Text

Official Exhibit - FPL-030-00-BD01 - Revised Andersen Memorandum Evaluation of Drawdown in the Upper Floridan Aquifer Due to Proposed Salinity Reduction-based Withdrawals, Dated November 13, 2014
	ML16015A204
Person / Time
Site:	Turkey Point
Issue date:	11/13/2014
From:	; Florida Power & Light Co
To:	; Atomic Safety and Licensing Board Panel
	SECY RAS
References
	50-250-LA, 50-251-LA, ASLBP 15-935-02-LA-BD01, RAS 28505
	Download: ML16015A204 (20)
	v • d • e

United States Nuclear Regulatory Commission Official Hearing Exhibit In the Matter of: FLORIDA POWER & LIGHT COMPANY (Turkey Point Nuclear Generating, Units 3 and 4)

ASLBP #: 15-935-02-LA-BD01 Docket #: 05000250 & 05000251 FPL-030 Exhibit #: FPL-030-00-BD01 Identified: 1/4/2016 Admitted: 1/4/2016

(*at:) TETRA TECH Rejected:

Other:

Withdrawn:

Stricken:

TECHNICAL MEMORANDUM From: Peter F. Andersen and James L. Ross, Tetra Tech To: Rory Rahming, Florida Power & Light Company Date: November 13, 2014

Subject:

Evaluation ofDrawdown in the Upper Floridan Aquifer Due to Proposed Salinity Reduction-based Withdrawals 1 INTRODUCTION 1.1 Background Florida Power & Light Company (FPL) is in the process of applying for a modification to site certification to reflect the proposed reduction of salinity of cooling canal system (CCS) waters at the Turkey Point Power Plant, located near Florida City, Florida. A component of this project is a series of 1000-1200 foot deep wells that will extract low salinity water from the Upper Floridan Aquifer and discharge it into the CCS for the purpose of reducing the sal inity of CCS water to levels commensurate with Biscayne Bay. As a step in the site certification process, FPL must demonstrate the feasib ility of withdrawing approximately 14 million gallons per day (MGD) of Upper Floridan Aquifer water without adversely impacting the wells of existing legal users of the Floridan Aquifer. This memorandum describes the calibration and simulation of a groundwater flow model of the Floridan Aquifer system that is used to determine potential groundwater level (drawdown) changes resulting from the use of the Floridan Aquifer as a source of water for CCS salinity reduction.

1.2 Scope The scope of this analysis is to calibrate a regional groundwater flow model of the Upper Floridan Aquifer, as defined from regional hydrogeologic data, including two documented Floridan Aquifer Performance Tests (APTs). The modeling shall meet the minimum requirements of the South Florida Water Management District (SFWMD) Basis of Review (BOR) for water use permitting. Once calibrated, the model wi ll be used to evaluate the anticipated drawdown of the Upper Floridan Aquifer potentiometric surface at the plant site and regional settings. The drawdown information will be used to assess the likely impacts to the wells of existing legal users.

1.3 R eport Organization Following this introduction, the memorandum provides a summary of the existing regional groundwater model developed by the SFWMD that was modified andre-calibrated. This existing model is referred to as the East Coast Floridan Aquifer System Model - Phase 2 (ECFAS2). The calibration to the two APTs is then discussed, including changes that were made to the ECF AS2 model and the resulting quality of calibration. Pred ictive regional simulations and corresponding results follow.

2 METHODOLOGY 2.1 Geneml The methodology for conducting this study follows standard groundwater modeling protocols.

As outlined in Anderson and Woessner (1992) the steps involved with model application include:

Definition of purpose

Conceptual model development

Code selection Model design

Calibration I verification

Prediction

Presentation of results 2.2 Regional Model The primary purpose of the regional model analysis is to assess potential regional drawdown resulting from pumping water from the Upper Floridan Aquifer as a source of low-sal inity water for the CCS. Some of the early steps in the modeling process, most notably conceptual model development, model design, and, to some degree, calibration, were abbreviated in this application because the ECFAS2 model (Golder Associates, 2008) was available to use as the framework for the analysis. The abbreviated relevant steps are summarized in this section. The resulting revised model marks an FPL adaptation to the ECFAS2 model, and is herein referred to as the Adapted Floridan model.

The conceptual model of the natural system is consistent with that described in the existing ECFAS2 model documentation (Golder Associates, 2008). Additional data to modify the hydraulic parameters are available from site specific data collection and testing. Two APTs performed at the site are documented in JLA Geosciences (2006) and Dames and Moore (1975) and serve to supplement the conceptual model presented in the existing ECFAS2 model documentation (Golder Associates, 2008).

The design of the original model was generally unchanged. However, the modeled domain was truncated in the north such that the longitudinal extent of the revised model is less than that of the original. Additionally, the finite difference grid spacing was modified to account for well locations used in the APTs that are simulated in the model re-calibration. Grid modifications are described in Section 3.2. Additionally, since relative changes in flow cond itions (i.e. drawdowns) are the focus of both model calibration and predictions, only the groundwater flow component of the original model is evaluated and employed, herein. Logistically, this decision faci litated efficient model calibration and predictive simulations, as consideration of density-dependent flow and transport resulted in very long run times. The original groundwater flow and transport model was calibrated to regional water levels and saltwater concentrations. To account for site-specific conditions, the model was re-calibrated to two APTs conducted at the site.

2 TETRA TECH

3 Regional Model Simulations 3.1 ECFAS2 Model The SFWMD, through contractors, developed a density-dependent groundwater flow and saltwater transp01t model of the East Coast of F lorida in two phases. The first phase, ECFAS 1 (HydroGeologic, 2006), simulated the southern half of the study area (the Lower East Coast of Florida); the second phase (ECFAS2) expanded the model domain northward to include more of the East Coast of Florida (Golder Associates, 2008). Both phases of the ECFAS model are available from the SFWMD; only the former has been peer-reviewed. Nevertheless, these model s represent the best available framework from which to base a permitting-level analysis of regional Floridan Aquifer impacts resulting from pumping.

The ECFAS2 model encompasses the ECFAS1 region and represents a revision to the earlier work. Consequently, the ECFAS2 model was used as the framework for this analysis. The ECFAS2 model covers the much of the East Coast of Florida, from southern Indian River County to the Florida Keys. This area is discretized into uniform 2400 by 2400 ft cells.

Vertically, the model extends from land surface to the Boulder Zone, a depth of approximately 3000 ft. The vertical section is discretized into 14 layers, with the Upper Floridan Aquifer represented as 2 layers. Boundary conditions are specified to represent flow into and out of the model domain, usually along the perimeter of the study area. Both flow (hydraulic heads) and saltwater transport (TDS concentrations) are simulated and are dependent upon one another (density-dependent flow and transp01t) . Field data from numerous borings were used to establish the structure of the model layering, which represents the hydrostratigraphic layers. In addition, field data from APTs were used to guide the initial choice of hydraulic parameters that were used in the model calibration. The model was calibrated to both hydraulic heads and concentrations.

