Seagriculture EU: Resolving ‘Offshore’ Aquaculture

Kelson CEO Toby Dewhurst shared results from the Kelson team on how a site’s exposure energy and distance from shore affected the cost of seaweed at Seagriculture EU in Trondheim, Norway. He reported the ICES Working Group on Open-Ocean Aquaculture’s recommendation to resolve ambiguity in the term “Offshore aquaculture” by distinguishing between distance from the baseline and “exposure” as measured by a physical index. Furthermore, Dewhurst showed results from field validation of predictions of forces on large-scale seaweed farms using open-source tools. Lastly, he provided results on the levelized cost of seaweed for food and carbon as informed by those validated models.

Do you agree with the conclusion that decoupling these terms is necessary? Let us know if this is relevant to your work.

Have you tried our exposure calculator? If not, check it out!

Introduction

Hi Everybody, I'm Toby. I'm an ocean engineer with Kelson Marine. At Kelson we have a unique skill set designing and de-risking offshore seaweed farms. A lot of that comes from our work with Advanced Research Projects Agency – Energy (ARPA-E), where our job has been to develop the tools and techniques to accurately and efficiently predict the forces on these large structures from waves and currents. This unique skill set allows us to work with clients around the world on designing seaweed farms. Several clients have asked me, “So, is anyone making any money doing this?” It's a hard industry that we've collectively picked. There's conflicting data about the profitability of seaweed, and it's made even more controversial by the question of whether you can make money moving farming into deeper, more offshore, and more exposed sites. That's part of the question that we wanted to answer today, but before we can answer the question of whether offshore farming is economical, we have to answer the question “what is offshore?”

Defining “Offshore”

I'm part of the IEC's working group on open ocean aquaculture. This is a group of farmers, engineers, scientists, and other stakeholders, some of whom are here in this room. Part of our mission has been to deliver clear, concise, and useful terminology so that stakeholders can effectively communicate risks, costs, and other data around offshore or open ocean or exposed seaweed farming. When we talk about the term “offshore,” take a site off the coast of Maine where we've had several experimental farms over the past five years. It's less than a kilometer from shore, but it's exposed to 6,000 kilometers of open ocean to the southeast. The economics of that site are going to be the same as other near shore sites. After about three years of research and debating within our team, we came to one simple conclusion when it comes to defining this. We need to decouple the term offshore into two separate quantifiable metrics: Distance from Shore and Exposure Energy.

Distance From Shore

Distance From Shore is readily quantified, but we need to develop a new index to capture the complex interactions of currents, waves, and water depth into a single number that stakeholders can use. We actually proposed six different metrics that we've been evaluating and testing. The first two are combination of the current velocity and the wave induced fluid dynamic velocity. The next is the specific exposure energy, which is essentially energy per unit mass, and conveniently is also proportional to drag forces normally on structures. We also looked at two metrics which are the depth integrated energy and energy flux. The first four proposed metrics here are exclusively dependent on the site, and the last two also take into account the characteristic dimensions of the structure where the last one is. This is effectively the representation of the ratio of drag force to buoyant force on these components. For simplicity, I'm going to focus on one of these which I also happen to like the best. For a region, we're going to study the German bite of the North Sea, and we're studying this area in particular because it's an area of interest for a co-location with offshore wind arks.

Exposure Energy

Distance from Shore is shown on the left, but when we map that specific Exposure Energy for each site what we find is kind of fascinating. As we know, sheltered sites that are protected have low exposure energies. Most interesting to me is that the highest exposure energies actually happen in these near shore sites that are exposed to large waves coming into shallow water which create large horizontal fluid velocities, as well as where these constrictions create large tidal or storm-driven currents. This is why we need to decouple the term offshore into Distance from Shore and Exposure Energy.

Site Exploration

With that definition, let's look at a few sites. Going back to that initial site that's just an easy half kilometer Distance from Shore, we've been testing things like composite lines to mitigate whale entanglement, spatially dense farms, low-cost helical anchor installation technologies, and other novel technologies under the ARPA-E program. That near shore site has an exposure to energy of about five joules per kilogram, and this is a combination of strong storm-driven currents and large waves coming into shallow water.

We went ahead and mapped this, not just for this site, but for about 30 different existing and proposed aquaculture sites around the world. Some of you in this room provided data for this, so thank you very much, and this will all be in an upcoming publication that you can look at in more detail. One of the drawbacks of combining all this data into a single index is that you have to know some wave kinematics and calculate things like energy flux. To make that easier, we've developed an online calculator. You can go to kelsonmarine.com resources and put in your water depth and design conditions, and you can see where your site ranks on these six different metrics.

