Deep-sea mining on the NCS: context and emerging technologies

In October 2022, the Norwegian Ministry of Petroleum and Energy published its Impact Assessment for exploring and extracting deep-sea minerals on the NCS (Ministry of Petroleum and Energy, 2022b). The hearing document and ministerial press release highlight that deep-sea minerals may become pivotal in transitioning towards a low-emission society and an important emerging industry for Norway. The Norwegian Ministry of Petroleum and Energy further states that more knowledge and data about the deep-sea environment and potential mineral resources are required for responsible resource management. The ministry suggests that opening the NCS for commercial mineral exploration is pivotal for acquiring more and better data—as the current state of exploration is limited. The ministry emphasizes that an opening for commercial actors commencing with mineral exploration will significantly increase the data retrieval and knowledge generation across disciplines studying the deep-sea in the relevant region. The Norwegian parliament is scheduled to vote over the opening of the NCS for mineral exploration and extraction in the spring parliamentary session of 2023. Upon a potential opening, commercial entities may apply for exploration licenses. The Norwegian petroleum directorate will administer the licensing process (Ministry of Petroleum and Energy, 2022c, 2022a).

Bang and Trellevik (2022) present a comprehensive stochastic simulation model and analysis of the emergence of a Norwegian deep-sea mining industry targeting SMS. They indicate that the cost of exploration will strongly affect the overall profitability of Norway’s emerging deep-sea mineral industry. Furthermore, their findings suggest that net-present value is strongly affected by discounting—as the cost of exploration is substantial and accumulated at a much earlier stage than the income for extracted minerals.

This article explores how three different but established techno-operational concepts for seabed and sub-seabed surveys and exploration may affect the aggregated commercial performance of the emerging SMS industry. Specifically, this study simulates how evolving technologies may affect the cost efficiency of SMS mineral exploration, thereby altering the effect of discounting on the net-present value of the aggregated SMS industry on the NCS. This study applies the model developed by Bang and Trellevik but further develops and applies this model framework to test the effect of techno-operational concepts. The concepts are abstracted as alternative parameterization of structural elements already implemented in the model. The parameters are tested separately and in combination, and the sensitivity to the parameterization is analyzed. This provides a methodical and formal approach to disentangle the effect of different techno-operational pathways in the complex and uncertain future of SMS exploration and extraction; as such, this study provides techno-operational policy advice for innovation and development investment in the emerging SMS industry.

The concepts explored in this study are unmanned surface vessels (USVs), fleet or swarm operation of autonomous underwater vehicles (AUVs), and enhanced remote geophysical methodology for assessing mineralization on the seabed. These techno-operational concepts are all recognized as focus areas for innovation and development, currently pursued by the subsea-service and mineral exploration industry (Argeo 2022; ECA Group 2022; Fugro 2022; Konberg Maritime 2022; Malehmir et al. 2012; Ocean Infinity 2022; Sahoo et al. 2013, 2019; Yu et al., 2019).

The contribution of this study is applied insight for policymakers, industry, and research communities involved in deep-sea minerals on the Norwegian continental shelf and beyond. This study does not consider how the projected emergence of DSM on the NCS would affect the economies of local communities. The study assumes a high level of aggregation and is limited to exploring aggregated effects of innovation and the proliferation of specific emerging technologies at the projected industry level. Through stochastic simulation and sensitivity analysis, this study facilitates clear thinking and qualified decision-making in a highly uncertain domain, aligning with Zeckhauser’s thoughts on “investing in the unknown and unknowable” (Zeckhauser, 2010). As such, this study also contributes a methodological framework for clear thinking and methodic assessment of emerging technologies of unknown or uncertain effect in industrial domains where a baseline of current technological efficacy may be assessed.

This study suggests that developing a remote geophysical methodology for assessing mineral deposits will significantly improve the economic outlook of deep-sea massive sulfide seamount deposits. USV technology, in combination with geophysical methods, is also a profitable endeavor. At the same time, operating swarms of AUVs during high-resolution surveys may reduce the net present value of SMS exploration and extraction and henceforth be counterproductive.

Methods

This study builds on a published model and implements emerging techno-operational concepts in an established model framework. Information about emerging technologies is drawn from reviewing academic literature, technical and operational information provided by industry stakeholders, and by way of qualitative research.