Even though the model was calibrated, Golder Associates (2008) found that the model 's size resulted in exceptionally long run times such that the scope of the calibration had to be reduced from what was originally envisioned.

3.2 Adapted Floridan Model The ECFAS2 model was not usable in its avai lable state because it covers a very large area and does not provide the resolution required to accurately assess site-specific features and impacts.

Several structural modifications were made to the model and are described herein. Modifications to the calibration of the model are discussed in this section. As previously mentioned, only the groundwater flow capabilities of the ECFAS2 model were germane to the analyses of drawdown described herein, as regional changes in water quality attributable to the proposed wells, as well as the impact of such changes on drawdown, are anticipated to be negligible. Moreover, model run times were dramatically reduced by eliminating the density-dependence.

Since the Adapted Floridan model simulates groundwater flow and is adapted from the SEAWAT-based ECFAS2 model, the USGS simu lation software MODFLOW-2000 (Harbaugh, et al, 2000), a commonly applied groundwater flow model, was used to simulate the regional model. MODFLOW-2000 is capable of addressing the requirements of the SFWMD BOR inasmuch as it:

3 TETRA TECH

simulates groundwater flow,

is capable of addressing mu ltiple hydrostratigraphic layers and subdividing these layers such that drawdown can be computed at multiple levels within each layer, and

is in the public domain, peer-reviewed, and widely used .

The most significant structural change to the model was the grid spacing, which was originally set at 2400 ft. For calibration purposes, the grid was refined in the immediate vicinity of the Turkey Point APTs, such that the well spacing for the APTs could be accurately represented and changes in head over small distances resolved. The revised grid spacing in the model for the calibration is shown in Figure la. The minimum grid spacing used in the Adapted Floridan model, near pumping and monitoring wells, is as little as 1.5 ft. The original model grid spacing, shown in Figut*e lb, was used in subsequent predictive runs because it was adequate for assessment of impacts at the desired scale and was practical from a run-time perspective.

The original model layering was retained because it appeared to be generally appropriate for the level of detail required. The Intermediate Confining Unit (ICU), which overlies the Upper Floridan, was represented using a single layer.

The additional pumping wells that were included as a part of the calibration of the Adapted Floridan model also represent modifications to ECF AS2. The well locations and rates are described in the calibration and model results sections below. The time stepping of the models was also modified to provide adequate resolution for the duration of the APTs and to account for intermittent pumping (Section 3.2.1.1 and 3.2.1.2).

3.2.1 Additional calibration of model Although the ECFAS2 model may represent the regional conditions fai rly well, it may not represent site-specific conditions particularly well. This hypothesis was tested by running the model using documented pumping stresses on the system and comparing the modeled response to that which was observed during the test. In general, as discussed below, the compari son was not good. In order to obtain a reasonable representation of site-specific conditions, two additional calibrations, one to a short-term APT and another to a longer term APT, were performed. The ability to match aquifer system response to these APTs prov ides confidence that the model can predict the response to future proposed pumping. Modeled water levels were checked to ensure that the match to regional calibration targets had not been degraded as a result of the local changes. The methodology and resu lts of each of the additional calibrations are described below.

3.2.1 .1 JLA APT JLA Geosciences (2006) conducted an APT in support of the Unit 5 site certification. Floridan water supply well PW-1 was pumped for 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months and drawdown was measured in two other water supply wells and a shallow observation well. The drawdown response documented during this test was believed to represent a good series of targets to match as a part of a calibration because it was local to the area of proposed pumping and was conducted under quality-controlled conditions. However, it was recognized that the short duration of the test and extent of monitoring points would provide data that may only be representative of a relatively small area.

4 TETRA TECH

Simulation of the APT was accomplished using the revised model grid. Well PW -1 was represented with a single well pumping at a rate of 4500 gpm in model layers 3 and 4, which represent the Upper Floridan Aquifer, in the cell at row 166, column 143 Timestepping ranged from a minimum of 5 to a maximum of 567 minutes, Drawdown response was noted in wells PW-3 (layers 3 and 4, row 168, column 171), PW-4 (layers 3 and 4, row 180, column 157), and OBS-1 (layer 2, row 166, column 143) at distances of 3036, 1686, and 0 feet, respectively from the pumped well. Note that OBS-1 is co-located with the pumping we ll , but is screened near the base of the Biscayne Aquifer and did not experience drawdown in response to the APT.

Comparison of modeled to observed conditions for the original model, prior to adjustment, was not good, with a residual standard deviation of greater than 100 ft. However, as shown in F igure 2, this match improved cons iderably (residual standard deviation of 0.36 ft) after adjustment of hydraulic parameters as a pmt of the calibration. In general, hydraulic conductivities were increased from their original values during calibration. Goodness-of-fit ca libration metrics are shown in Table 1 and indicate that the model provides a reasonable fit to observed data.

Table 1. Goodness of fit metrics for the JLA APT calibration Metric Numerical Value Mean Error, ft 0.22 Mean Absolute Error, ft 0.33 Residual Standard Deviation, ft 0.36 Range of Targ_ets , ft 6.36 Residual Standard Deviation I Range *100 5.6%

Note that this calibration was conducted iteratively with the Dames and Moore APT described below and hence the calibrations strike a balance between matching the results of both APTs with the same set of parameters.

3.2.1.2 Dames and Moore APT Dames and Moore (1975) conducted an APT in support of a feasibility study for using Floridan Aquifer water to cool the original Turkey Point nuclear units. Floridan Aquifer production test well (PTW) was pumped for 90 days and drawdown was measured in eight monitoring wells at various distances from the pumped well and depths in the aquifer. The drawdown response documented during this test was believed to represent a good series of targets to match as a part of a calibrati on because of its long duration and use of monitoring points that were distant from the pumping well. Thus, this test was complementary to the shorter duration, more local JLA APT described above.

As in the simulation of the JLA APT, the simulation of the Dames and Moore APT was accomplished using the refined model grid. Well PTW was represented with a single well pumping at a rate of 5000 gpm in cell layers 3 and 4, row 220, and column 97. Timestepping ranged from a minimum of 73 minutes to a maximum of 11.8 days. Drawdown response was noted in wells OW-A (row 229, column 108), OW-B (row 238, column 120), OW-C (row 207,column 82), and OW-D (row 258, co lumn 181) at distances of 100 feet, 500 feet, 2000 feet, and 48,000 feet, respectively from the pumped well. Drawdown was recorded in the Upper and Middle Floridan aquifers at each of the four observation well sites, which are represented by layers 3 and 4, and 7 and 8, respectively 5

TETRA TECH

Comparison of modeled to observed conditions for the original model, prio r to adjustment, was not good (residual standard deviation in excess of I 0 ft), as was the case for the JLA APT. As shown in Figure 3, this match also improved considerably (with a residual standard deviation of 0.77 ft) after adjustment of hydraulic parameters as a part of the calibration. Goodness-of-fit calibration metrics are shown in Table 2 and indicate that the model provides a reasonable fit to observed data.

