Marine Hydrokinetic Energy in the Gulf Stream Off North Carolina: An Assessment Using Observations and Ocean Circulation Models

  • Caroline F. Lowcher
  • Michael Muglia
  • John M. Bane
  • Ruoying He
  • Yanlin Gong
  • Sara M. Haines


There has been global interest in renewable energy for meeting energy demands, and as these demands increase, it will become of greater importance to utilize low-carbon energy sources to mitigate anthropogenic impact on the environment. Onshore hydropower is responsible for half of the electricity generated by a renewable source in the USA. In the ocean, marine hydrokinetic (MHK) energy in western boundary currents (WBCs) can be considered for electricity generation by submarine turbines. WBCs are a continuous and sustainable source of energy that could be transmitted to shore to support coastal communities in future years. The Gulf Stream is the WBC of the North Atlantic subtropical gyre, and it flows for part of its course along the upper continental slope off the southeastern USA. This large-scale current has maximum flow speeds exceeding 2 m s−1, and this together with its proximity to the coastline distinguishes it as a potential source of MHK energy. Using current data from a moored acoustic Doppler current profiler (ADCP) and a regional ocean circulation model, MHK power densities offshore of North Carolina were found to average 798 W m−2 for the ADCP and 641 W m−2 for the model during a nine-month period at a potential turbine site, a difference of about 20%. The model was shown to have similar current speeds to the ADCP for slowly varying currents (fluctuations of weeks to months due to Gulf Stream path shifts), and lower speeds for higher frequency current variations (fluctuations of several days to a couple of weeks due to wavelike Gulf Stream meanders). This article considers the Gulf Stream as a prospective renewable energy source and assesses the power density of this WBC at multiple locations offshore of North Carolina. Understanding the Stream’s power density character, including its spatial and temporal variations along the North Carolina coast, is essential in considering the Gulf Stream as a future alternative energy resource.


Power density Alternative energy Ocean turbine Gulf Stream Cape Hatteras 


Marine hydrokinetic (MHK) energy, the kinetic energy in moving seawater, is one potential renewable energy source for electricity generation. Almost all MHK energy is in ocean surface waves, tidal motions, and large-scale (non-tidal) currents. In this chapter, we consider the potential for harvesting MHK energy from large-scale ocean currents to generate electricity using moored submarine turbines. A worldwide survey of ocean currents using Hybrid Coordinate Ocean Model (HYCOM) simulations shows that the western boundary currents (WBCs) in the wind-driven subtropical gyres are essentially the only large-scale currents with flow speeds sufficiently fast to be considered for commercial installation of electricity-generating turbines (VanZwieten et al. 2013; Bane et al. 2017). These WBCs include the Gulf Stream, Kuroshio Current, North Brazil Current, Agulhas Current, Somali Current, and East Australian Current (Imawaki et al. 2013).

We provide herein an assessment of the MHK energy and power density in the Gulf Stream offshore of North Carolina in the region of the Carolina Capes (Cape Hatteras, Cape Lookout, and Cape Fear), based on direct observations of ocean currents and ocean circulation model simulations. Our study region off North Carolina is one of the two locations along the Gulf Stream off the southeastern US coast where the Gulf Stream flows close to the coastline (Fig. 1). The other location is within the Florida Straits (see Haas et al. chapter “Ocean Current Energy Resource Assessment for the Gulf Stream System: The Florida Current” in this volume). The proximity of the Gulf Stream to the shoreline and coastal population centers in these areas, and its limited meandering relative to the rest of the southeastern US coast, makes these two locations desirable for possible future commercial development. Chapter “Ocean Current Energy Resource Assessment for the Gulf Stream System: The Florida Current” by Haas et al. in this volume provides a model-based assessment of MHK energy in the Florida Current, which is the portion of the Gulf Stream within the Florida Straits.
Fig. 1

A sea surface temperature “snapshot” showing the warm Gulf Stream flowing poleward along the southeastern US coastline in March 2008. Note the wavelike meandering of the path of the Gulf Stream, especially noticeable north of 32° N latitude. Raleigh Bay is located between Cape Hatteras and Cape Lookout, and Onslow Bay is between Cape Lookout and Cape Fear. The three cross-isobath transects that have been studied in detail are shown as solid black lines orthogonal to the coastline. The Gulf Stream is closest to the coastline off Cape Hatteras, North Carolina, and within the Florida Straits, south of 27.5° N latitude, where the continental shelf is relatively narrow. The 100-m isobath (the seaward edge of the continental shelf) is denoted by the dashed black line

To examine the power (kinetic energy per unit time) in the Gulf Stream flow off North Carolina, we use current observations from moored and boat-mounted acoustic Doppler current profilers (ADCPs) and ocean surface-current radars, and simulations of ocean currents by a regional ocean circulation model. This model-data approach gives estimates of how much power is available for the resource in this region, delineates how the power varies in space and time, and will help determine which areas off North Carolina are optimal for future deployments of MHK devices. This approach also provides for model-observation comparisons of current speed and power, thereby revealing the confidence that may be placed in model predictions of ocean power at times and at locations where observations are unavailable. In the following discussion, we present the characteristics of Gulf Stream power. The engineering, environmental, legal, financial, and other use considerations that will likely affect the development of commercial MHK energy harvesting facilities are not our focus here and thus are not discussed. Details about these subjects may be found in other publications, for example, Boehlert and Gill (2010), Neary et al. (2014), Brown et al. (2015), and Li et al. (2017). In the next section, we describe pertinent aspects of MHK energy, and in section “The Gulf Stream”, we give a description of the Gulf Stream off the southeastern US coast. Section “Assessing MHK Energy and Power Density” presents the available data sets to determine MHK energy, and section “MHK Energy from the Gulf Stream Along North Carolina” discusses the speed and power at particular locations off the coast of North Carolina, including their temporal and spatial variations.

