Far-Field Impacts of Tidal Energy Extraction and Sea Level Rise in the Gulf of Maine

The dynamics of tides in the Gulf of Maine are unique due to the tidal resonance, which generates the largest tidal range in the world (about 16 m). Consequently, a large tidal energy resource is available in this area, particularly in the Bay of Fundy, and is expected to be harvested in the future. Currently, more than 6 projects are operational or under development in this region (in both US and Canadian waters). Understanding the far-field impacts of tidal-stream arrays is important for future development of tidal energy extraction. The impacts include possible changes in water elevation, currents, and sediment transport. Accordingly, a number of previous studies have assessed the impacts of the tidal energy development in the Gulf of Maine. Further, due to the sea level rise (SLR), those impacts may also change during the project lifetime, which is usually more than 25 years. The objective of this study is to assess the combined effects of SLR and tidal energy extraction on the dynamics of tides in the Gulf of Maine. A tidal model of the Gulf of Maine was developed using Regional Ocean Model System (ROMS) at one arcminute scale. The model extends from 71.5W to 63.0W and from 39.5N to 46.0N. After validation of the model at NOAA tidal gauge stations and NERACOOS buoys, several scenarios; including SLR scenario, and tidal extraction scenario, were examined. Recent studies suggest that the global dynamics of tides will change due to SLR; therefore, SLR not only affects the bathymetry of the model inside the domain, it also changes the boundary forcing, which was considered in this effort. The results of the impacts of the tidal energy extraction with and without the SLR were presented, and compared with those from literature. Up to 4% decrease in tidal range and M2 amplitude was estimated in Minas Basin due to the 2.5 GW extraction scenario without SLR. On Massachusetts coastal area, the impacts of the same scenario can be considered negligible, 0.94%. In summary, the implementation of modified boundary forcing due to SLR, which was ignored in the previous works, can change the results of the impact assessment. Based on the results, the far-field impact is more threatening in coastal regions of US. However, the impact of energy extraction in Minas Passage is relatively small. Compared to the model validation, the impacts were inside the uncertainty level of the model. For example, maximum change in Boston coastal area was calculated up to 1.65 %, which is inside the level of uncertainty in models, about 10 %. Furthermore, the impact of SLR on the dynamics of tides is much more than energy extraction assuming 2.5 GW extraction in Minas Passage.

Ocean renewable energy resources (e.g, tidal range and tidal-stream) can help reduce carbon emissions, which are produced by fossil fuel based power plants . Currently, ocean renewable energy extraction is in the development phase from prototype design into commercial power generation.
Tidal energy generation is highly site-specific and generally is feasible where the tidal range and/or current velocity are large enough due to the ocean environment such as amplification by the sea bottom profile, estuaries profile, reflections by large peninsulas and headlands, and resonance effects . The Gulf of Maine, which is located in the north east of America continent, has a vast amount of energy due to the resonance effect. Previous studies in this area have explored the available tidal energy resource and also some researchers have assessed the impacts of energy extraction on the marine environment. Additionally, recent studies (e.g., Nicholls and Cazenave, 2010 ) shows the importance of the SLR scenario in ocean modeling due to its potential impacts on ocean dynamics. In this research, the impact of tidal energy extraction considering SLR was simulated to predict future change in the dynamics of tides.  . Tidal stream technology has been inspired largely by the wind turbine technology. TEC is categorized into horizontal and vertical axis turbines (Figure 1.   Present day turbine technology, which is designed for ∼2.5 m/s maximum tidal current velocity and water depth ranging from 25 and 50 m, are categorized as the first generation of TEC and are expected to lead tidal-stream energy generation within the next 10 years . A recent study by Lewis et al. in 2015  Illustration of tidal-stream turbine arrays .

