor in the Biscayne Aquifer, and (2) porosity within the Biscayne is isotropic and homogeneous. Departures from these primary assumptions are considered geologic noise and result in data departures from the regression trend line. As the number of monitoring points (wells) included in the regression is relatively small, additional data points that land off the trend line (due to geologic noise) can move the line and affect the regression equation. The TPGW monitoring points used in the baseline calibration data set (used 2018 through 2020) comprise data that are scattered geometrically around the regression trend line. The locations of these data points tend to fall in the same general areas of the plot from survey to survey. This suggests that the effect of the geologic noise on the AEM resistivity to lab chloride is consistent, even as the chloride concentration changes at the monitoring locations. The vertical position of data points on the regression line plots is principally determined by porosity, with water quality samples from lower porosity zones plotting above the regression line and data from higher porosity zones plotting below the regression line. The regression line in effect yields an average porosity value.

Examination of trend lines for regressions of AEM resistivity against porewater conductivity with and without TPGW-22 data included (Figures 4.2-6 and 4.2-4) illustrate that the three porewater resistivity values for TPGW-22 result in data points that all plot below the regression line, pulling the trend line down slightly. This results in a slight decrease in the AEM bulk resistivity value that correlates with 19,000 mg/L chloride and increases the calculated volume of the hypersaline plume. The shift in the regression equation and resulting volume calculation is artificial, and solely due to the suspected higher than average porosity at the new well location and the limited data set that gives relative weight to the new data. If, for example, monitor well TPGW-22 had been present during the baseline acquisition in 2018, the calculated baseline and Year 1-3 volumes would have been proportionally higher, and the declining trend would mirror the trend shown in Figure 4.2-7. Consequently, if chloride data from new monitoring wells that were not included in the 2018 baseline calibration procedure are included in subsequent years AEM-monitor well resistivity regressions, the new well data can move the regression line up or

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-9 down in a different manner than a regression based on the original group of monitoring wells, which makes resulting hypersalinity volumetric comparisons to the 2018 baseline non-sequitur.

To consistently assess the effect of the RWS on AEM estimated hypersaline pore-water volume, the year-to-year AEM survey calibration must use the same wells as used in the baseline survey.

For this reason, the 2021 survey calibration does not explicitly use TPGW-22 water quality data for calibration. However, TPGW-22 data are used in comparisons with borehole induction logs, on vertical profiles of AEM estimated pore-water chloride, and (x, y) contour and color-flood maps of AEM estimated pore-water chloride values.

Recalculating the regression line between AEM resistivity and porewater resistivity without including TPGW-22 data results in a volume calculation that confirms the observed trend in volume from 2018-2020 (Figure 4.2-7). In addition, groundwater monitoring data trends from September 2020 through September 2021 demonstrate continued salinity reductions that support the CSEM results produced from the 2021 AEM resistivity/porewater resistivity regression excluding the TPGW-22 data. Using the wells in the original calibration set as was used in 2018, 2019, and 2020, results in a 42% decrease in the volume of the hypersaline plume compared to the baseline 2018 volume. Again, this volume closely fits the trend in volume observed from 2018-2020 and aligns with trends observed in physical processes.

4.2.3.3 2021 Chloride Conversion The plot of CSEM derived chloride concentration (x-axis) and lab-determined chloride concentration (y-axis) (Figure 4.2-8) produces a regression equation with values of R2 = 0.65, r =

0.81, and p < 0.001. As previously described, R2 is the percent of the variance in the dependent variable explained by the variance of the independent variable, the p value measures the probability that the observed relationship is due to random variation, and r measures the strength of the correlation between CSEM-determined chloride concentration and lab-determined chloride concentration, with an r of 1.0 being a perfect correlation (i.e., 1:1).

The correspondence of chloride concentration calculated from CSEM resistivity and lab-derived values of chloride concentration from TPGW wells is graphically illustrated by superimposing the TPGW well-derived chloride values on CSEM-derived chloride concentration versus depth profiles and using the same color-coded contour intervals (see Attachment 3 of Appendix G).

On the vertical profiles, the agreement between the CSEM-predicted chloride value and the chloride value obtained from samples taken from the monitor wells can be visually compared. A quantitative comparison between CSEM-estimated chloride and lab-determined chloride at monitor wells is found in Table 4.3-1. Note that the sample area of each CSEM-estimated chloride value is different in every case from the corresponding sample area of the monitoring well. Accordingly, at any given point, the correspondence is affected by many factors, including porosity variations, the location of the screened interval with respect to the assigned CSEM layer, distance between monitoring well locations and flight lines, and EM noise.

Monitoring well chloride data collected in 2021 are also posted on the 2021 CSEM color-flood depth layers for comparison (see Appendix G-4). The monitoring well data are posted only on the applicable CSEM layers where the well-screened interval is contained within the layer.

Monitoring well chloride data for well screens that are at or across a layer interface are posted on

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-10 both layers. Although TPGW-12, TPGW-15, and TPGW-22 chloride values were not included in the CSEM calibration data set due to excessive EM noise or addition of a new station without historic data, they are posted on the appropriate CSEM layer maps.

Monitoring well data are from the Biscayne Aquifer only. Consequently, the calibrated equation relating CSEM resistivity to groundwater chloride concentrations is constrained to the data collected from the CA/CO monitoring network wells. For this reason, use of this empirically derived relationship between CSEM resistivity and chloride concentrations should not be applied to CSEM resistivities from geologic units below the Biscayne Aquifer not monitored under the CA/CO monitoring network.

4.2.4 CSEM Method Uncertainty Analysis The CSEM results define the extent of the hypersaline plume both horizontally and vertically.

The 95% confidence interval for the 19,000 mg/L contour for 2021 is about 16,000 mg/L to 22,000 mg/L. This uncertainty in the horizontal plume position is illustrated in Appendix G-6 for each of the CSEM layers. Appendix G-6 shows the 19,000 mg/L contour in 2018 and 2021 and the horizontal range for estimated chloride values between 16,000 and 22,000 mg/L.

However, the estimates of porewater chloride levels derived from CSEM resistivity do not always match measured chloride levels from discrete monitoring intervals. An important source of differences between CSEM-estimated chloride and laboratory-determined chloride is the assumption of constant porosity. The equation that relates CSEM bulk resistivity to chloride uses the regression equation derived from a plot of CSEM bulk resistivity vs. porewater resistivity measured at wells. This equation uses the best-fit relationship between CSEM bulk resistivity and measured porewater chloride levels. In areas with higher-than-average porosity, the CSEM estimates will overestimate chloride; and in areas with lower-than-average porosity, the CSEM estimates will underestimate chloride. Variation of porosity horizontally and vertically introduces some geologic noise to the CSEM chloride estimates. As there are no data on porosity variation in the survey area, it is difficult to assess the effect of porosity variations on estimated chloride values. However, this geologic noise is included in the overall assessment of the error in CSEM chloride estimates. The additional data collection and modeling FPL performs complements the CSEM survey and, together, provides certainty as to the remedial action progress.

Another source of noise is the spatial displacement of the water quality samples taken from discrete intervals in the monitoring wells and the location of the nearest CSEM data on a flight line. None of the monitoring wells available for calibration of the CSEM data are on a flight line, all well screens are shorter than the thickness of CSEM layers, and some well screens are divided by two CSEM layers (see Figure 4.2-9). In addition, the CSEM data average the instruments response to variations in chloride content over distances of a few tens of meters to over 100 m, while well samples come from small-diameter well screens about 1 m in length. It can be reasonably assumed that lateral changes in porewater chloride content are smooth and not abrupt over distances of tens of meters, and the effects of spatial errors are included in the overall assessment of CSEM estimated chloride values.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-11 The differences between sample-based and CSEM-based chloride estimates can be quantified.

The resulting apparent error can be used to identify and distinguish changes in chloride levels attributable to this statistically determined error from those changes that are significantly and statistically different between survey years. The evaluation of method error and its application in identifying statistically significant increases or decreases in hypersaline groundwater are discussed below.

4.2.4.1 Definition of CSEM Survey Error The basis for evaluating the accuracy of the CSEM surveys is the aquifer porewater chloride data obtained from the TPGW multi-level monitoring wells in and near the CCS. Each well cluster has three screened intervals corresponding to the upper, middle, and lower flow zones within the Biscayne Aquifer. Water quality data are available from the TPGW monitoring well system.

For comparison to CSEM data, water quality data closest in time to the CSEM survey were selected for wells that are close to a CSEM flight line (i.e., there are no wells on a flight line) and at screen intervals that correspond to CSEM layers. Where the well screen elevations span two CSEM layers, the average of the two CSEM layers is used for comparison. Figure 4.2-9 shows relative position and length of monitoring well screen (shown in red) vs. CSEM layers. Wells located near CSEM flight lines provide 27 water quality samples obtained in 2021 for comparison to the CSEM data. The water quality samples used in the error analysis are the same samples used to establish the baseline relationship between CSEM bulk resistivity and porewater chloride content, as these samples were used to establish the transfer function converting CSEM resistivity to chloride concentrations, as described in Section 4.2.3.

The CSEM estimates of porewater chloride levels are obtained as described in Section 4.2.3.

The error between CSEM-estimated chloride values and chloride levels obtained from monitoring wells is defined as the ratio of the CSEM-estimated chloride value to the closest monitoring well value:

= [ ]/[ ]

A perfect correlation between CSEM estimates of chloride and monitoring well data would have a value of 1.0. Values > 1.0 indicate an overestimation of chloride by the CSEM data, and values < 1.0 indicate an underestimation of chloride by the CSEM data.

An alternative error measure is the algebraic difference between the CSEM estimates and the water quality data:

= [ ] [ ]

However, the magnitude of this difference error varies with the chloride level, larger for high chloride values and smaller for low chloride values. The error defined as a ratio maintains the There is a 95% probability that the actual location of the 19,000 mg/L contour lies within the area bounded by the 16,000 and 22,000 mg/L contours.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-12 relative error between CSEM and water quality data across a wide range of chloride values. It is similar to a percent error, where:

% = 1

[ ]

[ ]

100 4.2.4.2 Distribution of Error An evaluation was performed to determine whether there is a bias in the distribution of error.

The error distribution is close to a normal distribution (Figure 4.2-10). For 2021, the relative error of AEM chloride estimates was assessed using 2021 data only to reduce any effects of changes in instrument response that may have occurred between 2018 and 2021. The mean error of the 2021 data is 0.89 and the standard error is 0.07. The equation for the 95% confidence interval is CI95 = (average error) * (1.96) * (s/n0.5) where s is the standard error and n is the number of samples. For the 2021 data set, n is 27, and the 95% confidence interval is 0.154, or about +/-15%. As applied to the 19,000 mg/L estimated chloride value, the 95% confidence interval is +/- 2,926 mg/L, or 16,074 mg/L to 21,926 mg/L.

For mapping purposes, these values were rounded to 16,000 to 22,000 mg/L. Plan-view maps showing the 19,000 mg/L contour and associated area within the 95% confidence interval for the 2018 and 2021 surveys are contained in Appendix G-6. Where the confidence intervals for the 2018 and 2021 19,000 mg/L contours do not overlap, there has been a statistically significant movement of the 19,000 mg/L contour between 2018 and 2021.

4.2.4.3 Global Error For 2021, point-to-point comparisons of CSEM estimated chloride values have a 95%

confidence interval of about +/-15%. For a CSEM chloride estimate of 19,000 mg/L, a +/-15%

interval is from about 16,000 mg/L to 22,000 mg/L. For global estimates, as the error in the CSEM-estimated porewater chloride content at a point is approximately normally distributed, the positive and negative errors often cancel. Consequently, for error summations such as the volume of the hypersaline plume within a layer or the volume of the entire plume, the CSEM estimate error is related to the mean error of 11%.

4.2.5 Method Minimum Reliable Chloride Concentration Although the CSEM time-domain resistivity survey at Turkey Point was designed to map the extent of aquifer porewaters with > 19,000 mg/L chloride concentration, FPL was asked by MDC to assess the method reliability for portions of the aquifer containing lower salinity concentrations. The 2018 Recovery Well System Startup Report (FPL 2018a) provides a discussion on the derivation of the minimum reliable chloride reporting limit. As discussed, prediction accuracy decreases significantly for porewater chloride ion contents < 10,000 mg/L.