The qualitative research includes the author’s attendance and participation at 11 academic, technical, and industry conferences addressing deep-sea minerals and subsea exploration technology. Several stakeholders have presented techno-operational concepts. Qualitative data is also elicited and qualified through 17 semi-structured, disconfirmatory interviews with diverse stakeholders within the marine minerals and subsea exploration industry and academia.

The academic literature on the three different technologies to be tested is plentiful at engineering research's technical and micro levels. However, the opposite is true for techno-economic analysis at an aggregated industry level; this relates to the novelty inherent in any emerging technology—and the time required for academia to provide empirical observation and evidence of techno-operational performance.

Therefore, the structural abstraction and parameterization of the techno-operational concepts are synthesized through triangulation between qualitative data, academic literature, and data provided openly by commercial stakeholders. In summary, this multi-faceted approach has rendered a well-defined techno-operational understanding of the concepts and their applicability as model abstractions, as well as a range of parameters for cost and expected efficiency of the emerging concepts that enable simulation and policy analysis. It is stressed that these concepts are still under development and far from being supplied at scale—the policy parameters must be considered approximate and uncertain. In acknowledgment of uncertainty—the study assumes an exploratory and conceptual approach. In order to mitigate uncertainty and shed light on the cost efficiency range potential of the different techno-operational concepts under different development trajectories, the study employs sensitivity analysis.

The base model by Bang and Trellevik presents a baseline result where the SMS industry may or may not prove profitable in terms of net-present value depending on ore-grade and investment policy. The ore-grade scenarios are 3, 4, and 5% concentrations of a mineral mix of copper, zinc, and cobalt. The investment policy “Wait and See” requires considerable mineral resources to be confirmed before investment in extraction technology is executed. In contrast, the “Anticipatory” policy commences with investment at an earlier stage and is less risk-averse. The model is simulated stochastically, and results are averaged over 1000 Monte Carlo simulations for each policy configuration. The baseline suggests that with a 3% ore grade and a “Wait and See” investment policy, the NPV is negative 980 million USD, while with a 5% ore grade and “Anticipatory” investment policy, the aggregated industry generates a 2.53 billion USD NPV over the simulated time-horizon. All scenarios with an ore grade above 4% yield positive NPV in the baseline results. The study concludes that discounting and the long time between accruing cost of exploration and retrieving revenue from extracted minerals is the major challenge for the profitability of the industry as, i.e., the net non-discounted value of the 3% “Wait and See” scenario generates a profit of 10.85 billion USD while the NPV value is negative. The “Results” section discusses the baseline results in further detail.

Simulation results presented in this study are generated with the same stochastic parameters, numerical integration method, and Monte Carlo parameters as the simulations presented by Bang & Trellevik. This allows for comparative policy analysis with the baseline results presented and, as such, a relative quantitative framework for understanding the plausible policy impact of the different techno-operational concept scenarios.

As for any study of the future—this study offers little certainty and lays no claim to accurate prediction. No commercial-scale deep-sea SMS industry is currently established anywhere (Bang & Trellevik, 2022; Kaluza et al., 2018). Nautilus Mining Ltd came close to extracting minerals from the Solwara 1 prospect. However, the company filed for bankruptcy in August 2019—and no other initiatives have to date come as close to the commencement of commercial DSM (Gross, 2022). Furthermore, while the techno-operational concepts studied here are well-described and implemented at varying maturity levels, neither of these concepts has been tested at scale for performance in an aggregated deep-sea mineral context. This study, therefore, with the perspective that “structure generates behavior,” elicits model structure and parametrization from the established techno-operational concepts and simulates how these may drive behavior and performance in the context of SMS exploration on the NCS (Forrester, 1980; Kwakkel & Pruyt, 2015; Lane, 2000; Lane & Oliva, 1998). Through simulation and sensitivity analysis, meaningful insight and clear thinking on possible future behavior may be obtained as the concepts provide a structural foundation for synthetic analysis. The technological concepts discussed here are all incremental innovations anticipated to materialize for some time. As such, the scope of this study lends itself well to the thesis that “…the future is embedded in the past; it is the projection of the past through the present” (Poli, 2010). This, however, implicitly infers complexity and significant uncertainty. Dynamic simulation is commonly used to analyze complex and uncertain problems, developing over time, and is, therefore, a valuable approach for developing insight in such domains (Pruyt, 2007).