Area of study
considered future TEC generations based on maximum tidal current velocity and water depth in their simulation. In the cited study, second and third generation of TEC are expected to aim towards a lower tidal current velocity limit. In details, limits for those turbines are: first generation (velocity > 2.5 m/s 25 < h < 50); TEC must be placed in a specific array configuration in order to optimize energy extraction. Wake effects from the blades disturb water flow in an array.
Thus, TEC array optimization must consider the wake effects to maximize energy extraction. Figure 1.4 illustrates possible tidal-stream turbine arrays in the ocean: single turbine, line array, staggered array and random (grid based) array. A recent study by Divett in 2013 suggested the staggered array as the most optimal array configuration, which 54% more efficient compared to other configurations . In the cited study, the distance between turbines was suggested as 7.5 and 10 times TEC blade diameter for the across and along the flow field, respectively.  Previous studies  have suggested that the Bay of Fundy and the Gulf of Maine are a unified system that produce resonance due to their geographical configurations. Consequently, the Gulf of Maine has a high potential for ocean renewable energy. A map of Gulf of Maine and previous research results of maximum power generation are shown in Figure 1.6, more details are provided in Table 1.3 for each site.
In terms of tidal energy resource assessment, several locations for both tidal range and tidal-stream energy generation have been explored by past studies as shown in Figure 1.6. Annapolis Tidal Power Station has been supplying 50 GWh annual electricity productions for Canada since the 1980's   , and 2-6.5 kW/m 2 was simulated on the Kennebec River .

Physical impacts of tidal energy extraction
Extracting energy from the water column will cause changes in the dynamics of the ocean. In the Gulf of Maine, any change that occurs in the dynamics of tides will create effects in the far field  due to basin's resonance.  Table  1.3 for details. Redrawn from the original images at /www.necwa.org Several methods have been used in the literature to simulate TEC in ocean models, such as the increasing bottom drag coefficient method and actuator disc theory .  modeled tidal-stream turbines at ∼10km 2 area in Minas Passage to set up a 6.95 GW power extraction using the additional bottom friction in Finite-Volume Community Ocean Model (FVCOM). The cited study predicted a decrease of tidal elevation in Minas Basin by 36%. A recent study by  supported previous research with 7.6 GW tidal-stream extraction scenario using the increasing turbine drag coefficient method in the water column. In the cited study, the maximum tidal current velocity reduction was predicted at 38.8%, the maximum M2 tidal amplitude decrease was simulated up to 2.4 m inside the Bay of Fundy, and 0.2 m M2 amplitude increase is predicted for the Massachusetts coastal area as the results of 7.6 GW tidal-stream extraction scenario.
With regard to tidal-stream energy extraction and SLR, a recent study by  included 2 m SLR to simulate the impact of tidal energy extraction at Minas Passage on the Gulf of Maine. The simulation was performed with a 2-D ocean model with a 1 arc minute grid resolution. The simulated scenario consisted of 7.1 and 5.2 GW tidal-stream energy extraction scenarios, including the consideration of coastal flooding due to SLR. In the cited study, the flood scenario was defined as SLR being allowed to overtake the coastal areas while the no-flood was defined as SLR without coastal flooding. Figure 1.7 displays the results from the cited study. Up to 0.5 m tidal amplitude increase was predicted on Massachusetts coastal area due to the maximum tidal-stream energy extraction scenario (7.1 GW) for both SLR scenarios.
In summary, previous studies

Sea level rise
Tides as long waves are easily modified by water depth, bathymetry, and topographic features . Therefore, tidal dynamics is sensitive to SLR, which changes ocean bathymetry and global dynamics of the ocean. NOAA has published a map of global SLR trend which is shown in Fig spectively. In terms of the impacts of SLR on global tidal dynamics, a recent study using a global ocean model predicted the change on the M2 and K1 components due to SLR . The cited study presented M2 and K1 amplitude changes due to a globally uniform 1 m SLR, which is shown in Figure 1 The cited study did not explore the other important tidal components such as the S2 and N2.