Based on this analysis, the reliable lower limit of the CSEM survey for mapping chloride concentration within the Biscayne Aquifer is 10,000 mg/L.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-13 4.2.6 Creation of a 3D Chloride Ion Voxel Grid A voxel is a 3D grid cell, or volume element. The CSEM-derived chloride values were interpolated to a uniform voxel grid to allow for more effective graphical visualization of the chloride ion distribution. Each voxel has lateral (x, y) dimensions of 328 ft x 328 ft (100 m x 100 m) and a thickness equivalent to the individual 3D CSEM resistivity layers (Table 4.2-1).

The bottom of layer 14 is at a depth of about 100 ft below land surface (30.3 m). As a result of the interpolation process to develop the voxel model, CSEM-derived chloride concentrations near monitoring wells located in blanked areas (due to excessive background EM noise) can be compared to water quality data obtained from those wells and used to assess remediation progress. For example, TPGW-12 and -15 were excluded from the CSEM bulk resistivity model because there were no usable data within the radial criterion of 175 m (due to EM noise and proximity of flight line data). However, these data points are useful for comparison with AEM-derived chloride in the plan-view and cross-sectional maps.

Depth slices, profile views, and 3D views of the CSEM-derived chloride concentrations are provided in Appendix G for the 2021 survey. Chloride concentrations between 10,000 mg/L and 19,000 mg/L are shown in shades of gray; chloride concentrations above 19,000 mg/L are shown with a colored scale, with red representing the highest concentrations (up to ~40,000 mg/L) and blue representing the lowest concentrations (~19,000 mg/L). An example of a chloride depth slice is shown in Figure 4.2-11, representing CSEM layer 12.

4.3 DISCUSSION OF FINDINGS 4.3.1 Natural Occurrence of Hypersaline Water Two sources of hypersaline groundwater occur within the CSEM survey area adjacent to the CCS. The predominant source is CCS groundwater while the other source is naturally occurring non-CCS sourced evaporated seawater that originates in the coastal wetland margins referred to as the white zone and documented by the USGS (Prinos et al. 2014). Salinities exceeding 40 PSU (> 22,000 mg/L) have been documented to occur in coastal waters in western Florida Bay and Taylor River, well outside of any influences from the CCS (SFNRC 2012), as well as north and south of the CCS as discussed in Section 4.3.3.2 below. Hypersaline surface water with fluid densities greater than underlying groundwater will sink into groundwater, resulting in both shallow and deep expressions of hypersaline groundwater. This is significant to the RWS remediation assessment as the CA and CO do not require FPL to extract naturally occurring hypersaline groundwater.

CSEM data demonstrate there is no net westward movement in the leading edge of the hypersaline plume. The data also demonstrate a statistically significant reduction of 42% in the volume of the hypersaline Biscayne Aquifer materials in 2021 as compared to 2018. In addition, there is no CCS-sourced hypersaline water in layers 1 through 4 (upper 15 feet of the aquifer), and there has been a statistically significant net retraction of the western extent of the CCS-sourced hypersaline plume in layers 5 through 14.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-14 Fitterman et al. (2012) used helicopter electromagnetic (HEM) surveys to map the distribution of saline groundwater in the C-111 and Model Lands basin areas of southeast MDC. The HEM data were presented as resistivity-depth profiles. Comparison of geophysically determined formation resistivity and salinity concentrations from well samples (Fitterman and Prinos 2011) shows that formation resistivities of 1 to 2 ohm-m represent geologic units saturated with groundwater close to or at normal seawater chloride concentrations of 19,000 mg/L. Formation resistivities with values of 1 ohm-m or less represent hypersaline groundwater with chlorinity >

19,000 mg/L. Fitterman et al.s (2012) HEM data show that at a depth of approximately 17 ft (5 m), hypersaline groundwater is present between Card Sound Road and U.S. Highway 1 (US 1) in a coast-parallel band 4,000 to 6,000 ft wide. The hypersaline groundwater in Fittermans coastal band is not from the CCS as there is no mechanism for dense hypersaline water to travel westward along the coast for a distance of 4 miles at a depth of only 17 feet. This hypersaline water corresponds to a coast-parallel zone of lower vegetative density in the coastal wetlands as viewed from satellite images known as the white zone. It is common in coastal wetlands in this area for evaporation of seawater to form hypersaline groundwater that moves downward into the sediments under a density gradient (Prinos et al. 2014). Salinities in shallow groundwater in coastal wetlands can reach 60 to 100 PSU (34,000 to 56,000 mg/L) (Stringer et al. 2010) and will migrate downward due to the increased density as compared to normal seawater. Close to the coast, evaporation of seawater can create a wide band of hypersaline groundwater. The HEM data of Fitterman et al. (2012) suggest that this band of naturally created hypersaline groundwater extends to the base of the Biscayne Aquifer between Card Sound Road and southwest past US 1. This band of naturally occurring hypersaline groundwater forms in the coastal margin both north and south of Turkey Point, as has been documented by surface water salinity monitoring stations in the Everglades Mitigation Bank and by CSEM surveys.

4.3.2 Spatial Extent CSEM-Derived Chloride Concentrations Color-flood maps that illustrate the 2D plan-view variation in CSEM-estimated chloride content of groundwater (i.e., representation of groundwater contours utilizing CSEM) for the 2021 survey (Year 3) are provided in Attachment 4 of Appendix G. As discussed in Section 4.2.4.3, the 95% confidence interval of +/-15% provides an estimate of the uncertainty of the location of the 19,000 mg/L contour. There is a 95% confidence that the 19,000 mg/L contour is within an area bounded by the 16,000 and 22,000 mg/L contours. It is not possible to illustrate the +/-15%

error zone on these maps. This makes interpretation of any differences in the 2018 and 2021 hypersaline boundaries using color-flood maps difficult to assess. Nevertheless, the color-flood maps provide insights into changes in chloride concentrations over time, which is also helpful in assessing impacts of the RWS on plume reduction. To better assess changes in the spatial extent of the hypersaline plume edges, contoured difference maps were created, as described below, to compare the 2018 baseline results with Year 3 (2021). These comparisons are shown for each CSEM layer in Appendix G-6, and an example is shown below in Figure 4.3-1.

As noted above, the 19,000 mg/L +/-15% error band is about 16,000 mg/L to 22,000 mg/L. In the plan view, this creates a zone of uncertainty with the 19,000 mg/L contour approximately in the middle of the error band. Where the 2018 uncertainty zone overlaps the 2021 zone, the separation of the 2018 and 2021 19,000 mg/L contours is not statistically significant at the 95%

confidence interval (i.e., there is a < 5% chance that the separation is due to random error).

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-15 Where the error zones for 2018 and 2021 do not overlap, the separation between the two contours is statistically significant.

4.3.2.1 Spatial Comparison and Volumetric Determination Methodologies As described above, the CSEM porewater chloride estimates were interpolated to a voxel grid with horizontal dimensions of 100 m x 100 m for each grid cell. The thickness of each cell is the thickness of a given CSEM layer (see Table 4.2-1). The voxels with estimated chloride values

> 19,000 mg/L can be counted and their volumes calculated. This allows an estimate of the volume of the hypersaline plume (> 19,000 mg/L) to be made. This comparison of hypersaline volume can be made layer by layer or for the entire thickness of the Biscayne Aquifer. The error analysis suggests that this summation should have an accuracy of about +/-11%.

4.3.3 Comparison of the 2018 and 2021 CSEM Survey Results 4.3.3.1 CSEM Layers and High Flow Zones Monitoring wells within and surrounding the Plant are constructed into high permeability zones in the upper, middle, and lower Biscayne Aquifer that are based on review of pilot well lithologic and borehole geophysical data. As the elevations of the screens in the monitoring wells vary in depth and the elevations of the CSEM layers are constant across the survey area, the CSEM layers that represent the upper, middle, and lower flow zones of the Biscayne Aquifer can vary.

Table 4.3-1 lists the TPGW monitoring wells used in the chloride calibration procedure (except for TPGW-12 and -15 located in areas affected by EM noise), and the CSEM layer associated with the screened interval. For example, the upper flow zone is present in CSEM layers 6 through 8, the middle flow zone is present in CSEM layers 10 and 11, and the lower flow zone is present in CSEM layers 12 through 14. The three CSEM layers that include the most upper, middle, and lower monitoring well intervals are CSEM layers 7, 10, and 13 respectively.

4.3.3.2 Comparison of 2018 and 2021 Plan View Contour Maps of Estimated Chloride Plan-view hypersaline boundary (19,000 mg/L) contour maps were prepared for each layer with the 2018 and 2021 CSEM surveys. These maps are provided in Appendix F-5 and Appendix G-5, and selected figures are presented at the end of this chapter. The maps also show the voxels in each layer that have estimated CSEM salinities of 16,000 to 22,000 mg/L for 2021. These voxels represent the approximate 2021 95% confidence intervals for the 19,000 mg/L contours.

The 95% confidence range for the 2018 data is 17,000 to 21,000 mg/L. Where the 2018 and 2021 confidence intervals do not overlap, the separation between the 2018 and 2021 contours is statistically significant at the 95% confidence level. Where they do overlap, any movement of the 2021 19,000 mg/L contour is not statistically significant at the 95% level.

Comparing the 2018 baseline survey to the 2021 survey, the volumetric extent of the plume has shown a statistically significant reduction of 42% from 2018 to 2021. Overall, the Year 3 survey shows there are net reductions in the extent and total volume of CCS-sourced hypersaline groundwater within the compliance area, in all layers compared to the 2018 baseline condition that existed prior to the RWS startup in May 2018. The survey indicates there is no CCS-

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-16 sourced hypersaline groundwater west of the L-31E canal or north of the FPL property in the upper 15 feet (4.7 m) of the aquifer in layers 1 - 4. Reductions in layers 1 - 3 and increases in layers 4 and 5 are caused by non-CCS sourced hypersaline surface water formed by evaporation of the Hurricane Irma storm surge that has been sinking through the upper Biscayne Aquifer over the past several years in the northeast corner of the compliance area. Layers 6 - 14 show a volumetric reduction in CCS-sourced hypersaline groundwater, the amount of which varies from layer to layer based on several factors, including the degree to which hypersaline groundwater existed in the layer, the hydraulic characteristics of the layer, the distance between the RWS wells and the edge of the plume, and the concentration of the plume. Regarding the spatial extent of the plume, all CSEM layers that identified CCS-sourced hypersaline groundwater within the compliance area in the 2018 baseline survey have shown net movement of the western and northern plume extent back toward the CCS/RWS extraction wells. However, there are localized areas in each layer where the leading edge of the plume does not show statistically significant retraction toward the CCS; and in some areas, there is small movement to the west or north. These deviations may be the result of localized heterogeneity in the aquifer that will be resolved over time as the remediation continues. The 2021 survey also identifies reductions in chloride concentration within the plume that indicate a progressive reduction of salt mass within the plume associated with the RWS operations. A more detailed description of changes to notable layers follows.