Model structure

The model encompasses five sectors (Fig. 1). (1) “Exploration process” tracks the seabed exploration for minerals from unexplored areas through 3 levels of declining geographic area and increasing levels of data resolution confirming or disconfirming mineral deposits. The final step of the exploration process aggregates area with positive finds going through environmental impact assessments. Areas deemed without commercially viable mineral resources are aggregated in a stock of the discarded area. The exploration process and the area flow through this process are governed by the sector (2) “Exploration technology.”

Fig. 1
figure 1

Simplified high-level model overview (Bang & Trellevik, 2022)

Full size image

Exploration technology includes the application of four different vessel configurations where regional surveys are executed with relatively small and low-cost vessels, performing seafloor surveys from hull-mounted or towed acoustic and magnetic sensors. These vessels cover large areas of seabed at relatively low cost—and at relatively low data resolution. Areas deemed attractive by exploration companies are then surveyed in greater detail. Finally, high-resolution surveys are executed from relatively large vessels where autonomous or remotely operated vehicles are operated close to the seabed. These operations cost more, as they involve advanced subsea equipment and large ships with considerable crew onboard and onshore support.

The same category of ships is applied to offshore platforms during coring operations and environmental impact assessments. Coring involves drilling into the seabed and retrieving geological core samples of the seabed. This process is tedious, costly, and covers a limited area per time unit. Still, it does provide explorations with high certainty, ground-truth, and data on geological composition and mineralization. Environmental impact assessment involves documenting and assessing the possible environmental impact of mining in areas with confirmed mineral deposits. The model prioritizes the use of these vessels first for environmental impact assessment, second for coring, and third for high-resolution surveys—to pass confirmed mineral deposits through to mining activities.

(3) “Mining process” is the sector in the model where ore is extracted from the seabed, and the pace and magnitude of this activity are governed by the (4) “Mining technology” sector. This model sector includes all logistics, vessels, and subsea equipment involved in bringing ore from the seabed through the water column to the deck and from there to the shore. The model does not include onshore processing or refinement of minerals, which would add considerable complexity to the model and the analysis. This study's scope is to explore the projected efficacy of subsea exploration technology—including onshore processing will not add to the analysis. The study therefore considers ore-value “on deck” rather than value of the refined mineral value in the global commodities market.

The (5) “Financial accounting” sector tracks spending and income for all technology sectors in the model. This sector is essential in that it governs investment in equipment—the model is simulated with two different spending policies where a “Wait and see” policy will have substantial confirmed resources available before investing in more technology. In contrast, the “Anticipatory” policy will show more risk-seeking behavior where investments are made with less confirmed mineral ore on the seabed.

Simulation

Throughout the exploration process, the area considered interesting for further evaluation or impact assessment is governed by four stochastic parameters. This stochasticity is the reason for running 1000 Monte Carlo simulations for every policy scenario and providing the average of this simulation as the result—as stochasticity will generate a distribution of varying results, reflecting the uncertainty related to the distribution of minerals on the seabed.

The model is simulated with three different average mineral mix ore-grade scenarios. These scenarios are simulated to accommodate the uncertainty of what average ore-grades will be proven for SMS deposits on the NCS. The three ore-grade scenarios are the low scenario of 3%, the medium scenario of 4%, and the high scenario of 5%. The mineral mix constitutes copper, zinc, and cobalt—the percentage indicated is the total percentage containing mineral mix in the entire ore body. The mineral mix includes assumptions of 77.8% Cu, 16.7% Zn, and 5.6% Co. It should further be noted that SMS deposits on the NCS have been demonstrated to include cobalt—which is an anomaly when considering SMS deposits studied in other regions (Bang & Trellevik, 2022; Pedersen & Bjerkgård, 2016). The model employs a price matrix based on historical commodity prices for the different commercial minerals in the mineral mix, embedded in the “FINANCIAL ACCOUNTING” sector of the model.

Policies

Three techno-operational concepts are implemented in the model and examined. First, the three concepts identified as Policies A, B, and C are introduced to the model by changing parameters embedded in the model structure in the “Exploration technology,” “Exploration process,” and “Financial accounting” model sectors. Then, the three policies are further tested in all possible combinations.