Introductory remarks
The Gulf of Maine has very good tidal energy potential due primarily to the extreme tidal range, up to 16 m, in the Bay of Fundy. Therefore, many studies have been conducted to better understand the tidal resource and evaluate the most efficient and effective methods of energy generation, and also to predict the future impacts of energy extraction.
Presently, available TEC devices in industry are mainly horizontal axis turbines designed by several companies such as SeaGen, Marine Current Turbine, and Open Hydro. Further research on TEC also focuses on array optimization. Single turbine, line array, staggered array, and random array designs are possible site optimization method which is based on methods used in offshore wind array. Passage, which was predicted to have very high tidal-stream velocities (up to 3.5 m/s). In general, previous studies predicted a tidal amplitude decrease inside the Bay of Fundy and a tidal amplitude rise in the US coastal area. Recent research in global ocean dynamics predicted that SLR is not only adding water elevation in the ocean but also changes the boundary forcing. Therefore, SLR may change the impacts of tidal energy extraction.

Objectives
The objectives in this study can be listed as follows: 1. Assessment of the impacts of the tidal energy extraction on tides in the Gulf of Maine.
2. Investigating the effect of SLR on tidal energy resource, including the changes in global dynamics of tides.
Firstly, this study aims to predict the impacts of tidal power extraction and SLR on the Gulf of Maine. Previous studies have focused on the dynamics of tides, resource assessment, and tidal energy extraction at several sites such as Passamaquoddy-Cobscook Bay, Kennebec River, Minas Passage and Minas Basin.
Furthermore, SLR, which is caused by global climate change, has emerged as an important factor that affects the dynamics of tides. Therefore, assessment of the combined effect of tidal energy extraction and SLR on the dynamics of tides pro-vides a better understanding of the impacts in the future, and will be beneficial to tidal energy development. Furthermore, recent studies in global dynamics of tides suggest that SLR not only affects the bathymetry of the model, it also modifies the boundary forcing. This study will analyze the changes on tidal dynamics due to tidal-stream energy extraction and SLR, including the changes in the dynamics of tide.

Tidal water elevation and tidal amplitude
Tidal water elevation and tidal components data are commonly used in model validation. In this thesis, 11 stations in the Gulf of Maine was used for validation.
Tidal water elevation was obtained from NOAA website (tidesandcurrents.noaa.gov) that provide historical data, prediction of water elevation for public and amplitude for tidal components. There are 6 NOAA tidal stations in the Gulf of Maine: Portland, Eastport, Nantucket, Boston, Chatham, Cutler Farris which are used in this thesis for model validation. For stations which are located in Canada, tidal amplitude at 5 locations (Yarmouth, Grindstone, Advocate Har-bour, Minas Basin and Economy) was obtained from previous studies, e.g, Wu, 2011). January-February 2011 period was selected as validation period due to time series data availability at all of the stations.

Tidal current velocity data
Tidal current velocity measurement was retrieved from NERACOOS website (www.neracoos.org). The website provides various measurement from their buoys which are operating in the Gulf of Maine. Due to data availability, we used 4 buoys (M01, N01, B01 and E01) for model validation. Similar to tidal water elevation, historical current measurement data was retrieved for January-February 2011 period.

SLR
Sea level rise data in this thesis were based on the literature study in Section 1.3.4. Model scenarios regarding SLR consider the effect of SLR on bathymetry and boundary effect. uniform +1 m water elevation without coastal flooding and the boundary effect is defined as 10% increase in M2 amplitude along the open ocean boundary.

Methodology
The methodology used in this work to examine the impacts of tidal-stream energy extraction and SLR follows these steps: • Application of a tidal model for the area using Regional Ocean Model System (ROMS).
• Tidal stream resources assessment assuming present situation.
• Impact of SLR on tidal stream resources.
• Impact of tidal-stream energy extraction and SLR on the dynamics of tides.

Theoretical background 2.3.1 Tidal constituents
Tidal constituents are key parameters in tidal modeling. 45 astronomical and 101 shallow-water constituents are known and are implemented in t tide . However, many of them have small amplitudes and/or extremely long periods. Therefore, in this thesis, 10 dominant tidal constituents are used for tidal simulation, as shown in Table 2.1.

