Aquacultural Engineering 32 (2004) 77–94 The design and analysis of a four-cage grid mooring for open ocean aquaculture David W. Fredriksson a,∗ , Judson DeCew a , M. Robinson Swift b , Igor Tsukrov b , Michael D. Chambers c , Barbaros Celikkol b b a Center for Ocean Engineering, University of New Hampshire, Durham, NH 03824 USA Department of Mechanical Engineering, University of New Hampshire, Durham, NH 03824 USA c Open Ocean Aquaculture Manager, University of New Hampshire, Durham, NH 03824 USA Abstract In an effort to expand open ocean aquaculture operations, a submerged, four-cage grid mooring system was designed, analyzed and deployed in 52 m of water at an exposed site maintained by personnel at the University of New Hampshire. Mooring system geometry, subsurface flotation and pretension requirements were specified using analytical techniques, which included standard chain catenary equations and equilibrium analysis. Mooring gear and ground tackle were sized, in part, by modeling the designed system using a finite element program. The model included representations of potential sets of mooring gear and four Sea StationTM submersible cages (totaling 7200 m3 of containment volume). Numerical simulations were performed using a wave height of 9 m, wave period of 8.8 s and a current of 1 m/s as input. Using the results of the simulations, along with practical experience, a system design load of 178 kN was obtained. The design load was used to specify mooring components and the gear was deployed during the summer of 2003. The species being raised in the system include Atlantic halibut (Hippoglossus hippoglossus), Atlantic cod (Gadus morhua) and haddock (Melanogrammus aeglefinus). Future work will incorporate the development of automated, high capacity feeding systems for servicing submerged cages, harvesting systems and uniquely engineered support vessels. © 2004 Elsevier B.V. All rights reserved. Keywords: Hydrostatic analysis; Numerical modeling; Component specification ∗ Corresponding author. Tel.: +1 603 862 0273; fax: +1 603 862 0241. E-mail addresses: dwf@cisunix.unh.edu (D.W. Fredriksson), jcdc@cisunix.unh.edu (J. DeCew), mrswift@cisunix.unh.edu (M.R. Swift), igor.tsukrov@unh.edu (I. Tsukrov), mdc6@cisunix.unh.edu (M.D. Chambers), celikkol@cisunix.unh.edu (B. Celikkol). 0144-8609/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.aquaeng.2004.05.001 78 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 1. Introduction As environmental and utilization issues put pressure on existing near shore aquaculture facilities, the need to move operations into more exposed sites is becoming necessary. The technologies required to perform economic open ocean aquaculture, however, are still in the process of being developed. The University of New Hampshire (UNH) operates an open ocean aquaculture site in 52 m of water approximately 10 km from the New Hampshire coast in the United States (Fig. 1). The site is permitted to perform research related to the operational, engineering, biological and environmental aspects of open ocean aquaculture. To support the research, two independent 600 m3 Sea StationTM fish cages (SS600) were initially designed and deployed at the site in 1999 using separate, robust mooring systems (Fredriksson et al., 1999; Tsukrov et al., 1999; Tsukrov et al., 2000; Fredriksson et al., 2000; Baldwin et al., 2000). For over 4 years, these systems were the focus of intense engineering and operational studies. From the engineering perspective, analyses were conducted to investigate system dynamics so that numerical and physical modeling techniques could be developed to cost-effectively engineer and specify equipment suitable for deployment. The first Fig. 1. The open ocean aquaculture site is located off the coast of New Hampshire, USA in the southwest corner of the Gulf of Maine (figure downloaded from http://spo.nos.noaa.gov and annotated). D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 79 experiment performed consisted of a SS600 cage drag test measuring towing line tension, and relative velocity inside and outside of the cage (Fredriksson et al., 2003a). Results of the test provided full-scale velocity–tension pairings and an estimate of the amount of the velocity reduction through the nets for comparison with the modeling tools. The second experiment was a free release test with the SS600 cage to investigate the added mass, damping and natural period characteristics of the cage in the heave direction (Fredriksson et al., 