Loading...
HomeMy WebLinkAboutNC0025348_Report_20180514/V(Mp2 -50 New Bern Wastewater Reclamation Facility Predictive Modeling Analysis; Martin Marietta Quarry (East Lake) Water Reclamation Facility, City of New Bern, Craven County, North Carolina Prepared For: Rivers and Associates, Inc. 107 East 2ntl Street Greenville, NC 27858 Prepared By Groundwater Management Associates, Inc. 4300 Sapphire Court, Suite 100 Greenville, North Carolina 27834 And -a Groundwater Management Associates, Inc. 2205-A Candun Drive Apex, North Carolina 27523 GMA Project #21342 May 14, 2018 ` N C PER _, TABLE OF CONTENTS EXECUTIVESUMMARY.......................................................................................... ES-1 1.0 INTRODUCTION.................................................................................................1 2.0 SCOPE OF WORK...............................................................................................1 3.0 REGIONAL HYDROGEOLOGIC SETTING .............................................................2 3.1 Regional Geology.............................................................................................. 2 3.2 Hydrogeology of the Site.................................................................................. 5 4.0 REVIEW OF PREVIOUS STUDIES.......................................................................6 ._ 4.1 The Pohlig Model............................................................................................6 4.2 Matyiko Master's Thesis.................................................................................6 4.3 Additional Data Sources.................................................................................7 a.. 5.0 PREDICTIVE MODELING....................................................................................7 5.1 Grid and Layer Design....................................................................................9 5.2 Model Boundaries........................................................................................10 5.3 Model Input.................................................................................................11 5.4 Steady -State Calibration..............................................................................12 5.5 Sensitivity Analysis......................................................................................13 5.6 Predicted Groundwater Flow Patterns.........................................................13 5.7 MT3DMS Transport Modeling.......................................................................13 5.7.1 M73DMS Settings.................................................................................14 OEM 5.7.2 Nitrogen Loading Scenarios......................................................................15 5.7.3 M73DMS Predictions................................................................................15 5.8 Model Limitations.........................................................................................18 NOW 6.0 CONCLUSIONS AND RECOMMENDATIONS......................................................18 7.0 REPORT CERTIFICATION .................................................................................19 8.0 LIST OF REFERENCES......................................................................................21 MIND FIGURE Figure 1: New Bern Wastewater Reclamation Facility 0000 Figure 2: Model Grid Design Figure 3: Cross Section through East Lake showing Vertical Distribution of Model Layers Figure 4: Recharge FUN Figure 5: Layer 1 Boundaries Figure 6: Layer 2 Boundaries Figure 7: Layer 3 Boundaries FM Figure 8: Layer 5 Boundaries Figure 9: Modeled Horizontal Hydraulic Conductivity Distributions Figure 10: Observation Well Locations RON Figure 11: Model Calibration Plots Figure 12: Modeled Potentiometric Surface for the Surficial Aquifer Figure 13: Modeled Potentiometric Surface for the River Bend Aquifer fRIN Figure 14: Modeled Potentiometric Surface for the Upper Spring Garden Aquifer Page i Figure 15: Modeled Potentiometric Surface for the Lower Spring Garden Semi -Confiner Figure 16: Modeled Potentiometric Surface for the Lower Spring Garden Aquifer Figure 17: Modeled Groundwater Flow within a Vertical Section through the East Lake Figure 18: Maximum Predicted Total Nitrogen (TN) Concentrations above Background During 2016 Figure 19: Chronological Total Nitrogen Levels in Monitoring Wells Surrounding the East Lake Figure 20: Predicted Total Nitrogen (TN) Concentrations above Background after 15 Years of Daily Discharge of 5 MGD of Effluent containing 5.0 mg/L of TN to the East Lake APPENDICES .. Appendix I: 2016 East Lake Usage Records me am am MR am an am am fm Page ii Predictive Modeling Analysis, East Lake Wastewater Reclamation Facility am EXECUTIVE SUMMARY '^ GMA has developed a three-dimensional, six -layer groundwater flow and transport model for the East Lake Wastewater Reclamation Facility operated by the City of New Bern. The groundwater flow model was constructed using the Groundwater Modeling System interface (GMS 10.2.3) with the United States Geological Survey (USGS) groundwater model, MODFLOW- 2000. MODFLOW is a modular, three-dimensional groundwater -flow model code that simulates groundwater flow using a finite -difference method applied to a block -centered rectangular grid. Model assumptions were based upon a wealth of regional and local hydrogeologic data available to GMA. Model input assumptions and conceptual model design elements were derived predominantly from the following sources: GMA's prior work on the City of New Bern Castle Hayne Aquifer Welifield, data provided by Martin Marietta Aggregates from the Clarks Quarry, design data provided by Rivers and Associates, Inc. for the East Lake Wastewater Reclamation facility, drilling and testing data collected by GMA from the Mackilwean Turf Farm, published data from an ECU master's thesis prepared by Sara Matyiko, results of prior regional modeling performed by Ken Pohlig of the NCDENR, and 10 years of groundwater and effluent monitoring data provided by the City of New Bern. GMA calibrated the groundwater model to simulate steady-state flow based on average elevations of groundwater and lake levels reported for 2016. Hydraulic conductivities, recharge, and river conductance values were calibrated using a "trial -and -error" approach to test the overall soundness of the conceptual model and to ensure that calibrated values represented real -world ranges. The model calibration has an acceptably high correlation coefficient (r2=0.92), and the root mean square of residuals was low (RMSR=0.79 ft). The groundwater - flow model provides spatial representations of head pressure variations and groundwater flow for the Castle Hayne Aquifer System and the Surficial aquifer for an approximately 44 square mile (mil) area. After model calibration, GMA used MT3DMS (Modular Transport, 3- Dimensional, Multi -Species model) in conjunction with MODFLOW to predict nitrogen concentrations moving from East Lake into and through the groundwater system. Results from the MODFLOW groundwater -flow model and MT3DMS simulations indicate the Im following: • Groundwater pumping associated with the nearby Martin Marietta Clarks Quarry and the City of New Bern production wells were determined to not have an effect on the groundwater flow from the East Lake Wastewater Reclamation Facility. The East Lake is functioning as a recharge boundary that drives groundwater radially and downward under the induced, artificial head created by the wastewater facility. • The induced head beneath the East Lake results in a groundwater divide in the upper layers of the Castle Hayne Aquifer beneath the lake. Groundwater recharge on the east side of that divide discharges to the Neuse River, and groundwater recharge west of the divide discharges to the West Lake. ,R ES-1 IF_ Predictive Modeling Analysis, East Lake Wastewater Reclamation Facility OR • The majority of groundwater flow out of East Lake is concentrated at the sides of the basin where hydraulic gradients are the steepest and where aquifer permeability is the highest. An estimated 62% of the groundwater recharge exiting the East Lake ultimately discharges to the Neuse River. The remaining 38% discharges to the West Lake, where it then flows as surface water back to the Neuse River, or it is lost as evapotranspiration. Travel times for the majority of groundwater flow from the recharge at East Lake to the areas of discharge to the Neuse River or to the West Lake range from 66 days to 5 years. • Model results indicate that approximately 4% of the total groundwater recharge entering from the East Lake is directed downward through deeper, less permeable, layers of the Castle Hayne Aquifer System before eventually discharging at the Neuse River or the West Lake. Travel times for this deeper portion of the groundwater flow from the East Lake are on the order of 100+ years. • Historical water -quality monitoring of wells surrounding the East Lake facility do not indicate an increase in total nitrogen (TN) concentrations above the background concentrations that were measured prior to the facility's first use. Further, base -level �. (i.e., preexisting) concentrations in some monitoring wells (e.g. EL1 and EL3) prior to use of the facility exceed the measured TN concentrations in the East Lake after the start of facility use. The field monitoring data indicate, on average, that the addition of wastewater into the East Lake is not resulting in a recognizable increase in TN concentrations in the groundwater monitoring wells surrounding the facility. • During 