Resonance in a basin
Tides can be regarded as long waves. Further, waves in the ocean are modified by water depth and coastal boundaries. Wave transformations, such as shoaling, refraction, and diffraction, apply to propagating waves in the ocean. Aside from that, coastal boundaries reflect incoming waves, causing interaction between the incident and reflected waves. This phenomenon may lead to standing waves, an where L is the length of basin and h is the depth of basin. Standing waves and resonance may also be produced in a basin with one open boundary that is forced harmonically. The resonant period of this case (T nf ) is expressed as, The application of standing wave and resonance theory in realistic conditions are more complex due to non-uniform bathymetry and irregular coastal basins. The study area in this thesis is the Gulf of Maine, which is known for an extreme high tidal range inside the basin due to resonance  Desplanque and Mossman, 2001).

Empirical equations for vertical velocity profile
In order to have a better comparison between model and observed data, velocity profiles were fitted to experimental data. The velocity profile, which can be obtained via measurements and/or 3-D ocean models, is a useful parameter for ocean studies. Many measurements have been conducted in effort to provide the vertical velocity profile in the ocean. However, the measurements are often not enough due to many factors, such as device specifications and maintenance.
Therefore, empirical methods were introduced to estimate the vertical current profile based on measured data. Power law is commonly used to give an estimate of velocity at specific water depth, which is expressed as where z is distance from seabed, d is total water depth and a is the profile coefficient. The value of a is set to 7 as recommended by previous research .

Simulations of tidal turbine in ocean models
TEC energy extraction theoretically is based on the kinetic energy concept that is defined as energy that is produced by a body due to its motion, which is defined as, where E k is the kinetic energy, m is the mass and u is the velocity. Current power is defined as the rate of change of current. Since mass flux can be expressed by volume flux times water density, kinetic power of a flow can be defined as, where P is power (watt), t is time, V -is the volume of water, ρ is the density of water, Q is flow rate (m 3 /s), and A t is the area of a turbine (m 2 ). From Equation 2.7, power is mainly dependent on the current velocity. The current power can also be expressed as power density, In Equation 2.7 and 2.8, the power is the available theoretical power in the ocean. shows that energy density rises significantly as current speed increases as it is proportional to u 3 . However, the technical power, which is defined as estimated power generation by turbine, is significantly lower due to energy loss. Practical power, P t , is estimated as, where C p is the efficiency of TEC.

Tidal modeling using ROMS
In this thesis, tidal simulations was done using Regional Ocean Modeling System (ROMS). The source code of ROMS is available online at www.myroms.org.
This section gives the overview of the model.

ROMS theoretical background
ROMS is a three dimensional terrain following ocean model based on conservation of mass and momentum. ROMS solves the Reynolds-averaged Navier Stokes equations using the hydrostatic and Boussinesq assumptions .
the vertical momentum equation with hydrostatic assumptions, ROMS momentum equations include local and convective acceleration, Coriolis force, pressure, turbulent and fluid shear stresses, forcing terms and diffusive terms.

Bottom stress parameterization
At areas close to the ocean bed, many hydrodynamic parameters, such as velocity, shear stress, Reynolds stresses, energy dissipation, and turbulent viscosity  . The method provides a force based on the drag force concept at the bottom boundary layer to represent the frictional mechanism. This formulation can be expressed as (see Table 2.2 for definition of parameters), For tidal simulation, common values for bottom drag coefficient are 0.0025 to 0.0040. The value is usually adjusted according to model validation.

Tidal turbines simulation in ROMS model
TEC implementation in ocean model has been studied in the past to predict future change in ocean dynamics. There are several methods such as bottom friction method in 2-D momentum equation , quadratic Rayleigh friction , and 3-D actuator disc concept .

