Brine Dilution Analysis for DeepWater Desal, LLC, Monterey Bay Regional Water Project at Moss Landing, CA

Transcription

1 Brine Dilution Analysis for DeepWater Desal, LLC, Monterey Bay Regional Water Project at Moss Landing, CA by Scott A. Jenkins, Ph.D. and Joseph Wasyl Submitted by: Scott A. Jenkins Consulting Kalapana St Poway CA Submitted to: K. Scott Jackson, Program Manager DeepWater Desal, LLC 7532 Sandholdt Rd. Suite 6 Moss Landing, CA Draft 7 March; Revised 25 March 2014; 2 nd Revision 11 April 2014; 3 nd Revision 23 June 2014

2 2 TABLE OF CONTENTS ABSTRACT ) Introduction ) Discharge and Dilution Compliance Issues ) Technical Approach ) Model Initialization ) Diffuser Dilution Analysis ) Conclusions ) References APPENDIX-A: Diffuser Turbulence Mortality Estimates..104

3 ABSTRACT: We evaluate the dilution and dispersion of combined brine and thermal effluent discharged from DeepWater Desal s proposed Monterey Bay Regional Water Project (MBRWP) at Moss Landing, CA. This problem was studied by numerical simulation of worst case and long term scenarios using a combination of hydrodynamic models: 1) a near field mixing zone model, Visual Plumes certified by the U.S. Environmental Protection Agency and the California State Water Resources Control Board for use in ocean outfall design; 2) a fully 3-dimensional far field dispersion model SEDXPORT that is a processed-based, stratified flow model with the complete set of littoral transport physics including tidal transport, wind &wave induced transport and mixing; and 3) two commercially computational fluid dynamics models (CFD) known as COSMOS/ FLowWorks for hydraulic design and Star-CD, Version 3.1, with QUICK space discretization for first order up-winding of the diffuser turbulence equations. We consider two potential offshore discharge sites located along the existing Moss Landing Power Plant (MLPP) oil pipeline route; one located in 35 m of local water depth (deep discharge site) at UTM coordinates N, E; while the other is located in 25 m of water depth (intermediate discharge site) at UTM coordinates N, E. At these two potential discharge sites, the study addresses hyper-salinity toxicity relative to existing and proposed discharge limits. The discharge/dilution strategy proposed in the MBRWP project description is based on high velocity diffusers. Our analysis is based on a proxy diffuser design utilizing a five-jet linear diffuser oriented orthogonal to the shoreline and consisting of five discharge risers emerging from a manifold and fitted with duckbill diffuser nozzles. Five such nozzles will discharge a maximum of 5.45 mgd each of brine, or an ultimate maximum total discharge of mgd of brine 3

4 at ppt end-of-pipe salinity. (These figures represent ultimate maximum brine source loading to the receiving waters. When operating at preferred efficiency, the SWRO will operate at 46% recovery, producing 26.1 mgd of brine with an average end-of-pipe salinity of 63.7 ppt). Potential hyper-salinity toxicity arising from the project discharge is evaluated relative to two regulatory discharge limits. Existing discharge limits applied to fully permitted desalination projects in Southern California have been based on the 0.3 TUa objective of Requirement III.C.4(b) of the present version of the California Ocean Plan, as it would apply to a Zone of Initial Dilution (ZID), typically a circle with 1,000 ft. (305 m) radius. However, a recently released study by the California Water Resources Control Board (SWRCB) Science Advisory Panel (The Brine Panel) has suggested amendments to the California Ocean Plan based on a numeric water quality objective that would limit brine discharges from ocean desalination to no more than 5% over ambient ocean salinity at the outer edge of a Regulatory Mixing Zone measuring 100 m (328 ft) in radius around the discharge, (referred to as The 5% Rule). In practice, The 5% Rule is significantly more stringent than Requirement III.C.4(b), and is used as the defining compliance criteria of this study. A question regarding implementation of the 5% Rule which remains uncertain is how ambient salinity is defined. In Monterey Bay, ambient salinity varies seasonally by as much as 8%. The technical approach to evaluating project compliance with discharge limits considers both worst case scenarios at the two potential discharge sites, as well as long term variability of the diffuser-induced dilution fields. Worst case scenarios consider three predominant hydrographic periods of Monterey Bay, namely: 1) the upwelling period, 2) the oceanic period dominated by relaxation states, and 3) the Davidson Current period. The long term analysis is based on 4

5 8,149 modeled generated by feeding the models a matrix of 7 controlling variables assembled from the oceanographic period record. All three hydrographic periods are embedded in this period of record. From the worst case simulations, we conclude that the intermediate discharge site has a slight advantage in dilution performance over the deep discharge site; but has a potential for occasional re-circulation of a small amount of brine at the proposed Tenera intake site (at UTM coordinates N, E). During periods of northward flowing currents (the predominant current direction, particularly during the upwelling period), the brine footprint on the seabed that corresponds to 5% above ambient salinity is 7% smaller at the intermediate site than at the deep water discharge site. This foot print is 2 % smaller at the intermediate site during occasional episodes of southward flowing mean currents (occurring during relaxation and Davidson current periods). This advantage is attributable to differences in wave induced mixing at the two sites, where the scrubbing action of oscillatory waves is more intense at the shallower intermediate site than at the deep water discharge site. These wave mixing effects are especially important in the high-energy wave environment of the Monterey Bay region. The intermediate site also has the advantage of a shorter pipeline run with associated lower capital, operating and maintenance cost. Total head requirements for the 36 inch pipeline and the 5-jet linear diffuser are 14.4 psi over ambient ocean pressure at the depth of the deep water discharge; and 11.1 psi over ambient ocean pressure at the depth of the intermediate discharge. However, during relatively rare and brief periods of southward flowing currents (typically occurring during the relaxation period or Davidson current period), there is less than a 1% chance for a potential re-circulation of 1% of the brine discharged from the intermediate site. The actual re-circulation is likely to be 5