2016, effluent was discharged to the East Lake reclamation facility during the months of May and November (with a small amount discharged during October). Consistent with historical water -quality monitoring, MT3DMS groundwater transport simulations under this usage scenario predict no measurable increase over base level TN concentrations at the groundwater monitoring well locations. These MT3DMS simulations were performed without applying a retardation factor, a very conservative assumption. The results suggest that, under present usage patterns, TN added to the East Lake is being rapidly diluted in the groundwater system to concentrations that cannot be distinguished from the naturally occurring background concentrations. • Per request of the City of New Bern and Rivers and Associates, Inc., GMA ran additional MT3DMS simulations that modeled effluent discharge to the East Lake at a rate of 5.0 million gallons per day (MGD) and a concentration of 5.0 mg/L of TN for four equally FM spaced months per year. This scenario represents the realistic maximum time period that the City would discharge to the East Lake. MT3DMS simulations predicted that, after discharging under this scenario for 15 years, TN loading to the Neuse River would be 8,963 pounds of TN per year, or 35% of the total input. • As an additional run, GMA also modeled effluent discharge to the East Lake at a rate of 5.0 MGD and a concentration of 5.0 mg/L of TN every day for 15 years. Under this usage scenario, after 15 years, the model predicted an annual TN load to the Neuse River of 29,558 pounds per year, or 39% of the TN input • Because the MT3DMS runs assume no retardation factor for TN movement through the groundwater, all reductions in concentration in these simulations are due to dilution only. Previous research indicates that denitrification represents a sink of approximately 5% of the nitrate -nitrogen load to the Neuse River and its tributaries. Applying this PM ES-2 Predictive Modeling Analysis, East Lake Wastewater Reclamation Facility research to the model results, GMA predicts an annual nitrogen load of 8,515 pounds per year of TN to the Neuse River under the maximum potential usage of the — reclamation facility (quarterly wastewater inputs). Because this load is only 30% of the TN added to the system, the load represents a potential TN credit of 70%. Available field data show close agreement with the MODFLOW 2000 groundwater -flow model simulations and the MT3DMS results. GMA asserts that the groundwater -flow model that we have produced is a reasonable representation of the groundwater conditions associated with the City of New Bern Wastewater Reclamation Facility. However, groundwater models must incorporate simplifying assumptions about the nature of groundwater flow, and as a result FM some local heterogeneities of the groundwater system may not be accurately represented. For example, this model assumes that groundwater flow through the aquifer beneath the lake occurs solely as homogeneous flow and does not reflect preferential flow paths (i.e., heterogeneities) that may exist within the limestone. Although there is some degree of uncertainty in the model predictions, this model reasonably simulates groundwater flow and advective transport of total nitrogen added to the East Lake of the New Bern Wastewater Reclamation facility based on the best available data. RM FM am FM MM OM MR ,M ES-3 ma 1.0 INTRODUCTION — The City of New Bern (the City) operates a wastewater reclamation facility at the location of the former Martin Marietta Glenburnie limestone quarry. The quarry ceased operations in 1996. The site, now known as the East Lake Reclamation Facility, is located along the south bank of the Neuse River on the north side of New Bern in Craven County, NC (Figure 1). The site includes a 145-acre abandoned quarry lake, called the East Lake, which receives treated wastewater effluent. An eleven foot high berm constructed around the lake allows for treated wastewater to be pumped to the lake where it then infiltrates into the subsurface. The wastewater moves away from the quarry lake, through the groundwater system where it is diluted, and eventually discharges into the Neuse River. In 2004, the North Carolina Division of Water Quality (NCDWQ) granted the City of New Bern a non -discharge permit in addition to their existing NPDES permit for discharging into the Neuse River. The non -discharge permit allowed the city to discharge tertiary treated wastewater into the quarry disposal system, and/or water could be irrigated onto the nearby Mackilwean turf farm. The quarry portion of the permit was 'performance limited' rather than'flowrate limited', and required the City to monitor the wastewater effluent water quality, the East Lake water quality and water level, and the water quality within a network of compliance monitoring wells surrounding the facility. RM The 2004 non -discharge permit did not provide the City nitrogen credit for wastewater diverted to the East Lake. In accordance with the City of New Bern's non -discharge permit, to apply for FM nitrogen credit, the City is to provide estimates of subsurface total nitrogen loading to the Neuse River that may result from flow out of the quarry lake. To accomplish this, Groundwater Management Associates (GMA) has developed a groundwater -flow model to simulate the 'm directions, rates, and concentrations of total nitrogen (TN) moving through the groundwater system from the quarry lake to a point of discharge. fm 2.0 SCOPE OF WORK Rivers and Associates, Inc. contracted GMA to evaluate the nitrogen loading of the Neuse River from the New Bern wastewater reclamation facility. GMA accomplished this work in three stages: • Stage 1 Site Characterization and Records Review — GMA reviewed available records to establish a baseline understanding of the operation of the facility, the hydrogeologic data, and prior modeling. Specific records reviewed included operational and monitoring records of effluent disposal to the quarry lake, prior hydrogeologic studies, including a thesis and a previous mounding modeled prepared by GMA, and prior nitrogen loading evaluations developed in 2000 by Ken Pohlig of the NC FM Page 1 nm a.o Department of Environment and Natural Resources (now the NC Department of Environmental Quality). Stage 2 Groundwater flow modeling — GMA developed a 3-dimensional finite - difference groundwater flow model (USGS MODFLOW-2000) using Groundwater Modeling System (GMS) software. Model properties were derived from prior hydrogeologic studies and information gathered during the records review process. The groundwater model was calibrated to simulate steady-state flow based on average elevations of groundwater and lake levels reported for 2016. • Stage 3 Groundwater transport modeling — Once the model was calibrated to local water -level data from the existing monitoring well network and the quarry lake levels, GMA used the MT3DMS module of MODFLOW to predict how quarry use would affect nitrogen concentrations in the adjacent groundwater system. FM This report provides modeling predictions of groundwater flow paths and TN loading estimates to the Neuse River resulting from the addition of tertiary treated wastewater additions to the fm East Quarry Lake reclamation facility. 'R 3.0 REGIONAL HYDROGEOLOGIC SETRNG IM The facility (Figure 1) lies within the Coastal Plain Physiographic Province of North Carolina (NCGS, 1985). North Carolina's Coastal Plain is a broad, relatively flat physiographic province separating the hilly Piedmont province from the Atlantic Ocean. Local elevation ranges from FM about 35 feet above sea level (ASL) in upland areas west of the facility to almost sea level at the shore of the Neuse River. Land surface topography in Craven County is primarily a product of Neogene and Quaternary fluctuations in sea level that repeatedly inundated and exposed the land over the past 23 million years (Horton and Zullo, 1991). These sea -level cycles sculpted the land surface into broad, relatively flat marine terraces bounded by low escarpments that represent former shorelines. Streams and rivers have incised these terraces to create the current topographic character of the area. rRa The regional hydrogeology of the area has been described in published reports by GMA (2014), Fine (2008), Lautier (2001), Giese and others (1997), Winner and Coble (1996), and Lyke and Winner (1990). For the purposes of this investigation, relevant aquifer and confining unit fm descriptions presented in GMA (2014) have been adopted, and these are presented below. MW 3.1 Regional Geology The Coastal Plain Province is underlain by marine, estuarine, and terrestrial sediments that were deposited along the continental margin over the past 200 million years. The East Lake Reclamation Facility is underlain by approximately 1,700 feet of Cretaceous to Recent aged sediments and sedimentary rocks that were deposited on top of pre -Mesozoic aged (>250 MW ,e Page 2 M million years) basement rocks. Basement rocks in the vicinity of the facility site are mapped as granite of the Carteret Batholith (Lawrence and Hoffman, 1993). The Mesozoic -aged sediments beneath the property are dominantly clastic in nature and include sequences of silt and clay interbedded with sand and gravel zones and minor amounts of shell. These sediments are associated with deltaic and marginal marine depositional environments that predominated along the eastern margin of North America from about 145 to 66 million years ago. The Mesozoic sediments have been hydrostratigraphically subdivided into four principal aquifers of the Cretaceous