2003b). In the third and most extensive study (Fredriksson et al., 2003c), field measurements of surface waves, fish cage motion response in heave, surge, pitch and anchor line tensions were made using instrumentation described in Irish et al. (2001). Transfer function calculations, using the measured data sets, were performed for the SS600 cage motion response and the anchor line tension. The normalized transfer functions were then compared with results from similar tests using the models. For design and analysis purposes, it was found that the modeling techniques are capable of simulating the primary physical processes associated with the dynamics of these fish cage and mooring systems. Recently, in an effort to expand bio-mass capacity at the site, the two small mooring systems were replaced with a larger four-grid mooring, enabling the deployment of additional containment structures. The new mooring system also allows auxiliary equipment, such as feeding platforms (Rice et al., 2003; Fullerton et al., 2004), to be installed at the site. The intent is to approach commercial level operations so that proper economic and environmental assessments can be initiated. The objective of this paper is to describe the engineering design process used to specify components of the four-grid mooring system in support of this goal. The approach includes: (1) a review of the specific design criteria for the open ocean site including the conceptual design, (2) application of standard analytical methods to investigate the system hydrostatic characteristics, (3) construction of a numerical model and comparison of static simulations with values calculated analytically and (4) analyzing the results of dynamic simulations using a deterministic design wave condition with a superimposed, co-linear current. Using the results of the model simulations, along with practical experience, a design load was determined, components specified, procured and the system deployed. A brief description of the deployment of the system is also provided. 2. Design criteria Design criteria specific for the site off the coast of NH were established prior to the engineering analysis and specification of components. First, it was required that the entire system fit into the site boundaries specified in the government permit for the previously installed equipment. It was also necessary that the new mooring be able to accommodate the two existing SS600 cages and have space for additional fish containment structures. The gear had to be able to withstand the waves and currents that occur at the site, especially those associated with extreme storms. Finally, the mooring system needed to be designed to minimize entanglement of marine mammals. The first design constraint required that the new mooring system be deployed in the existing permitted site approximately 10 km from the shore. The site is in 52 m of water and has a 30 acre (12.41 ha) area. The bottom composition consists of relatively heterogeneous materials, which include bedrock outcroppings, gravel and muddy sands (Grizzle et al., 80 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 2003). The site is fully exposed from nearly all directions, though a small set of islands is located approximately 2 km to the north (Fig. 1). In this study, a deterministic wave height of 9 m with a period of 8.8 s is used with a co-linear current of 1 m/s for design purposes. The design wave height of 9 m is estimated to be the energy based significant wave height (Hmo ) of a 50 year storm at the site (Fredriksson, 2001). The design wave period is approximately the average dominant wave period of the most frequent wave directional band. Included in the design condition is a superimposed, a co-linear current of 1 m/s (constant with depth). Although the largest velocity measured in 2 years of observation was 0.6 m/s (due to internal waves), the design coastal current value was chosen to encompass other coastal current components due to tidal forcing, surface winds and storm surge. The mooring system was specified to have a four-cage capacity. Two of the systems consist of the previously deployed SS600 fish cages. For mooring design purposes, two 3000 m3 Sea StationsTM (SS3000) were considered for the other two cage locations, though other commercial fish cages can be accommodated. The operational plan, however, was to deploy one SS3000 and reserve the fourth cage location for future deployments of experimental systems. Both the SS600 and SS3000 have a similar