MHK Energy and Power Density

Harvesting the Gulf Stream’s MHK energy with moored submarine turbines is a type of “in-stream” power generation, meaning it does not use water impoundment like a hydroelectric power plant that stores water behind a dam (VanZwieten et al. 2014). Kinetic energy flux in an oceanic current is the amount of MHK energy that flows through a unit cross-sectional area oriented perpendicular to the current direction per unit time. Kinetic energy flux is equivalent to the power density, P (in W m−2), of the current, and it is given by

$${\text{P}} = 1/2\,\uprho\,{\text{S}}^{ 3}$$
where ρ is the density of seawater (taken to be 1025 kg m−3 in our power assessment) and S is the current speed (in m s−1). The total power in the current flowing past a submarine turbine is P integrated over the surface area, A, swept by the turbine blades (the “blade circle”). For a current speed that is non-varying across the cross-sectional area swept by the turbine blades, the total power is P times A. To address the Gulf Stream’s power density off the North Carolina coastline, we must know the current speed, and in this chapter, this is provided by the ADCP measurements and the numerical model simulations.
Submarine electricity-generating turbines are basically underwater versions of wind turbines used for electrical power generation, which would likely be moored to the ocean bottom using anchor lines (e.g., Corren et al. 2013) as opposed to solid masts like those used for wind turbines. Although there are presently no large size submarine turbines that would be appropriate for commercial use in large-scale ocean currents, smaller versions are presently operating in many locations in inshore and estuarine waters (e.g., Neary et al. (2014) present a model study of a submarine turbine that has a rotor diameter of 33 m and a 1 MW power rating. The pertinent operational characteristics of this model turbine (Fig. 2) are start-up speed = 0.5 m s−1 (no power generated if the current speed is below this), maximum-rotation-rate speed = 1.7 m s−1 (maximum power of 1 MW generated if the current speed exceeds this), and maximum rotor power coefficient = 0.48 (power harvested from the current is 48% of the current’s power density times the blade circle area).
Fig. 2

Ocean current power, rotor power, and power coefficient versus current speed for the rotor on the model submarine turbine presented by Neary et al. (2014). This is a 33-m-diameter, three-blade rotor. (Figure from Bane et al. 2017, adapted from Neary et al. 2014)

Siting an array of subsurface turbines off North Carolina will take into account the power available at the site (section “MHK Energy from the Gulf Stream Along North Carolina”), ecological and bottom geological properties, proximity of the array to the coastline (for cable connections from the turbine array to shore), and onshore electrical grid connection locations. The array site would likely be on the upper portion of the continental slope, as shown in Fig. 3, because this positions the turbines relatively close to shore, in water depths that will allow reasonable mooring design, and near the high-speed core of the Gulf Stream current. Turbines will need to be positioned so that all hardware is at least 30 m or so below the surface in order to keep the machinery out of the active surface wave zone and below surface ship traffic.
Fig. 3

A schematic of the Gulf Stream current off the southeastern US coastline. The inshore edge of the Stream, indicated by the sea surface temperature (SST) front, is typically located near the shelf break (~40 km from the coastline near Cape Hatteras, ~100 km from the coastline off Georgia, and a few kilometers from the coastline off southeast Florida). The main Gulf Stream jet meanders as it flows over the continental slope, seaward of the shelf break. Surface-current arrows are drawn with a typical profile showing the lateral current shear. The lengths of the arrows and the subsurface isotach curves show the fastest speed at the surface in the core of the current. These speeds diminish in the horizontal and vertical directions away from the surface-current maximum. Subsurface current turbines (blue circles) are shown moored on the upper portion of the continental slope. For reference, offshore wind turbines might be moored on the shallower continental shelf, where current speeds are much lower than in the Gulf Stream. (Figure from Bane et al. 2017.)