Layers 1 through 5: North of CCS A shallow zone of > 19,000 mg/L aquifer porewater has previously occurred on the north-east side of Palm Drive and the east side of the L-31E levee. The shallow zone of hypersalinity layers 1 through 3, was observed in the 2018 baseline survey, which occurred approximately 6 months after the Hurricane Irma storm surge of 3-5 ft inundated the coastal reaches of the Model Lands basin. Seawater that flooded the coastal wetlands up to the L-31E levee and the Turkey Point entrance road north of the Plant was trapped and became concentrated via evaporation during the dry season. This evaporative-sourced hypersaline groundwater was limited to the upper 7 to 10 ft of the aquifer as shown in 2018 CSEM survey layers 1-3 (2019 RAASR, Appendix E Figures 1a, 2a, 3a). At the same time, at depths of 20 to 25 feet (CSEM layers 5 and 6), groundwater salinities were lower with chloride levels similar to or less than seawater. From 2018 to 2021, this more-dense, naturally sourced hypersaline groundwater has migrated downward under a density gradient, from layers 1, 2 and 3 to layers 4 and 5 resulting in temporary increases in non-CCS sourced hypersaline aquifer volume (see Appendix G-4 and Table 4.3-2 below). In layers 4 and 5, this shallow hypersaline water is not vertically connected to the main hypersaline plume at depth just north of the CCS (see Appendix G.3: 2D Chloride Concentration Profiles). The CSEM data do not show continuity of hypersalinity in the inundated area with the CCS-sourced hypersaline plume. It is expected that the volume of hypersaline aquifer material in layers 4 and 5 will decrease as the evaporative sourced plume sinks to deeper elevations of like salinity. As this northeast portion of the compliance area is tidally influenced, periodically producing hypersaline waters in the upper layers that migrate downward, it is expected that the measured hypersaline volumes by layer will continue to change cyclically. A new monitoring well cluster, TPGW-23, will be constructed on the west side of the L-31E canal north of TPGW-19. This monitoring site may provide additional data to further evaluate the source of hypersaline groundwater in this area.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-17 Layers 1 through 5: South of S-20 Similar to the coastal area north of the CCS, overland infiltration of marine water from Card Sound has been observed in the coastal wetlands south of the S-20 Discharge Canal and the L-31E Levee associated with storm surges, king tides. In the 2018 baseline CSEM survey, hypersaline water was mapped in the upper 3 to 6 feet of sediment (layers 1 and 2) after Irmas storm surge flooded the area in November 2017. However, at depths from 6 to 9 feet the salinity in the area was below 19,000 mg/L Cl. By 2019, some of the hypersaline water from Layers 1 and 2 had moved into layer 3. A similar event occurred in this area in September 2020 when hypersalinity was identified in layers 1 and 2 south of the S-20 Discharge Canal. In the subsequent June 2021 survey, the hypersaline groundwater in layers 1 and 2 had dissipated and increased in layer 3. Storm water releases from the S-20 structure provide freshwater flow that flush saltwater out of the area during the wet season. Beginning on October 8, 2019, SFWMD implemented a temporary deviation for the S-20 that increased the headwater stage when storm water releases were made. As a result of extended dry conditions and the temporary deviation, no substantive stormwater freshwater releases from the S-20 to this area were made from mid-September 2018 until mid-November 2020. Reductions to the frequency and duration of S-20 freshwater releases to this area will result in more frequent conditions favoring hypersalinity which impacts both the groundwater remediation efforts of FPL (by increasing the volume of non-CCS sourced hypersaline groundwater FPL removes and disposes) and wetlands that are being impacted by high salinity conditions.

Layer 7 Layer 7 most closely represents the upper flow zone in the Biscayne Aquifer near the CCS (see Figure 4.3-1). This layer shows a 74% reduction in the volume of hypersaline water between the 2018 and 2021 CSEM surveys (Table 4.3-2) and indicates there is no hypersaline water west of L-31E for a distance of approximately 1.5 miles south of the northwest corner of the CCS. This retraction is statistically significant. In the southern portion of the 2021 CSEM survey, there is a statistically significant eastward retraction of the 19,000 mg/L boundary from 2018 to 2021 ranging between 400 to over 1,200 m (0.25 to 0.75 mile).

Layer 9 Overall, layer 9 (Figure 4.3-2) shows a 40% reduction in the volume of hypersaline water from 2018 to 2021 with much of the discontinuous hypersaline area in the southwest survey area remaining unchanged while significant retraction has occurred in the area adjacent to the CCS.

Retraction of the contiguous hypersaline plume in the range of 1,300 to 2,600 feet have occurred west and north of the CCS during the first 3 years of remediation.

A broad area 1 to 2 miles west of TPGW-2 and TPGW-17 shows convoluted discontinuous patterns of hypersalinity in both the 2018 and 2021 surveys (Figure 4.3-2). As shown by the underlying 16,000 to 22,000 mg/L confidence intervals, nearly the entire area has estimated chloride values between 16,000 and 22,000 mg/L, within the range for normal Biscayne Bay salinity. There is little discernible statistically significant change in the position of the 19,000 mg/L contour between 2018 and 2021, and the contours are convoluted. Lithologic, acoustic imaging, and seismic borehole logs for TPGW-2 and TPGW-4 (JLA 2010) and data described in

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-18 Fish and Stewart (1991) (E-E geologic cross section) suggest that at least the upper part of layer 9 is a lower permeability sandstone or sandy limestone near the bottom of the Ft. Thompson Formation. The drillers log for TPGW-22, shows extensive silty materials. Silty materials have inherently higher porosities compared to bedrock, which may result in CSEM estimates of salinity being higher than actual salinity. The possible lower permeability of layer 9 may contribute to the broad area of slightly hypersaline to slightly less than hypersaline aquifer porewaters in the SW corner of the survey area. This is supported by the hypersaline CSEM chloride estimates recorded in the vicinity of well TPGW-4M from 2018 through 2021 while lab chlorides for this well from September 2010 to present have consistently averaged around 15,000 mg/L.

Layer 10 Layer 10 (Figure 4.3-3) most closely represents the middle flow zone of the Biscayne Aquifer near the CCS. Overall, the 19,000 mg/L contour shows a 33% statistically significant retraction of the hypersaline plume toward the RWS between 2018 and 2021. In some areas of layer 10, the confidence intervals for 2018 and 2021 19,000 mg/L contours overlap; however, the plan view, color-flood chloride map for 2021 shows large areas of reduced chlorinity since 2018, indicating additional statistically significant reduction to the hypersaline plume can be expected in the layer in the near future (see Appendix G-4).

East of TPGW-5, there is a roughly circular area about 1,100 m in diameter that suggests a statistically significant increase between 2018 and 2021 in an isolated hypersaline zone. This apparent expansion cannot be currently explained as it is discontinuous with the CCS-sourced hypersaline groundwater, and there is no known source of hypersaline water. There is also a small statistically significant area north of TPGW-19 where the plume has expanded. This is unusual as it occurs proximal to the RWS-1 extraction well and in an area where there are statistically significant retractions in the northern extent of the plume in layer 12. As part of the annual data collection and review, FPL will continue to monitor and evaluate these areas.

Layers 11 and 12 Layers 11 and 12 showed a 44% and 39% reduction, respectively, of the hypersaline plume since 2018; and much of the area where hypersaline voxels dropped below the 19,000 mg/L chloride was over 1 mile west of the CCS. This is significant as it indicates the remediation is working in the lower portion of the aquifer and at significant distances west of the CCS. Factors that could be contributing to reductions in chlorinity at distances further west than the capture radii of the RWS extraction wells include the prevailing east-southeasterly groundwater gradient, the halting of net westward flow of hypersaline groundwater from beneath the CCS, and reduction in fluid density west of the CCS (due to decreasing salinity concentrations), thus facilitating increased inflow of fresher groundwater from the west.

Layer 13 Layer 13 (Figure 4.3-4) most closely represents the lower flow zone within the Biscayne Aquifer. While there are no large areas in CSEM layer 13 with statistically significant movement of the hypersaline interface between the 2018 and 2021 CSEM surveys, over most of

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-19 the CSEM survey area, the 2021 19,000 mg/L contour is east of the 2018 contour in layer 13; however, these shifts are not significant based on the 95% confidence interval. There is an anomalous area on the south side of the Florida City Canal at the northwest corner of the CSEM survey. This area is separated from the main hypersaline plume by more than 4,000 m. This apparent high-chloride area does not appear in CSEM layers 12 or 14. This roughly circular area is considerably smaller in the 2021 results than in 2018. The source of this elevated salinity area at depth is not known; however, as part of the annual data collection and review, FPL will continue to monitor and evaluate these areas.

4.3.4 Volume of the Hypersaline Plume The volume of the hypersaline plume is estimated by summing the volume of the CSEM grid cells (voxels) that have an estimated chloride level > 19,000 mg/L. This has been done for all 14 layers of the baseline 2018 and Year 3 2021 surveys. It is very important to note that the volume of each voxel represents aquifer matrix measured by the resistivity survey to be hypersaline and consists of rock matrix plus aquifer porewater. Most of the voxel volume is rock, while a smaller percentage (approximately 20-30%) is groundwater. As a result, the reduction in volume of the Biscayne Aquifer saturated with aquifer porewater with chloride content

> 19,000 mg/L should not be construed as the volume of hypersaline water removed from within the aquifer. Table 4.3-2 lists the measured volume of hypersaline aquifer matrix by layer for the 2018 and 2021 CSEM surveys and summarizes the volumetric changes by layer from 2018 to 2021, expressed both as a percent change in a layer and as a fraction of the total volume.

The reduction of hypersaline volume between the 2018 baseline survey and the Year 3 survey is 42%. The volumes are illustrated in Figures 4.3-5 and 4.3-6.

As previously described, the CSEM layer geometry increases with depth, and the volumes of the voxels increase with depth. As a result, care should be exercised when comparing the percent reduction of the volume of the aquifer saturated with hypersaline water for different layers. The lower CSEM layers have substantially greater volume per voxel than the shallowest layers. To illustrate a relative change in hypersaline aquifer volume between the 2018 and 2021 CSEM surveys, the 2018 and 2021 hypersaline voxel volumes are tabulated and plotted (Table 4.3-2 and Figure 4.3-5). Figure 4.3-6 clearly illustrates that most of the reduction in hypersaline plume volume between 2018 and 2021 occurred in layers 6-14, corresponding to the three zones of preferential flow in the Biscayne Aquifer. Columns 4 and 5 of Table 4.3-2 list the relative percentage changes in hypersaline voxels by layer between 2018 and 2021 (e.g., positive percentages indicate an increase from 2018). The changes in layer volumes are reported as percentages of the 2018 total volume. The sum of the individual layer changes is the 42%

change in the total plume volume from 2018 to 2021. The layer percentages represent the relative contribution of each layer to the total volume change.

The voxel volume is primarily rock with approximately 20-30% groundwater. As a result, the voxel volume of hypersalinity reduced should not be construed as the volume of hypersaline water removed from within the aquifer. The voxel volume represents the volume of the Biscayne Aquifer that is saturated with aquifer porewaters with chloride values > 19,000 mg/L.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-20 For the 2021 survey, the error in estimated CSEM chloride at any voxel is about 15%. However, the error is approximately normally distributed; therefore, when averaged over a large number of voxels, the error in the number of voxels with estimated CSEM chloride > 19,000 mg/L approaches the average error of about 11% for the 2021 CSEM chloride estimates. There is a sufficiently large number of hypersaline voxels in CSEM layers 4-14 for the error in the estimated volume change in any layer to be close to the mean error.

4.3.5 Summary of Comparison of 2018 and 2021 CSEM Survey Results There has been a statistically significant reduction in the volume of the hypersaline Biscayne Aquifer materials of 42% in 2021 as compared to 2018. Of this total reduction in hypersaline aquifer volume, approximately 8 percent of the reduction occurred during the 8-month period from the September 2020 Year 2 survey and the June 2021 Year 3 survey indicating robust plume remediation is continuing west and north of the CCS in all layers and at distances over 1 mile from the CCS. There is no CCS-sourced hypersaline water in layers 1 through 4 (i.e., the upper 15 feet of the aquifer) west of the L-31E levee. The volume of the hypersaline plume decreases between 2018 and 2021 in CSEM layers 6 through 14, with the largest volume decreases in layers 7 through 14. There are areas of statistically significant net retraction of the western extent of the CCS-sourced hypersaline plume in layers 5 through 14, particularly at the northern end of the CCS as indicated by the > 19,000 mg/L contour maps. The area of hypersalinity located in layers 1 and 2 in the northeastern corner of the compliance area in 2018 was not sourced from the CCS and, since 2018, has sunk into layers 3 and 4 by 2019 and into layers 4 and 5 by 2021. Layers 1, 2, and 3 do not contain porewaters > 19,000 mg/L in 2021.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-21 Table 4.2-1.

Thickness and Depth to Bottom for each Layer in the CSEM Model.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-22 Table 4.2-2.

June 2021 Water Quality Data from TPGW Wells.