Policy A: Unmanned surface vessels (USVs)

are ships or crafts able to operate without any personnel onboard. USVs can be either remotely operated via data-link, pre-programmed, or autonomously. USVs can carry any number of sensors or instruments—and be used for a wide array of purposes (Rumson, 2021; Zhang et al., 2019). In the context of this study, USVs are imagined as a replacement for survey vessels employed in regional surveys of deep-sea mineral prospecting. In this context, USVs would be equipped with multi-beam echo sounders, side-scan sonars, and other sensors. The significant impact of replacing a survey vessel with a USV is the removal of all personnel offshore—and the logistics involved in maintaining crews offshore. Removing all offshore crews further affords significant risk mitigation and allows for elongated operating seasons in arctic waters (Rumson, 2021). The USV is henceforth significantly less costly to build and operate, and it can operate for a more extended period of the year and thereby be more productive. There are several companies involved in developing USV technology for this purpose, including Konberg Maritime (Kongsberg Maritime, 2022), ECA Group (ECA Group, 2022), Fugro (Fugro, 2022), Sea-Kit (Sea-Kit, 2022) and Ocean Infinity/Armada (Ocean Infinity, 2022). These companies unanimously claim increased operating windows and fractional cost relative to crewed offshore operations—and there is little reason to doubt that this is the case as remote operations of smaller unmanned crafts and vessels must be more cost-effective than the conventional alternative. It should be noted that also USVs require monitoring, maintenance, and repair—which may amount to considerable cost as they, in this scenario, will be operated far from people and workshops; these costs are included in the aggregated parameterization terms of efficiency and cost. The USV concept is identified in this study as “Policy A”; please refer to Table 1 for policy implementation in simulation runs.

Table 1 Policy parameters
Full size table

Policy B: Fleet operation of autonomous underwater vehicles (FAUV)

is a technological and operational concept where multiple AUVs are launched and operated from a single crewed surface vessel. Several companies are advancing this concept, most notably Ocean Infinity and Argeo (Argeo, 2022; Ocean Infinity, 2022). This concept significantly increases the seabed survey footprint of a single surface vessel, with marginal increases in vessel crews. The increased efficiency is made possible by advanced robotics and autonomous technology and has been under development for over two decades (Sousa et al., 1997; Wang et al., 2021). The Fleet AUV technology concept has yet to become commercially widespread, but several projects have been completed. Notably, in 2018 Ocean Infinity conducted highly effective seabed surveys employing as many as eight AUVs simultaneously in the search for the lost MH370 flight (Ocean Infinity, 2018, 2022; Xu and Jiang, 2021). This study identifies this concept as “Policy B”; please refer to Table 1 for policy implementation in simulation runs.

Policy C: Geophysical sampling

In Bang & Trellevik’s model, coring is a significant cost driver on the exploration side (2022). Coring is a common geological sampling technique on land involving drilling into the ground and retrieving continuous rock samples to identify and classify orebodies. A dense matrix of cores is required to ascertain the ore-grade throughout a deposition. While subsea coring has been executed for many decades, it remains complicated and costly, and there are only a limited number of successful coring campaigns targeted at SMS deposits (Holtedahl, 1959; Murton et al., 2019; Spagnoli et al., 2016). New technological concepts and applications may reduce the reliance on extensive subsea coring for identifying and evaluating deep-sea mineral deposits by supplementing or filling in gaps between physical cores with geophysical data that can be correlated to physical samples. These technologies include, but may not be limited to, modified seismic applications, electro-magnetic sampling, self-potential anomaly measurements, atomic dielectric resonance spectroscopy, and combinations of these technologies (Almqvist and Mainprice, 2017; Biswas 2018; Malehmir et al. 2012; Stove et al. 2013, 2009). Although there is little in the way of proven subsea application of such technologies in this category, this study assumes that these technologies will be adapted, mature, and become available as they are both theoretically possible and under development. These technologies are likely, and expected, to vastly increase the assessment area during a coring campaign—as geophysical sampling will be used to augment physical coring data by providing calibrated remotely measured or sampled data points for interpolation between physical core samples. Geophysical sampling will be executed from the same vessel and parallel with ongoing coring operations. Thus, geophysical sampling will enable surveys of a considerably larger area to be assessed for prospectivity within the same operational time frame as what is obtainable with conventional coring operations alone. To understand the space for innovation, this study includes geophysical sampling as a concept for exploratory model analysis. This study identifies this concept as “Policy C.” Please refer to Table 1 for policy implementation in simulation runs.