construction (Fig. 2). The cages are built around a central spar buoy and rim, both made of galvanized steel and can be submerged by ballasting a chamber inside the central spar. The structure is held in a semi-rigid configuration by tensioning stays between these primary components. The containment net is woven into the stays, therefore maintaining a constant volume. Component details can be found in Fredriksson et al. (2003c) and Kurgan (2003) for the SS600 and SS3000 cages, respectively. The information can also be obtained from Net Systems, Inc. located in Bainbridge, WA USA. Fig. 2. The SS600 and SS3000 cages each consist of a central spar and rim held together with tensioned stays. The spar on the SS600 has a length of 9 m while the spar on the SS3000 is about 15 m. The nominal rim diameter of the SS600 is 15 m while on the SS3000 it is 25 m. Each cage incorporates a ballast weight suspended with a pendent line from the spar weighing 19 and 53 kN for the SS600 and the SS3000 cages, respectively. D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 81 Fig. 3. An isometric view of the submerged four-cage grid system. Another important consideration is marine mammal entanglement. Though no gear deployed in the open ocean will eliminate this serious issue, efforts must be made to incorporate designs that minimize the effect on the marine mammal population. One approach utilizes a submerged grid using large diameter ropes (44–52 mm) that are pre-tensioned in the deployment process. In the previously deployed single-cage submerged grid, the design (minimum) pre-tensioned value was estimated at 2.2 kN (Fredriksson et al., 2000). In the 4 year deployment period, it was fortunate that not one entanglement occurred. While it is recognized that few large marine mammals actually enter the site, the criterion was doubled in the design of the four-grid mooring system. The mooring system concept was designed to be placed at a depth of approximately 15 m and consists of nine nodes (Figs. 3 and 4). Four sets of bridle lines connect each cage to the submerged grid. The grid is anchored to the bottom using 12 mooring legs each incorporating co-polymer rope and a chain catenary. Tension in the system is maintained using subsurface flotation at the nine nodal locations. Due to the 12 anchor design, flotation elements at the corners are required to be larger than those at the grid sides to accommodate the weight of chain for two anchor legs. During the deployment process, the anchors are set to form the required geometry, which submerges the flotation elements down to the desired depth and lifts chain up off the bottom. The chain catenary in the anchor legs provides compliance to the system, while maintaining static pre-tensioning. 3. Engineering approach 3.1. Hydrostatic analysis Using the design concepts introduced in the previous section, the hydrostatic and geometric configuration of the submerged grid mooring was estimated using a standard analytical approach. Tension loads in the anchor legs and the desired geometry of the system to 82 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 Fig. 4. A top view of the submerged grid mooring system. It consists of 8 corner anchor legs, 4 side anchor legs, 1 center anchor line, 12 grid lines, and 16 bridle lines. The anchors on the north and east sides are labeled for load identification. maintain the static shape of the grid were estimated using the inextensible cable (catenary) equations, as defined by Faltinsen (1990). A similar approach was used in the design of a double submerged grid aquaculture mooring as described in Fredriksson et al. (1999). These equations were also used to specify the chain forming the catenary in the anchor leg and the submerged flotation at the grid. The schematic shown on Fig. 5 defines the components of one anchor leg of the mooring system. For the four-grid mooring configuration, it was required to have the pre-tensioned subsurface grid at a depth of 15 m for relatively easy diver serviceability. In general, the approach assumes a certain geometric configuration. For instance, since the average depth of the water at the site is 52 m, the grid plane is approximately 37 m off the bottom (defined as dv on Fig. 5). The anchor legs, which are made of rope (lr ) and chain (lc ), are not identical. It was decided that the eight corner legs were to be made up of 36.5 m of chain and 78 m of rope resulting in a scope (anchor leg length to depth ratio) of 3.1. The four side anchor