6 6 much less, because the brine plume has dispersed as a thin bottom spreading layer of no more than 5 cm thickness at the proposed Tenera intake site, and the intake structure will probably stand several meters above the seabed. Regardless, the potential for brine re-circulation arising from the intermediate discharge site might tip a cost-benefit analysis of the two sites in favor of the deep discharge site, despite the higher operating pressure requirements of that site. At the intermediate MBRWP discharge site, long-term simulations find the median salinity at the outer limit of the Regulatory Mixing Zone is ppt, well within The 5% Rule requirements when using heated source water from the data center; and 99.9% of the 8,149 modeled solutions are less than or equal to 35.1 ppt (5% over long-term average salinity which is ppt). Maximum salinity at 100 m from the intermediate discharge site in any direction is ppt, exceeding the 5% Rule by only 0.7% when based long term average ambient salinity. (If the 5% Rule based on daily salinity, then this over-limit result exceeds the 5% Rule by only 0.3%). The probability of occurrence of this over-limit case is only 0.1%, or about 1 day in 3 years. The median salinity at the outer limit of the Regulatory Mixing Zone at the deep discharge site is ppt, also well within The 5% Rule requirements when using heated source water from the data center; and 99.8% of the 8,149 modeled solutions are less than or equal to 35.1 ppt (5% over long-term ambient mean). Maximum salinity at 100 m from the deep discharge site in any direction is ppt, exceeding the 5% Rule by only 1.3% when based long term average ambient salinity. (If the 5% Rule based on daily salinity, then this overlimit result exceeds the 5% Rule by only 0.8%). The probability of occurrence of this of this over-limit is only 0.2%, or about 1 day every 1.4 years. The magnitude of these occasional over-limit model results at either discharge site are within sampling error of standard oceanographic

7 temperature/conductivity measurements used for determination of practical salinity units (psu). The probability of recurrence of these over-limit values is equally negligible, on the order of one day every few years. Although NPDES permits are generally written for the specific conditions of each discharger, most allow for occasional over-limit monitoring results, as a provision for sampling errors. In addition, over-limit outcomes at the MBRWP discharge site are even less likely than indicated by the model results because the SWRO facility will operate most of the time under preferred efficiency conditions. The modeled results were obtained for the ultimate maximum brine discharge rate of mgd with 48% recovery; whereas the most efficient operating point for the MBRWP reverseosmosis facilities is at 46% recovery ratio with about 1 mgd less brine discharge than what produced the over-limit model results. Therefore, we conclude the diffuser dilution strategy at both the deep water and intermediate MBRWP discharge sites satisfies any of the presently permitted or potential future dilution standards for all foreseeable long-term ocean conditions at MBRWP Moss Landing. There is the possibility that the data center and the SWRO facilities at MBRWP will not come on-line at the same time, and consequently the source water for SWRO facilities would be unheated during some portion of the initial operating period. Hydrodynamic simulations of brine plume dilution arising from this the start-up scenario find slightly larger, more frequent over-limit discharge results at both the deep and intermediate discharge sites. However the preponderance of outcomes easily satisfies the 5% Rule, and the over-limit cases are still not statistically significant, particularly sense these cases represent start-up transients. 7

8 8 Brine Dilution Analysis for DeepWater Desal, LL, Monterey Bay Regional Water Project at Moss Landing, CA By Scott A. Jenkins, Ph.D. and Joseph Wasyl 1.0) Introduction: This is a hydrodynamic analysis of a diffuser-based discharge strategy prepared for Deepwater Desal LLC, (DWD), who is developing the Monterey Bay Regional Water Project (MBRWP) at Moss Landing, California, (CCRWP, 2013). At full build-out, the MBRWP will discharge an ultimate maximum mgd of heated brine through a sea-floor located diffuser at the end of a dedicated 36 inch diameter pipeline at depth of approximately 25 m to 35 m, approximately 1 to 1.5 miles offshore at the head of the Monterey Bay submarine canyon (Figure 1). The brine is the by-product of a seawater reverse osmosis desalination ("SWRO") plant at the MBRWP, operating at 48% maximum recovery and producing discharge concentrate having an average end-of-pipe salinity of ppt. (These figures represent ultimate maximum brine source loading to the receiving waters. When operating at preferred efficiency, the SWRO will operate at 46% recovery, producing 26.1 mgd of brine with an average end-of-pipe salinity of 63.7 ppt, see Table 1 ). The salinity of the brine may vary by as much as 8% from this average level, based on long-term variations of ambient ocean salinity in Monterey Bay shown in Figure-2. Most of this variation is due to salinity depression occurring during floods of the regional rivers, principally the Pajaro and Salinas Rivers (Figure 2). The temperature of the brine effluent will be heated to 20 0 C over ambient, ( 20 0 C) because the source water is used to cool a data center before entering the = Tb

9 Figure 1: Bathymetry of Monterey Bay in meters MSL, showing approximate location of the diffusers for the Deepwater Desal Monterey Bay Regional Water Project ("MBRWP") at Moss Landing. 9