Table 2. Goodness of fit mctrics fo1* the Dames and Moore APT calibration Metric Numerical Value Mean Error, ft -0.30 Mean Absolute Error, ft 0.66 Residual Standard Deviation, ft 0.77 Range of Targets, ft 11.8 Residual Standard Deviation I Range

100 6.5%

Though the wells shown in Figure 3 are not an exhaustive representation of the calibration targets, they are a microcosm of the quality of the model match to this APT. The lateral and vertical proximity to the pumping well precluded a reasonable match to the observed drawdown at well OW-A (Upper); as such, this well was omitted from the calibration.

3.2.1.3 Adjustments to the calibration The primary parameters that were changed as a result of the additional calibration were hydraulic conductivities of the Upper Floridan Aquifer (UFA), ICU, Middle Confining Unit (MCU), and the Middle Floridan Aquifer (MFA). These parameters were all raised from their orig inal values, as shown in Table 3.

Note that the parameter changes were made within zones that were near the Turkey Point site and mostly in areas potentially affected by drawdown from proposed salinity reduction wells, as shown in Figure 4, 5, and 6.

The changes made to the hydra ulic properties in the Adapted Floridan model are not expected to significantly impact the quality of the model match to the water level and water quality targets employed in the calibration of the ECF AS2 model. The changes made to the Adapted Floridan model were generally minor, and the preponderance of the ECFAS2 model calibration targets are located outside of the Adapted Floridan model doma in. Water levels at well ENP-100, located about 17 miles to west of Turkey Point, were simulated to be approximately 36ft, which slightl y underestimates observed water levels (approximately 37.5 ft NGVD29). The ECFAS2 model si mulates a water level of 39 feet at this location, after 350 days of simulation. Simulated water levels at W ASA-South, located to the n011h of Turkey Point, are approximately 35 ft NGVD 29; this well, however, was omitted from the ECFAS2 modeling analysis due to anomalous water levels (Golder Associates, 2008).

6 TETRA TECH

Tablc3 P arameter c 1anges resu I tmg

. f rom ca J'b 1 ration o f t I1c Ad apted Flone

'I an mo d e.I ECFAS2 model FPL Floridan Model Hydrologic Model Aquifer (original) (recalibrated)

Unit Layers Parameter magnitude magnitude 0.0006 0.001 Kz (fUd)

ICU 2 0.000075 0.001 Kh (fUd) 0.000075 0.001 5.2 100 Kz (fUd) 9 15 72.5 225 0.33 225 UFA 3,4 52 100 Kh (ft/d) 90 150 725 330 3.33 330 Ss 5.25E-07 8.00E-07 0.004 0.003 0.000002 0.08 Kz (fUd) 0.4 0.003 MCU 5,6 0.002 0.08 0.02 0.08 Kh (fUd) 0.00001 0.4 0.03 Kz (fUd) 5.2 30 450 900 MFA 7,8 300 600 Kh (fUd) 180 1200 52 600 0.0015 0.01 MC2 9,10,11 Kz (ft/d) 0.0002 0.02 3.3 Predictive Simulations Once calibration of the regional Floridan model was confirmed, equilibrium flow conditions were established by running the model, holding all flow boundaries (e.g. specified heads) constant until changes in the simulated flow field in the Florida n Aquifer System were negligible. No pumping was simulated in this equi librium model. The resulting equilibrated state formed the initial conditions for ensuing predictive simulations. Equilibrated regional water 7

TETRA TECH

levels, especially near Turkey Point, were generally lower than observed water levels; this is due to the exclusion of salt transport and the associated density-dependent flow. However, given that purpose of th is model is to provide estimates of relative changes in water level , the low simulated water levels were deemed irrelevant. Since the focus of the salinity reduction well evaluation is regional drawdown, the original 2400-ft grid spacing was employed for predictive simulations.

According the SFWMD BOR, predictive evaluations made with the calibrated model must be conducted using monthly stress periods that simulate average annual groundwater withdrawals subject to rainfall that alternates between average and 1-in-1 0 year draught conditions (3 months of average cond itions, followed by 12 months of drought conditions, followed by 6 months of average conditions). As such, the predictive models were conducted using 21 monthly stress periods .. Due to the specified head boundary in the topmost layer of the model, climatic stressess were not varied between average and drought conditions. Moreover, it is not anticipated that variation in rainfall would impact the Floridan aquifer over the 21-month timeframe of the simulation.

Additionally, the BOR stipulates that the 1-ft drawdown contour associated with the proposed pumping be simulated and the impacts to existing legal users' wells within that contour be evaluated. The process by which this was accomplished is described below.

3.3.1 Proposed Salinity Reduction Well Operation There are six proposed salinity reduction wells. At any one time, five of these well s will collectively pump 14 MOD of low salinity water from the Upper Floridan aquifer. The six wells will be spaced approximately 1900 ft apart, along the northernmost canal of the Cooling Canal System and along the Interceptor Ditch (Figure 7). In the model , the 14 MGD of pumping is distributed evenly amongst the five active wells and is assumed to be a constant rate of pumping over the course of the 25-year simulation. Two alternative pumping scenarios are considered in this modeling analysis and differ in the aJJocation of pumping to wells F-2 and F-6. The base scenario simulates pumping at wells F-1 through F-5 (no pumping at F-6); the alternative scenario simulates pumping at wells F-1 and F-3 through F-6 (no pumping at F-2).

The salinity reduction wells were simulated, starting from an equilibrium flow field. Aside from these salinity reduction wells, no pumping was simulated in this model such that the permitted users w ithin the 1-ft drawdown contour could be delineated. At the conclusion of the 21-month simu lation, the simulated drawdowns in the regional model are those attributable only to the five proposed salinity reduction wells. Figure 8a illustrates these regional drawdowns associated with the base pumping scenario. In this base simulation, the drawdowns at a distance from the site are affected by variations in hydraulic conductivity; this is evident upon inspection of the 1-ft drawdown contour, wh ich generally has an oblong shape, whose major axis is oriented north-to-south . Nearer to the site, the drawdown contours radiate outward from the wells in a more uniform manner, where the maximum drawdown is approximately 15.1 ft, near well F-3. In the alternative scenario, the maximum drawdown is approximately 14.4 ft, near well F-5.

As previously mentioned, the SFWMD BOR dictates tbat drawdown at permitted users' wells encircled by the 1-ft drawdown contour be determined. As illustrated in Figure 8a, the following permitted users fall within the 1-ft drawdown contour:

8 TETRA TECH

Card Sound Golf Club,

Ocean Reef Club,

the F loridan Keys Aqueduct Authority (FKAA),

Miami-Dade Water and Sewage Department South Miami Heights Wellfield, and

FPL Unit 5 Wells.

A second 21-month simulation was conducted wherein pumping by permitted users and the proposed salinity reduction wells was simulated. Predicted drawdowns attributable solely to permitted Floridan pumping were determined by comparing drawdowns from this simulation (Figure 8b) to those attributable to the proposed salinity reduction wells alone (Figure Sa).