Aquifer System (CAS). The CAS includes (from deep to shallow) the Lower Cape Fear Aquifer, the Upper Cape Fear Aquifer, the Black Creek Aquifer, and the Peedee Aquifer. FAR Overlying the CAS is a sequence of Cenozoic -aged (<66 million years) sediments of dominantly marine origin. These include significant beds of sands, shelly clays, and fossiliferous sandy limestones. These sediments have been stratigraphically subdivided into six mappable units, including (from deep to shallow): the Paleocene Beaufort Formation, two members of the Eocene Castle Hayne Formation, the Oligocene River Bend Formation, the Pliocene Duplin Formation, and unnamed Pleistocene to recent marine terrace deposits (GMA, 2014). The basal unit of the Castle Hayne Formation in Craven County is the Comfort Member. The Comfort Member is a hard sandy skeletal limestone and is the most widespread of the Castle Hayne units. The Spring Garden Member of the Eocene Castle Hayne Formation overlies the Comfort within the study area. The Spring Garden Member is a thinly bedded, very sandy, FAR shelly limestone to calcareous unconsolidated sand deposit that is interpreted to have been deposited in a low -energy, shallow marine environment. The upper surface of the Spring Garden unit is very uneven and is often thinly coated by a veneer of phosphate (Ward et al. `m 1978). f M9 The Oligocene River Bend Formation overlies the Castle Hayne Formation at the New Bern Quarry. The unit dips to the east and south in the New Bern area and is about 10 feet thick at the location of the New Bern Quarry (Ward et al. 1978). The River Bend Formation is composed of limey sand, sandy shell, barnacle hash, and sandy limestone, with the most prominent lithology being a pelecypod moldic sandy limestone. It is interpreted to be an inner to mid -shelf deposit that formed below wave base. Near the base of the formation, the River Bend typically contains little to no quartz sand and predominantly consists of barnacle plates and molluscan molds in calcarenite matrix (Ward et al. 1978). pan Together, the Comfort and Spring Garden Members of the Castle Hayne Formation, and the River Bend Formation comprise the Castle Hayne Aquifer System (CHAS) in Craven County. Sm Remnants of the Pliocene -age Yorktown Formation occur as laterally discontinuous deposits of fossiliferous sands and clays that often occur as paleo-channel fill material. Shallower deposits RM are primarily Pleistocene in age and are composed of sand, silt, shelly debris, and some clay. ,e Page 3 rm M" MR MR am MIN FMI FRO MR Table 1 presents a list of the principal aquifers that occur in Craven County and the associated ages and formations that comprise these aquifers. Table 1. Regional Aquifer Framework for Craven County (Adapted from GMA, 2014) Hydrostratigraphic Period Ma Geologic Age Formation Name Units cRecent Unnamed Holocene Surfiicial Aquifer v o� 0 - 2.5 to Pleistocene James City Fm. Upper Castle Hayne CL a� 2.5 - 5 Pliocene Yorktown/Duplin Fm. Yorktown Aquifer a Upper Castle Hayne CL z 23 - 34 Oligocene River Bend Fm. Upper Castle Hayne Aquifer Upper Castle Hayne Castle Hayne Fm. Aquifer (Spring Garden Mbr.) Middle Castle Hayne CL 34 - 56 Eocene Middle Castle Hayne 0 A uifer q a� a Lower Castle Hayne CL (Comfort Member) Lower Castle Hayne Aquifer Beaufort Fm. Beaufort CL 56 - 66 Paleocene Beaufort Aquifer Yaupon Beach Fm. Peedee CL 66 - 72 Upper Cretaceous Peedee Fm. Peedee CL (Maastrichtian) Peedee Aquifer 72 — 84 Upper Cretaceous Black Creek Group Black Creek CL (Campanian) Black Creek Aquifer W 84 — 86 Upper Cretaceous Middendorf Fm. Black Creek Aquifer 401 a� . (Santonian) (Pleasant Creek Fm.) Upper Cape Fear CL 86 — 90 Upper Cretaceous Upper Cape Fear CL (Coniacian to Cape Fear Fm. Upper Cape Fear Aquifer 90 — 94 possibly Turonian) Lower Cape Fear CL Lower Cape Fear Aquifer >252 Paleozoic and Crystalline Basement Bedrock Aquifer Late Proterozoic Ma — Million Years (mega-annum) CL — Confining Layer Fm. — Formation Mbr. — Member ,�, Page 4 no 3.2 Hydrogeology of the Site The City of New Bern Reclamation Facility is located within the Neuse River drainage basin on no the north side of the City of New Bern in Craven County, NC. The Reclamation Facility lies immediately adjacent to the Neuse River, with the East Lake being located as close as 120 feet F, from the southern river bank (Figure 1). The Neuse River flows from the northwest to the southeast. Major bodies of water that discharge to the Neuse River within the model area include the Trent River to the south and Bachelor Creek and one of its tributaries, Caswell no Branch, to the north. As presented in Table 1, the CHAS in Craven County is composed of three aquifer units: the Lower Castle Hayne (Comfort), the Middle Castle Hayne (Spring Garden), and the Upper Castle Hayne (River Bend). The middle and upper portions of the CHAS are the primary aquifers of interest for this investigation. All of the units evaluated in this report generally thicken from west to east. The Lower Castle Hayne confining unit is used as the lower limit of our investigation. This confining unit, consisting predominantly of silt and clay, lies beneath the Lower Spring Garden member and impedes the vertical flow of groundwater. Aquifers and confining units deeper (older) than the Lower Castle Hayne confining unit were not included in this modeling investigation because these deeper units are unlikely to be affected by the addition of wastewater to the East Lake. Me During pre -mining conditions, the groundwater flow direction at the facility location was to the east-northeast towards the Neuse River. When the quarry was active, mine dewatering lowered head beneath the quarry and facilitated movement of salty groundwater from the Neuse River westward towards the quarry. Under present usage conditions, the addition of effluent to the East Lake produces a local high water level that causes the downward movement of groundwater and thus serves as a groundwater recharge area. At the subject site, the Surficial aquifer and the Upper Castle Hayne Aquifer are unconfined, and the Middle Castle Hayne Aquifer is semi -confined. Major withdrawals in the study area are the Martin Marietta Clarks Quarry, located approximately three miles west of the East Lake Reclamation Facility, and the City of New Bern Castle Hayne Aquifer wellfield (Figure 1). The Clarks Quarry is an active limestone mine that is permitted to withdraw 16 million gallons per day (MGD) from the Surficial and Castle Hayne FOM aquifers as part of the mine's dewatering activities. Since 2002, average daily withdrawals from the mine have ranged from 6.7 to 11.9 MGD. New Bern's Castle Hayne wellfield consists of fifteen production wells scattered west of the site. The wellfield is permitted to withdrawal a maximum of 5.5 MGD from the Castle Hayne aquifer. The New Bern wells are screened in the lower portions of the Spring Garden member, and withdrawals are from the Middle Castle Hayne aquifer unit. Maximum withdrawals from individual wells within the New Bern Castle Hayne wellfield may range from 0.252 MDG to 0.720 MGD. em �, Page 5 4.0 REVIEW OF PREVIOUS STUDIES 4.1 The Pohlig Model GMA reviewed a previous groundwater flow (MODFLOW) and transport model (MT3D) of the facility developed by Ken Pohlig (Pohlig, 2000) of the NCDWQ. Pohlig constructed a five layer regional model. The model simulated a wastewater flow to the quarry of 6.5 MGD with a total nitrogen concentration of 4 mg/L. Nitrogen removal or uptake processes were not considered. Pohlig's model predicted that the addition of treated wastewater to the reclamation facility would cause a small flux of groundwater to migrate towards the Clarks Quarry with the rest discharging to a 2-mile stretch of the Neuse River via the surficial and Castle Hayne aquifers. Importantly, however, Pohlig modeled the entire reclamation facility property, a 550-acre footprint, as the quarry lake receiving wastewater input. This may have originally been the plan for the facility; however, an 11-foot high berm was constructed only around the 145-acre East Lake, and only the East Lake receives wastewater. Thus, the Pohlig model greatly overestimates the area of the Neuse River receiving wastewater discharge. Pohlig also chose to model the nitrogen inputs from the reclamation facility operation to the groundwater system as a constant concentration of total nitrogen in recharge assigned to model cells within the quarry property footprint. This approach does not adequately simulate the dilution of the wastewater effluent as it discharges to the East Lake. Surface water in the East Lake is regularly diluted by precipitation, and it is this diluted concentration of total nitrogen that actually enters the groundwater system beneath and around the East Lake. For example, in 2016, the City discharged wastewater to the lake with total nitrogen concentrations as high as 8.27 mg/L., but measured total nitrogen levels in the East Lake during 2016 were less than 1.15 mg/L at all monitoring locations. Thus, GMA believes that Pohlig's estimates of nitrogen loading to the Neuse River are erroneously high. 4.2 Matyiko Master's Thesis two GMA also reviewed published data from an East Carolina University Master's thesis prepared by Sara Matyiko, which described the hydrogeology of the CHAS at the former Glenburnie Quarry. elm Matyiko also attempted to predict the hydraulic impacts of discharging tertiary treated wastewater to the East Lake Reclamation Facility (Matyiko, 2001). Through drilling observations and analysis of geophysical logs, Matyiko divided the site into surficial sediments (0 - 8 feet below the land surface), an 'upper zone' consisting of alternating hard and soft gray moldic limestone beds (from 8 — 66 feet below the land surface), a softer `middle zone' composed of calcareous sandstone (from 66 — 83 feet depth), and a 'lower zone' consisting of alternating sand, clay, and limestone beds (from 83 —162+ feet depth). Matyiko estimated hydraulic conductivity, transmissivity, and storativity at the site using data collected during 24- hour constant -rate aquifer pumping tests at wells around the facility. Testing indicated that the transmissivity of the "upper zone', which corresponds to the Riverbend Formation, was as much as two orders of magnitude greater than the lower units. Testing also confirmed that the CHAS ,R Page 6 MR rs was, in fact, hydraulically connected with both the Glenburnie Quarry lakes and the Neuse River. Contrary to Ken Pohlig's model, Matyiko also documented a groundwater divide between the Clarks Quarry and the reclamation facility. 