10 Figure 2: Long-term salinity variation typical of the Central California, Monterey Bay. 10

11 95, 96, 105 SWRO plant. Temperature data from nearby Tenera and MBARI moorings show pronounced seasonal variations in ambient ocean temperatures. The maximum recorded daily mean temperature was o C during the summer of the 1997 El Niño and the minimum falling to 9.25 o C during the winter of the La Niña. The mean ocean temperature at 30 m depth is o C. On a percentage basis, the natural variability of the temperature of coastal waters in the vicinity of the MBRWP is significantly greater than that of salinity (on the order of T / T = 72% vs. S / S = 8%). These SWRO operating points and ambient ocean properties determine the nature of the brine plume, i.e., whether it sinks or floats once discharged into the receiving water. If the brine plume floats, its density is less than the receiving water, and it is referred to as a buoyant plume. If it sinks, the brine is denser than the receiving water, and it is referred to as a dense plume. The density of the receiving water and the density of the brine discharge b are both a function of temperature, T, and salinity, S, and the density contrast between these two,, will determine the net buoyancy of the plume and control its dynamics in the nearfield of the diffuser. Delta-T and Delta-S effects on density of the brine plume are quantified by the equation of state expressed in terms of the specific volume, 1/ b b or: b 11 d b 1 1 dtb ds T S b (1) The factor / T, which multiplies the differential temperature changes, is known as the coefficient of thermal expansion and is typically 2 x 10-4 per o C for seawater;

12 the factor / S multiplying the differential salinity changes, is the coefficient of saline contraction and is typically -8 x 10-4 per part per thousand (ppt) where 1.0 ppt = 1.0 g/l of total dissolved solids (TDS). For a standard seawater, the specific volume has a value = cm 3 /g and for Monterey Bay water the specific volume averages about cm 3 /g. If the percent change in specific volume by equation (1) is less than zero, then the brine is heavier than the receiving water, and lighter if the percent change is greater than zero. The heat load generated by the data center will likely be insufficient to achieve more than a Tb = 8 0 C to 10 0 C. This will result in a brine discharge temperature of nominallyt b = 16 0 C resulting from 35k mg/l feed concentration at 15 0 C with R.O. recovery ranging between 42%, 46%, and 48% for year 0. Table 1 indicates that the brine salinity will range between 59.3 ppt and 66.2 ppt for these R.O. recoveries, while the density of the brine will range between g/cm 3 and g/cm 3 when it exits the diffuser nozzles, and will be about 1.9 % to 2.4% denser than the receiving water, which typically has a density of g/cm 3. Consequently the brine plume will be a dense plume that will sink in the receiving water, based on the projected operating points. 12 Table 1: Brine Salinity and Density for Projected Operating Points SWRO Recovery, % Brine Salinity, ppt* Brine Density, g/kg, 16C Yr 0 Pressure, psi 42% , % , % , * practical salinity units in part per thousand, Includes isobaric energy recovery, ERI PX seawater 10C, 35 ppt = 1,027 g/kg seawater 60C, 63.7 ppt = roughly 1,027 g/kg

13 13 2.0) Discharge and Dilution Compliance Issues Hyper-salinity toxicity: This is the primary regulatory driver for how much initial dilution by turbulent mixing and entrainment the diffuser should be designed to achieve. The Project Description submitted by Deepwater Desal, LLC, to the CPUC suggests an initial dilution of 38 to 1.This is probably overly aggressive and requiring more powerful diffuser jets and higher operating pressures than necessary to meet either present or potential discharge limits on brine discharge. Recently, the California State Water Resources Control Board convened a 5- member science advisory panel (The Brine Panel) to produce a study that could provide a scientific basis for amending The California Ocean Plan in order to establish water quality objectives for brine discharges from ocean desalination plants. 1 The Brine Panel concluded that there is surprisingly limited scientific data on hyper-salinity tolerance of marine organisms on which a regulatory standard could be based at any particular site. Hyper salinity tolerance is extremely dependent on the mix of locally relevant species and on the seasonal variability of ambient ocean salinity. The hyper salinity tolerance data that does exist primarily focuses on short term lethal exposures (acute toxicity) using a laboratory method referred to as Whole Effluent Toxicity Testing (WET); while data on longer term sub-lethal exposure (chronic toxicity) is virtually non-existent (Sub-lethal exposures cause injuries that typically limit growth or reproduction). Given this paucity of data, The Brine Panel concluded that until additional data became available, it is better to err on the side of caution with a very conservative water quality objective that would limit brine discharges to 5% over ambient ocean salinity at the limit of a Regulatory Mixing Zone measuring 100 m (330 ft) in radius around the discharge point (referred to as The 5% Rule). 1 However the