These permitted well-based drawdowns are illustrated in Figure 9 for the base pumping scenario. Pred icted drawdowns at Floridan wells of these existing legal users due to the proposed operation of the salinity reduction wells are presented in Table 4. These drawdowns are calculated at the center of the model grid cells in which the respective wells are simulated. In addition to drawdowns attributable to the proposed wells for the base pumping allocation scenario, cumu lative drawdowns at nearby wells due to both pumping at permitted and proposed welJs are provided in Table 4. Withdrawals by nearby users were simulated at their respective permitted rates.

Table 4. Predicted drawdown at nearby users for the proposed Salinity Reduction Wells due to the base pumpmg scenano.

Distance Base Scenario Permitted Base Scenario Location from Drawdown at Facility Withdrawal Cumulative (L,R,C) well F-2 3,15, and 21 (MGD) Drawdown (ft)

(miles) Months (ft)

Card Sound Golf (3-Club (WUP 44- 0.58 8.8 1.85/2.20/2.21 9.83/11.47/1 1.54 4, 173,93) 00001)

Ocean Reef Club (3-1.42 8.8 1.85/2.20/2.21 9.83/1 1.47/11 .54 (WUP 44-00002) 4,173,93)

FKAA (3-9.70 10.3 1.76/2.15/2. 16 416.06/17.78/1 7.85 (WUP 13-00005) 4, 155,61)

South Miami Hts (3-4, 133-23.3 10.3 1.83/2.25/2.26 46.71/48.72/48.80 (WUP 13-00017) 135,79)

FPL Unit 5 Well (3-14.3 < 1.0 11.44/11 .85/11 .86 34.11/35.92/35.99 (PW-1) 4, 156,85)

A second evaluation was conducted in which the a lternative pumping allocation (wells F-1 and F-3 through F-6) for the salinity reductions wells was simulated. The resulting simulated drawdowns at legal users within the 1-ft drawdown contour are provided in Table 5; cumu lative drawdowns are also tabulated. Inspection of the drawdowns in Table 5 reveals that they are not significantly different from those produced by the base pumping allocation.

The cumulative drawdown due to permitted pumping, as illustrated in Figure 9, are not significantly different than those produced by the combination of proposed and permitted withdrawals (Figure 8b ). This suggests that the proposed pumping of Floridan water by the salinity reduction wells will not s ignificantly exacerbate drawdowns in the Upper Floridan aquifer beyond those induced by existing permitted pumping.

9 TETRA TECH

Table 5. Predicted drawdown at neuby users for the proposed Salinity Reduction Wells due to the a Iternatlve

. pumpmg

. scenano.

Alternative A lternative Distance Scenario Location Scenario Facility from well Drawdown at (L,R,C) Cumulative F-2 (miles) 3,15, and 21 Drawdown (ft)

Months (ft)

Card Sound Golf Club (3-4, 173,93) 8.8 1.86/2.21/2.22 9.84/11.48/11 .56 (WUP 44-00001)

Ocean Reef Club (3-4, 173,93) 8.8 1.86/2.21/2. 22 9.84/11.48/11.56 (WUP 44-00002)

FKAA (3-4, 155,61) 10.3 1.78/2.17/2.18 16.08/17.80/17.87 (WUP 13-00005)

South Miami Hts (3-4, 133-10.3 1.81 /2.22/2.24 46.68/48.69/48.78 (WUP 13-00017) 135,79)

FPL Unit 5 Well (PW-1) (3-4,156,85) < 1.0 9.92/10.33/10.35 32.59/34.40/34.47 4 Conclusions The evaluation of drawdown due to pumping at the proposed salinity reduction wel ls is based on the ECFAS2 model developed for the SFWMD. This model was subsequently adapted to site-specific conditions and re-calibrated to two APTs performed at Turkey Point. The resulting regional calibrated groundwater flow model provides assessment of drawdown at nearby existing Floridan water users.

In a regional sense, the proposed pumping of 14 MGD is projected to result in a maximum Upper Floridan Aquifer drawdown ranging between 14.4 ft (alternative scenario) and 15.1 ft (base scenario) at the Turkey Point site; simulated drawdowns at a distance from Turkey Point are not significantly different between the two pumping scenarios. The extent of drawdown, as defined by the 1-ft drawdown contour encompasses four existing legal users. Overall, the impacts to off-site permitted wells are minor. The maximum drawdown due to the proposed salinity reduction wells experienced by the nearest (non-FPL) users is 2.26 ft and occurs at the South Miami Heights wellfield, located approximately 10.3 miles away. Th is drawdown comprises approximately 4.6% of the cumulative drawdown simulated at this s ite. The drawdown contribution by the proposed salinity reduction wells is a conservative estimate (greater than would actually be experienced), since the drawdown in the wellbore at each nearby user due to localized pumping is undersimulated by the coarse-gridded regional model.

In addition to a demonstration of minimal drawdown induced at wells of permitted users within the 1-ft drawdown contour, the BOR also stipulates that the proposed pumping not impact the saltwater interface, as defined by the 250 mg/L isochlor. As the quality of Upper Floridan Aquifer water in this area already exceeds such a concentration, and no saltwater interface exists, this stipulation does not apply to the proposed project. Moreover, the operation of the salinity reduction well is not expected impact Upper F loridan water quality in a regional sense. Local changes in water quality are expected to be minor, as demonstrated by other Upper Floridan water users in the region (SFWMD, 20 12).

10 TETRA TECH

5 References Anderson, M.P., and Woessner, W.W., 1992, Appl ied Groundwater Modeling- Simulation of Flow and Advective Transport: San Diego, Ca, Academ ic Press, 381 p.

Dames and Moore, 1975. Floridan Aquifer Water Supply Investigation, Turkey Po int Area, Florida.

Golder Associates, 2008. East Coast Floridan Aquifer System Model, Phase 2, Southeastern Florida, final Model Documentation Report, October 2008, 259pp.

Harbaugh, A.W., Banta, E.R., Hill, M.C., and McDonald, M.G., 2000, MODFLOW-2000, the U.S. Geological Survey modular ground-water model -- User guide to modularization concepts and the Ground-Water Flow Process: U.S. Geological Survey Open-File Report 00-92, 121 p.

HydroGeoLogic, 2006. Development of a Density-Dependent Saltwater Intrusion Model for the Lower East Coast Project Area, April 2006. 166 pp JLA Geosciences, 2006. Well Completion Report for Floridan Aquifer Wells PW-1 , PW-3, and PW -4. FPL Turkey Point Expansion Project (Unit 5) Homestead, Florida.

Langevin, C.D., D.T. Thorne, Jr, A.M . Dausman, M.C. Sukop, and W. Guo, 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.

South Florida Water Management District (SFWMD), 2012, Overview and Current Use of the Floridan Aquifer System in the Lower East Coast, Public Workshop, Pompano Beach, Florida, Apri118, 2012.