4.3 Additional Data Sources In addition to the data collected from the Pohlig model and the Matyiko thesis, GMA used aquifer test data and drilling records from our previous work in the area to serve as the basis of constructing the model. Geophysical logs and aquifer testing data from the New Bern Castle Hayne Aquifer wellfield served as a major source for top of aquifer elevations and hydraulic conductivity values within the study area. Drilling and testing data collected by GMA from the Mackilwean Turf Farm, just north of the reclamation facility, provided the major source for Surficial aquifer parameters (GMA, 2006). 5.0 PREDICTIVE MODELING MR GMA developed a three-dimensional groundwater MODFLOW model using Groundwater Modeling System (GMS) software to simulate groundwater flow in the vicinity of the East Lake. MR GMS is a graphical user interface that facilitates model design and parameter input for programming MODFLOW, a modular, 3-dimensional, finite -difference, groundwater -flow model code developed by the United States Geological Survey (USGS) (Harbaugh and others, 2000). O, GMA's model predicts hydraulic head in the Surficial, Riverbend, and Spring Garden aquifer units under steady state conditions based on regional water -level information from 2016. The purpose of this report section is to describe the development, calibration, and application of the model to support non -discharge permit renewal for the City of New Bern. Specifically, GMA performed model simulations to: • Calibrate to baseline groundwater -flow conditions in the vicinity of the City of New Bern fim Wastewater Reclamation Facility, • Simulate the impacts on groundwater flow of effluent discharge into East Lake, • Simulate the transport of total nitrogen emanating from East Lake and moving through the groundwater system in response to effluent discharges to East Lake, and • Estimate nitrogen loading to the Neuse River from treated wastewater inputs to East Lake. Groundwater flow for the period of January 1, 2016 to December 31, 2016 was simulated for four aquifer units potentially affected by the addition of wastewater to the reclamation facility quarry — the Surficial, the Riverbend, and the Upper and Lower Spring Garden units. Because headwater streams and shallow surface water features far from the reclamation facility are unlikely to affect ground -water flow patterns in the vicinity of East Lake, these features were not explicitly modeled. Therefore, simulations for the Surficial aquifer relate only to the areas in ,M Page 7 n r� the refined portion of the model in the vicinity of the reclamation facility. Table 2 lists a summary of the assumptions and design elements used to develop the model. ow ROM r" FM em rmm MR M rM on M i aDie z. bummarlf or the moaeii uesin ana Assumptions Design and Assumptions Parameters Area Area surrounding the proposed lake including the Neuse River to the northeast and the Clarks Quarry to the west (see Figure 1) Code & Solver MODFLOW 2000 — PCG Solver with GMS 10.2.3 Calibration Period Steady state model calibrated to average 2016 head values Dimensions & Model origin: x= 2,558,000, y = 488,000 (NAD83 State Plane NC - Feet) Orientation X extent: 40,000 ft Y extent: 30,500 ft Grid rotated 320 west consistent with predominant flow direction. Grid Spacing 101 rows & 120 columns (Figure 2) Cell size: ranging from approximately 167 ft x 167 ft to 500 ft x 500 ft Layers 6 layers: SurFcial aquifer (SA), River Bend (RB), Upper Spring Garden (USG), Lower Spring Garden Semi -Confining Unit (LSGCU), Lower Spring Garden (LSG), Lower Castle Hayne Confining Unit (LCHCU) (Figure 3) Surfaces Based on City of New Bern well logs, DWR well logs, and site borehole data BOUNDARIES No -flow Unless otherwise specified, model extents were left as no flow boundaries by Boundaries default. Inactive Cells Cells within the Clarks Quarry excavation area that were excavated after the date of the NED source data were set to inactive in order to simulate excavation. Inactive cells function as no -flow boundaries. Groundwater Groundwater recharge set at 0.002 ft/day (8.76 in/yr) for majority of model. Recharge Range of recharge used in the model: 0.001 ft/day to 0.005 ft/day (4.38 in/yr to 22 in/yr). All recharge applied to the top of layer 1. Recharge details shown in Figure 4. Constant Head No constant head cells were used in this model. General Head General head boundaries (GHBs) were used in the SA (Layer 1) to simulate the lakes in and around the reclamation facility and to simulate lateral flow to the Trent River, south of the model. GHBs were set in the Riverbend Formation (Layer 2) based on regional groundwater contouring available from the NCDWR website for 2015 (NCDWR, 2017). Drains Drain cells were used to model the Clarks Quarry in the western portion of the model area. Drain cell elevations (set in layer 3) were assigned based on water level data from the Clarks Quarry monitoring wells. Rivers The Neuse River and perennial streams to the east of the reclamation facility were set as MODFLOW River boundaries with elevations trending linearly downstream between two topographic elevations. Pumping Wells Pumping well data were gathered from the NCDWR website (NCDWR, 2017b). ,M Page 8 5.1 Grid and Layer Design _ The study area is the East Lake Wastewater Reclamation Facility located at the former New Bern Martin Marietta Quarry in Craven County. To ensure realistic hydrologic boundaries for the groundwater model, and to reduce the potential for introducing non-realistic boundary effects during modeling, the model was expanded to an area of approximately 44 mi2 (Figure 1). The large areal extent of the model also ensured that the simulation included the closest large-scale groundwater users (e.g., the Martin Marietta Clarks Quarry and the City of New Bern wellfield) so that we could ascertain any predicted interactions between these large groundwater users and groundwater recharge from the East Lake. The finite -difference method requires that the model be discretized horizontally into a two- dimensional grid (Figure 2) and vertically into layers (Figure 3), resulting in a three-dimensional array of cells known as the model grid. GMA developed a 6-layer static groundwater model that represents the units underlying the wastewater treatment facility. Table 3: Model Layers Model Layer Lithology Hydrostratigraphic Unit Layer 1 Surficial Sands and Clays Surficial Aquifer Layer 2 Riverbend Fm. (moldic limestone) Upper Castle Hayne Aquifer Layer 3 Upper Spring Garden (skeletal limestone) Upper Castle Hayne Aquifer Layer 4 Lower Spring Garden (sand) Middle Castle Hayne Semi - Confining Unit Layer 5 Lower Spring Garden (limestone) Middle Castle Hayne Aquifer Layer 6 Lower Castle Hayne (clay) Lower Castle Hayne Confining Unit Elevations for the top of each model layer were constructed by contouring interpreted data using GMS. Hydrogeologic data used to construct layer surfaces in the model were primarily obtained from drilling data collected by GMA within the model area, from Clarks Quarry well _ records, and from the DWR online ground water database (https://www.ncwater.org/?pa(ie=20). The elevation of the land surface was based on the USGS 1/3 arc -second National Elevation Dataset available online (https://nationalmai).00v/elevation.htmi). Data were correlated to construct a three- dimensional representation of aquifers and confining units. The six hydrostratigraphic layers modeled are listed in Table 3. The model grid has 101 rows and 120 columns (Figure 2). The spatial discretization of the model grid determines the _ resolution. GMA has designed a model grid that is tighter in the area surrounding the East Lake to increase the resolution in this area. This increased resolution more accurately simulates groundwater flow from the lake to the nearby Neuse River, located between 120 and 1200feet to the northeast. Cell dimensions range from approximately 167 feet by 167 feet ,� Page 9 M" em (approximately 0.64 acres) in the refined model area to approximately 500 feet by 500 feet (approximately 5.7 acres) outside of the refined area. Principal grid axes (rotated 32' west) "" align with the predominant groundwater flow direction and are generally parallel the Neuse River. 