From (ft)

To (ft)

From (m)

To (m)

TPGW-1S 32.00 34.00 9.76 10.37 12,600 20.78 33,274 TPGW-1M 52.10 54.10 15.88 16.49 24,900 42.55 63,139 TPGW-1D 85.30 89.30 26.01 27.23 27,400 47.98 70,124 TPGW-2S 27.97 31.97 8.53 9.75 15,600 29.88 46,165 TPGW-2M 53.88 55.88 16.43 17.04 28,300 50.70 73,590 TPGW-2D 88.79 90.79 27.07 27.68 30,500 52.16 75,429 TPGW-4S 23.20 25.20 7.07 7.68 2,490 4.72 8,502 TPGW-4M 38.10 43.10 11.62 13.14 14,200 25.89 40,546 TPGW-4D 61.60 65.60 18.78 20.00 14,900 26.65 41,612 TPGW-5S 28.60 32.60 8.72 9.94 154 0.47 951 TPGW-5M 49.30 54.30 15.03 16.55 10,300 18.94 30,562 TPGW-5D 67.00 72.00 20.43 21.95 13,000 23.19 36,714 TPGW-6S 25.09 27.09 7.65 8.26 328 0.85 1,691 TPGW-6M 51.61 55.61 15.73 16.95 8,500 14.36 23,717 TPGW-6D 84.70 88.70 25.82 27.04 8,870 15.23 25,049 TPGW-12S 25.19 27.19 7.68 8.29 19,100 32.50 49,750 TPGW-12M 59.21 63.21 18.05 19.27 21,400 35.66 54,029 TPGW-12D 93.24 97.24 28.43 29.65 26,600 45.09 66,420 TPGW-15S 24.32 29.32 7.41 8.94 13,000 24.30 38,437 TPGW-15M 44.39 49.39 13.53 15.06 27,500 48.55 71,005 TPGW-15D 79.31 84.31 24.18 25.70 29,000 49.55 72,294 TPGW-17S 32.11 37.11 9.79 11.31 21,800 36.71 55,433 TPGW-17M 49.95 54.95 15.23 16.75 27,400 45.68 67,221 TPGW-17D 86.81 91.81 26.47 27.99 27,400 47.69 69,750 TPGW-18S 35.25 40.25 10.75 12.27 3,130 5.55 9,882 TPGW-18M 63.25 68.25 19.28 20.81 20,500 39.04 58,502 TPGW-18D 84.27 91.27 25.69 27.83 21,000 39.78 59,482 TPGW-19S 27.37 31.37 8.34 9.56 309 0.77 1,540 TPGW-19M 48.39 52.39 14.75 15.97 20,100 33.98 51,751 TPGW-19D 84.35 89.35 25.72 27.24 23,400 40.85 60,918 TPGW-22S 29.00 32.00 8.84 9.75 13,900 26.76 41,752 TPGW-22M 54.00 57.00 16.46 17.37 20,800 36.43 54,992 TPGW-22D 69.00 72.00 21.03 21.94 20,600 37.33 56,191 Well ID Well Screen (from Top of Casing)

Cl (mg/L)

Specific Conductance

(µS/cm)

Salinity (PSU)

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-23 Table 4.2-3.

CSEM (AEM) Resistivity Associated with 19,000 mg/L Chloride Listed by Survey Year.

Survey Year 19,000 CSEM Resistivity (ohm-m) 2021 1.656 2020 1.719 2019 1.879 2018 1.968

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-24 Table 4.3-1.

Correspondence Between TPGW Screened Zones and the CSEM Model Layer.

From (m)

To (m)

From (ft)

To (ft)

Groundwater Model Layer TPGW-4S 7.1 7.7 23.2 25.2 2,490 6

4 3,921 TPGW-15S 7.4 8.9 24.3 29.3 13,000 6

5 5,290*

TPGW-6S 7.6 8.3 25.1 27.1 328 6

5 231 TPGW-12S 7.7 8.3 25.2 27.2 19,100 6

5 18,850*

TPGW-2S 8.5 9.7 28.0 32.0 15,600 7

4 18,158 TPGW-5S 8.7 9.9 28.6 32.6 154 7

5 197 TPGW-6S 7.6 8.3 25.1 27.1 328 7

5 231 TPGW-15S 7.4 8.9 24.3 29.3 13,000 7

5 5,290*

TPGW-19S 8.3 9.6 27.4 31.4 309 7

5 1,556 TPGW-22S 8.8 9.8 29.0 32.0 13,900 7

5 7,812 TPGW-1S 9.8 10.4 32.0 34.0 12,600 8

5 11,287 TPGW-17S 9.8 11.3 32.1 37.1 21,800 8

5 25,832 TPGW-18S 10.7 12.3 35.3 40.2 3,130 8

4 11,387 TPGW-4M 11.6 13.1 38.1 43.1 14,200 9

8 19,933 TPGW-18S 10.7 12.3 35.3 40.2 3,130 9

4 11,387 TPGW-15M 13.5 15.1 44.4 49.4 27,500 9

8 12,380*

TPGW-15M 13.5 15.1 44.4 49.4 27,500 10 8

12,380*

TPGW-1M 15.9 16.5 52.1 54.1 24,900 10 8

21,224 TPGW-2M 16.4 17.0 53.9 55.9 28,300 10 8

19,091 TPGW-5M 15.0 16.6 49.3 54.3 10,300 10 8

10,482 TPGW-6M 15.7 17.0 51.6 55.6 8,500 10 9

3,274 TPGW-17M 15.2 16.8 50.0 55.0 27,400 10 8

35,987 TPGW-19M 14.8 16.0 48.4 52.4 20,100 10 7

26,387 TPGW-22M 16.5 17.4 54.0 57.0 20,800 10 9

15,780 TPGW-12M 18.1 19.3 59.2 63.2 21,400 11 8

14,938*

TPGW-4D 18.8 20.0 61.6 65.6 14,900 11 11 7,378 TPGW-22M 16.5 17.4 54.0 57.0 20,800 11 9

15,780 TPGW-4D 18.8 20.0 61.6 65.6 14,900 12 11 7,378 TPGW-5D 20.4 22.0 67.0 72.0 13,000 12 10 6,245 TPGW-18M 19.3 20.8 63.3 68.3 20,500 12 7

18,308 TPGW-22D 21.0 21.9 69.0 72.0 20,600 12 10 10,089 TPGW-15D 24.2 25.7 79.3 84.3 29,000 13 11 34,301*

TPGW-1D 26.0 27.2 85.3 89.3 27,400 13 11 17,937 TPGW-6D 25.8 27.0 84.7 88.7 8,870 13 11 7,252 TPGW-18D 25.7 27.8 84.3 91.3 21,000 13 10 19,420 TPGW-19D 25.7 27.2 84.3 89.4 23,400 13 11 8,635 TPGW-1D 26.0 27.2 85.3 89.3 27,400 14 11 17,937 TPGW-2D 27.1 27.7 88.8 90.8 30,500 14 10 32,066 TPGW-6D 25.8 27.0 84.7 88.7 8,870 14 11 7,252 TPGW-17D 26.5 28.0 86.8 91.8 27,400 14 11 38,002 TPGW-18D 25.7 27.8 84.3 91.3 21,000 14 10 19,420 TPGW-19D 25.7 27.2 84.3 89.4 23,400 14 11 8,635 TPGW-12D 28.4 29.6 93.2 97.2 26,600 14 11 17,310*

Note: CSEM-derived Cl produced from average resistivity within 175 m radius of borehole and converted to Cl using regression equation.

• TPGW-12 and TPGW-15 CSEM estimated chloride data are calculated from nearest Cl voxel due to locations within EM noise.

Well ID Depth Below Land Surface 2021 Chloride (mg/L)

CSEM Model Layer CSEM Estimated Layer Cl (mg/L)

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-25 Table 4.3-2.

CSEM-Derived Chloride Volume Estimates of Hypersaline Aquifer Material for 2018 and 2021 by Layer (m3).

Layer 2018 Volume

> 19,000 mg/L (m3) 2021 Volume

> 19,000 mg/L (m3)

% Volumetric Change by Layer (2018 to 2021)

% Volumetric Change by Layer (2018 to 2021)

Relative to Total Volume 1

1,920,000 0

-100

-0.4 2

2,026,000 0

-100

-0.4 3

1,587,000 0

-100

-0.3 4

858,000 2,106,000 145 0.3 5

2,595,500 3,799,000 46 0.3 6

12,672,000 5,024,000

-60

-2 7

22,896,000 5,850,000

-74

-4 8

20,480,000 4,920,000

-76

-3 9

33,770,000 20,306,000

-40

-3 10 68,281,500 45,668,000

-33

-5 11 92,840,000 51,837,500

-44

-9 12 89,822,500 54,442,500

-39

-8 13 63,181,000 40,568,500

-36

-5 14 43,050,000 28,665,000

-33

-3 Totals:

455,979,500 263,186,500

-42

-42

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-26 Figure 4.2-1. 2021 CSEM Survey Area, Flight Lines, and Monitoring Well Locations.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-27 Figure 4.2-2. Locations of the Decoupled and Removed Data (Red Lines) Along the CSEM Flight Lines and the Data Used in the Inversion (Blue Lines).

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-28 Figure 4.2-3. Relationship Between Borehole Induction Log Resistivity and CSEM Resistivity.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-29 Figure 4.2-4. Comparison of 2021 Formation Water Resistivity versus CSEM Resistivity.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-30 Figure 4.2-5. Comparison of 2021 Formation Water Resistivity versus Lab Chloride Concentration.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-31 Figure 4.2-6. AEM Resistivity v. Lab Water Resistivity Plot Including TPGW-22 Data; Locations of the TPGW-22 Data Identified Relative to the Trend Line.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-32 Figure 4.2-7. Compliance Area Hypersaline (>19,000 mg/L) Volume Trends.

Figure 4.2-8. Comparison of 2021 Lab Chloride and CSEM-Derived Chloride Concentrations.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-33 Figure 4.2-9. Monitor Well Screened Zone vs. CSEM Layer.

Figure 4.2-10. Normal Probability Plot of 2021 Error Distribution.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-34 Figure 4.2-11. Chloride Depth Slice (Layer 12) for 2021 CSEM Survey.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-35 Figure 4.3-1. Layer 7, 19,000 mg/L Chloride Concentration Contours for 2018 and 2021.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-36 Figure 4.3-2. Layer 9, 19,000 mg/L Chloride Concentration Contours for 2018 and 2021.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-37 Figure 4.3-3. Layer 10, 19,000 mg/L Chloride Concentration Contours for 2018 and 2021.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-38 Figure 4.3-4. Layer 13, 19,000 mg/L Chloride Concentration Contours for 2018 and 2021.

FPL Turkey Point RAASR Year 3 November 2021

4. CSEM Survey Summary 4-39 Figure 4.3-5. 2018 and 2021 Hypersaline Volume (> 19,000 mg/L) by Layer.

Figure 4.3-6. Normalized Percent Change: 2018 to 2021 (> 19,000 mg/L) from the Total 2018 Volume.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-1 5 GROUNDWATER MODEL 5.1 MODEL OVERVIEW AND EVOLUTION 5.1.1 Objectives The variable density flow and salt transport model developed for the design of the FPL Turkey Point Biscayne Aquifer RWS has been updated and recalibrated using data from the third year of operation of the RWS. This update represents the sixth version of the model. The objectives of the update and recalibration are to reduce the uncertainty of the model and to improve alignment of model responses with new and existing monitoring data in order to assess progress of the ongoing hypersaline groundwater plume remediation effort.

These objectives are accomplished by the following actions:

1. Incorporation of the salinity, water level, and mass extraction data (RWS and UIC test production wells) collected during the third year of the recovery system operation

2. Derivation of monthly salinity change targets at groundwater monitoring locations to further emphasize the accurate simulation of the movement of the saline and hypersaline interfaces during the third year of RWS operation

3. Incorporation of changes from baseline to Year 3 CSEM salinity targets into the model calibration process

4. Incorporation of vertical hydraulic conductivity profiles derived from inspection of drill cores and geophysical logs as templates for specification of depth-discrete hydraulic conductivity values

5. Performance of a sensitivity analysis with the model used in the prior RAASR (version 5) to investigate the causes of that models inability to align the saline-hypersaline interface (HSI) position along the base of the aquifer and retract the HSI in a manner consistent with monitor well and CSEM data (details of this sensitivity evaluation is described in Chapter 4 of Appendix H)

6. Incorporation of the results of the sensitivity analysis into the version 6 model calibration process.

Data from the 2018 baseline, Year 1, Year 2, and Year 3 remediation operations have been incorporated into and assessed in the Turkey Point variable density dependent solute transport model in order to provide a better understanding of the hydrogeology of the study area, improve the models ability as a predictive tool, and contribute to the assessment of the progress of remediation and predictions of future performance.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-2 The revised model was then used to predict RWS impacts on the degree of CCS sourced hypersaline groundwater plume retraction at Year 5 and 10 of remediation.