The Gulf Stream

The Gulf Stream, like most WBCs, is a narrow (~100 km in width), deep (~1 km in vertical extent) jetlike flow that has a maximum surface-current speed often in excess of 2 m s−1 (Imawaki et al. 2013). Current speed decreases in any direction away from this surface maximum. These characteristics are shown schematically in Fig. 3. Off Cape Hatteras, the inshore edge of the Gulf Stream is approximately 40 km (~25 miles) from shore, where surface-current speeds along the edge are roughly 0.2 m s−1. To reach the core, where the fastest speeds can be found, is an additional 30–40 km. This is based on the cross-current extent of the Stream, and the fact that its cyclonic shear zone is roughly the inner third of the Gulf Stream and the anticyclonic shear zone is the other two-thirds, thus leading to an asymmetry in the current’s jetlike shape. Cape Hatteras is the closest location to the Gulf Stream along the US east coast north of the Florida Straits where it is only several km from Fort Lauderdale. The volume transport in the Straits is approximately 30 Sv (Barringer and Larsen 2001), and the transport off Cape Hatteras is typically about 90 Sv (Halkin and Rossby 1985) (1 Sv = 106 m3 s−1). Another apparent feature in Figs. 1 and 3 is the wavelike, meandering path of the Stream (Webster 1961). This meandering pattern propagates poleward, and each meander wave has an alongshore wavelength between 100 km and a few hundred km (Tracey and Watts 1986). At a fixed location, it will take from around 3 days to as many as 20 days for one meander wavelength to propagate past (the “meander period”) (Bane et al. 1981). This means that the Gulf Stream jet will move closer to shore for about half of the meander period, then move offshore again for the next half-period. This lateral meandering motion of the Stream’s narrow jet will cause the current speed at a fixed turbine location to increase and decrease with a meander’s period. Another type of motion that the path of the Gulf Stream undergoes from time to time is a longer term lateral shift of the jet’s path (Bane and Dewar 1988; Quattrocchi et al. 2012). This can move the Stream laterally relative to a moored turbine site, either decreasing or increasing the power generated by that turbine for weeks to months (Bane et al. 2017). In section “MHK Energy from the Gulf Stream Along North Carolina”, we show observed and modeled power variations at a fixed turbine site that are caused by meanders and path shifts. These Gulf Stream motions will be the dominant causes for power generation variations over time.

Assessing MHK Energy and Power Density

Our study region is over the outer continental shelf and continental slope from 32.5° N to 36° N and 73.5–77° W, and we focus on three cross-isobath transects within this area (Fig. 4). At one location along the northern transect, we make comparisons between simulations from a regional ocean circulation model and ADCP observations. Each transect covers most of the average Gulf Stream’s cyclonic and anticyclonic shear zones. The northernmost transect, the Cape Hatteras (CH) transect, is located southwest of a bathymetric feature known commonly as The Point. The Point is offshore of Cape Hatteras at the location where the 100-m isobath crosses 35.5° N. This is where the shelf break changes from a northeastward to a more northward orientation. The Gulf Stream typically separates from its flow along the upper continental slope near The Point and begins its transit into the deeper water of the open Atlantic.
Fig. 4

2013 average current velocities at 75 m overlying a North Carolina base map. Stations in the cross-isobath transects are shown by yellow circles with both model and observational data at the blue circle. The transects are labeled CH for Cape Hatteras, CL for Cape Lookout, and OB for Onslow Bay. Green arrows show the current velocity. Isobaths are the black contour lines, and the 100-m isobath is in bold. The four blue stars show the locations of the cities Virginia Beach, Kitty Hawk, Morehead City, and Wilmington—the sites where MHK energy from the Gulf Stream can be connected to the power grid

There are 33 stations total from which we have gathered model data, 21 of these are displayed in Fig. 4. Of these displayed, 6 of which are shown along the CH transect as 5 yellow circles and one blue circle. These have been selected to cover the span of the average Gulf Stream. The geographical coordinates and total water depth for each station on the CH transect are listed in Table 1. The bottom-moored 150 kHz ADCP lies on the CH transect at the blue circle in Fig. 4. The middle transect (Cape Lookout = CL) is positioned in Raleigh Bay just off Cape Lookout, and the southernmost transect (Onslow Bay = OB) is in Onslow Bay. Tables 2 and 3 in the Appendix give the position and water depth of each station along these transects. Blue stars in Fig. 4 denote four sites where the present-day grid is substantial enough for MHK energy harvested from the Gulf Stream to be introduced. The sites are at Virginia Beach, Kitty Hawk, Morehead City, and Wilmington. Our power assessment for MHK energy in the Gulf Stream focuses on a depth of 75 m below the surface, given consideration of MHK device operating depths, turbine technology, commercial shipping interests, and surface wave influences on turbines and their moorings.
Table 1

Cape Hatteras (CH) transect stations

Cape Hatteras


Latitude (°)

Longitude (°)

Depth (m)

CH 1




CH 2




CH 3




CH 4




CH 5




CH 6




CH 7




CH 8




CH 9




CH 10




CH 11




CH 12




CH 13




CH 14




CH 15




CH 16




Table 2

Cape Lookout (CL) transect stations

Cape Lookout


Latitude (°)

Longitude (°)

Depth (m)

CL 1




CL 2




CL 3




CL 4




CL 5




CL 6




CL 7




CL 8




CL 9




CL 10




Table 3

Onslow Bay (OB) transect stations

Onslow Bay


Latitude (°)

Longitude (°)

Depth (m)

OB 1




OB 2




OB 3




OB 4




OB 5




OB 6




OB 7




The Gulf Stream power assessment presented here uses data sets from two different sources. One is the Regional Ocean Model System (ROMS) that has been run for an area that encompasses the Cape Hatteras region. The other is a 150 kHz ADCP that was moored on the ocean floor on the upper continental slope off Cape Hatteras. Each data source is described below.

