Based on the potential for mitigating anthropogenic climate change suggested by our analysis, we chose to perform initial calculations estimating the economic feasibility of implementing OTEC coupled with DACCS at the suggested scale. While we recognize that such estimates are highly uncertain and far too general to be definitive, emerging literature can help guide us. Below we first explore the cost of deploying OTEC, followed by the cost of installing DACCS in association with OTEC.

The cost of OTEC

The cost to build OTEC plants varies greatly depending on the location, scale, and technology used to build these plants. At this time there is limited quantitative cost data available for OTEC, with most cost references based on feasibility studies from limited sources. Ref.10 attempted to provide an overview of the available cost projections for a range of OTEC plants, spanning 5 to 250 MW capacity, based on data from12. They found that due to the large overhead cost, small-scale OTEC plants (those capable of producing less than 10 MW power) have comparatively higher capital costs per kW than larger (10 to 100 MW) OTEC plants. They suggested that the capital cost of an OTEC plant with 100 MW capacity would range from USD2010 5,000/kW to 15,000/kW. These capital costs include plant components such as pumps, heat exchangers, and components for thermodynamic cycles which are considered predictable costs since these components are commercially available for other offshore operations. Other costs, such as labour, were estimated based on USD2010 prices. Therefore, for an OTEC plant with 90 MW capacity, the capital cost per plant would be USD2010 500 million to 1.5 billion10. Using inflation-adjusted conversion rates from the U.S. Bureau of Labor Statistics, this is equivalent to 700 million to 2.1 billion USD2023. To produce 3 TW of electricity in ODHMit using 90 MW OTEC plants, we needed 33,000 plants, whereas 30,700 plants producing on average 98 MW, were needed in ODLMit.

Using the lower and the upper bound of these estimates (500 million USD2023 and 2.1 billion USD2023, respectively), the required 33,000 OTEC plants in ODHMit would cost between 17 trillion USD2023 and 69 trillion USD2023, respectively. The 30,700 plants in ODLMit would cost between 15 trillion USD2023 and 64 trillion USD2023. These numbers do not consider the decrease in price that would occur as the OTEC plants become more standardized (economy of scale) or the possibility of attaching OTEC plants to decommissioned oil and gas platforms. Ref.10 estimated that closed-cycle OTEC plants (where ammonia with a low boiling point is used as the working fluid) are less expensive than open-cycle (where seawater is used as the working fluid) designs. However, open-cycle OTEC plants produce fresh water as a by-product whereas closed-cycle plants do not, so the economic benefits of freshwater production may make open-cycle OTEC systems more economically viable in the long term, particularly for plants located near small island nations or the coast.

Once OTEC is operational, the LCOE also depends on interest rates and subsidies. For a ~ 100 MW plant, it has been estimated that the LCOE could range from 0.09 to 0.24 USD2023/kWh10,11,12. Additionally11, estimated that obtaining financing from government bonds (with lower interest rates than commercial loans) could reduce the LCOE of OTEC power generation to around 0.04 USD2023/kWh for a ~ 100 MW OTEC plant. For comparison, ref.13 recently estimated the LCOE for US-based coal-powered electricity as 0.09 to 0.14 USD2023/kWh. Therefore, the LCOE of OTEC-generated electricity could be considered competitive with the LCOE of coal-generated electricity.

The cost of DACCS

One of the major factors in determining the cost of DACCS is the cost of electricity and thermal energy used to power the system. Since the energy requirement for CO2 capture and storage can be significantly larger than for other emission control systems, how DACCS is powered is an important consideration to ensure the process is both economically profitable and environmentally beneficial. Due to the number of factors influencing the cost of DACCS, there is little consensus on estimates of the technology’s cost. Cost estimates based on simple scaling relationships, yield results from ~ 50 to ~ 1,200 USD2023 per tonne CO2 sequestered27,30,31,32,33,34.

In April 2022 the International Energy Agency published a report28 exploring the role that direct air capture could play in assisting the world reach net zero emission targets. They noted that only 18 such plants existed at the time of publication, although governments and industry had committed billions in direct air capture technology investment in the years ahead. While the cost per tonne of CO2 removed is presently high, the IEA estimated that for a large-scale plant built in 2022, the cost would be between 125 and 335 USD per tonne CO2 captured. They further argued that by 2030, costs for DAC could fall to below $100 per ton of CO2 capture.

Carbon Engineering, based in British Columbia in Canada suggested that direct air capture on land could be achieved for 94 to 232 USD2018 per tonne (equivalent to 114 to 280 USD2023; 27). This is much lower than the previous American Physical Society estimate35 of 600 USD2011 per tonne CO2 (equivalent to 830 USD2023), suggesting, and consistent with ref.28, that the price of direct air CO2 capture is rapidly decreasing with technological advances and will continue to do so.

If we use the range of land-based direct air capture costs from the IEA report (between 125 USD2023 and 335 USD2023 per ton of CO2 captured, respectively), the cost to extract the 2,446 Gt of CO2 in ODHMit through direct air capture converts to between 305 trillion and 819 trillion USD2023. In the case of ODLMit, the cost to extract the 471 GtCO2 would be between 59 trillion and 158 trillion USD2023. Considering the decrease in price that would occur with electricity coming from OTEC rather than the electrical grid, these costs would decrease further. We recognize that the above estimates are for land-based direct air capture and that these costs could be expected to be different in the marine environment.