Policies A, B, and C are implemented in the following way.

Policy A is modifying the yearly survey speed as the USVs are assumed to be able to operate for a more extended season than conventional survey ships. As a result, the build cost (CAPEX) and operational cost are also significantly reduced.

Policy B is modifying the yearly rate for high-resolution surveys as there is a cost impact for mobilizing and operating 8 AUVs. In addition, the survey footprint, or swath, is also increased by a multiple of eight, aligned with the operational configuration of Ocean Infinity’s MH360 search operation (Ocean Infinity, 2018). Thus, Policy B still utilizes vessels from the vessel pool included in the model at a different cost.

Policy C is implemented by increasing the operational cost of the coring ship—as there will be a cost impact of mobilizing geophysical sampling in addition to coring equipment. The area covered by coring is increased tenfold. This increase in efficiency is aligned with expectations expressed by interviewed stakeholders engaged with such techno-operational concepts. As mentioned above, Geophysical sampling is expected to operate simultaneously with conventional coring. Geophysical samples will be correlated and calibrated to physical cores retrieved and qualify a larger area with interpolated prospective analysis in between cores based on geophysical sampling samples. This will significantly increase the footprint covered by a coring campaign. Like Policy B, Policy C still utilizes vessels from the vessel pool included in the model, yet at a different cost.

As described above, the model is simulated with the same stochastic parameters, ore-grade scenarios, and investment policies as in the original model. In addition, each new policy is tested separately and in all possible combinations on top of the original model configurations. This renders the following simulation matrix for innovation policy baseline results (Table 2).

Table 2 Innovation policy baseline simulation matrix—42 different simulation configurations run across 1000 Monte Carlo simulations each
Full size table

Results

The following section reports the simulation results compared to the baseline results provided by Bang and Trellevik. It should be noted that these results represent prognostic simulated system behavior and not empirical data. The following data is included in all tables in this section. (1) Exploration capex, summarizing all capital investment expenditure for the exploration process. (2) Exploration opex, summarizing all operational costs associated with exploration. The policies will directly affect exploration capex and exploration opex. (3) Mining capex summarizes capital investment expenditures for mining units and associated logistical elements. The policies will indirectly affect this cost via efficiency in the exploration process. (4) Mining opex summarizes operational expenditure for the mining activities. Mining opex will only be indirectly affected by the introduced policies. (5) Total extraction is reported in million tons of mineral mix. (6) Total revenue is the aggregated revenue throughout the simulation period. (7) Net non-discounted value reports the aggregated revenue less the aggregated cost. (8) Net present value (NPV) is the profit adjusted for discounting through the simulation horizon. “Baseline results” refer to the results generated by the original model (Table 3).

Table 3 Overview of baseline simulation results. Average values across 1000 Monte Carlo runs (Bang & Trellevik, 2022)
Full size table

The baseline results from the original model indicate that it is not given that a deep-sea SMS exploration and extraction industry on the NCS will be profitable. NPV varies from a negative 980 million USD low to a profit of 2.53 billion USD throughout the simulation horizon. The results also indicate that the ore-grade of the extracted minerals is a significant driver for the industry’s profitability; however—of significance is also the investment policy pursued. The “Anticipatory” policy generates a significantly higher profit in the high (5%) ore-grade mineral mix than does the “Wait and See” policy, showing an NPV improvement of 1.2 billion USD with the less risk-averse policy. The results further show a significant difference between non-discounted and net-present values, indicating that discounting is a significant challenge for this industry’s prospective profitability. This arises from a long temporal horizon between cost propagating throughout the exploration process, and initial investment in the mining process before revenue is generated by bringing mineral commodities to shore. In short, this suggests that reducing the time horizon between initial exploration and minerals reaching markets is essential, as is the cost of exploration. Policies A, B, and C and their possible combinations explore this assumed room for improvement related to emerging techno-operational concepts (Table 4).

Table 4 Overview of simulation results with Policy A. Average values across 1000 Monte Carlo runs with baseline results in brackets
Full size table

Policy A’s results indicate that the net present value is marginally increased across ore-grades and investment policies. However, the exploration capex and exploration opex are only marginally reduced. The regional survey in the original model constitutes a relatively small portion of the total exploration process cost. Therefore, this policy cannot generate a positive NPV for the lowest ore-grade scenario (Table 5).