Increasing bottom friction to simulate energy extraction
The extracted power over a cross-sectional area can be theoretically treated as additional dissipation of energy due to bottom friction . Using this concept, the bottom friction coefficient could be modified to simulate the far-field effect of TEC in the flow field.
and the total bottom friction is expressed as,

Actuator disc concept
Energy at TECs are generated by the torque which is applied to the rotor and is induced by movement of the blades. Consequently, wake and turbulence are produced at the area where a TEC operates ). Recently, Roc , 2013 provided a method to incorporate wake due to stream turbine energy extraction in regional ocean model as an assessment tool for turbine array optimization. In the cited study, actuator disc concept were implemented into ROMS.
The modified ROMS momentum equation is expressed as (see Table 2.2 for list of variables), In Equation 2.25, TECs are represented by F t , which is the force produced by TECs during power generation. The formulation of F t is expressed as, where |u ∞ | is current velocity at a location far from the turbine and n is the normal vector with respect to current velocity. The numerical implementation of TEC in ROMS is done with sub-grids between the ocean model and TEC.

ROMS tidal model development
The ROMS model domain was discretized with 1 arc-minute horizontal resolution, and 11 layers in a terrain-following vertical coordinate provided by ROM- In this thesis, 3 ROMS scenarios were assumed to examine the change in tidal dynamics: • Tidal simulation at present condition.
• SLR scenario: +1 m change in bathymetry and boundary effect. Water elevation increase was assumed uniform and water do not flood coastal area.
• Energy extraction scenario combined with SLR scenario.

Tidal stream resource assesment
Quantifying the available resources is the first step for a tidal energy development. The dynamics of tides was modeled using ROMS (Regional Ocean Modeling System) followed by model validation to assure the accuracy of the results. Then, tidal-stream velocity was characterized at potential sites and was used to evaluate tidal energy. The effects of SLR on the tidal-stream energy resource was also examined in this part.

Impact of tidal stream turbines and SLR
Energy extraction using TEC may change the dynamics of tides in the Gulf of Maine. Therefore, the assessment of future tidal dynamics due to TEC is important to provide a better understanding of tidal energy extraction. In this study, the impact assessment of tidal-stream turbines and SLR was performed as previous research . The change in the dynamics of tide was first examined using a hypothetical scenario at Minas Passage, which is approximated with the added bottom friction method. Then, the impact of SLR and/or tidal energy extraction on the dynamics of tides in the Gulf of Maine was investigated.

CHAPTER 3
Results

Model validation
To determine model performance, a reference tidal simulation in the Gulf of Maine was set up using ROMS. A comparison between model results and observational data was performed to validate the model. Error calculation was done using root mean square error (RMSE) and scatter index (SI), which can be expressed as, where X ROM S is ROMS results, X obsv is observation data, and N data is the total number of data. Tidal elevation on the Gulf of Maine was first validated with tidal water level measurement from stations available in the area and followed by tidal amplitude comparison for 11 tidal stations; then the validation for current was performed at 4 NERACOOS buoy locations. The period for model validation was selected on January-February 2011 due to data availability.

Tidal amplitudes validation
Tidal amplitude validation was first performed by time series comparison between model and observed data.   of tidal stations). Due to data availability, 11 tidal stations and 5 tidal stations for M2 and S2, respectively, was used in model validation. Figure 3.2 shows the validation chart for both M2 and S2 components. Based on the results, RMSE for the amplitude and the phase of the M2 constituent are 7% and 9%, which was very convincing. For the S2 component, the comparisons were resulted in also a good agreement, the error was 19% and 9% for amplitude and phase. More details are provided in Table 3.1 and 3.2 for the M2 and S2 components, respectively.   The M4 component also shows tidal asymmetry, which is caused by topographic features and friction at the seabed . The computed co-tidal for M2, S2 and M4 are shown in Figure 3.3, 3.5, and 3.6. Additionally, zoomed preview at the Bay of fundy for M2 and S2 are shown in Figure 3 Figure 3.5. S2 co-tidal chart simulated using ROMS. Colorbar shows the amplitudes and white lines represent the phase.