5.1.2 Model Versions The current groundwater flow and salt transport model documented in Appendix H is the sixth version (V6) of a 3D regional model developed by FPL to evaluate various projects associated with the Turkey Point CCS. The model has undergone an evolutionary process as the objectives of the modeling changed, as progressively more data are added, and as the knowledge base expands. The evolution of the model to date is summarized below.

FPL originally developed a 3D SEAWAT (Langevin et al. 2008) model wherein density varied as a function of both salinity and temperature (Tetra Tech 2016a). This model is referred to as the V1 model. The purpose of the V1 model was to evaluate alternatives for compliance with the MDC CA and FDEP CO that required stopping the westward migration of hypersaline water and retracting hypersaline water north and west of the CCS to the western extent of the FPL property.

This model simulated the period from pre-development (early 1940s) through 2015 and was calibrated by manual methods to measured water levels and salinity. The model was used to evaluate a number of potential groundwater remediation projects that resulted in the selection of Alternative 3D, which involved implementing a groundwater recovery well system (RWS) consisting of 10 wells screened to the base of the Biscayne Aquifer capable of pumping 15 mgd of hypersaline groundwater that would be disposed in an existing UIC DIW. Subsequent to the submittal of the model and the preferred alternative to the agencies on May 16, 2016, FPL was requested to make several additions/improvements that are documented in a Tetra Tech report (2016a). These changes included responding to regulatory comments; primarily regarding boundary conditions (Tetra Tech 2016b); the use of an automated calibration tool (i.e., Parameter Estimation [PEST]) (Tetra Tech 2016c); and introducing a spatially variable hydraulic conductivity field into certain model layers (Tetra Tech 2017). Although these changes are considered to improve the model, they did not alter the choice or design of Alternative 3D.

Based on these model changes and the associated results related to the retraction of the hypersaline plume, the models application in the assessment of Alternative 3D were conditionally approved by MDC on September 29, 2016.

In addition to using the model to aid in the evaluation and selection of a groundwater remediation system, FPL was directed by FDEP to use the variable density 3D groundwater model developed under the MDC CA to allocate relative contributions of the CCS and other entities or factors on the movement of the saltwater interface. In order to conduct this evaluation, several modifications, including many of those required by MDC, were necessary.

FPL subsequently updated and recalibrated the V1 model to create the V2 model for apportionment of contributing factors on migration of saltwater in the region. Primary changes that were incorporated into the V2 model were: (1) inclusion of data from MDC wellfields; (2) land use time series, including rock mines and quarries; (3) separate simulation of recharge and evapotranspiration processes rather than a single net recharge term; (4) inclusion of detailed precipitation data; (5) consideration and incorporation of canal and surface water methodology from the Hughes and White (2014) model; and (6) incorporation of sea level rise. The model was recalibrated using PEST, and this resulted in a 34% reduction in the overall error statistic

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-3 that was used as the objective function in calibrating the model. The V2 model was used to simulate alternate historical scenarios and assess their individual and cumulative impacts on the saltwater interface as required under the FDEP CO. The results of the V2 modeling were presented to FDEP on June 19, 2019.

In compliance with paragraph 17.b.ii., and 17.d.11 of the MDC CA, as amended on August 20, 2019, FPL updated the variable density flow and salt transport model informed with data collected during the first year of operation of the RWS. This model, referred to as the V3 model, was updated using the V2 model as a starting point. Updates included (1) incorporation of compatible elements of the surface water routing package developed by USGS and implemented by Hughes and White (2014); (2) incorporation of geologic information obtained during the installation of the groundwater recovery system and newly installed TPGW monitoring wells; (3) incorporation of baseline CSEM salinity targets into the model calibration process; (4) inclusion of hydrogeologic data collected by other entities west of the RWS, and (5) calibration of the model using objective automated parameter estimation techniques. The model was calibrated to conditions leading up to the start of the RWS (May 2018) and verified to the first year of RWS operation (through May 2019). The verification consisted of simulating the RWS for one year of operation and comparing the simulation results to observed results (i.e., water levels, salinity at monitoring wells and CSEM targets, and mass extracted). The model was then used to simulate the effect of operating the RWS for 10 years. Compliance with the CA were supported by model predictions indicated by figures and tables of hypersaline plume retraction, mass removal, and declines in salinity versus time in monitoring wells during the 10-year simulation. The model documentation was included as Appendix H of the RAASR (FPL 2019c).

The next version of the model was called V4 and was originally intended to use data from and evaluate the second year of RWS operation for the second RAASR. However, COVID-19 delayed the scheduled acquisition of Year 2 CSEM data from May to September 2020. The V4 model was updated without the Year 2 CSEM data and used to support FPLs application to FDEP to increase the CCS freshening allocation referred to as the Supplemental Salinity Management Plan (SSMP). The updates to the V4 model included (1) calibration to data (e.g.,

salinity, water levels, and mass extracted) collected from the first year of operation of the RWS; (2) incorporation of geologic information obtained during and after the installation of the RWS and newly installed monitoring wells; (3) incorporation of CCS sediment sampling and testing information into the assignment of the CCS boundary condition; (4) incorporation of changes from baseline to Year 1 CSEM salinity targets into the model calibration process; and (5) calibration of the model using two separate search algorithms contained within objective automated parameter estimation techniques. Note that the original RAASR included a verification of model results to the Year 1 CSEM data but, due to time constraints, did not include calibration to Year 1 CSEM data. Calibration to Year 1 CSEM data was accomplished with the V4 model. The V4 model was used to assess impacts resulting from discharging additional water to the CCS from the Floridan Aquifer and to support the SSMP.

The development of the V5 model used the V4 model as its starting point and built upon the knowledge gained from calibration and execution of the V4 model. The V5 model included the Year 2 CSEM data that was collected in September 2020, later than originally planned, as

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-4 described above. A key change to the model calibration procedure was the manner in which the CSEM data were used as targets. Rather than using a random subset of discrete CSEM measurements as was done in earlier versions of the model, the target assignment involved averaging CSEM measurements over approximate 1000-ft by 1000-ft areas in each model layer.

These averaged CSEM targets were then compared to modeled salinities at the centroid of each CSEM target area. This revised procedure used fewer CSEM targets than previous versions, but they covered the same area and depth as prior versions CSEM targets. In addition, the averaging procedure created a smoother CSEM salinity distribution that contained less apparent outliers than the prior versions. Model calibration involved running two search algorithms, as was done in prior versions of the model. Following calibration, the model was used to make projections regarding the operation of the RWS over the next eight years. Compliance with the CA was supported by figures and tables of hypersaline plume retraction, mass removal, and declines in salinity versus time in monitoring wells during the 8-year simulation. The model documentation was included as Appendix H of the Second Remedial Action Annual Status Report (FPL 2021a).

Features of each model version, including the current V6 model, are summarized in Table 5.1-1.

5.1.3 Sensitivity Analysis with the Version 5 Model One of the recommendations in the second RAASR report (FPL 2021a) was to investigate the causes of the V5 models inability to align the saline-hypersaline interface position along the base of the aquifer and retract the HSI in a manner consistent with monitor well and CSEM data.

To address this recommendation, sensitivity analyses were performed using the V5 model prior to development of the V6 model (Appendix H, Chapter 5). The sensitivity analysis identified several factors that were important for more accurate calibration and prediction: (1) use of hydraulic conductivity distributions that were more consistent with those developed through visual core inspection by a geologist; (2) use of accurate vertical hydraulic conductivities; (3) obtaining an accurate end-of-calibration location of the hypersaline interface; (4) use of spatially variable porosity fields instead of the previously used layer-wide porosity specification; and (5) accurate vertical resolution of the plume through the use of a different numerical solution technique. The sensitivity analysis also suggested that the climate pattern used in the model predictions was important. These findings were considered in the development, calibration, and predictive simulations with the V6 model.

5.1.4 Description of Version 6 Model Detailed descriptions of the V6 model assembly, calibration, and predictive simulations are included in Appendix H. The V6 model uses the same basic framework as the prior V1 though V5 models. It simulates a 276-square-mile area that is subdivided from west to east into 274 columns and from north to south into 295 rows. The width of the rows and columns vary between 200 ft and 500 ft, with smaller grid cell dimensions located near the CCS. The model domain overlain by the model grid is shown on Figure 5.1-1.

Vertically, the modeled Biscayne Aquifer (which varies in thickness from approximately 50 ft in the west to approximately 100 ft in the east) is divided into 11 layers, for a total of approximately

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-5 890,000 grid cells. The uppermost model layers (layers 1 through 4) represent the Miami Oolite.

The thicker Fort Thompson Formation was divided into seven layers (layers 5 through 11).

Many of the well borings that were analyzed to define geologic contact elevations contained zones with large, connected voids. These features were represented in the model as thin layers with high hydraulic conductivity. Two zones of high hydraulic conductivity material (i.e., high-flow zones) were identified from the various well borings analyzed to define the two high-flow zones represented in the regional model. The upper high-flow zone (i.e., layer 4) occurs at the base of the Miami Oolite, and the lower high-flow zone (i.e., layer 8) is located in the approximate middle of the Fort Thompson Formation. A third, lower high-flow zone, located along the contact between the Fort Thompson and Tamiami formations, is monitored by the regional groundwater network; but it is not expressly represented as a unique hydraulic conductivity zone in the model. Hydraulic conductivities within the high-flow zones were calibrated using targets of hydraulic head, salinity, and a range in hydraulic conductivity estimated by a geologist through inspection of geologic logs. This methodology allowed the high-conductivity zones to be discontinuous or to be present in multiple layers besides layers 4 and 8. Figure 5.1-2 provides a cross-sectional view of the 11 layers of the groundwater flow model and how they correspond to the hydrogeologic formations. The location of this cross-section, along row 116 of the model, is shown on Figure 5.1-1.

The USGS groundwater flow and solute transport modeling tool SEAWAT Version 4 (Langevin et al. 2008) was used in this analysis as well as the prior (V1 through V5) modeling analyses.

This SEAWAT version includes: (1) solute transport simulations through the integrated MT3DMS (IMT) Process (Zheng and Wang 1998); and (2) variable-density flow (VDF) simulation through the VDF process. SEAWATs VDF package was used to simulate the density effects of both temperature and salinity. SEAWAT inputs and outputs are specified in terms of point-water heads (Langevin et al. 2008), which represent the hydraulic head at a given location based on salinity and temperature. SEAWAT solves the groundwater flow and transport equations after converting point-water heads to reference heads or equivalent freshwater head at the reference temperature.

The boundary conditions applied to the V6 model are also very similar to those applied in the prior models. Namely, specified head-boundary conditions are used to simulate the effects of temporal changes in Biscayne Bay and the various canals (including the CCS) on groundwater flow and transport. General-head boundaries are used to simulate the exchange of groundwater across the models lateral boundaries on all sides. NEXRAD-based rainfall rates and historical patterns (spatial and temporal) in land use are used to estimate the amount of groundwater recharge throughout the model domain. Reference evapotranspiration (ETo) data and land use/land cover data are used similarly to estimate groundwater evapotranspiration (ET) rates as a function of groundwater head. Consumptive use of groundwater for agricultural purposes and some industrial uses (e.g., the Blue Water Industries and Card Sound quarries) are simulated as specified withdrawals; and they were estimated based on land cover, estimated local rainfall/recharge, and ET rates. Municipal and other industrial groundwater uses are also simulated as specified withdrawals and are based on data as much as possible. The temperatures and salinities assigned to water entering from the various boundaries simulated are also based on actual data, as much as possible.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-6 The V6 model incorporates canal hydraulics and groundwater interactions based on Hughes and Whites (2014) modeling for MDC that were added in the V2 model. Details of these additions are included in Appendix H. Two key changes to the calibration approach include the characterization of porosity and hydraulic conductivity. Whereas a single porosity was calibrated within each model layer of the V5 model, a pilot point methodology that allowed variable porosity with a layer was utilized in the V6 calibration, based on insights gained from the sensitivity analyses discussed above. In addition, horizontal hydraulic conductivity estimates provided by JLA Associates, based on geologic cores and geophysical logs collected from TPGW monitoring wells and RWS production wells, were included in the V6 model calibration process. Unlike in V5 where the TPGW and RWS interpreted hydraulic conductivities were calibration targets, hydraulic conductivities at these locations were adjusted during the course of calibration. The percent differences in hydraulic conductivity across model layers 5-7 and 9-11 were maintained at each location while hydraulic conductivities in layers 1, 2, 4, 5, 8, and 9 were calibrated for each pilot point.