Zheng, C., and P. Wang, 1999, MT3DMS, A modular three-dimensional multi-species transport model for simulation of advection, dispersion and chemical reactions of contaminants in groundwater systems; documentation and user's guide, U.S. Army Engineer Research and Development Center Contract Report SERDP-99-1, Vicksburg, MS, 202 p 11 TETRA TECH

P.\FPL\SaliMy Reduction\GISIAg1_ModciGrid mxd Legend TITLE:

Model Grid Spacing in the Vicinity of Turkey Point for a) Calibration Model and b) Predictive Models D Calibration Model Extent LOCATION; D Original Model Extent Turkey Point, Florida

- - Cooling Canals

-Model Grid

[1t:] TETRA TECH 1-CH

--Ec::CKC::E=!-P'-'f,:_:A'----

DRAFTED SCS PROJ DATE 117-2826019 0311012014

- -j FIGURE:

1

7 PW3

~

6 I I 0 Simula ted Drawdown I I - Observed Drawdown 5 -

~4 - ,

~ I

~3 I J t 1 I II IIIII 0

II I 'II I I 7 OB 1

Ill Ill I I 1r1 I Slj 1111 .I I 0 I I Slmulal cd Drawdown fl 0 I I 6 -

II II I II J- obs e"':'e~ Orawdown , I Ill Ill I II I II I Il l 5 I I II II I I ll 0 1 2 3 4 5 6 7 g4 Elapsed Time {d) c

~3 I I II Ill Ill

-o 3:

~2 - I I I II Ill 7

PW4J I IJJJ, I I I 1 I 1 Ill ,, II II1 6

I . 0 Simulated Orawdown Ill I I II I 5 - ill II - Observed Orawdown 0 I ill! I IIIII I II IIIII

£4 I II II II II -1 0 1 2 3 4 5 6 7 L Iill I I I Elaps ed Time (d)

-o 3:

~2 II II I I 1 IIIII I I I I 0 I II I I ::r I Ill TITLE:

Modeled vs observed drawdown for the JLA APT -

-1 I II I I I II I I 1111 I LOCATION:

Calibrated Adapted Floridan Model 0 1 2 3 4 5 6 7 Turkey Point, Florida Ela psed Time (d)

I CHECKED. PFA FIGURE:

["R: TETRA TECH lnRAFTEDc FILE:

JlR 117*2826019 2 DATE: 031091201 4

8 now-BUpper 8 rtow-B Lower I I 7

0 Simulated l>rawdown 0 Slmula1ed Orawdown 1-7

- Observed Orawdown - Observed Drawdown I I I I I 6 6 l/'"*nl I ~~ ~1

E: s b ""II

E: s I I c

..,~4 I c

..,~4 I g3 II I II 3:

,53 II"" !A I """ ..... ~ f\,

2 2 -

u~ op o o c ) ( ! c 0 I~ h r

1 1 0 I.I ~~ ftl~ 0 II ~~ I 0 20 40 60 80 100 120 0 20 40 60 80 100 120 Elapsed Time (d) Elapsed Time (d) 8 7

dow-o Upper I I 0 I I Simulated Orawdown 8

7 i ow-o Lower I I I I O Simullled Orawdown I

Observed Orawdown r

6 II I I - Observed Orawdown 6

II I I I I

E: s I I I ;E:s I I Ill I c

..,~4 I II ..,L I II 3:

,53 I II I 3:

,53 I I I II I I 2 I II

.,..I 2

I I I

I .* .

1 0

0 I

~P'~

20 o ,s; 40 If. PI *"'r-- "Hi ~ ~

60 80

~;n~

100 120 1

0 0

-...-nbc 0 10 p O J C ?

20 40 60

. '*.1

. I 80

~~

100 120 Elapsed Time (d) Elapsed Time (d) nne:

Modeled vs observed drawdown for the Dames and Moore APT -Adapted Floridan Model lOCATION:

Turkey Point, Florida CH ECKED: PFA FIGURE*

["11;] T ETR A T E CH lnRAFTEDo FILE:

LR 117-2826010 3 DATE: 0)1091201-4

P.\FPL\Salinit)' Redudron\GIS\Ftg4_Con<A.2.mKd 0 __

4.5 9 Scattln MIH 18 I

Legend Original Conductivity - - Coo ling Cana ls Calib_rated Conductivity m'bistribution of Vertical Hydraulic Conductivity in the Interm ediate Confining Unit CJ 3.0E-06 ft/d a) Original Model, PriortoAddilional Calibration a nd b) After Additional Calibration CJ 3.0E-06 ft/d D 7.5E-05 ft/d D 6.0E-04 ft/d LOCATION:

Turkey Point, Florida D 6.0E-04 ft/d - 1.0E-03 ft/d CHECKE PFA

('"ft:) TETRA TECH FIGURE:

2.5E-03 ft/d 2.5E-03 ft/d ORA FTEO SCS 9.5E-03 ft/d PROJ 117-2826019 4 9.5E-03 ft/d DATE 03/1012014

P.\fPL\Salinit)' Redotfon'GS\FigS_Condl3.mxd Legend 0 5 10 Scale In Min 20 TITlE:

Original Conductivity D 90 ft/d Calibrated Conductivity Distribution of Hydraulic Conductivity in the Upper Floridan Aqui fe r a) Original Model, Prior to Additional Calibration and b) After Additional Calibration 725 ft/d 55 ft/d D 330 ftld LOCATION; o 224.995 wd n 52 ft/d D 150 ftld Turkey Point, Florida D 1oo ftld LJ 3.33 ftld

- - Cooling Canals CJ 100 ftld 55 ft/d

[1'1:] TETRA TECH I::CH':"E:.::C::KE::f.:-

ORAFTE PROJ DATE PO::FA:-------J SCS 117*2826019 03/1012014 FIGURE:

5

P;\FPL\Salinity Reduction\GIS\F1g6_Coodl6.mxd

_f 0

-- -- 5 20 Legend Calibrated Conductivity Original Conductivity - - Cooling Canals D 2.0E-03 ft/d TITLE: Distribution of Vertical Hydraulic Conductivity in t he Middle Confining Unit a) Original Model, Prior to Additional Calibration and b) After Additional Calibration CJ 2.0E-06 ft/d 3.0E-03 ft/d LOCATION:

D 2.0E-03 ft/d D 4.0E-03 ft/d Turkey Point, Florida D 4.0E-03 ft/d D

5.0E-03 ft/d 1.5E-02 ft/d D 1.5E-02 ft/d 8.0E-02 ft/d 4.0E-01 ft/d

(-n:] TETRA TECH CHEC ORAFTE PROJ OATE PFA scs 117-2826019 0311012014 FIGURE:

6

P:IFPL\Salini!y Reductlon\GIS\NewWells_SCA.mxd t

F-1J 1

Legend TITLE: Location of Proposed Salinity Reduction Wells

Proposed Wells Screened in the U er Floridan A uifer LOCATION:

Existing Unit 5 Wells Turkey Point, Florida

['A:) TETRA TECH CHECKE DRAFTED PROJ DATE PFA JlR 117-2826019 03/31/2014 FIGURE:

7

P.\FPL\SW'lnity Recmtion'1GIS\Redonat_OON_21Mo_SCA.mxd 0

5 10 Sea~ In Mles 0

- 5 10 Scat.lnMies TITlE:

a) Predicted UFA Drawdown Due to Proposed Wellfield, and Legend b) Cumulative UFA Drawdown Due to Permitted and Proposed Wellfields

Existing User Wells LOCATION:

Turkey Point, Florida Drawdown Contours (1 ft)

[-n::)

CHECK PFA FIGURE:

ORAFTE JLR TETRA TECH PROJ 117-2826019 8 DATE 11/5/2014

P:IFPL\Salinily Reduction'I31S\Permilled_Q_ DDN_ 21 Mo_ SCAmxd Club (WUP 44-00002) 0

-- 4

-- 8 16 Legend TITLE: Simulated Equilibrium Drawdown in Upper Floridan Aquifer

- - Drawdown Conto urs due to Permitted Pumping LOCATION;

- - Coo ling Canals Turkey Point, Florida CHECKE PFA FIGURE;

Existing User Wells

["jl;:) TETRA TECH DRAF TED JLR PROJ 117-2826019 9 DAT E W S/2014

United States Nuclear Regulatory Commission Official Hearing Exhibit In the Matter of: FLORIDA POWER & LIGHT COMPANY (Turkey Point Nuclear Generating, Units 3 and 4)

ASLBP #: 15-935-02-LA-BD01 Docket #: 05000250 & 05000251 FPL-030 Exhibit #: FPL-030-00-BD01 Identified: 1/4/2016 Admitted: 1/4/2016

(*at:) TETRA TECH Rejected:

Other:

Withdrawn:

Stricken:

TECHNICAL MEMORANDUM From: Peter F. Andersen and James L. Ross, Tetra Tech To: Rory Rahming, Florida Power & Light Company Date: November 13, 2014

Subject:

Evaluation ofDrawdown in the Upper Floridan Aquifer Due to Proposed Salinity Reduction-based Withdrawals 1 INTRODUCTION 1.1 Background Florida Power & Light Company (FPL) is in the process of applying for a modification to site certification to reflect the proposed reduction of salinity of cooling canal system (CCS) waters at the Turkey Point Power Plant, located near Florida City, Florida. A component of this project is a series of 1000-1200 foot deep wells that will extract low salinity water from the Upper Floridan Aquifer and discharge it into the CCS for the purpose of reducing the sal inity of CCS water to levels commensurate with Biscayne Bay. As a step in the site certification process, FPL must demonstrate the feasib ility of withdrawing approximately 14 million gallons per day (MGD) of Upper Floridan Aquifer water without adversely impacting the wells of existing legal users of the Floridan Aquifer. This memorandum describes the calibration and simulation of a groundwater flow model of the Floridan Aquifer system that is used to determine potential groundwater level (drawdown) changes resulting from the use of the Floridan Aquifer as a source of water for CCS salinity reduction.

1.2 Scope The scope of this analysis is to calibrate a regional groundwater flow model of the Upper Floridan Aquifer, as defined from regional hydrogeologic data, including two documented Floridan Aquifer Performance Tests (APTs). The modeling shall meet the minimum requirements of the South Florida Water Management District (SFWMD) Basis of Review (BOR) for water use permitting. Once calibrated, the model wi ll be used to evaluate the anticipated drawdown of the Upper Floridan Aquifer potentiometric surface at the plant site and regional settings. The drawdown information will be used to assess the likely impacts to the wells of existing legal users.

1.3 R eport Organization Following this introduction, the memorandum provides a summary of the existing regional groundwater model developed by the SFWMD that was modified andre-calibrated. This existing model is referred to as the East Coast Floridan Aquifer System Model - Phase 2 (ECFAS2). The calibration to the two APTs is then discussed, including changes that were made to the ECF AS2 model and the resulting quality of calibration. Pred ictive regional simulations and corresponding results follow.

2 METHODOLOGY 2.1 Geneml The methodology for conducting this study follows standard groundwater modeling protocols.

As outlined in Anderson and Woessner (1992) the steps involved with model application include:

Definition of purpose

Conceptual model development

Code selection Model design

Calibration I verification

Prediction

Presentation of results 2.2 Regional Model The primary purpose of the regional model analysis is to assess potential regional drawdown resulting from pumping water from the Upper Floridan Aquifer as a source of low-sal inity water for the CCS. Some of the early steps in the modeling process, most notably conceptual model development, model design, and, to some degree, calibration, were abbreviated in this application because the ECFAS2 model (Golder Associates, 2008) was available to use as the framework for the analysis. The abbreviated relevant steps are summarized in this section. The resulting revised model marks an FPL adaptation to the ECFAS2 model, and is herein referred to as the Adapted Floridan model.

The conceptual model of the natural system is consistent with that described in the existing ECFAS2 model documentation (Golder Associates, 2008). Additional data to modify the hydraulic parameters are available from site specific data collection and testing. Two APTs performed at the site are documented in JLA Geosciences (2006) and Dames and Moore (1975) and serve to supplement the conceptual model presented in the existing ECFAS2 model documentation (Golder Associates, 2008).

The design of the original model was generally unchanged. However, the modeled domain was truncated in the north such that the longitudinal extent of the revised model is less than that of the original. Additionally, the finite difference grid spacing was modified to account for well locations used in the APTs that are simulated in the model re-calibration. Grid modifications are described in Section 3.2. Additionally, since relative changes in flow cond itions (i.e. drawdowns) are the focus of both model calibration and predictions, only the groundwater flow component of the original model is evaluated and employed, herein. Logistically, this decision faci litated efficient model calibration and predictive simulations, as consideration of density-dependent flow and transport resulted in very long run times. The original groundwater flow and transport model was calibrated to regional water levels and saltwater concentrations. To account for site-specific conditions, the model was re-calibrated to two APTs conducted at the site.

2 TETRA TECH

3 Regional Model Simulations 3.1 ECFAS2 Model The SFWMD, through contractors, developed a density-dependent groundwater flow and saltwater transp01t model of the East Coast of F lorida in two phases. The first phase, ECFAS 1 (HydroGeologic, 2006), simulated the southern half of the study area (the Lower East Coast of Florida); the second phase (ECFAS2) expanded the model domain northward to include more of the East Coast of Florida (Golder Associates, 2008). Both phases of the ECFAS model are available from the SFWMD; only the former has been peer-reviewed. Nevertheless, these model s represent the best available framework from which to base a permitting-level analysis of regional Floridan Aquifer impacts resulting from pumping.

The ECFAS2 model encompasses the ECFAS1 region and represents a revision to the earlier work. Consequently, the ECFAS2 model was used as the framework for this analysis. The ECFAS2 model covers the much of the East Coast of Florida, from southern Indian River County to the Florida Keys. This area is discretized into uniform 2400 by 2400 ft cells.