Tidal current validation
Following the tidal amplitude and phase validation, tidal current was validation also performed by comparison with available velocity data.  and simulated depth-averaged current is shown in Figure 3.8. According to time series comparison, the predicted tidal-stream profile qualitatively showed acceptable results in terms of current velocity magnitude. However, observed data showed irregular peaks which indicates measured currents velocity are not only tidal related, but are also affected by other ocean currents (e.g.,wind generated current). To fur-  ther examine the tidal-stream characteristic, a tidal analysis was done for the M2 tidal component using t tide MATLAB toolbox . From the results, the simulated tides showed good agreement with buoy data, which is visualized in Figure 3.9. Ellipse shapes between buoys were qualitatively similar with small errors in inclination angle, minor axis and major axis. For instance, at N01 station, maximum current velocity was calculated at 0.85 m/s and 1.1 m/s for observed and ROMS results, respectively. Other locations, B01 and E01, also showed good agreement between model results and observations. In general, the model overestimated tidal currents velocity in the area. A noticeable error was found at buoy M01, which showed significant error at minor axis prediction. More details are provided in Table 3.3. Based on the tidal analysis results, simulated depth-averaged tidal currents is not agreed very well with observation data. Major axis comparison shows that ROMS overestimate the current velocity in the domain, SI calculated at 36%. The inclination angle between data show good results with 6% error. A noticeable error was found at the minor axis with 120% scatter index between observation and model. To further assess ROMS performance, the tidal ellipse chart for the Gulf of Maine was plotted and was compared to the previous study by Hasegawa et al. in 2011 (Figure 3.11) that is based on Princeton Ocean Model (POM).

Increased bottom drag coefficient and tidal energy extraction
In this part, we set up an energy extraction scenario to test the increasing bottom drag coefficient method in ROMS. The extraction scenario was examined considering a suggested optimum configuration of array  to evaluate the spacing of TEC. Total horizontal area of the Passage is ∼10 km 2 which consists of six numerical cells (1287 m x 1287 m), as shown in Figure 3.12.
Minas Passage is able to fit 300 TECs in total (12 by 25 units across and along central axis of the water flow, respectively). TEC are assumed to be a horizontal axis and have a 20 m diameter. Also, TEC was assumed ideal, C p = 1 The scenario was further simulated in the model using the increased bottom friction method, resulting in an 0.0047 additional bottom drag coefficient (C * d ) and an 0.0077 total bottom friction (C * * d ). Table 3.4 shows the summary of the extraction scenario.  energy influx and outflux were calculated at 2.67 GW and 1.50 GW , respectively, which resulted in 1.18 GW of total dissipated energy. The calculated flux agreed well with the 1.23 GW tidal-stream energy extraction scenario. Table 3.5 shows the summary of the energy flux calculation.
Based on the results in this part, we were convinced that the increasing bottom drag coefficient method is applicable for TEC array representation in ROMS

Tidal resource assessment in the Gulf of Maine
In this section, we will focus on tidal-stream energy resources in the study area.
ROMS model was run for a 30 days period, which is the suggested period to assess energy resource according to European Marine Energy Centre (www.emec.org.uk).
Then, the average power density was evaluated over the entire domain based on the outputs. The impacts of SLR on the dynamics of tide were also examined to predict future tidal resource in the domain.

Present tidal energy resources in the Gulf of Maine
The tidal resource can be evaluated for both maximum theoretical power and average theoretical power. Maximum theoretical power may indicate a promising site, however, tidal current velocity and direction are changing over a tidal cycle and during spring-neap cycle. Thus, for tidal energy development, average tidalstream energy resource also commonly used to represent the potential of a site.
direction are changing over a tidal cycle and during spring-neap cycle.
First, the maximum spring velocity was used to estimate maximum theoretical power density in the area (Figure 3.14). In general, based on the results, 3 to 8 kW/m 2 maximum theoretical power is available in the study region, which is a relatively good resource. The highest power density was predicted at Minas passage, having up to 4.5 m/s current velocity which results in up to 23.24 kW/m 2 and 7.70 GW available maximum theoretical power. The results at Minas Passage agree with previous studies which estimated ∼7 GW available maximum theoretical power , as shown in Table 3.6. Table 3.7 shows the summary of available maximum theoretical power in the domain.   Passage while other sites have significantly lower resources. However, many sites have sufficient velocity ranges as demonstration sites or small power generation Table 3.7. Summary of available maximum theoretical power in the Gulf of Maine (see Figure 3.14 for site locations).  (2016) predicted a 10% change in the M2 amplitude along the boundary due to a 1 m SLR in the Gulf of Maine. In this part, we examine how the tidal resource in the Gulf of Maine will respond to SLR; the change in bathymetry and the dynamics of tides. Here, we set up two simulations: 1. +1 m uniform change in bathymetry.
2. +1 m uniform change in bathymetry and boundary effects (see Section 2.1.5 for details).