ROMS Model

Simulated ocean currents were generated by a regional ocean circulation model implemented for the Mid-Atlantic Bight (MAB) and South Atlantic Bight (SAB), hereafter MABSAB, covering the area between 81.8–69.8° W and 28.4–41.8° N (Bane et al. 2017). The model is based on the ROMS (Shchepetkin and McWilliams 2005), a free-surface, terrain-following, primitive equations ocean model in widespread use for estuarine, coastal, and basin-scale applications. The horizontal resolution of this model is 2 km, sufficient to resolve the Gulf Stream’s variability. Model bathymetry was interpolated from the National Geophysical Data Center (NGDC) 2-Minute Gridded Global Relief Data. Vertically, there are 36 terrain-following layers that have high resolution near the surface and bottom in order to better resolve ocean boundary layers. Momentum advection equations were solved using a third-order upstream bias scheme for three-dimensional (3D) velocity and a fourth-order centered scheme for two-dimensional (2-D) transport, whereas tracer (temperature and salinity) advections were solved with a third-order upstream scheme in the horizontal direction and a fourth-order centered scheme in the vertical direction. The horizontal mixing for both the momentum and tracer used the harmonic formulation with 100 and 20 m2 s−1 as the momentum and tracer mixing coefficients, respectively. Turbulent mixing for both momentum and tracers was computed using the Mellor/Yamada Level-2.5 closure scheme (Mellor and Yamada 1982). For open boundary conditions, the model was nested inside the 1/12 degree global data assimilative HYCOM/Navy Coupled Ocean Data Assimilation (Chassignet et al. 2007; Gong et al. 2015) output superimposed with tidal forcing of 6 major tidal constitutes (M2, S2, N2, O1, K1, and Q1) derived from an Advanced Circulation (ADCIRC) tidal model (Luettich et al. 1991) simulation of the western Atlantic. The MABSAB hindcast ran from January 1, 2009 through December 31, 2014, and it did not incorporate the observations from the ADCP into the model computations of currents. Hourly model output was used for this study.

Moored ADCP Measurements

The moored 150 kHz ADCP location was selected based on its position within a region where Gulf Stream meanders have relatively low lateral amplitudes (Miller 1994). Higher power density levels at a fixed location tend to result from less lateral movements of the Stream. The ADCP was deployed for nine months, from August 1, 2013 through May 29, 2014. It was deployed in a water depth of 228 m at 35.1° N and 75.1° W. The instrument measures currents from 9 m above the bottom to 30 m below the surface with 4 m vertical resolution every 10 min. Measurements within 30 m of the surface are unreliable because of overwhelming acoustic reflection from the air–sea interface. The ADCP pod also contained a conductivity, temperature, and depth sensor.

Additional Observations

In addition to the moored ADCP measurements, hourly high-frequency (HF) radar surface-current measurements were collected, and sporadic (as weather allowed) vessel-mounted ADCP current measurements were made with downward-looking ADCPs along the CH transect. As these data get processed and analyzed, future studies will be able to make model comparisons with them in order to improve the spatial coverage for power density computations. As confidence is gained in the quality of the radar surface currents and their ability to accurately determine the variability in Stream location, they will serve as a valuable tool to infer MHK energy variability in this region. Because the structure of the Gulf Stream jet is remarkably consistent (Halkin and Rossby 1985), surface currents alone may provide accurate inferences about the MHK resource beneath them.

HF Radar Surface Currents

A network of land-based 5 MHz HF radars made consistent hourly (3 h averaged) surface-current measurements with 6 km2 spatial resolution (Fig. 5). During the moored ADCP time frame, an additional radar was added to the network on Core Banks to enhance spatial coverage of the Gulf Stream. Currents were measured within the Cape Hatteras MHK Gulf Stream focus area. The radars are essential because they provide consistent hourly estimates of the Gulf Stream location previously not available from other historical methods such as satellite sea surface temperatures (SST) and altimetry. The primary cause of variability in Gulf Stream power density at a given location is the variability in Gulf Stream position caused by the Stream’s meanders and path shifts. Thus, we are endeavoring to develop methods to determine the hourly Gulf Stream location using the radar network. These methods use the hourly location of the landward Gulf Stream edge from HF surface currents to identify the maxima in the relative vorticity in the surface currents and determine the maximum gradients in radial velocities (relative to individual radars). A demonstration of this method is shown in Fig. 6.
Fig. 5

The Gulf Stream MHK energy program added a radar to the southern extent of network coverage in 2013 to expand surface-current coverage over the area being considered for harvesting energy (black dots are the locations of preexisting Coastal Ocean Dynamics Applications Radar and blue “x” is the recently added radar on Core Banks). The blue line is the cross-shelf ADCP boat transect within the focus area that crosses the ADCP mooring. The boat transect is along the shoreward portion of the CH transect

Fig. 6

An example of a graphical user interface developed for radar/SST edge detection comparison to evaluate the efficacy of radar Gulf Stream edge detection methods. Four lines radiating from the Buxton, North Carolina HF radar site are the four bearings selected for edge detection by the radar; the max surface velocity gradients (blue diamonds) and max surface velocity (black squares) are overlaid on a high-quality SST image