14 scientific basis for this cautious recommendation was hyper-salinity tolerance data for sea grasses living in territorial waters of Spain, 9, 14, plants with associated marine communities that do not live in California waters. WET lab testing on locally relevant species in Southern California found chronic toxicity thresholds to be in the neighborhood of 20% over ambient ocean salinity. 22,51,52,55 Regardless, a question still remains regarding how ambient salinity should be defined in the implementation of the 5% Rule; whether ambient salinity is the long term average value for the receiving water, or an instantaneous (daily) measurement. We will consider the implications of both definitions in our dilution analysis. In the absence of clear scientific consensus on the appropriate hyper-salinity tolerance levels for the Monterey Bay marine communities, we proceed in the present analysis by addressing two regulatory discharge compliance questions in parallel: 1) Will the discharge strategy of MBRWP satisfy the 0.3 TUa objective of Requirement III.C.4(b) of the present version of the California Ocean Plan as it would apply to a Zone of Initial Dilution (ZID)? This was the dilution standard on which discharge permits were issued to the Carlsbad and Huntington Beach desalination projects by the Regional Water Quality Control Boards of the San Diego and Santa Ana Regions. And, 2) will the MBRWP discharge strategy satisfy suggested amendments to the California Ocean Plan based on the 5% rule? In the present study we seek the appropriate initial dilution parameters to be applied to the MBRWP diffuser design such that both of these regulatory standards are met without subjecting the marine life that is entrained by the diffuser jets (principally eggs larvae and juvenile fish) to excessive turbulent shear stress and strain rates (see APPENDIX-A). Generally, for moderate RO recovery ratios, (less than 50%), the 5% Rule is more stringent. However, for very high end-of-pipe discharge 14

15 salinities (greater than double concentrated) the present 0.3 TUa objective of Requirement III.C.4(b) is more challenging for compliance ) Technical Approach The technical approach evaluates worst case scenarios and long term variability of dilution fields associated with the MBRWP. This analysis involved the use of hydrodynamic transport models driven by historic wave, current, and wind and water mass data for known events. The production of brine by-product by the proposed desalination demonstration facility was overlaid on these events to determine the potential range of variability in dilution outcomes. Analysis of brine dilution issues uses a combination of hydrodynamic models: a near field mixing zone models, Visual Plumes certified by the U.S. Environmental Protection Agency and the California State Water Resources Control Board for use in ocean outfall design; and a fully 3-dimensional far field dispersion model SEDXPORT that is a processed-based stratified flow model with the complete set of littoral transport physics including tidal transport, and wind &wave induced transport and mixing, (cf. Jenkins and Wasyl 2007, 2008). SEDXPORT is used to predict the trajectory of the brine plume following initial dilution in the nearfield of the diffuser. SEDXPORT modeling system has been extensively peer reviewed by 8 independent experts and can be found in the public records of the State Water Resources Control Board, the California Coastal Commission and the Cities of Carlsbad, Huntington Beach and Moss Landing. SEDXPORT was also employed in the dilution studies for desalination projects by Los Angeles Department of Water and Power (Jenkins and Wasyl, 2005) and by the West Basin Municipal Water District (Jenkins and Wasyl, 2008b). SEDXPORT has also been incorporated into the Coastal Evolution Model available on-line at the Digital

16 Library of the University of California (Jenkins and Wasyl, 2005b; Additional details of the SEDXPORT model along with many of its important algorithms are found below. Analysis of diffuser turbulence mortality and turbidity issues issues will use a third type of hydrodynamic model belonging to a class of models known as computational fluid dynamics (CFD). We use three different CFD models: 1) The basic hydraulics of the offshore discharge pipeline, riser and diffuser internal flow simulations were performed using commercially available hydraulics design software known as COSMOS/ FLowWorks. 2) The subsequent turbulence kinetics 16 of the discharge plume was evaluated using a v 2 f mode computational fluid dynamics model, Star-CD, Version 3.1, with QUICK space discretization for the mean flow and first order up-winding of the turbulence equations. 70,71 It was used herein to compute the jet core velocities, shear stress, strain rates. 3) Diffuser induced seedbed scour and sediment re-suspension processes are modeled with the Vortex Lattice Scour Model. This process-based model was developed at Scripps Institution of Oceanography for the Office of Naval Research, and its algorithms have been published in the peer-reviewed literature (Jenkins and Inman, 2006; Jenkins et al. 2007). Because wave and current transport processes are predominant in the analysis of brine dilution and dispersion at the Moss Landing site, SEDXPORT is the primary analysis tool. SEDXPORT has been built in a modular computational architecture (Jenkins and Wasyl, 2007). The modules are divided into two major clusters: 1) those which prescribe hydrodynamic forcing functions; and, 2) those which prescribe the mass sources acted upon by the hydrodynamic forcing to produce dispersion and transport. The cluster of modules for hydrodynamic

17 forcing ultimately prescribes the velocities and diffusivities induced by wind, waves, and tidal flow for each depth increment at each node in the grid network. The finite element current model, TIDE_FEM, (Jenkins and Wasyl, 1990; Inman and Jenkins, 1996) will be employed to evaluate the tidal currents at Moss Landing. TIDE_FEM was built from some well-studied and proven computational methods and numerical architecture that have done well in predicting shallow water tidal propagation in Massachusetts Bay (Connor and Wang, 1974) and along the coast of Rhode Island, (Wang, 1975), and have been reviewed in basic text books (Weiyan, 1992) and symposia on the subject, e.g., Gallagher (1981). TIDE_FEM employs a variant of the vertically integrated equations for shallow water tidal propagation after Connor and Wang (1975). These are based upon the Boussinesq approximations with Chezy friction and Manning s roughness. The finite element discretization is based upon the commonly used Galerkin weighted residual method to specify integral functionals that are minimized in each finite element domain using a variational scheme, see Gallagher (1981). Time integration is based upon the simple trapezoidal rule (Gallagher, 1981). The computational architecture of TIDE_FEM is adapted from Wang (1975), whereby a transformation from a global coordinate system to a natural coordinate system based on the unit triangle is used to reduce the weighted residuals to a set of order-one ordinary differential equations with constant coefficients. These coefficients (influence coefficients) are posed in terms of a shape function derived from the natural coordinates of each nodal point in the computational grid. The resulting systems of equations are assembled and coded as banded matrices and subsequently solved by Cholesky s method, see Oden and Oliveira (1973) and Boas (1966). The hydrodynamic forcing used by TIDE_FEM 17