Table 5 Overview of simulation results with Policy B. Average values across 1000 Monte Carlo runs with baseline results in brackets
Full size table

Policy B is increasing the exploration capex across all scenarios and investment policies—but is simultaneously reducing the exploration opex. This indicates that the slightly more costly hi-resolution survey configuration, including eight AUVs per vessel, is more efficient in operation but more costly to procure. The same effect can be identified for mining capex and opex; Policy B is driving up investment costs but is slightly reducing or neutral on operational costs during the mining operations. As a result, net present value is only marginally affected by Policy B and has a neutral or slightly positive effect. Interestingly, NPV with Policy B is lower than what is evident with Policy A (Table 6).

Table 6 Overview of simulation results with Policy C. Average values across 1000 Monte Carlo runs with baseline results in brackets
Full size table

Policy C is vastly outperforming Policies A and B. Geophysical sampling technology is generating positive NPV within the lowest ore-grade scenario under the “Wait and See” regime, turning a profit of 1.54 billion USD. The Baseline scenario here is negative 980 million USD. In the 5% ore-grade and “Anticipatory” investment strategy, Policy C generates a net present value more than 2.5 times the baseline results. Exploration Capex and Exploration Opex are both significantly lower than what is found in Policies A and B; this is a significant factor in the overall performance of Policy B (Table 7).

Table 7 Overview of simulation results with Policy A + B. Average values across 1000 Monte Carlo runs with baseline results in brackets
Full size table

Policy A + B generated improved NPV results for both simulations in the 3% ore-grade scenario. In the 4 and 5% scenarios, apart from the 4% ore-grade and “Anticipatory” scenario, Policy A alone generates a higher Net Present Value than the combined A and B policies. The combined policy outperforms Policy B in terms of NPV across all simulations (Table 8).

Table 8 Overview of simulation results with policy A + C. Average values across 1000 Monte Carlo runs with baseline results in brackets
Full size table

The combined Policy A + B generates significantly lower Exploration Capex and Exploration Opex than Policies A or B or the combination of the two policies. Policy A +B generates higher NPV than Policy C. Henceforth, combining USVs with geophysical sampling is beneficial (Table 9).

Table 9 Overview of simulation results with Policy B + C. Average values across 1000 Monte Carlo runs with baseline results in brackets
Full size table

The combined Policy B + C generates a similar exploration capex to Policy A + C, yet a slightly lower exploration opex. Policy B + C produces an exploration opex of 0.88 billion USD across all scenarios, while Policy A + C shows an exploration opex of 0.95 billion USD across scenarios. Interestingly Policy B + C renders slightly higher mining capex and mining opex than Policy A + C Across all scenarios. Policy B + C produces a lower NPV than Policy A + B, most notably in the 5% ore-grade and “Anticipatory” scenario where Policy B + C generates an NPV of 5.70 billion USD, 760 million USD lower than Policy A + C (Table 10).

Table 10 Overview of simulation results with Policy A + B + C. Average values across 1000 Monte Carlo runs with baseline results in brackets
Full size table

Policy A + B + C Generates a similar exploration capex, but slightly lower exploration opex than Policy B + C. Exploration opex for Policy A +B + C is also lower than what is seen in Policy A + C and Policy C. Policy A + B + C generates a slightly higher NPV then Policy B + C, but a lower NPV then Policy C and Policy A + C.

In summary, the baseline policy simulations demonstrate that the most significant techno-operational concept in reducing exploration costs is remote sensing geophysical data collection. Any combination of policies, including geophysical sampling, will by far outperform both baseline results and the techno-operational policy scenarios where the exploration process is confided by conventional coring. The data further shows that introducing Policy B, FAUVs for Hi-Resolution surveys, does not improve the NPV from the baseline scenarios—except for the 4% ore-grade and “Anticipatory” configuration. Policy A demonstrates improved NPV, but not substantially so. The combined Policy A + B produces lower NPV than the baseline scenarios. The best-performing policy is the combined Policy A + C with NPV of 6.46 billion USD in the 5% ore-grade and “Anticipatory” scenario. This is a substantial increase in NPV relative to the baseline scenario of 2.53 billion USD.