ADCP Vessel Transects

Currents have been measured along the shoreward portion of the CH transect (Fig. 4) on several occasions from a small vessel and from the Research Vessel (RV) Neil Armstrong. The small vessel has a downward-looking 300 kHz ADCP that measures currents in the top 100 m of the water column with 4 m resolution. Thirteen transects have been conducted during the 2013–2016 period along the inshore side of the CH transect, extending 14 km from the 100–1000-m isobaths. The RV Armstrong is a research vessel that has three downward-looking ADCPs at three different frequencies: 38, 150, and 300 kHz. The Armstrong measured currents along the CH transect in April 2016 (Fig. 7). The 70-km transect started at the 100-m isobath in the cyclonic shear zone of the Gulf Stream and extended offshore through the anticyclonic shear zone. The cross-stream current measurements are valuable for observing the variability in MHK resource with depth along the CH transect, and for examining shears in the water column that are important engineering considerations for the development of energy extraction devices. In Fig. 7, the canonical velocity structure indicative of the Stream off Cape Hatteras is apparent in the top 1000 m. The notable counter flow below the Stream hugging the shelf slope is likely Upper Labrador Sea Water (e.g., Richardson 1977).
Fig. 7

Cross-stream measurements made by the three RV Armstrong ADCPs in April 2016. The white curves delineate the depths to which currents were measured at the 300 and 150 kHz frequencies (respectively). The 38 kHz ADCP measured to approximately 1500 m depth. Some discontinuity exists along current contours due to varying resolution at differing frequencies of the three instruments. Water below the ADCP range is black, and the brown is the ocean bottom measured from the onboard multibeam system

MHK Energy from the Gulf Stream Along North Carolina

Average Currents

The average model current velocities from 2013 at 75-m depth are displayed in Fig. 4. The average Gulf Stream jet is apparent in this figure; the inshore edge of the average jet is roughly along the 100-m isobath, while the offshore edge of the Stream is close to the 4000-m isobath. Near the center of each of the CH, CL, and OB transects, average 75 m velocities are slightly greater than 1 m s−1. Average velocities diminish toward either edge of each transect.

Figure 8 shows a 3D view of model currents in the upper 1000 m on November 1, 2013, that agrees well with Gulf Stream structure from vessel transect observations. The Stream’s jetlike structure is apparent on each transect. White arrows are the surface currents, red arrows are the wind stress, and the magenta X is the ADCP mooring location. The dark red core of the current meanders along the North Carolina continental slope as seen by its differing location on each of the transects. The core of the current is farthest offshore in the CH transect, and closer to shore in the CL and OB transects. This model day coincides with an occurrence when the Stream location was away from the shelf break near Cape Hatteras. Such offshore movement of the Stream occurs because of meandering and longer term path shifts. These lateral meanderings are described in greater detail in the next sections.
Fig. 8

Snapshot of the vertical slices at the three cross-isobath transects on November 1, 2013, constructed using model data. The core of the current is shown in dark red, white arrows show the surface currents, and red arrows show the wind stress. This is a frame from a movie that shows the model speeds at the transects from August 1, 2013 through April 28, 2014

Current Speed and Power Density Time Series

Figure 9 shows the current speed time series computed by the model (red time series) and observed by the moored ADCP (blue time series) from August 1, 2013 to April 28, 2014 at 75 m below the surface at the ADCP site. During this time period, the ADCP and model each had average current speeds of 0.94 m s−1 (black horizontal line in Fig. 9). The maximum observed current speed was just above 2 m s−1, while the model speed maximum was 2 m s−1. Current speeds fluctuate with periods that range from semi-daily (tidal) to multi-monthly (Gulf Stream path shifts). There are times when the model speeds are generally slower than the observed speeds, such as from August through mid-October and from the latter half of March to the end of April. In general, the model underestimates many of the higher frequency speed fluctuations seen in the ADCP observations, but successfully captures the lower frequency changes. It is notable that in November, both the model and observed current speeds decreased to near zero for almost 3 weeks before rapidly increasing to 1.5 m s−1 in the model and 1.8 m s−1 in the ADCP observations. During this time, a large Gulf Stream meander propagated through the area, causing the Stream’s core to be farther offshore and weaker current speeds over the ADCP site.
Fig. 9

Time series of the current speed at 75 m from August 1, 2013 through April 28, 2014. Model speeds are shown in red and moored ADCP speeds in blue. The horizontal black bar is the model and ADCP average speed of 0.94 m s−1

Time series of power density from the ADCP (blue time series) and model (red time series) is shown in Fig. 10. Unlike the close agreement in speed averages (Fig. 9), the ADCP has an average power density of 798 W m−2 (blue horizontal line in Fig. 10), while the model has an average power density of 641 W m−2 (red horizontal line). The observed average power density is 24% higher than the model’s, due mostly to the model’s tendency to have smaller amplitude speed peaks (primarily from Gulf Stream meanders) than the ADCP. Because power density is proportional to speed cubed, pronounced differences in speed (where the ADCP observations have higher amplitudes than the model, such as August 1 through mid-October 2013) result in even greater differences in the power density. During November, the current speeds (and hence power density values) decreased to near zero. With the Stream positioned farther offshore during this time, the ADCP location was no longer within the jet and measured very weak flow speeds.
Fig. 10