In order to perform reliable predictions and satisfactorily meet the objectives of this modeling effort, the SEAWAT model required calibration. Model calibration is the process of adjusting parameters and boundary conditions within reasonable ranges to match historical observations, such as hydraulic heads and salinities, reasonably well. The ability to replicate past conditions in the calibration period provides confidence that the model is capable of simulating future conditions in the model applications. The calibration model is subdivided into four time frames, each of which simulates the development and movement of the saltwater wedge under different hydrologic and anthropogenic stresses.

5.2 MODEL MODIFICATIONS/CALIBRATION 5.2.1 Model Calibration Process The process for developing a model capable of providing accurate projections of RWS operation involves calibration of the model to prior measured data that are similar to those of the projections it is to make.

Model calibration is the process of adjusting parameters and boundary conditions within reasonable ranges to match historical observations reasonably well. This model calibration involved matching: (1) water level and salinity observations; and (2) salinity estimates based on The model was calibrated to 81 years of data, including Years 1 through 3 of operation of the RWS.

The resulting calibration indicates the model is a reasonable tool to be used in conjunction with monitoring data and CSEM results to assess progress in meeting the groundwater remediation objectives of the MDC CA and FDEP CO. However, improvements in the model's alignment with CSEM and monitoring well data at the edge of the hypersaline plume in 2018, prior to remediation and during subsequent remediation years, is needed to improve reliability of long-range remediation forecasts.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-7 spring 2018, spring 2019, fall 2020, and summer (June) 2021 CSEM datasets1. The calibration period includes the V5 model calibration period (i.e., pre-development through September 2020) plus an additional 9 months. This nine-month period includes the June 2021 CSEM dataset. The calibration model is subdivided into four time frames, each of which simulates the development and movement of the saltwater wedge under different hydrologic and anthropogenic stresses.

These periods are defined as follows:

7. Pre-development steady-state flow model (prior to 1940)

8. Steady-state flow and transient transport calibration model (1940-1968)

9. Seasonal transient flow and transport calibration model (1968-2010)

10. Monthly transient flow and transport calibration model (2010-2021)

Calibration was performed primarily using automated PEST.

The final period includes the first, second, and third year of RWS operation. Operational pumping rates for each RWS well are input on a monthly basis. All available precipitation and boundary condition (i.e., canal stage) data are also used. Model results for water levels, salinities at monitoring wells, CSEM salinity distribution, and mass extracted by the RWS are compared to measured values.

5.2.2 Model Calibration Results The calibration results for water levels and relative salinities are shown in Table 5.2-1. Seasonal and monthly transient model water levels and relative salinities are shown separately in Appendix H. In general, the monthly data set is considered more reliable than the seasonal data set because it uses the multi-depth and short-screened TPGW wells. The monthly data set also has considerably more data despite its shorter duration (2010-2021). A robust assessment of model calibration quality and statistics is provided in Appendix H.

The model responds similarly to the hydrologic system during the first three years of RWS operation in that salinity changes are only observed in wells relatively close to the RWS. In addition, responses are noted only in the shallow and intermediate wells. Model versus measured change in salinity for the shallow and intermediate zones of wells TPGW-1, TPGW-2, TPGW-4, TPGW-15, TPGW-17, and TPGW-19 are shown on Figure 5.2-1. Comparison of model to measured salinity change is very good for these wells, though inexact. There is some over-simulation of salinity decline in two intermediate wells (TPGW-2M and TPGW-15M) though the match to salinity decline in the shallow zones of these wells is quite good.

The model shows a similar response as the CSEM data to the movement of the hypersaline (i.e.,

chloride >19,000 mg/L) interface. Table 5.2-2 compares the number of CSEM targets (i.e.,

1000-ft by 1000-ft composite cells in the model) and model targets that change from hypersaline to saline between the 2018 to 2019, 2019 to 2020, 2020 to 2021, and entire 2018 to 2021 periods.

There is fairly good agreement with the number of cells that change over the entire 2018 to 2021 1 Whereas the simulated salinities coincident with the 2016 CSEM survey were reviewed during the course of the calibration, these data were not used to guide the adjustment of model parameters.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-8 remediation period, although the model shows fewer cells (707) turning saline than does the CSEM data (815). In individual years, the match between the model and CSEM results is generally favorable. Despite any differences, both the model and CSEM data indicate retraction of the plume during the first 3 years of operation, although to different degrees and at different locations.

The model has an excellent match to measured total salt mass extracted by the RWS (Figure 5.2-2). The match for individual wells is also good (Figure 5.2-3), with the largest discrepancies noted in wells RWS-4, RWS-7, and RWS-8. There is also some over-simulation of mass extracted in RWS-1 in the final 6 months of operation, likely due to over-simulation of salinity in this area.

Based on the success in meeting the calibration goals, the sequence of calibration models was deemed satisfactory to employ in the execution of the forecast model.

5.3 REMEDIATION YEARS 5 AND 10 FORECAST 5.3.1 Description of Remediation Simulations The projected results of operation of the RWS during Years 4 through 10 were simulated in a similar fashion as the operations of the RWS during the Years 1 through 3. The initial conditions for the simulation were the ending conditions (i.e., salinity, temperature, and water level) of the final period of the calibration simulation (June 2021). Climate conditions (e.g., precipitation, evaporation, canal stages) for the 2018-2021 period used in the calibration were repeated during the predictive period. This sequence uses precipitation that is below normal for the first year, above normal for the second year, and close to average for the third year, based on site-specific meteorological data. For comparison, June to May precipitation for the first year is documented at Miami International Airport to be 52.7 inches, 81.7 inches for the second year, and 69.6 inches for the third year (NOWData 2021). Average annual precipitation during the period 2010 through 2020 was 70.6 inches, which was close to the average of 68.0 inches used in the 3-year sequence. This methodology uses data and conditions that represent a recent time frame of varied hydrologic conditions and is a likely reasonable approximation of future conditions. The RWS was simulated to operate according to the design: 1.5 mgd from each RWS well, for a total of 15 mgd withdrawal. The CCS was set at a salinity of 34 PSU for the duration of the predictive period. In addition, the UIC test production wells were set to withdraw a total of 3 mgd from beneath the CCS for the duration of the remediation period.

Predictive model runs indicate the RWS will fully retract the HSI in the upper two-thirds of the Biscayne Aquifer (layers 1-8) to the FPL property within 10 years of RWS operation. The simulations indicate the HSI retracts to the FPL property in the northern half of the CCS in layer 9, expands slightly in layers 10 and 11 during the first 5 years of RWS operation, but stabilizes between Years 5 and 10. However, the results in the lower model layers differ from CSEM-measured changes suggesting additional evaluation of the model forecast and CSEM trends are needed.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-9 5.3.2 Remediation Forecast Model forecasts of the position of the hypersaline/saline water interface in Years 5 and 10 of remediation were determined using the V6 model, and they are similar to prior forecasts.

Retraction to the L-31E canal is generally achieved in the upper 8 model layers (approximately 67% of the Biscayne aquifer thickness) but is not fully achieved in layer 9. The model forecasts minor westward expansion of the plume in layers 10 and 11 by Year 5 but stabilizes from Year 5 through Year 10 (Figures 5.3-1 a, b, c). This forecast is contrary to Year 3 CSEM survey results that show net retractions of the plume along the base of the aquifer. Layer 8 shows a small portion of hypersaline groundwater persists west of the L-31E canal adjacent to the southern end of the CCS, which has not been the case in prior model versions. In addition, the model identifies a small area of hypersalinity north of the plant in Year 10, which is not identified to be contiguous with the CCS in CSEM surveys.

At this point, it is not clear whether the incomplete retraction is a result of a physical phenomenon or inaccuracies in the model. The physical phenomenon of the lower high-flow zone overlying the RWS wells may create a preferred vertical flow path for RWS pumping that renders modeled horizontal retraction in these layers difficult. This conceptual model is supported by the layer 10 capture zones being smaller than the capture zones in overlying but unpumped layers (Figure 5.3-2). This pattern suggests primarily vertical flow in layer 10 near the RWS wells. In addition, there appears to be contribution from non-CCS, coastal, evaporative-formed hypersaline groundwater that is recharging the hypersaline plume north and south of the CCS. The V6 and earlier versions of the model have shown this process to be simulated with the surface-formed hypersalinity migrating vertically in the aquifer and recharging the lower model layers. Retraction in the northern and southern areas surrounding the CCS could be hampered by continued addition of hypersaline water from the plume to the south.

Additional evaluation of the models representation and extent of this process should be evaluated to determine the degree to which this source of hypersalinity could impact the CCS remediation objectives.

Particle tracking, conducted using the V6 model (Figure 5.3-2), confirmed the hypersaline water from the CCS is intercepted, captured, and contained beneath the CCS, and no longer migrates into the compliance zone. This analysis was conducted for all model layers from initiation of the RWS operations through 10 years of remediation. It indicates the hydraulic constraint to the migration of hypersaline groundwater from beneath the CCS was halted shortly after RWS operations began and are maintained through Year 10. Figure 5.3-2 shows the predicted capture zones of layers 4, 8, and 10; particle tracking for additional layers is shown in Appendix H.

These figures are generated by initializing particles in cells surrounding the RWS wells (in the areas shaded grey in Figure 5.3-2) for all layers and highlighting the starting locations (in orange) of each particle that ends at the RWS. It is clear that there is an extensive capture zone to the east and west in shallow high-flow zones despite these layers not being pumped. In contrast, capture zones with comparatively smaller radii surround each of the RWS wells in layers 10 and 11, indicate these wells obtain a smaller portion of their water laterally within those layers and more vertically from shallower overlying high flow layers. Moreover, Figure 5.3-2 illustrates that the RWS prevents water from the CCS from flowing through, and westward

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-10 past, the RWS. This includes the deep layer where further analysis revealed that the apparent gaps between RWS-6, RWS-7, and RWS-8 are stagnation areas where the particles do not follow a consistent pathway and are neither captured nor allowed to discharge westward past the RWS.

Stagnation zones develop in areas between adjacent extraction wells with overlapping drawdowns in pump and treat operations that result in low hydraulic gradients. Cohen et al.

(1994), provide examples of stagnation zones associated with different pumping schemes including the distribution of multiple stagnation zones at a complex field site.

Another potential inaccuracy in the model is that the simulated hydraulic conductivities in layers 9-11 south of the CCS and between the RWS and Tallahassee Road are significantly lower than estimated by inspection of geologic cores. The simulated hydraulic conductivities in layers 9-11 in wells TPGW-4 and TPGW-5 (located west of the CCS near Tallahassee Road) are 310 and 453 feet per day (ft/d), respectively, while estimated hydraulic conductivities for these wells are 2,460 and 4,890 ft/d. The low hydraulic conductivities estimated in the model in this area act to impede lateral movement of salinity in these layers. Such disparities are evident at TPGW-1, TPGW-2, and TPGW-3. However, comparison of estimated and calibrated hydraulic conductivities at nearby RWS wells indicates a reasonably accurate match. Higher hydraulic conductivities could also result in the plume moving farther west than documented to occur at the beginning of the remediation, a disparity that already exists in this and prior model versions.

Both the physical explanations and potential inaccuracies in the model that hamper saline-hypersaline interface retraction in the lower layers should be explored in the V7 model assessment.

5.3.3 Sensitivity Simulation A sensitivity simulation was made to test the effect of the accuracy of the location of the hypersaline plume at the start of the RWS. It has been recognized for some time that the model calibration tends to result in an initial plume that extends further west than supported by the CSEM data. Though conservative (i.e., a condition where plume retraction is more difficult), it is not clear if and how much additional plume retraction would result from a more accurate initial plume location. This condition was evaluated by replacing modeled salinities within the compliance zone with data from the 2021 CSEM study as initial conditions for the simulation of RWS operation.