Vertically, the model extends from land surface to the Boulder Zone, a depth of approximately 3000 ft. The vertical section is discretized into 14 layers, with the Upper Floridan Aquifer represented as 2 layers. Boundary conditions are specified to represent flow into and out of the model domain, usually along the perimeter of the study area. Both flow (hydraulic heads) and saltwater transport (TDS concentrations) are simulated and are dependent upon one another (density-dependent flow and transp01t) . Field data from numerous borings were used to establish the structure of the model layering, which represents the hydrostratigraphic layers. In addition, field data from APTs were used to guide the initial choice of hydraulic parameters that were used in the model calibration. The model was calibrated to both hydraulic heads and concentrations.

Even though the model was calibrated, Golder Associates (2008) found that the model 's size resulted in exceptionally long run times such that the scope of the calibration had to be reduced from what was originally envisioned.

3.2 Adapted Floridan Model The ECFAS2 model was not usable in its avai lable state because it covers a very large area and does not provide the resolution required to accurately assess site-specific features and impacts.

Several structural modifications were made to the model and are described herein. Modifications to the calibration of the model are discussed in this section. As previously mentioned, only the groundwater flow capabilities of the ECFAS2 model were germane to the analyses of drawdown described herein, as regional changes in water quality attributable to the proposed wells, as well as the impact of such changes on drawdown, are anticipated to be negligible. Moreover, model run times were dramatically reduced by eliminating the density-dependence.

Since the Adapted Floridan model simulates groundwater flow and is adapted from the SEAWAT-based ECFAS2 model, the USGS simu lation software MODFLOW-2000 (Harbaugh, et al, 2000), a commonly applied groundwater flow model, was used to simulate the regional model. MODFLOW-2000 is capable of addressing the requirements of the SFWMD BOR inasmuch as it:

3 TETRA TECH

simulates groundwater flow,

is capable of addressing mu ltiple hydrostratigraphic layers and subdividing these layers such that drawdown can be computed at multiple levels within each layer, and

is in the public domain, peer-reviewed, and widely used .

The most significant structural change to the model was the grid spacing, which was originally set at 2400 ft. For calibration purposes, the grid was refined in the immediate vicinity of the Turkey Point APTs, such that the well spacing for the APTs could be accurately represented and changes in head over small distances resolved. The revised grid spacing in the model for the calibration is shown in Figure la. The minimum grid spacing used in the Adapted Floridan model, near pumping and monitoring wells, is as little as 1.5 ft. The original model grid spacing, shown in Figut*e lb, was used in subsequent predictive runs because it was adequate for assessment of impacts at the desired scale and was practical from a run-time perspective.

The original model layering was retained because it appeared to be generally appropriate for the level of detail required. The Intermediate Confining Unit (ICU), which overlies the Upper Floridan, was represented using a single layer.

The additional pumping wells that were included as a part of the calibration of the Adapted Floridan model also represent modifications to ECF AS2. The well locations and rates are described in the calibration and model results sections below. The time stepping of the models was also modified to provide adequate resolution for the duration of the APTs and to account for intermittent pumping (Section 3.2.1.1 and 3.2.1.2).

3.2.1 Additional calibration of model Although the ECFAS2 model may represent the regional conditions fai rly well, it may not represent site-specific conditions particularly well. This hypothesis was tested by running the model using documented pumping stresses on the system and comparing the modeled response to that which was observed during the test. In general, as discussed below, the compari son was not good. In order to obtain a reasonable representation of site-specific conditions, two additional calibrations, one to a short-term APT and another to a longer term APT, were performed. The ability to match aquifer system response to these APTs prov ides confidence that the model can predict the response to future proposed pumping. Modeled water levels were checked to ensure that the match to regional calibration targets had not been degraded as a result of the local changes. The methodology and resu lts of each of the additional calibrations are described below.

3.2.1 .1 JLA APT JLA Geosciences (2006) conducted an APT in support of the Unit 5 site certification. Floridan water supply well PW-1 was pumped for 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months and drawdown was measured in two other water supply wells and a shallow observation well. The drawdown response documented during this test was believed to represent a good series of targets to match as a part of a calibration because it was local to the area of proposed pumping and was conducted under quality-controlled conditions. However, it was recognized that the short duration of the test and extent of monitoring points would provide data that may only be representative of a relatively small area.

4 TETRA TECH

Simulation of the APT was accomplished using the revised model grid. Well PW -1 was represented with a single well pumping at a rate of 4500 gpm in model layers 3 and 4, which represent the Upper Floridan Aquifer, in the cell at row 166, column 143 Timestepping ranged from a minimum of 5 to a maximum of 567 minutes, Drawdown response was noted in wells PW-3 (layers 3 and 4, row 168, column 171), PW-4 (layers 3 and 4, row 180, column 157), and OBS-1 (layer 2, row 166, column 143) at distances of 3036, 1686, and 0 feet, respectively from the pumped well. Note that OBS-1 is co-located with the pumping we ll , but is screened near the base of the Biscayne Aquifer and did not experience drawdown in response to the APT.

Comparison of modeled to observed conditions for the original model, prior to adjustment, was not good, with a residual standard deviation of greater than 100 ft. However, as shown in F igure 2, this match improved cons iderably (residual standard deviation of 0.36 ft) after adjustment of hydraulic parameters as a pmt of the calibration. In general, hydraulic conductivities were increased from their original values during calibration. Goodness-of-fit ca libration metrics are shown in Table 1 and indicate that the model provides a reasonable fit to observed data.

Table 1. Goodness of fit metrics for the JLA APT calibration Metric Numerical Value Mean Error, ft 0.22 Mean Absolute Error, ft 0.33 Residual Standard Deviation, ft 0.36 Range of Targ_ets , ft 6.36 Residual Standard Deviation I Range *100 5.6%

Note that this calibration was conducted iteratively with the Dames and Moore APT described below and hence the calibrations strike a balance between matching the results of both APTs with the same set of parameters.

3.2.1.2 Dames and Moore APT Dames and Moore (1975) conducted an APT in support of a feasibility study for using Floridan Aquifer water to cool the original Turkey Point nuclear units. Floridan Aquifer production test well (PTW) was pumped for 90 days and drawdown was measured in eight monitoring wells at various distances from the pumped well and depths in the aquifer. The drawdown response documented during this test was believed to represent a good series of targets to match as a part of a calibrati on because of its long duration and use of monitoring points that were distant from the pumping well. Thus, this test was complementary to the shorter duration, more local JLA APT described above.

As in the simulation of the JLA APT, the simulation of the Dames and Moore APT was accomplished using the refined model grid. Well PTW was represented with a single well pumping at a rate of 5000 gpm in cell layers 3 and 4, row 220, and column 97. Timestepping ranged from a minimum of 73 minutes to a maximum of 11.8 days. Drawdown response was noted in wells OW-A (row 229, column 108), OW-B (row 238, column 120), OW-C (row 207,column 82), and OW-D (row 258, co lumn 181) at distances of 100 feet, 500 feet, 2000 feet, and 48,000 feet, respectively from the pumped well. Drawdown was recorded in the Upper and Middle Floridan aquifers at each of the four observation well sites, which are represented by layers 3 and 4, and 7 and 8, respectively 5

TETRA TECH

Comparison of modeled to observed conditions for the original model, prio r to adjustment, was not good (residual standard deviation in excess of I 0 ft), as was the case for the JLA APT. As shown in Figure 3, this match also improved considerably (with a residual standard deviation of 0.77 ft) after adjustment of hydraulic parameters as a part of the calibration. Goodness-of-fit calibration metrics are shown in Table 2 and indicate that the model provides a reasonable fit to observed data.