Time series of the power density at 75 m from August 1, 2013 through April 28, 2014. Model time series is shown in red and ADCP time series in blue. The red horizontal bar shows the model power density average of 641 W m−2, and the blue horizontal bar shows the ADCP power density average of 798 W m−2

Year-to-Year Power Variations

Reasonably good agreement between the model and ADCP observations supports using the model to compute annual-average power levels at all of our stations off North Carolina. The colored vertical bars in Figs. 11, 12, and 13 are the average power densities from 2009 through 2013 for the stations in Fig. 4. The red vertical bars in Fig. 11 are the model’s annual power averages at the ADCP location (blue circle on CH transect). At this location, the 2010 annual average of 14 W m−2 is very low compared to the subsequent years, and is 3% of the 5-year average (467 W m−2) and only 1% of the power in 2012 (1055 W m−2). Meanwhile, the average power in 2012 is 230% of the overall average. In 2010, the Gulf Stream path had a long-term shift offshore of the ADCP location that produced much lower averaged current speeds. Two years later, the Stream had shifted onshore and was positioned against the continental shelf break with the core of the current over the ADCP site and hence greater speeds. For both of these years, it appears that the current had long-term lateral shifts in the Stream’s path that produced large fluctuations in available power. The other 3 years—2009, 2011, and 2013—had averages close to the 5-year average. This was likely a result of the Gulf Stream’s path staying over the ADCP location for much of the year, with minimal long-term lateral shifts in the Stream’s path.
Fig. 11

Five-year model annual power density averages from 2009–2013 at 75 m at multiple stations in the Cape Hatteras transect. Red vertical bars represent the ADCP location. The stations presented above are the same stations in the Cape Hatteras transect shown in Fig. 4 for average current velocities

Fig. 12

Five-year model annual-average power densities from 2009 through 2013 at 75 m at stations in the Cape Lookout transect. These stations are shown in the Cape Lookout transect in Fig. 4

Fig. 13

Five-year model annual power density averages from 2009 through 2013 at 75 m at multiple stations in the Onslow Bay transect. These are the same stations in the Onslow Bay transect with average current velocities shown in Fig. 4 for average current velocities

The maximum power density reached at the Cape Hatteras transect (Fig. 11) during these 5 years is 1,169 W m−2 at Station 13 (yellow bar) in 2013. In 2009, there is small lateral variation in power density, with greater lateral variation in the subsequent 4 years. The greatest annual-average power density is at Station 13 for most of the years. In 2011–2013, the power density peaks are greater and more pronounced, while stations not within the peaks experienced much less power and greater horizontal current shear. Comparing our ADCP location in 2010 to the other stations in the transect, we note that the ADCP station lies along the inshore flank of the Gulf Stream, within its cyclonic shear zone, and greater power is found at more offshore stations. Again in 2012, the Gulf Stream had shifted onshore as indicated by the skewed power envelope of the annual averages, the result being much higher power at the ADCP station.

The same analysis of the CL and OB transects is shown in Figs. 12 and 13. The maximum power density at the CL transect is 1,105 W m−2 at Station 5 in 2012. The distribution is skewed more landward. The power peak for each year fluctuates between Stations 5 and 6—the yellow and green bars. These stations both have a 5-year average of about 875 W m−2, and lie over depths of approximately 500 and 2800 m. This is important to note considering the costs of cabling and hardware required to reach greater power availability in the Stream at locations farther offshore. A depth of 2,800 m presents significant challenges for mooring turbines to capture power and then transporting the power back to shore over longer distances.

On the OB transect, the maximum power is 1,008 W m−2 at Station 4 in 2013. The annual OB transect power is more uniform and consistent throughout the 5 years. The lateral shear is apparent, and the greatest available power is at Station 4. The water depth at Station 4 is 654 m, presenting potential engineering challenges. In 4 of the 5 years, Station 3 has the second greatest power with a 5-year average of about 400 W m−2 and a water depth of 334 m. This is much shallower than the depths of Stations 5 and 6 in the CL transect, and based on its consistent power density over these 5 years, this suggests it would be a viable turbine location.

Current Directions

Current roses in Figs. 14 and 15 were constructed from ADCP observations and model currents at a depth of 75 m from August 1, 2013 to April 28, 2014. They are binned in 0.5 m s−1 speed and 15° directional increments. The Stream almost always flows northeastward for the time period considered. The flow direction is 45° about 70% of the time for the model and about 60% of the time for the observed currents. There are also a few instances on both roses when the current reverses to flow toward the southwest. Reversal speeds are small and occur less than 2% of the time. These flow reversals are likely caused by frontal filaments or cold-core eddies propagating along the cyclonic edge of the current (e.g., Bane et al. 1981). Flow of the Gulf Stream at the ADCP site is nearly unidirectional at about 45°. While there are lateral meanders of the Gulf Stream in which the whole current shifts onshore and offshore, the direction of flow within the current has small variation. This agrees with Kabir et al. (2015) who also found current direction to be quite uniform in the northeastward direction.
Fig. 14

Moored ADCP current rose from August 1, 2013 through April 28, 2014 at 75 m. Percentages show the time spent in a given direction

Fig. 15

Model current rose from August 1, 2013 through April 28, 2014 at 75 m at the ADCP location. Percentages show the time spent in a given direction