Impacts of energy extraction and SLR on tidal dynamics
Tidal energy extraction in general affects ocean dynamics and may result in adverse physical and environmental impacts. In this part, we set up two simulation Table 3.8. Summary of available average theoretical power and the impacts on the resources in the Gulf of Maine (see Figure 3.14 for site locations). The energy extraction scenario was set based on the testing scenario for the increasing bottom drag coefficent, 1.23 GW, which have a total of 300 TEC in the array and C p is assumed ideal. By using the increasing bottom friction method, additional bottom friction calculation (see Equation 2.23) is mostly dominated by C p , A t , and A cell , thus, energy extraction scenario can be set up by adjusting those parameters. For 0.74 MW extraction scenario, the 1.23 GW extraction case was modified by the implementation of the betz limit, C p is 0.6. The last extraction scenario, 2.5 GW was set up to match available estimated stream-energy in Minas Passage by FORCE. For the last scenario, the area of turbine blade was increased to extract 2.5 GW from water flow without modifying the total number of turbine and the turbine configuration. Further, we included SLR scenarios into energy extraction scenarios to predict future change in the dynamics of tide in the Gulf of Maine. Table 3.9 show the summary of energy extraction scenarios in this study.
All of the scenarios are located in ∼10 km 2 horizontal area in Minas Passage.
Boston and Minas Basin (see Figure 3.16) was selected to be the focus area based on basin configuration. For instance, Minas Basin is located at the end of the basin, while Boston is one of sites in the farthest area from Minas Passage.  Additionally, we also computed the change in the tidal range to see the total changes in water elevation. Figure 3.19 and 3.18 show the change in the tidal range for energy extraction and the combined scenarios. Similar to the results for M2, the tidal range differences rise as energy extraction in Minas Passage is higher. For instance the relative changes in Minas Basin is -0.79% at 740 MW energy extraction scenario and -3.59% at 2.5 GW energy extraction scenario. The inclusion of SLR scenario into the simulation also shows similar qualitative trend with the M2 tidal amplitude analysis because the resonance in the Gulf of Maine is determined by the M2 component. Table 3.11 shows more details for the impacts of energy extraction scenarios on the tidal range in Boston and Minas Passage. Table 3.12 shows summary of the model validation from previous research related to the impacts of tidal energy extraction in the Gulf of Maine.