18 18 is based upon inputs of the tidal constituents derived from Fourier decomposition of tide gage records. Tidal constituents are input into the module TID_DAYS, which resides in the hydrodynamic forcing function cluster. TID_DAYS computes the distribution of sea surface elevation variations in Monterey Bay based on the tidal constituents derived from the tide gage station at Santa Barbara, NOAA # Forcing for TIDE_FEM is applied by the distribution in sea surface elevation across the deep water boundary of the computational domain. Wave driven currents will be calculated from wave measurements by the Coastal Data Information Program (CDIP) arrays and/or buoys. These measurements will be back refracted out to deep water to correct for propagation and shoaling effects between the monitoring sites and Moss Landing. The waves were then forward refracted onshore to give the variation in wave heights, wave lengths and directions throughout the nearshore around the Moss Landing coast. The numerical refraction-diffraction code used for both the back refraction from these wave monitoring sites out to deep water, and the forward refraction to the Moss Landing coast site is OCEANRDS. This code calculates the simultaneous refraction and diffraction patterns of the swell and wind wave components propagating over bathymetry replicated by the OCEANBAT code. OCEANBAT generates the associated depth fields for the computational grid networks of both TID_FEM and OCEANRDS using packed bathymetry data files derived from the National Ocean Survey (NOS) depth soundings compiled by GEODAS. The structured depth files written by OCEANBAT are then throughput to the module OCEANRDS, which performs a refraction-diffraction analysis from deep water wave statistics. OCEANRDS computes local wave heights, wave numbers, and directions for the swell component of a two-component, rectangular spectrum.

19 19 The wave data are throughput to a wave current algorithm in SEDXPORT which calculates the wave-driven longshore currents, v(r). These currents were linearly superimposed on the tidal current. The wave-driven longshore velocity, v(r), is determined from the longshore current theories of Longuet-Higgins (1970). Once the tidal and wave driven currents are resolved by TIDE_FEM and OCEANRDS, the dilution and dispersion of brine and backwash constituents is computed by the stratified transport algorithms in SEDXPORT. The SEDXPORT code is a time stepped finite element model which solves the advection-diffusion equations over a fully configurable 3-dimensional grid. The vertical dimension is treated as a two-layer ocean, with a surface mixed layer and a bottom layer separated by a pycnocline interface. The code accepts any arbitrary density and velocity contrast between the mixed layer and bottom layer that satisfies the Richardson number stability criteria and composite Froude number condition of hydraulic state. The SEDXPORT codes do not time split advection and diffusion calculations, and will compute additional advective field effects arising from spatial gradients in eddy diffusivity, (the so-called gradient eddy diffusivity velocities after Armi, 1979). Eddy mass diffusivities are calculated from momentum diffusivities by means of a series of Peclet number corrections based upon TSS and TDS mass and upon the mixing source. Peclet number corrections for the surface and bottom boundary layers are derived from the work of Stommel (1949) with modifications after Nielsen (1979), Jensen and Carlson (1976), and Jenkins and Wasyl (1990). Peclet number correction for the wind-induced mixed layer diffusivities are calculated from algorithms developed by Martin and Meiburg (1994), while Peclet number corrections to the interfacial shear at the pycnocline are derived from Lazara and Lasheras (1992a;1992b). The momentum

20 diffusivities to which these Peclet number corrections are applied are due to Thorade (1914), Schmidt (1917), Durst (1924), and Newman (1952) for the windinduced mixed layer turbulence and to Stommel (1949) and List, et al. (1990) for the current-induced turbulence. The proposed MBRWP discharges a combined effluent of heat from a data center and concentrated seawater from a sea water reverse osmosis facility. We must solve for the dispersion and dilution of both heat and salt. For the sea salts, SEDXPORT solves the eddy gradient form of the advection diffusion equation for the water column density field: 20 u t 2 0V ( t) da (2) where u is the vector velocity from a linear combination of the wave and tidal currents, is the mass diffusivity, is the vector gradient operator and is the water mass density in the nearshore dilution field; and 0 is the density of the water discharged by the discharge at a flow rate V (t). The density of the discharge is a function of the salt and heat flux from the data center, where the heat flux is given as: Q 2 u Q H Q H Q Q0V ( t) da t dq C p dt where H is the heat diffusivity and C p is the heat capacity of seawater at constant pressure. and Q 0 is the heat discharged by the data center at a flow rate (t) (3) V.

21 Solutions to the density field of the discharge plume from the outfall are calculated from equations (1-3) by SEDXPORT, from which computations of local discharge salinity, S ( x, y, z), can be made using equation (1). The salinity field of the discharge plume can be used to solve for the dilution factor D BWW ( x, y, z) of the combined brine/ thermal effluent from the data center according to: 21 D BWW ( x, y, z) S o S o (4) S( x, y, z) where S o is the ambient seawater salinity in ppt, and S ( x, y, z) is the local salinity in the discharge plume from the model solution in ppt. Model solutions will find a significant variation in the salinity with water depth, z. Therefore we introduced a depth averaged dilution factor, H 1 D( x, y, H) D( x, y, z) dz (5) H( x, y) 0 Where H H( x, y) h is the local water depth, h is the local water depth below mean sea level and is the tidal amplitude. Solutions for the density and concentration fields calculated by the SEDXPORT codes from equations (1)-(3), are through put to the dilution codes of MULTINODE to resolve dilution factors according to (4) and (5). These codes solve for the dilution factor (mixing ratio) for each cell in the finite element mesh of the nearshore computational domain based on a mass balance between imported exported and resident mass of that cell. The diffusivity,, in (1) controls the