In this chapter, we have presented an observation-and model-based assessment of MHK energy in the Gulf Stream off the coast of North Carolina, as well as demonstrated the additional available observations from land-based HF radars and shipboard ADCPs that will be included in future studies. Data from a moored 150 kHz ADCP and a ROMS model were analyzed. At a depth of 75 m at the ADCP location, the ADCP data and model data yield the same 9-month speed average of 0.94 m s−1, and a fairly uniform 45° current direction. The power density time series shows that the ADCP observations occasionally reach 4,500 W m−2 with an average of nearly 800 W m−2. Recent offshore wind studies in North Carolina have demonstrated that winds produce an average hub-height power density of 600–800 W m−2, close to the ADCP power density average of 798 W m−2. This suggests that Gulf Stream subsurface turbines may become viable as turbine and mooring technology develop.

Kabir et al. (2015) used HYCOM and a grid bounded by 34.9° N to 35.2°N and 74.9–74.5°W to find that more than 50% of the days from November 2003 through December 2012 at 20 m depth exhibit a power density of 500 W m−2 or greater. This coincides with the offshore end of the CH transect near Station 14. From Fig. 10 it can be seen that approximately 48% of the time both the model and ADCP site yield power densities greater than or equal to 500 W m−2. Figure 10 and Kabir et al. (2015) use different models to conclude that nearly the same power densities of 500 W m−2 are available for the same percentage of the time. The small difference between the two percentages is likely due to the analyzed model grid location being farther offshore than the ADCP site.

Yang et al. (2014) define the recoverable resource as the amount that can be extracted within current technological limits. Given the present uncertainty in underwater turbine technology and engineering capabilities, it is difficult to provide with confidence a determination of the recoverable energy resource available from the Gulf Stream. However, here, we have provided a view into the energy resource available and demonstrated its spatial and temporal variations along the North Carolina continental slope. Quantifying the Gulf Stream power characteristics is an essential step for the development of ocean current energy in this region. Accurate approximations of the energy resource and ocean environment will lay the foundation for extracting Gulf Stream energy in the future. As society continues to exploit fossil fuels and non-renewable resources at great cost to the environment, renewable energy increasingly becomes a more attractive long-term alternative. In helping to reach the goal of 80% of US electricity generation from renewable energy resources by 2050 (National Renewable Energy Laboratory 2012), ocean energy can be an essential component of a renewable energy portfolio that takes advantage of continuing technological advances in resource characterization, turbine design, mooring technology, and ocean engineering.



We gratefully acknowledge funding from the North Carolina Renewable Ocean Energy Program for continued support of MHK energy research off North Carolina. We also appreciate Joe DeCarolis, Billy Edge, Mo Gabr, Harvey Seim, and Jim VanZwieten for their suggestions. Thank you to Patterson Taylor for supplying Fig. 6 and to Zhaoqing Yang and Kevin Haas for their comments that helped improve this chapter.