legs, however, were different, being composed of 27.4 m of chain and 78 m D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 83 Fig. 5. Anchor leg definition sketch. of rope and therefore having a scope of 2.9. The horizontal component of the corner and side anchor legs (dH ) was defined to be 107 and 96 m so that the entire system could fit into the 30 acre (12 ha) site, assuming that the grid lines are 65 m. For the static condition, the amount of chain designed to be in the corner and side anchor leg catenaries (lc ) was 11 and 6 m, respectively. Therefore, 25.5 and 21.4 m of chain is located on the bottom (lb ) for each corner and anchor assembly. Stud-link chain with a wet weight (wc ) of 460 N/m (reused from the previously deployed single submerged grid mooring system) was chosen. Using these values, the static pre-tension and the geometry of the submerged grid mooring were determined using the following analytical approach. The position of the grid corner, with respect to the anchors, is defined by: dH = lb + xc + lr cos θb (1) dv = yc + lr sin θb (2) and where θ b is the angle formed at the top of the catenary. In Eqs. (1) and (2), the horizontal (xc ) and the vertical (yc ) components of the chain catenary are defined as: 
Th lc wc xc = (3) sinh wc Th and yc = 
  Th wc x c cosh −1 wc Th (4) where the horizontal (Th ) and vertical (TV ) tension components at the flotation node are expressed as: Th = TA cos θb (5) 84 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 and Tv = wc lc = TA sin θb (6) where TA is the tension in the anchor line. In Eqs. (1) through (6), six unknowns were determined by solving the equations iteratively. To calculate the tension in the horizontal grid lines, equilibrium analysis was applied at each flotation node. 3.2. Numerical model A numerical model was also used to analyze the mooring. The numerical model employs the finite element analysis approach in which wave and current loadings on truss, buoy and net elements (representing aquaculture gear) drive the system dynamics. The computer program incorporates a nonlinear Lagrangian approach to account for large displacements of structural elements (Gosz et al., 1996; Tsukrov et al., 2003). In addition, the unconditionally stable Newmark direct integration scheme is adopted to solve the nonlinear equations of motion. Hydrodynamic forces on the structural elements are calculated using the Morison equation (Morison et al., 1950) modified to include the relative motion between the structural element and the surrounding fluid. Following Haritos and He (1992), the fluid force per unit length acting on a cylindrical element is represented as: f = C1 VRn + C2 VRt + C3 V̇n + C4 V̇Rn (7) where VRn and VRt are the normal and tangential components of the fluid velocity relative to the structural element, V̇n is the normal component of total fluid acceleration and V̇Rn is the normal component of fluid acceleration relative to the structural element. Bold letters are used to denote vectors. The coefficients in the formula above are given by C1 = 1/2ρw DCn VRn , C2 = Ct , C3 = ρw A and C4 = ρw ACa , where D and A are the diameter and the cross-sectional area of the element in the deformed configuration, ρw is the water density, Cn and Ct are the normal and tangential drag coefficients. A value of one was used for the added mass coefficient (Ca ) following the work of Bessonneau and Marichal (1998). Note that Cn and Ca are dimensionless, while Ct has the dimension of viscosity. Eq. (7) is known to adequately predict the hydrodynamic force on a submerged cylindrical element whose diameter is small compared to the length of the wave (Haritos and He, 1992; Webster, 1995; Tsukrov et al., 2000). The numerical procedure calculates Cn and Ct using a method described by Choo and Casarella (1971) that updates the drag coefficients, as a function of the Reynolds number (Ren ) according to,  8π  −2    Ren s (1 − 0.87 s ) (0 < Ren ≤ 1), Cn = (8) 1.45 + 8.55Re−0.90 (1 < Ren ≤ 30),  n    1.1 + 4Re−0.50 (30 < Ren ≤ 105 ) n and 2/3 Ct = πµ(0.55Re1/2 n + 0.084Ren ) (9) where Ren = ρw DVRn /µ, s = −0.077215665 + ln(8/Ren ) and µ is the water viscosity. D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 85 Recently, the consistent net element was constructed to adequately account for the hydrodynamic forces acting on any section of netting or an entire net panel (Tsukrov et al., 2003). It uses one-dimensional net elements since the development is straightforward and they are easily compatible with the existing finite element codes used to analyze mooring systems and floating structures. Any two-dimensional net panel can be modeled as a set of perpendicular or inclined one-dimensional net elements. The hydrodynamic behavior of the proposed net element is based on the Morison equation. Since drag force and inertia force in this equation are uncoupled, the parameters can be treated separately. 