Economics

Now that we've established some terminology around offshore or open ocean farming, we can start to look at the economics. Taking an example site 16 kilometers off the coast of Maine, this one was selected because it's being considered for a floating offshore wind installation. We derived the current and wave design conditions, as we do for all of these projects, and this one has an exposure energy of 7.1 joules per kilogram so it's more exposed than the sites we've considered previously.

A quick note on defining those design conditions. I'll say this again and again, but when we look at the wave and the current velocities for a site, we don't design for the typical, or even the 99th percentile of the one-year storm, what we design for is the extremes: the 5-, 10-, and 50-year storms. There's a world of difference between those two, so these exposure indices actually have a return period. Your mean exposure energy will be different from your 50-year exposure energy, so you can look at the characteristics of the site. Another side note here, we use fairly traditional and maybe slightly augmented ocean engineering techniques to apply these extreme values or estimate these extreme values, but because we're in this really difficult industry where the margins are so tight we've been exploring more and more how to cut down the uncertainty around these design frameworks and drive down costs. We've been working more and more with extreme probability contours, and what we're showing here is the 10-year probability contour for the combination of a significant wave height period and wind speed. While we usually assume that wind speed peak and peak wave happen on top of each other and their forces combine, what we're showing here is that they don't ever actually happen at the same time. Moving forward we're trying to incorporate more and more of this into our design process and even into engineering guidelines so we can keep driving down those uncertainties and driving down costs.

Farm Design

Now that we've defined those design characteristics for the site, we can get into actually designing the farm. Here we immediately run into the first challenge and the first impact on cost that happens with designing structures for far offshore deep-water locations. Because we're in 100 meters of water and because you need your mooring lines to be several times the water depth to keep the farm well tensioned, a basic farm design easily loses 80 percent of its area just to the moorings. This is an effect that's usually neglected in published techno-economic analyzes of large-scale offshore aquaculture, but commercially speaking even when we're dealing with fairly large sites this is always the constraint that we're balancing. How much grow line can we fit in a single area and where's the trade-off between that and increasing forces on the moorings and anchors?

Model Accuracy and Validation

We looked at a range of different farm designs and aspect ratios designed to maximize the amount of biomass in each site, and there's a huge amount of optimization to do in just packing as much grow line per area as possible. The question then is, with those high aspect ratio farms that have more and more biomass, which economic factor dominates? Does having more biomass pay off or does the increased force on the anchors and mooring lines drive your costs up so much that this becomes economically unfeasible? The only way we can answer that is with these simulation techniques so we can accurately predict the forces on these structures and calibrate the cost. We did that for a range of farm designs, and here we get into the second and the distinct effect of farming in high exposure energies as opposed to offshore sites. In the large waves that are associated with these high exposure sites, we have to be really careful about spacing between lines. It does not work to take yield per grow line per hectare numbers from near shore sites and just put those into offshore sites. Lines on line interactions and potential biomass loss are a huge design constraint as we look on these very large farms.

We conducted this analysis on each of these different farm designs, went all the way through to details like sizing the anchors appropriately, taking into account the holding capacity, directional characteristics, etc., and got to the total cost of the components for each one of these structures. What we found is that these farms are expensive, no surprise, and as expected the largest, highest aspect ratio farm is much more expensive than the more traditional farms. When we normalized by the amount of biomass that each farm can fit in the specified permit area, which is constrained and which is the same for each one of these designs, we see a very different story.

As we get to the larger, higher aspect ratios farms, the capex cost per biomass becomes more and more efficient. Now I hope you're asking yourself a question that our clients never ask us, even though they should ask us before they ever take any of our results for granted, and that is how accurate are your models? The hydrodynamics of seaweed structures are complex. They don't fit standard engineering drag, added mass, and hydrodynamic formulations, and we've done a fairly extensive testing on full one-to-one scale models of macrocysters, saccharina, etc., in the tank. Because they don't fit traditional ocean engineering models, we've had to do this in custom software. I'm really grateful to ARPA-E for seeing the need for this. They've recently funded us to take that whole workflow which was previously tied to commercial solvers and now implement this in open-source solvers so that we can have our non-traditional hydrodynamics but also be able to optimize for speed. Right now, we're at about 10 to 16 times faster than our old workflow, so if you're in this room and I've ever told you that you have to wait two weeks for a full set of dynamic results those days are almost over. We're down to about a day for a full set of dynamic load cases.