22 strength of mixing and dilution of the seawater and storm water constituents in each cell and varies with position in the water column relative to the pycnocline interface. Vertical mixing includes two mixing mechanisms at depths above and below the pycnocline: 1) fossil turbulence from the bottom boundary layer, and 2) wind mixing in the surface mixed layer. The pycnocline depth is treated as a zone of hindered mixing and varies in response to the wind speed and duration. Below the pycnocline, only turbulence from the bottom wave/current boundary layer contributes to the local diffusivity. In the nearshore, breaking wave activity also contributes to mixing. The surf zone (zone of initial dilution) is treated as a line source of turbulent kinetic energy by the subroutine SURXPORT. This subroutine calculates seaward mixing from fossil surf zone turbulence, and seaward advection from rip currents embedded in the line source. Both the eddy diffusivity of the line source and the strength and position of the embedded rip currents are computed from the shoaling wave parameters evaluated at the breakpoint, as throughput of OCEANRDS. Altogether there are eight primary variables that enter into the SEDXPORT solutions for the simultaneous dispersion and dilution of the brine and backwash constituents from the MBRWP. These eight variables may be organized into forcing functions and boundary conditions. The forcing function variables affect the strength of ocean mixing, ventilation and available dilution volume in shallow water. These include: * Waves * Ocean Water Levels (tides and sea level anomalies) * Currents * Winds 22

23 The boundary condition variables control the source strength (concentrated sea salts) and background conditions. Some of these variables change daily (primary boundary conditions), while others vary slowly in time (stationary boundary conditions). The primary boundary conditions are: * Combine Flow Rates of data center and SWRO facility * Ocean Salinity * Discharge Temperature * Ocean Temperature The local bathymetry typically has a seasonal variation inshore of closure depth (about 15 m depth). In the following sub-sections, concurrent 24 year long records (the longest available period of record for all of the eight controlling variables) are reconstructed. These long-term records contain 8,149 consecutive days of daily mean values between 1983 and 2008, depending on the number of unfilled data gaps. Certain of these data sets do not span the period of record. In this case the known data were folded back over time, (preserving seasonality in the folding process), in order to obtain full-length 24 year long time series to be input to the model ) Model Initialization Long-term monitoring of ocean properties in the coastal waters surrounding Moss Landing and the greater Monterey Bay has been on going for about 30 years. These data were accessed from a variety of data archives, including a few NPDES monitoring reports filed with the Regional Water Quality Control Board; the California Coastal Water Quality Monitoring Inventory maintained by the California Department of Water Resources; the Coastal Data Information Program

24 (CDIP); the United States Geological Survey, (USGS), National Water Information System; the California Data Exchange Center; National Geophysical Data Center (NGDC); National Oceanic and Atmospheric Association (NOAA), National Data Buoy Center; the National Ocean Service Water Level Observation Network; Monterey Bay Aquarium Research Institute (MBARI); and the California Spatial Information Library. In attempting to reconstruct 24-year long, continuous, unbroken records of all the controlling variables for the dilution and dispersion modeling problem, certain gaps were found in some of the data bases obtained from these sources. These gaps were filled either by developing proxy records from surrogate sites having similar coastal morphology, or by folding the existing records over the gaps while preserving characteristic seasonality ) Bathymetry: Bathymetry provides a controlling influence on all of the coastal processes that affect dispersion and dilution. The bathymetry consists of two parts: 1) a stationary component in the offshore where depths are roughly invariant over time; and 2) a non-stationary component in the nearshore where depth variations do occur over time. The stationary bathymetry generally prevails at depths that exceed closure depth, which is the depth at which net on/offshore sediment transport vanishes. Closure depth is typically -12 m to -15 m MSL in the Monterey Littoral Cell, [Inman et al. 1993]. The stationary bathymetry was derived from the National Geophysical Data Center GEODAS digital database. Gridding is by latitude and longitude with a 3 x 3 arc second grid cell resolution in a 1201 X 1201 farfield grid yielding an xy-computational domain of km x 86.5 km. The farfield grid mesh is indicated by the black lines in Figure 3. A nearfield grid was constructed with a 1 x 1 arc second grid cell resolution in a 601 X 601 numerical array, yielding an xy-computational domain of km x 18.57

25 Figure 3: Farfield grid with 3 x 3 arc second grid cell resolution in a 1201 X 1201 farfield grid yielding an xy-computational domain of km x 86.5 km. 25

26 26 km,(figure 4). The farfield grid (Figure 3) computes the effects of regional scale refraction and circulation due to the shallow banks of the Bay and inner continental margin (Figure 5). The nearfield grid (Figures 4) is nested inside the farfield grid and is used to calculate the brine dilution and dispersion and potential recirculation to the intake. For the non-stationary bathymetry data inshore of closure depth (less than -15 m MSL) nearshore and beach surveys were conducted by the US Army Corps of Engineers in 1985, 1990, and These surveys were used to update the GEODAS database for contemporary nearshore and changes. 4.2) Intake and Discharge Structures: The MBRWP will utilize a refurbished, existing oil pipeline for the Moss Landing Power Plant (MLPP) (Figure 6). A new intake site is proposed by Tenera 105 at UTM coordinates: N, E. The intake pipeline will convey source water from this location to an existing onshore wet well located at the MLPP. From the wet well the source water will be pumped to an approximately 110-acre site (referred to as the Tank Farm Parcel ) located along Dolan Road, east of the MLPP, where the SWRO plant and data centers will be constructed. The discharge pipeline for brine concentrate from the SWRO plant and raw seawater from other uses in the Project, will be routed via a new pipeline from the site to a submerged discharge below the euphotic zone using the same approximate route as the existing intake pipeline (Figures 4 and 6). At full build-out, the MBRWP will discharge mgd of heated brine through a sea-floor located diffuser at the end of new pipeline, which is sized to be 36 inch diameter and extend to either one of two potential discharge sites. The two discharge sites are located along the existing MLPP oil pipeline