The results of this simulation are similar to the original simulation in the upper 7 layers. The plume for this simulation is less extended to the west in the lower layers than in the original model, with a salinity below 45 PSU in layer 10 that is within the compliance zone. However, A sensitivity simulation evaluating an initial location of the hypersaline plume that is closer to the RWS (less extended to the west) than the predictive simulation produced greater retraction in layers 8 and 9 and salinity reduction in layers 10 and 11. The sensitivity simulation suggests the importance of accurate quantification of the initial plume and configuring the model to be able to reproduce the CSEM plume orientations prior to and during remediation.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-11 consistent with the original model, there is only a small amount of retraction in layers 10 and 11 in years 5 through 10. This simulation highlights the importance of having an accurate initial hypersaline interface location for the predictive simulations. However, plume retraction in the lower layers is still incomplete, despite having an initial condition that is now aligned with the CSEM data. The sensitivity simulation also highlights the importance of having the model construction and hydraulic properties be capable of reproducing the CSEM initial plume more accurately.

5.3.4 Model Recommendations Six recommendations for future evaluations of system performance are offered:

1. The continuity and hydraulic conductivity of the lower high-flow zone should be investigated and incorporated accurately into the model.

2. The hydraulic properties vertically adjacent to the lower high-flow zones should be investigated and incorporated accurately into the model.

These two recommendations are important because it appears from the capture zones and the models inability to retract the hypersaline interface within the pumped layers that the high-flow zone located in the middle model layers serves as a preferential pathway that minimizes the models ability to draw water from the lower layers resulting in a lack of retraction of the hypersaline plume in these layers.

3. Detailed evaluation and incorporation of CCS sediment and sampling information should be used to guide how the modeled connection of the CCS to groundwater needs to be revised and represented.

4. FPLs CCS water and salt-balance model should continue to be used to inform the model with respect to amounts and timing of canal and groundwater seepage to and from the CCS.

The third and fourth recommendations seek to obtain through calibration a more accurate present-day location of the hypersaline interface. The modeled plume appears to be more extensive than suggested by the CSEM data. Although conservative, the larger plume represented in the model is further outside the capture radius of the RWS and results in more hypersaline volume and mass than may actually be present.

5. CCS salinities and climate conditions should continue to be monitored and the model updated and recalibrated with more data reflective of longer RWS operations. The longer period of RWS operation and consequent changes to salinities over a progressively larger area will help inform the model and increase its accuracy in simulating the effect of the RWS and forecasting longer-term performance.

6. Evaluations should be conducted to verify the degree to which model-generated, non-CCS hypersaline groundwater impacts remediation objectives.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-12 Table 5.1-1. Summary of Groundwater Model Versions.

Version 1 Version 2 Version 3 Version 4 Version 5 Version 6 Date June 2016 June 2018 October 2019 September 2020 April 2021 September 2021 Purpose Design RWS FDEP attribution analysis Year 1 verification of RWS Year 1 Calibration to RWS Year 2 Calibration to RWS Year 3 Calibration to RWS Calibration Method Manual Automated (PEST)

Automated (PEST)

Automated (PEST)

Automated (PEST)

Automated (PEST)

Hydraulic Conductivity Representatio n

Uniform within layers Heterogeneous in layers 4,8,9,10,11 Heterogeneous in layers 4,8,9,10,11 Heterogeneous in layers 4,8,9,10,11 Heterogeneous in layers 4,8,9,10,11 Heterogeneous in layers 4,8,9,10,11.

Contrast between 5,6, and 7; 9,10, and 11 based on JLA Recharge Formulation Net Precipitation and Evapotranspiration Precipitation and Evapotranspiration Precipitation and Evapotranspiration Precipitation and Evapotranspiration Precipitation and Evapotranspiration CSEM Data Used?

No Yes Yes Yes Yes Yes Predictions 10-year forward 40-year backward 10-year forward 9-year forward 8-year forward 7-year forward Primary Change /

Focus Assess alternatives for compliance; selected alternative 3D Differentiate between Recharge and ET, detailed surface water representation, CSEM targets evaluate causal factors of regional saltwater intrusion 2nd round of CSEM and recent TPGW &

RWS wells as targets; verify with stress (RWS).

Recent TPGW &

RWS wells as targets; incorporation of geologic information at TPGW and RWS locations 3rd round of CSEM and recent TPGW &

RWS wells as targets; revision to CSEM targets to eliminate localized significant changes in salinity 4th round of CSEM and recent TPGW &

RWS wells as targets; porosity as a spatially variable parameter, sensitivity analysis to guide calibration

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-13 Table 5.2-1. Calibration Statistic Summary for the Version 6 Model.

Model Target Type Units ME MAE RMSE MAE ÷ Range Seasonal (1968-2010)

Hydraulic Head ft

-0.129 0.432 0.567 6.3%

Relative Salinity R.S.

-0.113 0.163 0.243 9.7%

Monthly (2010-2021)

Hydraulic Head ft

-0.210 0.302 0.398 4.9%

Relative Salinity R.S.

0.128 0.182 0.232 9.5%

CSEM (2018, through 2021) 2018 CSEM Survey R.S.

0.082 0.226 0.312 11.0%

2019 CSEM Survey R.S.

0.087 0.204 0.276 9.9%

2020 CSEM Survey R.S.

0.101 0.207 0.295 10.0%

2021 CSEM Survey R.S.

0.143 0.211 0.300 10.3%

Note: One Relative Salinity (R.S.) Unit = 35 PSU = 19,400 mg/L Cl

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-14 Table 5.2-2. Comparison of CSEM and Model Representations of Remediation Progress.

Category 2018-2019 2019-2020 2020-2021 2018-2021 CSEM Model CSEM Model CSEM Model CSEM Model Number of hypersaline targets that change to saline 387 243 285 315 276 167 815 707 Number of initially (2018) hypersaline targets 1728 2076 1728 2076 1728 2076 1728 2076 Percent of initial (2018) hypersaline targets that change to saline 22%

12%

16%

15%

16%

8%

47%

34%

Note:

The CSEM calculated reduction of 42% in section 4 is based on the number of 100 m X 100 m CSEM voxels that changed from hypersaline to hypersaline. The model target cell sizes are 1,000 x 1,000 feet which, when overlain over the CSEM defined hypersaline layers, result in a small difference between the CSEM calculated percent reduction and the model cell-based accounting results.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-15 Figure 5.1-1. Model Study Area Overlain by the Active Model Grid; Red Dashed Line Represents the Location of the Model Cross Section Shown in Figure 5.1-

2.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-16 Figure 5.1-2. Model Cross Section Showing Model Layering and Hydrogeologic Formations (Location of Cross Section Shown in Figure 5.1-1).

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-17 5.2-1. Comparison of Model and Observed Changes in Relative Salinity with Time by Well Between April 2018 and June 2021.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-18 Figure 5.2-2. Comparison of Model and Observed Total Mass Extracted by the RWS Between May 2018 and June 2021.

0 20 40 60 80 100 120 140 160 180 200 220 05/18 07/18 10/18 01/19 05/19 07/19 10/19 01/20 04/20 07/20 10/20 01/21 04/21 Salt Removed (lbs x 10E+06)

Date Salt Removed Per Month (Millions of Pounds)

Published (Data-Based) Estimate Calibrated

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-19 Figure 5.2-3. Comparison of Model and Observed Mass Extracted by Well Between May 2018 and June 2021.

0 5

10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-1 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-2 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-3 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-4 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-5 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-6 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-7 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-8 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-9 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated 0

5 10 15 20 25 05/18 08/18 11/18 02/19 05/19 08/19 11/19 02/20 05/20 08/20 11/20 02/21 05/21 Salt Removed (lbs x 10E+06)

Date RWS-10 (Million Pounds per Month)

Published (Data-Based) Estimate Calibrated

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-20 Figure 5.3-1a. Location of Initial, Year 5, and Year 10 Hypersaline Interface in Model Layer

4.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-21 Figure 5.3-1b. Location of Initial, Year 5, and Year 10 Hypersaline Interface in Model Layer 8.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-22 Figure 5.3-1c. Location of Initial, Year 5, and Year 10 Hypersaline Interface in Model Layer

11.

FPL Turkey Point RAASR Year 3 November 2021

5. Groundwater Model 5-23 Figure 5.3-2. Predicted 10-yr Capture Zones for Model Layers 4, 8, and 10.

FPL Turkey Point RAASR Year 3 November 2021

6. Summary and Recommendations 6-1 6

SUMMARY

AND RECOMMENDATIONS 6.1 OVERALL

SUMMARY

FPL submits this Year 3 RAASR, which covers the third year of RWS operation from October 1, 2020, to September 30, 2021, in compliance with the monitoring and reporting objectives of the MDC CA and the FDEP CO. While the COVID-19 pandemic impacted the Year 2 RAASR schedule and resulted in that report being submitted in two parts and covering 16 months of data, the schedule for the Year 3 RAASR has returned to normal with a single report covering a 12-month period. This report includes RWS operations and groundwater monitoring data collected from October 2020 to September 2021, Year 3 CSEM survey results based on the June 2021 survey, groundwater monitoring well contour mapping, and the updated/recalibrated regional 3-dimensional, density-dependent solute transport groundwater model (Version 6 [V6]). This report also includes a Performance and Compliance evaluation (Appendix I) addressing the first 3 years of RWS operation with respect to achieving the objectives of intercepting, capturing, containing, and documenting the extent of retraction of the plume. The following is a summary of the major findings of this evaluation:

• After 3 years of remediation operations, the CA objectives to intercept, capture, contain and demonstrate statistically valid reductions in the salt mass and volumetric extent (retraction) of hypersaline groundwater from the CCS have been met. The CO requirement to halt the westward migration of hypersaline water from the CCS within 3 years was documented to have been achieved in the April 2021 Year 2 Part 2 RAASR, and it has been confirmed by analyses conducted in this report.

• The RWS operated 98.1% of the reporting period; there were only 7 days (1.9%) when the entire system was shut off for maintenance. The individual wells collectively operated, on average, 92.5% this reporting period, a slight increase over Year 2 (91.7%).

• With each year of RWS operation, the net number of monitoring wells with statistically significant declining trends in chloride, salinity, and/or tritium have increased in all three FPL employs three types of data and associated analyses (monitoring, electromagnetic surveys, and modeling) to assess progress in meeting the objectives of the Consent Agreement and Consent Order. Analyses of data through Year 3 of remediation demonstrate that the net westward migration of the hypersaline plume has been halted, hypersaline groundwater from the CCS is being intercepted, captured, contained, and retracted by RWS operations. The CSEM data shows that the volume of hypersaline water in the compliance area has been reduced by 42% since remediation began in 2018.

FPL Turkey Point RAASR Year 3 November 2021

6. Summary and Recommendations 6-2 depth intervals, indicating RWS operations are effectively reducing hypersalinity throughout the vertical extent of the plume.

• Approximately 5.92 billion gallons of hypersaline water and 2.32 billion pounds of salt were removed during this reporting period. Since inception of the remediation system, approximately 18.45 billion gallons of hypersaline groundwater and 7.32 billion pounds of salt have been extracted from the Biscayne Aquifer.

• In total, 20 of 23 monitoring wells showed a statistically significant declining trend in one or more parameters (i.e., quarterly chloride, quarterly tritium, and weekly average automated salinity) and at least one or more parameters that were the lowest this reporting period compared to baseline and Years 1 and 2.

• The Year 3 CSEM results, compared to the 2018 baseline survey results, indicate the volumetric extent of the hypersaline plume has been reduced by 42% after 3 years of RWS operation. There are areas of statistically significant net retraction of the CCS-sourced hypersaline plume throughout the Biscayne aquifer west and north of the Plant site.

• Based on CSEM data, the greatest reduction in hypersalinity volume is occurring in the lower portion of the aquifer while the more significant reductions in groundwater monitoring well salinities are being measured in the upper portion of the aquifer as the plume shrinks from top to bottom.

• Two sources of hypersaline groundwater, coastal naturally occurring evaporated seawater and the CCS, occur within the survey area. Coastal evaporated seawater has been identified in both CSEM and modeling north of Plant, east of the plant entrance road, and east of the L31 levee, as well as in the area south of the S-20 structure east of the L31E levee and south of the CCS. Hypersaline surface water and shallow groundwater with fluid densities greater than underlying groundwater sink downward, resulting in both shallow and deep expressions of hypersaline groundwater.