5.2 Model Boundaries Boundary conditions are necessary to define how the model interacts both internally and with areas outside the model domain. Incorrectly assigned or unrealistic boundary conditions are often the greatest source of error in groundwater modeling. To minimize any potential errors in boundary specification, GMA utilized boundaries corresponding to natural physical boundaries, wherever possible. A summary of the boundaries specified in this model is provided below and boundaries are shown in Figures 5-8. Unless otherwise specified below, model extents were left as no flow boundaries by default. However, the extent of the model was chosen to be large enough so that model edge boundaries would only minimally affect the solutions for the area of interest. Assumptions used to construct the model are summarized in Table 2. General Head Boundaries GMA used general head boundaries (GHBs) to simulate the lakes within the model area and to simulate lateral boundary flows to and from distant boundaries located outside of the model domain. GHBs are assigned by defining a hydraulic head and a conductance value. A GHB will remove or add water to adjacent cells with lower hydraulic head, based on an assigned hydraulic head, and a threshold conductance, expressed in square feet per day. The flow to adjacent cells is not allowed to exceed the conductance of the general head boundary. Because the primary purpose of this model was to predict nitrogen loading to the Neuse River r, and because East Lake water levels can be artificially maintained, GMA modeled the East Lake as a fixed lake level model using a GHB (Figure 5). As a surface water feature, East Lake is not represented explicitly within the model (i.e. the lake itself is not represented by grid cells); 'm rather a general head representing the artificially maintained head within the lake was assigned to the thin section of layer 1 corresponding to the lake bottom. In the initial calibrated steady- state model, the elevation of the East Lake water surface was set at 5.71 feet ASL, which was the average lake level in 2016. ,e" GHBs were also used in the Surficial aquifer layer to simulate the surrounding lakes (West Lake, North Lake, and the private lake located north of North Lake) and to simulate lateral flow to the Trent River, which is located outside of the model domain to the south. GMA used regional groundwater contouring available from the NCDWR Ground Water Management Branch Map Interface (NCDWR, 2017) for 2015 to assign GHBs to the UCHA (Figure 6). fm Page 10 MR Rivers and Drains fm The Neuse River was simulated as a MODFLOW river boundary with elevations trending linearly downstream (Figure 5). During model calibration, GMA determined that accurate simulation of water levels in the monitoring wells at the reclamation facility required incorporating the no channels and creeks that lie between the quarry lakes and the Neuse River. These features were also modeled as river boundaries. ` l Drain cells were used to simulate the Clarks Quarry in the western corner of the model extent. Both river and drain cells represent head dependent flux boundaries; but of the two, only river cells are capable of recharging the groundwater system. In MODFLOW, if the head in a drain cell falls below the specified elevation, the flux from the drain model cell drops to zero. In T other words, drains can only take water out of the model. MR 5.3 Model Input Model input included recharge estimates (Eimers et al., 1994 and Giese et al., 1997), hydraulic conductivity values for the aquifers and confining units modeled (Matyiko, 2001; GMA 2008- 2013), withdrawals by pumping, and average observed groundwater levels for 2016. M" RecharcL The discharge from the surficial aquifer to the small tributaries outside of the immediate reclamation facility vicinity was not simulated as part of this model. Instead, recharge in this area was reduced to represent only the fraction estimated to infiltrate to the deeper confined MR aquifers, a practice employed in previous models within the region (e.g. Eimers et al., 1994 and Geise et al., 1997). The distribution of recharge throughout the modeled area is shown in Figure 4. Recharge was reduced in the dewatered area at the Clarks Quarry. GMA also determined during model calibration that increased recharge rates were needed in the topographical high area just west of the reclamation facility. All modeled recharge rates were within the range of published estimates of groundwater recharge to the unconfined Surficial aquifer in this region (e.g. 5-20 in/year (0.001-0.005 ft/day), Heath 1980). ran Hydraulic Conductivity Hydraulic conductivity (K) is a measure of an aquifer's ability to transmit water. More ,M specifically, hydraulic conductivity is a measure of the volume of water transmitted in a unit of time through a unit area of the aquifer measured at right angles to the direction of flow under a hydraulic gradient of one. Hydraulic conductivity is equal to the transmissivity of an aquifer divided by the aquifer thickness (Heath, 1983). GMA assigned uniform values of hydraulic conductivity to the Upper Spring Garden (layer 3), the Lower Spring Garden Semi -Confiner (layer 4), and the Lower Castle Hayne confining layer (layer 6) (Table 4). The distribution of hydraulic conductivities applied to the surficial aquifer (layer 1), the Riverbend portion of the Upper Castle Hayne Aquifer (layer 2), and the Middle Castle Hayne Aquifer (layer 5) is shown in Figure 9. Modeled K values were based on measured site specific data from aquifer testing 029 Page 11 Om where available (e.g. Matyiko 2001, New Bern Castle Hayne wellfield data). Estimates of transmissivity for the MCHA were known from multiple aquifer tests conducted at the City of 'W New Bern Castle Hayne wellfield. Using these data, GMA assigned a distribution of horizontal K (Kh) values for the MCHA aquifer (Figure 9). MR Values of vertical hydraulic conductivity (K„) are generally small and are typically at least an order of magnitude less than the Kh. GMA modeled vertical hydraulic conductivity (K„) as one order of magnitude less than Kh for all layers except the Upper Spring Garden (layer 3), which was slightly lowered during the calibration process (Table 4). on Table 4: Hydraulic Conductivity and Porosity of the Model Layers em fm Model Aquifer/Confining Unit Kh (ft/day) Kv (ft/day) Porosity Layer Layer 1 Surficial Aquifer SA See Figure 9 0.25 Layer 2 River Bend RB See Fi ure 9 0.15 Layer 3 Upper Spring Garden USG 3.0 0.08 0.15 Layer 4 Lower Spring Garden Semi - 1.0 0.1 0.15 Confining Layer LSGSU Layer 5 Lower Spring Garden (LSG) See Fi ure 9 0.3 Layer 6 Lower Castle Hayne 0.001 0.0001 0.2 ConfiningLayer LCHCU MR Groundwater Withdrawals The removal of groundwater via pumping results in a lowering of hydraulic head near the pumping wells and must be accounted for during modeling. Pumping wells withdrawing within the study area were identified using NCDWR Central Coastal Plain Capacity Use Area permitting Mel records and GMA's regional knowledge (Figures 6-8). The majority of pumping wells located within the model area belong to the City of New Bern's Castle Hayne Aquifer wellfield. Estimated daily pumping rates were based on average pumping data for 2016, which were available from the NCDWR website (NCDWR, 2017b). 5.4 Steady -State Calibration Steady-state flow was simulated to represent conditions in 2016. GMA assigned average head observations from 2016 to 16 monitoring and inactive production wells within the study area for Pa use during model calibration (Figure 10). Head observations were included for wells screened within the Surficial aquifer (n=1), the Riverbend Formation (UCHA; n=8), the Upper Spring Garden (UCHA; n=4), and the Middle Castle Hayne aquifers (n=3). Although multiple monitoring wells exist at the Clarks Quarry site, GMA eliminated all but one of the Clarks Quarry wells from the calibration well dataset. GMA did not use these wells due to the coarseness of the grid in this region of the model relative to the complexities of the excavation geometry and pumping schedules. Calibration of the model to wells between the Clarks Quarry and the East Lake Reclamation facility (e.g. well PW6 and the Craven 30 Supply well) ensure that the cone of MIR depression from the mine is well simulated. ,M Page 12 Model input, specifically, hydraulic conductivities, confining -unit vertical conductance, and recharge rates, were adjusted during model calibration so that simulated heads best matched measured head data for 2016. To calibrate the model, parameter estimates were manually adjusted and optimized to increase the goodness -of -fit between observed and calculated heads. All parameter adjustments were kept within realistic limits according to available site specific and published information. A comparison between modeled and observed groundwater levels indicate a very good fit (r = 0.921, Figure 11). The differences between the observed and calculated heads ranged from 0.03 feet to 1.54 feet. The mean difference was 0.64 feet and the root mean square residual was 0.79 feet. Calibrated head data were used to prepare potentiometric surface maps for the layers 1 through 5 to simulate groundwater head elevations and flow directions within these aquifers (Figures 12-16). 5.5 Sensitivity Analysis As part of the steady-state model calibration process, GMA performed a preliminary sensitivity analysis to evaluate the relation between parameter variability and the modeled hydraulic head in the model. Model response was tested for sensitivity to river conductance values, recharge, and hydraulic conductivity values using the PEST (Parameter ESTimation) utility. PEST analysis indicated that the model was sensitive to horizontal hydraulic conductivity values in the Riverbend and Upper Spring Garden layers and that the model was also sensitive to recharge values. 