CHAPTER 4 Discussion
Throughout this effort, tidal-stream energy assessment, and the impact of tidal energy extraction and SLR in the Gulf of Maine have been explored. It is shown that SLR, as well as tidal energy extraction, are affecting the dynamics of tides in this region.
Application of ROMS in this study demonstrated convincing results to simulate ocean dynamics. Lewis et. al (2013) considered ∼1 km grid as sufficient resolution to assess the first TEC generation and suggested higher resolution for better simulation results. Based on model validation, the implementation of regular horizontal uniform 1 arc-minute grid (∼ 1km 2 ) in the Gulf of Maine shows good results for 3-D regional tidal simulation. For instance, 7% and 9% scatter index for validation of M2 component amplitude and phase, respectively. Further, higher resolution using regular horizontal grid and/or the implementation of sub grids may present better results for both tidal water elevation and tidal current velocity simulation in the domain, which has a very complex bathymetry and topography. However, the implementation of very high resolution and sub grid are complex and computationally more expensive.
Regarding the tidal-stream resource assessment, the inclusion of SLR scenario in this study: +1 m uniform water level and boundary effect, significantly affects the resource compared to the present day. In the Gulf of Maine, the effects of +1 m SLR to the bathymetry of the domain is relatively very small throughout the domain. Recent research in global ocean dynamics suggested that SLR not only affects the bathymetry, but also the dynamics of tides. According to  , about 10% increase in M2 amplitude was predicted due to + 1 m global uniform SLR. Consequently, the implementation of the effects of SLR on tidal dynamics along the open ocean boundary become important as the tidal model is forced by tidal components. The results predicted up to 74% changes in tidal-stream resources throughout the domain except Minas Passage, which is predicted to have 37% increase in the resources. Based on the results, future energy extraction may benefit from SLR in terms of the available resource.
The simulation of tidal energy extraction in Minas Passage was conducted with the increasing bottom friction method in ROMS. The method allows TEC array representation using added bottom drag coefficient in the tidal model. The method is relatively simple compared to the actuator disc concept in ROMS, which is recently proposed by Roc (2010). By using the increasing bottom drag coefficient method, the bottom drag coefficient of the domain is spatially modified in the designed location to represent TEC array. The method was tested and the energy flux calculation showed good results, 4% error, between the estimate of energy extraction and the total energy dissipation by additional friction at the seabed.
However, the increasing bottom drag coefficient distribute the energy dissipation uniformly inside the cell area so that the method neglects the hydrodynamics effects in the near-field produced by the blades, which is not the objective of this study. The actuator disc concept in ROMS provides more advanced approach for TEC representation with turbulence correction at TEC array location. The proposed method is more complex in terms of domain discretization that uses sub grids between ocean grid (∼ 1 km) and turbine grid (∼ 20 m) and is also computationally more expensive.
The inclusion of SLR change the results of energy extraction scenario in Minas Passage. Based on the results (Table 3.10 and Table 3.11), the changes in tidal amplitude throughout the domain is relatively small at the present day. The inclu-sion of SLR significantly affects the results as maximum difference rise up to 8% in Boston. In detail, the impacts of SLR on present day without energy extraction scenario dominated the change in tidal amplitude (up to 7% change). Furthermore, the combination of energy extraction scenario and SLR showed non-linear relation between them. Therefore, future energy extraction activity in Minas Passage need to be explored regarding several topics, such as, total energy extraction scenario, spatial area of turbine array, SLR value related to TEC lifetime design, and TEC array configurations.

CHAPTER 5 Conclusion
Ocean renewable energy resources (e.g, tidal range and tidal stream) can help reduce carbon emissions . Tidal power generation is highly site-specific and generally is feasible where tidal range and/or current velocity are large enough due to ocean environment such as amplification by sea bottom profile, funneling in estuaries, reflections by large peninsulas, headlands and resonance effects . The dynamics of tides in the Gulf of Maine are unique due to the tidal resonance, which generates the largest tidal range in the world (about 16 m). Accordingly, a number of previous studies have assessed the impacts of the tidal energy development in the Gulf of Maine. Further, due to the sea level rise (SLR), those impacts may also change during the project lifetime, which is usually more than 25 years. In this research, the impact of tidal energy extraction considering sea level rise was simulated.
A tidal model of the Gulf of Maine was developed using Regional Ocean Model Basin, respectively, for both tidal range and the M2 components. The application of actuator disc theory in ROMS will be considered in the future. In summary, the implementation of modified boundary forcing due to SLR, which was ignored in the previous works, can change the results of the impact assessment. Table 5.1 shows the summary of the impacts of tidal energy extraction and SLR.
Based on the results, the far-field impact is more threatening in coastal regions of US. However, the impact of energy extraction in Minas Passage is relatively small. Compared to the model validation, the impacts were inside the uncertainty level of the model. For example, maximum change in Boston coastal area was calculated up to 1.65 %, which is inside the level of uncertainty in models, about 10 %. Furthermore, the impact of SLR on the dynamics of tides is much more than energy extraction assuming 2.5 GW extraction in Minas Passage.