• The groundwater model has been updated with Year 3 RWS data, groundwater monitoring well data, and CSEM data, and has been recalibrated to Years 1, 2, and 3 remediation data. These updates have reduced total model error (defined as the cumulative difference between model estimated value and actual measured values) and improved model representations of plume reductions associated with 3 years of remediation. However, significant differences remain between measured aquifer conditions and the model representation of these conditions. It is expected that these differences will continue to be reduced as the model is informed by subsequent annual remediation results.

• The Year 3 recalibrated V6 model forecast simulations for Years 5 and 10 of remediation shows an improved hypersaline retraction result in the northern portions of layers 9, but the southern portion and layers 10 and 11 do not show retraction by simulated Year 10 of

FPL Turkey Point RAASR Year 3 November 2021

6. Summary and Recommendations 6-3 remediation. While improvements in modeled orientation of the plume extent are achieved by the V6 model, further improvements are needed to reliably represent the dynamics of the hypersaline plume responses to the RWS in the lower portions of the aquifer. It is expected that these differences will continue to be reduced as the model is informed by subsequent annual remediation results.

6.2 REFINEMENTS FPL has implemented actions, including the following, to enhance the ongoing remediation and to further the objectives of the CO and CA:

Cooling Canal System Salinity Reduction

• A modification to the site certification license PA 03-45F was issued to FPL (October 19, 2021), authorizing use of the new F-7 CCS freshening well and increasing the UFA allocation for wells F-1 through F-7 from 14 mgd (5,110 million gallons per year) to 30 mgd (10,950 million gallons per year) with a maximum monthly allocation of 34 mgd (1,033.6 million gallons per month). This additional allocation will offset evaporative losses of fresh water from the CCS that exceed the 14 mgd previously allocated, thereby limiting dry season salinity increases that occur under the previous allocation and stabilizing CCS salinities at 34 PSU. FPL began increasing freshening inflows starting in November 2021.

Hypersaline Groundwater Remediation

• Extraction of up to 3 mgd of hypersaline groundwater from beneath the CCS using two existing Biscayne aquifer underground injection control test production wells was initiated during the Year 2 reporting period.

• Modifications to the UIC test production well connection and well operations are being implemented to maximize UICTPW production capacity to 3.6 mgd.

Groundwater Monitoring Network Expansion

• TPGW-22 was incorporated into the remediation compliance monitoring network on February 16, 2021. Data on groundwater salinity and analytic chemistry is being collected from all three monitoring well intervals, although automated water level data has yet to be reliably produced.

• An application for a SFWMD right-of-way permit for monitoring well site TPGW-23 was filed on May 13, 2021, and issued by SFWMD on July 22, 2021. The U.S. Army Corps of Engineers 408 permit application is under review at the time this report was filed.

FPL Turkey Point RAASR Year 3 November 2021

6. Summary and Recommendations 6-4 6.3 RECOMMENDATIONS Given the significant progress of remediation since initiation of the RWS, FPL does not propose any changes to the agencies approved remediation plan at this time. However, based on review of Year 3 CSEM survey data and the updated/recalibrated Year3 V6 forecast modeling, FPL shall incorporate the following actions and recommendations:

• Continue to utilize sensitivity analysis to investigate methods to better align model calculated plume orientations with CSEM and monitoring well data, particularly in the lower model layers, in order to produce more reliable long-range remediation forecasts.

• Investigate the continuity and hydraulic conductivity of the lower high flow zone and incorporate it accurately into the model. Similarly, the hydraulic properties vertically adjacent to the lower high flow zones should be investigated and incorporated accurately into the model.

• Detailed evaluation and incorporation of CCS sediment and sampling information should be used to guide how the modeled connection of the CCS to groundwater needs to be revised and represented.

• FPLs CCS water and salt-balance model should be used to inform the three-dimensional groundwater flow and salt transport model with respect to amounts of seepage to and from the CCS.

• CCS salinities and climate conditions should continue to be monitored and the model updated and recalibrated with more data reflective of longer RWS operations. The longer period of RWS operation, and consequent changes to salinities over a progressively larger area, will help inform the model and increase its accuracy in simulating the effect of the RWS and compliance with regulatory requirements.

• Evaluations should be conducted to verify the degree to which model-generated, non-CCS hypersaline groundwater impacts remediation objectives.

It is important to note that the aquifer system is complex and subject to many external factors beyond the CCS and RWS; therefore, continued monitoring, model updates, and scientific data analyses are performed to improve our understanding of the impact of RWS operations in concert with these other factors. FPL will continue to monitor and evaluate progress in meeting the requirements of the CA and CO and make recommendations for modifications as needed.

FPL Turkey Point RAASR Report Year 3 November 2021

7. References 7-1

7. REFERENCES Arcadis 2020. Review of Aerial Electromagnetic Surveys at Turkey Point Power Plant, Southern Florida; prepared for Miami-Dade County Department of Environmental Resource Management.

Cohen, R.M., Vincent, A.H., Mercer, J.W., Faust, C.R. and Spalding, C.P., 1994. Methods for Monitoring Pump-and-Treat Performance. U.S. Environmental Protection Agency, Office of Research and Development. EPA/600/R-94/123.

ENERCON. 2016. PTN Cooling Canal System, Electromagnetic Conductance Geophysical Survey, Draft Final Report, Florida Power and Light Turkey Point Power Plant, 9700 SW 344th Street, Homestead, FL 33035.

Fish, J.E. and M. Stewart. 1991. Hydrogeology, aquifer characteristics, and ground-water flow of the surficial aquifer system, Dade County, Florida. U.S. Geological Survey, Water Resources Inv. 91-4000.

Fitterman, D.V. and S.T. Prinos. 2011. Results of time-domain electromagnetic soundings in Miami-Dade and Southern Broward Counties, Florida. U.S. Geological Society Open-File Report 2011-1299, ix, 42 p.

Fitterman, D.V., M. Deszcz-Pan, and T. Scott. 2012. Helicopter Electromagnetic Survey of the Model Land Area, Southeastern Miami-Dade County, Florida. U.S. Geological Society Open-File Report 2012-1176:77.

Florida Power & Light Company (FPL). 2012. Florida Power & Light Company Comprehensive Pre-Uprate Monitoring Report for the Turkey Point Monitoring Project.

Prepared for Florida Power & Light Company by Ecology and Environment, Inc. October 31, 2012.

__________. 2013. Florida Power & Light Company Quality Assurance Project Plan (QAPP) for the Turkey Point Monitoring Project. Prepared for Florida Power & Light Company by Ecology and Environment, Inc. June 2013.

__________. 2016. Florida Power & Light Company Comprehensive Post-Uprate Monitoring Report for the Turkey Point Monitoring Project. Prepared for Florida Power & Light Company by Ecology and Environment, Inc. March 2016.

__________. 2017. Florida Power & Light Company Annual Monitoring Report for the Turkey Point Monitoring Project. Prepared for Florida Power & Light Company by Ecology and Environment, Inc. September 2017.

FPL Turkey Point RAASR Report Year 3 November 2021

7. References 7-2

__________. 2018a. Florida Power & Light Company Recovery Well System Startup Report.

Prepared for Florida Power & Light Company by Ecology and Environment, Inc. October 2018.

__________. 2018b. Florida Power & Light Company Turkey Point Recovery Well System (RWS) Second Quarter Status Report. December 2018.

__________. 2018c. Florida Power & Light Company Annual Monitoring Report for the Turkey Point Monitoring Project. Prepared for Florida Power & Light Company by Ecology and Environment, Inc. August 2018.

__________. 2019a. Florida Power & Light Company Turkey Point Recovery Well System (RWS) Third Quarter Status Report. March 2019.

__________. 2019b. Florida Power & Light Company Turkey Point Recovery Well System (RWS) Fourth Quarter Status Report. June 2019.

__________. 2019c. Florida Power & Light Company Turkey Point Remedial Action Annual Status Report. November 2019.

__________. 2019d. Florida Power & Light Company Annual Monitoring Report for the Turkey Point Monitoring Project. Prepared for Florida Power & Light Company by Ecology and Environment, Inc. August 2019.

__________. 2020a. Florida Power & Light Company Remedial Action Annual Monitoring Report for the Turkey Point Monitoring Project. Prepared for Florida Power & Light Company by WSP Inc. August 2020.

__________. 2020b. Florida Power & Light Company Turkey Point Remedial Action Annual Status Report, Year 2, Part 1. Prepared for Florida Power & Light Company by WSP.

November 2020.

__________. 2021a. Florida Power & Light Company Turkey Point Remedial Action Annual Status Report, Year 2, Part 2. Prepared for Florida Power & Light Company by Stantec.

April 2021.

__________. 2021b. Florida Power & Light Company Annual Monitoring Report for the Turkey Point Monitoring Project. Prepared for Florida Power & Light Company by Stantec. August 2021.

Hughes J.D. and White, J.T., 2014. Hydrologic Conditions in Urban Miami-Dade County, Florida, and the Effect of Groundwater Pumpage and Increased Sea Level on Canal Leakage and Regional Groundwater Flow. U.S. Geological Survey Scientific Investigations Report 2014-5162, 175 pp. https://doi.org/10.3133/sir20145162.

FPL Turkey Point RAASR Report Year 3 November 2021

7. References 7-3 JLA Geoscience, Inc. 2010. Geology and Hydrogeology Report for FPL, Turkey Point Plant Groundwater, Surface Water, and Ecological Monitoring Plan, FPL, Turkey Point Plant, Homestead, Florida. Prepared for Florida Power & Light Company. October 2010.

Langevin, C.D., Thore, D.T., Dausman A.M., Sukop, M.C., and Guo, W., 2008. SEAWAT Version 4: A Computer Program for Simulation of Multi-Species Solute and Heat Transport: USGS Techniques and Methods Book 6, Chapter A22, 39 p.

Meals, D.W., Spooner, J., Dressing, S.A. and J.B. Harcum. 2011. Statistical Analysis for Monotonic Trends, Tech Notes 6, November 2011. Developed for U.S. Environmental Protection Agency by Tetra Tech, Inc., Fairfax, VA, 23 p. Available online at https://www.epa.gov/polluted-runoff-nonpoint-source-pollution/nonpoint-source-monitoringtechnical-notes.

NOWDataNOAA Online Weather Data, 2021. http://www.weather.gov/wrh/climate?wfo-mfl Accessed November 9.

Prinos, S.T., M.A. Wacker, K.J. Cunningham, and D.V. Fitterman. 2014. Origins and delineation of saltwater intrusion in the Biscayne aquifer and changes in the distribution of saltwater in Miami-Dade County, Florida. U.S. Geological Survey Scientific Investigations Report 2014-5025. 101 pp. http://dx.doi.org/10.3133/sir20145025.

South Florida Natural Resources Center (SFNRC). 2012. Hydrology and Salinity of Florida Bay. Status and trends: 1990-2009. Technical Series 2012:1.

South Florida Water Management District (SFWMD). 2015. Applicants Handbook for Water Use Permit Applications within the South Florida Water Management District.

Stringer, C.E., M.C. Rains, S. Kruse, and D. Whigham. 2010. Controls on water levels and salinity in a barrier island mangrove, Indian River Lagoon, Florida. Wetlands. 30(4):725-734.

Tetra Tech, 2016a. A Groundwater Flow and Salt Transport Model of the Biscayne Aquifer, Technical Memorandum provided to Florida Power & Light, June 10, 2016.

Tetra Tech, 2016b. Application of Parameter Estimation Techniques to Simulation of Remedial Alternatives at the FPL Turkey Point Cooling Canal System, Technical Memorandum provided to Florida Power & Light, July 14, 2016.

Tetra Tech, 2016c. Addendum to Regional Biscayne Aquifer Model Report (Tetra Tech, 2016),

Technical Memorandum provided to Florida Power & Light, October 12, 2016.

Tetra Tech, 2017. Biscayne Aquifer Groundwater Flow and Transport Model: Heterogeneous Hydraulic Conductivity Analyses, Technical Memorandum provided to Florida Power &

Light, January 2017.