27 Figure 4: Nearfield computational domain. Nearfield grid constructed with a 1 x 1 arc second grid cell resolution in a 601 X 601 numerical array yielding an xycomputational domain of km x km.utm coordinates in meters are shown for the deep and intermediate discharge sites along the existing MLPP oil pipeline shown in red. 27

28 Figure 5: Farfield refraction/diffraction computation for Monterey Bay province for the storm of, 5 December Deep water significant wave height = 5.78 m; significant period = 18.2 sec; predominant direction = 265 deg. (CDIP, 2008). 28

29 29 route, one located in 35 m of local water depth (deep discharge site) at UTM coordinates N, E; while the other is located in 25 m of water depth (intermediate discharge site) at UTM coordinates N, E, see Figure 4. Pipeline run to the deep water discharge site is approximately 2,416 m (7,927 ft.), as measured from the shoreline; while the pipeline run to the shallower intermediate site is 1,738 m (5,703 ft.). Estimated frictional losses in the pipeline to the deep water site are 11.9 psi; while pipeline head losses along the shorter run to the intermediate discharge site are 8.6 psi. The brine/thermal effluent will be discharged via diffusers to assure rapid and thorough mixing with ambient seawater. We pose a five-jet linear diffuser located orthogonal to the -25m and -35m MSL isobaths on the north bank of the Monterrey Submarine Canyon. The linear diffuser consists of five discharge riser/diffuser structures at 10 m spacing that extend above the seabed from a manifold pipe. To minimize the likelihood of snagging drifting kelp and bottom debris, and to reduce seabed scour around the manifold and discharge riser/diffuser structures, it would be advisable to bury the manifold under the seabed. For the same reasons, this design employs the fewest numbers of discharge riser/diffuser structures that are needed for adequate dilution performance without incurring excessive turbulence mortality (see APPENDIX-A). A physical example of such a linear diffuser is shown in Figure 7. The discharge manifold consists of a 54 inch diameter pipe buried below the seafloor (Figure 8) that delivers a maximum of 5.45 mgd the brine to each of 5 diffuser risers. Each riser is fitted with a Tideflex duckbill nozzle angled upward at a 60 degree angle. The duckbill nozzles are selfadjusting to variable flow rate to maintain optimal jet nozzle diameter, and each

30 Figure 6: Location of the abandoned fuel oil pipeline. The proposed route of brine discharge pipeline is shown in red and the feedwater intake pipeline is shown in green, terminating at an intake site proposed by Tenera,

31 Figure 7: Physical example of the proposed linear diffuser concept posed for the DeepWater Desal, LLC, Monterey Bay Regional Water Project at Moss Landing, CA. Five discharge riser at 10 m spacing will extend above the seabed from a buried manifold pipe. Each riser pipe is fitted with a Tideflex duckbill nozzle. 31

32 Figure 8. Dimensional drawing and 3-d SolidWorks model of one of the five discharge riser/diffuser structure to be used in the DeepWater Desal, LLC, Monterey Bay Regional Water Project at Moss Landing, CA. 32

33 duckbill stands 7 ft above the seafloor atop it s riser (Figure 8) to isolate the discharge nozzles from burial effects and to protect against damage from bottom debris moving about in the wave and current surge. The Tideflex duckbill nozzles are sized converging restrictors inside the 16 inch diameter riser pipes in order to produce discharge velocities of u 0 = 5 m/s. The engineering drawings of the discharge riser/diffuser structures in Figure 8 were gridded into a series of lattice panels to form a nearfield grid. Figure 9 shows a cross-section simulation of a discharge flow of mgd design passing (into the page) through the buried 36-inch feeder pipe that distributes 5.45 mgd to the Port-1 riser/diffuser structures. Maximum discharge velocity is 5 m/sec (16.4 ft/sec). Total head requirements for the 36 inch pipeline and the 5-jet linear diffuser are 14.4 psi over ambient ocean pressure at the depth of the deep water discharge; and 11.1 psi over ambient ocean pressure at the depth of the intermediate discharge. Some head losses are apparent at the junction of the feeder pipe with the riser pipe. The simulation in Figure 9 was based on discharge into a perfectly quiet ocean with no waves or current motion in the receiving water. 33

34 Figure 9. Still water simulation of discharge flow through riser/diffuser structure proposed for the DeepWater Desal, LLC, Monterey Bay Regional Water Project at Moss Landing, CA. Discharge flow for Port-1 of the mgd design. 34