5.6 Predicted Groundwater Flow Patterns The simulated groundwater elevations for the Surficial, Riverbend, and Upper Spring Garden indicate that the East Lake is functioning as a recharge boundary that drives groundwater radially and downward under the induced, artificial head created by the wastewater facility (Figures 12-17). The recharge effects from the East Lake diminish with depth, but a subtle mound is evident in the model extending down into Layer 6 (the Lower Spring Garden unit). The induced head beneath the East Lake results in a groundwater divide in the upper layers of the Castle Hayne Aquifer beneath the lake (Figure 17). Groundwater recharge on the east side of that divide discharges to the Neuse River, and groundwater recharge west of the divide discharges to the West Lake. Flow budget analysis indicates that approximately 62% of the groundwater recharge exiting the East Lake ultimately discharges to the Neuse River. The remaining 38% discharges to the West Lake, where it then flows as surface water back to the Neuse River, or it is lost as evapotranspiration. As shown on the potentiometric surface map for the surficial aquifer (Figure 12), the majority of groundwater flow out of East Lake is concentrated at the sides of the basin where hydraulic VIM gradients are the steepest and where aquifer permeability is the highest. Travel times for the majority of groundwater flow from the recharge at East Lake to the areas of discharge to the Neuse River or to the West Lake range from 66 days to 5 years. Model results indicate that approximately 4% of the total groundwater recharge entering from the East Lake is directed ,� Page 13 downward through deeper, less permeable, layers of the Castle Hayne Aquifer System before eventually discharging at the Neuse River or the West Lake. Travel times for this deeper portion of the groundwater flow from the East Lake are on the order of 100+ years. Model results indicate that groundwater pumping associated with dewatering at the Martin Marietta Clarks Quarry does not have an effect on the groundwater flow from the East Lake Reclamation Facility. The cone of depression predicted by this model does not extend to the facility property. The model also predicts that pumping from the City of New Bern's production wells does not affect groundwater flow from the East Lake. New Bern's Castle Hayne production wells are screened in the productive limestone section of the Lower Spring Garden unit. Impacts from pumping extend upward into the Lower Spring Garden Semi -confining unit, but do not greatly perpetuate beyond that (Figures 13-16). 5.7 MT3DMS Transport Modeling MT3DMS (Modular Transport, 3-Dimensional, Multi -Species model) is a modular three- dimensional transport model for the simulation of advection, dispersion/diffusion, and chemical reactions of dissolved constituents in groundwater flow systems (Zheng and Wang, 1999). MT3DMS can be used in conjunction with MODFLOW to predict concentrations of a solute plume moving through groundwater. GMA used MT3DMS in conjunction with the calibrated MODFLOW model developed for the site to predict TN concentrations moving from East Lake into and through the groundwater system. In this study, the flow simulation was steady state (i.e. with one stress period and one time step), but the transport simulation was transient (i.e. with multiple stress periods and time steps). Because the flow model was steady state, the velocity and groundwater sink/source information from the MODFLOW model is only updated once at the beginning of the simulation ' -9 and remains the same, while changes in sink/source concentrations are determined by the stress periods (time intervals in which external sinks/sources are constant) set up with the transport model. GMA modeled nitrogen additions to the East Lake using a mass loading approach. Nitrogen mass was added to the general head boundary representing the surface of the East Lake. Initial nitrate concentrations were assumed to be zero at the beginning of all simulations. GMA also assumed that all of the nitrogen added to the lake enters the groundwater system and that degradation of nitrogen did not occur — an extremely conservative approach. 5. Z 1 M73DMS Seft'nQs Dispersion is the spreading of a solute over a greater area than would be predicted solely from advective transport. Dispersion is caused by differences in the velocity that water travels at the pore level as well as differences in the rate at which water travels through different strata in the flow path (Fetter, 1980). Dispersion is important because it allows the solute to be diluted by unaffected groundwater and affects the first arrival time of a plume. Longitudinal dispersivities for the model were estimated using published field dispersivities for similar aquifer types Page 14 rim PER (Gelhar et al. 1992). Horizontal dispersivities were assumed to be one order of magnitude smaller than longitudinal dispersivities, and vertical dispersivities were assumed to be an order of magnitude smaller than that (Gelhar et al. 1992). Advection refers to the transport of miscible solutes at the same velocity as the average groundwater flow. GMA used the Third Order Total -Variation -diminishing ('TVD) scheme (ULTIMATF) solver for the Advection package in MT3DMS. This TVD method is a mass conservative, Eulerian solution technique that is generally more accurate at solving advection- dominated problems (Zheng and Wang, 1999). f' 5.72 Nitrogen Loading Scenarios To evaluate the potential nitrogen loading to the Neuse River, GMA modeled nitrogen transport FM resulting from wastewater discharge to the East Lake under three scenarios: 1) Actual wastewater discharge to the East Lake during 2016, 2) Quarterly wastewater discharge to the East Lake, which represents the practical maximum usage of the reclamation facility by the City, and 3) Continuous discharge of wastewater to the East Lake every day for 15 years During 2016, the City discharged wastewater to the East Lake during the months of May and November (Appendix I). The 2016 usage pattern required that GMA simulate five stress periods in the transport model to represent the two months of nitrogen inputs to East Lake. Per request of the City of New Bern and Rivers and Associates, Inc., GMA ran a subsequent MT3DMS simulation that modeled effluent discharge to the East Lake at a rate of 5.0 million gallons per day (MGD) and at a concentration of 5.0 mg/L of TN for four equally -spaced months per year. This quarterly usage scenario represents the realistic maximum time period that the City would discharge to the East Lake. As an additional run, GMA also modeled effluent discharge to the East Lake at a rate of 5.0 MGD and a concentration of 5.0 mg/L of TN every day for 15 years. 5.7.3 M73DMS Predictions MT3DMS groundwater transport simulations under the 2016 usage scenario predict no measurable increase over base level TN concentrations at the groundwater monitoring well locations (Figure 18). The results suggest that, under present usage patterns, TN added to the East Lake is being rapidly diluted in the groundwater system to concentrations that cannot be distinguished from the naturally occurring background concentrations. rM These findings are consistent with historical water -quality monitoring of wells surrounding the East Lake facility. Historical monitoring well sampling does not indicate an increase in total r' nitrogen (TN) concentrations above the background concentrations that were measured prior to the facility's first use (Figure 19). Further, base -level (i.e., preexisting) concentrations in some FM monitoring wells (e.g. EL1 and EL3) prior to use of the facility exceed the measured TN ,, Page 15 F" MM concentrations in the East Lake after the start of facility use. The field monitoring data indicate, on average, that the addition of wastewater into the East Lake is not resulting in a recognizable increase in TN concentrations in the groundwater monitoring wells surrounding the facility. Even under the extreme assumption of daily wastewater discharge to the East Lake, groundwater concentrations of TN at the point of discharge to the Neuse River are only predicted to be 0.001 mg/L (Figure 20). To predict mass loading of TN to the Neuse River, GMA evaluated the MT3DMS mass budget for each model simulation. MT3DMS simulations predict that after 15 years under the quarterly usage scenario, approximately 8,963 pounds of TN will be discharged to the Neuse River each year (Table 5). This represents only 35% of the total annual nitrogen discharged to the East Lake under a quarterly usage scenario. Thus, the city should be eligible for credit for at least 65% of the nitrogen added each year. Under the unrealistic assumption that wastewater would be discharged to the reclamation facility continuously for 15 years, the model predicts an annual TN load to the Neuse River of 29,558 pounds, which represents 39% of the TN input under this scenario. Please note: we recognized a previous incorrect error in the technical memo (GMA, 2017) with regard to mass loading calculations. In the technical memo, the loading calculations were erroneously predicted to be much lower than what we report here. In this final report we are correcting those mass loading calculations. As previously stated, GMA performed all MT3DMS simulations without applying a retardation factor for nitrate, a very conservative assumption that ignores the denitrification processes that almost certainly occur along the groundwater flow path from the East Lake to the Neuse River. Denitrification is an anaerobic respiratory pathway that reduces nitrate (NO,,) to nitrous oxide (N20) or di -nitrogen gas (N2). This microbially-facilitated process occurs under