4. Results 4.1. System hydrostatics To size the grid flotation elements and to determine the required geometry and pretension values of the mooring, the analytical techniques described in Section 3.1 were applied. Using a total vertical force of 5.1 and 2.8 kN for the corner and side grid flotation nodes, the static anchor leg tensions (TA ) for the corner and side were calculated using Eqs. (1) through (6), resulting in values of 12.52 and 6.62 kN, respectively. At the grid node locations, equilibrium analysis was applied to obtain grid line tensions of 11.4 and 6.01 kN for the exterior (outside square) and interior (connecting to the center node) grid lines, respectively. The next step was to build a numerical model of the entire system and perform a hydrostatic simulation (i.e. no wave or current loading was applied). Geometric and material properties used in the model were based on the components described in Section 3.1 and the properties calculated as part of the analytical analysis (i.e. the size of the corner floats and geometry). Cage characteristics used in the model are discussed in Section 2. For the hydrostatic numerical model tests, the entire fish cage and mooring system was released and allowed to come to static equilibrium for a period of 30 s. Tensions in the anchor and grid lines were calculated and the results provided in Table 1. After the transient portion of the simulation, the corner and side anchor line pretension values were calculated to be 12.84 and 7.30 kN, respectively. The exterior and interior grid lines were determined to have tensions of 11.69 and 6.67 kN, respectively. Results of the numerical model compared reasonably well (within 10%) with those calculated analytically. The static tension results are different for the two methods because the analytical approach utilizes inextensible catenary equations while the numerical approach considers mooring line elasticity. Table 1 Analytical and numerically modeling static load results Corner anchor line Side anchor line Grid line (outside edges) Grid line (interior) Analytical (kN) Numerical (kN) Difference (%) 12.52 6.62 11.41 6.01 12.84 7.30 11.64 6.67 2.5 9.3 1.9 10.0 86 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 Fig. 6. Selected mooring line load results using one of the UNH design conditions. The geometric difference due to stretching the mooring lines slightly changes the static tension. 4.2. Dynamic simulations Dynamic simulations were performed using one of the UNH design conditions consisting of a deterministic wave with a height of 9 m and a period of 8.8 s coming from the northeast direction. Model simulations were used to calculate mooring line loads. Time series results for the anchor line tension from the numerical model are shown on Fig. 6. Maximum steady state values were calculated to be 147 and 132 kN for the side and corner anchor line assemblies, respectively. One advantage of the data set resulting from performing numerical model simulations is that the tension from a variety of components in the mooring system can be analyzed. From this data set grid line tensions were also investigated. This information is important in the understanding of how the cages transfer loads to the anchor legs and the ground tackle. It was found that the grid line tensions in the northeast quadrant of the mooring were major load bearing components with values ranging from 70 to 120 kN when the design condition was applied (Fig. 7). In addition, it was found that a majority of the southwest SS3000 cage loads are transferred through the grid to the two side anchors. This information was vital to D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 87 Fig. 7. The maximum load distribution in the mooring using one of the UNH design conditions. The current and waves are applied from the northeast direction. determining how the loads are distributed throughout the grid so each component can be specified. 