35 35 4.3) Wave Climate: Waves are the principle driving mechanism of mixing and current ventilation in the very near-shore region off Moss Landing. This wave dominated region consists primarily of the surfzone but extends seaward into the wave shoaling zone a few surf zone widths beyond the point of wave breaking. The waves also create local mixing due to the orbital motion around the intake and discharge structures. Waves are among the most difficult of the controlling variables to get long unbroken records. The availability of wave data in the Monterey Bay and Central California region limited the period of record for this long term model analysis to Waves have been routinely monitored at several local and regional locations by the Coastal Data Information Program, (CDIP, 2008). These data are supplemented by ocean observations with Datawell buoys by the National Data Buoy Center (NDBC). The nearest CDIP and NDBC directional wave monitoring sites are: a) CDIP, Harvest, California Station ID: 071 Location: N; W Water Depth: 549 m b) CDIP, Point Reyes, California Station ID: 029 Location: N; W Water Depth: 301 m c) CDIP, Coquille River Inner, Oregon Station ID: 037 Location: N; W Water Depth: 64 m

36 d) National Data Buoy Center, Monterey, CA Station Location: N; W Site elevation: sea level Water depth: 2115 m 36 Air temp height: 4 m above site elevation Anemometer height: 5 m above site elevation Barometer elevation: sea level Sea temp depth: 0.6 m below site elevation These data sets possessed gaps at various times due to system failure and a variety of start ups and shut downs due to program funding and maintenance. The undivided data sets were pieced together into a continuous record from and entered into a structured preliminary data file. The data in the preliminary file represent partially shoaled wave data specific to the local bathymetry around each monitoring site. To correct these data to the nearshore of Moss Landing, they are entered into a refraction/diffraction numerical code, back-refracted out into deep water to correct for local refraction and bay sheltering, and subsequently forward refracted into the immediate vicinity of MBRWP and Moss Landing. Hence, wave data off each monitoring site was used to hindcast the waves at MBRWP and Moss Landing coastal region. The backward and forward refractions of CDIP data to correct it to MBRWP and Moss Landing were done using the numerical refraction-diffraction computer code, OCEANRDS. The primitive equations for this code are lengthy, so a listing of the FORTRAN codes of OCEANRDS appear in Jenkins and Wasyl, (2005), Appendix B. These codes calculate the simultaneous refraction and diffraction patterns propagating over a Cartesian depth grid. A regional outer grid (Figure 3) was used in the back refraction calculations to correct for continental margin effects, while a high resolution inner grid (Figure 4) was used for the forward

37 refraction over the local bathymetry at the MBRWP and the nearfield of Moss Landing (Figure 10). OCEANRDS uses the parabolic equation method (PEM), Radder (1979), applied to the mild-slope equation, Berkhoff (1972). To account for very wideangle refraction and diffraction relative to the principle wave direction, OCEANRDS also incorporates the high order PEM Pade approximate corrections modified from those developed by Kirby (1986a-c). Unlike the recently developed REF/DIF model due to Dalrymple, et al. (1984), the Pade approximates in OCEANRDS are written in tesseral harmonics, per Jenkins and Inman (1985); in some instances improving resolution of diffraction patterns associated with steep, highly variable bathymetry such as found near the Monterey Submarine Canyon (Figure 10). These refinements allow calculation of the evolution and propagation of directional modes from a single incident wave direction; which is a distinct advantage over the more conventional directionally integrated ray methods which are prone to caustics (crossing wave rays) and other singularities in the solution domain where bathymetry varies rapidly over several wavelengths. An example of a reconstruction of the wave field throughout the Monterey continental margin region is shown in Figure 5 using the back refraction calculation of the CDIP data from the NDBC buoy # Wave heights are contoured in meters according to the color bar scale and represent 6 hour averages, not an instantaneous snapshot of the sea surface elevation. Note how the shelf bathymetry induced longshore variations in wave height throughout the Monterey Bay region. Figure 11 shows the significant wave heights inside Monterey Bay offshore of Moss Landing, with corresponding periods and directions, resulting from the series of back-refraction calculations for the complete CDIP and NBDC data sets at Δt = 6 hour intervals over the period of record. The data in 37

38 38 Figure 11 are values used as the deep water boundary conditions on the nearfield grid (Figure 4) for the forward refraction computations into the MBRWP and Moss Landing Harbor region (like those in Figure 10). The deep water wave angles in Figure 11c are plotted with respect to the direction (relative to true north) from which the waves are propagating at the deep water boundary of the nearfield grid (Figure 4). Inspection of Figure 11a reveals that a number of large swells of 6 m to 8 m heights have affected the Monterey Bay and Moss Landing region nearly every year during the period of record. Waves of this size create powerful wave induced currents and vigorous mixing when scrubbing over the seabed and intake and discharge structures. The largest local swells computed from our analysis of the period of record resulted from the storm of 5 December 2007 that produced 10 m high shoaling waves at various locations in and around Moss Landing and the greater Monterey Bay region. (Figures 5 and 10) Examining Figure 10 in closure detail, it is noted how the refraction effects over local bathymetry create areas to north and south of the discharge where heights increase to 7.4 m. In these areas, the bay and shelf bathymetry has focused the incident wave energy and these regions of intensified wave energy are referred to as bright spots. The increased wave heights in these bright spots increase the mixing and turbulence generated over the seabed boundary layer and by oscillatory wakes of the intake riser and discharge diffuser structures. This increases the mixing and dilution rates of the brine that might spread into the bright spots. Conversely, the dark areas in Figure 10 in the immediate nearfield of the intake and discharge are areas where wave heights have been diminished and are termed shadows. The shadow area over the intake and discharge sites at Moss Landing is caused by the refraction effects of the Monterey Submarine Canyon. The waves propagate onshore up the deep water center axis of the canyon