Table 2. Goodness of fit mctrics fo1* the Dames and Moore APT calibration Metric Numerical Value Mean Error, ft -0.30 Mean Absolute Error, ft 0.66 Residual Standard Deviation, ft 0.77 Range of Targets, ft 11.8 Residual Standard Deviation I Range

100 6.5%

Though the wells shown in Figure 3 are not an exhaustive representation of the calibration targets, they are a microcosm of the quality of the model match to this APT. The lateral and vertical proximity to the pumping well precluded a reasonable match to the observed drawdown at well OW-A (Upper); as such, this well was omitted from the calibration.

3.2.1.3 Adjustments to the calibration The primary parameters that were changed as a result of the additional calibration were hydraulic conductivities of the Upper Floridan Aquifer (UFA), ICU, Middle Confining Unit (MCU), and the Middle Floridan Aquifer (MFA). These parameters were all raised from their orig inal values, as shown in Table 3.

Note that the parameter changes were made within zones that were near the Turkey Point site and mostly in areas potentially affected by drawdown from proposed salinity reduction wells, as shown in Figure 4, 5, and 6.

The changes made to the hydra ulic properties in the Adapted Floridan model are not expected to significantly impact the quality of the model match to the water level and water quality targets employed in the calibration of the ECF AS2 model. The changes made to the Adapted Floridan model were generally minor, and the preponderance of the ECFAS2 model calibration targets are located outside of the Adapted Floridan model doma in. Water levels at well ENP-100, located about 17 miles to west of Turkey Point, were simulated to be approximately 36ft, which slightl y underestimates observed water levels (approximately 37.5 ft NGVD29). The ECFAS2 model si mulates a water level of 39 feet at this location, after 350 days of simulation. Simulated water levels at W ASA-South, located to the n011h of Turkey Point, are approximately 35 ft NGVD 29; this well, however, was omitted from the ECFAS2 modeling analysis due to anomalous water levels (Golder Associates, 2008).

6 TETRA TECH

Tablc3 P arameter c 1anges resu I tmg

. f rom ca J'b 1 ration o f t I1c Ad apted Flone

'I an mo d e.I ECFAS2 model FPL Floridan Model Hydrologic Model Aquifer (original) (recalibrated)

Unit Layers Parameter magnitude magnitude 0.0006 0.001 Kz (fUd)

ICU 2 0.000075 0.001 Kh (fUd) 0.000075 0.001 5.2 100 Kz (fUd) 9 15 72.5 225 0.33 225 UFA 3,4 52 100 Kh (ft/d) 90 150 725 330 3.33 330 Ss 5.25E-07 8.00E-07 0.004 0.003 0.000002 0.08 Kz (fUd) 0.4 0.003 MCU 5,6 0.002 0.08 0.02 0.08 Kh (fUd) 0.00001 0.4 0.03 Kz (fUd) 5.2 30 450 900 MFA 7,8 300 600 Kh (fUd) 180 1200 52 600 0.0015 0.01 MC2 9,10,11 Kz (ft/d) 0.0002 0.02 3.3 Predictive Simulations Once calibration of the regional Floridan model was confirmed, equilibrium flow conditions were established by running the model, holding all flow boundaries (e.g. specified heads) constant until changes in the simulated flow field in the Florida n Aquifer System were negligible. No pumping was simulated in this equi librium model. The resulting equilibrated state formed the initial conditions for ensuing predictive simulations. Equilibrated regional water 7

TETRA TECH

levels, especially near Turkey Point, were generally lower than observed water levels; this is due to the exclusion of salt transport and the associated density-dependent flow. However, given that purpose of th is model is to provide estimates of relative changes in water level , the low simulated water levels were deemed irrelevant. Since the focus of the salinity reduction well evaluation is regional drawdown, the original 2400-ft grid spacing was employed for predictive simulations.

According the SFWMD BOR, predictive evaluations made with the calibrated model must be conducted using monthly stress periods that simulate average annual groundwater withdrawals subject to rainfall that alternates between average and 1-in-1 0 year draught conditions (3 months of average cond itions, followed by 12 months of drought conditions, followed by 6 months of average conditions). As such, the predictive models were conducted using 21 monthly stress periods .. Due to the specified head boundary in the topmost layer of the model, climatic stressess were not varied between average and drought conditions. Moreover, it is not anticipated that variation in rainfall would impact the Floridan aquifer over the 21-month timeframe of the simulation.

Additionally, the BOR stipulates that the 1-ft drawdown contour associated with the proposed pumping be simulated and the impacts to existing legal users' wells within that contour be evaluated. The process by which this was accomplished is described below.

3.3.1 Proposed Salinity Reduction Well Operation There are six proposed salinity reduction wells. At any one time, five of these well s will collectively pump 14 MOD of low salinity water from the Upper Floridan aquifer. The six wells will be spaced approximately 1900 ft apart, along the northernmost canal of the Cooling Canal System and along the Interceptor Ditch (Figure 7). In the model , the 14 MGD of pumping is distributed evenly amongst the five active wells and is assumed to be a constant rate of pumping over the course of the 25-year simulation. Two alternative pumping scenarios are considered in this modeling analysis and differ in the aJJocation of pumping to wells F-2 and F-6. The base scenario simulates pumping at wells F-1 through F-5 (no pumping at F-6); the alternative scenario simulates pumping at wells F-1 and F-3 through F-6 (no pumping at F-2).

The salinity reduction wells were simulated, starting from an equilibrium flow field. Aside from these salinity reduction wells, no pumping was simulated in this model such that the permitted users w ithin the 1-ft drawdown contour could be delineated. At the conclusion of the 21-month simu lation, the simulated drawdowns in the regional model are those attributable only to the five proposed salinity reduction wells. Figure 8a illustrates these regional drawdowns associated with the base pumping scenario. In this base simulation, the drawdowns at a distance from the site are affected by variations in hydraulic conductivity; this is evident upon inspection of the 1-ft drawdown contour, wh ich generally has an oblong shape, whose major axis is oriented north-to-south . Nearer to the site, the drawdown contours radiate outward from the wells in a more uniform manner, where the maximum drawdown is approximately 15.1 ft, near well F-3. In the alternative scenario, the maximum drawdown is approximately 14.4 ft, near well F-5.

As previously mentioned, the SFWMD BOR dictates tbat drawdown at permitted users' wells encircled by the 1-ft drawdown contour be determined. As illustrated in Figure 8a, the following permitted users fall within the 1-ft drawdown contour:

8 TETRA TECH

Card Sound Golf Club,

Ocean Reef Club,

the F loridan Keys Aqueduct Authority (FKAA),

Miami-Dade Water and Sewage Department South Miami Heights Wellfield, and

FPL Unit 5 Wells.