Sensitivity analysis

The results of the baseline policy simulation nominate Policy C, or the introduction of geophysical sampling, as the most significant driver for increased profitability within deep-sea SMS mining. This techno-operational concept is not mature, and the parameterization of the policy is, therefore, subject to deep uncertainty. Sensitivity analysis of Policy C, in isolation from other policies, is therefore of value. To reduce complexity and, as such, provide more clarity on results, sensitivity is only tested for 3% ore-grade in the “Wait and See” configuration. This configuration does not generate a positive NPV in the baseline results, yet it does so with the same configuration under Policy C. It is, therefore, interesting to identify a lower limit of efficiency for the policy’s ability to turn a profit and, as such, de-risk the deep-sea SMS mining industry at large. Two parameters govern Policy C: the area concluded per coring campaign and the annual cost of coring operations. Both parameters are tested in isolation.

Table 11 indicates that by increasing the coring efficiency by 82.5%, deep-sea SMS mining will be marginally profitable, with an NPV of 1 million USD, in the lowest ore-grade and passive investment regime. In the original model, a coring campaign is, on average, estimated to cover an area of 0.2125 km2. The sensitivity results of Policy B suggest that by increasing the area to 0.387 km2 the industry would, on an aggregated level, be profitable at a 3% average ore-grade and risk-averse investment regime. The sensitivity analysis of Policy B further indicates that the policy is not very sensitive to the annual cost of the operation; at 9.5 times the annual cost, the industry generates a negative NPV of 2 million USD—at nine times the annual cost, it generates a profit of 7 million USD. The ability to effectively confirm or disconfirming mineral prospects as commercially viable appears to be of greater importance than the annual cost of these operations.

Table 11 Overview of sensitivity results with Policy C at 3% average ore-grade and “Wait and See” setting, average values across 1000 Monte Carlo runs with Policy C baseline results in brackets
Full size table

Policy B, fleet-operated AUVs for high-resolution surveys, is the techno-operational policy of the poorest performance. In order to establish limit values for this policy to yield positive results in the lowest ore-grade and passive investment regime is therefore interesting. Policy B is governed by the yearly cost of high-resolution survey vessels and by the swath, or footprint on the seabed, obtained by the AUVs. To test the sensitivity of Policy B, these two parameters are tested in isolation (Table 12).

Table 12 Overview of sensitivity results with Policy B at 3% average ore-grade and “Wait and See” setting. Average values across 1000 Monte Carlo runs with Policy B baseline results in brackets
Full size table

The model is not sensitive to high-resolution survey swath. Even with 40 times the swath, the NPV is largely unaffected, as is mining opex, total extraction, and total revenue. exploration capex is marginally higher, and exploration opex is marginally lower in this extreme configuration, as is net non-discounted value. The same tendency is evident also for significant changes in the yearly rate of high-resolution surveys. The model behavior is insensitive to vastly reduced rates of high-resolution surveys. The efficiency of high-resolution surveys does not affect model behavior to any considerable extent, nor does the cost of this step in the exploration process.

Discussion and policy analysis

This study indicates that their innovation and development in the exploration processes for deep-sea SMS deposits on the Norwegian Continental Shelf may significantly reduce the commercial risk and boost profitability. It should be noted that these simulation runs include uncertain parameters, certainly pertaining to the efficacity of the emerging technologies being projected. By employing broad arrays of sensitivity analysis on the stochastic model—the uncertainty is included and explored in the analysis. The simulation results identify that the successful development of geophysical sampling tecno-operational concepts will significantly impact the profitability of this emerging industry on an aggregated level. The Policy Simulation Baseline results indicate that at the high (5%) ore-grade, under the “Anticipatory” investment scenario—a geophysical sampling technology applied to enhance the footprint of conventional coring may produce about 250% improvement of NPV. Sensitivity analysis furthermore shows that scaling up the area confirmed or disconfirmed by geophysical sampling enhanced coring by 82.5% will render a positive NPV also in the low ore-grade (3%) and risk-averse (“Wait and See”) policy scenario. The sensitivity analysis demonstrates that the increased cost of geophysical sampling is less critical. This is a significant finding for a budding industry—as it suggests a distinct focus on research and development and indicates that considerable budgets for developing and scaling up such techno-operational concepts may be beneficial as the resulting operational cost efficiency may be significant.