anoxic conditions and requires a labile, organic carbon (C) source to serve as food for the denitrifying rum bacteria. Research on the main stem of the Neuse River and its tributaries indicates that denitrification represents a sink of approximately 5% of the nitrate -nitrogen load to these waterbodies (Whalen et al., 2008). Applying this finding, nitrogen loading to the Neuse River under the maximum potential usage of the reclamation facility (quarterly wastewater inputs), is predicated to be 8,515 pounds per year (Table 5). SM rMq ow "M Page 16 rn V Table S. Summary of Nitrogen (N) Loading to the Neuse River from the Wastewater Reclamation Facility MT3DMS Model Output IN OUT Potential N Estimated N Minimum Credit assuming N added to Estimated N % of loading % of Potential a 5% Reduction East Lake loading to the Neuse River Total to West Lake Total N Credit in Load due to Denitrification3 Modeling (lb/yr) (lb/yr) Input (and surrounding Input Scenario lakes) (lb/yr) Max every day at 15 yearsi 76,176 29,558 39% 17,498 23% 61% 66% Quarterly use u rt years2 25,638 8,963 o 35 /0 5,521 0 22 /0 0 65 /0 0 70 /o i Modeled effluent discharging to the East Lake at a rate of 5.0 MGD and a concentration of 5.0 mg/L of TN every day for 15 years. Z Modeled effluent discharging to the East Lake at a rate of 5.0 MGD and a concentration of 5.0 mg/L of TN for four equally spaced months per year. This scenario represents the realistic maximum time period that the City would discharge to the East Lake 3Reduction in load due to denitrification based on research conducted in the Neuse River by Whalen and others (2008). rom MR 5.8 Model Limitations Mm The steady-state, finite -difference model described in this report reasonably simulated groundwater flow and nitrogen loading from the East Lake into the groundwater system. However, due to the inherent complexities of groundwater flow systems in both space and time, MN and considering limitations on available data and computing capabilities, there are some model limitations. Any model is limited by the quantity and quality of the supporting data. This model has the benefit of representing an area that is well studied. As with any model, however, there is a fair degree of uncertainty in the hydraulic properties of the aquifers and confining units. Site - specific parameters were used whenever available. At locations without known data, GMA used hydraulic property values obtained from the literature or based on GMA's experience with the hydrogeology of this region. This model, like all models, is also limited by grid spacing. Although the grid spacing of this model adequately allows for simulation of the hydrologic system within the area, data input and simulation results are averaged over the entire cell. Consequently, small heterogeneities within the system, such as small streams or well clusters, can result in discrepancies between modeled and observed values. It should also be noted that because East Lake does not occupy space within the model grid, the exchange of water with the underlying aquifer is assumed to occur only vertically through the bottom of the lake. Furthermore, this model does not accurately predict the groundwater flow patterns in the surFcial aquifer outside of the refined model area surrounding the reclamation facility. Surface drainage features were only modeled in the refined area where such features could potentially be affected, and model calibration only included Sunccial aquifer head data from this area. Because this is a steady-state model calibrated to average groundwater levels and pumping rates for 2016, the model also does not account for potential seasonal variation in groundwater recharge or withdrawals rates. Although there is uncertainty in the model predictions, GMA contends that this model reasonably simulates groundwater flow and advective transport of dissolved constituents added to the East Lake of the New Bern Wastewater Reclamation facility based on the best available Mr data. We further contend that the model reasonably predicts nitrogen loading to the Neuse River resulting from the input of wastewater to the lake. MR 6.0 CONCLUSIONS AND RECOMMENDATIONS r, GMA has completed a predictive modeling analysis of groundwater flow and nitrogen loading from the City of New Bern Wastewater Reclamation Facility's East Lake. GMA's groundwater ' model focused on determining groundwater recharge effects of wastewater additions to the East Lake and, ultimately, nitrogen loading to the Neuse River as a result of groundwater discharge. Based upon our modeling efforts, GMA concludes the following: Page 18 era MW • Usage of the East Lake Reclamation Facility for wastewater disposal by the City of New Bern delays and reduces total nitrogen inputs from the City's wastewater to the Neuse River. • The artificial head maintained in the East Lake drives groundwater radially and downward and results in a groundwater divide in the upper layers of the Castle Hayne Aquifer beneath the lake. Groundwater recharge on the east side of that divide discharges to the Neuse River, and groundwater recharge west of the divide discharges to the West Lake. • An estimated 62% of the groundwater recharge exiting the East Lake ultimately discharges to the Neuse River. The remaining 38% discharges to the West Lake, where it then flows as surface water back to the Neuse River, or it is lost as evapotranspiration. • Groundwater pumping associated with the nearby Martin Marietta Clarks Quarry and the City of New Bern production wells do not have an effect on the groundwater flow from the East Lake Wastewater Reclamation Facility. • Historical water -quality monitoring of wells surrounding the East Lake facility do not indicate an increase in TN concentrations above the background concentrations that were measured prior to the facility's first use. • Consistent with historical water -quality monitoring, MT31DMS groundwater transport „tea simulations predict little to no increase over base level TN concentrations at the groundwater monitoring well locations. TN added to the East Lake is being rapidly diluted in the groundwater system to concentrations that cannot be distinguished from the naturally occurring background concentrations. • MT3DMS simulations that modeled effluent discharge to the East Lake at a rate of 5.0 million gallons per day (MGD) and a concentration of 5.0 mg/L of TN for four equally FWD spaced months per year predict a TN load to the Neuse River of 8,963 pounds per year, or 35% of the total annual input. This scenario represents the realistic maximum time period that the City would discharge to the East Lake. Research on the Neuse River indicates that denitrification represents a sink of approximately 5% of the nitrate -nitrogen load to the river. Therefore, under the maximum potential usage of the reclamation facility (quarterly wastewater inputs), GMA estimates an annual nitrogen load of only 30% of the TN added to the East Lake, which corresponds to a potential TN credit of 70%. 7.0 REPORT CERTIFICATION FM This hydrogeologic characterization and predictive modeling report was prepared by Groundwater Management Associates, Inc., a professional corporation licensed to practice geology (NC Corporate License No. C-121) and engineering (NC Corporate License No. C-0854) in North Carolina. We, Emma H. Shipley, James K. Holley, and Richard K. Spruill, North Carolina Licensed Geologists for GMA, do certify that the information contained in this report is correct and accurate to the best of our knowledge. M Ma Page 19 _ Emma Shipley, .G. Project Hydrogeologist James K. Holley, P.G. ej Senior Hydrogeologist Richard K. Spruill, Ph.D., P.G. Principal Hydrogeologist _ Page 20 rW 8.0 LIST OF REFERENCES Eimers, JL, CC Daniel, and RW Coble. 1994. Hydrogeology and Simulation of Ground -water Flow at U.S. Marine Corps Air Station, Cherry Point, North Carolina, 1987-90. Fetter, C.W., 1988, "Applied Hydrogeology", Second Edition, Merrill Publishing Company, Columbus, Ohio, p. 567. �► Fine, JM, 2008, "Hydrogeologic Framework of Onslow County, North Carolina, 2008": U.S. Geological Survey Scientific Investigations Map 3055, 1 sheet. Giese, GL, IL Eimers, and RW Coble, 1997, "Simulation of Ground -Water Flow in the Coastal Plain of North Carolina": U.S. Geological Survey Professional Paper 1404-M. Gelhar, GL, C Welty, and KR Rehfeldt, 1992, "A Critical Review of Data on Field -Scale Dispersion in Aquifers', Water Resources Research, vol. 28, no. 7, p. 1955-1974. Groundwater Management Associates Inc. (GMA), 1998, "Hydrogeologic Evaluation, Abandoned Glenburnie Quarry", prepared for Rivers and Associates, Consulting Engineer for the City of New Bern, NC. Groundwater Management Associates Inc. (GMA), 2006, "Monitoring Well Installations and the Abandonment of Monitoring Wells at the Former martin Marietta Glenburnie Road Quarry and the Mackilwean Turf Farm near New Bern, Craven County, North Carolina", Prepared for the City of New Bern, NC, 5 pages plus figures. fW Groundwater Management Associates Inc. (GMA), 2008-2013, Various test well and production well reports for the 16 well sites for the City of New Bern wellfield. Groundwater Management Associates Inc. (GMA), 2014, "Hydrogeologic Framework of Onslow County, North Carolina", Prepared for the Onslow Water Resources Group, Jacksonville, North Carolina, 75 pages plus figures, tables, and appendices. M' Horton, JW, and VA Zullo, 1991, Geology of the Carolinas: Carolina Geological Society 50th Anniversary Volume, 1st Ed., 424 pages. Lawrence, DP, and CW Hoffmann, 1993, "Geology of basement rocks beneath the North Carolina Coastal Plain", North Carolina Geological Survey, Bulletin 95, 60p, one Plate. 