4.3. Component specification and deployment The use of the numerical model is important in the design of these mooring systems. It is imperative, however, to understand that the tension values calculated are only approximations. Modeling variability associated with choosing correct material and geometric properties, appropriate forcing and other physical characteristics, not necessarily represented, creates some uncertainty. For example, the numerical modeling approach does not take into consideration (explicitly) the effects of net “blockage” (for example see Aarsnes et al., 1990; Fredriksson et al., 2003a,b,c) or the change in drag and mass due to biological fouling. Therefore, developing an appropriate design load also incorporates knowledge obtained from practical experience. A team of engineering and operational personnel discussed 88 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 Table 2 The mooring system particulars Component Description M.B.L.a Anchor (12) Construction Mass Drag Embedment 1000 kg 178 kNb Side anchor chain (4) Construction Length Mass Stud-Link 27.4 m 706 kg/m Corner anchor chain (8) Construction Length Mass Stud-Link 37.5 m 706 kg/m Anchor line (12) Construction Length Specific gravity Diameter 8-plait co-polymer 78 m 0.94 48 mm Side grid flotation (4) Construction Mass Diameter Steel 136 kg 0.9525 m 894 kN 894 kN 390 kN 370 mc 47 mc Corner and center grid flotation (5) Construction Mass Diameter Urethane Foam Comp. 295 kg 1.45 m Corner rope ring/chain Construction Mass Length 25.4 mm steel long-link 61.33 kg 2.0 m Grid line Construction Length Specific gravity Diameter 8-plait co-polymer 65 m 0.94 48 mm Shackles Construction Mass Diameter Galvanized Steel 7.25 kg 38 mm a b c d 444 kNd 390 kN 756 kN Minimum breaking load. Holding power. The flotation elements are rated at working depth. The corner rope ring was tested by manufacturer to a load of 444 kN without failure. D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 89 the modeling results, along with deployment and maintenance implications. Based on these discussions, the maximum loads calculated with the numerical model were increased by 17% to obtain a design load of 178 kN. Along with the cost and operational factors, this design load was used to specify mooring system components. Many of the mooring parts used in the previous single-cage grid deployments, including 8 of the 12 embedment anchors (and chain), as well as 4 of the side flotation elements, were reused to reduce costs. The anchors were chosen to have the smallest safety factor relative to other mooring components. Even though conservative environmental conditions were used in the design process, often more extreme or unplanned events occur at the site. If a more extreme event is encountered, the intent is to have the anchors “drag” to relieve system stress before actual structural damage occurs in the other components (e.g. mooring rope, shackles). Depending upon the direction of the waves and currents, the grid lines are important members for the transfer and distribution of loads to the anchor lines. Therefore, all of the mooring rope (grid and anchor) were sized using the same design load requirements. This also helped to reduce costs since the rope could be purchased in bulk quantities. The lines are held in place by 38 mm shackles, which have been found to be the limit of easy diver serviceability. Prior to deployment, each shackle pin was welded to prevent them from becoming undone. The grid corner flotation was sized not only to tension Fig. 8. Component details of the corner grid mooring assembly. Some items are not to drawn to scale. 90 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 Fig. 9. Component details of the side grid mooring assembly. Some items are not to drawn to scale. D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 91 two anchor legs, but also to offset any biological fouling that may occur, allowing greater flexibility in routine cleaning operations. The mooring system components are listed in Table 2. Schematics of the grid corner and side and center anchor leg assemblies are shown in Figs. 8–10. The location of the mooring grid within the permitted site was determined using bottom topography information obtained courtesy of Center for Coastal and Ocean Mapping/Joint Hydrographic Center (CCOM/JHC). The gear was successfully deployed on the first week of July 2003, using the F/V Nobska operated by Stommel Fisheries from Woods Hole, MA. Anchor locations were first determined using Differential Global Positioning System instrumentation based on the design geometry calculated using the catenary equations. The gear was deployed “slack” with each of the grid floats at the surface. A 15 m line with indicator floats was attached to each of the grid floats. Next, the anchors were pulled out with the fishing vessel to the predetermined positions. The anchors were set when only the indicator floats of the 15 m lines were visible, “indicating” that the grid was at the proper depth. This technique allows the vessel operator to accurately position the anchors and grid since the inextensible catenary equations do not take into consideration stretching of the rope and bottom contour variability. Fig. 10. Component details of the center grid mooring assembly. Some items are not to drawn to scale. The center node is held down by a 1800 kg steel deadweight. 92 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 In the second week of July 2003, the two SS600 cages used with the previous mooring, containing Atlantic halibut (Hippoglossus hippoglossus) and haddock (Melanogrammus aeglefinus) were connected to the new grid. A previously constructed feed platform was attached to one of the SS600 cages containing the haddock the following week. In August, a new submersible SS3000 cage was deployed. Shortly thereafter, approximately 30,000 Atlantic cod (Gadus morhua) were transferred from a shore side facility into a nursery net located inside of the cage. The next significant challenge will be to deploy a floating feed buoy system to automatically feed the codfish in the subsurface SS3000 cage. The design and construction of this feed buoy is currently underway (some of the engineering details are provided in Fullerton et al., 2004). 5. Conclusion The design of UNH submerged grid moorings for open ocean aquaculture has evolved from the single-cage grid system to the four-cage grid mooring described here. At large portion of the design effort involved the interpretation of the numerical model results. Efforts were made to accurately model the system by comparing static simulation values to those obtained using inextensible catenary equations. Results from the dynamic simulations have to be used carefully since effects of biological fouling and net blockage are not explicitly taken into account. Design load considerations, however, do incorporate operational experience, which is difficult to quantify, but not less important. The mooring system and the three submersible cages deployed are now being used to investigate the open ocean grow-out capability of halibut, haddock and cod. The grid mooring system approached is being investigated, in part, to determine the feasibility of moving many similar (multi-cage) surface grid systems used for inshore aquaculture to more exposed sites. In addition to the mooring design issues, other significant engineering challenges exist to make open ocean aquaculture more economically feasible. Only a few have considered the implications of automated, high capacity feeding to submerged cages. As grow-out takes place, it will become more evident that other technologies such as harvesting systems and uniquely engineered support vessels need to be developed in support of cost-effective open ocean aquaculture. The development of these and other technologies will require an integrated systems engineering approach considering biological and environmental design criteria. The engineering and operational costs must be balanced with the appropriate seafood product and market and reflect responsible environmental practices. The challenges are substantial, but many believe that open ocean aquaculture can be performed economically at a scale that can make a significant impact on the global need for seafood. Acknowledgements Funding for this research was provided by the NOAA grant no. NA16RP1718 to the UNH Cooperative Institute for New England Mariculture and Fisheries (CINEMAR). The authors would like to thank Rich Langan (Director of CINEMAR), Oleg Eroshkin, Caleb D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 93 Thibeau, Glen Rice (University of New Hampshire), Jim Irish and Walter Paul (Woods Hole Oceanographic Institution) for their support on this project. References Aarsnes, J.V., Rudi, H., Loland, G., 1990. Current Forces on Cage, Net Deflection. Engineering for Offshore Fish Farming. Thomas Telford, London, pp. 137–152. Baldwin, K., Celikkol, B., Steen, R., Michelin, D., Muller, E., Lavoie, P., 2000. Open Aquaculture Engineering: Mooring and Net Pen Deployment. Mar. Tech. Soc. J., vol 34, no. 1. Washington DC, pp 53–67. Bessonneau, J.S., Marichal, D., 1998. Study of the dynamics of submerged supple sets (applications to trawls). Ocean Eng. 25 (27), 563–583. Choo, Y.I., Casarella, M.J., 1971. Hydrodynamic resistance of towed cables. J. Hydronautics. 