The results also demonstrate that developing and scaling up fleet AUV techno-operational concepts may be of little value in terms of the aggregated profitability of the industry over time. The sensitivity analysis furthermore indicates that neither the cost nor the actual efficiency of FAUV is of considerable importance to the deep-sea SMS exploration and extraction industry on the NCS. This is also an interesting find, as this is a techno-operational concept currently receiving much attention and investment in the subsea industry (Argeo, 2022; Ocean Infinity, 2022). There may be several reasons for this, but it is likely related to the ship utilization—and the size of areas expected to require high-resolution surveys. By the time the exploration process moves into high-resolution surveys, earlier survey initiatives dramatically reduce the area of interest. Therefore, increasing the footprint on the seabed by adding several AUVs may not offer much benefit as the conventional footprint in the original model is already sufficient—and, indeed, relatively efficient as it is.

The introduction of Unmanned Surface Vessels for regional surveys is likely to positively affect aggregated NPV, although the impact is mainly marginal. This can be directly related to regional surveys already being relatively cost-efficient in comparison with the other stages of the exploration process. It should, however, be noted—that removing conventional regional survey vessels is likely to have a positive impact on risk to personnel and emissions to the environment while operating in the Arctic, as the USVs will not be crewed—and will consume considerably less fuel (Rumson, 2021). On the other hand, removing crews will hurt the employment rates in the subsea industry. This study considers neither of these effects as they lie beyond the model boundaries.

The policy combination that generates the highest NPV is Policy A + C. This policy combination generates a net present value of 6.46 billion USD in the high ore-grade, “Anticipatory” scenario and an NPV of 1. 61 billion USD in the low ore-grade, “Wait and Wait and See” scenario. These are considerable improvements to the 2.53 billion USD and − 0.98 billion USD results produced by the original model and, as such, indicate a considerable innovation space for these two techno-operational concepts within the realm of SMS mining on the NCS.

It is simultaneously interesting to note that Policy A + B + C generates an NPV of 760 million USD below Policy A + C in the high ore-grade (5%) and “Anticipatory” scenario—while it demonstrates a 40 million lower NPV in the low ore-grade (3%) and risk-averse scenario (“Wait and See”). Similarly, overall, Policy B + C performs poorer in terms of NPV than Policy C alone. Furthermore, Policy A + B renders lower NPV across all scenarios than Policy A alone. It appears that focusing on developing and scaling up fleet-operated AUVs is not only moot but counterproductive for the aggregated deep-sea SMS industry on the Norwegian continental shelf. This is also a significant finding—as it informs stakeholders developing new techno-operational concepts within this emerging industrial segment in which concepts to allocate a low priority or to avoid altogether.

Conclusion

This study has explored the possible impact of three emerging techno-operational concepts within the subsea and sub-seabed survey as these apply to the nascent SMS industry on the Norwegian continental shelf. These concepts are unmanned surface vessels for regional surveys, fleet operation of AUVs for high-resolution surveys, and geophysical sampling in conjunction with geological core samples for resource evaluation of mineral deposits.

Significant possible gains are available in the techno-operational innovation space within SMS exploration on the Norwegian continental shelf. Most predominantly stand for developing a geophysical methodology that enhances the area covered and qualified, or disqualified, by geological core sampling. This process's cost is less significant—the cost of the combination of coring and geophysical sampling can be increased by about nine times and still yield profits in a low ore-grade and risk-averse investment scenario. By only increasing the area covered by a coring and geophysical campaign by 82% relative to the baseline scenarios produced by the original model—profits may be generated. This appears amply possible within the offshore industries. Gains are also likely by the introduction of unmanned surface vessels for regional surveys—but these gains are less significant. The most profitable endeavor is developing USVs for regional surveys and geophysical sampling. Simultaneously, the emergence of fleet-operated AUVs is less beneficial to the SMS mining industry on the NCS. In fact—not only does this concept appear pointless, but it is also counterproductive as it may reduce the aggregated net present value of the industry. Nevertheless, these findings are of value to an emerging industry currently placing its bets on techno-operational concepts and gearing up for a possible opening of the Norwegian continental shelf (Energi24.no, 2021; Ministry of Petroleum and Energy, 2022a).