Speed is only good as your results are accurate, so how do we know that even with all these sophisticated simulation tools that we're actually predicting the right results? Well, we measure it at full scale in the ocean, and again ARPA-E has been very helpful in funding these really difficult experiments. Partners like Ocean Rainforest and Trophic and others have been awesome collaborators as we actually measure the tensions on these farms while simultaneously measuring the waves and currents at these exposed sites. We've gotten to do this for saccharina, macrocystis and a whole range of different sites, waves, wind, and current conditions. So far, we're within about 15 percent of our particular predicted values, and we're doing this on larger and larger scales. This year Ocean Rainforest in the US put out the first truly offshore farm in the US which was really awesome to see. This was particularly gratifying for us because we've been working with them on this for about three years, even talking with regulators to provide some assurance that this was designed properly, that it’s robust, safe, and worthy of being put off state waters in California. Ocean Rainforest has done an amazing amount of work to make this happen, again all funded by ARPA-E.

This site is incidentally 6.6 kilometers from shore and the exposure energy is about 6.6. There are tons to go into when we start to look underwater at what happens at these farms in extreme storms. We could talk about the non-quadratic drag, we could talk about the wake dynamics, or the low frequency mooring tensions. What I found as I've studied the physics of these farms for about six years is that hardly ever anyone ever really cares about the physics, so we'll bring it back to the economics. When you take all this, now that we have some confidence in these models, how does all this analysis tie into the actual cost of producing seaweed in exposed locations far from shore?

Economic Viability

Well, we took all of those results from our analyses of those different farm designs and put them into a very detailed and very rigorous economic model that was put together by Struan Coleman, who's now at the World Wildlife Foundation. It includes everything from permitting costs, monitoring, anchor installation, not just anchor and mooring costs. It also includes the financial cost of financing these farms, which is actually a very significant portion because this is a high capex expenditure. Our result for the fairly optimized structural farm is a real result of about five and a half thousand dollars per dry metric ton farmgate, so prior to processing.

If we want to get a little bit more controversial, what happens if you want to do the same thing but sink the seaweed for carbon credits? Well, the first thing you have to do is account for the carbon that you are emitting in the process of growing the seaweed. It turns out that almost two-thirds of the carbon that you sequester is actually offset by the carbon you emit in the process. This becomes a very difficult metric that drives the costs of these structures, or the cost of levelized cost of carbon over a 30-year project, to results that are one or two orders of magnitude higher than markets are paying right now.

Driving Down Costs

So how do we drive these costs down, whether it's for food, for fuel, or even for carbon? Well, there are a lot of steps that we need to make happen. Scott Lindell and his team are working with a group at the University of Maine to help improve the hatchery process and maximize the yield per meter of grow line. We need people to keep working on reducing labor costs, that's a huge driver of costs. And what we're working on is maximizing the farm design – cutting down the uncertainty, cutting down the cost, and maximizing the amount of grow line in a given farm area, while keeping the costs sustainable.

Summary

• We need to decouple Distance from Shore from Exposure Energy. That will help us communicate effectively between stakeholders, regulators, insurers, investors, farmers, engineers, and scientists.
• Validation to me is everything. When I say the physics of macro algae don't follow standard ocean engineering tools, they actually don't scale with the non-dimensional numbers that we usually use to go from small scale to big scale: Froude number and Reynolds number. That's why we do all of our validation in real world, full-scale farms.
• Lastly we cannot take numbers, whether it's yield per hectare or cost per ton from near shore cultivation or sheltered site cultivation, and plug those into models for large-scale offshore or exposed aquaculture.

As a side note, we've gotten some pushback in the US about using the term exposed. In general, Europeans are more comfortable with exposure than Americans are, so I think we're okay.

Next Steps

So what's next now? Well, we're part of a multi-year process now to do this techno-economic optimization, and to do this we're closely coupling the techno-economic model with the engineering so that we can keep finding these pathways toward economic viability. A side note for you biologists and phycologists in the room: one thing we need is robust relationships between exposure energy and biomass and growth rates. Right now, that's missing. There's anecdotal evidence, but there's some conflicting evidence in the literature about whether more exposure leads to higher growth rates and when more exposure leads to lower growth rates. We need good data so that we can really optimize this for these exposed, open ocean sites.

Lastly, on the US side, we're working on new engineering guidelines specifically for low trophic shellfish and seaweed aquaculture to help drive down the uncertainty and drive down the risks for these low-trophic cultivation structures. We’re hoping this will result in new engineering guidelines so that we can reduce that uncertainty, reduce the costs while maintaining structures, and optimize and improve structures that are safe, reliable, and economically feasible.