126–131. Faltinsen, O.M., 1990. Sea Loads on Ships and Offshore Structures. Cambridge University Press, New York, 328 p. Fredriksson, D.W., E. Muller, M.R. Swift, I. Tsukrov, B. Celikkol, 1999. Offshore grid mooring/net pen system: design and physical model testing. OMAE99-3064. In: Proceedings of the 18th International Conference Offshore Mechanics and Arctic Engineering 99, St. John’s, Newfoundland Canada, ASME. Fredriksson, D.W., Swift, M.R., Muller, E., Baldwin, K., Celikkol, B., 2000. Open ocean aquaculture engineering: system design and physical modeling. Mar. Tech. Soc. J. 34 (1), 41–52. Fredriksson, D.W., 2001. Open Ocean Fish Cage and Mooring System Dynamics. Ph.D. Dissertation submitted to the University of New Hampshire in partial fulfillment of the Engineering: Systems Design Program. Durham, NH, September 2001. Fredriksson, D.W., M.J. Palczynski, M.R. Swift, J.D. Irish, B. Celikkol., 2003a. Fluid dynamic drag of a central spar cage. In Bridger, C.J., Costa-Pierce, B.A. (Eds.), Open Ocean Aquaculture: From Research to Commercial Reality. The World Aquaculture Society, Baton Rouge, Louisiana, United States, pp 151–168. Fredriksson, D.W., Swift, M.R., Irish, J.D., B, Celikkol., 2003b. The heave response of a central spar fish cage transactions of the ASME. J. Off. Mech. Arct. Eng. 125, 242–248. Fredriksson, D.W., Swift, M.R., Irish, J.D., Tsukrov, I., Celikkol, B., 2003c. Fish cage and mooring system dynamics using physical and numerical models with field measurements. Aquat. Eng. 27 (2), 117–270. Fullerton, B., Swift, M.R., Boduch, S., Eroshkin, O., Rice, G., 2004. Design and analysis of an automated feed buoy for submerged cages. Aqua. Eng. (In press). Grizzle, R.E., Ward, L.G., Langan, R., Schnaittacher, G.M., Dijkstra, J.A., Adams, J., 2003. Environmental monitoring at an open ocean aquaculture site in the gulf of maine: results for 1997–2000. In: Bridger, C.J., Costa-Pierce, B.A. (Eds.), Open Ocean Aquaculture: From Research to Commercial Reality. The World Aquaculture Society, Baton Rouge, Louisiana, United States, pp. 105–117. Gosz, M., Kestler, K., Swift, M.R., Celikkol, B., 1996. Finite element modeling of submerged aquaculture net-pen systems. In: Proceedings of an International Conference Open Ocean Aquaculture, May 8–10, 1996, Portland, Maine, Marie Polk (Ed. New Hampshire/Maine Sea Grant College Program Rpt. #UNHMP-CP-SG-96-9, 523–554). Kurgan, C., 2003. Fish Cage and Multiple Mooring Systems Finite Element Modeling. Master’s Degree Thesis submitted in partial requirement for the Mechanical Engineering degree program. University of New Hampshire, Durham, NH. Haritos, N., He, D.T., 1992. Modelling the response of cable elements in an ocean environment. Finite Elements Anal. Design 19, 19–32. Irish, J.D., Carroll, M., Singer, R., Newhall, A., Paul, W., Johnson, C., Witzell, N., Rice, G., Fredriksson, D.W., 2001. Instrumentation of Open Ocean Aquaculture Monitoring. Woods Hole Oceanographic Institution, Technical Report WHOI-2001-15, p. 95. Morison, J.R., Johnson, J.W., O’Brien, M.P., Schaaf, S.A., 1950. The forces exerted by surface waves on piles. Petroleum Transactions, American Institute of Mining Eng. Rice, G.A., Chambers, M.D., Stommel, M., Eroshkin, O., 2003. The design, construction and testing of the University of New Hampshire Feed Buoy. In: Bridger, C.J. Costa-Pierce, B.A.(Eds.), Open Ocean Aquaculture: From Research to Commercial Reality. The World Aquaculture Society, Baton Rouge, Louisiana, United States, pp. 197–203. 94 D.W. Fredriksson et al. / Aquacultural Engineering 32 (2004) 77–94 Tsukrov, I., Ozbay, M., Fredriksson, D.W., Muller, E., Swift, M.R., Celikkol, B., 1999. Offshore grid mooring/net pen system: finite element analysis. OMAE99-3065. In: Proceedings of the 18th International Conference on Offshore Mechanics and Arctic Engineering ’99, St. John’s, Newfoundland Canada, ASME. Tsukrov, I., Ozbay, M., Fredriksson, D.W., Swift, M.R., Baldwin, K., Celikkol, B., 2000. Open Ocean Aquaculture Engineering: Numerical Modeling. Mar. Tech. Soc. J., vol. 34, no. 1. Washington DC, pp 29–40. Tsukrov, I., Eroshkin, O., Fredriksson, D.W., Swift, M.R., Celikkol, B., 2003. Finite element modeling of net panels using consistent net element. Ocean Eng. 30, 251–270. Webster, W.C., 1995. Mooring-induced damping. Ocean Eng. 22 (6), 571–591.