'm Lyke, WL and Winner, MD, 1990, Hydrogeology of aquifers in Cretaceous and younger rocks in the vicinity of Onslow and southern Jones Counties, North Carolina": U.S. Geological Survey Water -Resources Investigation Report 89-4128. MM Matyiko, S, 2001, "Hydrogeologic Evaluation of the Castle Hayne Aquifer System North of New Bern, North Carolina: Implications for Disposal of Tertiary -treated Wastewater into the 'W Martin Marietta Glenburnie Quarry," Thesis, Master of Science in Geology, East Carolina University, Greenville, North Carolina. M ,M-9 Page 21 North Carolina Division of Water Resources (NCDWR), 2017, "Ground Water Management Branch Map Interface" https://www.ncwater.ora/GWMS/`openlavers/ol.php?menulist=bl. Accessed October 2017. North Carolina Division of Water Resources (NCDWR), 2017b, "Central Coastal Plain Capacity Use Area: Query the Data" htto://www.ncwater.orci/?page=49&menu=QuervtheData. Accessed October 2017. .. North Carolina Geological Survey (NCGS), 1985, Geologic Map of North Carolina, one plate. Pohlig, KO, 2000, Groundwater Flow and Transport Analysis for a Water Reclamation Project. NC Division of Water Quality Report. Pollock, DW 2012. User Guide for MODPATH Version 6 — A Particle -Tracking Model for .. MODFLOW: U.S. Geological Survey Techniques and Methods 6-A41, 58 p. Ward, LW, Lawrence, D.R., and Blackwelder, B.W., 1978, Stratigraphic Revision of the Middle Eocene, Oligocene, and Lower Miocene — Atlantic Coastal Plain of North Carolina, Geological Survey Bulletin 1457-F, Whalen, SC, Alperin, MI, Nie, Y, and EN Fischer, 2008, Denitrification in the mainstem Neuse River and tributaries, USA, Fundamental and Applied Limnology, 171:249-261. Winner, MD and Coble, RW, 1996, "Hydrogeologic Framework of the North Carolina Coastal Plain Regional Aquifer System Analysis: U.S. Geological Survey Professional Paper 1404- I. Zheng, C and PP Wang. 1999. MT3DMS: A Modular Three -Dimensional Multispecies Transport Model. US Army Corps of Engineers Contract Report SERDP-99. Zheng, C. 2010. MT3DMS v5.3: Supplemental User's Guide. Technical Report for the U.S. Army Corps of Engineers. 50 pages. Page 22 FIGURES '1 (y A ^ At BLUE RIDGE COASTAL PLAIN I I PHYSfOGRAPHIC PROVINCES b _ e i f s A 1 :I• 11 el r , flr ' r r •i F, .YA ��M" _ i" tv"!'�' 0 ��S prb ys�y��J f4° is � f liv 4Y� I MM•��' k�.t� File: , WASTEWATER I 11N=4MI P P A Af WEST EAST NEUSE LAKE LAKE RIVER LAYER r-A--% r-41--I , =% 1 (SA) SURFICIALAQUIFER 2 (LICHA ERIVERBEtD FM. 3 (UCHA) UPPER SPRING GARDEN 4 (MCHSC) LOWER SPRING GARDEN (SAND) LILL 5 (MCHA) 6 (LCHC) - - - LOWERISPRING GARDEN (LIMESTONE) `'" '' " `''" LOWER CASTLE HAYNE (CLAY.) VERTICAL EXAGGERATION = 25x -LEGEND- MODELAREA GMA HYDROSTRATIGRAPHICABBREVIATIONS: —PROFILE SA = SURFICIAL AQUIFER UCHA=UPPER CASTLE HAYNE AQUIFER MCHSC = MIDDLE CASTLE HAYNE SEMI -CONFINER MCHA= MIDDLE CASTLE HAYNE AQUIFER-� HYDROGRAPHY LCHC = LOWER CASTLE HAYNE CONFINER GROUNDWATER MINIOFMFM gf50<IgTFS, INC. File: DRAWINGS/21342/ CROSS SECTION THROUGH EAST LAKE DATE: 4/18/2018 FIG3 GRID LAYERS .MXD SHOWING VERTICAL DISTRIBUTION OF MODEL LAYERS PROJECT21342 NEW BERN, CRAVEN CO., NC FIGURE 3 m f .a i _ 1 i t CLARKS QUARRY & _ IEAST LAKE I L IP•i -1_ F' .,7 LEGEND HEAD (FT MSL) 14 -2 -18 GMA -34 0 7000'� "iign� ��...'- -50 1 IN = 7,000 FT Z:\Drawings\21342\ FIG14_USG_POT.mxd MODELED POTENTIOMETRIC SURFACE FOR THE UPPER SPRING GARDEN AQUIFER DATE: 4/18/2018 PROJECT: 21342 EAST LAKE RECLAMATION FACILITY, NEW BERN, CRAVEN, CO., NC I FIGURE 14 LEGEND PROJECT: 21342 i v_* EAST LAKE i HEAD (FT MSL) nW 14 -2 -18 GMA -50 � O SO 1 IN = 7,000 FT MODELED POTENTIOMETRIC SURFACE DATE: 4/18/2018 FOR THE LOWER SPRING GARDEN SEMI -CONFINER EAST LAKE RECLAMATION FACILITY, NEW BERN, CRAVEN, CO., NC I FIGURE 15 E R pert; 1p FIGURE 17: MODELED GROUNDWATER FLOW WITHIN A VERTICAL SECTION THROUGH THE EAST LAKE WEST LAKE EAST LAKE NEUSE RIVER - - AI 00, RR. 4�� 7�i /T ------- - 4, 1 Ox VERTICAL EXAGGERATION 500 FT -LEGEND- HEAD (FT MSQ NOTE: VIEW IS LOOKING NORTH -10 t RIVER CELL ♦ GENERAL HEAD CELL - GMA 6 OSSEFFNM 4 GROUNDWATER MANAGEMENT ASSOCIATES, INC. 2 10 File: DRAWINGS/21342/ NEW BERN WASTEWATER RECLAMATION FACILITY DATE: 12/8/2017 VERT PROFILE Fig PROJECT21342 CITY OF NEW BERN, CRAVEN CO., NC FIGURE 18: MAXIMUM PREDICTED TOTAL NITROGEN (TN) CONCENTRATIONS ABOVE BACKGROUND DURING 2016 TN: 12/1/201612:00:01 AM 0.0005 mg/L - 0.0004 mg/L 0.0003 mg/L 0.0002 mg/L 0.0001 mg/L EAST LAKE 0.0 mg/L 1 WEST LAKE NEUSE RIVER 1 Ox VERTICAL EXAGGERATION LEGEND- GMA + RIVER CELL ♦ GENERAL HEAD CELL OROUNCW TER MANAOEMENI ASSOCI4iES, INC. File: DRAWINGS/21342/ NEW BERN WASTEWATER RECLAMATION FACILITY DATE: 12/8/2017 TN MT3DMS Fi PROJECT21342 CITY OF NEW BERN, CRAVEN CO., NC 5 4.5 4 3.5 3 E w 2.5 rn 0 z 2 0 r 1.5 1 0.5 0 FIGURE 19. CHRONOLOGICAL TOTAL NITROGEN LEVELS IN MONITORING WELLS SURROUNDING THE EAST LAKE Initiated Operation of Shallow Well Facility —�—Shallow Wells Avg Baseline = 2.45 mg/I t Deep Wells Avg Deep Well Baseline = 1.82 mg/I ✓d y�r✓O�pc ✓d q� ✓4 O� ✓d y� ✓G�p� ✓d yp ✓4jp� ✓d yq ✓4�p✓d y� ✓` 0�f✓d yp2f2y' �74�f✓oY y0S✓✓0 yb ✓G 0 'Sf 7S7'7�6�j7��0"�7j7 ?6E �76'000600'0O Oj0;0�'OjbB0d08'08OAOQO0'o9-O-O;OQ7 DATE FIGURE 20: PREDICTED TOTAL NITROGEN (TN) CONCENTRATIONS ABOVE BACKGROUND AFTER 15 YEARS OF DAILY DISCHARGE OF 5 MGD OF EFFLUENT CONTAINING 5.0 MG/L OF TN TO THE EAST LAKE TN: 12/1/201612:00:01 AM 0.008 mg/L 0.006 mg/L 0.004 mg/L 0.002 mg/L EAST LAKE 0.0 mg/L WEST LAKE NEUSE RIVER MH 1 Ox VERTICAL EXAGGERATION -LEGEND- GMA * RIVER CELL ♦ GENERAL HEAD CELL GROUNOWRIFR MRNRGFMFNI RSSOCIRlES, INC. File: DRAWINGS/21342/ NEW BERN WASTEWATER RECLAMATION FACILITY DATE: 12/13/2017 TN MT3DMS Fig PROJECT21342 CITY OF NEW BERN, CRAVEN CO., NC APPENDIX I m" 2016 EAST LAKE USAGE RECORDS 11 Jan 2016 1 Feb 2016 1 Mar2016 I Apr 2016 May 2016 1 Jun 2016 1 Jul 2018 1 Aug 2016 1 Sep 2016 1 Oct2016 I Nov 2016 1 Dec 2016 1 FINAL Final Measure W WTP: Influent Total influent 157.37 183.13 137.77 111.79 115.94 124.68 128.87 123.96 168.91 182.03 106.90 121.10 1662.45 Total Avg, influent daily 5.07 6.31 4.44 3.72 3.74 4.15 4.15 3.99 5.63 5.87 3.56 3.90 4.54 Average Max. influent daily9.12 12.59 5.69 4.85 4.74 6.53 4.88 5.13 11.34 12.88 4.21 5.26 12.88 Maximum Max Date 23Jan 5-Feb 4-Mar 15-A r 31-Ma 7-Jun 14Jul 3- 3Se 9-Oct 2-Nov 6-Dec Total Inn. BOD tons Removed 152.6 125.33 743.85 119.521i120.77 123.64 118.66 109.57 137.46 152.14 131.24 150.43 1585.21 Total Total Inn. TSS tons Removed 85.64 59.47 101.32 98.49 77.58 49.82 57.6 68.76 74.48 69.50 79.43 101.55 923.64 Total Rainfall WWTP (") 3.22 5.85 3.49 1.8 5.39 4.35 9.98 7.W 10.56 6.61 1.09 3.06 64.56 Total Total Effluent 157.81 183.46 139.19 110.61 115.97 126.56 127.69 127.74 170.83 182.74 106.72 121.81 1671.131 Total Avg. effluent daily 5.09 6.32 4.49 3.68 3.74 4.21 4.11 4.21 5.69 5.89 3.55 3.92 4.57 Avem e Max. effluent daily 9.04 12.65 5.72 4.87 4.81 6.79 5.21 5.24 11.44 12.60 4.27 5.40 12.60 Maximum Max Date Total to River 157.81 183.46 139.19 110.61 7.86 126.56 127.69 127.74 170.83 180.58 2.83 121.81 1456.97 Total Total to Quarry 0.00 0.00 0.00 0.00 108.11 0.00 0.00 0.00 0.00 2.16 103.89 0.00 214.16 Total Avg. Effluent BOD m 2 < 2 < 2 < 2 < 2 < 2 < 2 < 2 < 2 < 2 < 2 < 2 < 2 Average Avg. Effluent TSS m 1L 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 < 2.5 Average Avg. Effluent TN m 9.11 5.24 5.9 4.73 2.61 1.32 1.35 1.69 3.21 5.37 8.27 7.76 4.71 Average Avg. Effluent Phos. m 1.12 0.428 0.68 0.732 0.304 0.4 0.682 1.44 1.04 1.12 1.42 1.24 0.88 Average Rainfall Quarry 3.23 6.82 3.2 1.85 3.75 4.87 9.25 6.5 11.05 5.58 0.91 3.82 60.83 Total Quarry Level 6.5 6.3 5.8 5.2 5.6 6.1 5.5 5.1 5.2 5.3 5.8 6.1 5.710 Average Total um MMQ to Mackllvieen O.00D 0.000 0.000 0.720 0.704 1.519 1.031 0.214 0.349 0.917 1.011 0.000 6.4651 Total Total s rasd at MackInvean I O.coul U.ucul 0.000 0.7201 0.7utI 1.5191 1.0311 U.2141 0.3491 0.917 1.011 0.000 6.465 Total Anything out of compliance I No I No I No I No I No I No I No I No I No I No No No11 �jl am LAYER 1: k SURFICIAL AQUIFER S LAYER 2: UPPER CASTLE HAYNE AQUIFER .;x LAYER 5: MIDDLE CASTLE HAYNE AQUIFER LEGEND HORIZONTAL HYDRAULIC CONDUCTIVITY (Kh; FEET/DAY) GMA _0.01 -18 -31 -63 80 -100 -150 -5.0 -23 -50 -65 -85 -113 -270 Z:\Drawings\21342\ MODELED HORIZONTAL HYDRAULIC CONDUCTIVITY DATE: 4/18/2018 FIG 9 Kh.pdf DISTRIBUTIONS PROJECT. 21342 EAST LAKE RECLAMATION FACILITY, NEW BERN, CRAVEN, CO., NC FIGURE 9 m m Clarks MW5 PW3 EL6 PW8 EL5"e- EL4 PW6 ED EL1 Clarks RS USG OWt Craven 30 Supply PW5 OW2 OW3 iP 'L 0i J LEGEND OBSERVATION WELL USED FOR CALIBRATION GMA 0 7000' 1 IN = 7,000 FT zFIGIOw_013S t842\ OBSERVATION WELLS DATE: 4/18/2018 FIG10 OBS WELLS.mxd PROJECT: 21342 EAST LAKE RECLAMATION FACILITY, NEW BERN, CRAVEN, CO., NC FIGURE 10 Computed vs. Observed Values Head 11 } EL1 10 X EL2 p EL3 B • EL4 ❑ EL6 8 0 ■ ELB PW3 7 ♦ PINS o p PW8 8A PWB . V Clarks RS USG E u 6 V Clarks MWS U } OWt 4 + i OW2 � O OW3 3 ■ • • • Craven 30 Supply 2 • cb 1 r2=0.92 o B 0 1 2 3 4 5 8 7 ll B 10 11 Observed Residual vs. Observed Values Head 2.0 } EL1 X EL2 1.6 0 EL3 EL4 ❑ EL5 1.0 • ■ • o ■ EL6 ;. PW3 0.6 ♦ PINS o + p PW8 ° •A PINS 0 A V Clarks RS USG x ♦ Clarks MW5 + OW1 x OW2 O OW3 •1.0 • Craven 30 Supply •1.S -2.0 2 3 4 5 6 7 8 B 10 11 Observed STATISTICS Mean Residual (Head, in feet) 0.30 GMA Mean Absolute Residual (Head, in feet) 0.64 Root Mean Squared Residual (Head, in feet) 0.79 ZFIG10ingsbratio\ MODEL CALIBRATION PLOTS DATE: 4/18/2018 FIG 10 Calibration.pdf PROJECT: 21342 EAST LAKE RECLAMATION FACILITY, NEW BERN, CRAVEN, CO., NC I FIGURE 11 4 M 00, ,z �IT mil- V '�. I r 7w,�1,�• ` '4 , f i a N