  1. Bane, J. M., Brooks, D. A., & Lorenson, K. R. (1981). Synoptic observations of the three-dimensional structure and propagation of Gulf Stream meanders along the Carolina continental margin. Journal of Geophysical Research, 86, 6411–6425.CrossRefGoogle Scholar
  2. Bane, J. M., & Dewar, W. K. (1988). Gulf Stream bimodality and variability downstream of the Charleston bump. Journal of Geophysical Research, 93, 6695–6710.CrossRefGoogle Scholar
  3. Bane, J. M., He, R., Muglia, M., Lowcher, C. F., Gong, Y., & Haines, S. M. (2017). Marine hydrokinetic energy from western boundary currents. Annual Review of Marine Science. doi:10.1146/annurev-marine-010816-060423. Invited Paper.
  4. Barringer, M. O., & Larsen, J. C. (2001). Sixteen years of Florida current transport at 27° N. Geophysical Reseach Letters, 28, 3179–3182.CrossRefGoogle Scholar
  5. Boehlert, G. W., & Gill, A. B. (2010). Environmental and ecological effects of ocean renewable energy development: a current synthesis. Oceanography, 23, 68–81.CrossRefGoogle Scholar
  6. Brown, A., Beiter, P., Heimiller, D., Davidson, C., Denholm, P, Melius, J., et al. (2015). Estimating renewable energy economic potential in the United States: methodology and initial results. Tech. Rep. NREL/TP-6A20-64503, Natl. Renew. Energy Lab, Golden, CO.Google Scholar
  7. Chassignet, E. P., Hurlburt, H. E., Smedstad, O. M., Halliwell, G. R., Hogan, P. J., Wallcraft, A. J., et al. (2007). The HYCOM (Hybrid Coordinate Ocean Model) data assimilative system. Journal of Marine Systems, 65, 60–83.CrossRefGoogle Scholar
  8. Corren, D., Hughes, S., Paquette, J., Sotiropoulos, F., & Calkins, J. (2013). Improved structure and fabrication of large, high-power KHPS rotors. Tech. Rep. DOE/GO18168, Verdant Power, New York, NY.Google Scholar
  9. Gong, Y., He, R., Gawarkiewicz, G. G., & Savidge, D. K. (2015). Numerical investigation of coastal circulation dynamics near Cape Hatteras, North Carolina, in January 2005. Ocean Dynamics, 65, 1–15.CrossRefGoogle Scholar
  10. Halkin, D., & Rossby, T. (1985). The structure and transport of the Gulf Stream at 73° W. Journal of Physical Oceanography, 15, 1439–1452.CrossRefGoogle Scholar
  11. Imawaki, S., Bower. A., Beal, L., & Qiu, B. (2013). Western boundary currents. In G. Siedler., S. M. Grifies., J. Gould & J. A. Church (Eds.), Ocean Circulation and Climate: A 21st Century Perspective, 2nd ed. (pp. 305–38).Oxford, UK: Academic.Google Scholar
  12. Kabir, A., Lemongo-Tchamba, I., & Fernandez, A. (2015). An assessment of available ocean current hydrokinetic energy near the North Carolina shore. Renewable Energy.
  13. Li, B., Bane, J., DeCarolis, J. F., Neary, V., de Queiroz, A. R., & Keeler, A. G. (2017). The economics of ocean current energy: a Gulf stream case study. Submitted to Nature Energy.Google Scholar
  14. Luettich, R. A., Birkhahn, R. H., & Westerink, J. J. (1991). Application of ADCIRC-2DDI to Masonboro Inlet, North Carolina: A brief numerical modeling study, Contractors Report to the US Army Engineer Waterways Experiment Station, August 1991.Google Scholar
  15. Mellor, G. L., & Yamada, T. (1982). Development of a turbulence closer model for geophysical fluid problems. Reviews of Geophysics, 20, 851–875.CrossRefGoogle Scholar
  16. Miller, J. L. (1994). Fluctuations of Gulf Stream frontal position between Cape Hatteras cand the Straits of Florida. Journal Geophysical Research, 99, 5057–5064.CrossRefGoogle Scholar
  17. National Renewable Energy Laboratory. (2012). Renewable Electricity Futures Study. Hand, M. M., Baldwin, S., DeMeo, E., Reilly, J. M., Mai, T., Arent, D., Porro, G., Meshek, M., Sandor, D. (Eds.), 4 vols. NREL/TP-6A20-52409. Golden, CO: National Renewable Energy Laboratory.Google Scholar
  18. Neary, V. S., Previsic, M., Jepsen, R. A., Lawson, M. J., Yu, Y.-H., et al. (2014). Reference model 4 (RM4): ocean current turbine. In V. Neary, M. Previsic, R. A. Jepsen, M. J. Lawson, Y.-H. Yu, et al. (Eds.), Methodology for Design and Economic Analysis of Marine Energy Conversion (MEC) Technologies (pp. 180–228). Albuquerque, NM: Sandia Natl. Lab.Google Scholar
  19. Quattrocchi, G., Pierini, S., & Dijkstra, H. A. (2012). Intrinsic low-frequency variability of the Gulf Stream. Nonlinear Processes in Geophysics, 19, 155–164.CrossRefGoogle Scholar
  20. Richardson, P. L. (1977). On the crossover between the Gulf Stream and the Western Boundary Undercurrent. Deep Sea Research, 24, 139–159.CrossRefGoogle Scholar
  21. Shchepetkin, A. F., & McWilliams, J. C. (2005). The regional ocean modeling system (ROMS): a split-explicit, free-surface, topography-following-coordinate oceanic model. Ocean Modelling, 9, 347–404.CrossRefGoogle Scholar
  22. Tracey, K. L., & Watts, D. R. (1986). On Gulf Stream meander characteristics near Cape Hatteras. Journal Geophysical Research, 91, 7587–7602.CrossRefGoogle Scholar
  23. VanZwieten, J. H., Duerr, A. E. S., Alsenas, G. M., & Hanson, H. P. (2013). Global ocean current energy assessment: an initial look. In Proceedings of the 1st Marine Energy Technology Symposium, April 10–11, Washington, DC.
  24. VanZwieten, J., McAnally, W., Ahmad, J., Davis, T., Martin, J., Bevelhimer, M., et al. (2014). In-stream hydrokinetic power: review and appraisal. The Journal of Energy Engineering, 141, 04014024.CrossRefGoogle Scholar
  25. Webster, F. (1961). A description of Gulf Stream meanders off Onslow Bay. Deep Sea Research, 9, 130–143.CrossRefGoogle Scholar
  26. Yang, X., Haas, K. A., & Fritz, H. M. (2014). Evaluating the potential for energy extraction from turbines in the Gulf Stream system. Renewable Energy, 72, 12–21.CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Caroline F. Lowcher
    • 1
  • Michael Muglia
    • 2
  • John M. Bane
    • 3
  • Ruoying He
    • 4
  • Yanlin Gong
    • 4
  • Sara M. Haines
    • 3
  1. 1.Scripps Institution of OceanographyUniversity of CaliforniaSan Diego, La JollaUSA
  2. 2.University of North Carolina Coastal Studies InstituteWancheseUSA
  3. 3.Department of Marine SciencesUniversity of North CarolinaChapel HillUSA
  4. 4.Department of Marine, Earth, and Atmospheric SciencesNorth Carolina State UniversityRaleighUSA

Personalised recommendations