The technology for injecting CO2 into depleted or nearly depleted oil and gas reservoirs is mature and has been successfully used for decades for enhanced oil recovery. This method of carbon injection occurs when liquidized CO2 is injected into the oil-bearing formation to lower the viscosity of the oil, therefore allowing the oil to flow more easily to the oil well36. The 2019 cost for land-based enhanced oil recovery CO2 injection was estimated to be about 40 USD2019 per tonne of CO2 stored37,38, although the IEA report28 notes that in the US, land-based storage can be achieved for less than 10 USD while offshore storage cost below 35 USD per ton of CO2 stored. Assuming 35 USD2023 per ton of CO2 stored, the total cost to store 2,446 GtCO2 is 86 trillion USD2023 and to store 471 GtCO2 is 16 trillion USD2023.

Total cost

The capture and compression/liquefaction of CO2 are expected to be the dominant costs involved in DACCS, followed by transport costs39. With OTEC powering DACCS directly above tropical depleted to semi-depleted oil and gas fields, the cost associated with the transport of the liquified CO2 would be eliminated. We can also assume, due to the presence of pre-existing technology, that the cost of CO2 injection into marine sites is similar to that of DACCS on land and so use estimates from existing enhanced oil recovery operations. Taken together, we estimate the total cost for 33,000 90 MW OTEC plants that sequester 2,446 Gt (ODHMit) of CO2 over a 70-year period (Figs. 1b and 2a) using existing enhanced oil recovery technology ranges from 407 trillion to 973 trillion USD2023 (equivalent to between 166 and 398 USD2023 per ton of CO2 sequestered). The total cost for 30,700 98 MW OTEC plants that sequester 417 GtCO2 (ODLMit) over a 70-year period ranges from 90 trillion to 238 trillion USD2023. This is equivalent to between 215 and 570 USD2023 per ton of CO2 sequestered. Expected costs could be lower than these estimates as we took the conservative approach of not fully including cost savings associated with energy production from OTEC.

Technological feasibility

The current experimental efficiencies of OTEC systems are relatively low compared to theoretical estimates likely due to the latter not accounting for the efficiencies of external components such the generator, mechanical transmission, and inverter40. Efforts are currently underway to improve system efficiency, primarily through turbine design which is the most influential component40. Non-seawater working fluids are known to increase efficiency, especially in the case of refrigerant grade R717 ammonia which is considered to be the most effective working fluid, producing up to a six-times greater power output compared to alternatives40. Additionally, the use of solar energy to enhance the heating of the working fluid could increase net power generated by 20–25%41,42.

Onshore, offshore, and floating OTEC facilities offer distinct infrastructure options for implementation. Floating facilities have the potential to be relocated depending on the available ocean vertical temperature gradient. OTEC plants could also repurpose decommissioned marine oil and gas platforms43,44. This oil and gas industry infrastructure has undergone decades of improvement and is known to be durable under harsh marine conditions. The use of such existing infrastructure might allow for an acceleration of the implementation of OTEC while creating a transitional pathway from the reliance on fossil fuels44.

Environmental feasibility

Exploring the potential environmental consequences of widespread marine OTEC implementation is the subject of several recent studies. Modelling efforts have found ocean surface cooling in regions of OTEC implementation9,26,45, ocean surface warming at high latitudes9,26,45, and heating in the ocean interior26,45 due to OTEC-induced vertical mixing. Enhanced OTEC-induced vertical mixing also leads to an increased meridional gradient of depth-integrated steric height along the western boundary of the North Atlantic46, thereby reinforcing or even increasing the strength of the Atlantic Meridional Overturning Circulation (AMOC; 9, 26, 45). This might be viewed as a positive environmental effect of OTEC deployment as it counteracts AMOC weakening associated with increased atmospheric greenhouse gas loading. On the other hand, the optimal location for OTEC deployment is within the warm pool of the western equatorial Pacific. Net surface cooling in this area would almost certainly affect El Niño and monsoonal systems which future research might explore.

Near-surface discharge associated with OTEC also transports large volumes of deep-nutrient rich seawater into the photic zone45,47,48. Artificial upwelling in nutrient depleted waters has been found to alter the assemblage of phytoplankton communities, resulting in a greater percentage of micro-phytoplankton compared to pico-phytoplankton, while primary production rates remained low49. Nevertheless, a mitigative solution could involve increasing the density of the discharge plume so that neutrally buoyant settling of nutrient rich waters occurs well below the photic zone thereby preventing interference with natural primary productivity levels or shifts in the composition of phytoplankton communities44,48,49.

Other potential biological effects associated with OTEC implementation include entrainment or impingement of marine organisms in water intake pipes47,48, disruption of seabed communities47, and changes to marine organism foraging patterns as offshore and floating OTEC infrastructure might act as fish aggregating devices that attract marine fauna potentially creating artificial reef communities47,48,50,51. Deep ocean organisms are highly susceptible to temperature, pH, and salinity changes47. Therefore, it is important to conduct field assessments at potential OTEC facility sites to establish baselines for comparison with induced changes and improve spatial planning of cumulative environmental effects47,48.